Skip to main content
  • Meeting abstracts
  • Open access
  • Published:

42nd International Symposium on Intensive Care & Emergency Medicine


Effective FiO2 delivered by a new frugal CPAP system with low oxygen needs: from bench to clinical observations

E. De Beaufort1, G. Carteaux2, F. Morin3, F. Beloncle4, A. Lesimple5, D. Savary3, A. Mercat4, L. Brochard6, J. C. Richard4, A. Mekontso-Dessap2

1Université Paris-Est Créteil, Laboratoire CARMAS, IMRB, Créteil, France, 2Assistance Publique-Hôpitaux de Paris, CHU Henri Mondor-Albert Chenevier, Service de Médecine Intensive Réanimation, Créteil, France, 3Faculté de Santé, Centre Hospitalier Universitaire d’Angers, Université d’Angers, Département de Médecine d’Urgence, Angers, France, 4Faculté de Santé, Centre Hospitalier Universitaire d’Angers, Vent’ Lab, Université d’Angers, Département de Médecine Intensive-Réanimation et Médecine Hyperbare, Angers, France, 5Université d’Angers, CNRS, INSERM 1083, MITOVASC, Angers, France, 6St. Michael’s Hospital, Interdepartmental Division of Critical Care Medicine, Toronto, Canada

Critical Care 2023, 27(S1): P001

Introduction: In the context of a pandemic with a massive influx of hypoxemic patients, the high oxygen consumption usually required to achieve optimal inspired fraction of oxygen (FiO2) with non-invasive respiratory supports may jeopardize healthcare organization and oxygen delivery capabilities in hospitals.

Methods: A new CPAP device aimed at optimizing oxygen delivery with minimal oxygen needs and named “Bag-CPAP”, was designed and evaluated:

  1. 1.

    A bench study compared performances of Bag-CPAP and four other non-invasive devices to those of an ICU ventilator. With the objective of delivering a PEEP level between 5 and 10 cmH2O, FiO2 actually delivered at two target ranges (40–60% and 80–100%) and oxygen consumption were measured for two different simulated patients.

  2. 2.

    An observational clinical study evaluating the Bag-CPAP was conducted in France, including 20 adult patients with de novo acute respiratory failure. Actual FiO2 and PEEP as well as oxygen saturation (SpO2), respiratory rate (RR) and dyspnea score were assessed.

Results: 1. Bench study: All systems tested reached the minimal FiO2 target range (40–60%) and four were able to reach the higher target range. The ratio between FiO2 and oxygen consumption was the highest with the Bag-CPAP and the lowest with high-flow oxygen device whatever the FiO2 target.

2. Clinical study: The Bag-CPAP was well tolerated and reached the two FiO2 target ranges, 53% [52–53%] for moderate FiO2 and 94% [93–96%] for high FiO2 with an oxygen flow rate of at least 5 L/min and 15 L/min, respectively (Fig. 1). SpO2 significantly increased and dyspnea significantly decreased within the first hour of treatment (93% [90–97%] vs. 97% [95–98%], p = 0.001 and 4 [2–7] vs. 3 [2–4], p = 0.017, respectively). There was no significant effect on RR.

Conclusions: This study shows that the Bag-CPAP has the highest oxygen saving properties. The clinical observation confirmed bench experiments suggesting good clinical tolerance and the capacity to alleviate respiratory distress.

figure a

Fig. 1 (abstract P001). FiO2 measurements according to oxygen flow rate during Bag-CPAP treatment and FiO2 target range (moderate: green or high: pink)


Influence of intraoperative liberal versus conservative O2 fraction on pulmonary function in adults undergoing spine surgery

SC Meena, K Jain, A Ghimere, R Chauhan, A Luthra, V Kumar

Pgimer Chandigarh, Department of Anesthesia and Intensive Care, Chandugarh, India

Critical Care 2023, 27(S1): P002

Introduction: Spine surgery patients are very prone to development of postoperative pulmonary complications. We aimed to compare the impact of different intraoperative fractional inspiratory concentration of oxygen on postoperative pulmonary outcomes including oxygenation index and lactate values after thoracolumbar spine instrumentation.

Methods: Total 75 American Society of Anesthesiologists I–II adult patients aged 18–60 years who were scheduled to undergo thoraco-lumbar spine instrumentation were randomly allocated into group A: (conservative oxygen) FiO2 = 0.30 (30% oxygen with air) and group B: (Liberal Oxygen) FiO2 = 0.80 (80% oxygen with air). We aimed to compare the changes in PaO2/FiO2 ratio, lactate values, C-reactive protein, interleukin-6, PONV and SSI at the different study time intervals.

Results: At the end of surgery after extubation, the mean (SD) of PaO2/FiO2 ratio in 0.3 FiO2 and 0.8 FiO2 groups were 424.69 (53.69) and 315.57 (48.62) respectively. The values of lactate in both the groups increased at the end of surgery from a mean of 0.93 pg/ml in both the groups preoperatively to 1.18 pg/ml and 1.74 pg/ml in 0.3 and 0.8 FiO2 groups postoperatively respectively. No significant changes in the levels of interleukin 6 and CRP were found in our study when compared between baseline values and the values after intervention.

Conclusions: This study demonstrated the superior efficacy of conservative oxygen inspiratory fraction (0.3% FiO2) as compared to liberal FiO2 (0.8% FiO2), in terms of better PaO2/FiO2 ratio and low lactate values in patients undergoing surgery for thoraco-lumbar spine.


Effect of non-invasive respiratory support on interstitial lung disease with acute respiratory failure: a systematic review and meta-analysis

N Sanguanwong1, N Jantarangsi2, J Ngeyvijit3, N Owattanapanich4, V Phoophiboon5

1Chulalongkorn University, Department of Physiology, Bangkok, Thailand, 2Buddhachinaraj hospital, Department of Internal Medicine, Phitsanulok, Thailand, 3Chulalongkorn University, Division of Pulmonary and Critical Care Medicine, Bangkok, Thailand, 4Mahidol University, Division of Trauma Surgery, Bangkok, Thailand, 5Chulalongkorn University, Critical Care Medicine, Bangkok, Thailand

Critical Care 2023, 27(S1): P003

Introduction: The data of non-invasive respiratory support (non-invasive positive pressure ventilation; NIPPV and high flow nasal cannula; HFNC) in interstitial lung diseases (ILDs) with acute respiratory failure (ARF) are scarce [1, 2]. Our study aimed to identify the benefits of non-invasive respiratory support on this patient subgroup.

Methods: MEDLINE, EMBASE and the Cochrane Library searches were conducted from 1946 to June 2022. An additional search of relevant primary literature and review articles was also performed. A random effects model was used to estimate the PF ratio, PaCO2, mortality, intubation rate and hospital length of stay.

Results: Of 380 studies, 9 articles were included in the meta-analysis. The non-invasive respiratory support demonstrated a significant improvement in PF ratio compared to conventional oxygen therapy (COT), the mean difference was 54.58 (95% CI [9.48–9.69]) (Fig. 1). Compared to HFNC, there was a significant increase in PF ratio in NIPPV (mean difference 0.45; 95% CI [0.12–0.79]). There were no mortality and intubation rate benefits when comparing between NIPPV and HFNC, mean difference were 1.1; 95% CI [0.83–1.44] and 1.86; 95% CI [0.42–8.33], respectively. In addition, there was a significant decrease in hospital length of stay in HFNC compared to NIPPV (mean difference 9.27; 95% Cl [1.45–17.1]).

Conclusions: Non-invasive respiratory support might be an alternative modality in ILDs with ARF. NIPPV demonstrated a potential to improve the PF ratio compared to HFNC. There was no evidence to support a benefit of NIPPV and HFNC in terms of mortality and intubation rate.


  1. 1.

    Rochwerg B et al. Intensive Care Med. 2020;46:2226–37.

  2. 2.

    Rochwerg B et al. Eur Respir J. 2017;50:1602426.

figure b

Fig. 1 (abstract P003). Effect of non-invasive respiratory support on PF ratio


Causes of mortality of severe COVID-19 patients receiving high-flow oxygen therapy

Z Göçerler, R Akkurt, E Paksoy Şenol, E Karakoç, B Yelken

Eskişehir Osmangazi University Medical Faculty, Anesthesiology and Reanimation Intensive Care Unit, Eskişehir, Turkey

Critical Care 2023, 27(S1): P004

Introduction: High-flow oxygen therapy (HFOT) is used to provide oxygenation and reduce the need for intubation in severe pneumonia cases caused by SARS-CoV-2 virus. In this study, causes of mortality during hospitalization in the intensive care unit (ICU) in patients receiving HFOT therapy were investigated.

Methods: The data of 215 adult patients, who were admitted to ICU of a university hospital between April 2020 and October 2021, with severe COVID pneumonia and received HFOT were enrolled retrospectively in our study.

Results: Total mortality among patients was 158 (73.4%). The overall mean age was 72 years, 61 in the survivor group and 73 in the mortality group (p < 0.001). It was determined that mortality rates decreased as the length of stay (LoS) in ICU and HFOT duration of the patients increased (p = 0.008 and < 0.001, respectively). The increase in respiratory rate on arrival was found to be significantly associated with increased mortality. Although the goal of HFOT is to improve oxygenation, no significant mortality-related difference was found in terms of pO2, pCO2 and P/F values calculated at the time of admission. The ROC curve was applied to examine the differential effect of Apache-II, SOFA, ROX and Procalcitonin measurements according to the survival. The area under the curve (AUC) and cut-off values were calculated as follows: APACHE-II (63.9%, 5,) SOFA (62.8%, 2), ROX index (66.8%, 4.72), Procalcitonin (65.7%, 0.23) (Fig. 1).

Conclusions: Unlike reports in literature on mortality in ICU, LoS was found to have a decreasing effect on mortality rate [1]. In addition to the well-known scoring systems such as APACHE-II and SOFA, the ROX index, which is used to predict the success of HFOT, has emerged as a predictive value for mortality at admission to the intensive care unit [2].


  1. 1.

    Moitra VK et al. Crit Care Med 2016;44:655–662.

  2. 2.

    Chandel A et al. Respir Care 2021;66:909–919.

figure c

Fig. 1 (abstract P004). ROC curve of APACHE-II, SOFA, ROX and procalcitonin relation with mortality


Patient-ventilator interaction during non invasive ventilation with helmet: a comparison between pressure support ventilation and the new neural pressure support (NPS) software

A Costa1, S Zorzi2, A Rivolta2, G Cammarota3, R Vaschetto4, A Pagni5, DC Davide Colombo5

1Ospedale SS. Trinità Borgomanero, Borgomanero, Italy, 2Eastern Piedmont University, Novara, Italy, 3Perugia University, Anesthesia and Intensive Care, Perugia, Italy, 4Eastern Piedmont University, Anesthesia and Intensive Care, Novara, Italy, 5Ospedale SS. Trinità Borgomanero, Anesthesia and Intensive Care, Borgomanero, Italy

Critical Care 2023, 27(S1): P005

Introduction: The NIV success is highly affected by patient comfort and patient-ventilator synchronism. The helmet is one of the most comfortable interfaces, even if synchronism is lower due to dead space. The use of neurally adjusted ventilatory assist, in which the electrical activity of the diaphragm (EAdi) drives the ventilator, has shown improvement in comfort and synchronism but still some limitation with helmet, due to the long pressurization time. pressure-support ventilation (PSV) is still the most used and diffused assisted mode, due to its simplicity and effectiveness in helmet-NIV. We tested a new ventilation software called neural pressure support (NPS), which merges the rapid pressurization of PSV along with the EAdi trigger, on a population of 24 critical care patients ventilated via helmet-NIV.

Methods: PSV and NPS 30-min trials were started in random order. A physician expert in mechanical ventilation optimized ventilator setting blinded for EAdi and 2 new 30-min trials were started. Tracings were recorded and the last 2 min of each analyzed. As primary endpoint we evaluated TimeSync, defined as the time during which both EAdi and Flow tracings were in the inspiratory phase. Trigger delay and TimeSynch to neural inspiratory time (TS/TN) was calculated.

Results: Mean TimeSync was statistically higher in NPS compared to PSV (0.632 s vs 0.469 s, p < 0.01). Inspiratory trigger delay was lower in NPS as compared to PSV (0.124 s vs 0.278 s, p < 0.01) (Fig. 1). Optimization by the expert appears to have no statistical significance for both TimeSync and Delay (0.56 vs 0.54, p = 0.55; 0.21 s vs 0.18 s, p = 0.12). TS/TN was higher in NPS versus PSV (0.76 vs 0.57, p < 0.01) with no difference after the optimization (0.67 vs 0.66, p = 0.6693).

Conclusions: Based on this preliminary data NPS seems to increase patient-ventilator synchronism during helmet-NIV both improving TimeSynch and shortening trigger delay. These improvements seem to be not affected by the expertise of the physician that set the ventilator.

figure d

Fig. 1 (abstract P005). Representative flow (V), airways pressure (P) and electrical activity of the diaphragm (EAdi) tracings from one patient both in PSV and in NPS. In light transparency is highlighted the time during which both patient and ventilator are in inspiratory phase. In dark transparency is highlighted the trigger and cycling off delay. A longer inspiratory synchrony time is appreciable in NPS versus PSV


Tidal volume measurement during non-invasive respiratory support by helmet continuous-flow CPAP is feasible and accurate in a bench model

A Coppadoro1, G Bellani2, G Foti2

1San Gerardo Hospital, Department of Anesthesia and Intensive Care, Monza, Italy, 2University of Milan-Bicocca, Department of Medicine and Surgery, Monza, Italy

Critical Care 2023, 27(S1): P006

Introduction: We investigated a novel technique designed to measure tidal volume during non-invasive helmet continuous-flow CPAP, a device for non-invasive respiratory support largely used during the recent COVID-19 pandemic to treat acutely ill hypoxic respiratory failure patients.

Methods: An active lung simulator coupled with a helmet CPAP was used to compare measured and reference tidal volumes at PEEP 5, 10 and 15 cmH2O and different levels of distress (pMusc 10, 15 20 and 25 cmH2O; respiratory rate 15, 20, 25 breaths per minutes). Tidal volume measurement was based on helmet outflow-trace analysis. Helmet inflow was increased from 60 to 75 and 90 L/min to match patients’ inspiratory flow; an additional subset of tests was conducted in condition of purposely insufficient inflow (i.e.: high respiratory distress and 60 L/min inflow).

Results: Explored tidal volumes ranged from 250 to 910 mL. The Bland-Altman analysis showed a bias of −3.2 ± 29.3 mL for measured tidal volumes as compared to reference, corresponding to an average relative error of −1 ± 4.4% (see Fig. 1). At univariate analyses, tidal volume underestimation correlated with respiratory rate (rho = .411, p = .004) but not with peak inspiratory flow, distress, or PEEP. When the helmet inflow was purposely maintained insufficient as compared to the simulated inspiratory flow, the Bland-Altman analysis showed a significant tidal volume underestimation (bias −93.3 ± 83.9 mL), corresponding to an error of −14.8 ± 6.3%.

Conclusions: We showed that tidal volume measurement is feasible and accurate in a model of bench continuous-flow helmet CPAP therapy by the analysis of the outflow signal, provided that helmet inflow is maintained adequate to match patient’s inspiratory efforts. Insufficient inflow resulted in tidal volume underestimation.

figure e

Fig. 1 (abstract P006). Tidal volume measurement error was lower than ± 10% during helmet CPAP therapy, provided that the inflow was maintained adequate to patient’s respiratory distress (filled symbols). A standard helmet inflow (60 L/min), a value lower than peak inspiratory flow in case of high respiratory distress (pMusc 20 and 25 cmH2O, empty circles), lead to underestimation of tidal volumes; the higher the peak inspiratory flow, the higher the underestimation error (r = − 0.905 p < .001)


Non-invasive ventilation in ≥ 70 year old COVID-19 patients in a level two unit: retrospective cohort analysis, comparison with COVIP substudy and defense of a multifaceted work of breathing control driven strategy

J Rua, P Costa, AR Nogueira, JM Mateus

Centro Hospitalar Universitário de Coimbra, Serviço de Medicina Intensiva, Coimbra, Portugal

Critical Care 2023, 27(S1): P007

Introduction: The debate about optimal management of patients with COVID-19 ARDS remains, including medical treatment, ventilatory strategies, awake proning and others. COVIP is a multicentric observational study with over 3000 patients under NIV. A substudy by Polok and al. evaluated patients (PTS) ≥ 70 years old. At our intermediate care unit (IU) we used a strategy of high dose corticosteroid started when the work of breathing (WOB) increased, prolonged awake prone positioning (> 12 h) and high CPAP ventilatory strategy. We describe our cohort of ≥ 70 years old NIV PTS and compare it to COVIP substudy results.

Methods: Descriptive retrospective study. Data were collected from electronic medical records of 95 COVID-19 PTS aged 70 years old or above under NIV at the IU between September/20 and March/21. Categorical data are presented as frequency (percentage) and were compared using χ2-test. Continuous variables were compared using Mann-Whitney U test. Cohort results were compared with those from Polok et al. COVIP substudy (COVIPss).

Results: 95 of PTS were submitted to NIV. Median age was 76 years and 49.5% were male, versus 75.7 and 71.4% in COVIPss. Median admission SOFA score was 4 and CFS was 3 with 14% considered frail (CFS > 5). In COVIPss median SOFA was 5 and 17% of PTS were frail. The preferred mode was CPAP with median maximum pressure of 13. Mean PaO2/fiO2 ratio after start of NIV was 125, 30% < 100. NIV failure occurred in 46.3% versus 74,7% in COVIPss. Our intra-unit mortality was 31.6%. 14 PTS (14.7%) were submitted to invasive mechanical ventilation and 57% of those died. In COVIPss mortality at 30d was 52.9% in NIV and 47.7 in IMV groups.

Conclusions: We argue that NIV is a valid option for COVID ARDS management if supported by a multifaceted strategy such as ours, using prone and CPAP for WOB control. We agree with COVIPss authors as NIV trial should be short and intubation promptly if WOB not controlled. Comparison with COVIP substudy NIV failure and mortality results, support our belief.


Delayed intubation with high flow nasal cannula in COVID: a comparison between a first and second pandemic wave

J Ringoir, A Van Hoorn, J Poelaert

UZ Brussel, Intensive Care Unit, Jette, Belgium

Critical Care 2023, 27(S1): P008

Introduction: High flow nasal cannula (HFNC) treatment is an efficient treatment for hypoxemia in acute respiratory distress syndrome (ARDS). Before the COVID pandemic, non-invasive ventilation was associated with higher mortality in ARDS, and early intubation was advocated. We hypothesized that HFNC treatment was more restrictive in the first wave of the COVID pandemic compared to the second wave respecting the pre-COVID conceptual consensus of early intubation in ARDS.

Methods: We conducted this retrospective observational single-center study in a tertiary ICU in Brussels during the COVID pandemic. The first flare-up ranged from March to May 2020, and the second flare-up from September to January 2021. All patients with COVID pneumonia and HFNC before intubation were included. We considered a delayed intubation a ROXi < 3.85 at the start of HFNC. ROXi is the ratio of SpO2/FiO2 to the respiratory rate. The primary outcome was delayed intubation based on the ROXi in the number of days. The secondary outcome was mortality.

Results: We included 60 patients in the first wave and 70 in the second wave. The duration of HFNC treatment before intubation was longer during the second wave, based on ROXi < 3.85: 1.6 days versus 2.8 days, p < 0.05 (Fig. 1). There was no significant difference in mortality, 18% versus 29%. The length of intubation was similar in both groups. The CCI, SOFA, APACHE III and SAPS II scores were similar in both groups.

Conclusions: The duration of HFNC treatment in COVID-19-related ARDS before intubation has significantly been extended in the second pandemic wave. The delayed intubation based on the ROXi was in this study without significantly increased mortality. However, a trend toward higher mortality after prolonged HFNC was seen in the second pandemic wave.

figure f

Fig. 1 (abstract P008). Comparison duration HFNC between first and second COVID wave


Outcomes following application of high flow nasal cannula and non-invasive ventilation during the second COVID-19 wave in Singapore

AS Nazem, ZY Tan, F Khan, D Bruce-Hickman

Ng Teng Fong General Hospital, Intensive Care Medicine, Singapore, Singapore

Critical Care 2023, 27(S1): P009

Introduction: COVID-19 pneumonia can result in significant morbidity and mortality in affected individuals. Our audit aims to study the respiratory outcomes of patients with COVID-19 pneumonia following the use of HFNC and NIV during the second wave of the pandemic.

Methods: We analysed the outcomes of 94 patients admitted to a tertiary combined HDU/ICU with severe COVID-19 pneumonia requiring non-invasive support between July and December 2021.

Results: 94 patients were admitted during the study period. ICU mortality rate was 22% (21/94), total in-hospital mortality was 38% (36/94). HFNC was used as first line respiratory support in 58/94 cases, of which 39.6% required intubation. Of those patients who failed HFNC, time to intubation was significantly higher in those patients who passed away than those who were intubated and survived (mean 6.08 days vs 2.86 days, p < 0.05 one sided T-test). In all patients, very late intubation defined as intubation > 5d post admission to ICU, occurred in 6/41 patients, of which the mortality rate was 100%. ROX score performed at 12 h post intubation was unable to discriminate those who succeeded with HFNC and those who required intubation (mean ROX 7.24 vs 7.9, p > 0.05). NIV was used in combination with HFNC pre-intubation in 5/23 HFNC cases with 100% mortality rate. Extubation failure rates were low (5/94) and use of tracheostomy was uncommon (4/94; all 4 survived ICU stay, 3 eventually died in hospital).

Conclusions: HFNC failure with prolonged use of HFNC and use of multiple non-invasive device strategies before intubation was associated with a high risk of mortality. Conventional measurements of HFNC failure in the form of a 12 h ROX score could not assist the clinician in predicting those patients at risk of HFNC failure.


Mortality rate, intensive care unit length of stay and time to orotracheal intubation of COVID-19 patients under different non-invasive ventilatory therapies: retrospective cohort study

GM Martins1, PLS Leme Silva2, APC Pereira Cruz2, VM Marques2, SM Christovam2, IP Prado2, CS Dos Santos Samary2, PR Rocco2, FC Cruz2

1Barra D’ Or Hospital, Intensive Care Unit, Rio De Janeiro, Brazil, 2Barra D’ Or Hospital, Rio de Janeiro Federal University, Rio De Janeiro, Brazil

Critical Care 2023, 27(S1): P010

Introduction: COVID-19 may lead to heterogeneous needs for ventilator therapy, whether oxygen therapy (OT), noninvasive ventilation (NIV), high-flow nasal catheter (HFNC) or their combination (NIV + HFNC). The purpose of the study was to describe, retrospectively, the mortality rate, intensive care unit length of stay (ICU-LOS) and time to orotracheal intubation of COVID-19 patients under OT, NIV, HFNC or combined (NIV + HFNC). A retrospective cohort study was done analyzing official medical data from March 2020 up to July 2021. (CAAE: 52534221.5.0000.5249).

Methods: The inclusion criteria were age > 18 years-old, and positive swab test for COVID-19 or computed tomography consistent of COVID-19. The exclusion criteria were hospital LOS less than 3 days, patients whose therapy (OT, NIV, HFNC or NIV + HFNC) lasted less than 48 h, and missing data about the outcome variables. The primary outcome was mortality rate, while secondary outcomes were ICU-LOS and time to orotracheal intubation. Chi-Square test was used to assess mortality rate. The Mann-Whitney U test was applied to assess differences in ICU-LOS and time to orotracheal intubation (p < 0.05).

Results: Overall, 1371 patients were enrolled. 880, 120, 35, and 148 patients were submitted to OT, NIV, HFNC or NIV + HFNC, respectively. The mortality rates were 8.4%, 29.6%, 22.2%, and 33.2% for OT, NIV, HFNC or NIV + HFNC, respectively (p < 0.001). The ICU-LOS was higher in NIV + HFNC (median [IQR] 15 days [16]) than NIV (9 days [10]) and OT (4 days [5], p < 0.001). The time to orotracheal intubation was higher in NIV (6 days [6]), HFNC (6 days [4.5]), and NIV + HFNC (6 days [6]) than OT (2 days [4]), p < 0.001. Mortality rate and ICU-LOS were higher in those patients requiring the combination of NIV and HFNC.

Conclusions: Although the type of ventilator therapy may be associated to increased mortality rate and ICU-LOS, we cannot assure causality due to exploratory nature of the retrospective study, but a marker of severity.


Ventilator avoidance among critically ill COVID-19 patients with acute respiratory distress syndrome

J Fletcher, P Hahn

University of Michigan Health West, Critical Care, Wyoming, USA

Critical Care 2023, 27(S1): P011

Introduction: Misinformation citing mechanical ventilation, not the virus, as causing death in COVID-19 patients with respiratory failure has led to ventilator avoidance (“initial refusal of intubation”) during the pandemic.

Methods: Prospective observational cohort study (March 2020–June 2021) evaluating the incidence and significance of “initial refusal of intubation” in patients with critical COVID-19 defined as ARDS requiring > 55% sustained FiO2 on high flow nasal canula (HFNC), non-invasive positive-pressure ventilation (NIPPV) or requiring intubation. Outcomes included in-hospital mortality and 1-year modified Rankin Scale (mRS) score. Logistic regression was used to estimate the age and Charlson Comorbidity Index adjusted odds ratio (OR) of in-hospital death. The Wilcoxon rank-sum test was used to evaluate differences in the mRs.

Results: The cohort was predominantly non-Latino white (76%), male (65%), unvaccinated (99.4%), mean age of 66, and good pre-COVID-19 functional status (median mRs score of 0). Overall, 315 patients were critically ill due to COVID-19 with an in-hospital mortality of 41.9% (132/315; 95% CI 36–47%). In patients in whom intubation was recommended 39% initially refused (40/102; 95% CI 30–49%). Utilization of HFNC (90%) and NIPPV (72%) were similar between groups, however actual use of mechanical ventilation differed (98.4% in those that did not initially refuse compared to 20% in those that initially refused (p = 0.001)). In-hospital mortality was 79.3% (49/62) in those who initially did not refuse intubation compared to 77.5% (31/40) in those who refused (adjusted OR 1.3; 95% CI 0.5–.5). The distribution of 1-year mRS was not significantly different between groups (p = 1.0) (Fig. 1).

Conclusions: Among critically ill patients with COVID-19 associated ARDS, ventilator avoidance was common however, it was not associated with increased in-hospital mortality or a difference in 1-year functional outcome.

figure g

Fig. 1 (abstract P011). Distribution of 1-year mRS scores in participants dichotomized by initial refusal of intubation


Diaphragmatic sonographic evaluation in critical illness patient to predict invasive mechanical ventilation use

K Suttapanit, P Supatanakij, S Wongkrasunt, S Savatmongkorngul

Faculty of Medicine Ramathibodi Hospital, Department of Emergency Medicine, Bangkok, Thailand

Critical Care 2023, 27(S1): P012

Introduction: Diaphragmatic dysfunction is one of the incidents in critically ill patients and associated worsening outcomes. The diaphragmatic function can be evaluated with bedside ultrasound. The diaphragmatic excursion is one of the methods in diaphragmatic ultrasound for detecting diaphragm dysfunction [1–3]. This study aimed to evaluate the performance of the diaphragmatic excursion to predict invasive mechanical ventilation (IMV).

Methods: In this prospective study, patients 18 years and older who visited the emergency department (ED) with critical illness were enrolled to measure the right-side diaphragmatic excursion within 10 min after arrival. Multivariable logistic regression and the area under the receiver operating characteristic (AUROC) curve was used to assess the performance of the diaphragmatic excursion to predict IMV.

Results: 314 patients were enrolled in this study, and 113 (35.9%) patients required IMV. The diaphragmatic excursion was an independent risk of receiving IMV with adjusted odd ratios 0.08 [95% confidence interval (CI), 0.04–0.17, p value < 0.001] and had the area under the receiver operating characteristic (AUROC) 0.850 (95% CI 0.807–0.894). The cut-off of the diaphragmatic excursion was 1.2 cm with a sensitivity of 82.3% (95% CI 74.0–88.8%) and specificity of 78.1% (95% CI 71.7–83.6%) (Table 1). Ventilator days in the diaphragmatic excursion < 1.2 cm (cm) was significantly longer [13 days (interquartile range (IQR); 5.27) versus 5 (IQR; 3, 8), p value 0.006].

Conclusions: The right-side diaphragmatic excursion had a good performance in predicting IMV in critical illness. The optimal cut-off of the diaphragmatic excursion was 1.2 cm. The benefit of implementing patient management should be assessed in the future.


  1. 1.

    Supinski GS et al. Chest. 2018;153:1040–51.

  2. 2.

    Dres M et al. Intensive Care Med. 2017;43:1441–52.

  3. 3.

    Tuinman PR et al. Intensive Care Med. 2020;46:594–605.

Table 1 (abstract P012). Prediction performance of diaphragmatic excursion for predicting intubation in critical illness in ED

Diaphragmatic excursion (cm)



Predicted probability

AUROC = 0.850 (95% CI 0.807–0.894)


≥ 2.00

3.4% − 0.4 to 7.3%)


97.3% (92.4–99.4%)

41.8% (34.9–48.9%)

12.8% (3.2–22.3%)


92.2% (85.4–96.3%)

62.2% (55.1–68.9%)

25.6% (12.5–38.6%)


82.3% (74.0–88.8%)

78.1% (71.7–83.6%)

58.5% (45.2–71.8%)

≤ 0.90

54.9% (45.2–64.2%)

89.1% (83.9–93.0%)

73.8% (64.4–83.2%)


A new simple method to identify the activation of the expiratory muscles. Index based on the determination of expiratory time constant

JF Martínez Carmona1, C Joya Montosa1, J Luna Castro1, MJ Delgado Amaya1, A Rodríguez Carmona2, J Castaño Pérez3, G Besso Centeno4, M Rodríguez Delgado5, JA Benítez Lozano, in memoriam1, JM Serrano Simón6

1Hospital Regional Universitario de Málaga, Intensive Care Unit, Málaga, Spain, 2Hospital El Carmen, Intensive Care Unit, Mendoza, Argentina, 3Hospital Virgen de las Nieves, Intensive Care Unit, Granada, Spain, 4Clínica del Aconcagua, Intensive Care Unit, Villa Mercedes San Luis, Argentina, 5Hospital La Merced, Intensive Care Unit, Osuna, Sevilla, Spain, 6Hospital La Merced, Osuna, Sevilla and Hospital Reina Sofía, Córdoba, Intensive Care Unit, Cordoba, Spain

Critical Care 2023, 27(S1): P013

Introduction: We hypothesized the expiratory time constant (tE) may be used to provide determinate expiratory muscle activation (EMA) in mechanically ventilated patients. Determination of tE by the quotient between exhaled tidal volume (Vt) and flow rate values between 0.10 and 0.5 s, during passive lung emptying, are linear fit. However, in EMA, tE obtained from exhaled Vt at 66% is dissociated from tE determined by 25–75% of the expiratory flow-volume curves (tE 25–75%). EMA increases the pleural pressure, with partial airway pressure (Paw) collapse and expiratory flow limiting. Our purpose was to evaluate the index tE25–75%/tE66%, to identify EMA.

Methods: A retrospective study, made by Córdoba University Hospital (January-May 2020), using available recordings of flow, Paw, esophageal (Pes) and gastric pressure (Pgas) of consecutive patients admitted to ICU, during partially assisted mechanical ventilation. Index tE was obtained cycle by cycle, during 60 min, by monitoring device and personal software. Sensitivity, specificity and ROC curve was carried out versus Pgas (gold std).

Results: 41 patients were included. 19 presented EMA: 64.16 (± 11.07) years. 80% male. 6 Pneumonia, 2 Single lung transplant, 3 Neurological, 2 Pancreatitis, 1 Trauma, 3 Cardiac and 2 abdominal surgery. Days on ventilator: 7.5 (1.5–9.5). Ers 19.06 (± 6.97) cmH2O/L. Rrs 13.50 (± 9.93) cmH2O/L/s. Peepi 7.66 (± 1.58) cmH2O. ΔPgas exp 9.67 (± 4.94) cmH2O. Settings: 6 A/C (flow 40–60 L/min), 13 PSV (10 ± 4)/PEEP5 cmH2O. Index tE: Passive exhalation 1.07 (± 0.11), EMA 2.18 (± 0.84)sec, (p < 0.001). Sensitivity and specificity (95% CI), threshold > 1.55: 0.84 (0.72–0.92), 0.95 (0.86–0.99), respectively; PPV 0.94, VPN 0.87. ROC curve: AUC = 0.98. A cut-off ≥ 1.4 was associated with the best index for success, sensivity of 0.95 (95% CI 0.85–0.99), specificity of 0.90 (95% CI 0.79–0.95), PPV 0.89, NPV 0.95.

Conclusions: Our index tE allows us to identify with accuracy the expiratory muscle activity, only by the flow signal, regardless of underlying disease.


Early diaphragm dysfunction in critically ill patients under invasive mechanical ventilation: preliminary data from a prospective observational study

V Silva1, C Pação2, D Palacios3, M Alves2, G Nobre de Jesus3, J Santos Silva3

1Hospital de Santa Maria, Centro Hospitalar Universitário Lisboa Norte, EPE, Intensive Care Department, Lisboa, Portugal, 2Faculdade de Medicina de Lisboa, Lisboa, Portugal, 3Hospital de Santa Maria, Centro Hospitalar Universitário Lisboa Norte, EPE, Lisboa, Portugal

Critical Care 2023, 27(S1): P014

Introduction: Diaphragmatic dysfunction (DD) is a keystone factor in difficult weaning from invasive mechanical ventilation (IMV). Diaphragmatic ultrasound (DUS) is the preferred method for the evaluation of diaphragm function in the ICU setting, namely through the diaphragm thickening fraction (DTF). However, its potential role in the decision-making process of mechanical ventilation weaning is yet to be established. We aimed to assess the incidence of early DD and its role as a predictor of prolonged IMV.

Methods: We conducted a prospective, non-interventional study in a universitary hospital ICU. Adult patients subject to at least 48 h of IMV were consecutively enrolled. Exclusion criteria was a prior period of IMV in the past 3 months. DUS was performed at 48 h of IMV. End-inspiratory and end-expiratory diaphragm thickness were measured using M-mode, with a high-frequency linear probe placed at the zone of apposition of the diaphragm. The mean values of 3 measurements were used to calculate DTF. Interobserver measurement variability was not evaluated.

Results: 32 patients were included. 34.4% were female, average age was 59.9 years. Mean SAPS II and SOFA at admission were 52.87 and 7.84, respectively. Mean DTF was 21.10% (± 14.59). Average IMV duration was 10.13 days (± 8.43). 75% of patients had DD at 48 h of IMV. A weak negative correlation (r < − 0.1) was observed between DTF, days of endotracheal intubation (p value 0.609) and days of IMV (p value 0.756). Using DTF cut-off values of 20% and 30%, DTF at 48 h of IMV did not correlate with prolonged IMV (p value of 0.427 and 0.704, respectively).

Conclusions: In our preliminary data, there was a high prevalence of DD at 48 h of IMV, as suggested in previous literature. DD at 48 h when measured through DTF did not seem to predict prolonged IMV. DUS is well-established for diaphragm functional assessment, but further research is needed to clarify its application in clinical practice.


The effect of differing loading conditions on ultrasound assessment of diaphragm function

IS Morris, EC Goligher

University Health Network, Department of Adult Critical Care Medicine, Toronto, Canada

Critical Care 2023, 27(S1): P015

Introduction: Diaphragm dysfunction is a common cause of failure to wean from invasive mechanical ventilation [1]. Ultrasound assessment of diaphragm activity has been proposed as a potential noninvasive method to assess diaphragm function and force generation [2]. The influence of varying physiological conditions on these measurements has not been systematically evaluated.

Methods: Healthy adult volunteers underwent esophageal and gastric balloon placement to measure transdiaphragmatic pressure (Pdi). Diaphragm thickening fraction (TFdi) and diaphragm excursion (EXdi) were measured under varying resistive load (none, marked, complete airway occlusion), varying postural load (semi-recumbent vs. supine) and varying thoracoabdominal motion (abdominal- vs. rib cage-predominant inspiration). The influence of physiological conditions on these measurements was analyzed by linear mixed effect regression.

Results: Nine participants were included. Pdi swing increased with resistive load (p < 0.001) and decreased in the supine position (p < 0.001) (Table 1). By contrast, TFdi did not change significantly when marked resistance was applied (p = 0.44) but decreased with complete airway resistance (p = 0.055). TFdi also decreased in the supine position (p = 0.004) and with rib cage-predominant inspiration (p < 0.001). EXdi decreased with increasing resistance (p < 0.001) and supine position (p = 0.013).

Conclusions: Changing physiological conditions has differential effects on Pdi, TFdi, and EXdi. Position, mechanical load, and thoracoabdominal motion should be considered when interpreting these parameters to assess diaphragm function. TFdi is relatively less affected by physiological conditions than EXdi.


  1. 1.

    Dres M et al. Intensive Care Med 2017;43:1441–1452.

  2. 2.

    Goligher EC et al. Am J Respir Crit Care Med 2020;202:950–961.

Table 1 (abstract P015). The effect of differing loading conditions on measures of diaphragm function


Pdi cmH2O: mean (SEM); p value

TFdi %: mean (SEM); p value

EXdi cm: mean (SEM); p value

Resistive load: none (reference)

53 (2.0)

52 (6.0)

2.9 (0.21)

Resistive load: marked

84 (2.8); < 0.001

57 (6.4); 0.44

1.4 (0.12); < 0.001

Resistive load: complete

87 (3.2); < 0.001

42 (4.8); 0.055

1.2 (0.10); < 0.001

Position: semi-recumbent (reference)

82 (1.54)

56 (5.7)

2.0 (0.13)

Position: supine

67 (2.18); < 0.001

45 (3.5); 0.004

1.5 (0.18); 0.013

Thoracoabdominal motion: abdomen predominant (ref)


115 (15.1)


Thoracoabdominal motion: rib cage-predominant


60 (7.5); < 0.001


  1. Pdi, transdiaphragmatic pressure; TFdi, diaphragm thickening fraction; EXdi, diaphragm excursion; SEM, standard error of the mean); reference, reference category; n/a, not applicable (measured)


Temporary diaphragm pacing electrodes to decrease mechanical ventilation and improve diaphragm function during cardiac procedures

RPO Onders, M Elmo, N Carl

University Hospitals Cleveland Medical Center, Department of Surgery, Cleveland, USA

Critical Care 2023, 27(S1): P016

Introduction: Studies have shown diaphragm pacing (DP) prevents ventilator induced diaphragm dysfunction (VIDD) secondary to mechanical ventilation (MV) improving extubation and facilitates injured phrenic nerves recovery. We report our experience utilizing a temporary DP system during cardiac procedures.

Methods: This is a retrospective analysis of prospective IRB approved databases of non-randomized interventional experiences at a single institution. Subgroup analysis was then limited to patients implanted with the temporary DP system (TransAeris, Synapse Biomedical) during cardiac procedures. Prior to sternotomy closure, the pleural space is opened, two electrodes are placed in each hemi- diaphragm. The electrodes also record diaphragm burst electromyography (dEMG). Once in the intensive care unit DP ensued using a disposable multi-channel stimulator generating a fused diaphragm contraction preventing VIDD.

Results: From 1/1/2020 to 6/30/22 252 patients were implanted with either the chronic long term DP (NeuRx) or the temporary DP system. 21 patients were implanted with the temporary DP system during a cardiac procedure. All had risk factors that predict prolonged MV including: prior open cardiac surgery, LVEF less than 30%, COPD, history of CVA, or pre-operative intra-aortic balloon pump In 3 patients, there was rapid weaning with bilateral dEMG and no stimulation occurred. The remaining 18 patients underwent diaphragm stimulation to facilitate weaning and maintain diaphragm strength. In patients with pre-existing unilateral dysfunction, DP prevented paradoxical diaphragm movement and associated sleep dysfunction. There were no adverse events related to electrode implantation, stimulation or subsequent removal.

Conclusions: Prolonged MV after cardiac surgery leads to significant morbidity, mortality and cost. DP was safely used to wean from MV, prevent VIDD and the electrodes can identify and improve recovery of phrenic nerve injuries.


Comparison of ultrasound assessment for diaphragmatic workload during spontaneous breathing trial between tube compensation and pressure support ventilation

N Kulkanokwan1, S Morakul2, C Pisitsak2, W Mongkolpun3, P Theerawit1

1Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Division Critical Care Medicine, Department of Medicine, Bangkok, Thailand, 2Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Department of Anesthesiology, Bangkok, Thailand, 3Siriraj Piyamaharajkarun Hospital, Siriraj Hospital, Mahidol University, Department of Critical Care Medicine, Bangkok, Thailand

Critical Care 2023, 27(S1): P017

Introduction: Spontaneous breathing trail (SBT) with pressure support ventilation (PSV) might provide over assistant while SBT with T-pieces may increase workload. Therefore, tube compensation (TC) might be an optional SBT mode but its effect on patients’ effort has not been well shown. This study aimed to compare the patient’s effort, measured by diaphragm ultrasonography, between TC and PSV and PEEP during SBT.

Methods: This was a crossover study. Patients requiring mechanical ventilation ≥ 48 h with the weaning requirements were included. TC (30 min) was performed and PSV 5 cmH2O (30 min) was done after TC in one group. In the other group, PSV was performed before TC. During the washout period, baseline ventilation was resumed until tidal volume, respiratory rate, and heart rate were return to baseline. Before and immediate after 30 min of PSV and TC, diaphragmatic workload parameters were assessed by relative change (%) of diaphragm thickness fraction (∆DTF), diaphragm excursion (∆DE) and tissue Doppler imaging (TDI) parameters of peak contraction velocity (∆PCV), peak relaxation velocity (∆PRV), velocity–time integral (∆VTI) and TDI-derived maximal relaxation rate (∆TDI-MRR).

Results: For our results, the interaction between period and treatment was significant. So, we used only the first period in our interpretation. Among 48 patients who met the inclusion criteria, nobody failed SBT before 30 min. ∆DTF was significant greater with TC than with PSV ([65.84 (17.52, 83.38)] % vs [0.49 (−40.85, 30.47)] %, p < 0.01, respectively). ∆DE was significant greater with TC [14.75(0.00,36.45)] than with PSV [−9.47(−33.14, −1.01)] % (p < 0.01). Change of TDI variables, except ∆TDI-MRR, did not differ between TC and PSV (Table 1).

Conclusions: SBT with TC was associated with higher diaphragm effort than with PSV with the same PEEP.

Table 1 (abstract P017). Ultrasound-measured diaphragmatic parameters in the tube compensation and the low-pressure support periods during spontaneous breathing trial




Difference between treatments (95% CI)

p value

∆DTF (%)

0.51 (−40.85, 30.46)

54.24 (10.59, 70.25)

61.28 (23.27, 99.29)


∆DE (%)

−9.47 (−33.14, to 1.01)

14.75 (0.00, 36.45)

37.42 (19.32, 55.53)

 < 0.001

∆PCV (%)

0.67 (−8.63, 14.21)

2.56 (−9.38, 16.02)

5.31 (−6.54, 17.17)


∆PRV (%)

2.15 (−21.93, 15.75)

16.43 (−7.83, 31.26)

12.89 (−9.10, 34.87)


∆VTI (%)

11.47 (−8.23, 23.36)

−4.92 (−24.58, 12.04)

−17.81 (−37.38, 1.75)


∆TDI-MRR (%)

−9.48 (−37.70, 31.67)

24.09 (−24.23, 58.77)

44.13 (0.11, 88.15)



A hemodynamic echocardiographic evaluation predicts prolonged mechanical ventilation in septic patients

T Giraldi1, JR Matos-Souza2, D Cecilio-Fernandes 3, T Santos3

1University of Campinas, Emergency Medicine, Campinas, Brazil, 2University of Campinas, Cardiology, Campinas, Brazil, 3University of Campinas, School of Medical Sciences, Campinas, Brazil

Critical Care 2023, 27(S1): P018

Introduction: Prolonged mechanical ventilation (PMV) is common among critically ill septic patients and leads to serious adverse effects. Transthoracic echocardiography (TTE) is an efficient tool for the assessment of septic shock. Our study investigated the relationship between TTE parameters and PMV in mechanically ventilated septic shock patients.

Methods: TTE was performed in the first 24 h of intensive care unit admission, acquiring data on cardiac output (CO), cardiac index (CI), s′ wave (s′), E wave (E), e′ wave (e′) and E/e′ ratio. We compared data on patients who met the criteria for PMV with data on patients who did not.

Results: Sixty-four patients were included, 26 of whom met the criteria for PMV. CO, CI and s′ were higher in PMV patients (5.49 vs. 4.20, p = 0.02; 2.95 vs. 2.34, p = 0.04; and 12.56 vs. 9.81, p = 0.01, respectively). Patients with PMV had greater mean CO, CI and s′ (Fig. 1). The area under the ROC curve for CO in predicting PMV was 0.707 (p < 0.01). A cutoff value of 4.82 L/min was associated with the highest diagnostic accuracy and predicted PMV with a sensitivity of 61.5% and a specificity of 78.9%; with positive likelihood ratio (PLR) value of 2.91 and a negative likelihood ratio (NLR) of 0.49. CI had an area under the ROC curve of 0.69 (p = 0.01). A cutoff value of 2.54 L/min/m2 was associated with the highest diagnostic accuracy and predicted PMV with a sensitivity of 69.2%, a specificity of 71.1%; PLR of 2.39 and NLR of 0.43. The s′ wave had an area under the ROC curve of 0.679 (p = 0.03). A cutoff value of 13.85 cm/s was associated with the highest diagnostic accuracy and predicted PMV with a sensitivity of 42.1% and a specificity of 90.6%; it provided a PLR of 4.48 and a NLR of 0.36.

Conclusions: Increased values of CO, CI and s′ wave on ICU admission were related to PMV. The s′ correlated with CI and could estimate systemic vascular resistance. This echocardiographic hemodynamic protocol could be added to the initial assessment of mechanically ventilated septic patients.

figure h

Fig. 1 (abstract P018). Examples of CI and s′-wave measurements in a patient with PMV (A) and in one without PMV (B). The left ventricular outflow tract ventricular-time integral (LVOT VTI) (arrows) is illustrated in (A.1) and (B.1), and tissue Doppler peak systolic velocity (s′ wave) (arrowheads) in (A.2) and (B.2). Cardiac index values for patients A and B were, respectively, 4.81 and 1.49 L/min/m2, and s′ wave values for patients A and B were 21.16 and 7.42 cm/s, respectively




Metabolic acidosis in pre-intubation predicts post-intubation hypotension during emergency airway management

M Suga, T Nishimura, S Maemura, T Ochi, S Ishihara.

Hyogo Emergency Medical Center, Department of Emergency and Critical Care Medicine, Kobe, Japan.

Critical Care 2023, 27(S1): P020

Introduction: Post-intubation hypotension (PIH) in the emergency department (ED) endotracheal intubation is an adverse event that is associated with poor outcomes. Metabolic acidosis in a pre-intubated scenario represents a significant challenging situation and a risk for patient deterioration. However, little is known about the relationship between metabolic acidosis and PIH. We aimed to evaluate the association between pre-intubation metabolic acidosis and PIH during emergency airway management.

Methods: We conducted a single-center, retrospective observational study at an urban ED from Nov 2016 to Mar 2022. Patients who required emergent intubation at ED were included. Patients younger than 18 years, with systolic blood pressure(sBP) < 90 mmHg, received vasopressors or mechanical circulatory support before intubation were excluded. The primary endpoint was the incidence of PIH, defined as sBP < 90 mmHg or the administration of any vasopressors within 30 min after intubation. Multivariable logistic regression models were used to evaluate the association between PIH and metabolic acidosis, adjusted with known risk factors (shock index (SI) before intubation, age, gender, use of non-depolarizing neuromuscular blocking agents, and intubation for respiratory failure).

Results: In total, 257 patients were enrolled in the current study, and 87 patients (33.9%) developed PIH. The median age was 67 years [interquartile range (IQR): 49–76], and 119 (46.3%) of 257 patients had metabolic acidosis. Compared to patients without metabolic acidosis in pre-intubation, those with metabolic acidosis showed a significantly higher incidence of PIH (51.2% vs. 18.8%, p < 0.001). In the patients overall, multivariable logistic regression demonstrated that age (odds ratio [OR] 1.03; 95%confidence interval [CI] 1.0–1.1), metabolic acidosis (OR, 3.4; 95% CI 1.9–6.3), shock index (OR 8.1; 95% CI 2.4–27.6) was significantly related to PIH.

Conclusions: Metabolic acidosis in pre-intubation could be a predictor of PIH at the ED.


Intubation times with the bougie versus stylet in the immobilized cervical spine: a randomized trial

A Hazarika1, D Namratha1, D Jain1, N Bhatia1, A Ray1, S Singh Dhatt2, M Karthigeyan3, C Ahuja4

1Postgraduate Institute of Medical Education and Research, Department of Anaesthesia & Intensive Care, Chandigarh, India, 2Postgraduate Institute of Medical Education and Research, Dept of Orthopaedics, Chandigarh, India, 3Postgraduate Institute of Medical Education and Research, Dept. of Neurosurgery, Chandigarh, India, 4Postgraduate Institute of Medical Education and Research, Dept. of Radiology, Chandigarh, India

Critical Care 2023, 27(S1): P021

Introduction: In cervical spine injury patients, the neck is stabilized by neck collar and traction, limiting neck movements. This leads to failure in aligning the laryngeal, pharyngeal and oral axes, making the visualization of the larynx difficult. Endotracheal intubation (ETI) in patients with an immobilized cervical spine is often challenging, urging the use of airway adjuncts like bougie and stylet and a video-laryngoscope (VL). Owing to the scarcity of comparative literature, we aimed to compare the intubation characteristics when using either a bougie or a stylet with a VL in these patients.

Methods: This randomized controlled study involved eighty-six adult ASA I/II patients with cervical spine immobilized with a collar or traction, scheduled for cervical spine surgery. ETI was performed with the C-MAC VL, assisted with bougie (Group ETB) or stylet (Group ETS). The primary outcome was time to successful ETI. First-attempt success (FAS) rate, overall successful ETI, cervical spine motion detected using fluoroscopy and complications were secondary outcomes. Any movement of the atlanto-occipital joint during ETI was noted by fluoroscopy image at two time points; T1-at the neutral position (during bag and mask ventilation after induction) and T2-point of insertion of the ETI through the glottis aperture. Fluoroscopy data was analysed only for successful ETI.

Results: The time for ETI in group ETB was 52.38 ± 20.85 s (n = 43), and in group ETS was 52.39 ± 32.85 s (n = 43), p = 0.958. There was no significant difference in FAS rate, overall success of intubation or cervical spine movement between the groups (Fig. 1). No complications were encountered.

Conclusions: In patients with an immobilized cervical spine, there was no significant difference in the intubation times when comparing the bougie and the stylet. The FAS rate was also similar in both groups with minimal motion at C1, C2. Both bougie and stylet are equally useful adjuncts when used with a VL, while intubating patients in whom neck movements are restricted.

figure i

Fig. 1 (abstract P021). The overall time to successful intubation, time to successful 1st attempt intubation, overall success rate and change in angulation between cervical vertebra C1 and C2 as measured during intubation by fluoroscopy


Sealing performance of taper shaped cuff for pediatric use

P Lichtenthal1, D Wade2, U Borg2

1University of Arizona, Department of Anesthesiology, Tucson, USA, 2Medtronic, Respiratory Interventions, Boulder, USA

Critical Care 2023, 27(S1): P022

Introduction: Studies have demonstrated that current barrel-shaped high volume, low-pressure endotracheal tube (ETT) cuffs allow aspiration of fluids through channels formed in the cuff [1]. Leakage around the ETT cuff is known to contribute to microaspiration [2]. A tapered shaped ETT cuff, in the adult population, has demonstrated less fluid leakage [3] and possibly a decrease in tracheal damage [1]. A taper shaped ETT cuff, intended for the pediatric population, was recently introduced. The object of this study was to compare fluid sealing performance of a taper shaped ETT cuff (Falcon II, Medtronic, Boulder, CO, USA) to that of barrel shaped ETT cuffs (Shiley™ Hi-Lo, Medtronic, Boulder CO USA; MicroCuff, Avanos) using excised animal tracheas.

Methods: Excised rabbit tracheas, including larynx, were utilized. Tracheas were wrapped in saline soaked gauze to prevent drying and suspended by a laboratory clamp stand. Prior to intubation, the cuffs were inflated to 50 cmH2O to eliminate any wrinkles and confirm integrity. After placement in the trachea, the cuffs were inflated to 25 cmH2O and monitored continuously. Saline (0.5–1.0 ml) was placed on top of the ETT cuff and leak volume was collected at 1, 5 and 10 min using a beaker placed on a digital scale. The smallest ETT size available for each device tested was employed for testing (2.5 mm for taper shaped ETT cuff and 3.0 for barrel shaped ETT cuffs).

Results: No statistically significant differences in leaked saline amount were observed between the taper shaped ETT cuff and barrel shaped ETTs.

Conclusions: The findings of the study suggest comparable performance between the taper shaped ETT cuff and barrel shaped ETT cuff, when assessing leak volume using a bench top study design. Further studies are required to elucidate the possibility of decreased injuries to vocal folds and trachea as a result of cuff shape.


  1. 1.

    Chang JE et al. Anesth Analg. 2017;125:1240–1245.

  2. 2.

    Pozuelo-Carrascosa DP et al. Eur Respir Rev 2020;29:190,107.

  3. 3.

    D’haese J et al. Acta Anaesthesiol Scand. 2013;57:873–80.


Benefits of fully closed loop ventilation modes in patients with body mass index > 35 kg/m2

R Komnov, A Eremenko, A Alferova, D Riabova, P Titov, A Urbanov, M Fominukh, S Gerasimenko

Petrovsky Russian Research Center of Surgery, Post Cardiac Surgery Intensive Care unit, Moscow, Russian Federation

Critical Care 2023, 27(S1): P023

Introduction: In recent years, there is an active automation of processes in medicine and robotic (Intellectual) ventilation modes are being actively introduced into the daily practice of intensive care units. These modes allow for automated respiratory support, minimizing clinician involvement.

Methods: In this randomized trial 32 adult patients with BMI > 35 were included. Half of them were ventilated with INTELLiVENT-ASV® mode and 16 with conventional ventilation modes after uncomplicated cardiac surgery. Hamilton G5 ventilators were used and 8 physicians were involved into the study. Care of both groups was standardized, except modes of postoperative ventilation. We compared:

  • The physician’s workload, through accounting number of manual ventilator settings and time they spent near the ventilator in every group;

  • Safety of respiratory support by considering driving pressure and mechanical power during mechanical ventilation, positive end expiratory pressure, FiO2 and tidal volume level during all phases of respiratory support;

  • Duration of tracheal intubation in the ICU.

Results: In the Intellivent group the number of manual ventilator settings and physician’s time spent near the ventilator before tracheal extubation were significant lower: 1 (0–2) versus 6 ± 2, and 64 ± 17 s versus 36 ± 90 s respectively (p < 0.0001 in both cases). Intellivent-ASV mode was more protective compared to conventional mode through significant reduction in the driving pressure, mechanical power, tidal volume, FiO2 and PEEP levels, but without difference between PaO2/FiO2 ratio (Table 1). There were no significant differences in the duration of respiratory support in the ICU: 276 ± 103 min (Intellivent group) versus 300 (225–175) min (control) (p = 0.2427).

Conclusions: Intellivent-ASV® mode in patients with BMI > 35 after uncomplicated cardiac surgery allows to personalize respiratory support, provides more protective mechanical ventilation and reduces the physician’s workload.

Table 1 (abstract P023). Main outcomes


Intellivent—ASV group (n = 16)

Control group (n = 16)


BMI (Body mass index); (CI mean 95%)

39 ± 2 (37–40)

38 ± 2 (37–39)


Ratio PaO2/FiO2; (CI mean 95%)

320 ± 37 (300–339)

323 ± 52 (295–351)



29 (28–30)

35 (30–40)


Tidal volume (Vt), ml/kg/PBW (CI mean 95%)

6.4 ± 0.5 (6.1–6.6)

9 ± 1.3 (8.3–9.7)


∆P (driving pressure), cmH2O (CI mean 95%)

7 ± 1 (7–8)

9 ± 2 (8–10)


PEEP, cmH2O; (CI mean 95%)

9 ± 2 (8–10)

11 ± 2 (10–12)


Mechanical power, J/min; (CI mean 95%)

10 ± 3 (8–12)

15 ± 5 (12–18)



Automatic close-loop ventilation for patients with traumatic brain injuries

S Vosylius1, A Ramaskaite2, J Semenkovaite2, E Januskeviciute3, J Meskauskiene1, N Aliukoniene1, A Balakirev4

1Republican Vilnius University Hospital, ICU, Vilnius, Lithuania, 2Vilnius University, Faculty of Medicine, Vilnius, Lithuania, 3Vilnius University, Clinic of Anesthesiology and Intensive Care, Vilnius, Lithuania, 4Acrux Cyber Services, Acrux Cyber Services, Vilnius, Lithuania

Critical Care 2023, 27(S1): P024

Introduction: Automatic close-loop mode of mechanical ventilation compared to traditional modes provides ventilated patients with monitoring parameters within safety limits and reduces the risk of ventilator induced lung injury. The aim of this study was to assess the ventilator parameters for the patients with traumatic brain injury under adaptive support ventilation (ASV).

Methods: This prospective observational study was conducted in the intensive care unit (ICU) of Republican Vilnius University Hospital. We included 11 patients ventilated at least 48 h with ventilators Hamilton S1 using INTELLiVENT®-ASV mode applying automatic control of minute volume, positive end expiratory pressure, and FiO2. IntelliSync + tool was applied for better synchronization between the inspiratory and expiratory times of patient respiratory system and ventilator. The patients were continuously monitored using Acrux DeepBreath software. The study protocol was approved by the regional Vilnius bioethics committee (2021/9-1380-851).

Results: A total duration of mechanical ventilation for all patients was 1541 h (from 62 to 289). A total of 1,847,197 breath cycles were analysed after exclusion of 2.3% of cycles defined as artefacts related to nursing, and therapeutic interventions. 95.2% of all breath cycles (for single patient: min. 86.8–max. 98.8%) meet criteria for lung protection (Pmax ≤ 30cmH2O, VT/IBW ≤ 12 ml/kg). Median of monitoring parameters: MVe = 9.8 l/min, VT = 572 ml, respiratory rate −17 breaths/min, VT/IBW = 7.5 ml/kg, Pmax = 19 cmH2O, PEEP = 5 cmH2O). Gas exchange was optimal: PetCO2 = 36 cmH2O, SpO2 = 98% with FiO2 = 35%. Monitoring parameters from ventilators during time scale were stable over 48 h (Table 1).

Conclusions: The patients with traumatic brain injury ventilated using automatic ventilation mode (INTELLiVENT®-ASV and IntelliSync + tool) during the first 48 h had optimal gas exchange and monitoring parameters within safety limits for lung protection.

Table 1 (abstract P024). Monitoring ventilator parameters during time scale over 48 h





































Pressure-controlled versus volume-controlled ventilation: is there a difference in ventilation homogeneity?

T Sebrechts1, P Vets1, T Bleeser.2

1ZNA Middelheim, Intensive care unit, Antwerp, Belgium, 2UZ Leuven, Department of Anesthesiology, Leuven, Belgium

Critical Care 2023, 27(S1): P025

Introduction: During mechanical ventilation, volume-controlled ventilation (VCV), which performs a constant inspiratory flow, may create a more homogenous ventilation compared to pressure-controlled ventilation (PCV), which performs a decelerating inspiratory flow. Ventilator associated lung injury can be reduced by a more homogenous ventilation. Two parameters concerning ventilation homogeneity can be conducted by using electric impedance tomography (EIT). Regional ventilation delay (RVD) displays regional varying time constants [1]. Intratidal variation (ITV) describes impedance changes [2].

Methods: In this observational prospective study, included patients were intubated and sedated because of a respiratory disease, and had no signs of spontaneous breathing. EIT measurements were performed during PCV first, followed by VCV during five minutes each. Respiratory frequency, inspiratory time and PEEP were held constant. In RVD, five regions of interests were investigated (global, ventral, midventral, middorsal and dorsal). In ITV, there were four regions (ventral, midventral, middorsal and dorsal). Wilcoxon signed-rank test was used.

Results: Five patients were included. ITV measurements did not show any statistically significant or clinically relevant differences between PCV and VCV. RVD measurements showed a trend to a higher RVD in the middorsal part in every patient when VCV was applied [median(IQR) + 0.95% (+ 0.84; + 1.04), p = 0.0625]. Other results were neither statistically significant, nor clinically relevant.

Conclusions: There was no evidence for relevant differences in ventilation homogeneity between PCV and VCV, as measured by ITV. For RVD there was a trend that only the middorsal part, against expectations, experiences a less homogenous ventilation during VCV, with no relevant differences in other regions. Additional research with a larger patient group is required.


  1. 1.

    Muders T et al. Yearbook of Intensive Care and Emergency Medicine 2009, pp 405–412.

  2. 2.

    Lowhagen K et al. Minerva Anestesiol 2010;76:1024–35.


Ventilator management and risk of air leak syndrome for patients with COVID-19 pneumonia

N Miyake, Y Igarashi, R Nakae, T Mizobuchi, T Masuno, S Yokobori

Nippon Medical School Hospital, Emergency and Critical Care Medicine, Tokyo, Japan

Critical Care 2023, 27(S1): P026

Introduction: Coronavirus disease 2019 (COVID-19) pneumonia is reportedly associated with air leak syndrome (ALS), including mediastinal emphysema and pneumothorax, and has a high mortality rate. In this study, we compared values obtained every minute from the ventilator to clarify the relationship between ventilatory management and the risk of developing ALS.

Methods: This was a single-center, retrospective, observational study for a 21-months period. Patient background, ventilator data, and outcomes were collected from adult patients with COVID-19 pneumonia on ventilator-assisted respiratory management. The primary outcome was the development of ALS within 30 days of starting ventilator management.

Results: Of the 105 patients, 14 (13%) developed ALS. The mean positive-end expiratory pressure (PEEP) difference was 0.33 cmH2O (95% confidence interval (CI) 0.31–0.33), and it was higher in the ALS than in the non-ALS group (9.18 ± 2.20 versus 8.85 ± 2.63, respectively). For peak pressure, the mean difference was −0.18 cmH2O (95% CI −0.20 to −0.15), (20.70 ± 5.30 vs. 20.87 ± 5.65) and the mean pressure difference of −0.05 cmH2O (95% CI −0.04 to −0.07) (12.80 ± 3.13 vs. 12.85 ± 3.55, respectively) was also higher in the non-ALS group. The difference in the single ventilation volume per ideal body weight was 0.65 ml/kg (95% CI 0.63–0.66) (7.83 ± 3.16 vs. 7.18 ± 2.96, respectively), and the difference in dynamic lung compliance was 8.57 mL/cmH2O (95% CI 8.43–8.70) (50.32 ± 31.68 vs. 49.68 ± 15.16, respectively), and both were higher in the ALS group. The percentage of times that the ventilation volume per body weight exceeded 8 was higher in the ALS group (53.7% vs. 38.6%, p < 0.001).

Conclusions: There was no association between higher ventilator pressures and the development of ALS. The ALS group had higher dynamic lung compliance and higher tidal volumes, which may indicate a pulmonary contribution to ALS, and ventilatory management that limits tidal volume may prevent the development of ALS.


VentConnect: telemetry system that enables remote live transmission of ventilator screen: qualitative study of clinicians’ expectations and experience

V Zvonicek1, M Macik2, J Jirman2, M Nemy2, L Vyslouzilova2, F Duska3

1University Hospital Kralovské Vinohrady, ICU, Prague, Czech Republic, 2Czech Technical University in Prague, Prague, Czech Republic, 3University Hospital Kralovské Vinohrady, Prague, Czech Republic

Critical Care 2023, 27(S1): P027

Introduction: To maximise the input of intensivists onto the management of ventilated patients during the COVID pandemic, we have developed and implemented telemetry system VentConnect [1]. The aim of this study is to identify stakeholder’s expectations and experience from this technology.

Methods: The telemetry device VentConnect (scheme at Fig. 1) enabled transmission of HDMI signal from mechanical ventilators to a password protected interface on any web browser. We implemented it between December 2020 and March 2021 on a total of 31 beds where patients were treated during COVID Pandemic. Afterwards, we performed “Structured User Interviews with ICU doctors”. Questionnaire responses we clustered and calculated.

Results: Eight doctors were interviewed, 4 fully qualified intensivists, and 4 in training. By far the most demanded was the ability to see flow curve or flow pattern (100%), followed by inspiratory pressures (75%) and check tidal volume (63%). Other parameters were mentioned less frequently such as driving pressure (25%) and interferences (38%). With regards users experience, answers were overwhelmingly positive, highlighting mostly the ability to continuously monitor the progress of patients without the need to donning personal protective equipment. In some, however, curiosity was the only motivator for use. Three juniors expressed apprehension that their supervisors might criticise their ventilator setting which would otherwise had gone unnoticed. Two participants thought that the temptation to check patient 24/7 would impair their ability to rest and relax during their off time.

Conclusions: Telemetry system that enabled clinicians to remotely check ventilator screen met the expectation of clinicians, who mainly demanded to check flow patterns, tidal volumes and pressures. Concerns were mainly about psychological impact of using this technology. These need to be addressed.

Acknowledgement: Supported by Grant Ministry of Health, Czech Republic, AZV NU22-06-00625.


  1. 1.

figure j

Fig. 1 (abstract P027). Scheme of VentConnect System


Understanding the influences on clinical decisions regarding ventilation of critically ill COVID-19 patients

K Matsumoto, H Malcolm, W Thomson, M Osman, Z Puthucheary, J Prowle, T Stephens

Critical Care and Perioperative Medicine Research Group, William Harvey Research Group, Queen Mary University of London, London, UK

Critical Care 2023, 27(S1): P028

Introduction: National Service Evaluations of COVID-19 ARDS care in the US and UK showed significant variability in clinical practice, and adherence to existing guidelines. To better understand the basis for this, we explored factors influencing decision-making around mechanical ventilation in COVID-19.

Methods: We conducted interprofessional focus groups identifying factors that influenced decision-making through thematic analysis. From this, we developed a questionnaire to validate these themes with a larger sample of critical care professionals across the UK. Kruskal–Wallis or Mann–Whitney U tests were used for data analysis.

Results: There were 179 complete responses from doctors, nurses and physios. In their usual practice, 66% of clinicians reported adherence to national ARDS guidelines. However, 80% thought COVID-19 ARDS presented differently to their previous clinical knowledge/experience of ARDS and 72% thought deviating from usual practice was necessary. Doctors were more likely to think deviation was necessary (p < 0.001) but there was no difference across level of ICU experience (p = 0.845). Clinicians reported their ventilatory decision-making was most influenced by disease factors, followed by team then contextual and least by environmental factors (p < 0.001). Disease factor was seen as most important across profession and experience level. During COVID-19, 68% of clinicians reported not being confident in their ventilatory decision-making; however, clinicians who felt COVID-19 ARDS presentation fitted with their previous clinical knowledge/experience of ARDS reported greater confidence (p < 0.001). Confidence was not affected by experience (p = 0.522) or profession (p = 0.294) (Fig. 1).

Conclusions: Clinicians were influenced by the uncertain understanding of COVID-19 ARDS, especially when they considered previous experiences to be less relevant. In the event of another novel disease, developing a consistent, understandable clinical models of disease should be prioritised to optimise decision making.

figure k

Fig. 1 (abstract P028). Whisker and Boxplot chart of findings. Charts compare confidence scores (out of 20) based on clinician’s perception of COVID-19 (different or similar to their previous clinical knowledge/experience of ARDS), ICU experience and profession. p values are based on Mann–Whitney U test or Kruskal–Wallis test


Prolonged mechanical ventilation: a retrospective regional study to identify need for specialist weaning service

G Thomas1, J Qiu1, A Campbell2, V Sundaram1, E Farley-Hills3, R Pugh4

1Department of Anaesthetics, Bodelwyddan, UK, 2Wrexham Maelor Hospital, Critical Care, Wrexham, UK, 3Ysbyty Gwynedd, Department of Anaesthetics, Bangor, UK, 4Department of Anaesthetics, Glan Clwyd Hospital, Bodelwyddan, UK

Critical Care 2023, 27(S1): P029

Introduction: Prolonged mechanical ventilation (PMV) has been defined as the need for ≥ 21 consecutive days of mechanical ventilation for ≥ 6 h per day [1], representing approximately 5% ventilated ICU patients but consuming 30% ICU bed days [2, 3]. Specialist weaning units offer dedicated multi-professional care to patients within a cohorted environment while freeing acute critical care capacity. Elsewhere in the UK, the impact of weaning unit on critical care capacity has been modelled using a transfer threshold of 21 days [2], whereas for Wales the benefit of centralising patients ventilated > 100 days has been evaluated [4].

Methods: The aim of this study was to evaluate over a 5-years period: numbers of patients requiring prolonged ventilation regionally, their underlying characteristics, critical care utilisation, outcome and discharge disposition. We conducted a retrospective analysis, utilising the critical care database of three North Wales ICUs from April 1st 2017 to 31st March 2022.

Results: During the study period, 3640 critically ill patients received invasive respiratory support over 22,245 ventilator days. Those requiring ventilation for ≥ 30 days represented a small proportion (3%) of all ventilated patients but consumed nearly a quarter of all ventilator days (Table 1). The majority were admitted for primary respiratory aetiology (55%) with a smaller proportion (15%) for intra-abdominal pathology. Two patients were transferred to an external weaning unit for specialist management after median 203 days.

Conclusions: Patients requiring a mechanical ventilation for ≥ 30 days account for a significant proportion of ventilator days. The majority of such patients have a respiratory primary diagnosis and survive to hospital discharge; their needs may be sufficiently homogenous and the impact on wider critical care capacity sufficiently great that cohorting within an appropriately resourced specialist weaning unit may be appropriate. This requires further exploration.


  1. 1.

    MacIntyr NR et al. Chest 2005;128:3937–3954.

  2. 2.

    Lone NI and Walsh TS. Crit Care 2011;15: R102.

  3. 3.

    Hill AD et al. Ann Am Thorac Soc 2017;14: 355–362.

  4. 4.

    Welsh Government. Task and Finish Group Report on Critical Care, 2019.

Table 1 (abstract P029) Patient characteristics and outcomes


 ≥ 30 days

 ≥ 60 days

 ≥ 100 days

No. (% vs. all patients)

113 (3%)

14 (< 1%)

3 (< 1%)

Ventilator days (% vs. all vent days)

5144 (23%)

1329 (6%)


Median age (years)




Primary diagnosis

Respiratory (55%)

Respiratory (57%)

Spinal (67%)

Survival to hospital discharge (%)

83 (73%)

12 (86%)

3 (100%)

Transfer to specialist weaning unit





Accuracy, consumption and NO2 production of inhaled NO administration devices: a comparative bench study

A Vuillermoz1, M Lefranc2, N Prouvez2, A Mercat1, JC Richard2, F Beloncle1

1CHU Angers, Service de Réanimation Médicale, Médecine Hyperbare, Angers, France, 2Air Liquide Medical Systems, Med2Lab, Antony, France

Critical Care 2023, 27(S1): P030

Introduction: Several inhaled nitric oxide (iNO) devices are available: basic continuous administration (MiniKINOX), “guided” (OptiKINOX) and “synchronized-to-ventilator” (NO-A, SoKINOX). We compared accuracy, consumption and NO2 formation in simulated ventilation conditions.

Methods: Four iNO devices were evaluated on a Michigan test lung (acute-respiratory-distress-syndrome model), connected to two ventilators (V800, Dräger and R860, GE Healthcare), seated at three different bias flows. Four NO concentration-settings were tested: 5, 10, 14 and 20 ppm. Actual NO and NO2 concentrations were measured in the test lung by NO and NO2 electrochemical cells (SoKINOX). NO–N2 mixture flow was measured by TSI-flowmeter to assess gas consumption.

Results: Actual NO concentration measured with “synchronized-to-ventilator” devices was more accurate than with “guided” and “basic continuous administration” devices in which NO concentration was mainly lower than the target and depended on the ventilator bias flow (Fig. 1). Despite differences in iNO administration, NO2 never exceeded the maximal safety threshold (1 ppm). “Synchronized-to-ventilator” devices which are synchronized to inspiratory triggers consumed less gas than “basic continuous administration” due to an adaptation of iNO administration. The consumption based on iNO concentration targeted was not lower with “synchronized-to-ventilator” than “guided” devices.

Conclusions: Last generation iNO devices, synchronized with ventilator insufflation and measuring actual flow in the inspiratory line, regulate iNO-delivery more precisely compared to older generation devices. Ventilator settings often blind for the clinician (eg. bias flow) may impact iNO accuracy.

figure l

Fig. 1 (abstract P030). Deviation from NO measured value to NO setting, expressed as percentage, from four devices, in three mechanical conditions (NA: Not applicable)


Bridging the gap: association of ventilator adjustments between the operating room and intensive care unit and 28-day mortality in surgical patients

D Von Wedel1, S Redaelli1, A Suleiman1, LJ Wachtendorf1, D Shay1, BA Azizi1, G Chen1, D Talmor1, EN Baedorf-Kassis2, MS Schaefer1

1Beth Israel Deaconess Medical Center, Harvard Medical School, Department of Anesthesia, Critical Care and Pain Medicine, Boston, USA, 2Beth Israel Deaconess Medical Center, Harvard Medical School, Department of Pulmonary, Critical Care & Sleep Medicine, Boston, USA

Critical Care 2023, 27(S1): P031

Introduction: Intensity of mechanical ventilation can be quantified through concepts such as mechanical power (MP), integrating driving pressure (DP), tidal volume (Vt), and respiratory rate (RR). Low intensity of ventilation during surgery and in the ICU has been associated with improved patient outcomes [1, 2]. However, there is a lack of information on the effect of ventilation adjustments when transitioning surgical patients to the ICU. We hypothesized that changes in MP from the operating room (OR) to the ICU are associated with 28-day mortality.

Methods: 1970 adult patients undergoing general anesthesia between 2008 and 2022, receiving continued, controlled mechanical ventilation in the ICU, were included in this retrospective study. Median MP was calculated for the last intraoperative hour and first six hours of ICU ventilation [3]. The association of MP changes from OR to ICU and 28-day mortality was assessed using multivariable logistic regression adjusted for patient demographics, comorbidities, and intraoperative factors. Relative importance of changes in individual parameters was determined by dominance analysis.

Results: 199 patients (10.1%) died within 28 days of surgery. Median (IQR) MP in the OR was 9.1 (6.7–11.7) and 9.7 (7.9–12.0) J/min in the ICU. Upon transition to ICU ventilation, Vt and DP were decreased, while RR was increased, and overall MP remained constant (Fig. 1A). An increase in MP from OR to ICU was associated with higher mortality (aOR 1.10 per 1 J/min increase, [1.05–1.14], p < 0.001, Fig. 1B). Dominance analysis revealed that this was primarily driven by increases in RR (Fig. 1C).

Conclusions: Adjustment of ventilator parameters leading to increased MP should be avoided when transitioning from OR to ICU. Clinicians should consider to avoid lowering Vt and DP at the cost of using high RR.


  1. 1.

    Santer P et al. Anesthesiology 2022;137:41–54.

  2. 2.

    Serpa Neto A et al. Intensive Care Med 2018;44:1914–1922.

  3. 3.

    Chiumello D et al. Crit Care 2020;24:417.

figure m

Fig. 1 (abstract P031). A Changes in MP and its components from the OR to ICU. B Association of parameter changes with 28-day mortality. Estimates for MP were obtained from the primary model. A secondary model was used for estimates of MP components, including changes of RR, DP, and Vt, as independent variables in a single model. C Dominance analysis was performed in the secondary model. Dominance statistics (pseudo-R2) are reported as determined by contribution to 28-day mortality prediction


Clinical value of prophylactic minitracheostomy in high-risk patients with acute cervical spinal cord injuries

S Maemura, M Suga, S Ishihara

Hyogo Emergency Medical Hospital, Department of Emergency and Critical Care Medicine, Kobe, Japan

Critical Care 2023, 27(S1): P032

Introduction: Cervical spinal cord injury (CSCI) is at high risk for sputum retention and respiratory exacerbation (RE). Prophylactic minitracheostomy (PMT), a technique to assist in removing airway secretions while maintaining the glottic function, allows repeated airway suctioning and can reduce incidents of intubation. However, The clinical value of PMT in acute CSCI patients is unclear. In our tertiary emergency center, PMT is performed on patients at high risk of sputum retention to prevent unexpected RE and intubation. This study aimed to evaluate the clinical value of PMT in acute CSCI patients.

Methods: We performed a retrospective review of patients with acute traumatic CSCI admitted to our hospital from April 2011 to April 2021. A total of 44 patients undergoing PMT were enrolled. The medical record was reviewed for age, ISS (Injury Severity Score), probability of survival (Ps), level of spinal cord injury, American Spinal Injury Association (ASIA) Impairment Scale (AIS), intubation because of RE, the success of decannulation, in-hospital mortality, complications related to the procedure.

Results: The median age was 67 years [interquartile range (IQR): 49–76]. The median ISS and Ps were 25 [IQR: 20–30], 0.86 [IQR: 0.69–0.94]. 33 patients out of 44 (75%) were classified as AIS grade A. 8 out of 44 (18.2%), all classified as AIS grade A, required emergency intubation at RE. That means PMT was ineffective in about one-fourth of severe CSCI patients (AIS grade A). One patient (2.3%) died in-hospital regardless of respiratory problems. One patient was intubated on the next day of PMT because of a tracheal hemorrhage. Other minor complications, including mild subcutaneous emphysema and bleeding, were found in a few patients, but all recovered quickly.

Conclusions: PMT for a CSCI patient with a high risk for sputum retention can prevent intubation, but we should keep in mind its failure, especially in severe CSCI patients. Identifying patients at risk of sputum retention who will benefit from PMT is important.


Does tracheostomy predict outcome in severe COVID-19?

R Pinto Medeiros1, M Bourbon Ruão2, B Vale 2, S Salvador3, A Novo2, R Pereira2, Á Moreira da Silva2

1Centro Hospitalar do Porto, Departamento de Anestesia e Cuidados Intensivos, Porto, Portugal, 2Centro Hospitalar do Porto, SCI, Porto, Portugal, 3Centro Hospitalar do Porto, Anesthesiology, Porto, Portugal

Critical Care 2023, 27(S1): P033

Introduction: The association of tracheostomy timing and clinical outcomes in ventilated COVID-19 patients remains controversial. Data from the pre-pandemic era has demonstrated the use of tracheostomy for ventilator weaning [1]. However, the use of tracheostomy in COVID-19 patients was a subject of discussion [2]. Nevertheless, evidence of the impact of tracheostomy on the outcome in critically ill COVID patients is still lacking. This study aims to evaluate the impact on Intensive Care Unit (ICU) outcome (survival) of tracheostomy in COVID-19 ventilated patients.

Methods: Monocentric descriptive observational study. Demographic and clinical data, timing of tracheostomy and outcome (ICU mortality) from 1st January to 31st December 2021 were registered. Analysis of descriptive statistics for continuous variables and survival analysis (log rank test).

Results: 115 patients were included (72% males), all mechanically ventilated, 7 (6%) were subjected to tracheostomy. The mean age was 67.2 years (range 36–84 years). The ICU mortality was 62% (71). The group of patients not submitted to tracheostomy had a mean survival time of 24.4 days (SD ± 1.5) and median survival time of 22 days (SD ± 1.7). The group of patients that were subjected to tracheostomy, the mean survival time was 68.5 days (SD ± 12.2) and median survival time was 50 days (SD ± 2). This comparison is significative (Log Rank test, p = 0.0001).

Conclusions: The present study demonstrates a better survival likelihood of the tracheostomized subpopulation. Tracheostomy was only done in 6% of patients, which elucidates a need to further prospective, randomized studies to assess the impact on the outcome of tracheostomy in ventilated COVID19 patients.


  1. 1.

    Trouillet JL et al. Ann Intensive Care 2018;8:37.

  2. 2.

    Ji Y et al. Crit Care 2022;26:40.


Early versus late tracheostomy in critically-ill COVID-19 patients

AS Szafran1, K Dahms1, K Ansems1, N Skoetz2, I Monsef2, T Breuer1, C Benstoem1

1Medical Faculty, RWTH Aachen University, Department of Intensive Care Medicine and Intermediate Care, Aachen, Germany, 2Department I of Internal Medicine, Faculty of Medicine and University Hospital Cologne, University of Cologne, Cochrane Haematology, Cologne, Germany.

Critical Care 2023, 27(S1): P034

Introduction: In the context of the COVID-19 pandemic and a qualitatively questionable evidence base, a thorough understanding of the available evidence for the management of critically ill patients with COVID-19 is needed. Previous reports have described ICU stays and weaning as prolonged in patients who developed COVID‐19 ARDS. Prolonged MV is the most common indication for tracheostomy in ICU patients. Pre-pandemic evidence on the benefits of early tracheostomy is conflicting but suggests shorter hospital stays and lower mortality rates compared to late tracheostomy.

Methods: We followed the standard Cochrane methodology to conduct the systematic review and meta-analyses. RCTs and observational studies evaluating concepts of early tracheostomy compared to late tracheostomy in critically-ill COVID-19 patients were included. Methodological quality was assessed using the Cochrane RoB 2 tool and ROBINS I tool. The quality of evidence in our prioritized categories was assessed using GRADE.

Results: We included 1 RCT with 150 participants and 24 observational studies with 6372 critically-ill COVID-19 patients (Fig. 1). RoB assessment for the RCT was considered to be low. RoB for the observational studies was considered to be critical.


  • We found low-certainty evidence that early tracheostomy may result in little to no difference in overall mortality in critically-ill COVID-19 patients requiring mechanical ventilation (MV). In terms of clinical improvement, early tracheostomy results in little to no difference in duration to liberation from MV. We are not certain about the impact on clinical worsening and ICU length of stay (LOS).

  • Our sensitivity analysis suggests that early tracheostomy may reduce the duration of invasive MV and the ICU LOS.

  • The overall evidence in this review is graded as low and can only be interpreted as a indication that early compared to late tracheostomy may be recommended to shorten the duration of MV in critically-ill COVID-19 patients and to reduce all-cause mortality.

figure n

Fig. 1 (abstract P034). Summary of findings


Tracheostomy management in spinal cord injured patients and the role of diaphragm pacing

RPO Onders, M Elmo, N Carl

University Hospitals Cleveland Medical Center, Department of Surgery, Cleveland, USA

Critical Care 2023, 27(S1): P035

Introduction: Tracheostomies have significant risks both for the patient and health care providers including: being an aerosol generating procedure, hemorrhage, tracheomalacia, infections, granulation tissue, mucous production, pneumonias and the most severe—obstruction and dislodgement leading to death or hypoxic brain damage. In spinal cord injured (SCI) patients 75% of all patients require intubation for mechanical ventilation (MV) acutely with 20% of all cervical SCI patient requiring tracheostomy MV. We report on diaphragm pacing (DP) to decrease acute and long term need for tracheostomies in SCI patients.

Methods: This is a retrospective analysis of an IRB approved prospective, non-randomized interventional experience at a single institution. Subgroup analysis was limited to endotracheal intubation and tracheostomy management of all patients with traumatic cervical SCI that were implanted laparoscopically with DP (NeuRx, Synapse Biomedical) within 30 days of injury to wean from MV.

Results: For the study group of early DP, there were 13 (all male) patients with average age at implant 49.3 years (range 17–70). Injury mechanisms included falls (6), MVA (4), GSW (2) and diving (1). One patient withdrew care 14 days post injury and is exclude from further analysis. 2 patients had tracheostomies before DP due to specific nature of their injuries. 3 patients had weak diaphragms with nerve injuries identified at surgery requiring prolonging weaning with tracheostomies. In total, 4 of 12 patients did not require tracheostomy (33%). For patients with intact diaphragms and no need for tracheostomies because of their injury-57% did not require tracheostomy (4 of 7). 3 additional patient(7 of 12 or 58%) had decanullation within 6 months of injury. Hospital length of stay for patients without tracheostomy was 19 days (range 19–27) and those requiring tracheostomies was 31 days (21–42).

Conclusions: Early DP allowed up to 57% of previously non-weanable patients to be extubated without tracheostomies.


Early versus late tracheostomy in multiple trauma patients in the intensive care unit: systematic review and meta-analysis

K Ansems, E Aleksandrova, AS Szafran, K Dahms, T Breuer, C Benstoem

Medical Faculty, RWTH Aachen University, Department of Intensive Care Medicine and Intermediate Care, Aachen, Germany

Critical Care 2023, 27(S1): P036

Introduction: The objective of this systematic review and meta-analysis is to assess the effects and safety of concepts of early tracheostomy (ET) compared to concepts of late tracheostomy (LT) in patients with multiple trauma on the ICU.

Methods: We followed standard Cochrane methodology. Randomized controlled trials (RCTs) and observational studies comparing ET with LT in adult patients with polytrauma in the ICU were included, regardless of disease severity, gender, ethnicity, or setting. Methodological quality was assessed using the Cochrane RoB 2 tool and ROBINS-I tool. The quality of evidence, in the categories we prioritized, was assessed using GRADE.

Results: We included one RCT (n = 60) and nine observational studies with a total of 1.694 multiple trauma patients who underwent tracheostomy. Statistical analysis of the RCT data showed that ET (≤ 8 days) was likely to have little or no effect on all-cause mortality. We found no differences in duration of mechanical ventilation, length of ICU stay, or number of patients with ventilator-associated pneumonia. Data from observational studies showed that the evidence is very uncertain about the effect of ET (≤ 7 days) on ICU mortality. ET (≤ 7 days) may reduce the duration of mechanical ventilation, ICU length of stay, VAP. ET (≤ 7 days) may have little to no effect on hospital length of stay but the evidence is very uncertain.

Conclusions: The evidence is very uncertain about the effect of ET performed ≤ day 7 on ICU mortality, reduction of the duration of MV, ICU length of stay, hospital length of stay and the incidence of VAP in patients with multiple trauma on the ICU. We strongly recommend performing high-quality (prospectively registered) observational studies and implementing the New Berlin definition for multiple trauma to ensure a more standardized patient population across studies. Additionally, we encourage authors to pool patients by similar injury patterns and provide information about the tracheostomy timing of the individual groups.


Tracheostomy outcomes in COVID-19 and non-COVID-19 critical patients

RV Jorge1, A Orfão2, J Pacheco Pereira3, J Patrício3, P Patrício3

1Hospital Beatriz Ângelo, Intensiva Medicine Department, Loures, Portugal, 2Hospital Professor Dr. Fernando Fonseca, Internal Medicina Department, Amadora, Portugal, 3Hospital Beatriz Ângelo, Intensive Medicine Department, Loures, Portugal

Critical Care 2023, 27(S1): P037

Introduction: Tracheostomy is a common surgical procedure in the setting of acute respiratory failure. And improves outcomes for critically patients requiring prolonged mechanical ventilation. Initially avoided due to it’s high risk to biosafety, tracheostomy soon became a routine procedure in the critical support of critical ill patients affected by COVID-19. The aim of this review was to compare tracheostomy done in COVID-19 and non-COVID-19 pneumonias in an UCI.

Methods: This retrospective, observational study included 78 patients (23 female, 55 male; age range: 23–90 years, mean age: 66) with severe pneumonia who were admitted to the intensive care unit (ICU) of Hospital Beatriz Ângelo (Portugal) between 01/03/2012 until 31/12/2021, to whom a tracheostomy was performed. Patients underwent orotracheal intubation with invasive mechanical ventilation, followed by percutaneous or open surgical tracheotomy. Indications, timing of the procedure, and time needed to complete weaning and decannulation, as well as complications, were reported and compared between patients with COVID-19 (N = 38) and non-COVID-19 (N = 40) pneumonias.

Results: In both groups, weaning from difficult ventilation was the most common indication for the procedure, followed by prolonged mechanical ventilation in the COVID-19 group (42%) and protection of the airway/secretions management in the non-COVID-19 group (22.5%). Timing of the procedure was 14.6 and 16.4 days after mechanical ventilation in the COVID-19 and non-COVID group, respectively. The non-COVID-19 group reported more days to decannulation (39.3 vs 15.1 days) as well as more days to wean off from mechanical ventilation (20.3 vs 14.1 days) and more major complications (12.5% vs 2%). Hospital discharge rate was similar in both groups (COVID-19 with 42.1% and 42.5% to non-COVID-19).

Conclusions: Although the differences between both groups are multifactorial, it’s useful for self-evaluation observations, as well as sharing practices and outcomes for further analysis.


Rapidly progressive brain atrophy in mechanically ventilated patients: a retrospective descriptive study using CT volumetry

R Nakae1, T Sekine2, T Tagami3, Y Murai4, G Warnock5, A Morita4, S Yokobori1

1Nippon Medical School Hospital, Department of Emergency and Critical Care Medicine, Tokyo, Japan, 2Nippon Medical School Musashi Kosugi Hospital, Department of Radiology, Kawasaki, Japan, 3Nippon Medical School Musashi Kosugi Hospital, Department of Emergency and Critical Care Medicine, Kawasaki, Japan, 4Nippon Medical School Hospital, Department of Neurological Surgery, Tokyo, Japan, 5PMOD Technologies Ltd., Zürich, Switzerland

Critical Care 2023, 27(S1): P038

Introduction: Mechanical ventilation has the potential to decrease cerebral blood flow; however, changes in brain volume associated with mechanical ventilation are not well understood. We assessed brain atrophy in mechanically ventilated patients using brain CT scans, and their findings’ relationship to outcomes.

Methods: We retrospectively analyzed mechanically ventilated patients admitted to an ICU who underwent at least two head CT scans during hospitalization. Patients with stroke, traumatic brain injury, or hypoxic encephalopathy were excluded. The first brain CT scan was routinely performed on admission, and the second and further brain CT scans were obtained whenever prolonged disturbance of consciousness was observed. Brain volume was estimated using an automatic segmentation method, and patients with a brain volume change < 0% from the first CT scan to the second CT scan were defined as the “brain atrophy group” and those with ≥ 0% were defined as the “no brain atrophy group.” Duration of mechanical ventilation, length of ICU stay, and mortality were compared between the groups.

Results: A total of 52 patients were included, 44 in the brain atrophy group and 8 in the no brain atrophy group. Analysis of the brain atrophy group showed a significant decrease in brain volume (first CT scan: 1.042 ± 0.124 L vs. second CT scan: 0.998 ± 0.123 L, t (43) = 9.705, p < 0.001). The mean change in brain volume in the brain atrophy group was − 43.5 cm3 over a median of 33 days. The median time on mechanical ventilation was 33 (IQR: 18–60) days and 25 (19–43) days, length of ICU stay was 30 (16–61) days and 30 (17–56) days, and mortality rate was 31.8% (14/44 cases) and 50.0% (4/8 cases) in the brain atrophy group and the no brain atrophy group, respectively.

Conclusions: Many ICU patients on mechanical ventilation who developed prolonged mental status changes showed signs of brain atrophy. Patients with rapidly progressive brain atrophy were more likely to have a longer duration of mechanical ventilation.


A quality improvement project to analyse the impact of an extubation checklist on ICU re-intubation rates and efficiency

PI Jenkinson, D Lau, J Sokhi, M Gabrel

Chelsea and Westminster Hospital, Intensive Care Medicine, London, UK

Critical Care 2023, 27(S1): P039

Introduction: Reintubation within 48 h of extubation, termed “failed extubation”, is associated with increased morbidity, mortality and length of ICU stay. An initial audit of our mixed adult ICU demonstrated a failed extubation rate of 23.3%, far higher than the recommended rate of 5–10% [1]. Our aim was to determine whether an extubation checklist would reduce reintubations and improve flow through the ICU.

Methods: All extubations at Chelsea and Westminster ICU during two 3-month periods, pre and post-intervention, were analysed. Patients who underwent primary tracheostomy, self-extubation, transfer to a different facility or extubation as part of end-of-life care were excluded. Our intervention consisted of a checklist, to be completed prior to extubation, which was based on current literature and case analyses of failed extubations in the first 3-month period. The primary outcome was rates of failed extubation and the secondary outcome was extubation time. Compliance with the intervention was determined by measuring rates of completed checklists and doctors were surveyed to assess reasons for non-compliance.

Results: Pre-checklist (Oct–Jan 2022) there were 30 primary extubations, of which 7 failed. Post-intervention (Feb–May 2022) there were 28 primary extubations, of which 2 failed. This showed that rates of failed extubation reduced by 16.2% following checklist introduction (Fig. 1) and on average, patients were extubated 1 h and 45 min earlier in the day. Compliance with the intervention was 39%. Our survey results showed that 55% of non-compliance was due to administration issues.

Conclusions: Our findings suggest that an extubation checklist is a simple and effective tool to reduce rates of failed extubation. Although compliance was low, our survey indicated that this was largely due to administrative issues which could be easily addressed. More research is warranted to assess the effect on patient mortality, morbidity and length of ICU stay.


  1. 1.

    Krinsley J et al. Crit Care. 2012;16:111.

figure o

Fig. 1 (abstract P039). Bar chart to compare the number of successful versus failed extubations pre and post checklist introduction


A quality improvement project (QIP) to introduce a standardised extubation checklist at the Royal Free Hospital ICU (phase 1)

LY Goh, A Carnachan, C Greenfield, M Bakir, A Walecka

Royal Free Hospital, Intensive Care Unit, London, UK

Critical Care 2023, 27(S1): P040

Introduction: We aim to introduce a standardised extubation checklist to determine suitability for extubation. We hope that with clearly defined parameters we can reduce extubation failure and reintubation, which is associated with worse outcomes and increased mortality [1].

Methods: Retrospective data was collected from 20 extubations in December 2021, which included associated airway risk factors, prior respiratory support requirements, and number of reintubations. A survey of nursing staff highlighted the need for clear guidance and protocol. A preliminary checklist was developed based on review of the Difficult Airway Society (DAS) guidelines [2]. After comparing with a pre-existing extubation standard operating procedure (SOP), this checklist was expanded on and approved for implementation. We will evaluate the use of this checklist by auditing documentation of its use, staff feedback and further assessment of reintubation rates (Phase 2).

Results: Data collected showed that six of the 20 patients (30%) studied were reintubated within 48 h of their initial extubation. Reasons for reintubation included stridor, epistaxis, drowsiness, and failed non-invasive ventilation (NIV) trial. The preliminary checklist created defined a list of airway risk factors and included steps from identifying readiness for extubation until post-extubation care. The SOP incorporated standard extubation criteria for use in all patients, and an extended criteria for complex patients as well as clearly defined targets for each parameter. The final checklist contains elements from both, and patients intubated for more than 10 days qualify for the extended criteria (Fig. 1).

Conclusions: This project highlights the importance of contributions from the multidisciplinary team to ensure the successful uptake and implementation of an intervention.


  1. 1

    Thille AW et al. Am J Respir Crit Care Med. 2013;187:1294–302.

  2. 2

    Difficult Airway Society. DAS extubation algorithm. 2011.

figure p

Fig. 1 (abstract P040) Standard and extended extubation checklists




Artificial neural network predicts extubation failure with uncertainty

YI Igarashi

Nippon Medical School, Department of Emergency and Critical Care Medicine, Tokyo, Japan

Critical Care 2023, 27(S1): P042

Introduction: Ventilator liberation is one of the most critical decisions in the intensive care unit; however, prediction of extubation failure is difficult, and the proportion thereof remains high. Machine learning can potentially provide a breakthrough in the prediction of extubation success.

Methods: This is a retrospective study using a two-center database. Using 27 features and data from 24 h prior to extubation, we developed models with and without uncertainty output using an artificial neural network consisting of three connected layers and one dropout layer. Five-fold cross validation was performed on each model to compare the accuracy of models with and without uncertainty output. Outcome was extubation failure, which was defined as reiintubation within 72 h after extubation.

Results: In MIMIC-III, 1203 patients underwent extubation, 1075 successfully and 128 unsuccessfully. In the NMS, 88 patients were extubated, 77 successfully and 11 unsuccessfully. The accuracy of the model with uncertainty was higher than that of the model without incertainty. In addition, the accuracy increased with external validation, with accuracy 0.51, precision 0.17, recall 0.73, negative predictive value 0.93, and F1 score 0.27.

Conclusions: An ANN model with uncertainty was developed to predict extubation, and the sensitivity of the model was increased. This model is expected to reduce the extubation faliure.


Greater concentration of N-methyl-d-aspartate receptor in the brainstem in moderate-ARDS pigs ventilated for 12 h with diaphragm neurostimulation

T Bassi1, E Rohrs2, J Witmann2, K Fernandez2, M Nicholas2, M Orowska2, M Gani1, D Evans1, S Reynolds2

2Lungpacer Medical Inc., Burnaby, Canada, 2Royal Columbian Hospital, New Westminster, Canada

Critical Care 2023, 27(S1): P043

Introduction: Studies have shown that mechanical ventilation (MV) and acute respiratory distress syndrome (ARDS) are associated with hippocampal cellular apoptosis. NR2A is an N-methyl-d-aspartate receptor linked to cellular survival and death. A greater concentration of synaptic NR2A is associated with cell survival and improved working memory. Temporary transvenous diaphragm neurostimulation (TTDN) in association with MV has been shown to mitigate hippocampal apoptosis in preclinical models. This study focused on investigating whether 12 h of MV with moderate ARDS would lead to a change in NR2A tissue concentration in the brainstem and whether diaphragm neurostimulation would affect that.

Methods: Juvenile pigs (4–5 months, 50–87 kg) underwent protective MV (volume control, PEEP 5 cmH2O, tidal volume 8 ml/kg) for 12 h, after being induced with moderate ARDS by injecting oleic acid into the pulmonary artery until PaO2/FiO2 was between 100 and 200. Subjects were assigned to three groups (n = 6 per group): lung injury with MV only (MV), lung injury with MV and with TTDN every other breath (MV + TTDN50%), and lung injury with MV and with TTDN every breath (MV + TTDN100%). Diaphragm neurostimulation was delivered using methods published previously. After the study, the brainstem (pons and medulla oblongata) was harvested and marked with enzyme-linked immunoassay to measure the tissue concentration of NR2A. The Kruskal–Wallis test was used for statistical analyses. p values < 0.05 are considered statistically significant.

Results: Brainstem NR2A concentrations were 4 pg/ml (3–5) for the MV group, 5 pg/ml (5–8) for the MV + TTDN50% group, and 7 pg/ml (6–8) for the MV + TTDN100% group, with a significant difference between the MV and the MV + TTDN100% groups, p = 0.0125 (Fig. 1).

Conclusions: In a preclinical model of moderate ARDS, diaphragm neurostimulation on every breath in association with MV for 12 h led to significantly greater concentrations of NR2A in the brainstem, compared to the group receiving MV, only, for 12 h.

figure q

Fig. 1 (abstract P043) Brainstem NR2A concentrations


Heart rate fluctuations and mortality in patients with acute respiratory failure: a retrospective analysis of the MIMIC-III database

L Tao1, J. Liu 2, C. Li1, Y Pan1

1The Affiliated Suzhou Hospital of Nanjing Medical University, Suzhou Municipal Hospital, Suzhou Clinical Medical Center of Critical Care Medicine, Suzhou, China, 2The Affiliated Suzhou Hospital of Nanjing Medical University, Suzhou Municipal Hospital, Suzhou Clinical Medical Center of Critical Care Medicine, Department of Critical Care Medicine, Suzhou, China

Critical Care 2023, 27(S1): P044

Introduction: There is limited evidence for an association between heart rate fluctuations in patients with acute respiratory failure and 28-day mortality. Therefore, the objectives of this paper were to investigate if changes in heart rates are connected with 28-day mortality among patients with acute respiratory failure after controlling for other factors.

Methods: A total of patients 6719 with acute respiratory failure was recruited into the Medical Information Mart database for Intensive Care (MIMIC)-III in this investigation, which was conducted as a cohort study. Within 24 h of admission, variations in heart rate were estimated as the difference between maximum heart rate and minimum heart rate. Individuals were divided into three groups: those with low heart rate fluctuations [less than thirty beats per minute], those with central rate fluctuations (30–49 beats per minute), and those with high heart rate fluctuations (more than fifty beats per minute). We used models of multivariate logistics regression, Generalized additive models and linear regression models of two segments to examine the influence of 24-h heart rate variations on death on the 28 day.

Results: Sixty-seven hundred and ninety patients suffering from acute respiratory failure were included in the study. The fluctuations in heart rates were independent due to 28-day mortality (dominance ratio [OR] 1.63; 95% confidence interval [95% CI (1.4–1.89); p < 0.001) adjusted for possible confusion. There was a non-linear association between heart rate fluctuations and 28-day mortality with a threshold of approximately 21 beats per minute. The effects sizes and ci were 0.98 (0.96, 1.00) and 1.01 (1.00, 1.01) respectively.

Conclusions: In patients suffering from acute respiratory failure, there is a non-linear association between heart rate fluctuations over 24 h and 28-day mortality, with a particularly marked increase in death on the 28th day when heart rate fluctuations exceed 21 beats per minute.




Mechanical power in patients receiving mechanical ventilation in the surgical intensive care unit and the association with increased mortality

A Piriyapatsom, A Trisukhonth, O Chintabanyat

Department of Anesthesiology, Faculty of Medicine Siriraj Hospital, Mahidol University, Anesthesiology, Bangkok, Thailand

Critical Care 2023, 27(S1): P046

Introduction: Mechanical power (MP) represents the amount of energy delivered by mechanical ventilation (MV) to overcome elastic static, elastic dynamic and resistive components of the respiratory system during each inspiration. Evidences show the correlation between the MP and clinical outcomes in mechanically ventilated patients. The aim of this study was to determine the MP and the associated outcomes in patients admitted to surgical ICU (SICU) and receiving MV for more than 12 h.

Methods: This was a retrospective analysis of the prospective observation cohort study. Patients whose age ≥ 18 years old admitted to the SICU and receiving MV for at least 12 h were included. Demographic and clinical data were recorded. Ventilator parameters at the initiation and at 24 h of MV support were collected and were used to calculate the MP. Cox regression was analyzed to determine the association between the MP and 90-day mortality.

Results: There were 306 patients included in this study. Overall, 90-day mortality was 19.3%. The MP at the initiation and at 24 h of MV was significantly higher in non-survivors (median 11.00 (IQR 7.89–14.18) J/min vs. 7.97 (6.51–10.42) J/min and 8.86 (7.01–12.55) J/min vs. 7.24 (5.61–9.43) J/min, both P < 0.001, respectively). The Cox regression analyses demonstrated that MP ≥ 10 J/min at the initiation of MV and MP ≥ 10 J/min both at the initiation and at 24 h of MV were associated with increased 90-day mortality (HR 1.855, 95% CI 1.074–3.203 and HR 1.906, 95% CI 1.080–3.361, respectively). In subgroup of patients with MP ≥ 10 J/min at the initiation of MV support, further increasing in MP at 24 h of MV was associated with increased 90-day mortality (HR per J/min 1.093, 95% CI 1.007–1.187).

Conclusions: In mechanically ventilated patients in the SICU, the MP was an independent factor associated with 90-day mortality. Change in the MP during MV support potentially affected 90-day mortality in these patients.


The impact of prone episodes on oxygenation in COVID-19 mechanically ventilated patients

F Neyroud, A Jackson, J Barnsely, E Hunter, M Radharetnas, R Beecham, A Dushianthan

University Hospital Southampton, Intensive Care, Southampton, UK

Critical Care 2023, 27(S1): P047

Introduction: Prone positions have been used extensively to improve oxygenation in patients with acute respiratory distress syndrome (ARDS). During the COVID-19 pandemic there was widespread adoption of proning in patients with acute severe hypoxic respiratory failure. Few studies explore the use of prone positioning in mechanically ventilated COVID-19 patients.

Methods: This study was part of the REACT COVID observational study at University Hospital Southampton (UHS) [1]. Eligibility included admission to UHS with a positive COVID-19 RT-PCR between 03/2020 and 03/2022. Data was collected from all available electronic clinical data sources using semi-automated and manual data extraction.

Results: 184 patients received invasive mechanical ventilation with documented evidence for 931 prone episodes. We performed detailed analysis for 763 prone episodes. The rest were excluded due to insufficient data. The median duration of each cycle was 16 h (IQR 15–17 h). 459 cycles were done within 7 days of intubation (early), 202 in 7–14 days (intermediate) and 102 after 14 days (late). The change in oxygenation defined as delta PaO2/FiO2 ratio (ΔPF) for early, intermediate, and late cycles were 2.4 ± 5.2 kPa, 1.6 ± 3.7 kPa and 1.4 ± 4.0 kPa, (p = 0.03) respectively. The overall ΔPF for all groups after a cycle was 2.1 ± 4.7 kPa. There was an increase in PaCO2 following proning with an overall change of 0.30 ± 1.0, however, this was not statistically significant (p = 0.30).

Conclusions: Following proning, there was significant improvement in oxygenation. Cycles lasted for 16 h consistent with current ARDS guidelines [2]. Although the results suggest a diminishing response in those proned at later times, the ΔPF ratio was still significant. Overall, this suggests a beneficial effect on oxygenation. However, findings cannot be translated into survival benefit. Further research including randomised controlled trials is recommended.


  1. 1.

    Burke H et al. BMJ Open. 2021;11:e043012.

  2. 2.

    Guérin C et al. N Engl J Med. 2013;368:2159–68.


Association between pulmonary extravascular water, dead space, and clinical outcome in intubated patients with COVID ARDS

S Leighton1, M Pontony2, R Pérez2, J Graf2, R López2

1Clínica Alemana de Santiago, Intensive Care Unit, Santiago, Chile, 2Clínica Alemana de Santiago, Departamento Paciente Crítico, Santiago, Chile

Critical Care 2023, 27(S1): P048

Introduction: The aim of this study was to determine whether there is an association between extravascular lung water index (EVLWi) and physiological respiratory dead space (VDphys/VT) and to determine if these factors are associated with the possibility to being discharged alive on day 28.

Methods: We analyzed a prospective cohort of patients with COVID ARDS supported with invasive mechanical ventilation (IMV) admitted in our ICU who were monitored with volumetric capnography and transpulmonary thermodilution (TPTD). First day TDTP and VDphys/VT were considered. Bohr-Enghoff formula was used to obtain VDphys/VT. This protocol was approved by the local IRB and informed consent was waived.

Results: 31 patients with matched TPTD and VDphys/VT during the first 24 h were analyzed in who EVLWi correlated with VDphys/VT (r = 0.599 p = 0.002), however, EVLWi did not associated with PaFi. Patients with EVLWi > 10 ml/kg had higher APACHE II and VDphys/VT. These patients had a lower cumulative incidence to be discharged alive on day 28 with aHR 7.3 [1.4–39.1] p = 0.02 (adjusted by APACHE II and VDphys/VT, Fig. 1A). Remarkably, patients with EVLWi > 10 ml/kg + VDphys/VT > 57% had worse outcome compared to those who had EVLWi > 10 ml/kg + VDphys/VT < 57% (25% vs 75%, p = 0.032, Fig. 1B).

Conclusions: In patients with COVID ARDS supported with IMV, VDphys/VT give prognostic data additional to EVLWi.

figure s

Fig. 1 (abstract P048). A: Outcome comparison between EVLWi < 10 ml/kg and > 10 ml/kg adjusted by APACHE II and VDphys/VT. B: outcome comparison in patients with EVLWi > 10 ml/kg + VDphys/VT < 57% versus patients with EVLWi > 10 ml/kg + VDphys/VT > 57%


Effect of positive end-expiratory pressure (PEEP) on the distance of pleura to subclavian vein and on the cross-sectional area of the subclavian vein in mechanically ventilated patients

Y Vercammen, E Mangelschots, P Vets

ZNA Middelheim, Intensive Care, Antwerp, Belgium

Critical Care 2023, 27(S1): P049

Introduction: Pneumothorax is a well-known complication of subclavian vein (SCV) catheterization because of the adjacent localization of the pleura [1]. Although there are no strict recommendations on the use of PEEP during catheterization, there is a common clinical practice of reducing PEEP to 0 cm H2O in order to prevent a pleural puncture. We investigated the effect of applying PEEP on the distances from SCV to pleura (DSCV-Pleura) as well as the effect of PEEP on the cross-sectional area of the SCV (CSASCV). We hypothesize that administration of PEEP will not affect the DSCV-Pleura thereby not increasing the risk of pneumothorax. In addition, applying PEEP may facilitate SCV catheterization by increasing the CSASCV.

Methods: We included 20 adult patients admitted at the ICU of ZNA Antwerp in 2022 following cardiac surgery who were mechanically ventilated. Using ultrasound we measured the cross-sectional diameter of the SCV (CSDSCV) and DSCV-Pleura after administration of PEEP 0 till 15 cm H2O. Differences of CSDSCV and DSCV-Pleura after applying PEEP were assessed with repeated measures ANOVA and paired sample t-test.

Results: There was no significant difference of DSCV-Pleura between PEEP 0 and PEEP 15 (p = 0.336). Measurements of CSDSCV showed no significant difference of CSASCV between PEEP 0 and PEEP 15 (p = 0.085) (Table 1).

Conclusions: The administration of PEEP did not affect the DSCV-Pleura neither the CSASCV, thereby questioning the common clinical practice of reducing PEEP to 0 cm H2O during SCV catheterization. Our findings suggest we could maintain a level of PEEP during SCV catheterization in mechanically ventilated patients with the aim of preventing alveolar derecruitment and desaturation [2]. Moreover, there was no proof of facilitation of catheterization by applying PEEP. Further research could help us to confirm these findings and improve medical practice.

Table 1 (abstract P049) CSDSCV and DSCV-pleura after applying PEEP


DSCV-Pleura (mm)



7.70 ± 3.03 (6.73–8.70)

10.26 ± 2.14 (9.58–10.95)


7.38 ± 2.38 (6.62–8.14)*

9.90 ± 2.43 (9.12–10.68)**

  1. Data of DSCV-pleura and CSDSCV are shown as mean ± SD (95% confidence interval). CSD = cross-sectional diameter; SCV = subclavian vein; DSCV-pleura = distance between the subclavian vein and pleura. PEEP 0 = PEEP 0 cmH2O, PEEP 15 = PEEP 15 cmH2O. CSA (cross-sectional area) was calculated with the formula: CSA = r2 x Π (radius r = CSD/2). The paired sample t-test was used to compare means of DSCV-pleura and CSDSCV between PEEP 0 and PEEP 15. A p value of < 0.05 was considered to be statistically significant. *p = 0.336 > 0.05, **p = 0.084 > 0.05


  1. 1.

    McGee DC et al. N Engl J Med 2003;348:1123–3.

  2. 2.

    Martinsson A et al. Br J Anaesth 2021;126:1067–74.


Association between lactate level in bronchoalveolar lavage fluid and outcome and regional pulmonary inflammation in acute respiratory distress syndrome

XL Liu, WC Chang

Jiangsu Provincial Key Laboratory of Critical Care Medicine, Department of Critical Care Medicine, Zhongda Hospital, School of Medicine, Southeast University, Nanjing, China, Nanjing, China

Critical Care 2023, 27(S1): P050

Introduction: Altered metabolism and inflammation was reported to be associated with activated glycolysis and induced production of lactate. Although as a hallmark in the development of sepsis, the lactate concentration in bronchoalveolar fluid (BALF) in acute respiratory distress syndrome (ARDS), which was characterized by pulmonary endothelial injury due to inflammation, has not be widely investigated.

Methods: BALF were collected from patients with ARDS and post-operative, non-ARDS, patients. Lactate and inflammatory mediators including interleukin (IL)-1beta, IL-6, IL-4, IL-10, endothelial adhere molecules including von Willebrand Factor (vWF) and nitric oxide (NO), were evaluated in BALF supernatant by fluorescent immunoassays. Mechanical ventilation and ICU stay days, and mortality was also recorded. All the patients were followed-up to 90 days with dead or alive recorded.

Results: A total of 28 patients with ARDS and 8 patients with non-ARDS were enrolled in this study. The results indicated that the normalized BALF lactate level in ARDS patients was significantly higher compared with non-ARDS patients (mmol/l, 20.6[13.8, 34.3] vs. 11.04[5.2, 21.4], p = 0.048), IL-1beta (ng/ml, 3277[653.5, 13865] vs. 537.7[337.9, 780.6], p = 0.027), IL-6 (ng/ml, 1798[689, 4136] vs. 65[48, 203], p < 0.001), IL-10 (mmol/l, 33.8[24.3, 68.2] vs. 14.6[7.3, 19.1], p < 0.001). BALF lactate was also associated with mortality in ARDS patients (HR 4.8, 95% CI 0.9–23.7, log-rank p = 0.0167), by a cut-off value of adjusted BALF lactate.

Conclusions: BALF lactate was elevated in ARDS patients, and was associated with inflammatory cytokines, and endothelial adherence molecules. BALF lactate was also associated with mortality in ARDS patients.


Reduced concentration of IL-8 in the brainstem in moderate-ARDS pigs ventilated for 12 h with diaphragm neurostimulation

T Bassi1, E Rohrs2, K Fernandez2, M Ornoskwa2, M Nicholas2, J Witmann2, M Gani1, D Evans1, S Reynolds2

1Lungpacer Medical Inc., Burnaby, Canada, 2 Royal Columbian Hospital, New Westminster, Canada

Critical Care 2023, 27(S1): P051

Introduction: Studies have shown that mechanical ventilation (MV) and acute respiratory distress syndrome (ARDS) are associated with neuroinflammation.IL-8 is a proinflammatory cytokine associated with cognitive impairment. Temporary transvenous diaphragm neurostimulation (TTDN) in association with MV has been shown to mitigate hippocampal neuroinflammation and apoptosis in preclinical models. This study focused on investigating whether 12 h of MV with moderate ARDS would lead to neuroinflammation, measured as IL-8 tissue concentration in the brainstem, and whether diaphragm neurostimulation in association with MV would affect IL-8 brainstem tissue concentration.

Methods: Juvenile pigs (4–5 months, 50–87 kg) underwent protective MV (volume control, PEEP 5 cmH2O, tidal volume 8 ml/kg) for 12 h, after being induced with moderate ARDS by injecting oleic acid into the pulmonary artery until PaO2/FiO2 was between 100 and 200. Subjects were assigned to three groups (n = 6 per group): lung injury with MV only (MV), lung injury with MV and with TTDN every other breath (MV + TTDN50%), and lung injury with MV and with TTDN every breath (MV + TTDN100%). Diaphragm neurostimulation was delivered using methods published previously. After the study, the brainstem (pons and medulla oblongata) was harvested and marked with ELISA to measure the tissue concentration of IL-8. The Kruskal–Wallis test was used for statistical analyses. p values < 0.05 are considered statistically significant.

Results: Brainstem IL-8 concentrations were 36 pg/ml (28–48) for the MV group, 29 pg/ml (18–64), for the MV + TTDN50% group, and 0 pg/ml (0–18) for the MV + TTDN100% group, with a significant difference between the MV and the MV + TTDN100% groups, p = 0.0319 (Fig. 1).

Conclusions: In a preclinical model of moderate ARDS, diaphragm neurostimulation on every breath in association with MV for 12 h led to significantly lower concentrations of IL-8 in the brainstem, compared to the group receiving MV, only, for 12 h.

figure t

Fig. 1 (abstract P051). Brainstem IL-8 concentrations


Alveolar macrophage-derived secretory autophagosomes exacerbate inflammation in lipopolysaccharide-induced acute respiratory distress syndrome

X Xu, L Liu, Y Yang

Jiangsu Provincial Key Laboratory of Critical Care Medicine, Department of Critical Care Medicine, Zhongda Hospital, School of Medicine, Southeast University, Jiangsu Provincial Key Laboratory of Critical Care Medicine, Department of Critical Care Medicine, Zhongda Hospital, School of Medicine, Southeast University, Nanjing, China

Critical Care 2023, 27(S1): P052

Introduction: This study investigated the effects and the mechanisms through which alveolar macrophage(AM) -derived secretory autophagosomes (SAPs) contribute to acute respiratory distress syndrome(ARDS)-associated lung injury and inflammation.

Methods: SAPs were isolated from cell culture supernatants of AMs treated or untreated with lipopolysaccharide (LPS) and intratracheally injected to determine whether SAP exacerbated lung injury in mice with ARDS. The SAP proteome was analysed using mass spectrometry. Lung fibroblasts were treated with SAPs and collected for transcriptome sequencing. A transwell assay was conducted to study the chemotactic activity of fibroblasts induced by SAPs.

Results: We found that AMs contributed to ARDS-associated lung injury by releasing a novel type of pro-inflammatory vesicles named SAPs, which were characterised as double-membraned vesicles approximately 200 nm in diameter and light-chain-3 expression. Proteomic analysis of SAPs and gene ontology (GO) enrichment analysis revealed that LPS-SAPs contain several differentially expressed proteins involved in effecting an inflammatory response. In addition, we conducted transcriptome sequencing of lung fibroblasts treated with SAPs to explore the effect of SAPs on fibroblasts in ARDS, which demonstrated that the expression of several chemokines (chemokine ligand (CXCL)2, CXCL15, and so forth) significantly increased after SAP stimulation. Furthermore, the transwell assay indicated that SAPs could enhance the chemotactic activity of fibroblasts, suggesting that SAPs may exacerbate inflammation in ARDS by promoting neutrophil or monocyte recruitment.

Conclusions: This is the first study to show that AM-derived SAPs contribute significantly to ARDS through promoting inflammatory cell recruitment by activating fibroblasts, which may be beneficial for therapeutic development of ARDS.


Comparison of effect of chest percussion and expiratory rib cage compression on blood gases and static lung compliance in mechanically ventilated patients with acute respiratory distress syndrome: a randomised crossover study

A Srivastava1, R Kumar 1, M Gurjar 1, B Poddar 1, A Azim 1, S Gutte 1, P Mishra 2, V Kapoor2

1Sanjay Gandhi Postgraduate Institute of Medical Sciences, Critical Care Medicine, Lucknow, India, 2Sanjay Gandhi Postgraduate Institute of Medical Sciences, Biostatistics, Lucknow, India

Critical Care 2023, 27(S1): P053

Introduction: This study compared the effect of two techniques of chest physiotherapy, chest percussion (CP) and expiratory rib cage compression (ERCC), on blood gases (PaO2 and PaCO2), static lung compliance (Cstat) and secretion yield in mechanical ventilated patients with moderate-severe acute respiratory distress syndrome (ARDS).

Methods: After institutional ethics committee approval, all adult ICU patients with moderate-severe ARDS on invasive ventilation were screened for this randomized crossover study (CTRI/2022/01/039650). Exclusion criteria were: age < 18/ > 65 years, suspected/confirmed raised intracranial pressure, acute coronary syndrome, arrhythmias, abdominal compartment syndrome, pneumothorax or pregnancy. Included patients received both the techniques on the same day, six hours apart. PaO2 and PaCO2 (mmHg) and Cstat (ml/cmH2O) were obtained at baseline, 20 and 60 min after each physiotherapy technique. Tracheal secretion yield measurement was also done after both techniques.

Results: During study period (Jan–Oct 2022), 212 patients screened and 19 included with median age 49 (33–59) years and 9 were males. On the study day, median SOFA score was 10 (8–10) and median PaO2/FiO2 ratio was 155 (130–174). Before CP technique, baseline PaO2, PaCO2 and Cstat were 80, 41.9 and 24.8 which were changed at 20 min -2.3 (p = 0.11), 0.2 (p = 0.11) and -1 (p = 0.39); and at 60 min -0.4 (p = 0.79), -1.3 (p = 0.27) and 0.7 (p = 0.26) respectively. While baseline values in ERCC were 75.9, 41.5 and 24.3 which changed at 20 min 0 (p = 0.79), 0.6 (p = 0.39) and 0.2 (p = 0.8); and at 60 min 1.1 (p = 0.23), 0.6 (p = 0.43) and 1 (p = 0.26) respectively.The median volume of secretion yield (ml) was 4.0 (3.0–6.0) and 5.0 (3.5–6.0) (p = 0.88), respectively.

Conclusions: In moderate-severe ARDS patients, both CP and ERCC techniques did not affect arterial blood gases and static lung compliance significantly at either 20 or 60 min (Fig. 1). Also the secretion yield was similar after both techniques.

figure u

Fig. 1 (abstract P053). Primary outcomes


Postoperative acute respiratory distress syndrome in therapeutic randomized controlled trials

VG Giannakoulis, E Papoutsi, A Kotanidou, II Siempos

First Department of Critical Care Medicine and Pulmonary Services, Evangelismos Hospital, National and Kapodistrian University of Athens Medical School, Athens, Greece

Critical Care 2023, 27(S1): P054

Introduction: Large studies to inform the epidemiology of postoperative acute respiratory distress syndrome (ARDS) are scarce. We, therefore, aimed to perform a secondary analysis of seven high-quality therapeutic randomized controlled trials.

Methods: A secondary analysis of participants enrolled in 7 trials of ARDS and PETAL Networks [1–7].

Results: We analyzed 5316 patients with ARDS, including 256 with postoperative ARDS. We found that prevalence of postoperative ARDS gradually declined between 2000 and 2011, and it was stabilized afterwards (R2 = 0.821). Patients with postoperative ARDS, were older (p < 0.001), had more frequently sepsis (p < 0.001) and multiple transfusions (p < 0.001) as ARDS risk factors and presented lower 90-day mortality when compared to patients with ARDS not related to surgery [24.6% versus 30.9%, adjusted hazard ratio 0.66, 95% confidence intervals (CI) 0.51–0.85, p = 0.001]. In adjusted analysis, older age (p < 0.001), immunosuppression (p = 0.009), and increased cumulative fluid balance during the first 3 days (p < 0.001) were factors associated with 90-day mortality of patients with postoperative ARDS.

Conclusions: Our findings put the epidemiology and clinical features of postoperative ARDS in perspective, and therefore, fill a missing gap in the literature. We also reveal that increased fluid balance (a potentially modifiable, practice-dependent factor) is associated with higher 90-day mortality in postoperative ARDS. Therefore, patients with postoperative ARDS may benefit from a conservative fluid management strategy.


  1. 1.

    Brower RG et al. N Engl J Med. 2000;342:1301–8.

  2. 2.

    Brower RG et al. N Engl J Med. 2004;351:327–36.

  3. 3.

    Wiedemann HP et al. N Engl J Med. 2006;354:2564–75.

  4. 4.

    Matthay MA et al. Am J Respir Crit Care Med. 2011;184:561–8.

  5. 5.

    Rice TW et al. JAMA. 2012;307:795–803.

  6. 6.

    Truwit JD et al. N Engl J Med.2014;370:2191–200.

  7. 7.

    Moss M et al. N Engl J Med.2019;380:1997–2008.


Ultra lung protective ventilation and biotrauma in severe ARDS patients on veno-venous extracorporeal membrane oxygenation: a randomized controlled study

C Guervilly1, T Fournier2, J Chommeloux3, S Valera2, R Lacroix4, S Hraiech2, A Roch2, M Schmidt3, L Papazian5

1Médecine Intensive Réanimation, Hôpital Nord, APHM, Intensive Care Unit, Marseille, France, 2Médecine Intensive Réanimation, Hôpital Nord, APHM, Marseille, France, 3Service de Médecine Intensive-Réanimation, Institut de Cardiologie, APHP, Sorbonne Université Hôpital Pitié- Salpêtrière, Paris, France, 4Laboratoire d’Hématologie et de Biologie Vasculaire, APHM,Université Aix-Marseille, INSERM 1263, Institut National de Recherche pour l’Agriculture, l’Alimentation et l’Environnement, Centre de Recherche en CardioVasculaire et Nutrition, Marseille, France, 5Centre Hospitalier de Bastia, Service de Réanimation, Bastia, France

Critical Care 2023, 27(S1): P055

Introduction: Ultra-lung-protective ventilation may be useful during veno-venous extracorporeal membrane oxygenation (vv-ECMO) for severe acute respiratory distress syndrome (ARDS) to minimize ventilator-induced lung injury and to facilitate lung recovery. Objective was to compare pulmonary and systemic biotrauma evaluated by numerous biomarkers of inflammation, epithelial, endothelial injuries, and lung repair according to two ventilator strategies on vv-ECMO.

Methods: Prospective randomized controlled study. Patients were randomized to receive during 48 h either ultra-lung-protective ventilation combining very low tidal volume (1–2 ml/kg of predicted body weight), low respiratory rate (5–10 cycles per minute), positive expiratory transpulmonary pressure, and 16 h of prone position or lung-protective-ventilation which followed the ECMO arm of the EOLIA trial (control group) (Fig. 1).

Results: The primary outcome was the alveolar concentrations of interleukin-1-beta, interleukin-6, interleukin-8, surfactant protein D, and blood concentrations of serum advanced glycation end products and angiopoietin-2 48 h after randomization. Enrollment was stopped for futility after the inclusion of 39 patients. Tidal volume, respiratory rate, minute ventilation, plateau pressure, and mechanical power were significantly lower in the ultra-lung-protective group. None of the concentrations of the pre-specified biomarkers differed between the two groups 48 h after randomization. However, a trend to higher 60-day mortality was observed in the ultra-lung-protective group compared to the control group (45 vs 17%, p = 0.06).

Conclusions: Despite a significant reduction of the mechanical power, ultra-lung-protective ventilation during 48 h did not reduce biotrauma in patients with vv-ECMO-supported ARDS. The impact of this ventilation strategy on clinical outcomes warrants further investigation.

figure v

Fig. 1 (abstract P055). Study design


Pulmonary endothelial cells-derived extracellular vesicles promote neutrophil extracellular trap formation during ARDS

SF Z, XJ Wu, HB Qiu

Zhongda Hospital, School of Medicine, Southeast University, Department of Critical Care Medicine, Nanjing, China

Critical Care 2023, 27(S1): P056

Introduction: The acute respiratory distress syndrome (ARDS) is a common cause of respiratory failure in critically ill patients which causes high mortality, but its molecular mechanisms are poorly understood. Neutrophil extracellular traps (NETs) originate from decondensed chromatin released to immobilize pathogens, and they can trigger contribute to sustained inflammation. However, their role in NET formation remains unclear. Extracellular vesicles (EVs) mediate intercellular communication between cells in physiological and pathological conditions. Our previous studies have demonstrated that EVs released during severe bacterial pneumonia were inflammatory and induced acute lung injury. Moreover, Endothelial cell (EC)-derived EVs increased markedly. Study has shown that EVs sustain NET formation. However, whether EC-derived EVs play a role in NET formation in ARDS has yet to be addressed.

Methods: Polymorphonuclear neutrophils (PMNs) were cocultured with EVs isolated from the supernatant of pulmonary microvascular ECs stimulated with phosphate buffer saline (PBS) or lipopolysaccharide (LPS). A cecal ligation and puncture (CLP) mouse model was used to mimic septic ARDS in vivo; then, NET formation and molecular pathways were detected by transcriptome of EVs-Treated PMNs and mass spectrometry of EVs.

Results: In the animal CLP model, EVs depletion reduced NET formation, lung inflammation and injury. NET components were significantly increased in response to treatment with PMECs-derived EVs in vitro and vivo. NETs inhibition reduced lung inflammation and lung injury after EVs injection intravenously. Mechanistic studies demonstrated that exosomal mitochondrial associated protein induced NET formation through the Fpr1 mediated-NF-KB pathway.

Conclusions: EC-derived EVs promote excessive NET formation and aggravate inflammation and subsequent lung injury. This finding suggests a previously unidentified role of EC-derived exosomes in ARDS and may lead to new therapeutic approaches.


Macrophage-derived LC3 positive extracellular vesicles induce lung microvascular endothelial cell- mediated excessive inflammatory by transferring tRF-1024

X Zhu, L Liu

Southeast University, Jiangsu Provincial Key Laboratory of Critical Care Medicine, Nanjing, China

Critical Care 2023, 27(S1): P057

Introduction: Acute respiratory distress syndrome (ARDS) is an unsolved clinical conundrum globally due to its elusive pathogenesis. Pulmonary microvascular endothelial cells (PMVECs) activation and injury are the hallmarks of ARDS pathogenesis. Thus, the elucidation of endothelial activation mechanisms might uncover new approaches for treating ARDS. Extracellular vesicles (EVs) have emerged as an important means of cell–cell signaling by loading nucleic acid and other cargo. tRNA-derived small RNA (tsRNA) is an emerging class of small non-coding RNA molecules that involve in the regulation of multiple physiological processes and diseases. This study describes macrophages-derived LC3+EV mediate PMVECs activation in ARDS and how this process is regulated by LC3+EVs carried-tsRNA.

Methods: Flow cytometry absolute count standard was used to quantify LC3+EV in alveolar lavage fluid of patients from ARDS and control groups. Intratracheal instillation model was applied to investigate the effects of endothelial function activation with LC3+EV treatment. RNA sequencing was used to identify the critical tsRNA related to endothelial function. Further evaluation of tsRNA functional was determined through transfecting mimics.

Results: Clinically, the number of LC3+EVs had a good predictive value for 28-day mortality of ARDS patients, indicating that LC3+EV is a critical mediator of acute lung injury. Functionally, LC3+EVs up-regulated the level of PMVECs adhesion molecules and increased neutrophil infiltration in vitro and in vivo experiments. Mechanistically, we found the tRF-1024 was closely related to endothelial adhesion function in LC3+EVs and overexpression of tRF-1024 enhanced neutrophil adhesion of PMVECs.

Conclusions: Our study uncovers a novel LC3+EV carried-tRF1024 mechanism for the rapid recruitment of neutrophils by PMVEC activation. These findings provide novel insights into PMVEC activation and open a new avenue for developing novel therapeutic strategies in ARDS.


BM-MSCs attenuate LPS-induced acute respiratory distress syndrome by promoting resolvin E1/protectin D1

S Yang1, J Liu1, C Li2, YY Pan2

1Department of Critical Care Medicine, The Affiliated Suzhou Hospital of Nanjing Medical University, Suzhou Municipal Hospital, Suzhou, China, 2Department of Critical Care Medicine, Suzhou, China

Critical Care 2023, 27(S1): P058

Introduction: Acute lung injury (ALI) and its severe form acute respiratory distress syndrome (ARDS) are respiratory failure caused by excessive alveolar inflammation with high mortality. In this study, we investigated effects of bone morrow mesenchymal stem cells (BM-MSCs) on lung injury of lipopolysaccharide (LPS)-induced ALI and explored the associated mechanisms.

Methods: BM-MSCs were isolated, cultured, identified by staining with CD34 and CD44 surface markers. LPS-induced ALI mouse model was generated by injecting with LPS and divided into ALI group and ALI + BM-MSCs group. Mice treated without any reagents were assigned as Control, mice transplanted with BM-MSCs were assigned as BM-MSCs group. Regulatory T (Treg) and Th17 percentages were evaluated using flow cytometry. Pro-resolving mediators (resolvin E1 (RvE1), protectin D1 (ProD1)) in lung tissue and cytokines (interleukin-6 (IL-6) and IL-17) in serum were analyzed by ELISA. Myeloperoxidase (MPO) activity was determined. Cultured cells demonstrated typical characteristics of BM-MSCs.

Results: BM-MSCs transplantation (ALI + BM-MSCs) obviously alleviated LPS-induced ALI in mice. BM-MSCs transplantation significantly decreased MPO activity in LPS-induced ALI in mice compared to Control group (p < 0.05). BM-MSCs transplantation markedly increased Treg percentages, and decreased dendritic cells (DCs) and Th17 cells percentages compared to those of Control group (p < 0.05). BM-MSCs transplantation remarkably enhanced RvE1 and ProD1 levels in LPS-induced ALI (ALI + BM-MSCs) compared to ALI group (p < 0.05). BM-MSCs transplantation significantly attenuated IL-6 and IL-17 levels in serum of mice treated with LPS (ALI + BM-MSCs) compared to those of ALI group (p < 0.05).

Conclusions: BM-MSCs transplantation effectively attenuated LPS-induced pathological injury of ALI in mice, at least partly through promoting pro-resolving mediators RvE1 and ProD1 and modulating balance of Treg/Th17.


Derivation and validation of clinical phenotypes and associated treatment responses in critically ill COVID-19 patients

N Bruse1, A Motos2, R Van Amstel3, E De Bie 1, J Kennedy4, N De Keizer5, L Van Vught3, A Torres2, P Pickkers1, M Kox1

1Radboudumc, Department of Intensive Care Medicine, Nijmegen, Netherlands, 2Institute of Health Carlos III, Pulmonary Department, Madrid, Spain, 3Academic Medical Center, Department of Intensive Care Medicine, Amsterdam, Netherlands, 4University of Pittsburgh, Department of Anesthesiology and Intensive Care Medicine, Pittsburgh, USA, 5Academic Medical Center, National Intensive Care Evaluation (NICE) Foundation, Amsterdam, Netherlands

Critical Care 2023, 27(S1): P059

Introduction: Due to variability in the host response, a uniform treatment strategy for severe COVID-19 may be inadequate. We applied unsupervised clustering methods to large cohorts of COVID-19 ICU patients to derive and validate clinical phenotypes, and to explore treatment responses in these phenotypes.

Methods: Phenotypes were derived in 13.279 critically ill COVID-19 patients admitted to 82 Dutch ICUs from September 2020 to February 2022. Twenty-one features were selected from clinical characteristics measured within 24 h after ICU admission. Phenotypes were assigned using consensus k means clustering. External validation was performed in 6225 critically ill COVID-19 patients admitted to 55 Spanish ICUs from February 2020 to December 2021. Individual patient data on corticosteroids therapy enabled us to investigate phenotype-specific responses in this cohort.

Results: Three distinct clinical phenotypes were derived (Fig. 1A). Patients with phenotype 1 (43%) were younger, had lower APACHE IV scores, higher BMI as well as a lower P/F ratio and 90-day in-hospital mortality (18%, Fig. 1A). Phenotype 2 patients (37%) were older and had slightly higher APACHE IV scores compared with phenotype 1, a lower BMI, and higher mortality compared to phenotype 1 (24%, p = 2.95e−07). Phenotype 3 (20%) included the oldest patients with the most comorbidities and highest APACHE IV scores, severe renal and metabolic impairment, and the worst outcome (47% mortality, p = 6.6e−16 and p = 6.6e−16 versus phenotypes 1 and 2, respectively). Phenotype distribution and outcome were very similar in the validation cohort (Fig. 1B). This cohort also revealed that corticosteroid therapy only benefited phenotype 3 (65% vs. 54% mortality, p = 2.5e−03, Fig. 1C).

Conclusions: COVID-19 ICU phenotypes based on clinical data are related to outcome and treatment responses. This can inform treatment decisions as well as randomized trials employing precision medicine approaches.

figure w

Fig. 1 (abstract P059). Phenotype distribution and 90-days in-hospital mortality for critically ill COVID-19 patients in the derivation (A) and validation (B) cohorts. C Comparison of 90-days in-hospital mortality for the three phenotypes in the validation cohort between patients who did and did not receive corticosteroid treatment after the first 24 h of ICU admission. p values were calculated using log-rank tests. COVID-19: coronavirus disease 2019


Machine learning uncovers blood test patterns subphenotypes at hospital admission discerning increased 30-day ICU mortality rates in COVID-19 elderly patients.

NRG Romero-Garcia1, LZ Zhou2, RB Badenes1, TGM Garcia-Morales2, AGC Gomez-de la Camara2, FGR Garcia-Ruiz2, JMGG Garcia-Gomez2, CS Saez2

1Hospital Clinico Universitario de Valencia, Anaesthesiology and Critical Care, Valencia, Spain, 2Universidad Politécnica de Valencia, Valencia, Spain

Critical Care 2023, 27(S1): P060

Introduction: Since March 2020, a number of SARS-CoV-2 patients have frequently required intensive care unit (ICU) admission, associated with moderate survival outcomes and an increasing economic burden. Elderly patients are among the most numerous, due to previous comorbidities and complications they develop during hospitalization [1]. For this reason, a reliable early risk stratification tool could help estimate an early prognosis and allow for an appropriate resources allocation in favour of the most vulnerable and critically ill patients.

Methods: This retrospective study includes data from two Spanish hospitals, HU12O (Madrid) and HCUV (Valencia), from 193 patients aged > 64 with COVID-19 between February and November 2020 who were admitted to the ICU. Variables include demographics, full-blood-count (FBC) tests and clinical outcomes. Machine learning applied a non-linear dimensionality reduction by t-distributed stochastic neighbor embedding (t-SNE) [2]; then hierarchical clustering on the t-SNE output was performed. The number of clinically relevant subphenotypes was chosen by combining silhouette and elbow coefficients, and validated through exploratory analysis.

Results: We identified five subphenotypes with heterogeneous inter-clustering age and FBC patterns (Fig. 1). Cluster 1 was the ‘healthiest’ phenotype, with 2% 30-day mortality and characterized by moderate leukocytes and eosinophils. Cluster 5, the severe phenotype, showed 44% 30-day mortality and was characterized by the highest leukocyte, neutrophil and platelet count and minimal monocytes and lymphocyte count. Clusters 2–4 displayed intermediate mortality rates (20–28%).

Conclusions: The findings of this preliminary report of Eld-ICU-COV19 patients suggest the patient’s FBC and age can display discriminative patterns associated with disparate 30-day ICU mortality rates.


  1. 1.

    COVID-ICU Group. Intensive Care Med. 2021;47:60–73.

  2. 2.

    Van der Maaten L et al. J Mach Learn Res. 2008;9:2579–605.

figure x

Fig. 1 (abstract P060). Differences in standardised average blood cell counts between the five Eld-ICU-COV19 subphenotypes


Latent class analysis of imaging and clinical respiratory parameters from patients with COVID-19-related ARDS identifies recruitment subphenotypes

DFL Filippini1, E Di Gennaro2, RBE Van Amstel 1, LFM Beenen3, S Grasso4, L Pisani5, LDJ Bos6, MR Smit1

1Amsterdam UMC, Intensive Care, Amsterdam, Netherlands, 2University of Bari, Faculty of Medicine, Bari, Italy, 3Amsterdam UMC, Radiology and Nuclear Medicine, Amsterdam, Netherlands, 4University of Bari, Emergency and Organ Transplantation, Bari, Italy, 5Mahidol Oxfornd Research Unit, Critical Care Africa Asia Network, Bangkok, Thailand, 6Amsterdam UMC, Intensive Care, Plumonology, LEICA, Amsterdam, Netherlands

Critical Care 2023, 27(S1): P061

Introduction: Patients with COVID-19-related acute respiratory distress syndrome (ARDS) require respiratory support with invasive mechanical ventilation and show varying responses to recruitment manoeuvres. In patients with ARDS not related to COVID-19, two pulmonary subphenotypes that differed in recruitability were identified using latent class analysis (LCA) of imaging and clinical respiratory parameters [1]. We aimed to validate these phenotypes and evaluate if similar subphenotypes are present in patients with COVID-19-related ARDS.

Methods: This is the retrospective analysis of mechanically ventilated patients with COVID-19-related ARDS who underwent CT scans at positive end-expiratory pressure of 10 cmH2O and after a recruitment manoeuvre at 20 cmH2O. LCA was applied to quantitative CT-derived parameters, clinical respiratory parameters, blood gas analysis and routine laboratory values before recruitment to identify subphenotypes.

Results: 99 patients were included. Using 12 variables, a two-class LCA model was identified as best fitting. Subphenotype 2 (recruitable) was characterized by a lower PaO2/FiO2, lower normally aerated lung volume and lower compliance as opposed to a higher non-aerated lung mass and higher mechanical power when compared to subphenotype 1 (non-recruitable) (Fig. 1). Patients with subphenotype 2 had more decrease in non-aerated lung mass in response to a standardized recruitment manoeuvre (p = 0.024) and were mechanically ventilated longer until successful extubation (adjusted SHR 0.46, 95% CI 0.23–0.91, p = 0.026), while no difference in survival was found (p = 0.814).

Conclusions: A recruitable and non-recruitable subphenotype were identified in patients with COVID-19-related ARDS. The subphenotypes are similar to non-COVID-19-related ARDS and are promising for identification of recruitable patients in future practice as they can be classified with only few clinically available parameters before the recruitment manoeuvre.


  1. 1.

    Wendel Garcia PD et al. Crit Care 2021;25:154

figure y

Fig. 1 (abstract P061). Profile plot of the two subphenotypes identified by the latent class analysis (LCA). All variables used in the LCA are plotted on the x-axis, with the y-axis displaying the standardized mean difference (SMD) of the corresponding variables in both of the LCA derived subphenotypes. SMDs are calculated by standardizing the variable to a mean of 0 and a standard deviation of 1. The variables on the x-axis are ordered by the y-value of the recruitable phenotype in a descending way


Distinct clinical phenotypes of critically ill COVID-19 patients: a cohort observational study

JP Cidade1, VSD Cés de Sousa Dantas2, AF Thompson3, R Miranda3, P Póvoa1

1Hospital São Francisco Xavier, Intensive Care Department, Lisbon, Portugal, 2Instituto D´Or de Pesquisa e Ensino, Intensive Care Department, Rio de Janeiro, Brazil, 3Hospital Copa D´Or, Intensive Care Department, Rio de Janeiro, Brazil

Critical Care 2023, 27(S1): P062

Introduction: COVID-19 presents a complex pathophysiology and evidence collected points towards an intricated interaction of viral-dependent and individual immunological mechanisms. The identification of phenotypes, through clinical and biological markers, may provide a better understanding of the subjacent mechanisms and an early patient-tailored characterization of illness severity.

Methods: Multicenter prospective cohort study performed in 5 hospitals of Portugal and Brazil, during one year, between 2020–2021. All adult patients with an Intensive Care Unit admission with SARS-CoV-2 pneumonia were eligible. COVID-19 was diagnosed using clinical and radiologic criteria with a SARS-CoV-2 positive RT-PCR test. A two-step hierarchical cluster analysis was made using several class-defining variables.

Results: 814 patients were included. The cluster analysis revealed a three-class model, allowing for the definition of three distinct COVID-19 phenotypes: 244 patients in phenotype A, 163 patients in phenotype B, and 407 patients in phenotype C. Patients included in the phenotype C were significantly older, with higher baseline inflammatory biomarkers profile, and significantly higher requirement of organ support and mortality rate (Table 1 (abstract P062)). Phenotypes A and B demonstrated some overlapping clinical characteristics but different outcomes. Phenotype B patients presented a lower mortality rate, with consistently lower C-reactive protein, but higher procalcitonin and interleukin-6 serum levels, describing an immunological profile significantly different from phenotype A (Table 1).

Conclusions: Severe COVID-19 patients exhibit three different clinical phenotypes with distinct profiles and outcomes. Their identification could have an impact in patients’ care, justifying different therapy responses and inconsistencies identified across different randomized control trials results.

Table 1 (abstract P062). Demographic, clinical and outcome variables of patients according to the COVID-19 phenotype


PHENOTYPE A (n = 244; 24%)

PHENOTYPE B (n = 163; 16%)

PHENOTYPE C (n = 407; 60%)


Age, years (median (IQR))

62 (47–79)

63 (47–80)

81 (65–97)

 < 0.001

SOFA at admission (median (IQR))

3 (2;5)

1 (0; 3)

10 (5; 13)

 < 0.001

Max C-Reactive protein, mg/dL (mean ± SD)

25.3 ± 10.4

18.6 ± 12.5

32.3 ± 11.0

 < 0.001

Max registered Procalcitonin, ng/mL (median (IQR))

0.34 (0.06; 0.74)

1.30 (0.70; 1.40)

9.73 (0.86;13.54)

 < 0.001

IL-6 serum levels, (median (IQR))

35.4 (6.6; 42.7)

41.0 (16.0; 49.0)

57.9 (5.7; 61.0)


ICU length of stay, days, (median (IQR))

6 (5; 11)

10 (3; 10)

14 (11; 15)

 < 0.001

Mortality, (n, %)

17 (7.0%)

2 (1.2%)

49 (12.0%)



Comparison of the efficacy of tocilizumab and itolizumab for the treatment of severe COVID-19: a retrospective cohort study

A Kumar, N Kumar, A Kumar, A Pattanayak

All India Institute of Medical Sciences Patna, Anaesthesiology, Patna, India

Critical Care 2023, 27(S1): P063

Introduction: Itolizumab, a CD6 inhibitor has been found to be effective in COVID-19 in some studies [1] but there is no randomised controlled trial at present to prove its effectiveness.

Methods: The study population was adults (> 18 years) with severe COVID-19 pneumonia admitted in the ICU who received either tocilizumab or itolizumab in their course of stay in ICU. The primary outcome of the study was a clinical improvement (CI). The secondary outcomes were time for clinical improvement, improvement in PO2/FiO2 ratio, best PO2/FiO2 ratio, need for mechanical ventilation (MV) after administration of study drugs, time to discharge and survival days.

Results: 126 patients were included in the study; 92 received tocilizumab, and 34 received itolizumab. CI was seen in 46.7% and 61.7% of the patients in the tocilizumab and itolizumab groups, respectively and was statistically non-significant. The time to CI was also non-significant between the tocilizumab and itolizumab groups (median 12 vs 11 days). The number of days required to achieve the improvement of 100 in the PO2/FiO2 ratio was significantly less with itolizumab as compared to tocilizumab. (6 vs 8 days, p value = 0.028). The best PO2/FiO2 ratio achieved was also significantly better with itolizumab as compared to tocilizumab (315 vs 250, p value = 0.043). The incidence of serious adverse events due to the study drugs was significantly higher with itolizumab as compared to tocilizumab (14.7 vs 3.26%). The estimated median time for CI was 12 days and 11 days in the tocilizumab and itolizumab groups, respectively and was non-significant (log-rank p value = 0.262) (Fig. 1).

Conclusions: The clinical improvement and survival rates with itolizumab are similar to tocilizumab. Better oxygenation can be achieved with itolizumab and can be a substitute for tocilizumab in managing severe COVID-19 infection.


  1. 1.

    Kumari P et al. Indian J Crit Care Med 2021;25:467–469.

figure z

Fig. 1 (abstract P063). Kaplan–Meier estimate of the cumulative probability of clinical improvement by treatment group


Stress rate in mechanically ventilated patients with ARDS: a novel concept of ventilator induced lung injury

S Redaelli1, E Costa2, D Von Wedel1, A Suleiman1, R Munoz1, A Santarisi1, D Talmor1, M Amato2, EN Baedorf-Kassis3, MS Schaefer1

1Beth Israel Deaconess Medical Center, Harvard Medical School, Department of Anesthesia, Critical Care and Pain Medicine, Boston, USA, 2Heart Institute (Incor), Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo, Laboratório de Pneumologia LIM-09, Disciplina de Pneumologia, Sao Paulo, Brazil, 3Beth Israel Deaconess Medical Center, Harvard Medical School, Department of Pulmonary, Critical Care & Sleep Medicine, Boston, USA

Critical Care 2023, 27(S1): P064

Introduction: Excessive lung stress, measured by driving pressure (DP) may result in ventilator-induced lung injury (VILI) and mortality in patients with ARDS [1]. Integrative concepts like mechanical power (MP) have been associated with mortality in ARDS, although it is unclear whether information gained from including respiratory rate (RR) [2] reflects rate of deformation (strain rate), rather than repetition [3]. While strain is difficult to measure in clinical settings, stress rate could be a surrogate, but has not yet been studied.

Methods: This secondary analysis included adult patients with ARDS from the EPVent2 trial [4]. We defined stress rate as DP over inspiratory time. We investigated its correlation with dynamic parameters including RR and MP on study day 1. As exploratory analysis, the association of DP and stress rate with 28-day mortality was investigated through logistic regression adjusting for age, sex, BMI and SOFA score.

Results: 193 patients were included. Characteristics of patients by high (> 15.3 cmH2O/s) and low (≤ 15.3 cmH2O/s) stress rate are presented in Table 1. The median (IQR) DP at baseline after study initiation was 12 (10–14) cmH2O. Median inspiratory time was 0.8 (0.7–0.9) s and stress rate was 15 (11–19) cmH2O/s. Stress rate was highly correlated with DP (r = 0.80, p < 0.001) and moderately correlated with RR (r = 0.38, p < 0.001). Low level correlation was observed with MP (r = 0.21, p = 0.02). In adjusted analyses, stress rate was not associated with 28-day mortality (aOR 1.04 [0.99–1.09], p = 0.09; per 1 cmH2O/s increase), opposed to DP (aOR 1.11 [1.02–1.21], p = 0.02, per 1 cmH2O increase).

Conclusions: Stress rate, as surrogate for strain rate showed poor correlation with mechanical power, suggesting that incorporating respiratory rate into this concept reflects repetition rather than rate of deformation.


  1. 1.

    Amato M et al. N Engl J Med 2015;372:747–755.

  2. 2.

    Schaefer MS et al. Intensive Care Med 2021;47:130–132.

  3. 3.

    Protti A et al. Crit Care Med 2016;44:838–45.

  4. 4.

    Beitler J et al. JAMA 2019;321:846–857.

Table 1 (abstract P064). Patient characteristics and distribution of variables by high (> 15.3 cmH2O/s) and low (≤ 15.3 cmH2O/s) stress rate


Stress rate ≤ 15.3 cmH2O/s, n = 97

Stress rate > 15.3 cmH2O/s, n = 96


Age, years

58.0 (47.0–65.0)

57.5 (42.5–70.5)

− 0.11

Female, n (%)

39 (40%)

49 (51%)

− 0.22

SOFA score

6.0 (4.0–9.5)

9.0 (6.0–10.0)

− 0.41

Airway driving pressure, cmH2O

10.0 (8.0–11.0)

14.0 (13.0–17.0)

− 1.58

Inspiratory time, s

0.9 (0.8–1.0)

0.7 (0.6–0.8)


Respiratory rate, breaths per min

26.0 (21.0–29.0)

30.0 (24.0–34.0)

− 0.59

Mechanical power, J/min

24.9 (20.2–31.0)

31.7 (23.6–38.6)

− 0.34

  1. Data are presented as median (IQR). SOFA, Sequential Organ Failure Assessment


Transpulmonary pressure as a predictor of successful lung recruitment: reanalysis of a multicenter international randomized clinical trial

A Santarisi1, A Suleiman1, S Redaelli1, D Von Wedel1, JR Beitler2, D Talmor1, V Banner-Goodspeed1, B Jung3, MS Schaefer1, E Baedorf-Kassis3

1Beth Israel Deaconess Medical Center, Department of Anesthesia, Critical Care and Pain Medicine, Boston, USA, 2Columbia University College of Physicians & Surgeons, Center for Acute Respiratory Failure and Division of Pulmonary, Allergy, and Critical Care Medicine, New York, USA, 3Beth Israel Deaconess Medical Center, Department of Pulmonary, Critical Care & Sleep Medicine, Boston, USA

Critical Care 2023, 27(S1): P065

Introduction: Recruitment maneuvers (RM) are used in ARDS patients to improve oxygenation and lung mechanics [1]. 50% of patients do not respond due to unpredictable transpulmonary pressures (PL) suggesting that individualized care based on PL may optimize recruitment and prevent overdistension [2].

Methods: This retrospective cohort study included ARDS adult patients, previously enrolled in the EPVent2 trial, receiving RM and with available waveforms recordings [3]. We estimated dose–response associations of end-recruitment end-inspiratory PL and change in lung elastance after RM (ΔEL). Analysis was done through a generalized propensity score (GPS) with logistic link function model, multivariable logistic and linear regression models. Negative ΔEL represented successful recruitment, and positive ΔEL overdistension.

Results: 123 patients were included. 69 (65.1%) had a negative ΔEL. The means (± SD) of PL were 14.0 cmH2O (± 4.0) and 15.1 cmH2O (± 4.9) in patients with negative and positive ΔEL, respectively. In the GPS model, a higher PL was associated with higher probability of overdistension (p = 0.003) and predicted an inverse S-shaped dose–response curve, where the probability of successful recruitment is below 50% when PL values were > 18.1 cmH2O (Fig. 1). In the adjusted logistic regression, higher values of PL were associated with decreased probability of successful recruitment (ORadj 0.73 per 1 cmH2O PL; 95% CI 0.62–0.86; p < 0.01). Volume of recruitment (VRM) was dependent on PL (p < 0.01, R2 = 0.49) and was inversely dependent on ΔEL controlling for the baseline lung elastance (p < 0.01, R2 = 0.43).

Conclusions: Higher PL results in more VRM, but also more overdistension. Evaluation of PL should be considered while performing lung recruitment.


  1. 1.

    Hess DR et al. Respir Care. 2015;60:1688–1704.

  2. 2.

    Baedorf Kassis E et al. Intensive Care Med. 2017;43:1162–1163.

  3. 3.

    Beitler JR et al. JAMA. 2019;321:846–857.

figure aa

Fig. 1 (abstract P065). Dose–response associations of end-recruitment end-inspiratory PL and change in lung elastance after RM (ΔEL), where positive ΔEL represents lung overdistension. Analysis was done through a generalized propensity score (GPS) with logistic link function model


Are there any alternatives to the standard esophageal balloon catheter to monitor esophageal pressure in ICU patients?

A El Kik, H Eid, T Mehanna, M Riachy

Hôtel-Dieu de France – Saint Joseph University, Pulmonary and Critical Care Department, Beyrouth, Lebanon

Critical Care 2023, 27(S1): P066

Introduction: Some intensive care physicians practicing in underdeveloped countries are limited in the use of esophageal pressure due to the unavailability and/or to the high cost of the esophageal balloon catheters. The aim of this study is to identify medical devices with inflatable balloons that could be alternatives to traditional esophageal balloon catheters.

Methods: Four different balloons were studied: the esophageal (EB) and gastric balloons (GB) of a Blakemore tube (Rüsch©, 18 Ch), the balloon of a Swan-Ganz catheter (Edwards Lifesciences©) (SGB) and the balloon of a Foley catheter (Bardia©, 24 Ch) (FB). Each balloon was connected apart to an ICU ventilator (Engström Carestation©) and was studied in vitro. The pressure–volume curves was drawn for each balloon by measuring the pressure for every imposed volume on room atmosphere. A plateau pattern is searched for each balloon to avoid any additional pressure due to balloon wall stretching [1]. Then each balloon was submerged in a water column of a progressive depth of 10, 20 and 30 cm in order to create an external pressure of 10, 20 and 30 cmH2O respectively. The pressure–volume curves were again obtained under these conditions.

Results: Zero-plateau curve was only observed for the esophageal (EB) (volume ≤ 45 mL) and gastric (GB) (volume ≤ 10 mL) balloons of the Blakemore tube. When outside pressure of 10, 20 and 30 cm H2O were applied to EB and GB, the plateau was successfully achieved. The volume varied from 20 to 40 mL for the EB with a plateau at 9, 17 and 27 cmH2O respectively. The volume varied from 4 to 10 mL for the GB with a plateau at 7, 20, and 26 cmH2O respectively (Fig. 1).

Conclusions: The use of EB to monitor the esophageal pressure is limited by the need of high volumes due to its intrinsic elastance. The GB is a good alternative to standard catheter with an adapted volume that ranges between 4 to 10 mL.


  1. 1.

    Mojoli F et al. Crit Care 2016;20:98.

figure ab

Fig. 1 (abstract P066). Pressure–volume curves of the Blakemore tube’s gastric balloon (GB)


Impact of the COVID-19 pandemic on acute pancreatitis in the Czech Republic: pilot data from the PANACOTA (pancreatitis acuta in COVID time analysis) study

M Harazim, R Kroupa

Department of Gastroenterology and Internal Medicine, University Hospital Brno, Czech Republic

Critical Care 2023, 27(S1): P067

Introduction: The COVID-19 pandemic had a significant impact on healthcare. Acute pancreatitis (AP) is a clinically serious disease that almost always requires hospitalisation and treatment that cannot be delayed or planned. International studies have highlighted the potential negative impact of COVID-19 infection on the onset and severity of the disease. The aim of this study is to analyse the impact of the COVID-19 pandemic on the epidemiology, treatment and prognosis of acute pancreatitis at the population level.

Methods: Comparison of parameters of hospitalizations for acute pancreatitis, care provided and treatment outcomes during the pandemic (2020 + 2021) with previous years (2010–2019) using data from the National Registry of Covered Health Services. Presentation of pilot data.

Results: Hospitalizations for AP increased slightly from 2010 to 2015 to about 7000/year with a relatively stationary number in 2016–2021. Hospitalizations for AP in each month of 2020 and 2021 did not correlate with the number of COVID-19 positive patients and showed a similar trend (with peaks in summer and at the end of the year) as in the years prior to the pandemic. Concurrently with AP, 2.3% and 3.7% of patients had COVID-19 infection in 2020 and 2021, respectively. AP mortality in the pre-pandemic and pandemic years was virtually the same at 3.9%. There was no difference in length of hospital or ICU stay. There was a downward trend in the number of ERCPs and surgical interventions performed.

Conclusions: According to pilot population data, the COVID-19 pandemic did not have an impact on the incidence or prognosis of acute pancreatitis in the Czech Republic. Changes in trends in interventions performed correlate with recent expert recommendations and are unlikely to be related to the reduction in care due to the pandemic.


A nationwide cohort study of pregnant and postpartum patients with severe COVID-19 pneumonitis in ICU in Israel

E Fatnic

Hadassah Medical Organization and Faculty of Medicine, Hebrew University of Jerusalem, Jerusalem, Israel, Department of Anesthesiology, Critical Care and Pain Medicine, Jerusalem, Israel

Critical Care 2023, 27(S1): P068

Introduction: This study included pregnant patients with severe COVID to test the hypothesis that the impact of delivery on maternal outcome depends upon illness severity at the time of delivery; we hypothesized that patients not yet requiring IPPV would improve following delivery (due to improvement in respiratory mechanics), while patients already on IPPV, or close to requiring ventilation, would deteriorate (due to maternal cardiovascular intolerance to autotransfusion).

Methods: This multicenter, prospective/retrospective cohort study evaluated Israeli ICU admissions of pregnant women with COVID-19 pneumonitis from 1-Feb-2020 to 31-Jan-2022. We assessed maternal, neonatal outcomes and longitudinal maternal clinical data. The primary outcome was maternal outcome (no-IPPV, IPPV, ECMO, death). The primary longitudinal outcome was SOFA score, the secondary longitudinal outcome was the novel PORCH score (PEEP, Oxygenation, Respiratory-support, Chest-X-ray, Haemodynamic-support). Patients were classified into: “no-delivery”, “postpartum admission”, “delivery-critical” and “delivery-not-critical” groups.

Results: 84 patients in 13 ICUs were analysed; there were 34 “no-delivery”, 4 “postpartum”, 32 “delivery-critical”, 14 “delivery-not-critical” patients. “Delivery-critical” and “postpartum” had worse outcomes with, 26/32(81%) and 4/4(100%) requiring IPPV; 12/32(38%) and 3/4(75%) requiring ECMO; 1/32(3%) and 2/4(50%) dying. “Delivery-not-critical” and “no-delivery” had far better outcomes with, respectively, 6/34(18%) and 2/14(14%) requiring mechanical ventilation; no patients required ECMO or died. SpO2, S/F ratio, P/F ratio in “Delivery-critical” deteriorated on the day of delivery, continued to deteriorate, and took longer to recover; “delivery-not-critical” improved rapidly following delivery. The day of delivery was a highly significant covariate for PORCH (p < 0.0001), not SOFA (p = 0.09).

Conclusions: Interventional delivery should be considered for maternal indications before patients deteriorate and require IPPV.


Sars-CoV-2 in critical immunocompromised patients. Evolution and prognosis

A Rivas Bilbao, C Hoya Gonzalez, D Iglesias Posadilla, J Higuera Lucas, M Corera Cia, M Vazquez, E Iñigo Morras, S Kadi Ayad, M Rubio Gaztelu, A Jimeno Rodriguez

Hospital Universitario de Cruces, Medicina Intensiva, Baracaldo, Spain

Critical Care 2023, 27(S1): P069

Introduction: In this study, we share the results of immunosuppressed patients who suffered from acute respiratory distress syndrome (ARDS) secondary to COVID-19 pneumonia managed in our ICU.

Methods: We tracked all patients admitted to ICU of a Tertiary Hospital diagnosed with severe SARS-COV2 pneumonia from March 1, 2020 to January 31, 2022. The definition of Immunocompromised patient is based on history of transplantation, active neoplasia, autoimmune diseases or HIV. Collected data includes: sex, age, type of immunosuppression, vaccination, mechanical ventilation, ECMO VV, incidence of superinfections and mortality.

Results: From a cohort of 425 patients, 55 met the inclusion criteria. 33% were women and 67% male. The average age was 58 years for women and 62 years for men. Out of these patients, 27% had solid organ transplants. 40% suffered from neoplasic disease. 27% had autoimmune diseases and were under treatment with immunosuppressants. 3 had HIV. Only the 29% had received at least 1 dose of COVID 19 vaccine. 80% required orotracheal intubation. 3.64% (2) required Veno-Venous ECMO. 61% presented bacterial superinfection, with the most frequent germs being Pseudomonas aeruginosa and Enterococcus. 36% had viral superinfection, being cytomegalovirus the most frequent one. 32% had fungal superinfection, mainly by Aspergillus fumigatus. 27% did not suffer any superinfection. 40% of the total sample died. After logistic regression, in our model (AUC 83,4% (Se 57.1%, Sp 87.9%), we identified need of intubation as independent variable of mortality (OR 27,06 IC95% 1.76–415.55, p = 0.018).

Conclusions: Immunocompromised patients with ARDS secondary to COVID-19 pneumonia present high mortality, with statistically significant difference when mechanical ventilation is needed. The most frequently isolated germs causing superinfection in this group of patients are bacterias. We believe that this group of patients require special care in our ICU units and an in-depth analysis and study to optimize their prognosis.


Geodemographic factors associated with COVID-19 mortality in indigenous population in Colombia

J Cárdenas1, J Leon2

1Hospital Universitario San Vicente Fundación, Critical Medicine and Intensive Care/Medical Intensive Care Unit, Medellin, Colombia, 2Centro de Tratamiento e Investigación sobre Cáncer Luis Carlos Sarmiento Angulo (CTIC), Hospitalization Unit/Internal Medicine, Bogotá DC, Colombia

Critical Care 2023, 27(S1): P070

Introduction: Since 2019 there have been over 80,000 confirmed COVID-19 cases in the indigenous ethnic groups in Colombia. Age, sex, and region of residency might be factors that contribute to COVID-19 mortality in these ethnic populations. The objective of this research is to describe COVID-19 whether these are associated with COVID-19 mortality in this population. According to the 2018 national population census, there are 1.905.617 persons who identified themselves as indigenous, 50.1% of whom are women, younger than 64 years old and live in rural areas [1].

Methods: This is a retrospective cohort study, using data collected through the national retrospective cohort of confirmed COVID-19 cases. The study population were the confirmed COVID-19 cases in the indigenous population in Colombia since 03/2019 until 10/2022. A Cox Regression Model was used to estimate the HR by age, sex, and geographical location.

Results: There were 83,436 confirmed COVID-19 cases in the indigenous population in this period. The association between age and COVID-19 mortality shows that older individuals and males have higher mortality risk. The geographical location was explored as a risk factor for COVID-19 mortality. Results are shown in Table 1. Most of the regions have HR very close to 1.0, and none reached statistical significance.

Conclusions: Age and sex remain significant factors associated with COVID-19 mortality, as they are in other population studies [2]. The region of residency is not a factor significantly associated with COVID-19 mortality in this study, as this characteristic does not seem to reflect socioeconomic inequalities that have been proven to impact COVID-19 mortality in Colombia [2].


  1. 1

    Ethnic groups population census, DANE. Available at:

  2. 2

    Cifuentes MP et al. J Epidemiol Community Health. 2021, 75: 60–615.

Table 1 (abstract P070). Geographic regions and risk of death due to COVID-19

Region of residency

HR (95% CI)

p value

Andine region



Amazonic region



Caribbean region



Orinoquia region



Pacific region




Impact of COVID-19 pandemic in solid organ transplant in Bogota, Colombia

J Leon1, J Cardenas2

1Centro de Tratamiento e Investigación sobre Cáncer Luis Carlos Sarmiento Angulo (CTIC), Hospitalization Unit/Internal Medicine, Bogota DC, Colombia, 2Fundación Cardioinfantil, Critical Medicine and Cardiovascular Intensive Care/Cardiovascular Intensive Care Unit, Bogotá DC, Colombia

Critical Care 2023, 27(S1): P071

Introduction: During the COVID-19 pandemic the number of solid organ transplants (SOT) lessened globally due probably to decreased donation and detour of resources to other prioritized activities [1, 2]. The aim of this paper is to analyze the SOT behavior during the COVID-19 pandemic in this city.

Methods: An exploratory analysis was performed on the data of the Statistics on donation and transplant of organs and tissues in Bogota from 2018 until the third trimester of 2022 [3].

Results: 416 SOT from 365 organ donors were performed in Bogota during 2018. The first COVID-19 case in Colombia was documented in March 2020. During the following two years there was a decline in the number of performed SOT and total organ donors, as shown in Table 1. During 2020 there was a drastic reduction on the total SOT, compared with 2018. During 2021 there was a slight recovery in the total SOT, and in the first three trimesters of 2022 there was a drastic increase in the total SOT with 380 procedures performed and 380 organ donors. There was a steeper reduction in the number of heart and lung transplants during 2020 and 2021. The SOT waiting lists remained stable from 2018 to 2022, with 1804 patients in 2021, up to 1950 patients in 2022.

Conclusions: The reduction in SOT might be due to COVID-19 in donors, reduced incidence of brain death and lessened capability to preserve viable organ donors. The SOT waiting list didn’t increase possibly caused by high mortality due to end-stage organ failure. The SOT increase during 2021 coincides with a decrease in COVID-19 lethality in Bogota. This decline in SOT was observed globally during 2020, whilst the rapid recovery in SOT and availability of organ donors during 2022 is a phenomenon that has not been described yet to our knowledge.


  1. 1.

    Dominguez-Gil B et al. Am J Transplant 2020;20:2593–2598.

  2. 2.

    Kute VB et al. Transplant Proc. 2022;54:1412–1416.

  3. 3.

    Statistics on donation and transplant of organs and tissues in Bogota. Health Secretary.

Table 1 (abstract P071). SOT behavior in Bogota, 2018–2021






Total SOT





Total organ donors





Kidney transplants





Liver transplants





Heart transplants





Lung transplants





Waiting list






Regional effects on efficacy and safety of vilobelimab in mechanically ventilated patients with severe COVID-19

EHT Lim1, D Van de Beek2, S De Bruin1, S Rückinger3, C Thielert4, R Guo5, BP Burnett4, MC Brouwer2, NC Riedemann4, APJ Vlaar1

1Amsterdam UMC, Intensive Care Medicine, Amsterdam, Netherlands, 2Amsterdam UMC, Neurology, Amsterdam, Netherlands, 3Metronomia Clinical Research GmbH, Munich, Germany, 4InflaRx GmbH, Jena, Germany, 5InflaRx Pharmaceuticals Inc., Ann Arbor, USA

Critical Care 2023, 27(S1): P072

Introduction: Poor outcomes in COVID-19 patients (pt) are associated with C5a-C5aR axis activation. A C5a-specific monoclonal antibody, vilobelimab (VILO), improves outcomes in critically ill COVID-19 pt in a Phase 3 randomized, double-blind, placebo (PLC)-controlled study [1].

Methods: COVID-19 pt within 48 h of intubation were randomly assigned to receive 6, 800 mg infusions of VILO or PLC at a 1:1 ratio on top of standard of care. Predefined subgroup analyses by region and country were performed.

Results: Forty-six (46) hospitals on 4 continents randomized 369 pt: VILO (n = 178), PLC (n = 191). VILO significantly reduced 28- (HR 0.67; 95% CI 0.48–0.96; p = 0.027) and 60-Day mortality (HR 0.67; 95% CI 0.48–0.93, p = 0.0163) using a predefined, unstratified per protocol analysis. Mortality rates at 28- and 60-days and VILO treatment effects, however, differed substantially between regions: Western Europe HR for 60-day mortality 0.59 [0.37–0.95], South Africa plus Russian Federation HR 0.62 [0.28–1.38] and South America HR 0.80 [0.46–1.39] (Fig. 1). The weak signal in South America is predominately driven by Brazil (n = 74), which showed a significant age imbalance with a median 9-years younger PLC group (44.5-years-old vs 53.5-years-old) with low 60-day mortality of ~ 32.5% in the PLC group versus ~ 43.3% in Western Europe. Adjusting for age group categories (≤ 30, 31–40, 41–50, 51–60, > 60; Cox regression) for 60-day mortality changed the HR in Brazil (0.96 [0.44–2.10] for continuous age-adjustment) to values near the estimate for the entire study population (HR 0.77 [0.35–1.69] for age in categories), suggesting a “by chance” imbalance and not a statistically evident weaker effect in Brazil.

Conclusions: Regional efficacy differences between the rest of the world and South America were driven by age imbalances between treatment groups, which do not diminish the robust efficacy signal for VILO in severe COVID-19.


  1. 1.

    Vlaar APJ et al. Lancet Respir Med. 2022;10:1137–1146.

figure ac

Fig. 1 (abstract P072). Forest plot for 60-days all-cause mortality by region (Western Europe, South America, South Africa/Russian Federation)—FAS. *Hazard ratio from the Cox proportional hazards regression model with outcome 60-day all-cause mortality as a censored time-to-event variable and explanatory variables treatment arm and age. Harzard ratios are only displayed for subgroups with at least 10 patients and for which at least one event per treatment group has been observed


High-dose corticosteroids and biomarker profiles in COVID-19 acute respiratory distress syndrome at the intensive care unit: a retrospective cohort study

K Daenen1, K Tong-Minh2, J Huijben1, V Dalm3, E Van Gorp2, H Endeman1

1Erasmus MC, Intensive Care, Rotterdam, Netherlands, 2Erasmus MC, Viroscience, Rotterdam, Netherlands, 3 Erasmus MC, Internal Medicine, Rotterdam, Netherlands

Critical Care 2023, 27(S1): P073

Introduction: COVID-19 patients with non-resolving ARDS may benefit from treatment with high-dose steroids (HDS), because of a presumed persistent systemic and alveolar hyperinflammation. At ICU admission, it is unknown which patients require HDS. Obtaining insights in their inflammatory state by using biomarkers and determining their association with the effect of HDS could support decision-making. The goal of this study is to compare the patient characteristics and biomarker profiles at ICU admission of the patients that received HDS to the patients who did not.

Methods: This was a retrospective cohort study including COVID-19 patients admitted to the ICU of the Erasmus MC between 2020 and 2022. The primary intervention was treatment with HDS, defined as 1000 mg methylprednisolone or > 40 mg prednisolone for three consecutive days. We compared demographics, comorbidities, biomarkers and mortality between patients treated with HDS and without. Logistic regression multivariate analyses was used to analyze which biomarkers were associated with initiation of HDS.

Results: We included 151 patients, of which 48 were treated with HDS at a median of 6 days after ICU admission (Table 1). There were no significant differences in demographics and comorbidities. Patients treated with HDS had a significantly longer ICU length of stay (p < 0.001) and higher hospital mortality rate (p < 0.001). LDH (p = 0.02) and ferritin (p = 0.05) levels on admission were significantly higher in patients treated with HDS, whereas their CRP was significantly lower (p = 0.02). In multivariate regression analysis, these were not independently associated with the initiation of HDS.

Conclusions: At ICU admission, demographics and comorbidities did not differ between patients treated with HDS and without. There were no factors associated with the initiation of HDS in COVID-19 patients in the ICU. Further studies on the association between the inflammatory state, mortality, and the effect of HDS are required to understand which patients may benefit from HDS.

Table 1 (abstract P073). Patient characteristics and outcomes


Non high-dose steroids (N = 113)

High-dose steroids (N = 48)

Overall (N = 151)

p value

Age in years (median (IQR))

65.0 [15.0]

62.0 [11.0]

64.0 [14.3]


Comorbidities (total (%))

53 (46.9%)

25 (52.1%)



LDH in U/L (median (IQR))

401 [178]

448 [190]

420 [182]


Ferritine in U/L (median (IQR))

1090 [1260]

1550 [1720]

1260 [1630]


CRP in mg/L (median (IQR))

169 [154]

97.3 [169]

159 [178]


ICU length of stay in days (median (IQR))

15.0 [16.0]

27.0 [18.0]

17.0 [19.0]

 < 0.001

Mortality (total (%))

25 (22.1%)

25 (52.1%)

50 (31.1%)

 < 0.001

  1. IQR Interquartile range; ICU intensive care unit


Antithrombotic strategies in COVID-19 patients: a systematic review and Bayesian network meta-analysis

HB Chen1, P Chen2, C Gao3, JF Xie3, HB Qiu3

1Jiangsu Provincial Key Laboratory of Critical Care Medicine, Department of Critical Care Medicine, Zhongda Hospital, School of Medicine, Southeast University, Nanjing, China, 2The Hospital of the Second Mobile Corps, Chinese People’s Armed Police Forces, Fuzhou, China, 3Jiangsu Provincial Key Laboratory of Critical Care Medicine, Nanjing, China

Critical Care 2023, 27(S1): P074

Introduction: COVID-19 is a public health emergency of international concern. Clinicians are likely to adopt various antithrombotic strategies to prevent embolic events, but the optimal antithrombotic strategy remains uncertain. We performed a Bayesian network meta-analysis to evaluate various antithrombotic strategies comprehensively.

Methods: We systematically searched PubMed, Cochrane Library, Web of Science, EMBASE and Clinical trials. gov to screen trials comparing different antithrombotic strategies. The primary outcome is 28-day mortality, and the secondary outcomes include major thrombotic event, major bleeding and in-hospital mortality, etc. We assessed the risk of bias using the Cochrane Collaboration’s tool and the quality of evidence according to the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach. We successively performed traditional pairwise and Bayesian network meta-analysis using R v4.2.1 software.

Results: Twenty-six eligible randomized controlled trials were included, giving a total of 35 paired comparisons with 32,041 patients randomized to 7 antithrombotic strategies. In comparison to standard of care (SoC) strategy, therapeutic anticoagulation (TA) (RR 0.36, 95% CrI 0.13–0.86) and prophylactic anticoagulation (PA) (RR 0.35, 95% CrI 0.12–0.85) strategy significantly reduced the mortality of COVID-19 patients (Fig. 1). The antiplatelet (AP) strategy was associated with high risk of major bleeding when compared with SoC strategy (RR 2.5, 95% CrI 1.1–8.9), and the TA (RR 0.43, 95% CrI 0.17–0.98), PA (RR 0.27, 95% CrI 0.10–0.63) and PA with Fibrinolytic agents (FA) strategy (RR 0.12, 95% CrI 0.01–0.81) was associated with low risk of major thrombotic event.

Conclusions: This network meta-analysis indicates that the TA and PA strategies probably reduce mortality and confer other important benefits in COVID-19 patients. These findings provide guidance on how to choose optimal antithrombotic strategies for COVID-19 patients.

figure ad

Fig. 1 (abstract P074). The forest plot of network effect sizes of various antithrombotic strategies compared with standard of care. CrI = credible interval, SoC = standard of care, PA = prophylactic anticoagulation, TA = therapeutic anticoagulation, AP = antiplatelet, FA = fibrinolytic agents


A new score to predict severe respiratory failure and guide anakinra treatment in COVID-19 pneumonia: analysis of the randomized SAVE-MORE trial

E Kyriazopoulou1, G Poulakou2, H Milionis3, S Metallidis4, M Fantoni5, A Angheben6, P Panagopoulos7, G Dalekos8, M Netea9, E Giamarellos-Bourboulis2

1National and Kapodistrian University of Athens, 4th Department of Internal Medicine, Athens, Greece, 2National and Kapodistrian University of Athens, Athens, Greece, 3University of Ioannina, Ioannina, Greece, 4Aristotle University of Thessaloniki, Thessaloniki, Greece, 5Fondazione Policlinico Gemelli IRCCS, Rome, Italy, 6Tropical Diseases and Microbiology, IRCSS Sacro Cuore Hospital, Negrar, Italy, 7University of Thrace, Alexandroupoli, Greece, 8University of Thessaly, Larissa, Greece, 9Radboud University, Nijmegen, Netherlands

Critical Care 2023, 27(S1): P075

Introduction: Anakinra treatment is approved for the treatment of COVID-19 pneumonia in hospitalized adults in need of oxygen and at risk for progression into severe respiratory failure (SRF) defined as circulating concentrations of the biomarker suPAR (soluble urokinase plasminogen activator receptor) ≥ 6 ng/mL by the EMA and has been authorized for emergency use by FDA under an EUA [1]. This is based on the results of the randomized SAVE-MORE trial where suPAR ≥ 6 ng/mL was used to select patients at risk for SRF [2]. The suPAR test is not commercially available in the USA and an alternative method of patient selection was needed.

Methods: In collaboration with the US FDA, an alternative method to select patients most likely to have suPAR ≥ 6 ng/mL based on commonly measured patient characteristics was developed. Patients with at least 3 of the following criteria are considered likely to have suPAR ≥ 6 ng/ml: age ≥ 75 years, severe pneumonia by WHO criteria, current/previous smoking status, Sequential Organ Failure Assessment score ≥ 3, neutrophil-to-lymphocyte ratio ≥ 7, hemoglobin ≤ 10.5 g/dl, history of ischemic stroke, blood urea ≥ 50 mg/dl and/or history of renal disease.

Results: The positive predictive value of this new score was 95.4% in SAVE-MORE population. However, a lower sensitivity meant a small proportion of patients with suPAR ≥ 6 ng/ml will not be identified by the new score. The adjusted hazard ratio for survival at 60 days for patients meeting this score and who receive anakinra is 0.45 (Fig. 1).

Conclusions: The developed score predicts accurately patients with suPAR levels ≥ 6 ng/mL and may be used as an alternative to guide anakinra treatment in patients with COVID-19 pneumonia. Based on these subgroup results, patients in SAVE-MORE who met the new score appeared to show beneficial efficacy results with treatment of anakinra consistent with the overall studied population.


  1. 1.

    Fact sheet for healthcare providers: emergency use authorization for kineret:

  2. 2.

    Kyriazopoulou E et al. Nat Med 2021;27:1752.

figure ae

Fig. 1 (abstract P075). Cox survival analysis by Day 60 of patients enrolled in the SAVE-MORE trial and meeting the new score


Outcomes of a community-based bamlanivimab infusion program for COVID-19

E Michelson, S Crawford, S Monks

Paul Foster School of Medicine, Texas Tech University HSC, El Paso, Emergency Medicine, El Paso, USA

Critical Care 2023, 27(S1): P076

Introduction: The U.S. FDA issued an Emergency Use Authorization in 11/2020 for Bamlanivimab (Bam), a monoclonal antibody infusion, in high-risk outpatients with COVID-19. We examined the effectiveness of Bam given in a coordinated community wide initiative in reducing ED visits, hospitalizations, and deaths during the first 30 days post-treatment.

Methods: Bam treated adults at a community-based (CB) infusion center in the SW USA were identified in a regional EMR database covering all local hospitals. Individual patients were de-identified but tracked for 30 days for ED visits, hospital admissions, ICU admissions, and deaths. Patient demographics including age, sex and ethnicity were recorded. Our local IRB approved the project as exempt.

Results: A total of 2242 patients received Bam between 11/2020 and 3/2021, while the Wuhan variant was predominant. The mean age was 52, 54% were female and 64% identified as Hispanic or Latino. During the 30 days post infusion 13% of patients visited the ED, but only 2% of the total patients were admitted, 0.6% were admitted to the ICU and 14 died (0.6%). For comparison, a historical placebo control group of COVID patients in Chen et al. [1] is shown in Table 1. Hospitalizations among patients over 60 occurred at the same rate as the total cohort versus more than double the rate in older patients in the control group.

Conclusions: A CB Bam infusion program reduced ED visits, and hospitalizations, particularly among older patients, with a smaller reduction in ICU admissions and death. Our population is older and including a higher percentage of Hispanic patients who tend to have worse outcomes from COVID-19 [2]. While Bam has now been replaced by newer monoclonal antibodies, our results support the potential community benefit of coordinated monoclonal antibody treatment of patients with COVID-19, with a need for more research to identify patient subgroups who may receive the greatest benefit.


  1. 1.

    Chen P et al. N Engl J Med 2021;384:229–237

  2. 2.

    De Ramos P et al. Am J Public Health 2022;112:1034–1044

Table 1 (abstract P076). Comparison of historical control to patients treated with Bam


Historical control 143 (Chen et al. [1])

CB Bam Infusion Center 2242

Average age









Hospitalized (age > 65)


2% (age > 60)

Hospitalized ICU







Population pharmacokinetics of dexamethasone in critically ill COVID-19 patients: inflammation might play a role

L Li1, S Sassen1, N Hunfeld2, T Smeets1, T Ewoldt2, B Koch1, H Endeman2

1Erasmus Medical Center, Pharmacy, Rotterdam, Netherlands, 2Erasmus Medical Center, Intensive Care, Rotterdam, Netherlands

Critical Care 2023, 27(S1): P077

Introduction: One of the common causes of COVID-19 related death is acute respiratory distress syndrome (C-ARDS). Dexamethasone is the cornerstone in the therapy of C-ARDS and reduces mortality probably by suppressing inflammatory levels in ICU patients. Its anti-inflammatory effects may be concentration-related. However, no pharmacokinetic studies of dexamethasone have been conducted in ICU patients. Therefore, we designed a population pharmacokinetic study to gain a deeper understanding of the pharmacokinetics of dexamethasone in critically ill patients in order to identify relevant covariates that can be used to personalize dosing regimens and improve clinical outcomes.

Methods: This was a retrospective pilot study at the ICU of the Erasmus Medical Center. Blood samples were collected in adults at the ICU with COVID who were treated with fixed dose intravenous dexamethasone (6 mg/day). The data were analyzed using Nonlinear Mixed Effects Modelling (NONMEM) software for population pharmacokinetic analysis and clinically relevant covariates were selected and evaluated.

Results: A total of 51 dexamethasone samples were measured in 18 patients. A two-compartment model with first-order kinetics best fitted the data. The mean population estimates for drug clearance and inter-compartment clearance were 2.85 L/h (IIV 62.9%) and 2.11 L/h, respectively, and central and peripheral volumes of distribution were 15.4 L and 12.3 L, respectively. The covariate analysis showed a significant correlation between dexamethasone clearance and CRP. Dexamethasone clearance decreased significantly with increasing CRP in the range of 0–50 mg/L and a correlation was observed with CRP up to 100 mg/L.

Conclusions: The dexamethasone PK parameters of ICU COVID patients were quite different from those come from healthy populations. Inflammation might play an important role in dexamethasone clearance and the dosing should be more individualized in order to achieve best therapeutic effect in ICU patients.


Pharmacokinetic/pharmacodynamic analysis of vilobelimab and anaphylatoxin c5a as well as antidrug antibodies in Panamo: a phase 3 study in critically ill, mechanically ventilated COVID-19 patients

EHT Lim1, APJ Vlaar1, S De Bruin1, S Rückinger2, C Thielert3, R Guo4, BP Burnett3, MC Brouwer5, NC Riedemann3, D Van de Beek5

1Amsterdam UMC, Intensive Care Medicine, Amsterdam, Netherlands, 2Metronomia Clinical Research GmbH, Munich, Germany, 3InflaRx GmbH, Jena, Germany, 4InflaRx Pharmaceuticals Inc., Ann Arbor, USA, 5Amsterdam UMC, Neurology, Amsterdam, Netherlands

Critical Care 2023, 27(S1): P078

Introduction: C5a-C5aR axis activation is associated with increased mortality in severe COVID-19. Vilobelimab (VILO), a C5a-specific monoclonal antibody, improved mortality in severe COVID-19 patients (pts) in a Phase 3 multicenter, randomized, double-blind, placebo (PLC)-controlled study [1]. A pharmacokinetic/pharmacodynamic (PK/PD) analysis was undertaken to assess VILO and C5a as well as antidrug antibodies (ADA) levels in the study.

Methods: Forty-six (46) hospitals on four continents randomized 369 COVID-19 pts (VILO [n = 178], PLC [n = 191]) within 48 h of being mechanically ventilated to receive 6, 800 mg infusions of VILO or PLC at a 1:1 ratio on top of standard of care. Blood samples were taken at screening, Day 8 and at hospital discharge for VILO and C5a and at screening and hospital discharge for ADA. Enzyme‐linked immunosorbent assays were used to analyze levels.

Results: Screening blood samples for VILO and C5a were available for VILO (n = 93) and PLC (n = 99) from sites in Western Europe. On Day 8 after 3 infusions, mean VILO trough concentrations were 21799.3–302972.1 ng/mL (geometric mean 137881.3 ng/mL) (Fig. 1). At screening, C5a was highly elevated and comparable between groups: VILO median 118.3 ng/mL, mean 130.3 ng/mL, PLC median 104.6 ng/mL, mean 123.2 ng/mL. By Day 8, C5a levels were reduced by 84.6% in the VILO group (median 14.5 ng/mL [mean 16.8 ng/mL], p < 0.001) versus a 19.6% increase in the PLC group (median, 119.2 ng/mL, mean 129.8 ng/mL). Beyond Day 8, though PD sampling was sparse, C5a levels remained elevated for PLC whereas C5a slowly rose but did not reach screening levels for VILO. Treatment-induced ADA were observed in 1 pt in the VILO group (Day 40 discharge) and 1 pt in the PLC group (Day 25 discharge), both appeared independent of treatment.

Conclusions: The PK/PD analysis shows that VILO efficiently inhibits C5a in pts with severe COVID-19 resulting in a robust clinical effect on mortality reduction without inducing ADA.


  1. 1.

    Vlaar APJ et al. Lancet Respir Med. 2022;10:1137–1146.

figure af

Fig. 1 (abstract P078). Vilobelimab [ng/mL] and C5a [ng/mL] at screening and Day 8—safety analysis set


Lung cavitation as a rare complication of severe COVID-19: report of 3 cases

E Ntikoudi, E Bourgani, S Giannakaki, A Aiginitou, M Karagianni, C Merkouri, M Poulou, E Kourtelesi, M Daganou, A Flevari

Sotiria Thoracic Diseases Hospital, Polyvalent ICU Department, Athens, Greece

Critical Care 2023, 27(S1): P079

Introduction: Lung cavitation is a rare radiological finding of COVID-19 pneumonia associated with unfavorable outcome. Its pathogenesis is unclear and it is characterized by diffuse alveolar damage, intra-alveolar hemorrhage and necrosis of parenchymal cells.

Methods: We retrospectively reviewed the radiological findings of COVID-19 patients admitted to our ICU during the pandemic in order to identify the development of lung cavitary lesions.

Results: From 11/2020 until 10/2022 1000 patients were admitted to our COVID-19 ICU (92% on invasive mechanical ventilation). According to our data there were three cases of lung cavity formation. The first case was a 78 years male with history of hypertension. Chest CT (Day26) showed a 11.6 cm cavity in the right middle lobe (Fig. 1). The second case was a 52 year old female with history of diabetes mellitus, obesity, hypertension and rheumatoid arthritis. Follow up chest CT (D29) revealed progressive development of multiple bilateral cavitary lesions. The third case was a 61 year old male with no medical history, who developed (D17) multiple cavitary lesions in both lower lobes, concomitant with left-sided pulmonary embolism. The presence of other well defined etiologies of cavitary lesions such as mycobacterial and fungal infections as well as neoplasmatic or autoimmune diseases had been widely excluded. However, since pulmonary cavitation is a late complication of severe COVID disease, we cannot overlook the fact that all patients suffered from superinfections by XDR Acinetobacter baumanii and/or Klebsiella pneumonia, as most of our patients with prolonged length of stay. Moreover, two of the three patients developed pneumothorax. All patients finally died.

Conclusions: Although bacterial co-infection does not allow absolute association between cavitary formation and coronavirus disease, it seems that destructive triggers, such as bacteria or mechanical ventilation, may aggravate COVID underlying lung lesions leading to cavitation.

figure ag

Fig. 1 (abstract P079). CT finding of first case (78 year male). Lung cavity formation during COVID-19 pneumonia


Lung inhomogeneities in COVID-19 related acute respiratory distress syndrome

A Santini1, A Protti1, F Pennati2, C Mercalli1, G Picardo1, L Pugliese1, N Martinetti1, C Chiurazzi3, A Aliverti2, M Cecconi1

1Humanitas University, Department of Biomedical Sciences, Pieve Emanuele (MI), Italy, 2Politecnico di Milano, Dipartimento di Elettronica, Informazione e Bioingegneria, Milan, Italy, 3IRCCS Humanitas Research Hospital, Department of Anesthesia and Intensive Care Units, Rozzano (MI), Italy

Critical Care 2023, 27(S1): P080

Introduction: In acute respiratory distress syndrome (ARDS) inhomogeneities in lung aeration can act as local multipliers of pressure during inspiration (“stress risers”), increasing the risk of lung damage even in presence of airway pressures considered “safe” [1]. In this study we aimed to describe lung inhomogeneities in COVID-19 related ARDS (C-ARDS) and to relate these to disease severity and lung morphology.

Methods: We enrolled patients with C-ARDS within 3 days from mechanical ventilation start, deeply sedated and paralyzed. Lung CT scan was obtained at PEEP of 5 cmH2O to measure lung weight compartments (non-, poorly-, well- and over-aerated). Lung inhomogeneities were computed as the gas/tissue ratio of each voxel compared to the neighboring voxels. We considered values > 1.61 as pathologic lung inhomogeneities, as previously described [1]. The fraction of total lung volume with pathologic inhomogeneities (“extent”) and the average severity of inhomogeneities contained in that fraction (“intensity”) was calculated. Respiratory system compliance and blood gas analysis were obtained at the same PEEP level of the CT scan. Some results have been presented in another publication [2].

Results: Forty patients were studied in the supine position 1 (0–1) days after ICU admission. The extent of pathologic lung inhomogeneities represented 18 ± 4% of total lung volume. The intensity of pathologic lung inhomogeneities was on average 2.53 ± 0.12. “Extent” was positively correlated with the amount of poorly aerated lung weight (r2 = 0.51, p < 0.001) (Fig. 1) and negatively correlated with the amount of non-aerated lung weight (r2 = 0.22, p = 0.002). No correlation was found between “extent” and “intensity” and PaO2/FiO2, dead space fraction or respiratory system compliance.

Conclusions: In C-ARDS lung inhomogeneities represent roughly 20% of total lung volume. In these regions local stress is increased with risk of secondary lung damage.


  1. 1.

    Cressoni M et al. Am J Respir Crit Care Med 2014;189:149–58.

  2. 2.

    Protti A et al. Chest 2022;161:979–988.

figure ah

Fig. 1 (abstract P080) Fraction of poorly-aerated lung tissue and extent of lung inhomogeneities at PEEP 5 cmH2O


Risk factors for nosocomial infections in COVID-19 ICU population

P Costa1, J Alves1, A Martinho1, C Silva1, L Simões1, G Almeida1, J Guimarães2, S Teixeira1, JP Baptista1, P Martins1

1Coimbra’s University Hospital, Intensive Care Department, Coimbra, Portugal, 2Coimbra’s University Hospital, Cardiology Department, Coimbra, Portugal

Critical Care 2023, 27(S1): P081

Introduction: The aim of this study is to identify the factors associated with an increased risk of developing nosocomial infections (NI) in COVID-19 patients admitted with pulmonary involvement in the ICU. NI in COVID-19 ICU population are an important cause of morbidity and mortality worldwide and its prompt identification might lead to its prevention and better outcomes.

Methods: This is a retrospective observational study of patients admitted with COVID-19 pneumonia in the ICU of a tertiary center in Portugal, between March 2020 and December 2021. We considered NI as any infection acquired > 48 h post ICU admission. Clinical, analytical and baseline patient data were evaluated. Logistic regression analysis was performed to correlate patient related variables with the development of NI.

Results: A total of 338 patients were enrolled, from which 167 (47.9%) presented with NI. Baseline characteristics are described in Table 1. In the logistic regression analysis, older age (OR 1.13; 95% CI 1.03–1.25; p = 0.013), coronary artery disease (CAD) (OR 28.7; 95% CI 1.92–429; p = 0.02), obesity (OR 3.14; 95% CI 0.86–11.42; p = 0.008), chronic liver disease (CLD) (OR 104.33; 95% CI 1,.04–1008.49; p = 0.04), use of dexamethasone (OR 21.89; 95% CI 3.04–157.85; p = 0.002) and days in RASS < 3 (OR 1.4; 95% CI 1.05–1.86; p = 0.02) were associated with an increased risk of developing NI in the ICU. Surprisingly, SOFA at admission, days of invasive mechanical ventilation, days of sedation and PaO2/FiO2 ratio at admission, although statistically significantly different between groups, did not correlate with the risk of infection.

Conclusions: We identified prolonged deep sedation, corticosteroid use, and patient characteristics (CAD, obesity, CLD, older age) as independent risk factors for NI development in COVID-19 critically ill patients. It is also noteworthy to point out for the presence of confounding variables, including the excessive workload in the ICU during this period, leading to an increase in NI numbers.

Table 1 (abstract P081). Baseline patient characteristics


With NI (n = 162)

Without NI (n = 176)

p value

Male sex (%)




Age (years)

59.7 (± 12.2)

58.6 (± 13.5)


Charlson comorbidity index

2.2 (± 1.6)

2.1 (± 1.7)


SOFA at admission

6.9 (± 2.8)

5.3 (± 2.9)

 < 0.001

Invasive mechanical ventilation (days)

21.5 (± 13.2)

9.0 (± 4.7)

 < 0.001

Sedation (days)

20.6 (± 12.0)

6.1 (± 5.7)

 < 0.001

PaO2/FiO2 ratio at admission

105.9 (± 47.0)

118.5 (± 48.3)



Incidence and microbiological characteristics of ventilator-associated pneumonia assessed by bronchoalveolar lavage and endotracheal aspirate in COVID-19 patients: COV-AP study

D Mangioni1, J Fumagalli2, A Meli2, M Tomasello1, C Bobbio1, C Matinato3, A Muscatello1, M Panigada4, A Bandera1, G Grasselli2

1Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico, Infectious Disease Unit, Milan, Italy, 2Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico, Anesthesia, Critical Care and Emergency, Milan, Italy, 3Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico, Microbiology Laboratory, Clinical Laboratory, Milan, Italy, 4Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico, Anesthesia, Critical Care and Pain Medicine, Milan, Italy

Critical Care 2023, 27(S1): P082

Introduction: Diagnosis of ventilator-associated pneumonia (VAP) in COVID-19 patients remains challenging. Also, the lack of gold standard for microbiological sampling undermines clinical judgement and management. We studied incidences of microbiologically-confirmed VAP comparing endotracheal aspirate (ETA) and bronchoalveolar lavage (BAL) in COVID-19 patients. Etiological agreement between ETA and BAL was then assessed.

Methods: Single-center prospective cohort study (NCT04766983). Patients were enrolled within 48 h from intubation; surveillance ETA (ETASURV) was performed twice weekly. ETA (ETACX) and BAL (BALCX) samples were collected upon VAP suspicion (Johanson’s criteria). CDC definitions were used for microbiological confirmation. ETA-BAL agreement (interrater reliability and Cohen’s kappa) and clinical/microbiological data were assessed for the first episodes of suspected VAP per patients.

Results: Ninety intensive care (ICU) patients enrolled from 01/2021 to 05 06/2022, of which 26 females (28.9%); median age was 60 [52–66] years. In-ICU mortality was 30/90 (33.3%), median length of stay in survivors 19 (10–32) days. Fifty-three patients (58.9%) had ≥ 1 episode of suspected VAP after 6 [5; 10] days from ICU admission. ETASURV were available in 52 cases, 2 [1; 3] days before VAP suspect, and tested positive in 28 (53.8%). ETACX and BALCX resulted positive in 35 (66.0%) and 29 (54.7%) of episodes. Main microbiological results are displayed in Fig. 1, panel A. Etiological agreement between techniques is shown in Fig. 1, panel B. Incidence rate of VAP suspicions per 1000 ventilator-days was 60.2 (95% CI 43.9–76.4), while incidence rates of microbiologically-confirmed VAP were 27.4 (18.3–36.5) with ETACXand 18.9 (95% CI 12.0–25.8) with BALCX, respectively.

Conclusions: We observed different incidence of VAP in COVID-19 ICU patients depending on sampling method. Etiological agreement between techniques yielded limited interrater reliability. The potential clinical impact needs further studies.

figure ai

Fig. 1 (abstract P082). Panel A: Microbiological findings of the different respiratory samples collected at the first episode of VAP suspicion; Panel B: Observed agreement for VAP diagnosis and interrater reliability [Cohen’s Kappa (95% CI)] for the different type of respiratory samples collected at the first episode of VAP suspicion. ETAsurv: last surveillance endotracheal aspirate obtained before the episode of VAP suspicion; ETAcx/BALcx: endotracheal aspirate/bronchoalveolar lavage obtained at VAP suspicion


Critical illness in patients with omicron variant COVID-19: different but crucial

A Hana, PD Wendel Garcia, C Friedrich, D Nowak, S Keiser, RA Schüpbach, MP Hilty

University Hospital of Zurich, Institute of Intensive Care Medicine, Zurich, Switzerland

Critical Care 2023, 27(S1): P083

Introduction: During the COVID-19 pandemic, various virus variants evolved worldwide. COVID-19 omicron (CO) was associated with a decrease in length of hospital stay, ICU admission and death [1] as compared to COVID-19 delta (CD). To estimate impact of CO on ICUs worldwide, we investigated characteristics, disease course and outcome of critically ill CO patients.

Methods:  Of 8562 critically ill COVID-19 patients included in the prospective international multicenter RISC-19-ICU registry [2,3], characteristics and outcome were compared for 1890 CD and 272 CO patients admitted to ICU between 1–2021 and 9–2022. Mixed model analysis corrected for individual center effects and adjusted for age, sex, vaccination status, use of steroids and anticoagulants was used.

Results: There was no difference in age, sex and BMI between groups. CO patients had more comorbidities [mean difference (MD) 0.7, 95% CI (0.5–1.0), p = 0.02], especially arterial hypertension, and higher SAPS II score [MD 0.0 (0–0.1), p < 0.001], SOFA score [MD 0.1 (0.1–0.3), p < 0.0001]. CO patients presented with higher cardiovascular system SOFA subscore, but better PF-ratio at ICU admission and smaller risk for intubation and mechanical ventilation throughout their ICU stay [OR 0.5 (0.3–0.8)]. There was no difference in ECMO treatments, ICU length of stay [MD 0.6 (0–11.4), p = 0.72] or ICU survival [HR 1 (0.7–1.5), p = 0.88] between the two groups.

Conclusions: We show that critically ill CO patients present with more comorbidities, less severe respiratory disease but more severe hemodynamic instability at ICU admission as compared to CD patients, although the ICU length of stay and mortality was similar. These differences could be explained by differences in disease characteristics caused by CO, or by the increasing prevalence of CO as co-factor to preexisting disease. Continued monitoring of critically ill CO patients in ICUs worldwide is warranted.


  1. 1.

    Iuliano AD et al. MMWR Morb Mortal Wkly Rep. 2022;71:146–152.

  2. 2.

    Wendel-Garcia PD et al. Crit Care 2022;26:199.

  3. 3.

    Hilty MP et al. Intensive Care Med 2022;48:362–365.


Clinical characteristics and outcome of older patients with severe COVID-19: a retrospective observational study

KS Singh, AK Kumar

All India Institute of Medical Sciences Patna, Anaesthesiology, Patna, India

Critical Care 2023, 27(S1): P084

Introduction: The objective of the study was to assess the relevant laboratory and clinical factors that may affect the prognosis of patients with severe COVID-19 in older population. Very few studies have specifically investigated the age-specific factors that affect the outcome of the patient in elderly patients [1,2].

Methods: The study population was adults (≥ 60 years) with severe COVID-19 admitted to the intensive care unit of a tertiary care hospital between March 2021 and June 2021. Binomial logistic regression analysis was used to analyze all variables as potential predictors for the death of older patients.

Results: In total, 113 older patients with severe COVID-19, with a median age of 68 years (interquartile range (IQR) 63–74), mortality rate was 61.9%. At admission, the median PO2/FiO2 ratio of the patients was 100 (IQR 90–150) and 33.6% were on mechanical ventilation. Binomial logistic regression showed that total leucocyte counts (TLC), platelet counts (PC), lactate dehydrogenase (LDH), D-dimer, and interleukin-6 levels were all significantly associated with death. Patients with poor outcomes had significantly lower PO2/FiO2 ratios at admission, higher TLC, lower PC, higher serum creatinine, and higher LDH as compared to survivals. Patients with poor outcomes had a significantly higher percentage of chronic kidney disease patients (20% vs 4%), and patients who didn’t receive remdesivir (27% vs 2.3%).

Conclusions: Increased TLC, LDH, D-Dimer, IL-6, and decreased PC are associated with poor outcomes. The use of remdesivir may increase the chance of survival in older patients.


  1. 1.

    Yan Y et al. BMJ Open Diabetes Res Care. 2020;8:e001343.

  2. 2.

    Singhal S et al. BMC Geriatr. 2021;21:321.


APACHE II score as a predictor of outcome in critically ill COVID-19 patients

M Matsko, C Bostantzoglou, A Katsogianni, K Georgiopoulos, V Karaouli, P Athanasiou, M Karagianni, C Merkouri, M Daganou.

ICU, Sotiria Hospital of Chest Diseases, Athens, Greece.

Critical Care 2023, 27(S1): P085

Introduction: APACHE II severity scoring system has been successfully used for mortality risk assessment in the ICU, however its validity in the subgroup of COVID-19 patients has been questionable. We aimed to examine the predictive value of APACHE II score in a cohort of critically ill COVID-19 patients.

Methods: We performed a retrospective analysis of prospectively collected data in a cohort of COVID-19 patients admitted to our 50-bed ICU between October 2020 and April 2022. Using a ROC analysis we assessed the performance of APACHE II score and identified the optimal cut-off value for mortality prediction.

Results: Our cohort included 783 patients (66% male) with positive PCR forSARS-Cov-2 and respiratory failure. Mean age was 66 years. Invasive mechanical ventilation was used in 92%of patients and 89.3% had at least one comorbidity. The mean APACHE II score of the whole cohort was 20.3 (± 8.5). ICU mortality was 44.7%. Death rate was similar between sexes but significantly higher in those who were older and those suffering from COPD, chronic renal or heart failure, atrial fibrillation or any kind of malignancy. Non-survivors had a significantly higher APACHE II score compared to survivors (25.2 ± 7.9 vs 16.3 ± 6.7, p < 0.001). ROC analysis showed an AUC 0.81 (95% CI 0.78–0.84, p < 0.001). At a cut-off value of 19.5 APACHE II score could predict death with a sensitivity of 77.1% (95% CI 72.4–81.4%), a specificity of 70.7% (95% CI 66.1–74.9%), PPV 68% (95% CI 63.2–72.6%) and NPV 79.3% (95% CI 74.9–83.2%).

Conclusions: APACHE II score is an effective tool for mortality prediction in critically ill COVID-19 patients. A cut-off value of 19.5 can be used for risk stratification in this patient population.


Non-depolarizing neuromuscular blockade in patients with severe acute respiratory distress syndrome associated with COVID-19: profile of usage and potential deleterious impact on morbidity and mortality

F Bessa1, G Cruz2, P Nave2, SM Fernandes3, GN Jesus1, JM Ribeiro1

1Centro Hospitalar Universitário Lisboa Norte, Intensive Care Department—ECMO referral centre, Lisbon, Portugal, 2Centro Hospitalar Universitário Lisboa Norte, Intensive Care Department, Lisbon, Portugal, 3Faculdade de Medicina de Lisboa, Clínica Universitária de Medicina Intensiva, Lisbon, Portugal

Critical Care 2023, 27(S1): P086

Introduction: Nondepolarizing neuromuscular blockade (NDMB) is a key intervention to avoid ventilation-induced lung injury in acute respiratory distress syndrome (ARDS). In patients with moderate-severe ARDS associated with SARS-CoV-2 infection (CARDS), NDMB were used for prolonged periods of time, with high cumulative doses. We hypothesize that administration of NDMB might contribute to an increased incidence of risk factors later associated with long COVID-19.

Methods: We designed a non-interventional, retrospective study in a large university urban hospital. From January to December 2021, data related to prescription of NDMB, respiratory physiology, mechanical ventilation (IMV) and clinical outcomes were collected from patients’ electronic records with a diagnosis of CARDS. Primary outcome was day-90 mortality. Secondary outcomes were ICU length of stay (LOS), ICU-acquired weakness and days of IMV. Mann–Whitney U test was used to compare continuous variables and logistic regression was used to evaluate the association of NDMB use with outcomes, adjusted or not for confounders.

Results: 116 patients diagnosed with CARDS were included, 87% with severe ARDS and overall mortality was 37.1%. Median age was 57 years (IQ:47–67) and 65.5% were male. P:F ratio at day-1 was 86 (IQ:43). Ventilator-free days (VFDs) at 28 days was 13 (IQ:0–19) in survivors and ICU-LOS was 19 (IQ:10–36). Median time and cumulative dose of NDMB were, respectively, 117 h and 1177.468 mg in patients who survived (n = 70) compared to 197 h and 1898.775 mg in patients who died (n = 41). In addition to days of NDMB exposure (OR 1.05, CI 95% 1.00–1.11), the cumulative dose of cisatracurium, expressed in logs, was correlated with risk of mortality in the ICU, with odds ratio 1.49 (CI 95% 1.08–2.04).

Conclusions: Patients with severe forms of CARDS received prolonged infusions of NDMB, with high cumulative doses. Both time of exposure and total doses were independently associated with higher risk of mortality.


Prolonged prone position in COVID-19-associated ARDS: a prospective study

G Karlis1, S Kakavas2, M Poulou1, D Bakali1, G Katsagani1, M Daganou1, D Markantonaki1

1Thoracic Diseases General Hospital Sotiria, Polyvalent ICU, Athens, Greece, 2Henry Dunant Hospital, Athens, Greece

Critical Care 2023, 27(S1): P087

Introduction: Ventilation in prone-position (PP) improves survival in moderate-to-severe ARDS. However, optimal duration of the intervention to gain maximum benefit is unknown. We sought to examine the efficacy and safety of a prolonged PP protocol in COVID-19-associated ARDS.

Methods: This was a prospective observational study. We included consecutive intubated and mechanically ventilated patients with ARDS and positive PCR for SARS-CoV-2 who underwent at least one session of PP from March 2021 to August 2021. PP was undertaken if P/F < 150 with FiO2 > 0.6 and PEEP > 10cmH2O. Oxygenation parameters and respiratory mechanics were recorded before PP, at the end of PP session and 4 h after supine repositioning. Patients with PP longer than 24 h (prolonged group) were compared to patients who were proned for less than 24 h (control group). The duration of PP was at the discretion of the treating intensivist.

Results: We recorded 56 patients (62.7% male). Five patients were excluded because PP was terminated in less than 4 h. Mean age of the 51 studied patients was 61.4 years. Patients in the prolonged group had significantly higher BMI than controls. Baseline oxygenation and respiratory mechanics were similar between groups. PP duration was 39.8 versus 20.5 h (p < 0.001). Increase of P/F was higher in the prolonged PP group during proning (103.8 ± 70.8 vs 66 ± 53.9, p < 0.05) and after supination (76.3 ± 64.6 vs 48.6 ± 34.9, p = 0.058). No change in respiratory mechanics was observed in either group. 28-day survival was 75% in the prolonged PP group and 69.5% in the control group (p = 0.665). Duration of mechanical ventilation, number of PP cycles and rate of complications were similar between groups.

Conclusions: In patients with ARDS due to COVID-19 prolonged PP resulted in better oxygenation, but had no impact on outcome. However, it is both feasible and safe and can be an alternative in conditions of increased work load as was the case during the recent pandemic.


Incidence of barotrauma in an intensive care unit: new data from COVID-19?

JN Patrício1, R Jorge1, AA Fernandes2, A Gonçalves1, CS Pereira1

1Hospital Beatriz Angelo, Intensive Care Unit, Loures, Portugal, 2Hospital Beatriz Angelo, Internal Medicine, Loures, Portugal

Critical Care 2023, 27(S1): P088

Introduction: During COVID-19 pandemic, the massive use of ventilatory support made its complications even more common. This study aimed to analyse the incidence of barotrauma in COVID-19 patients as well as its consequences.

Methods: Retrospective cohort study. All patients undergoing mechanical ventilation in an intensive care unit (ICU) during 2020–2021 were included. The time of both noninvasive and invasive ventilation was considered together. Statistical analysis was performed using IBM SPSS Statistics 28.0.

Results: A total of 967 patients were included, with 42 cases of barotrauma being reported (28 men and 14 women, median age 69 years [interval 22–94] and median APACHE 13). Out of those, 40 had severe COVID-19. Regarding patients with and without COVID-19, the incidence of barotrauma (episodes/1000 days of ventilation) was 0.64 and 9.22 (RR 14.86, p < 0.001) and the barotrauma rate (episodes/number of patients) was 0.4% and 8.5% (RR 21.25, p < 0.001), respectively. The most common type of barotrauma was subcutaneous emphysema (52.4%, CI 95% 37.3–67.5%), followed by pneumomediastinum (47.6%, CI 95% 32.5–62.7%) and pneumothorax (35.7%, IC 95% 21.2–50.2%). The median time to diagnosis was 11.5 days after initiation of ventilatory support [interval 1–67]. In the COVID-19 group, barotrauma was associated with longer ventilation (14.06 vs 7.91 days, p < 0.001), longer ICU stay (16.74 vs 8.17 days, p < 0.001) e higher mortality rates (45.0% vs 26.2%, RR 1.72, p 0.011).

Conclusions: We found a higher susceptibility to developing barotrauma as a potential complication of COVID-19 patients undergoing mechanical ventilation. From those, subcutaneous emphysema and pneumomediastinum seem to be more prevalent than pneumothorax. Barotrauma seems to be associated with longer periods undergoing mechanical ventilation, longer ICU stays and higher hospital mortality rates.


Mortality in patients with COVID-19-related ARDS treated with venovenous extracorporeal membrane oxygenation: a retrospective observational study

R Antolini1, E Casarotta2, A Damia Paciarini3, G Giaccaglia1, E Vitali1, A Salvucci1, A Carsetti1, S Pantanetti3, V Gabbanelli3, A Donati1

1Università Politecnica delle Marche, Ancona, Italy, 2Università Politecnica delle Marche, Biomedical Sciences, Ancona, Italy, 3AOU delle Marche, Ancona, Italy

Critical Care 2023, 27(S1): P089

Introduction: Venovenous extracorporeal membrane oxygenation (VV ECMO) is a technique that provides blood oxygenation and CO2 removal, allowing a protective ventilation strategy until the resolution of respiratory failure. A delay in ECMO initiation could worsen the outcome and prolong the duration of treatment. The study aims to describe the incidence of mortality in our intensive care unit (ICU) in patients with severe COVID-19-related acute respiratory distress syndrome (ARDS) treated with VV ECMO.

Methods: We performed an observational retrospective study, including patients with severe COVID-19-related ARDS admitted to our ICU and treated with VV ECMO between February 2020 and February 2022. We collected data on demographic characteristics, comorbidities, mechanical ventilation, therapies, laboratory results, VV ECMO and ICU mortality. SOFA score, SAPS II and Charlson Comorbidity Index were calculated at ICU admission.

Results: The average age of our cohort of 60 patients was 54.4 ± 7.7 years and 51 (85%) were males. The mean value of the SOFA score at ICU admission was 7 ± 2.3 points, and the median value of the SAPS II score was 41 [31–48] points. The incidence of mortality in the whole cohort was 48.3%. The differences between the two groups of patients, Survivors and Non-survivors, are presented in Table 1. Through a multivariate logistic regression model we found that age (OR 1.09 [95% CI 1.00–1.19], p = 0.03) and lymphocytes (OR 0.09 [95% CI 0.01–0.59], p = 0.01) were significantly associated with ICU mortality. Mechanical ventilation before ECMO placement higher than 10 days and superinfections at ICU admission were not significantly associated with the outcome in the same model.

Conclusions: In patients with COVID-19-related ARDS treated with VV ECMO, advanced age and lymphopenia at ICU admission are risk factors for ICU mortality. A longer duration of mechanical ventilation before ECMO placement and traditional ICU prognostic scores seem not to be relevant for the prognosis.

Table 1 (abstract P089). Demographic and clinical characteristics of the two groups of patients


Survivors (n = 31)

Non-survivors (n = 29)

p value

Age, years

53 [48–59]

59 [50–61]


Charlson Comorbidity Index

1.2 ± 0.8

1.4 ± 0.9


SOFA score, points

7.1 ± 2.5

7 ± 2.2


Time between intubation and ECMO start, days

6 [3–9]

6 [2–9]


Gram + superinfection

25 (80.6)

24 (82.7)


Gram- superinfection

30 (96.8)

23 (79.3)


Lymphocytes, × 103/mmc

0.89 [0.46–1.25]

0.56 [0.40–0.71]

 < 0.01


Can we detect in vivo hemolysis during extracorporeal membrane oxygenation support using endogenous carbon monoxide production?

I Seises1, M Serres-Gomez2, M Simon-Velasco2, KL Nanwani-Nanwani1, MP Sanz-de-Pedro2, B Fernandez-Puntero2, MJ Alcaide-Martin2, I Losantos-Garcia3, P Millan-Estañ1, R Gomez-Rioja2

1University Hospital La Paz, Intensive Care Medicine, Madrid, Spain, 2University Hospital La Paz, Clinical Laboratory, Madrid, Spain, 3University Hospital La Paz, Biostatistics Department (IdiPaz), Madrid, Spain

Critical Care 2023, 27(S1): P090

Introduction: Extracorporeal membrane oxygenation (ECMO) has been widely used in patients with ARDS due to COVID-19. In vivo hemolysis (ivH) is one of its complications, characterised by peaks of plasma free hemoglobin (fHb). However, an increase in carboxyhemoglobin (COHb) has also been observed due to Hb metabolism by heme-oxygenase that releases carbon monoxide. The aim of this study is to evaluate the incidence of ivH events and their relation to COHb in COVID-19 patients undergoing ECMO.

Methods: Single-centre observational retrospective study that included 33 COVID-19 patients with ARDS who received VV-ECMO treatment in the ICU from March 2020 to September 2021. Daily analytical monitoring was carried out including arterial blood gas test with cooximetry and biochemical parameters, incorporating the estimation of fHb using quantitative hemolysis index (HI). Significant ivH was considered with fHb > 50 mg/dL after discarding in vitro hemolysis. Daily maximum values of HI and COHb were recorded and paired in order to evaluate their correlation by generalised linear model.

Results: The total prevalence of patients having ivH in our cohort was 27.3%. Mortality during ECMO treatment in our study was 57.6%, higher within the group of patients with ivH events (77.8% vs 50%). A total of 777 daily maximum values of fHb from all the patients were obtained. Values of COHb were significantly higher during ivH episodes. Furthermore, positive significant correlation was obtained between daily analytical values of fHb and COHb (B coefficient 42.156; p = 0.042), as shown in Fig. 1. The cut-off value of COHb to be discriminative for hemolysis (fHb > 50 mg/dL) was 3.85% COHb (90.5% sensitivity and 83.3% specificity).

Conclusions: Point-of-care carboxyhemoglobin is a cheap and widely available parameter that could be useful when detecting in vivo hemolysis during ECMO treatment.

figure aj

Fig. 1 (abstract P090). Correlation between daily analytical values of fHb and COHb


Bilateral non-reactive pupillary dilation during continuous muscle relaxant therapy: a retrospective study

Y Giladi1, A Giladi2, M Frist3, PD Levin1

1Shaare Zedek Medical Center, Intensive Care Unit, Jerusalem, Israel, 2The Hubrecht Institute, Utrecht, Netherlands, 3Shaare Zedek Medical Center, Anesthesiology, Jerusalem, Israel

Critical Care 2023, 27(S1): P091

Introduction: Fixed dilated pupils in sedated ICU patients can be the first sign of brain injury. Reversible pupil dilation under rocuronium treatment has been described in case reports [1, 2]. We report a retrospective analysis of patients receiving continuous rocuronium or vecuronium to investigate the occurrence of pupil dilation.

Methods: Nursing data on pupil size were collected retrospectively from all adult patients receiving a continuous infusion of rocuronium or vecuronium, both during and before/after the infusion. Reversible bilateral pupil dilation ≥ 4 mm that was not responsive to light was defined as dilation. Patients with an organic cause of dilated pupils were excluded.

Results: Data on 197 patients admitted from 2019 to 2022 were analyzed. Rocuronium only was administered to 42 (21%) patients, vecuronium only to 134 (67%) patients and both to 23 (12%) patients. For rocuronium, transient pupil dilation occurred in 2/42 (5%) patients during treatment versus 2/42 (5%) different patients before or after treatment. Pupils were dilated for 34 ± 19 (20 ± 5%) hours during therapy versus 10 ± 6 (8 ± 4%) hours before or after therapy. For vecuronium transient pupil dilation occurred in 3/134 (2%) patients during treatment versus 4/134 (3%) other patients before or after treatment. Pupils were dilated for 12 ± 6 (23 ± 31%) hours during therapy versus 52 ± 28 (13 ± 10%) before or after therapy.

Conclusions: Pupil dilation seems to be more common during muscle relaxant therapy than before or after. Overall pupil dilation under muscle relaxants is rare, but seemed more common during rocuronium therapy than vecuronium. The small numbers precluded statistical significance. Familiarity with this potential side effect of rocuronium (which can be reversed with sugammadex) is important as it may prevent unnecessary CT exams in ventilated patients.


  1. 1.

    Zakynthinos GE et al. J Crit Care 2021;65:259–260.

  2. 2.

    Fernandes FA et al. Braz J Anesthesiol 2022;72:829–831.


Readiness for malignant hyperthermia crisis in intensive care unit

EDW Du, TCL Tan, ZXY Zhang

Changi General Hospital, Anaesthesiology and Intensive Care, Singapore, Singapore

Critical Care 2023, 27(S1): P092

Introduction: Malignant hyperthermia (MH) is a rare, life-threatening condition that manifests as a hypermetabolic response triggered by exposure to volatile anaesthetic agents and succinylcholine [1]. Prompt recognition and early treatment of a MH crisis improve survival rate. Inhaled sedation with volatile anaesthetic agents has been made feasible in intensive care unit (ICU) with the development of the anaesthetic conserving device (AnaConDa®) [2]. With its rising popularity among the intensivists, it calls for a MH crisis preparation plan for the ICU [3]. The objective of this study is to assess MH awareness, knowledge and crisis preparedness among ICU staff.

Methods: An online questionnaire was sent to staff working in the operating theatre (OT) and ICU in Changi General Hospital Singapore between August to October 2022. Staff included are doctors working in the anaesthesia department, nurses and attendants in the OT, ICU nurses and respiratory therapists. The survey consists of 16 single or multiple-choice questions on MH knowledge and crisis preparedness.

Results: Out of the 283 responders, 49.5% (140) are unsure or unaware of MH, among which ICU staff accounts for 71.5% (100). Among 114 ICU staff, only 14.0% correctly identified the signs and symptoms of MH or knew dantrolene is the drug of treatment. 17 people had prior encounters with MH either in ICU or OT settings, among them only 53% (9) chose the correct drug and dose of dantrolene. Majority of the ICU staff are not aware of the hospital MH protocol or availability of MH cart and 99.1% expressed lack of confidence in handling MH crisis (Fig. 1).

Conclusions: MH is a low-frequency, high fatality emergency in ICU. Current data showed a lack of awareness and preparedness for MH crisis among the staffs especially ICU colleagues. MH education plan is needed to promote early recognition, intervention and treatment of MH crisis.


  1. 1.

    Hopkins PM et al. Anaesthesia. 2021;76:655–664.

  2. 2.

    Rosenberg H et al. Can J Anesth 2015;62:319–320.

figure ak

Fig. 1 (abstract P092). Malignant hyperthermia knowledge and crisis preparedness among ICU staff


Pain is a problem during critical care

A Wlodarczyk-Abou Elseoud1, K Hamunen2, ML Kalliomäki3, M Bäcklund4, A Laukkanen4, O Mäkinen5, AM Kuivalainen4

1HUS Helsinki University Hospital, Department of Intensive Care, Helsinki, Finland, 2HUS Helsinki University Hospital, Department of Pain Medicine, Helsinki, Finland, 3Tampere University Hospital, Department of Prehospital Emergency Care, Pain Management and Anaesthesiology, Tampere, Finland, 4HUS Helsinki University Hospital, Department of Intensive Care, Helsinki, Finland, 5Tampere University, Tampere University, Tampere, Finland

Critical Care 2023, 27(S1): P093

Introduction: Many patients in the Intensive Care Unit (ICU) suffer from insufficient pain treatment. This leads to delay in the recovery and may evolve into chronic pain [1,2]. More information about pain intensity fluctuation and applied pain treatment throughout the stay in the ICU is needed. We conducted observational study to examine risk factors and incidence of moderate to severe pain during the first 10 days of stay in the ICU.

Methods: We recruited mixed population of 711 critically ill adult patients from two university hospitals in Finland. Numeric Rating Scale (NRS) and Verbal Rating Scale (VRS) were used to assess pain in patients able to communicate and Critical Care Pain Observational Tool (CPOT) for non-communicating patients. The primary endpoint of the study was the incidence of moderate to severe pain during intensive care (NRS ≥ 4/CPOT ≥ 3/VRS ≥ moderate pain). We collected data including pain intensity and other pain related factors during first 10 days in the ICU.

Results: 76% patients experienced moderate to severe pain during at least one of ten days of follow up in the ICU. Age below 64 years old, female gender and history of chronic pain was associated with higher number of days in moderate to severe pain. There was not significant difference between surgical and non-surgical groups of patients in the days they experienced moderate to severe pain. Pain scores throughout 10 days follow up are presented in Fig. 1.

Conclusions: Our results indicate that majority of patients experienced moderate to severe pain during ICU stay. We need to improve the pain assessment, especially in high-risk ICU patients. Further studies of pain trajectories during critical illness as well as identifying patient and treatment related risk factors for severe pain, will help improve pain treatment protocols and patients’ outcomes.


  1. 1.

    Puntillo KA et al. Am J Respir Crit Care Med 2014;189:39–47.

  2. 2.

    Payen JF et al. Anesthesiology 2009;111:1308–16.

figure al

Fig. 1 (abstract P093). Patients’ pain scores during 10 days follow up in the ICU. Patient had moderate to severe pain if NRS ≥ 4/CPOT ≥ 3/VRS ≥ moderate pain


Multimodal versus opioid analgesia in cardiac surgery (MONDAY)

NB Belkaid1, SB Bogaert2, AM Moerman3, SB Bouchez4, HP Peperstraete1

1Ghent University Hospital, Department of Intensive Care Medicine, Ghent, Belgium, 2AZ Groeninge, Department of Anesthesiology and Intensive Care, Kortrijk, Belgium, 3Ghent University Hospital, Department of Anesthesiology, Ghent, Belgium, 4OLV Ziekenhuis, Department of Anesthesiology, Aalst, Belgium

Critical Care 2023, 27(S1): P094

Introduction: Cardiac surgery performed by sternotomy is associated with moderate to severe acute postoperative pain. Our goal is to compare standard opioid based regimen to a multimodal pain management to determine which therapy provides the most comfort, the fastest extubation time, the least pain and the least delirium.

Methods: In this prospective randomized double-blinded trial we evaluated two groups of 50 patients, divided in an opioid group and a multimodal group. In the multimodal group pregabaline 75 mg, dexmedetomidine, ketamine and lidocaine was given versus liberal fentanyl and remifentanil in the opioid group. Postoperative pain was scored every 8 h by the NRS-scale/CPOT score in the awake/sedated patient. Delirium postoperatively is measured by the ICDSC-score. Time to extubation, LOS in ICU, LOS in hospital and rescue pain medication are compared between both groups.

Results: Ninety-six patients were successfully enrolled. There were no differences in baseline characteristics between both groups. The multimodal pain scheme was administered in 46 patients, while 50 patients received the opioid based scheme. After linear mixed model analysis no significant differences were seen in CPOT-score (p = 0.305), NRS-score (p = 0.182) (Fig. 1), consumption in rescue pain medication and ICDSC-score (p = 0.267). In the opioid group, mean LOS in ICU was 34 h in the opioid group versus 51 h in the multimodal group, (p = 0.032). Time to extubation or LOS in hospital showed no significant difference between both groups.

Conclusions: This trial shows a significant difference in LOS in ICU in favor of the opioid group. No significant difference could be found in pain, delirium, intubation time or LOS in hospital between multimodal analgesia versus an opioid-based analgetic protocol in cardiac surgery by sternotomy. This is not in line with previous studies but is probably because this trial is underpowered. It shows that there is a place for multimodal analgesia in cardiac surgery and it can be a good alternative for opioid based protocols.

figure am

Fig. 1 (abstract P094). Mean NRS-score in both groups


Analgesic potential of a serratus anterior plane block, placed on ICU arrival, after totally endoscopic aortic valve replacement (TEAVR)

J Dubois1, I Callebaut2, D Vranken2, I Gruyters2, H Jalil2, M Van Tornout2, M Brands 2, J Herbots2, B Stessel2, J Vandenbrande2

1Jessa, Critical Care, Hasselt, Belgium, 2Jessa, Anesthesiology & Intensive Care, Hasselt, Belgium

Critical Care 2023, 27(S1): P095

Introduction: Regional anesthesia is of growing importance in postoperative pain control. We examined whether a Serratus Anterior Plane Block (SAPB), placed on arrival at the Intensive Care Unit (ICU), in addition to standard-of-care after TEAVR could improve the quality of postoperative pain management.

Methods: We performed the first double-blinded randomized controlled trial on 75 patients undergoing TEAVR. Patients allocated to the intervention arm received a SAPB on ICU arrival when the patient was sedated. SAPB was placed by a block team, leaving the attending anesthesiologist, surgical team and critical care team blinded. Ultrasound-guided SAPB was performed on the right hemithorax at T2-T5 level: 30 ml of bupivacaine 0.25% was injected under the serratus anterior (SA) muscle and 10 ml bupivacaine 0.25% was administered between the latissimus dorsi and SA muscle. Control patients were approached by the block team for the same time span, though no SAPB was performed. Further analgesic treatment included acetaminophen and ketorolac unless contra-indications, as well as piritramide patient-controlled analgesia (PCA). Sedation was provided by means of dexmedetomidine.

Results: Thirty-seven patients were allocated to the control versus 38 to the SAPB group. Patients receiving SAPB suffered significantly less pain at 4, 8 and 24 h postoperatively (p values: 0.05, 0.03 and 0.04 respectively) (Fig. 1). Also 24 h-cumulative piritramide PCA consumption was significantly lower in the SABP group (median 12 mg, IQR 8–26 mg) versus control group (median 20 mg, IQR 15-31 mg) (p < 0.01). No differences in intraoperative use of sufentanil (p = 0.20) and ketamine (p = 0.59) neither postoperative use of NSAIDS (p = 0.18) and dexmedetomidine (p = 0.39) were observed. Time to extubation neither ICU length of stay was influenced by SAPB.

Conclusions: SAPB added to standard-of-care after TEAVR improved postoperative pain control, by significantly decreasing opioid consumption and pain scores in the first 24 h after surgery.

figure an

Fig. 1 (abstract P095). Evaluation of pain by a blinded assessor at predefined time points after ICU arrival. Pain was assessed by means of the Numeric Rating Scale (NRS) in which 0 = no pain versus 10 = worst pain ever experienced. Statistically significant differences are marked with an asterisk (*)


Impact of fentanyl versus remifentanil in critical care outcomes after cytoreductive surgery plus hyperthermic intraperitoneal chemotherapy (CRS-HIPEC)

G Madrid, A Obando, JD Guerra, JP Caceres

University Hospital Fundacion Santa Fe de Bogota, Bogotá, Colombia

Critical Care 2023, 27(S1): P096

Introduction: Anesthetic techniques and intraoperative care have an impact on outcomes after major surgery [1]. Remifentanil and fentanyl are strong opioids widely used intraoperatively due to their potency and pharmacokinetic properties. We designed a study to evaluate associations between these opioids and critical care and clinical outcomes in patients who underwent CRS-HIPEC.

Methods: A historical cohort of patients taken to CRS-HIPEC from 2007 to 2020 at a university hospital in Bogota, Colombia was analyzed retrospectively. We compared outcomes of patients who received intraoperative remifentanil (group R) or fentanyl (group F). Primary outcome was ICU length of stay (LOS), and secondary outcomes were total LOS, ICU readmission, inpatient mortality, time vented, need for tracheostomy, time to ambulation and oral feeds, need for transfusion, estimated blood loss, intraoperative crystalloids, and units of red blood cells and plasma transfused in surgery. We excluded patients with incomplete medical records or whose procedure was palliative.

Results: A total of 173 patients were included in the analysis. Mean age was 52 ± 11.4 years and female sex was preponderant (73%). 75 patients (43%) received intraoperative remifentanil and 98 (57%) fentanyl. Group R was associated with a shorter ICU stay (mean 5.9 days vs. 6.4 days in group F, p < 0.018). There were also statistically significant associations regarding days until extubation (Group R 3.1 vs 3.7 Group F, p < 0.01), need for transfusion, and estimated blood loss (Table 1).

Conclusions: Our results suggest an association of remifentanil and shorter ICU LOS and earlier extubation. We believe that the difference, although statistically significant, is arguably clinically significant.


  1. 1.

    Myles PS et al. Anesth Analg 2002;95:805–12.

Table 1 (abstract P096). Results






Mean (SD)

Median IQR

Mean (SD)

Median IQR

ICU stay*—days

6.4 (15)

3 (3)

5.9 (6.5)

3 (4)

Days on vent*

3.7 (16)

1 (1)

3.1 (12)

1 (1)

Estimated blood loss*—ml

1756 (1944)

1000 (1490)

2137 (1835)

1765 (2000)


Yes (%)

No (%)

Yes (%)

No (%)


60 (61)

38 (39)

64 (85)

11 (15)

  1. *Correlations with Mann–Whitney test and p < 0.05; ^Correlations with Pearson χ2t and p < 0.05


Ketamine-fentanyl versus fentanyl for analgosedation in postoperative surgical intensive care unit patients.

N Kitisin1, K Wongtangman1, P Kumtip2, O Chaiwat1, N Raykateeraroj1, P Somnuke1

1Faculty of Medicine, Siriraj Hospital, Mahidol University, Department of Anesthesiology, Bangkok, Thailand, 2Suratthani Hospital, Department of Anesthesiology, Suratthani, Thailand

Critical Care 2023, 27(S1): P097

Introduction: Ketamine at sub-anesthetic dose has been shown to provide sedative-analgesic effect and decrease opioid requirement in postoperative patients. We plan to evaluate the effects of low-dose ketamine on opioid consumption in adult surgical critically ill patients and observe its side effects.

Methods: A randomized, double-blinded controlled trial was conducted in the postoperative ICU patients. Fentanyl was initially titrated until Numerical Rating Scale (NRS) was < 4 and Richmond Agitation-Sedation Scale (RASS) -2 to 0. Patients received fentanyl 20 mcg/hour continuously with either ketamine (Group K) or placebo (Group C). Ketamine 0.09 mg/kg/hour (1.5 mg/kg/min) was administrated in Group K, and 0.9%NaCl was administrated at the same rate in Group C. The ICU nurse assessed NRS and RASS score in every 4 h and gave extra fentanyl bolus intravenously to keep the patients’ NRS < 4 and RASS 0 to − 2.

Results: 118 patients were included in the study. The patients’ characteristics were not significantly different between two groups. The fentanyl consumption at 4–8 h and 8–12 h after randomization in a ketamine group have significantly less fentanyl consumption compared to a control group (91.8 mcg vs 121.3 mcg, p = 0.002 and 95.6 mcg vs 121.1 mcg respectively, p = 0.012) (Fig. 1). The cumulative fentanyl consumption at 24 h was not different between ketamine and control group (445.8 mcg vs 499 mcg, p = 0.38). The ICU length of stay, duration of mechanical ventilation and adverse effects were not significantly different.

Conclusions: The use of low dose ketamine combined with fentanyl can reduce fentanyl consumption in postoperative SICU patients especially in the first 12 h without increase any adverse events.

figure ao

Fig. 1 (abstract P097). Mean fentanyl consumption (microgram) in each 4 h after randomization


Enteral clonidine in adult intensive care patients experiencing drug tolerance and iatrogenic withdrawal with weaning of dexmedetomidine

JR Berkan, TS Lam, JT Jancik

Hennepin Healthcare, Pharmacy, Minneapolis, USA

Critical Care 2023, 27(S1): P098

Introduction: Iatrogenic withdrawal has been reported in up to 73% of patients exposed to dexmedetomidine (DEX) for as few as 3 days [1, 2]. Use of enteral clonidine (CLN) to wean off DEX and manage withdrawal has become more prevalent in the intensive care unit (ICU). This study aimed to evaluate the relationship between DEX exposure and initiation of CLN in adult patients in the ICU.

Methods: A single-centered, retrospective cohort study was conducted in adult ICU patients admitted between January 1, 2021 and July 31, 2022 who received at least 48 h of continuous DEX. Patients weaned off DEX alone were compared to patients weaned with enteral CLN. The primary outcome was time (days) from start of wean to cessation of DEX. Secondary outcomes included DEX exposure and CLN utilization. Withdrawal (RASS > 1, HR > 100, and re-initiation of DEX or start of CLN within 48 h of DEX cessation) was assessed in patients weaned off DEX alone.

Results: A total of 267 patient encounters were included. Total days on DEX was greater in the CLN group (6 vs. 3 days). Days on DEX prior to wean were similar between groups, with no difference in weight-based dose of DEX (1 vs. 1 mcg/kg/hr). Time from start of wean to cessation of DEX was longer in the CLN group (3 vs. 0 days). The most common initial CLN regimen was 0.1 mg every 8 h, with a median total daily CLN dose of 0.4 mg at time of DEX cessation. At 48 h, 18.1% of patients weaned off DEX alone were re-initiated on DEX or started on CLN.

Conclusions: Use of enteral CLN to wean DEX may reduce rates of DEX resumption due to iatrogenic withdrawal.


  1. 1.

    Bhatt K et al. Crit Care Expl. 2020;2:e0245.

  2. 2.

    Glaess S et al. Am J Health Syst Pharm. 2020;77:515–522.


Bispectral index is useful for assessing sleep quality in postoperative patients in ICU

P Sirilaksanamanon1, S Charuluxananda1, T Thawitsri1, N Chirakalwasan2

1King Chulalongkorn Memorial Hospital, Faculty of Medicine, Chulalongkorn University, Department of Anesthesiology, Bangkok, Thailand, 2King Chulalongkorn Memorial Hospital, Faculty of Medicine, Chulalongkorn University, Department of Medicine, Bangkok, Thailand

Critical Care 2023, 27(S1): P099

Introduction: Monitoring and improving sleep quality may help to enhance recovery from critical illness. Although polysomnography is the gold standard for assessing sleep quality, it remains impractical for routine use due to its complexity and cost. We aimed to investigate diagnostic accuracy of alternative monitor which is Bispectral Index (BIS) in assessing sleep quality in postoperative patients in ICU.

Methods: This study was an observational study. We included 33 adult patients who underwent elective major surgery and required postoperative ICU admission with mechanical ventilation. After postoperative admission to the ICU, polysomnography and BIS were recorded overnight and the data were obtained for analyses. The primary outcome of this study was BIS index best predicting good sleep quality. We defined postoperative sleep quality by polysomnographic sleep efficiency more than 85%.

Results: Receiver operating characteristics (ROC) curve (Fig. 1) was created to determine the averaged BIS index associated with good postoperative sleep quality. Area under the ROC curve (AUC) was 0.65. The optimal cut-off with best differentiation derived from Youden’s Index was 75 with sensitivity 68% and specificity 56%. When comparing the subjects with good and poor postoperative sleep quality, major postoperative outcomes including length of mechanical ventilation and ICU stay were not significantly different. Although all subjects with postoperative delirium had poor postoperative sleep quality, incidence of delirium between group was not significantly different (0% vs 10.3%; p = 0.184).

Conclusions: Bispectral Index (BIS) monitoring is a feasible tool for assessing sleep quality with acceptable accuracy in postoperative patients requiring mechanical ventilation in ICU. However, larger controlled studies are required to investigate the impact of sleep quality monitoring on clinical outcomes.

figure ap

Fig. 1 (abstract P099). ROC curve of BIS associated with good postoperative sleep quality


Pursuing the real midazolam clearance during continuous renal replacement therapy in ICU patients with COVID-19: are midazolam and metabolites dialyzable?

TJL Smeets1, HRH De Geus2, BCP Koch1, D Gommers2, H Endeman2, NGM Hunfeld1

1Erasmus MC University Medical Center, Hospital Pharmacy, Rotterdam, Netherlands, 2Erasmus MC University Medical Center, Intensive Care Adults, Rotterdam, Netherlands

Critical Care 2023, 27(S1): P100

Introduction: Midazolam based continuous iv sedation became again prominent during the COVID-19 pandemic. However, this sedation therapy is associated with a high incidence of benzodiazepine-related delirium and an increased number of days spent in coma. Given the high midazolam dose requirements in some patients and due to the renal clearance (CL) of the active metabolite 1-OH-midazolam-glucuronide (OHmidazolamGluc), ICU patients with COVID-19 and continuous renal replacement therapy (CRRT) may are at risk of prolonged sedation. Therefore, the aim of the study was to investigate the CL of midazolam and metabolites in 5 critically ill COVID-19 patients with CRRT.

Methods: Pre-filter blood samples and ultrafiltrate samples were collected simultaneously. Midazolam, 1-OH-midazolam (OHmidazolam) and OHmidazolamGluc plasma samples were analysed by an UPLC-MS/MS method. CL of midazolam and metabolites were calculated by the delivered renal dose and sieving (SC) coefficient. Subsequently, the CL and delivered renal dose were corrected for downtime therapy and filter integrity by a filter urea ratio.

Results: We included 4 cases of CVVHD and 2 cases of CVVHDF. Midazolam, OHmidazolam and OHmidazolamGluc concentrations in μg/l ranged from 0 to 6070, 0 to 295 and 1727 to 39,000, respectively. SCs ranged from 0.02 to 0.03 for midazolam, 0.05 to 0.06 for OHmidazolam and 0.23–0.43 for OHmidazolamGluc. The CL in ml/min by the delivered renal dose was 0.82–1.67 for midazolam, 2.20–3.46 for OHmidazolam and 4.0–27.65 for OHmidazolamGluc. The CL in ml/min by the corrected renal dose was 0.68–1.50, 1.83–2.33 and 3.40–25.4, respectively. The urea ratios were 0.53 to 1.0.

Conclusions: Midazolam and OHmidazolam are not removed efficiently by CRRT and OHmidazolamGluc approximately up to 43%. Type of CRRT, filter integrity and downtime of CRRT affect the CL of midazolam and metabolites. Our results have implications for more personalized titration of midazolam in COVID-19 patients with CRRT, mainly to avoid oversedation.


Incidence of nephrogenic diabetes insipidus during prolonged sevoflurane sedation in the intensive care unit: a retrospective analysis

KVE Van Eeckhout1, BS Sol2, JG Glissenaar3, JP Poelaert1, BB Bravenboer4, JDF De Filette5

1Universitair Ziekenhuis Brussel, Intensive Care Unit, Jette, Belgium, 2AZ Alma, Endocrinology, Eeklo, Belgium, 3Universitair Ziekenhuis Brussel, Anesthesiology and Perioperative Medicine, Jette, Belgium, 4Universitair Ziekenhuis Brussel, Endocrinology, Jette, Belgium, 5CHU Brugmann, Endocrinology, Brussels, Belgium

Critical Care 2023, 27(S1): P101

Introduction: Sevoflurane is a halogenated inhalational anesthetic used for maintenance anesthesia. It is used in the intensive care unit (ICU), not in the least because of its potentially beneficial effects in patients with Acute Respiratory Distress Syndrome. However, polyuria and nephrogenic diabetes insipidus (NDI) were observed with its long-term administration. The objective of this study is to assess the incidence of NDI during prolonged sevoflurane sedation.

Methods: A retrospective analysis of patients receiving prolonged sevoflurane sedation (> 24 h) at the University Hospital of Brussels from 2013 to March 2021 was performed. The diagnosis of NDI was ‘confirmed’ in the presence of polyuria (urine output > 40 ml/kg/24 h), hypernatremia > 145 mmol/l, plasma osmolality > 295 mOsm/kg, urine osmolality < 300 mOsm/kg, and no or minimal response to exogenous desmopressin administration. Partial DI was established when urine osmolality was 300–600 mOsm/kg. A ‘possible’ diagnosis of DI was suggested when polyuria, hypernatremia and high plasma osmolality occurred during sevoflurane sedation in the absence of any other reasonable pathology.

Results: A total of 106 patients were included. Table 1 illustrates the baseline characteristics of the study population. NDI was confirmed in 9/106 patients (8.5%), of which 5 patients with partial NDI and 4 patients with complete NDI after desmopressin administration. Possible DI was observed in 16/106 patients (15%), but could not be confirmed due to absence of urine osmolality and/or density.

Conclusions: The occurrence of NDI during prolonged sevoflurane sedation is not uncommon and may be underestimated by clinicians. Close monitoring of diuresis, plasma sodium, plasma and urinary osmolality during prolonged sevoflurane administration is advisable for early diagnosis and treatment of NDI.

Table 1 (abstract P101). Demographics of patients receiving prolonged sevoflurane sedation in the intensive care unit


No NDI group

Confirmed NDI group

Possible NDI group

Age (years), mean ± SD

59.6 ± 14.9

55.0 ± 15.4

55.7 ± 9.1

Male sex (%)




SAPS II on ICU admission, median (IQR)

41 (28)

22 (24)

31.5 (29)

Baseline plasme Na (mmol/l), mean ± SD

141 ± 4.6

140 ± 3.52

141 ± 2.38

Baseline urine output (ml/24 h), mean ± SD

1963 ± 1321

1491 ± 862

2249 ± 1754

Sevoflurane end-tidal concentration (%), mean ± SD

1.11 ± 0.28

1.38 ± 0.50

1.22 ± 0.30

Sevoflurane exposure time (h), mean ± SD

127 ± 98

190 ± 79

166 ± 63


A systematic review and meta-analysis of the incidence of adverse events in adults undergoing procedural sedation in the emergency department

P Samara

University of Leicester, MRes Applied Health Research, Leicester, UK

Critical Care 2023, 27(S1): P102

Introduction: Procedural sedation is used in the emergency departments to facilitate the implementation of painful and distressing procedures. Several medications are used and their safety is quantified by estimating their adverse events. This systematic review and meta-analysis aimed to evaluate the incidence of adverse events of pharmacological agents used for procedural sedation of haemodynamically stable adults in the emergency department.

Methods: Three electronic databases were searched (Medline, Embase, Cochrane) to identify relevant literature after 2010. Randomised controlled trials comparing medications used for procedural sedation and their side effects were included in the review. After evaluating their quality, data extraction was completed for study characteristics and the outcomes of interest (agitation, hypotension, bradycardia, oxygen desaturation, apnoea, nausea/vomiting, intubation, laryngospasm, aspiration and subclinical respiratory depression. A meta-analysis of the results of the individual studies was performed using the fixed-effects model and reported as incidence with 95% confidence intervals.

Results: The search revealed 1004 studies, eight of which fulfilled the eligibility criteria and were included in this review. A total of 1103 patients were included in this review and meta-analysis, which focused on the study of ketamine, propofol, ketamine combined with midazolam and ketamine combined with propofol. Generally, the incidence of agitation was the highest at 14.90% (95% CI 12.67, 17.35) and the incidence of nausea and vomiting was the lowest at 4.30 (95% CI 12.67, 17.35). The rates of adverse events were lower with ketofol and higher with ketamine alone. There were no serious events present in this review.

Conclusions: A decision on the choice of medication used for procedural sedation depends on many factors. According to this review, there is potential in the combination of propofol and ketamine because it reduces the incidence of some of the adverse events present when using these drugs alone.


Methadone for sedation weaning in patients receiving extracorporeal membrane oxygenation: a retrospective observational study

C Remmington1, V Liew2, F Hanks1, L Camporota3, A Sousa3, O Stubbs3, N.A Barrett3

1Guy’s and St Thomas’ NHS Foundation Trust, Pharmacy Department, London, UK, 2Royal Melbourne Hospital, Department of Intensive Care, Melbourne, Australia, 3Guy’s and St Thomas’ NHS Foundation Trust, Department of Adult Critical Care, London, UK

Critical Care 2023, 27(S1): P103

Introduction: The aim of this study is to describe the effect of methadone on analgesia and sedation weaning and reported adverse effects in ECMO. Weaning analgesia and sedatives can be challenging following recovery from critical illness. Methadone is increasingly being used to facilitate weaning, however, limited data exists describing its use as a weaning strategy in adult patients receiving extracorporeal membrane oxygenation (ECMO) [1, 2].

Methods: Single-centre retrospective observational study including adult patients receiving ECMO for severe respiratory failure who were administered methadone between 2016 and 2022. We recorded demographics, methadone and sedation doses, adverse reactions and Richmond Agitation Sedation Scale (RASS) scores.

Results: Twenty-five patients were included. Median (IQR) age and weight were 45 (34, 52) years and 80 (75, 95) kg respectively. More patients were male (68% vs 32%). Median (IQR) APACHE II score were 17 (14, 20). Median (IQR) daily starting and peak doses were 30 (20, 40) and 40 (30, 50) milligrams (mg). Median (IQR) treatment duration and time to starting methadone from admission to ICU were 7 (5, 17) and 26 (5, 39) days. Median (IQR) daily doses of fentanyl, midazolam and propofol at day 0 and 5 days post methadone initiation were 6200 (4613, 8600) vs 3150 (0, 6042) micrograms/day; p < 0.001, 241 (151, 410) vs 130 (10, 273) mg/day; p = 0.30 and 4000 (1874, 5593) vs 3090 (0, 4373) mg/day; p = 0.04, respectively. Median (IQR) percentage of time in target RASS -2 to + 1, improved from 50 (5, 89) % to 88 (0, 100) %; p = 0.004. Adverse effects were: QTc prolongation in 2/25 patients, constipation in 2/25 patients and hypotension requiring noradrenaline in 1/25 patient.

Conclusions: Methadone was safe and effective at reducing analgesia and sedation in ECMO patients. Further studies are required to assess the effectiveness of methadone in this population.


  1. 1.

    Al-Qadheeb NS et al. Ann Pharmacother. 2012; 46:1160–6.

  2. 2.

    Wanzuita R et al. Crit Care. 2012;16:R49.


Sedation in the ICU: association between dexmedetomidine and 28-day mortality in patients with COVID-19

P Costa1, J Alves1, A Martinho1, C Silva1, L Simões1, G Almeida1, J Guimarães2, S Teixeira1, JP Baptista1, P Martins1

1Coimbra’s University Hospital, Intensive Care Department, Coimbra, Portugal, 2Coimbra’s University Hospital, Cardiology Department, Coimbra, Portugal

Critical Care 2023, 27(S1): P104

Introduction: The aim of our study is to evaluate the relationship between dexmedetomidine (DEX) use as a sedative agent in mechanical ventilated ICU patients and 28-day mortality. DEX, a selective alfa-2 adrenergic receptor agonist, widely used for its sedative and analgesic properties, has been linked to increasing parasympathetic tone, reducing the inflammatory response and oxidative stress [1]. Since severe COVID-19 is associated with an hyperinflammatory state, it is hypothesized that DEX might improve outcomes in these patients.

Methods: This is a retrospective observational study of mechanically ventilated patients admitted with COVID-19 pneumonia in the ICU of a tertiary center in Portugal, between March 2020 and December 2021. Logistic regression analysis was performed to evaluate the association of DEX use and 28-day mortality from time of intubation.

Results: A total of 277 patients were analyzed, 151 in the DEX group and 126 in the no DEX group. Patients in the DEX group were younger (53.3 vs. 63.3 years, p < 0.001), had less comorbidities (2.8 vs. 3.5, p = 0.01), lower SOFA at admission (6.2 vs. 7.1, p = 0.01) but had a prolonged ICU stay (21.4 vs. 15.9, p < 0.001). Male gender (65.6 vs. 69.0, p = 0.54), incidence of obesity (56.3 vs. 46.8, p = 0.12), coronary artery disease (4.0 vs. 7.9, p = 0.16) and atrial fibrillation (4.0 vs. 7.1, p = 0.25) were similar between groups. PaO2/FiO2 ratio at admission (111.1 vs. 108.1, p = 0.61), days spent in RASS < 3 (13.7 vs. 12.4, p = 0.31) and opioid use (14.8 vs. 13.1, p = 0.16) were also similar. From time of intubation, 28-day mortality in the cohort receiving DEX was 14.7% compared to 59.5% in the no DEX group (OR 0.12; 95% CI 0.07–0.21; p = 0.01).

Conclusions: Use of DEX was associated with lower 28-day mortality in COVID-19 critically ill patients requiring invasive mechanical ventilation in our study analysis. Considering the limitations of a retrospective observational study, RCTs are needed to confirm the results.


  1. 1.

    Ma J. Anesth Analg. 2020;130:1054–1062.


Safety and efficacy of phenobarbital for the treatment of alcohol withdrawal syndrome in patients admitted to the medical intensive care unit

CE Monzel, KA Considine, TS Lam, JT Jancik

Hennepin Healthcare, Pharmacy, Minneapolis, USA

Critical Care 2023, 27(S1): P105

Introduction: Sixteen to thirty-one percent of medical intensive care unit (MICU) patients have alcohol use disorder [1]. Benzodiazepines are first line therapy for alcohol withdrawal syndrome (AWS); however, phenobarbital (PHB) is emerging as an alternative therapy. PHB has demonstrated safety and comparative efficacy in reduction of withdrawal symptoms [2,3]. This study evaluated the safety and efficacy of PHB for the treatment of severe AWS in the MICU.

Methods: This retrospective, single-center, cohort study included patients ≥ 18 years of age admitted January 1, 2021, through July 31, 2022, to the MICU service for management of AWS. Patients who received diazepam monotherapy were compared to patients who received ≥ 1 dose of PHB ± diazepam for AWS. The primary outcome was to assess the difference in rate of intubation within 72 h of PHB administration compared to diazepam.

Results: A total of 67 PHB encounters and 98 diazepam monotherapy encounters were included, which yielded intubation rates of 55% and 41.4% within the 72-h timeframe, respectively. The most frequent indication for intubation within the 72-h timeframe was for airway protection which occurred in 54.6% of PHB intubations and 70.8% of diazepam monotherapy intubations. The mean PHB dose administered was 479.98 mg/dose (SD ± 223.3) or 6.48 mg/kg/dose (SD ± 2.97) based on ideal body weight. Mean diazepam dose administered over 72 h was 97.25 mg (SD ± 101.54) in the diazepam monotherapy group.

Conclusions: The use of PHB for control of AWS was associated with a higher rate of intubation compared to the diazepam group.


  1. 1.

    Dixit D et al. Pharmacotherapy. 2016;36:797–822.

  2. 2.

    Oks M et al. J Intensive Care Med. 2020;35:844–850.

  3. 3.

    Ammar MA et al. Ann Pharmacotherapy. 2021;55:294–302.


Valproic acid for agitation: retrospective analysis of psychiatric consults

M Perreault1, N Adessky2, A Regazzoni2, G White2, FW Chen2, L Tourian3, M Chagnon4, A Gursahaney5, M Alharbi3, D Williamson6

1The Montreal General Hospital, Pharmacy Suite C1-200, Montreal, Canada, 2McGill University Health Center, Dept of Pharmacy, Montreal, Canada, 3McGill University Health Center, Dept of Psychiatry, Montreal, Canada, 4Université de Montreal, Dept of Mathematics and statistics, Montreal, Canada, 5McGill University Health Center, Dept of Critical Care, Montreal, Canada, 6Université de Montreal, Faculté de Pharmacie, Montreal, Canada

Critical Care 2023, 27(S1): P106

Introduction: Agitation is a common problem in the intensive care unit (ICU). Treatment options are based on clinical experience and sparse quality literature. Our aim was to evaluate the effect of valproic acid (VPA) as adjuvant treatment for agitation in the ICU as well as identify independent predictors of response.

Methods: A retrospective single center observational study evaluated adult ICU patients for whom a psychiatric consultation was requested for agitation (defined as a RASS ≥ 2) from 2015 to 2021. A descriptive analysis of the proportion of agitated patients per day of follow-up, incidence of agitation-related-events (ARE) defined as device removal, as well as evolution of co-medications use over time are presented. A logistic regression model was used to assess predictors of VPA response for agitation and GEE models were used to evaluate the independent effect of VPA as an adjuvant therapy for agitation.

Results: 175 patients met inclusion criteria with 78 receiving VPA and 97 as a control group. Reason for admission in the VPA group was mostly TBI (62.8%) and causes of agitation were multifactorial. 97.4% were mechanically ventilated. The percentage of agitation-free patients was 6.5% on Day 1, 14.1% on Day 3 and 39.5% on Day 7 where Day 1 = day of VPA initiation.The percentage of patients with at least one ARE was highest on Day 1 with 14.3% and lowest on Day 4 with 6.76%. Multivariate regression model for variables identified female sex (n = 12) as predictor of response on Day 7 (OR 6.10 [1.18–31.64]). A significant effect of VPA on agitation was not observed when controlling for confounding variables.

Conclusions: Although VPA was associated with a decrease in agitation and ARE in the 7 days following initiation, its effect when controlling for confouding variables did not yield significant results. Prospective placebo-controlled studies are warranted to assess the natural evolution of agitation and the true impact of VPA on agitation and ARE in this population.


Valproic acid for the treatment of ICU agitation and delirium

RM Nelson, TS Lam, KA Considine, JT Jancik

Hennepin Healthcare, Pharmacy, Minneapolis, USA

Critical Care 2023, 27(S1): P107

Introduction: Agitation and delirium are common in the intensive care setting. Rates of agitation and delirium in ICU patients have been reported up to 70% and 89% respectively [1,2]. Previous studies have demonstrated that use of valproic acid (VPA) has been associated with reductions in agitation and delirium in the ICU [3–5]. The purpose of this study was to assess the impact of VPA on resolution of ICU agitation and delirium when prescribed at comparable or higher doses than previously published.

Methods: This single-center retrospective analysis was conducted in adult patients admitted between 01 Jun 2018 and 31 May 2022 who were prescribed VPA for agitation or delirium in the ICU. Agitation was defined as a RASS greater than or equal to 2 for at least 2 occurrences per calendar day. Delirium was defined as a positive CAM-ICU for at least 2 occurrences per calendar day. The primary objective was time to resolution of agitation or delirium from VPA initiation.

Results: The analysis included 177 patient encounters. Agitation or delirium was present in 61.0% and 94.4% of patients respectively. Agitation resolved in 85.2% of patients with median time to resolution of 3 days (IQR 4). Delirium resolved in 75.4% of patients with median time to resolution of 3 days (IQR 4.5). Agitation or delirium recurred in 53.2% and 61.9% of patients respectively during the same ICU stay. The median starting dose of VPA was 22.2 mg/kg/day (IQR 11.4).

Conclusions: VPA at doses greater than 20 mg/kg/day may be useful to aid in resolution of agitation or delirium in the ICU.


  1. 1.

    Aubanel S et al. Anaesth Crit Care Pain Med. 2020;39:639–646.

  2. 2.

    Reade MC et al. N Engl J Med. 2014;370:444–454.

  3. 3.

    Gagnon DJ et al. J Crit Care. 2017;37:119–125.

  4. 4.

    Crowley KE et al. Clin Ther. 2020;42:e65–e73.

  5. 5.

    Quinn NJ et al. Ann Pharmacother. 2021;55:311–317.


Risk factors for cognitive impairment in critically ill surgical patients

KJ Jeong1, YJ Jung1, SB An2, HJ Lee2, SK Hong2

1Asan Medical Center, Nursing, Seoul, South Korea, 2Asan Medical Center, Division of Acute Care Surgery, Department of Surgery, Seoul, South Korea

Critical Care 2023, 27(S1): P108

Introduction: As the ICU survival rate increases, the number of patients experiencing long-term cognitive impairment is increasing. The aim of this study is to evaluate cognitive impairments at the time of discharge from the surgical intensive care unit and identify the incidence and risk factors of cognitive impairments.

Methods: A prospective study was conducted from July 1, 2021, to May 31, 2022. The participants were adult patients who stayed in the surgical intensive care unit for more than 48 h. At the time of discharge from the ICU, cognitive function was evaluated with with the Mini-Mental State Examination (MMSE) and the risk factors (general, disease, treatment-related characteristics) of cognitive impairment were analyzed.

Results: In this study, the prevalence of cognitive impairments was 41.8% (64/153). As a result of analyzing the factors related to the occurrence of cognitive impairments, age, education level, living with family, and history of cerebrovascular diseases, CCI, delirium, use of sedatives, length of stay in ICU, use of insulin were different between groups (Table 1). According to multivariate regression, lower education level, history of cerebrovascular diseases, prolonged delirium, prolonged hyperglycemia, and absence of caregiver were identified as a predictive factors of cognitive impairment at the time if discharge from ICU.

Conclusions: The cognitive impairment was relatively high. The correction of modifiable risk factors for cognitive impairment such as prevention of delirium and hyperglycemia can attenuate cognitive dysfunction. It is necessary to develop and apply an individual and step-by-step cognitive rehabilitation program considering the characteristics of the patient by assessing the level of education and the history of cerebrovascular diseases such as stroke.

Table 1 (abstract P108). Risk factors of cognitive impairment when discharging in SICU



95% CI


Education(< high school)



 < .001*

Living with family




History of stroke




Presence of delirium




Use of insulin




Duration of ICU stay days





Greater serum concentration of IL-6 at hospital admission is associated with increased risk of developing delirium during hospitalization

T Bassi1, E Rohrs2, M Nicholas2, S Reynolds2

1Lungpacer Medical Inc., Burnaby, Canada, 2Royal Columbian Hospital, New Westminster, Canada

Critical Care 2023, 27(S1): P109

Introduction: Delirium is a prevalent syndrome in hospitalized patients. Chronic inflammation has been associated with an increased risk of developing delirium during hospitalization. The focus of our study was to investigate existing literature to determine whether, at hospital admission, serum concentration of a biomarker for systemic inflammation, interleukin (IL)-6, was associated with the likelihood of developing delirium during hospitalization. We conducted a systematic review and meta-analysis to investigate that question.

Methods: Our meta-analysis followed the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) protocol. Independent extraction with multiple reviewers' consensus was used to determine the studies included. The weight and heterogeneity of the manuscripts were calculated using inverse covariance with a random-effects model. The inclusion criteria were articles that investigated links between serum concentration of IL-6 at hospital admission and delirium during hospitalization.

Results: After excluding duplicates, sixteen studies that investigated the role of IL-6 serum concentration at hospital admission were included. Our search found evidence that patients who developed delirium during hospitalization had, at hospital admission, a mean serum concentration of IL-6 significantly greater than patients who did not develop delirium during hospitalization. The mean difference in IL-6 serum concentration was independent of other confounding variables such as the patient’s severity of illness. The mean difference in serum concentration of IL-6 at hospital admission between patients who developed delirium and those who did not was 24.05 pg/ml, p < 0.00001.

Conclusions: Our findings indicate that a greater serum concentration of IL-6 at hospital admission is associated with an increased risk of developing delirium during hospitalization.

figure aq

Fig. 1 (abstract P109) Forest plot showing mean difference in IL-6 serum concentrations at hospital admission between patients who develop delirium and who do not develop delirium during hospitalization


Thiamine supplementation and delirium occurrence in the intensive care unit

MA Mumin1, C McKenzie2, L Aitken3, D Hadfield4, F Hanks5, J Yick6, V Page7, B Blackwood8, D McAuley8, P Spronk9

1Pharmacy, King’s College Hospital NHS Foundation Trust, Pharmacy, London, UK, 2Perioperative and Critical Care Theme, Southampton NIHR Biomedical Research Centre, University Hospital Southampton, Southampton, UK, 3School of Health and Psychological Sciences, City, University of London, London, UK, 4Critical Care, King’s College Hospital NHS Foundation Trust, London, UK, 5Pharmacy, Guy’s and St. Thomas’ NHS Foundation Trust, London, UK, 6Pharmacy, King’s College Hospital NHS Foundation Trust, London, UK, 7Watford General Hospital, Watford, UK, 8Wellcome-Wolfson Institute for Experimental Medicine, Queen’s University Belfast, Belfast, UK, 9Gelre Ziekenhuizen Apeldoorn, Apeldoorn, Netherlands

Critical Care 2023, 27(S1): P110

Introduction: We hypothesized that routine intravenous (IV) thiamine supplementation in intensive care unit (ICU) will reduce delirium occurrence. Thiamine di-phosphate (TDP), the active form of thiamine, is an essential cofactor in glucose metabolism, glutamate transformation and acetylcholinesterase activity; all reported in delirium occurrence [1].

Methods: An anonymized, approved patient dataset was obtained from consecutive ICU admissions over two periods in Gelre Hospitals, Netherlands; October 2014–October 2016 (period 1) and April 2017–April 2019 (period 2). Routine scheduled prescription of IV thiamine, 300 to 600 mg within 72 h of admission, followed by 100 mg daily for 5 days, began in October 2016 after reporting of low whole blood thiamine [2]. Delirium was defined as 1 or more positive confusion assessment method-intensive care unit (CAM-ICU) score in 24 h. PREDELIRIC scores were calculated as a measure of prediction of delirium [3].

Results: There was a total of 3074 ICU admissions recorded. In period one, 111 of 1508 (7%) received IV thiamine, compared to 937 of 1566 (60%) patients during period 2 (Table 1). Data eligible for PREDELIRIC risk assessment was available for only 575 (38%) and 549 (35%) patients in period 1 and 2 respectively. Our analysis reported a reduction in delirium incidence during period 2 (0.16), compared to period 1 (0.20); OR: 0.78 [95% CI 0.65–0.94; p = 0.009]. There was no change in the mean PREDELIRIC score between period 1 and period 2, 0.30 (± 0.11).

Conclusions: IV thiamine administration may reduce delirium occurrence in the ICU; although it did not appear to confer benefit in those at greater risk of delirium. Limitations include low adherence to IV thiamine regime in period 2, and observational and retrospective data. Future work should focus on mechanism of thiamine in delirium and randomised control efficacy trials.


  1. 1.

    Kealy J et al. JNeurosci 2020;40:5681–5696.

  2. 2.

    van Snippenburg W et al. J Iintensive Care Med 2017;32:559–564.

  3. 3.

    van den Boogaard M et al. BMJ 2012;344:e420.

Table 1 (abstract 110). Overall delirium incidence


Number of patients receiving thiamine

PREDELIRIC score (mean ± SD)

Delirium incidence (mean)

OR (95% CI)

Period 1 (Oct’14–16)

111 (7%)

0.30 ± 0.11



Period 2 (Apr’17–19)

937 (60%)

0.30 ± 0.11


0.78 (0.65–0.94)


Differences in delirium presentation in non-intubated COVID-19 and non-COVID-19 patients in the ICU

MA Xanthoudaki, A Valsamaki, D Papadopoulos, P Katsiafylloudis, A Papadogoulas, K Oikonomou, E Papapostolou, D Tsimpida, D Xenou, P Papamichalis

General Hospital of Larissa, Intensive Care Unit, Larissa, Greece

Critical Care 2023, 27(S1): P111

Introduction: Delirium in intensive care unit (ICU) patients is a common disorder which is characterized by non-specific brain dysfuncion and is associated with increased mortality [1]. Coronavirus disease 2019 (COVID-19), caused by the novel coronavirus SARS-COV-2 is associated with increased delirium rates through various mechanisms including hypoxaemia, circulation of inflammatory mediators and direct central nervous system damage [2]. Due to life-threatening respiratory manifestations in COVID-19 patients delirium is often underestimated. Aim of the study was to record the pharmacologic treatment of delirium in ICU patients treated with non-invasive ventilation and the differences in delirium treatment between COVID-19 and non-COVID-19 patients.

Methods: Consecutive patients with respiratory failure were retrospectively included. All patients were hospitalized in the ICU and non-invasive ventilation was applied. Patient characteristics, length of ICU stay and medication were recorded.

Results: Totally 64 patients were included (33 patients with COVID-19). Pharmacologic management of delirium included administration of remifentanyl, dexmedetomidine, propofol and combinations of the above. COVID-19 patients were younger (61 vs 70.4 years, p = 0.02), had longer ICU hospitalization (p < 0.001) and needed longer intravenous medication administration (p = 0.001). Additionally, the proportion of patients with COVID-19 treated with intravenous medication due to delirium was higher than non-COVID-19 patients (81.8% vs 48.4%, p = 0.008). COVID-19 seemed to have worse outcomes than non-COVID-19 patients and this difference tended to be statistically significant (p = 0.09).

Conclusions: Delirium is quite common and requires aggressive treatment among COVID-19 patients, despite being younger than non-COVID-19 patients. Physicians should be vigilant to assess and manage delirium to improve the long-term outcomes of COVID-19 patients.


  1. 1.

    Pisani MA et al.Arch Intern Med 2007;167:1629–1634.

  2. 2.

    Kotfis K et al. Crit Care 2020;24:176.


Pathogenesis of encephalopathy and delirium: neocortical neuronal membrane potential and action potential firing are changed by coronavirus-associated cytokines

SR Rajayer, SM Smith

OHSU, PACCM, Portland, USA

Critical Care 2023, 27(S1): P112

Introduction: Encephalopathy and delirium are common following coronavirus infection [1], and the associated neuroinflammation often results in long-term behavioral and cognitive impairment. Neurovirulent cytokines (NVC) are strongly implicated in the pathogenesis of coronavirus encephalopathy [2]. We hypothesized that characterizing the abnormal signaling in NVC exposed neurons will enable us to identify targets to treat encephalopathy and prevent its downstream effects.

Methods: We incubated primary mouse neocortical cultures in NVC known to be increased in coronavirus encephalopathy (TNF-α, IL-1β, IL-6, IL-12 and IL-15). Using whole-cell patch clamp methods, we tested how neuronal function was impacted by 22–28-h exposure to NVC.

Results: We found that NVC depolarized the resting membrane potential (RMP), reduced the firing threshold of neocortical neurons, and increased baseline spontaneous action potential (AP) firing. NVC altered the sensitivity (or input–output properties) of single neurons to changes in their microenvironment. Specifically, decreasing external Ca2+ and Mg2+ from physiological to low (1.1–0.2 mM) levels increased evoked AP firing in control, but not following exposure to NVC. AP firing threshold and spontaneous firing rates returned to control levels 1 h after NVC wash-out. However, the RMP and attenuated sensitivity of evoked APs to changes in the microenvironment remained persistently abnormal suggesting two distinct mechanisms were at play. Interestingly, hyperpolarizing the RMP reversed this altered response.

Conclusions: Sustained exposure to NVC reversibly depolarizes neocortical neuronal RMP, altering excitability and the ability of neurons to respond to microenvironment changes. By characterizing the pathogenesis of the underlying changes in neuronal function in our model of coronavirus encephalopathy we will identify intervenable drug targets.


  1. 1.

    Uginet M et al. J Med Virol 2021;93:4374–81.

  2. 2.

    Li Y et al. J Virol 2004;78:3398–406.


Long-term outcomes of delirium in critically ill surgical patients: a multicenter prospective cohort study

O Chaiwat1, P To-adithep2, K Chittawatanarat3, S Morakul4

1Faculty of Medicine Siriraj Hospital, Anesthesiology, Bangkok, Thailand, 2Faculty of Medicine Siriraj Hospital, Department of Anesthesiology, Bangkok, Thailand, 3Department of Surgery, Faculty of Medicine Chiang Mai Hospital, Chiang Mai University, Chiang Mai, Thailand, 4Department of Anesthesiology, Faculty of Medicine Ramathibodi Hospital, Bangkok, Thailand

Critical Care 2023, 27(S1): P113

Introduction: Postoperative delirium in critically ill surgical patients was associates with short and long term worsening outcomes. There has been limited data regarding long-term outcomes for delirious patients in Thailand. The aims of this study were to identify the mortality rates and dependency rate (functional outcomes) of delirious patientsat 12 months after surgical intensive care unit (SICU) admission and to determine the independent risk factors of 12-month mortality and dependency rate in a cohort of SICU patients.

Methods: A prospective, multi-center study were conducted in 3 university-based hospitals. Critically-ill surgical patients who were admitted to surgical intensive care unit and followed up at 12 months after ICU admission were enrolled. The 12-month mortality was retrieved from medical records or from the Bureau of Registration administration. We assessed dependency state by basic activity of daily living (ADL) and instrumental ADL by telephone interview.

Results: A total of 630 eligible patients were recruited. 170 patients (27%) had postoperative delirium. The overall 12-month mortality rate in this cohort was 25.2%. Delirium group showed significantly higher mortality rates than non-delirium group at 12-months after ICU admission (44.1% vs 18.3%, p 0.001). Independent risk factors of 12-month mortality were ages, diabetes mellitus, preoperative dementia, high SOFA score and postoperative delirium. The overall dependency rate was 52%. Independent risk factors of dependency rate at 12-months included ages ≥ 75 years, cardiac diseases, duration of mechanical ventilation and postoperative delirium.

Conclusions: Postoperative delirium in SICU demonstrated higher mortality rates at 12-months after SICU admission than non-delirium group. Postoperative delirium was an independent risk factor of death at 12-months after SICU admission and becoming dependence after discharge from hospitals.


Study on the value of neutrophil to lymphocyte ratio and neutrophil to prealbumin ratio in prognosis and early warning of organ injury in patients with sepsis

J Liu, Q He, C Li, Y Pan, Y Wu

Department of Critical Care Medicine, The Affiliated Suzhou Hospital of Nanjing Medical University, Suzhou Municipal Hospital, Suzhou Clinical Medical Center of Critical Care Medicine, Suzhou, China

Critical Care 2023, 27(S1): P114

Introduction: To explore the early warning value of neutrophil to lymphocyte ratio (NLR) and neutrophil to prealbumin ratio (NPRI) in hospital death and organ dysfunction of septic patients.

Methods: The clinical data of 175 patients with sepsis were analyzed retrospectively. Patients were divided into survival group and death group according to whether they died in hospital. According to the scores of each system in SOFA score, they were divided into positive group and negative group of organ dysfunction. Then analyze and compare the indexes of each group and their value in predicting patients' death and organ dysfunction.

Results: The scores of NLR, NPRI, SOFA and the number of organ injuries in the death group were higher than those in the survival group (p < 0.05), while the prealbumin and hospitalization time were lower than those in the survival group (p < 0.05). Univariate binary logstic regression analysis showed that NLR [OR = 1.035, 95% CI (1.015–1.055), p = 0.001], SOFA [OR = 1.198, 95% CI (1.040–1.380), p = 0.012], p = 0.025], the number of organ injuries [OR = 1.619, 95% CI (1.031–2.541), p = 0.036] were the risk factors for the hospital death of sepsis patients, and prealbumin [OR = 0.992, 95% CI (0.985–0.998). Multivariate binary Logstic regression analysis showed that NLR [OR = 1.037, 95% CI (1.016–1.059), p = 0.000] and SOFA [OR = 1.236, 95% CI (1.024–1.491), p = 0.027] were predictive sepsis patients ROC curve analysis showed that NLR [AUC = 0.706, 95% CI (0.524–0.887), p = 0.005], NPRI [AUC = 0.7586,95% CI (0.621–0.895), p = 0.001] p = 0.004] and the number of organ injuries [AUC = 0.655, 95% CI (0.610–0.818), p = 0.043] are all predictive values for the hospital death of sepsis patients.

Conclusions: NLR and SOFA scores are independent risk factors for nosocomial death in patients with sepsis, and prealbumin is an independent protective factor for nosocomial death in patients with sepsis. NLR, SOFA score, NPRI and the number of organ injuries can better predict the hospital death of sepsis.


Peripheral dendritic cells as early biomarkers to predict persistent organ failure in patients with predicted severe acute pancreatitis

S Yang, J Liu

Department of Critical Care Medicine, The Affiliated Suzhou Hospital of Nanjing Medical University, Suzhou Municipal Hospital, Suzhou Clinical Medical Center of Critical Care Medicine, Suzhou, China

Critical Care 2023, 27(S1): P115

Introduction: To investigate whether circulating dendritic cells (DC) and its subtypes were able to serve as early biomarkers in predicting persistent organ failure (POF) in predicted severe acute pancreatitis (pSAP).

Methods: We measured the proportions of dendritic cells (DC) and its subtypes. We conducted a prospective study to study whether dendritic cells (DC) and its subtypes can predict persistent organ failure (POF) in patients with predicted SAP. A total of 59 predicted SAP patients who were admitted within 72 h after symptom onset. In addition, 12 age-matched healthy normal controls served as controls (Con). The fasting peripheral venous blood samples were taken in the group 24 h. Flow cytometry (FCM) assay in peripheral blood dendritic cells (DC) and myeloid dendritic cells (mDC) and plasmacytoid dendritic cells (pDC).

Results: A total of 12 healthy volunteers and 59 AP patients were enrolled in the study. AP patients were divided into TOF and POF groups, and there was no significant difference between the two groups. The changes of DC, mDC and pDC in peripheral blood of patients with AP: compared with the control group, the peripheral blood DC and mDC of the AP group decreased significantly (p = 0.000). DC and mDC in peripheral blood of patients with AP were negatively correlated with the severity of the disease (APACHE, II, Ranson scores and revised Marshall scores) (p < 0.05). DC prediction in patients with early AP prediction of POF AUC 0.808 (0.667–0.949).

Conclusions: The percentages of DC, mDC and pDC in peripheral blood mononuclear cells of patients with AP are significantly lower. The expression of DC, mDC and pDC in peripheral blood monocytes correlated with the severity of the disease. The quantity of DC in peripheral blood mononuclear cells of patients with AP may be a good index to predict the occurrence of POF.


Real-time, leukocyte-related, four times validated, patient data partitioning

A Rivas1, C Libertin2, P Satashia3, S Isha3, A Jenkins3, A Hanson3, D Sanghavi3, P Moreno Franco3

1Center for Global Health, School of Medicine, University of New Mexico, Albuquerque, USA, 2Mayo Clinic, Infectious Disease, Jacksonville, USA, 3Mayo Clinic, Critical Care Medicine, Jacksonville, USA

Critical Care 2023, 27(S1): P116

Introduction: Indicators that assess relationships among leukocytes may inform more and/or earlier than those measured in isolation.

Methods: Blood leukocyte differential counts collected from 101 Mayo Clinic COVID-19 patients were related to later outcomes following two approaches: (i) as unstructured data (e.g., lymphocyte percentages) and (ii) as data structures that assess intercellular interactions. Analyzing the same primary data, it was asked whether information contents differed among methods and/or when two sets of structured indicators are used.

Results: While unstructured data did not distinguish survivors from non-survivors (Fig. 1, rectangle A), one data structure (here identified with letters expressed in italics) exhibited one perpendicular inflection that differentiated two patient groups (B). Two survivor-related observations were also distinguished from the remaining data points (B). A second data structure also revealed a single line of observations and a perpendicular data inflection (C), while more (four) patient groups were identified (D). Four validations were conducted: (i) increasing mortality levels among contiguous data subsets (0, 7.1, 16.2, or 44.4%) suggested construct validity (D); (ii) internal validity was indicated because 22 of the 45 survivors detected by the first data structure were also captured by the second one; (iii) the analysis of patients that differed in address, co-morbidities and other aspects supported external validity; and (iv) quasi non-overlapping data intervals predicted statistical validity (E, F). The structured approach also uncovered new and/or dissimilar information: different leukocyte-related ratios explained the clusters identified in these analyses (E, F).

Conclusions: Structured data may yield more information than methods that do not assess multicellular interactions. Possible applications include daily, longitudinal, and personalized analysis of hospital data.

figure ar

Fig. 1 (abstract P116). Data partitioning validated by qausi non-overlapping immune profiles


Associations between red cell distribution width and outcomes of critically ill patients ≥ 90 years

P Theile, J Müller, R Daniels, S Kluge, K Roedl

University Medical Centre Hamburg-Eppendorf, Department of Intensive Care Medicine, Hamburg, Germany

Critical Care 2023, 27(S1): P117

Introduction: Different clinical conditions are associated with changes in red cell distribution width (RDW). High levels of RDW(> 14.5%) were described as predictive marker for unfavorable outcome/mortality in critically ill patients. There is a lack of data in the very elderly critically ill, why we aimed to investigate the association of RDW with outcome in critically ill patients ≥ 90 years.

Methods: Retrospective analysis of all consecutive critically ill patients ≥ 90 years of the Medical University Centre Hamburg-Eppendorf (Hamburg, Germany) with RDW on admission. Clinical course and laboratory were analysed from all patients. High RDW was defined as (> 14,5%). We clinically assessed factors associated with mortality. Univariable and multivariable cox regression analysis was performed to determine the prognostic impact of RDW on 28-day mortality.

Results: We could identify 863 critically ill patients ≥ 90 years with valid RDW values and complete data. 32% (n = 275) died within 28 days and 68% (n = 579) survived 28d. Median RDW levels on ICU admission were significantly higher in non-survivors compared with survivors (15.6% vs. 14.8%, p < 0.001). The proportion of high RDW (> 14.5%) was significantly higher in non-survivors(73% vs. 57%, p < 0.001). Patients with low RDW presented with a lower Charlson Comorbidity index (p = 0.014). SAPS II on admission was lower in patients in the low RDW (35 vs. 38 points, p < 0.001). 32% (n = 104) in the low and 35% (n = 190) in the high RDW group were mechanically ventilated (p = 0.273). Use of vasopressors (35% vs. 49%, p < 0.001) and RRT (1% vs. 5%,p = 0.007) was significantly higher in the high RDW group. Cox regression analysis demonstrated that RDW was significantly associated with 28d mortality [crude HR 1.768, 95% CI (1.355–2.305); p < 0.001; adjusted HR 1.372, 95% CI (1.045–1.802); p = 0.023].

Conclusions: High RDW was significantly associated with mortality in critically ill patients ≥ 90 years. RDW can be used as simple parameter for risk stratification in very elderly critically ill patients.


Sepsis outcomes: is there a place for leukocyte cell population data?

S Šundalić1, I Košuta1, R Radonić1, I Baršić Lapić2, I Rako2, D Rogić2, A Vujaklija Brajković1

1University Hospital Centre Zagreb, Department of Internal Medicine, Zagreb, Croatia, 2University Hospital Centre Zagreb, Department of Laboratory Diagnostics, Zagreb, Croatia

Critical Care 2023, 27(S1): P118

Introduction: Hematological analyzers rapidly provide information on the morphological and functional characteristics of leukocytes, known as leukocyte cell population data (CPD) which is automatically available without the need to collect additional samples [1]. Previous research mostly described the role of CPD in sepsis diagnosis with little emphasis on prognosis. We evaluated the utility of CPD on clinical outcomes.

Methods: This prospective observational study was conducted in a medical ICU of a tertiary level hospital. Complete blood count, CPD data and biochemistry were collected in antibiotic-naive sepsis cases at admission and repeated on the third day. The Sysmex XN-3000 analyzer was used for leukocyte CPD analysis. The association of procalcitonin (PCT), interleukin-6 (IL-6) and CPD on survival, need for mechanical ventilation (MV) and development of acute kidney injury (AKI) was evaluated. AKI was defined by the Kidney Disease Improving Global Outcomes (KDIGO) Criteria. Patients with leukopenia were excluded (< 4.00 × 109/l white blood cells).

Results: Of the 98 cases, 50 were male, and 44 had hemoculture-positive sepsis. The comorbidities included arterial hypertension (n = 55, 56.1%), diabetes (n = 32, 32.7%) and nonhematologic malignancies (n = 24, 24.5%) (Table 1). Binary logistic regression revealed significant association of IL-6 (p = 0.04, OR 1.01, 95% CI 1.001–1.021) and the following leukocyte CPD data on 3rd day: cytoplasmic granulation of neutrophils (NEUT-GI) (p = 0.02, OR 0.85, 95% CI 0.735–0.974), and reactivity of neutrophils (NEUT-RI) (p = 0.04, OR 1.15, 95% CI 1.008–1.321) with AKI development. Leukocyte CPD and PCT were not found to be associated neither with survival nor need for MV.

Conclusions: Leukocyte CPD parameters such as NEUT-GI and NEUT-RI could be used for early detection of patients prone to developing AKI.


  1. 1.

    Urrechaga E. Ann Transl Med. 2020;8:953–953.

Table 1 (abstract P118). Descriptive data at intensive care unit admission presented as median value and interquartile range


Reference range

Median ± IQR



SOFA score


7 ± 6



L (× 109/l)


15.95 ± 13.8



CRP (mg/l)

 < 5

200.45 ± 152.3



PCT (mcg/l)

 < 0.25

1.98 ± 33.73



IL-6 (pg/ml)

 < 7

158.75 ± 859.38





153.45 ± 6.9





60.45 ± 12.7




Assessment of a rapid cellular host response test in predicting sepsis in a subpopulation of patients who are administered steroids

HE Bunch1, M Sorrells2, R Sheybani3, H Tse3, A Shah3, CB Thomas1, HR O’Neal1

1LSU Health Sciences Center/Our Lady 0f the Lake Regional Medical Center, Baton Rouge, USA, 2Cytovale LLC, Medical Affairs, San Francisco, USA, 3Cytovale LLC, San Francisco, USA

Critical Care 2023, 27(S1): P119

Introduction: Sepsis is defined as host immune dysregulation in response to an infection causing organ dysfunction. In patients taking steroids, identification of sepsis may be difficult due to the numerous effects of steroids on the immune system. The IntelliSep test is an investigational diagnostic that quantifies immune activation by measuring the biophysical properties of leukocytes from a blood sample in < 10 min. The test provides a single score, the IntelliSep Index (ISI), between 0.1 and 10.0, stratified into three interpretation bands of increasing disease severity risk: green, yellow, and red. Previously, we demonstrated the utility of the test on risk-stratifying patients with suspicion of sepsis. The purpose of this study was to analyze the performance of the ISI on a subpopulation of patients taking steroids.

Methods: Adult patients presenting to the ED with signs or suspicion of infection were prospectively enrolled at multiple US sites (Feb. 2016–Oct. 2021). EDTA-anticoagulated blood was assayed and patients were followed by retrospective chart review for outcome information. From this population, we compared ISI risk stratification to sepsis-3 criterion between a population of patients not receiving steroids (Steroid− group) to a smaller population of patients receiving steroids (Steroid+ group).

Results: 1108 patients were included in this study: 985 in the Steroid− group and 123 in the Steroid+ group. These two populations were stratified across interpretation bands with 49.0% and 55.3% Green Band, 25.7% and 23.6% Yellow Band, and 25.3% and 21.1% Red Band for the Steroid− and Steroid+ groups, respectively. The groups showed no differences in risk stratification across ISI interpretation bands and showed similar sepsis detection based on Sepsis-3 criterion (Fig. 1).

Conclusions: The ISI, a quantitative measure of immune dysregulation, may aid clinicians in the rapid stratification for risk of sepsis in patients with signs or suspicion of sepsis, independent of steroid use.

figure as

Fig. 1 (abstract P119). Risk stratification of sepsis-3 across interpretation bands for Steroid− and Steroid+ groups (Steroid− : p < 10E−9 across all bands; Steroid+ : p = 0.22 Green-Yellow, p < 10E−5 Green–Red, p < 0.01; No significant differences were observed between Steroid+ and Steroid− groups at each band)


An artificial neuronal network for the early prediction of sepsis in intensive care units

A Ogica, A Krannich, D Thiele, K Ulrich, A Grieser, C Storm

TCC GmbH, Hamburg, Germany

Critical Care 2023, 27(S1): P120

Introduction: Machine learning algorithms showed promising results in predicting sepsis and septic shock before onset. Current meta-analysis revealed a few models with good accuracy using more than 10 predictors [1]. We propose an artificial neural network (ANN) to predict the risk of sepsis a few hours in advance for the use in intensive care units (ICU).

Methods: Data were collected from an ICU of a tertiary care university centre. Septic patients and controls over 18 years old were included. To build a prognostic model only patients were selected with data being available at least 24 h before the onset of sepsis. Controls were matched accordingly. The vital signs were continuously monitored using high frequency automated measurements. At each timepoint of the measurement the median of the values from the last two hours was calculated to smoothen the measurement curves and account for single missing values. We used several clinically relevant vital signs combinations with less than 10 variables to select predictors for the ideal model. The area under curve (AUC), the accuracy and the precision were calculated for each model to evaluate the prediction performance.

Results: We identified about 5000 patients from the database matching the inclusion criteria. Of these 1.5% had sepsis identified by the qSOFA criteria. The best time point to predict sepsis was ten hours before the onset of sepsis. An ANN achieved a high prediction accuracy of over 80% ROC AUC beside lower performing models like logistic regression.

Conclusions: We identified a model capable to predict the risk of sepsis hours ahead before onset. Compared with other models with the same accuracy, ours contains less than 10 predictors. The use of machine learning algorithms will enhance the diagnostic accuracy thus reducing the time to receive appropriate therapy.


  1. 1

    Fleuren et al. Intensive Care Med 2020;46:383–400.


Effect of C-reactive protein (CRP)-guided corticosteroid use on outcomes in hospitalized patients with community acquired pneumonia (CAP): a target trial emulation

YE Odeyemi1, W Farah 1, A LeMahieu2, E Barreto3, H Yadav1, O Gajic1, P Schulte2

1Mayo Clinic, Pulmonary and Critical Care Medicine, Rochester, USA, 2Mayo Clinic, Clinical Trials and Biostatistics, Rochester, USA, 3Mayo Clinic, Department of Pharmacy, Rochester, USA

Critical Care 2023, 27(S1): P121

Introduction: Existing evidence for adjuvant corticosteroid use in CAP is equivocal. CRP is an inflammatory marker that can be used to individually tailor steroid use in patients with CAP.

Methods: Using a retrospective cohort of hospitalized patients with CAP from 2009 to 2020, we conducted a target trial emulation comparing CRP-guided corticosteroid use to usual care. CRP-guided corticosteroid use arm specified measurement of CRP within 48 h of CAP admission and administration of corticosteroids for those with CRP ≥ 150 mg/dl but not those with CRP < 150 mg/dl. Remaining patients were classified as ‘usual care’. In eligible patients, inverse probability of treatment weighting methodology was used to balance potential confounders between CRP-guided and the usual care cohorts. The primary outcome was worst clinical status using the adapted WHO clinical progression scale. Secondary outcomes included need for non-invasive and invasive ventilation, and mortality. Weighted proportional odds logistic regression and logistic regression compared outcomes, using generalized estimating equations for robust variance estimates. Estimates reflect a per-protocol analysis of the specified target trial.

Results: In total, 184 patient received CRP-guided corticosteroid treatment and 4369 received usual care. Median age was 74, and the sample was 54% male and 94% white race. There was no difference in the primary outcome between the CRP-guided and usual-care groups. Specifically, in the unweighted analysis: OR 0.74 (95% CI 0.52–1.05; p = 0.093) and weighted analysis (OR 0.84 (95% CI 0.57–1.24, p = 0.38). There was no difference between groups in any of the secondary outcomes (Table 1).

Conclusions: In this target trial emulation, CRP-guided corticosteroid use in hospitalized patients with CAP was not associated with any differences in patient outcomes.

Table 1 (abstract P121). Clinical outcomes


CRP-guided arm (n = 184)

Usual care arm (n = 4369)

Unweighted, OR (95% CI), p value

Weighted, OR (95% CI), p value

Worst condition, n (%)


0.74 (0.52, 1.05), 0.093

0.84 (0.57, 1.24), 0.381

Supplemental oxygen

148 (80%)

3278 (75%)


Non-invasive mechanical ventilation

13 (7%)

453 (10%)


Invasive mechanical ventilation

20 (11%)

484 (11%)


Non-invasive mechanical ventilation or higher,

36 (20%)

1091 (25%)

0.83 (0.67, 1.03), 0.095

0.91 (0.71, 1.16), 0.452

Invasive mechanical ventilation or higher, n (%)

23 (13%)

638 (15%)

0.85 (0.63, 1.15), 0.301

1.00 (0.71, 1.41), 0.995

Mortality, n (%)

3 (12%)

154 (4%)

0.53 (0.34, 1.19), 0.126

0.60 (0.24, 1.45), 0.253


Epigenetic assays identify leukocyte subpopulations that differentiate outcomes in a West African sepsis cohort

J Chenoweth1, J Brandsma1, S Krishnan1, J Michalska2, U Hoffmueller2, J Achenbach2, A Rosselló Chornet2, G Oduro3, N Adams4, D Clark1

1Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc., ACESO, Bethesda, USA, 2Precision for Medicine GmbH, Berlin, Germany, 3Komfo Anokye Teaching Hospital, Accident and Emergency Department, Kumasi, Ghana, 4Naval Medical Research Center, Directorate for Department of Defense Infectious Disease Research, Silver Spring, USA

Critical Care 2023, 27(S1): P122

Introduction: Cellular immunophenotyping has emerged as a powerful tool for sepsis endotyping and overcoming the problem of heterogeneity. However, technologies such as single‐cell RNA‐sequencing and cytometry by time‐of‐flight are not readily accessible in low resources settings. More recently developed epigenetic quantitative polymerase chain reaction methods enable immune cell characterization and counting by evaluating cell type-specific unmethylated DNA. This approach is suited for low-resource settings and low- and middle-income countries.

Methods: We conducted comprehensive epigenetic cellular immunophenotyping of innate and adaptive immune cells in sepsis patients enrolled in an observational study in Ghana. Fourteen epigenetic assays were used to analyze whole blood of 103 subjects upon admission to the emergency department with up to 5 serial samples (0, 1, 3 and 28 days, 6 and 12 months).

Results: Comparison between sepsis patients and healthy donors showed significant differences at enrolment, including decreased proportions of natural killer (NK) cells, and helper CD4 + and cytotoxic CD8 + T-cells, whereas neutrophil levels were elevated. Inspection of the data by mortality showed significantly decreased levels of T cells and CTLA4 + cells in the early non-survivor group (died within 28 days), but otherwise their immune cell profiles were comparable to the late non-survivor or survivor groups. Recovery from neutrophilia and lymphopenia initiated within 72 h in the sepsis survivors, but not in the early or late non-survivor groups. Finally, some cell types took up to a year to re-equilibrate back to normal levels, whereas others recover more quickly.

Conclusions: Taken together these results show that epigenetic immune cell profiling is a promising new tool for diagnostic and prognostic profiling of sepsis subjects in low resource settings. Future analysis will use these data and available medical data to identify clinically relevant endotypes.


The role of IRS-1 gene polymorphism G972 to the increase of IGFBP-1 concentration and mortality risk in hyperglycemia severe sepsis patients

T Maskoen1, TH Achmad2, T Bisri1

1Faculty of Medicine Universitas Padjadjaran, Anesthesiology and Intensive Therapy, Kota Bandung, Indonesia, 2Faculty of Medicine Universitas Padjadjaran, Biochemistry, Kota Bandung, Indonesia

Critical Care 2023, 27(S1): P123

Introduction: Hyperglycemia and insulin resistance are common in severe sepsis and associated with poor outcomes. The pathogenesis of both hyperglycemia and insulin resistance have not been clearly understood up till now. The advancement of molecular biology made the expert start investigating the possibility of sepsis pathogenesis that also come from genetic concept likes polymorphism [1].

Methods: This was a cohort molecular epidemiology study to investigate the role of polymorphism G972A gene IRS-1 to increase of IGFBP-1 concentration (as a marker of insulin resistance), and mortality risk in hyperglycemia severe sepsis patients. We divided into two groups, IGFBP-1 group and IRS-1 group. The differences of two varibles were analyzed by X2 and mortality risk was measured with relative risk. Significance of differences was determined by p < 0.05.

Results: In IGFBP-1 group that consist of 40 patients, the role of polymorphism to increase of IGFBP-1 concentration was analyzed, and in IRS-1 group that consist of 70 patients, the role of polymorphism to mortality was analyzed. In IGFBP-1 group, the role of IRS-1 gene polimorhism G972A to increase IGFBP-1 concentration was 27% (RR: 0.73; 95 CI: 0.1–4), and it was also found that mortality risk in polymorhism patients was 1.6 higher than in patients without polymorphism (RR: 1.6; 95% CI 1–4).

Conclusions: In this study, we found that in hyperglycemia severe sepsis patients, IRS-1 gene polymorphism G972A had a slight role to increase IGFBP-1 concentration, and it is also act as mortality risk factor. However further study needs to be conducted to strengthen the exact mechanism of pathogenesis as a whole.


  1. 1

    Lin M, Albertson T. Crit Care Med 2004;32:569–79.


Motor, a circular RNA-encoded micropeptide, regulates LPS tolerance of monocyte in sepsis

W Huang

Zhongda Hospital, School of Medicine, Southeast University, Department of Critical Care Medicine, Nanjing, China

Critical Care 2023, 27(S1): P124

Introduction: LPS tolerance, referred to an impaired capacity of monocyte to produce proinflammatory cytokines upon bacterial agonists, is a key feature of immune suppression in sepsis. However, the underlying mediators and molecular pathways remain poorly understood. Emerging studies have uncovered that translatable circular RNAs (circRNAs), as novel findings in the life sciences, perform unexpected function in many fundamental biological and pathological processes. This study aimed to explore the function and potential mechanism of circRNA-encoded peptide in regulating LPS-tolerance of monocyte during sepsis.

Methods: CircRNA microarray was used to identify the expression pattern of circRNAs in the blood of sepsis and septic shock patient. Coding potential of circRNAs was assessed by bioinformation analysis. Human monocytes (THP-1) and mice cecal ligation and puncture (CLP) model were used to evaluate the function of micropeptide encoded by circRNAs.

Results: We identified a novel circRNA-encoded micropeptide named “Motor” for the first time in sepsis. Motor was found lowly expressed in septic shock patients and lower Motor level correlated with poor prognosis, indicating that Motor plays a protective role in sepsis. Silencing Motor by siRNA dramatically suppressed chemokines and cytokines production globally in human monocyte THP1 cells upon LPS stimulation. Consistently, overexpression of Motor promoted pro-inflammatory cytokines releasing. Mechanically, transcriptome profiling indicated that Motor regulated NF-κB signaling. In an experimental model, exosome-based delivery of Motor enhanced the pro-inflammatory function of monocyte and relieved sepsis-associated organ damage and mortality.

Conclusions: Our results highlight the pro-inflammatory function of Motor in converting LPS tolerance of monocytes and the clinical potential of targeting Motor for monocyte-cell-directed immunotherapy in sepsis.


Health economic impact of the use of calprotectin immunoassay for early detection of infections in intensive care patients

A Havelka1, M Lipcsey2, M Hultström2, A Larsson3

1Gentian Diagnostics, Stockholm, Sweden, 2Uppsala University, Department of Surgical Sciences, Anesthesiology, Uppsala, Sweden, 3Akademiska University Hospital, Department of Medical Sciences, Clinical Chemistry, Uppsala, Sweden

Critical Care 2023, 27(S1): P125

Introduction: Early diagnosis of bacterial infections in critically ill patients is challenging, as the clinical manifestation is non-specific. Neutrophil activation is a major response to bacterial infection and calprotectin is an important marker for neutrophil mediated inflammation. Earlier studies have shown the ability of calprotectin to predict bacterial infections before onset of clinical symptoms. With early diagnosis of bacterial infections delayed treatment will be avoided as well as deterioration due to severe infections/sepsis.

Methods: A decision tree model is employed to estimate the cost-effectiveness of calprotectin analysis for early detection of bacterial infection and thus, the earlier start of antibiotic treatment compared to other diagnostic biomarkers. The comparators included in the analysis are white blood cell count, procalcitonin, C-reactive protein, and no testing. The analysis is based on patients admitted to an ICU in a Swedish hospital. The model allows for different diagnostic outcomes based on correctly and incorrectly diagnosis of bacterial infection and timing of antibiotic treatment: patient survival, length of stay in ICU and in general ward and total costs.

Results: The base-case results show that predictively measuring of calprotectin in an ICU saves total costs, reduces mean duration of in-patient care and mortality, compared with the comparators. Using GCAL® immunoassay reduces total costs by approximately 13,000–18,000 EUR per patient, overall mortality rate by 0.11, and mean length of stay in an ICU and general ward by 1.3–2 days and 6–8 days, respectively.

Conclusions: The base-case scenario identified GCAL® Calprotectin Immunoassay as cost-effective for a patient cohort presenting in a Swedish ICU. Compared to the comparators, GCAL® saves total costs, reduces the mean duration of in-patient care, and reduces in-hospital mortality in those patients. In the sensitivity analysis, when key model inputs are varied, GCAL® Calprotectin Immunoassay remains the dominant option.


Serum catestatin is correlated with clinical parameters in non-critical COVID-19 patients

I Jerković1, V Kovacic2, T Ticinovic Kurir3

1Department for Urgent and Intensive Medicine with Clinical Pharmacology and Toxicology, Internal Medicine Clinic, University Hospital Centre Split, University of Split School of Medicine, Split, Croatia, 2University Hospital Centre Split, University of Split School of Medicine, Department for Urgent and Intensive Medicine with Clinical Pharmacology and Toxicology, Internal Medicine Clinic, Split, Croatia, 3Department of Pathophysiology, University of Split School of Medicine, Split, Croatia

Critical Care 2023, 27(S1): P126

Introduction: Catestatin (CST) is a peptid with imunomodulatory, antiinflammatory, and antimicrobial activities. Acute coronavirus disease 2019 (COVID-19) caused by the SARS-CoV-2 virus can cause a systemic disease range unpredictably from mild flu-like disease to multiple organ failure. Despite many studies and scientific interest for COVID 19, there is lack of information regarding correlation between serum CST levels and clinical course od COVID 19. There are only few studies investigated CST plasma levels at COVID 19 patients, but mostly at ICU-patients, and those studies revealed that COVID 19 patients release significant amounts of CST in the plasma and CST predicts a poor COVID-19 outcome. In our work the aim was to demonstrate plasma CST levels and correlation with clinical outcome in a group of severe COVID 19 patients admitted in non-ICU department.

Methods: The subjects were patients admitted during second surge of COVID 19 in April and May 2020 in non-ICU unit for COVID 19 patients (high dependency unit) in Infectology department of University Hospital Split, Croatia. The reason of admission was pulmonary infiltrates and COVID 19 positivity confirmed with nucleic acid test. In study were included 32 subjects (25 females, 7 males) (Table 1). An enzyme-linked immunosorbent assay was used for serum CST levels assessment.

Results: We found significant positive correlation between serum CST levels and: C-reactive protein (r = 0.423, p = 0.008), D-dimers (r = 0.395, p = 0.013), hsTNT (high sensitivity troponin T) (r = 0.603, p < 0.001), proBNP (N-terminal-pro brain natriuretic peptide) (r = 0.569, p < 0.001), and hospitalisation days (r = 0.388, p = 0.014). There was significant difference between groups of participants with SOFA < 3 (n = 18) and SOFA > 3 (n = 14) in catestatin serum levels (7.25 ± 3.66 vs. 11.05 ± 9.52 ng/ml; p = 0.065).

Conclusions: This study confirmed that serum CST levels could have important role as clinical prognostic parameter among non-ICU COVID 19 patients.

Table 1 (abstract P126). Clinical features of participants


Mean ± Std. Deviation



catestatin (ng/ml)

8.91 ± 7.00



age (years)

80.50 ± 9.86



SOFA score

3.16 ± 2.41



Sat O2 (%)

92.06 ± 6.89



creatinine (μmol/l)

105.12 ± 140.58



CRP (mg/l)

32.60 ± 40.76



hospitalisation (days)

30.59 ± 10.97




Which biomarker(s) to predict worsening of SARS-CoV2 infection? An ancillary study from the COVIDeF cohort

M Cancella de Abreu1, J Ropers2, N Oueidat3, L Piéroni4, C Frere5, M Fontenay6, L Velly1, K Torelino2, P Hausfater1

1Sorbonne Université APHP Hôpital Pitié-Salpêtrière, Emergency Department, Paris Cedex 13, France, 2Sorbonne Université APHP Hôpital Pitié-Salpêtrière, Unité de Recherche Clinique, Paris Cedex 13, France, 3Sorbonne Université APHP Hôpital Pitié-Salpêtrière, Biochemistry Laboratory, Paris Cedex 13, France, 4Sorbonne Université APHP Hôpital Tenon, Biochemistry Laboratory, Paris, France, 5Sorbonne Université APHP Hôpital Pitié-Salpêtrière, Hemtology Department, Paris Cedex 13, France, 6Institut Cochin, Hematology Laboratory, Paris, France

Critical Care 2023, 27(S1): P127

Introduction: COVID-19 has been responsible for millions of deaths and intensive care unit (ICU) admissions all over the world. Identifying the patients at risk of developing a severe form is crucial for an optimized orientation and allocation of resources. The main objective of our study was to identify among a selection of biomarkers, those predictive of short term worsening in COVID-19.

Methods: This is an ancillary study using clinical data and collected biobanking from the multicentric cohort study COVIDeF, which included prospectively from March 31th 2020 to March 30th 2021, patients admitted with a suspected Sars-CoV2 infection in the Assistance-Publique-Hôpitaux de Paris network, France. Patients with confirmed COVID-19 were divided in 2 groups: a severe (ICU admission or invasive or non-invasive ventilation or ARDS or death) and a control group (no worsening). The routine blood tests and following biomarkers: troponin, C Reactive Protein (CRP), procalcitonin, Mild-Regional pro-Adrenomedulin (MR-proADM), pro-endothelin, SuPAR, NT-proBNP, calprotectin, PF4, D-dimers, were measured in plasma or serum and compared between both groups using a conditional logistic regression.

Results: Among the 1040 first patients included in the COVIDEF cohort, we selected 512 patients having a blood sample drawn at admission before worsening, of which 60 secondarily worsened (severe group). The mean age was 59.5 (± 19.5) years and 50.2% were females. Among the biomarkers tested, three were independently associated with worsening: CRP (mg/l) OR 1.01 [IC 1.01–1.02], procalcitonin (ng/ml) OR 0.4428 [0.21–0.95] and MR-proADM (pg/ml) OR 3.012 [1.06–8.53].

Conclusions: Among a selection of biomarkers of interest, MR-proADM appears to best identify at admission COVID-19 patients at risk of worsening. Future interventional studies should test the efficacy and security of this biomarker to rule-in and rule-out severe outcome and the usefulness for allocating resources.

Acknowledgement: On behalf of the COVIDeF Study Group.


Correlation between serum levels of IL-6 and positive fluid balance in critical COVID-19

OI Aguilera Olvera1, GA Aguirre Gómez1, GL Arcos López1, CD Del Ángel Argueta2, YN López Esquivel1, JC Muñoz Cháves1, GL Velázquez Estrada1, JA Villalobos Silva1, A Zarate Gracia1

1High Specialty Regional Hospital “Bicentenario 2010”, Critical Care, Victoria, Mexico, 2High Specialty Regional Hospital “Bicentenario 2010” Anesthesiology, Victoria, Mexico

Critical Care 2023, 27(S1): P128

Introduction: IL-6 has been correlated as a prognostic biomarker for worsening sepsis and COVID-19 as well as positive fluid balance for duration of mechanical ventilation [1, 2].

Methods: We performed a retrospective cohort study to analyze the correlation between high levels of serum IL-6 and positive fluid balances in the first 24 h of ICU arrival with mechanical ventilation days. We included adult patient records of critical COVID-19 during 2020 from the High Specialty Regional Hospital “Bicentenario 2010”, all patients were intubated, received treatment according to guidelines inforced in that time. We obtained mean and standard deviation for continuous variables and frequencies for categorical variables, calculated Kolmogorov–Smirnov for non-parametric test and Spearman correlation, OR for severe hypoxemia, RRT.

Results: We analyzed 102 patient records, 72% were male, mean age 54.8 years (SD 19.4), tracheostomy was performed in 8.8% of cases, mean APACHE II 16.7 (SD 8.4), values of inflammatory markers were C-reactive protein 108 mg/dl (SD 95), IL-6 118 pg/ml (SD 240), mean paO2/FiO2 was 150 mmHg (SD 82), 93% were on vasopressors, fluid balance mean was 1542 ml (SD 839), severe hypoxemia was present on 62.7% (P/F below 150 mmHg), prono was used in 47.1%, with an overall mortality occurred in 69%. We found no correlation between serum IL-6 levels and positive fluid balance with mechanical ventilation days and outcomes (rs -0.11 p = 0.23, Fig. 1). Elevated serum IL-6 + positive fluid balance at 24 h ICU arrival was associated with severe hypoxemia (OR 2.82, CI 95% 1.14–6.97, x2 p = 0.022), OR for discharge was non-significant (0.48 CI 0.19–1.20 p = 0.11), RRT (1.09 CI 95% 0.27–4.37, p = 0.9).

Conclusions: In our study no correlation was found between serum IL-6 levels, positive fluid balance and mechanical ventilation days, but there was a significant association with severe hypoxemia.


  1. 1.

    Gorham J et al. PloS ONE 2020;15:e0244628.

  2. 2.

    Wang W et al. Crit Care Med 2022;50:307–316.

figure at

Fig. 1 (abstract P128). Correlation between serum IL-6 levels + positive fluid balance with severe hypoxemia


Usage of highly elevated procalcitonin levels for identification of urosepsis in intensive care unit patients

ID Basset1, V Kehl2, G Lorenz3, RM Schmid1, T Lahmer1

1Klinikum rechts der Isar der Technischen Universität München, Klinik und Poliklinik für Innere Medizin II, München, Germany, 2Klinikum rechts der Isar der Technischen Universität München, Institut für KI und Informatik in der Medizin, München, Germany, 3Klinikum rechts der Isar der Technischen Universität München, Abteilung für Nephrologie, München, Germany

Critical Care 2023, 27(S1): P129

Introduction: Identification of a specific infection site is a cornerstone in sepsis treatment. Urosepsis is beside pneumonia one of the most common reasons for ICU administration in elderly patients. Procalcitonin (PCT) is a promising biomarker to help diagnose sepsis and guide treatment decisions. We aimed to assess the utility of highly elevated PCT concentrations in serum to predict urosepsis in critically ill patients.

Methods: We compared in a monocentric, retrospective study septic patients with PCT values > 50 ng/ml to patients with PCT < 15 ng/ml in terms of primary focus of infection. Kruskal–Wallis test was performed to show differences in PCT-values between different foci of infection in both groups. Chi-square automatic interaction detection (CHAID) was used to find cut-off-values to distinguish between foci of infection.

Results: 88 out of 255 included patients had a maximum PCT > 50 ng/ml within the first 48 h of ICU admission. Median age was 66 (60–76) years, 63.6% were male and median SOFA-Score was 12 (10–15). Patients with PCT > 50 ng/ml showed significantly more positive urine culture results compared to patients with PCT < 15 ng/ml (p < 0.001). In patients with PCT > 50 ng/ml, there was a significant difference in PCT values between patients with respiratory and genitourinary focus of infection (p = 0.019). Using CHAID, a genitourinary focus was only present in 12,7% of cases when PCT was < 180.5 ng/ml but in 47.1% when it was > 180.5 ng/ml (p = 0.011).

Conclusions: When reaching high values, PCT can help to differentiate genitourinary focus of infection in critically ill patients at admission to ICU, using a cut-off-value of 180.5 ng/ml. This may help timely administration of specific urological treatment and result in further decline of mortality rates in urosepsis.


Procalcitonin combined to point of care molecular respiratory panel testing for antibiotic stewardship in the emergency room: the PROARRAY study

L Velly1, M Cancella de Abreu2, D Boutolleau3, I Cherubini2, E Houas2, P Hausfater1

1Sorbonne Université APHP Hôpital Pitié-Salpêtrière, Emergency Department, Paris Cedex 13, France, 2Sorbonne Université APHP Hôpital Pitié-Salpêtrière, Paris Cedex 13, France, 3Sorbonne Université APHP Hôpital Pitié-Salpêtrière, Virology Department, Paris Cedex 13, France

Critical Care 2023, 27(S1): P130

Introduction: Molecular syndromic respiratory panel (RP) or procalcitonin (PCT)-driven algorithms have reported conflicting efficacy for antibiotic ATB) stewardship in LRTI. We hypothesized that combining real-time PCT measurement and virus identification would reduce ATB exposition in LRTI suspicions presenting to the emergency department (ED).

Methods: PROARRAY study is a prospective, randomized interventional trial, conducted in the adult ED of an academic 1600-bed hospital. Patients attending the ED with a suspicion of LRTI were randomized into the intervention arm (systematic PCT measurement and point of care BIOFIRE® RP2plus (then 2.1) testing, with the recommendation to withhold or withdraw ATB if PCT < 0.25 µg/L and/or identification of a virus) or a standard of care (SOC) arm. The primary endpoint was the duration of antibiotic exposure in the first 28 days.

Results: 451 patients were randomized (intervention: 225, SOC: 226), mean age 62.5 ± 19.4 years, hospitalization rate 59.9%, mean length of stay 7.4 ± 8.4 days. Main diagnoses were CAP (n = 129), COVID-19 (n = 91), AECOPD (n = 31). The BIOFIRE® RP2.1plus identified at least one viral species in 112 patients (49.8%). The duration of ATB exposition in ITT population was 6.00 [0.00;9.00] and 5.00 [0.00–9.00] days in the SOC and interventional arm respectively (p = 0.71). ATB was started in 31.3% and 34.1% respectively (p = 0.54). ATB exposure was below 6 days in 100 (47.2%) and 108 patients (50.59%) respectively (p = 0.58).

Conclusions: Displaying real-time PCT and RP results failed to significantly reduce the ATB exposition in LRTI suspicions. However, the ATB duration and rate of initiation were already low in SOC arm, which comprised PCT measurement in routine in our ED. Routine PCT measurement probably participated to the lower median ATB duration (6.0 days) than hypothesized (9.0 days) and argues for the main contribution of PCT in ATB stewardship. Moreover, as the intervention was done at ED’s level, we did not control for ATB stewardship in wards for inpatients.


Temporal analysis of biomarker, severity-of-illness score, and machine learning model performance for sepsis prediction

M Patton1, VX Liu2, M Might1

1University of Alabama at Birmingham, Hugh Kaul Precision Medicine Institute, Birmingham, USA, 2Kaiser Permanente Division of Research, Oakland, USA

Critical Care 2023, 27(S1): P131

Introduction: Adoption of electronic medical records has accelerated development of machine learning (ML) surveillance systems for predicting sepsis. To assess the progress of these new methods, we conducted a retrospective analysis of published models for sepsis detection.

Methods: We performed a systematic query of MEDLINE for predictive outcome research from 1995 to 2022. Studies were manually curated and selected based on a search criterion for a sepsis, bacteremia, or septic-shock outcome model with a performance evaluation using an area under the receiver operating characteristic curve (AUROC) or precision recall curve (AUPRC). We categorized models as biomarker, severity score, and/or ML algorithm.

Results: We identified 3393 sepsis prediction publications and selected 74 for further evaluation. The top performing biomarker and severity score-based models for sepsis detection were procalcitonin (AUROC range = 0.67–0.91) and the SIRS score (AUROC range = 0.65–0.81) for studies including > 1000 patient records. ML model performance for sepsis prediction as binary outcome exhibited a wide range in performance (AUROC range = 0.63–0.93). Newer approaches to assessing sepsis as a time-to-onset prediction had AUROCs ranging from 0.74 to 0.91, with neural networks exhibiting a higher performance range (AUROC = 0.64–0.91) compared to traditional regression methods (AUROC = 0.74–0.85) (Fig. 1). Studies incorporating novel data elements like clinical documentation often exhibit improved model performance. Inter-study comparison revealed a tremendously wide range in sepsis outcome rates from < 1% to as high as 50% across datasets and different definitions of sepsis, thereby underscoring the challenge of comparing performance across time and methods.

Conclusions: A review of sepsis prediction models published over 3 decades revealed a wide variability in model performance and identified numerous confounding differences between studies limiting robust comparison.

figure au

Fig. 1 (abstract P131). Temporal analysis of biomarker, severity-of-illness score, and machine learning model performance for sepsis prediction from 1995 to 2022


Predictive performance of pancreatic stone protein for ICU mortality in hospitalized adults with infection: a systematic review and individual patient level meta-analysis

J Prazak1, A Moser2, P Zuercher1, LG De Guadiana-Romualdo3, MJ Llewelyn4, R Graf5, T Reding5, P Eggimann6, YA Que1

1Inselspital, Department of Intensive Care Medicine, Bern, Switzerland, 2University of Bern, CTU, Bern, Switzerland, 3Santa Lucia University Hospital, Biochemistry Department, Cartagena, Spain, 4Brighton and Sussex Medical School, Falmer, UK, 5Universitätsspital Zürich, Department of Visceral and Transplantation Surgery, Zürich, Switzerland, 6Lausanne University Hospital, Department of Locomotor Apparatus, Lausanne, Switzerland

Critical Care 2023, 27(S1): P132

Introduction: Pancreatic stone protein (PSP) has been shown in several studies to be a promising biomarker for prediction of mortality among patients with infection. The aim of this systematic review and individual patient level meta-analysis was to evaluate PSP predictive performance on ICU mortality and compare it to those of procalcitonin (PCT) and C-reactive protein (CRP).

Methods: In a systematic search on PubMed and Cochrane databases we identified 46 studies. Among them, five studies measuring the PSP by the ELISA technique explored the performance of PSP in stratifying unselected hospitalized adult patients on admission to the ICU/ED according to prognosis of mortality. We used hierarchical regression models to predict ICU mortality and evaluated predictive performance of PSP, PCT and CRP reporting classification plots besides AUC [1].

Results: Among 678 patients included into the analysis, the estimated overall ICU mortality was 17.8% with a 95% prediction interval (PI) ranging from 4.1 to 54.6% with a substantial heterogeneity between studies (I-squared 87%). For the three studied biomarkers, sensitivity and 1-specifity with continuous risk thresholds are shown as classification plots (Fig. 1). For PSP, the sensitivity was 0.96, 0.52, 0.30 for risk thresholds 0.1, 0.2 and 0.3 (values for false positive rate were 0.84, 0.25, 0.10). The AUC was 0.69 for PSP, 0.61 for PCT and 0.52 for CRP.

Conclusions: Classification plots are a key tool to support clinical decision making for treating patients with infections, because they allow to incorporate continuous risk threshold information. Compared to CRP and PCT, PSP revealed better predictive ability for ICU mortality in the studied heterogeneous population.


  1. 1

    Verbakel JY et al. J Clin Epidemiol 2020;126:207–16.

figure av

Fig. 1 (abstract P132). Classification plots—ICU mortality, sensitivity (Panel A) and 1-specifity (Panel B) with continuous risk thresholds. Depicted risk thresholds are derived from Youden’s index


Serum-calprotectin for identification of severe COVID-19 in hospitalized patients

Å Parke1, M Hansson2, B Strunz3, K Strålin4

1MedH, Department of Medicin, Karolinska Hopital Huddinge, Huddinge, Sweden, 2MedH, Department of Clinical Chemistry, Karolinska University Hospital, Stockholm, Sweden, 3MedH, Center for Infectious Medicine, Department of Medicine Huddinge, Karolinska Institutet, Stockholm, Sweden, 4MedH, Department of Medicine Huddinge, Karolinska Institutet, Stockholm, Sweden

Critical Care 2023, 27(S1): P133

Introduction: The disease severity of COVID-19 varies from mild upper respiratory tract infection to life-threatening lower respiratory tract infection. Biomarkers can support severity assessment. We aimed to compare scalprotectin with routine biomarkers to study if s-calprotectin could identify patients with risk for severe COVID-19. We also aimed to study if ongoing corticosteroid therapy would alter levels of s-calprotectin.

Methods: We collected serum samples within 8 days from admission from 162 adult patients hospitalized for COVID-19 from April to October 2020. The serum tubes were centrifuged within 2 h and serum was aspirated and frozen in aliquots at −80 °C. scalprotectin was measured according to the routine method at Karolinska University Laboratory. Study endpoint was severe COVID-19 any time during the hospital stay. Severe COVID-19 was defined as treatment with either low-flow oxygen ≥ 10 l/min, high-flow nasal oxygen, non-invasive ventilation, or invasive mechanical ventilation. For patients hospitalized before the results of the RECOVERY Dexamethasone study were presented, a propensity score matched sub-study was performed, comparing patients treated and not treated with corticosteroids prior to study sample.

Results: Median s-calprotectin was, 2.7 ml/L (IQR, 1.6–4.2) for non-severe COVID-19 cases (n = 59) and 5.0 ml/l (IQR, 3.5–9.1) for severe COVID-19 (n = 103) (p < 0.001). ROC-curve analysis compering s-calprotectin, C-reactive protein (CRP), neutrophils, 1/lymphocytes, and D-dimer, showed the highest AUC for s-calprotectin, 0.775 (Fig. 1). In the sub-study of with corticosteroid treated (n = 35) and non-treated patients (n = 26), the median s-calprotectin was 4.0 (IQR, 2.5–5.6) and 3.8 (IQR, 2.7–3.8), respectively (p = 1.0).

Conclusions: The study showed a good performance of s-calprotectin for identification of severe COVID-19, in agreement with previous studies. It also indicated that s-calprotectin results are not affected by ongoing corticosteroid treatment.

figure aw

Fig. 1 (abstract P133). Performance of biomarkers for identification of severe COVID-19


Performance of a molecular assay in sepsis patients stratified by qSOFA

S Cermelli1, K Navalkar1, RF Davis1, R Balk2

1Immunexpress, Seattle, USA, 2Rush Medical College and Rush University Medical Center, Chicago, USA

Critical Care 2023, 27(S1): P134

Introduction: We categorized ICU patients suspected of sepsis using quick Sequential Organ Failure Assessment (qSOFA), expert panel diagnosis, and a molecular assay (SeptiCyte RAPID) to demonstrate clinical utility of measuring host response to identify patients with an infection & systemic inflammation. Sepsis-3 definition can be operationalized by clinicians by assessing a qSOFA ≥ 2 [1]. However, clinicians first need to identify patients with an infection, which can be challenging because of heterogenous clinical signs [2].

Methods: Data was analyzed from clinical trials that enrolled ICU adult patients suspected of sepsis (n = 419) [3]. A full qSOFA calculated for 173 of these patients enabled stratification into < 2 or ≥ 2 groups. Expert panel diagnosis was performed retrospectively and independently classifying cases into systemic inflammation in response to infection (Inf+), or not (Inf-). SeptiCyte RAPID score (0–15) measured mRNA expression of two genes.

Results: For 173 ICU patients suspected of sepsis, 71 (41%) were retrospectively diagnosed as Inf + and 102 (59%) were diagnosed as Inf- (Fig. 1). Inf+ patients consisted of 55 (77%) with qSOFA ≥ 2, and 16 (23%) with qSOFA < 2. Inf + patients could be differentiated from Inf- using SeptiCyte RAPID irrespective of qSOFA ≥ 2 (AUC 0.82) or qSOFA < 2 (AUC 0.81).

Conclusions: SeptiCyte RAPID was able to differentiate patients suspected of sepsis into those with a systemic inflammatory response to infection (Inf+), versus those with a systemic inflammatory response due to non-infectious causes (Inf-). SeptiCyte RAPID achieved this independent of the qSOFA score < 2 or ≥ 2. The key to rapidly identifying patients with sepsis is to first determine infection status & then use qSOFA to determine likelihood of poor outcome.


  1. 1.

    Singer M et al. JAMA. 20,156;315: 801–810.

  2. 2.

    Liu VX et al. Crit Care Explor. 2021;3: e0344.

  3. 3.

    Balk R et al. medRxiv 2022.

figure ax

Fig. 1 (abstract P134). Stratification of ICU patients (n = 173) suspected of sepsis based on qSOFA, expert panel diagnosis and SeptiCyte RAPID


Biomarkers for organ dysfunction and outcome prediction in septic patients admitted in emergency department with community-acquired pneumonia

A Degrassi1, G Coppalini2, N Fiotti3, G Biolo3

1ASUGI Institute, Department of Emergency Medicine, Trieste, Italy, 2Humanitas Research Hospital, Anaesthesiology and Intensive Care, Rozzano, Italy, 3ASUGI Institute, Department of Medical, Surgical and Health Sciences, Trieste, Italy

Critical Care 2023, 27(S1): P135

Introduction: Community-acquired pneumonia (CAP) is a lung infection acquired outside hospitals. Specific biomarkers could help identification of patients who may rapidly develop organ dysfunction and are at risk of early mortality, guiding physicians to decide for best patients allocation (ICU/medical ward) and relocate resources. Aim of our study is analyze the role of biomarkers in the prediction of specific organ dysfunction and mortality in septic CAP patients.

Methods: We conducted a secondary analysis of a prospective study held in the emergency department (ED) of five university hospitals in Italy [1]. Levels of CD64, PLA2GIIA, CD14 and CD25 were evaluated from samples collected within 12 h from ED arrival. Patients were followed until discharge for development of neurological, hepatic, renal, respiratory, cardiovascular and coagulation dysfunctions and in-hospital mortality was assessed.

Results: Of 368 patients included in the analysis, 87 (24%) died during their hospital stay. All biomarkers showed correlation with mortality, but only CD14 and CD25 had significant increase in non-survivor at multivariate analysis corrected for age and comorbidities (452.5 [292 -754] vs 769 [453–1402] p < 0.01 and 13215 [9003–20041] vs 20329 [12841–36729] p < 0.01). Higher levels of CD14 were associated with neurologic dysfunction, hepatic and cardiovascular injury. All biomarkers were increased when coagulation disorder and renal injury developed. CD14 was the best predictor of mortality (Fig. 1) and of both these complications at ROC curve analysis (AUC 0.684, OR 4.07 [CI 2.14–7.73] and AUC 0.757, OR 5.59 [CI 3.42–9.14]).

Conclusions: Higher levels of CD25 and CD14 were the best predictors of mortality in septic CAP patients. Regarding organ dysfunctions, CD14 was the best predictor of neurologic, cardiovascular, renal and coagulation disorders. Further studies will be needed to better characterize role of biomarkers for outcome stratification in CAP patients.


  1. 1.

    Mearelli F et al. Crit Care Med, 20,018;46:1421–1429.

figure ay

Fig. 1 (abstract P135). ROC curve of considered biomarkers to predict mortality in CAP patients. Best prognostic role was found to be CD14 with 0.692 AUC


Importance of damage-associated molecular patterns and pathogen-associated molecular patterns for severe outcome in bacteraemic infection

H Alpkvist1, I Ziegler2, P Mölling2, E Tina2, A Norrby-Teglund3, S Cajander2, K Strålin4

1Karolinska Institutet, Department of Medicine Huddinge, Division of Infectious Diseases, Stockholm, Sweden, 2Department of Infectious Diseases, Örebro University Hospital, Örebro, Sweden, 3Karolinska Institutet, Center for Infectious Medicine, Department of Medicine Huddinge, Karolinska Institutet, Stockholm, Sweden, 4Karolinska Institutet, Department of Medicine Huddinge, Division of Infectious Diseases, Karolinska Institutet, Stockholm, Sweden

Critical Care 2023, 27(S1): P136

Introduction: We aimed to study the importance of damage-associated molecular patterns (DAMPs), i.e. nuclear DNA (nDNA), and Pathogen-associated molecular patterns (PAMPs), i.e. bacterial 16S DNA, and their correlation to severe outcome in patients with bacteraemic infections.

Methods: Adult patients hospitalized with culture-proven bacteraemic infection caused by Streptococcus pneumoniae (n = 28), Staphylococcus aureus (n = 24), or Escherichia coli (n = 11) were included in a prospective study. Plasma samples were collected on admission (day 0) and on days 1–2, 3–4, 7 ± 1, 14 ± 2, and 28 ± 4. Plasma was analyzed for nDNA and 16S DNA using specific quantitative PCR methods. Sepsis was defined according to Sepsis-3. Study outcome was severe outcome, defined as admission to an intensive care unit (ICU) and/or death within 60 days from admission.

Results: Of 63 included patients with bacteraemic infection, 37 (59%) had sepsis at admission. Severe outcome was noted in 16 patients (25%), including 11 patients admitted to an ICU and 7 non-survivors. In the univariate analysis, severe outcome was was associated with DAMP levels (nDNA) and PAMP levels (16S DNA) on day 1–2, but not with gender, age, comorbidity or aetiology (Table 1). In a multivariate analysis, both DAMP and PAMP levels on day 1–2 were independently associated with severe outcome (Table 1). 16S DNA was most often cleared from the circulation within a few days, only 3 patients had measurable 16S DNA on day 14. nDNA, on the contrary, remained at high levels over time in non-survivors, and nDNA on day 28 was significantly higher in patients who died within 60 days.

Conclusions: High levels of DAMP and high levels of PAMP were independently associated with severe outcome in patients with bacteraemic infection.

Table 1 (abstract P136). Univariate and multivariate analysis for severe outcome


Univariate analysis odds ratio (95% CI)

Univariate analysis p value

Multivariate analysis odds ratio (95% CI)

Multivariate analysis p value


1.90 (0.60–5.96)




1.03 (0.98–1.08)



Charlson comorbidity index

1.17 (0.92–1.50)



Staphylococcus aureus versus Escherichia coli

1.10 (0.22–5.40)



Streptococcus pneumoniae versus Escherichia coli

0.73 (0.15–3.62)



Plasma nDNA, log10 copies/μL (DAMP) on day 1–2

8.99 (1.63–49.57)


7.88 (1.21–51.49)


Plasma 16S DNA, log10 copies/ml (PAMP) on day 1–2

1.47 (1.12–1.92)


1.41 (1.08–1.85)



The biomarkers sRAGE and df in combination are significant predictors of mortality in COVID-19 patients

S Pillai1, JC Zaldua1, L Butcher2, JA Carnicero2, K Hawkins1, J Whitley1, R Mothukuri1, K Morris2, JD Erusalimsky2, PA Evans1

1Welsh Centre for Emergency Medicine Research, Emergency Department, Morriston Hospital, Swansea, UK, 2The Cellular Senescence and Pathophysiology Group, Cardiff Metropolitan University, Cardiff, UK

Critical Care 2023, 27(S1): P137

Introduction: COVID-19 infection is associated with marked inflammatory response and the patients who are admitted to the hospital are at increased risk of developing venous thromboembolism. sRAGE (soluble receptor for advanced glycation end-products) are acutely elevated in host inflammatory response to infections [1]. Fractal dimension (df), the biomarker of clot microstructure that measures thrombogenicity has shown to be elevated in acute inflammatory conditions such as sepsis and severe sepsis. The aim of the study was to analyse these biomarkers in COVID-19 infection and whether these biomarkers help to predict mortality.

Methods: 120 suspected COVID-19 patients were recruited from the Emergency Department of a tertiary teaching hospital. One patient was excluded because they were anticoagulated, blood samples were taken to perform fractal dimension (df) and sRAGE.

Results: When compared to PCR -ve group, 95 patients in the PCR + ve group had significantly elevated sRAGE (p < 0.001), but not df (p = 0.43). When compared to those who survived, sRAGE was significantly elevated (p = 0.01) in 14 patients who died in PCR + ve group, but not df (p = 0.08). No significant correlation existed between sRAGE levels and df in those patients who survived (p = 0.72) or died (p = 0.92). Logistic regression analysis showed that sRAGE and df in combination acted as highly significant predictors of mortality in COVID-19 (p = 0.009) in PCR + ve group.

Conclusions: COVID-19 patients had a profound inflammatory response as evidenced by significantly elevated sRAGE levels. This inflammatory process was more profound in those who died. The thrombogenicity in COVID-19 patients and those who died with COVID-19 appears to be not significant as measured by df. sRAGE in combination with df can be utilised as significant predictors of mortality in COVID-19 patients.


  1. 1.

    Butcher L et al. Respir Res 2022;23:303


VAP BioFire utilisation in diagnosis of ventilator associated pneumonia—are we misfiring in our application of this tool?

D Clarke, R Bart Koranteng, S McKechnie

Oxford University Hospitals Foundation Trust, Oxford Critical Care, Oxford, UK

Critical Care 2023, 27(S1): P138

Introduction: Ventilator associated pneumonia (VAP) is a nosocomial infection with high associated morbidity and mortality. Timely, accurate diagnosis of VAP and appropriate antibiotic therapy can impact outcomes. The disease itself is poorly defined and diagnosis often limited by low yield sampling techniques and delays in growth of microbiological samples. The use of rapid diagnostics such as BioFire VAP FilmArray have shown promise in reducing time to identification of pathogen, with hopes this may lead to more targeted antibiotic therapy. However, the benefit of this is unclear and relies on ICU clinicians having trust in the sampling technique and test itself. There Is a signficant cost burden to the test which on our unit was part of a routine order-set in patients suspected of VAP. We aimed to reduce uncessary sample through a QI process.

Methods: We performed a retrospective analysis of clinical data from electronic health records of all patients admitted to our ICU between December 2021 to January 2022, and then a repeat analysis post intervention between December 2022. All patients ventilated greater than 48 h were included for analysis. Data on length of stay, mortality, duration of antibiotics, number of BioFire samples sent, and change to therapy post test were recorded.

Results: BioFire lead to a change in therapy in 21% of patients. This was an escalation of antibiotics 66% of the time. The test never led to a cessation of antibiotic therapy. Following intervention (a package of education and removal of the test from our routine order set) there was an 90% reduction in unnecessary sampling. There was no significant change to duration of antibiotic, ICU length of stay or mortality in patients diagnosed as VAP following the change.

Conclusions: The use of BioFire FilmArray tests rarely led to a clinical change on our unit. Removing the test from a routine order set did not seem to have a significant clinical impact. More work Is required to assess their use as an microbial stewardship tool.


Prognostic analysis of cuproptosis-related gene in sepsis

HB Chen, HB Qiu

Jiangsu Provincial Key Laboratory of Critical Care Medicine, Department of Critical Care Medicine, Zhongda Hospital, School of Medicine, Southeast University, Nanjing, China

Critical Care 2023, 27(S1): P139

Introduction: Sepsis is a common disease in intensive care unit with high morbidity and mortality, which is a serious threat to human health. Regulated cell death plays an important role in the development of sepsis and serves as a potential therapeutic target. Cuproptosis is the most recently identified copper-dependent regulated cell death form that relies on mitochondria respiration. However, its role in the development of sepsis remains unclear. The correlation between cuproptosis-related genes (CRGs) with sepsis prognosis is also unclear.

Methods: We screened the data of sepsis patients with survival data through the Gene Expression Omnibus datasets (GEO), and collected CRGs based on previous studies. We used consensus clustering to identify gene subtypes based on CRGs, then univariate Cox regression analysis were performed on the differentially expressed genes to identify prognostic genes, and LASSO Cox regression were used to identify core genes and calculate CRG score, in addition, we create a prognostic model by multivariate Cox regression.

Results: In total, 479 sepsis samples with survival data were collected from GEO database. Most of CRGs were differentially expressed between normal and sepsis patients, and we identified two subclusters based on CRGs, and patients in cluster B had better clinical outcomes (Fig. 1). We identified 463 differentially expressed genes between the two subclusters, and 191 of them were associated with the prognosis of sepsis patients. Then, a 17-gene signature was identified utilizing the LASSO Cox regression model, and we created and validated a CRG score based on the 17-gene expression. Multivariate Cox regression analysis demonstrated CRG score was an independent prognostic factor.

Conclusions: Our thorough examination of CRGs in sepsis patients revealed the potential effects of CRGs on prognosis of sepsis. The CRGs is helpful in prognostic prediction and in guiding treatment for sepsis patients.

figure az

Fig. 1 (abstract P139). CRGs expressed in sepsis and normal patients. CRGs = cuproptosis-related genes


Plasma nucleosome measurement as a biomarker for NETs in intensive care patients

J Micallef1, M Herzog1, J Candiracci1, L Morimont2, J Douxfils2, A Aswani3, D Genkin4, P Skorup5, M Lipcsey6

1Belgian Volition SRL, Isnes, Belgium, 2Department of Pharmacy, University of Namur, Namur, Belgium, 3Department of Critical Care Medicine, Guy’s and St Thomas’ NHS Foundation Trust, London, UK, 4Santersus AG, Kusnacht, Switzerland, 5Department of Medical Sciences, Uppsala University, Uppsala, Sweden, 6Department of Surgical Sciences, Uppsala University, Uppsala, Sweden

Critical Care 2023, 27(S1): P140

Introduction: Circulating NETs have been measured using a variety of proxy biomarkers including MPO, MPO-DNA, NE, histones and cfDNA. These assays are not standardised and typically not analytically validated. We have developed the first analytically valid automated immunoassay for nucleosomes which has been CE marked as a diagnostic tool to aid the detection and evaluation of diseases associated with NETosis in Europe.


Human studies—correlation of NETs levels with SOFA and APACHE II scores in sepsis patients. The Volition Nu.Q®-NETs assay was used to measure plasma nucleosome levels in samples taken from 46 patients with septic shock (Sepsis-3) within 2 days of admission to ICU and from 48 age, gender and comorbidity matched controls.

Animal studies to monitor therapeutic removal of NETs

The Volition Nu.Q®-NETs assay was used to monitor plasma nucleosome levels to assess the efficacy of a novel apheresis treatment to remove circulating NETs in a porcine model of sepsis induced by a 3 h infusion of E. coli. The apheresis column used contained histone H1 as a specific binder of cfDNA and NETs. The efficiency of NETs removal from pig plasma passed through the apheresis column was also assessed.


Human studies: Measured nucleosome levels correlated with disease severity as indicated by SOFA and APACHE II scores.

Animal studies: E. coli infusion resulted in a continuous increase in nucleosome levels over 7 h from 58 to 356 ng/ml in a control (sham treated) pig. In contrast, levels plateaued at 120 ng/ml in the treated pig with attenuated septic shock, reduced lactate (4.3 vs 9.1 mmol/l), reduced noradrenaline requirement (120 ug vs 12,800 ug) and increased urine output (875 vs 340 ml). 97.7–99.0% of E. coli induced NETs present in the plasma of the treated pig were removed by a single pass of plasma through the apheresis column.

Conclusions: Nucleosome measurements using the Volition Nu.Q®-NETs assay may be useful in the management of patients diagnosed with sepsis.


Prognostic value of detection of bloodstream bacterial DNA in sepsis

D Yu1, B Stammler Jaliff2, A Somell3, P Dinnétz4, M Ullberg1, J Wernerman5, V Özenci1, K Strålin2

1Karolinska Institutet, Department of Laboratory Medicine, Huddinge, Sweden, 2Karolinska University Hospital, Department of Infectious Diseases, Stockholm, Sweden, 3Karolinska University Hospital, Functional Area of Perioperative Medicine and Intensive Care, Stockholm, Sweden, 4Södertörn University, School of Natural Sciences, Technology and Environmental Studies, Stockholm, Sweden, 5Karolinska Institutet, Department of Clinical Science, Intervention and Technology, Huddinge, Sweden.

Critical Care 2023, 27(S1): P141

Introduction: Using advanced molecular biological methods, such as PCR/electrospray ionization—mass spectrometry (PCR/ESI–MS), specific bacterial DNA can be detected in the bloodstream more frequently than bacteria can be detected by blood culture (BC) [1]. In the present study, we aimed to study the prognostic value of detected bloodstream bacterial DNA in patients with suspected sepsis.

Methods: Single center prospective cohort study conducted in Stockholm, Sweden. Adult patients admitted to the intensive care unit (ICU) with suspected infection based on the Centre for Disease Control and Prevention criteria for classifying infections in critically ill patients were included and subjected to sampling with BC and a whole blood tube (5 mL) for bacterial DNA (PCR/ESI–MS) on ICU admission. Presence of bloodstream bacterial DNA, positive BC and clinical characteristics were included in a multivariate analysis determining odds ratio (OR) for 60-days all-cause death after ICU admission.

Results: A total of 153 patients with infection were included, of which 48 patients (31.4%) had septic shock and 97 patients (63.4%) had sepsis without shock. In total, 52 patients (34.0%) had positive BC, and 72 patients (47.1%) had bloodstream bacterial DNA detected. Death within 60 days occurred in 42 patients (27.5%) and the adjusted odds ratio for death was 1.72 (95% CI 0.62–4.74) for bloodstream bacterial DNA and 1.07 (95% CI 0.36–3.23) for blood culture.

Conclusions: Neither blood culture nor bloodstream bacterial DNA was significantly associated with mortality in sepsis. However, the absolute OR of 1.7 for bloodstream bacterial DNA was similar to that of a previous study showing an association between bloodstream bacterial DNA and mortality in sepsis [2]. Additional studies are needed to further assess the prognostic value of bloodstream bacterial DNA in sepsis.


  1. 1.

    Tassinari M et al. PLoS One 2018;13:e0197436.

  2. 2.

    O’Dwyer MJ et al. Clin Microbiol Infect 2017;23:208 e201–208 e206.


Impact of corrective audits on the prevalence of primary bloodstream infections

EP Perecmanis, GDS Santos

Caxias Dor Hospital, Intensive Care, Duque de Caxias, Brazil

Critical Care 2023, 27(S1): P142

Introduction: Assess the impact of corrective audits on the prevalence of primary bloodstream infections.

Methods: Study conducted with prospectively collected data from January 2021 to June 2022 with a total of 3271 patients included. The intervention was implemented in June 2021, totaling 728 patients before the intervention and 2543 patients after. The central venous catheter utilization rate, the incidence of primary bloodstream infections and the mortality rate standardized by the SAPS3 score were compared. The monthly incidence rate can be viewed on Fig. 1.

Results: The incidence density of microbiologically proven primary bloodstream infections prior to the implementation of the audits was 0.79, with a central venous catheter utilization rate of 49.41%. The average achieved by the SAPS3 score of these patients was 56.63 points with a probability of death of 29.88%. The standardized in-hospital mortality rate (observed/expected) was 0.27. After the intervention, the incidence of laboratory-proven infections was 0.20, with a central venous catheter utilization rate of 33.80%. The mean by the SAPS3 score in the period was 50.62, with an in-hospital mortality rate (observed/expected) of 0.27.

Conclusions: Although there was no reduction in the standardized mortality rate, the implementation of audits seems to be effective in reducing the prevalence of primary bloodstream infections.

figure ba

Fig. 1 (abstract P142). Primary bloodstream infection incidence


Who is in my brain?

T Aydin1, SC Cetintas2, AB Keles3, HB Tokman3, O Kizilkilic4, MM Hanci2, OF Tabak5, I Bavunoglu6, G M. Koksal7

1Cerrahpasa Faculty of Medicine, Internal Medicine, Intensive Care, Istanbul, Turkey, 2Cerrahpasa Faculty of Medicine, Department of Neurosurgery, Istanbul, Turkey, 3Cerrahpasa Faculty of Medicine, Department of Medical Microbiology, Istanbul, Turkey, 4Cerrahpasa Faculty of Medicine, Department of Radiology, Istanbul, Turkey, 5Cerrahpasa Faculty of Medicine, Department of Infectious Diseases and Clinical Microbiology, Istanbul, Turkey, 6Cerrahpasa Faculty of Medicine, Department of Internal Medicine, Istanbul, Turkey, 7Cerrahpasa Faculty of Medicine, Department of Internal Intensive Care, Istanbul, Turkey

Critical Care 2023, 27(S1): P143

Introduction: Primary amoebic meningoencephalitis (PAM) is rare but a life-threatening infection caused by free-living amoebas (FLA) [1]. Herein we present a complex PAM case that has been successfully treated.

Methods: 87 year old male was transferred to intensive care unit with an initial diagnosis of viral encephalitis. Medical history was normal except for a subdural hematoma operation 20 years ago. He was in Turkey for vacation. The patient was somnolent; had fever, altered sensorium and pneumonia. Leukocytosis and hyponatremia were present. Meropenem, vancomycin, clarithromycin and acyclovir treatments were started. Cultures and Herpes, Legionella, Listeria and tuberculosis tests were negative. The patient was orotracheal intubated due to generalized status epilepticus. Cranial imaging showed increased intracranial pressure, meningeal enhancement and subdural empyema in the right frontoparietal Burr-Hole region. He was operated. Incisions were made on the old scars and the Burr-Holes from previous surgery had been used. The CSF sample was colorless and had clean appearance. 20 leukocytes were counted and motile microorganisms with flagella were noticed during the count. A protozoa with vibratory movement was seen in the wet-mount preparation at 40× magnification (Fig. 1). Intravenous(IV) amphotericin-B, azithromycin, metronidazole and dexamethasone were administered. The patient, who was intubated for 11 days, was transferred from ICU on the 26th day healthily.

Results: PAM is an acute disease of the central nervous system caused by N. fowleri. Mortality of FLA infections may exceed 90%. Only 7 of 381 confirmed PAM cases survived and all of them got IV amphotericin-B [2].

Conclusions: Early diagnosis, neuroprotective management and effective treatment improve prognosis. Recently, miltefosine is proved to be effective.

Acknowledgement: Informed consent was obtained from the patient to publish the clinical details.


  1. 1.

    Jain R et al. Neurol India 2002;50:470–2.

  2. 2.

    Gharpure R et al. Clin Infect Dis. 2021;73:e19–e27

figure bb

Fig. 1 (abstract P143). The pathognomonic images of the protozoa are shown by wet-mount preparation (left) and Gram staining (right)


Toxic epidermal necrolysis, a difficult diagnosis

J Rábano Alonso, L Cachafeiro Fuciños, A Agrifoglio Rotaeche, M Sánchez Sánchez, I Seises García, J Vejo Gutierrez, J Soto Gomez—Cambronero, R Yébenes Calvo

Hospital Universitario La Paz, Intensive medicine, Madrid, Spain

Critical Care 2023, 27(S1): P144

Introduction: Toxic epidermal necrolysis(TEN) and Stevens-Johnson syndrome (SJS) are rare, potentially disabling and fatal cutaneous diseases. Despite recent advances mortality is high, diagnosis remains a challenge and controversy exists regarding its treatment [1]. Our study describes the main characteristics of patients admitted with TEN/SJS.

Methods: A retrospective descriptive study was developed in the Critical Burn Unit of Hospital La Paz between 2013 and 2021. The diagnosis of TEN/SJS was confirmed by biopsy. Medical care was provided by a multidisciplinary team and immunosuppressive drugs was decided on an individualized basis. Demographic data, SCORTEN, frequencies of ARDS, shock, AKI requiring renal replacement therapy (RRT), infections, immunosuppressive drugs used, doubts in the clinical diagnosis and mortality were collected.

Results: 28 patients were admitted; 13 TEN, 8 TEN-SJS overlap and 6 SJS. In 5 patients with TEN/SJS (confirmed by biopsy) there were diagnostic doubts at clinical evaluation, and 5 patients with clinical suspicion of TEN were eventually diagnosed with other serious skin conditions: graft-versus-host disease, paraneoplastic pemphigus, and drug reaction with eosinophilia and systemic symptoms. Among those with TEN we found: mean age 56 ± 17 years; 13 men; mean skin detachment area 33 ± 26% (5–90); SCORTEN 3; mortality 31%; 11 presented shock, 5 ARDS, 5 needed RRT and 14 had severe infections. Figure 1 shows immunosuppressive drugs.

Conclusions: Data and mortality are consistent with literature and SCORTEN. Heterogeneity in immunosuppressive drugs used highlights the lack of consensus in treatment. Finally, similarities between the clinical presentation with other serious skin diseases highlight the difficulty in diagnosis and the possible delay in treatment. In any case, high-quality multidisciplinary supportive care is essential in all patients with severe skin diseases. Further studies are needed to improve the care of these patients.


  1. 1.

    Jacobsen A et al. Cochrane Database Syst Rev 2022:CD013130.

figure bc

Fig. 1 (abstract P144). Relative frecuencies of diagnostic doubts among confirmed TEN/SJS cases and immunosuppressive drugs used


Dynamics of severe comorbidity and survival in Clostridioides difficile patients during 10 year period

D Adukauskienė, R Mickus, A Dambrauskienė

Medical Academy, Lithuanian University of Health Sciences, Kaunas, Lithuania

Critical Care 2023, 27(S1): P145

Introduction: Comorbidities are a well-established risk factor for contracting Clostridioides difficile infection (CDI) [1]. The aim of this study was to evaluate the prevalence of severe comorbidity and its impact on in-hospital mortality in patients (pt) with CDI.

Methods: We conducted a single-center, retrospective 10 years study in a reference Lithuanian university hospital. Pt with initial episode of CDI (diarrhoea and positive stool test for C. difficile toxin A/B) were included. The impact of severe comorbidity on 30 day in-hospital survival rate has been estimated in period of 2011–2020. Prevalence of severe comorbidity (Charlson Comorbidity Index (CCI) ≥ 5) and its relation to in-hospital survival have been estimated in both group 1 (2011–2016) and 2 (2017–2020). IBM SPSS 23.0, Kaplan Meier model, log-rank and Pearson’s χ2 tests were used for statistics, level of significance—p < 0.05.

Results: A total of 370 pt was enrolled. In 2011–2020 a lower 30 day in-hospital survival amongst CDI pt with CCI ≥ 5 was found (p = 0.000001, Fig. 1). In group 1 there were n = 197 pt and in group 2 n = 173 pt. In group 1 CCI ≥ 5 has been estimated in n = 121 (61.4%) pt and n = 169 (76.9%) have survived. In group 2 on contrary, both severe comorbidity and survival decreased: n = 80 (47.1%) pt and n = 143 (62.5%), accordingly (p = 0.005, p = 0.038).

Conclusions: Severe comorbidity has decreased, but in-hospital survival recently has not improved in CDI patients. Severe comorbidity as a prognostic factor has been associated with lower 30 day in-hospital survival in a period of 10 years.


  1. 1.

    Anjewierden S et al. Infect Control Hosp Epidemiol 2021;42:565–72.

figure bd

Fig. 1 (abstract P145). Kaplan–Meier plots estimating 30-day survival of CDI with and without severe comorbidity in a period of 10 years


In vitro activity of antibiotics against Clostridioides difficile: can we rely on treatment guidelines?

J Závora, G Kroneislová, V Adámková

General University Hospital, Clinical Microbiology and ATB Centre, Prague, Czech Republic

Critical Care 2023, 27(S1): P146

Introduction: Clostridioides difficile (CLDI) is a well-recognized causative agent of severe healthcare associated diarrhoea. Recently, there was an update of European CLDI infection (CDI) treatment guidelines removing metronidazole from recommended regimen if vancomycin and fidaxomicin are available. Vancomycin is the most used agent against CDI and fidaxomicin is recommended only for treatment of initial CDI or the first recurrence. Several other agents with potential anti-CDI effect, such as tigecycline or rifampicin, are being considered and tested.

Methods: We have chosen 60 isolates of CLDI obtained from stool samples of patients hospitalized in General University Hospital in Prague in years 2020–2021. All samples have been initially tested by quick immunochromatographic test for detection of antigen (GDH) and toxin A and/or B. When positive, culture on selective agar media was performed. Minimal inhibition concentration of vancomycin, metronidazole, tigecycline and rifampicin was then determined by gradient strip method (Etest®). There is no commercial method for testing susceptibility to fidaxomicin available.

Results: Interpretation criteria (breakpoints) are provided only for vancomycin and metronidazole. All the strains tested were susceptible to both antibiotics, the range of MIC was 0.064–2 mg/l in vancomycin and 0.064–1 mg/l in metronidazole. Therefore, several strains were approaching breakpoint and there is certainly potential of resistance in this cohort. It is not possible to interpret the susceptibility of tigecycline and rifampicin, but the MICs suggested possibility of low efficiency, especially with the range of 0.016–32 mg/l in rifampicin. The MICs overall were the lowest in tigecycline, 0.016–1 mg/l.

Conclusions: The established anti-CDI therapy regimen is appropriate, but it is important to investigate potential patterns of resistance. There are alternative antibiotic agents with activity against CLDI (the best results given by tigecycline), but further testing is crucial.


Septic shock due to C. canimorsus treated with IgM-enriched immunoglobulin as adjuvant therapy in an immunocompetent woman

J Braunsteiner, S Siedler, D Jarczak, S Kluge, A Nierhaus

University Hospital Hamburg-Eppendorf, Department of Intensive Care, Hamburg, Germany

Critical Care 2023, 27(S1): P147

Introduction: Capnocytophaga canimorsus is a facultative anaerobic Gram-negative bacillus and is part of the microflora of canine and feline oral cavities. C. canimorsus possesses active and passive mechanisms of immune evasion. Invasive infections are rare and have been primarily described in immunocompromised patients.

Methods: We report a case of a healthy, immunocompetent 51-year-old woman with purpura fulminans due to septic shock caused by C. canimorsus after having been scratched by cats.

Results: The patient was admitted to a community care hospital and treated with empirically targeted antibiotic therapy. 48 h after admission, C. canimorsus could be identified as causative pathogen, and her antibiotic regimen was adapted. She was referred to our department because of multi-organ dysfunction and refractory shock. The blood gas analysis showed hyperlactatemia with 17 mmol/l. The inflammatory markers were increased (WBC 16.2 G/l, procalcitonin 49.6 µg/l, CRP 247 mg/l, IL-6 3321 ng/l). Further analysis revealed pathologically low IgM levels (0.31 g/l). Adjuvant therapy with IgM-enriched immunoglobulin (Pentaglobin®) was initiated to restore normal values. The further course was complicated by acute respiratory failure, multiple regional infarctions, and recurrent septic episodes. IgM had decreased to 0.32 g/l on day 78, which was again treated with Pentaglobin®. After prolonged intensive care therapy, her condition improved, and she was transferred to a rehabilitation center.

Conclusions: Data on immunoglobulin therapy in sepsis is conflicting and there is a lack of evidence in patients with invasive C. canimorsus infection. However, following the administration of immunoglobulins the patient showed sustained clinical improvement. Patients with both a high inflammatory load and IgM deficiency might benefit from therapeutic IgM-administration.

Acknowledgement: We obtained informed consent from the patient to publish the clinical details. All participating physicians also consented to the publication.


Sepsis after being licked in the face by a dog

M Bouwens, Y Acardag, JJ Arends

Albert Schweitzer Ziekenhuis, Intensive Care, Dordrecht, Netherlands

Critical Care 2023, 27(S1): P148

Introduction: We present a case where sepsis was caused by a dog licking its owners face. Sepsis is usually caused by pet bites.

Methods: Case report.

Results: A 58-year-old male presented at our hospital with severe sepsis. Patient history did not present any potential diagnostical clues. The patient owned a dog, but denied being bitten. He did not use any immunosuppressive medication. No focus for an infection was found during physical examination. Imaging studies were non-conclusive. He was admitted to the intensive care unit (ICU) with rapid progressive multi–organ failure and treated with mechanical ventilation, inotropes and broad–spectrum antibiotics. Initial blood and urine cultures remained negative. On day 8 of his ICU stay a 16S rRNA gene-sequencing test revealed an bacteremia with Capnocytophagus canimorsus, a slow–growing, gram–negative rod, present in dog or cat saliva and associated with pet bites [1]. Antibiotic treatment was adapted, and the patient made a quick recovery. Revision of the CT–images showed a small spleen (52.4 mm). Additional blood tests showed Howell-Jolly bodies, indicating functional asplenism. After cessation of mechanical ventilation, the patient admitted to letting his dog lick his face as a sign of friendship. The Dutch national guidelines for asplenic patients advice vaccination for capsulated bacteria and the use preventive antibiotics when being bitten by a pet [2].

Conclusions: Facial pet-licking can cause severe sepsis in functional asplenic humans.

Acknowledgement: Written consent for publication was obtained from this patient.


  1. 1.

    Brenner DJ et al. J Clin Microbiol. 1989;27:231–5.

  2. 2.

    Dutch National Institute for Public Health and the Environment:


Evaluation of risk factors associated with mortality in patients with ventilator-associated pneumonia caused by Pseudomonas aeruginosa with “difficult-to-treat” resistance

I Seises, M Rodríguez-Aguirregabiria, JM Añon-Elizalde, KL Nanwani-Nanwani, MJ Asensio-Martin

University Hospital La Paz, Intensive Care Medicine, Madrid, Spain

Critical Care 2023, 27(S1): P149

Introduction: Evaluation of prognostic factors in patients with ventilator-associated pneumonia (VAP) due to P. aeruginosa. The effectiveness of novel antipseudomonal antibiotics was reviewed.

Methods: Retrospective, single-center cohort analysis between April 2018 and June 2022. Data were obtained from the ENVIN-HELICS and electronic medical records. Demographic variables, underlying diseases and diagnosis to admission were registered. We considered each treatment appropriate according to Tamma PD et al. [1] criteria. We registered ventilator-associated tracheobronchitis (VAT) and pneumonia (VAP) episodes together with the recurrency of the infection.

Results: From 61 patients included, 77% were admitted for ARDS due to COVID-19. The mean APACHE-II was 14.3 ± 6.6. 7 patients required ECMO and 4 required RRT. The median length of stay in the ICU was 52 (ICR 36–84) days. 91 respiratory infections were recorded: 60 VAP and 31 VAT. On the first episode, carbapenem-resistance to meropenem was 40%; rising up to 58% on the second one. 6 patients developed a third episode (VAT) with a 100% of carbapenem-resistance. 13 (14%) respiratory infections showed resistance to the novel β-lactamase inhibitor cephalosporins (8 to ceftalozane-tazobactam and 5 to ceftazidime-avibactam). No resistance to cefiderocol was detected. During ICU stay, 21 patients (34%) developed secondary bacteremia from other foci and 7 (11%) invasive mycoses. Overall mortality was 49.2%. On the univariate analysis we found statistical significant relationships between mortality and COVID-19 admission, SOFA ≥ 7 points on the first VAP or the development of secondary bacteremia (Table 1).

Conclusions: COVID-19 admission, SOFA ≥ 7 points on the first VAP or other secondary bacteremia were associated with mortality. The 14.3% of respiratory infections were resistant to the new β-lactamase inhibitor cephalosporins. No resistance to cefiderocol was detected.


  1. 1.

    Tamma PD et al. Clin Infect Dis. 2022;75:187–212.

Table 1 (abstract P149). Univariate analysis of risk factors associated with ICU mortality

Charlson index ≥ 3

Alive 12 (37.5%)

Deceased 20 (62.5%)

OR 2.36

p 0.099

Admission due to severe COVID-19 pneumonia

Alive 19 (40.4%)

Deceased 28 (59.6%)

OR 3.68

p 0.41

SOFA-score ≥ 7 on the 1st episode of infection

Alive 10 (31.3%)

Deceased 22 (68.8%)

OR 4.18

p 0.007

ECMO therapy

Alive 1 (14.3%)

Deceased 6 (65.7%)

OR 6.46

p 0.061

Bacteremia by other microorganisms

Alive 5 (23.8%)

Deceased 16 (76.2%)

OR 4.80

p 0.007

Carbapenemase colonization prior to infection

Alive 5 (31.3%)

Deceased 11 (66.8%)

OR 2.51

p 0.129

Inappropriate treatment on first episode of infect

Alive 5 (33.3%)

Deceased 10 (66.7%)

OR 2.18

p 0.204

Resistance to new β-lactams with β-lactamase inhib

Alive 1 (20%)

Deceased 4 (80%)

OR 4

p 0.198


Are new antibiotics efficient against DTR (difficult-to-treat resistance) isolates? Prevalence and susceptibility of invasive DTR strains

G Kroneislová, J Závora, V Adámková

General University Hospital Prague, Clinical Microbiology and ATB Center, Prague, Czech Republic

Critical Care 2023, 27(S1): P150

Introduction: Bacterial resistance surveillance is one of the main outcomes of microbiological laboratories and its results are an important part of antimicrobial stewardship (AMS). We point out a relatively new ‘difficult-to-treat’ (DTR) category of resistance, which—when used—has a significant influence on predictability of morbidity and mortality in patients with invasive infections.

Methods: The susceptibility of 100 unique isolates of the most common Gram-negative bacilli (Enterobacterales, Pseudomonas aeruginosa, Acinetobacter baumannii) obtained from blood cultures of patients hospitalised in General University Hospital in Prague in 2017–2021 was tested, 50% of these fulfilled DTR criteria (nonsusceptibility to all beta-lactams—incl. carbapenems—and fluoroquinolons) and 50% were susceptible.

Results: 10 Enterobacterales strains met DTR criteria and 20% of those were resistant to colistin (COL), 20% to cefiderocol (FCR), 70% to imipenem/cilastatin/relebactam (I/R), 30% to ceftazidime/avibactam (CAT), 50% to fosfomycin (FOS) and 50% tigecycline (TGC). For Enterobacterales we also tested aztreonam/avibactam (AZA) for which there are no breakpoints yet. The highest MIC of AZA observed was 0.75 mg/l, in the susceptible cohort the range was 0.023–0.064 mg/l and in the DTR cohort (incl. class B beta-lactamase producers) it was 0.047–0.75 mg/l. In P. aeruginosa (15 DTR strains) the resistance to COL were 13%, to FCR 6.6%, to I/R 86.6%, to CAT 46.6%, and to ceftolozane/tazobactam 26.6%. All DTR isolates of A. baumannii were susceptible to COL and FCR and resistant to I/R and ampicillin/sulbactam.

Conclusions: The cooperation of microbiologists and clinicians is crucial when treating severe infections, especially in the era of increasing antimicrobial resistance. Not even new antimicrobial agents are 100% efficient against DTR strains. Therefore, it is necessary to perform susceptibility testing of these antibiotics, use the data for surveillance (including local surveillance) and conform to AMS standards.


Amoxicillin/clavulanate versus 3rd generation cephalosporins for community-acquired pneumococcal pneumonia: is it an arbitrary choice?

P Teixeira1, L Costa2, JG Pereira3, SM Fernandes1

1Faculdade de Medicina da Universidade de Lisboa, Clinica Universitária de Medicina Intensiva, Lisboa, Portugal, 2Hospital de Braga, Braga, Portugal, 3Hospital de Vila Franca de Xira, Serviço de Medicina Intensiva, Hospital de VIla Franca de Xira, Portugal

Critical Care 2023, 27(S1): P151

Introduction: European guidelines recommend ceftriaxone or amoxicillin/clavulanate (AAC) as equally effective first-line antibiotic therapy for community-acquired pneumonia (CAP). This study aims to identify factors, including patients' clinical characteristics, that may influence the choice of beta-lactam empirical therapy for CAP and to relate this with patients' prognosis.

Methods: This is a sub-analysis of a retrospective multicentric cohort of pneumococcal CAP. Clinical characteristics and disease severity at admission were compared according to their initial empirical antibiotic therapy: ceftriaxone, AAC, or other. Mortality and days of hospitalization were compared between groups, adjusting for severity, age, and ICU admission. STATA v15.0 software was used for statistical analysis.

Results: We included 797 patients. 40.9% received ceftriaxone, 41.1% AAC, and 17.9% other therapy. Patients receiving ceftriaxone were more likely to be septic (61% vs 34% vs 48%, p < 0.001). Patients admitted to the hospital due to cardiovascular disease (8% vs 13% vs 6%, respectively; p = 0.01) and with hyponatremia (29.9 vs 76.8 vs 29.8%, p < 0.001) were more likely to receive AAC. Patients receiving “other” antibiotic therapy were less likely to receive corticosteroids or a macrolide. Age, comorbidities, multilobar pneumonia or pleural effusion did not influence empirical antibiotic choice. No differences in mortality were noted between different empirical therapy choices (non-adjusted: OR 0.70 (0.46–1.06), p = 0.07; adjusted for sepsis, hyponatremia, and age: OR 0.79 (0.5–1.2), p = 0.28). As expected, sepsis (OR: 1.6, p = 0.02) and age (OR:1.1, p < 0.001) were relevant risk factors for death.

Conclusions: The choice between empirical therapy for CAP does not appear to be random despite comparable outcome effect between first line antibiotics. This data may inform educational strategies on antibiotic stewardship.


The combination of ceftazidime-avibactam plus meropenem can be a reliable therapeutic option for ventilator-associated pneumonia (VAP) due to pan-drug resistant Klebsiella pneumoniae

E Kourtelesi1, A Aiginitou1, C Merkouri1, M Poulou1, E Bourganie1, S Giannakaki1, E Dikoudi1, M Karagianni1, M Daganou1, A Flevari2

1Sotiria Thoracic Diseases Hospital, Athens, Greece, 2Sotiria Thoracic Diseases Hospital, Polyvalent ICU, Athens, Greece

Critical Care 2023, 27(S1): P152

Introduction: Over the last two decades, the prevalence of pandrug-resistant Klebsiella pneumoniae (PDR-Kp) has increased worldwide. Such strains along with the burden of co-morbidities, especially in ICU patients, pose therapeutic challenges and contribute to high mortality rates. Combination of antimicrobial drugs is one of the few options available to overcome this challenge. Among others the combination of ceftazidime-avibactam (CZA) plus meropenem (MEM) has been proposed to be effective and safe [1].

Methods: A prospective single cohort study was conducted from March to August 2022. Among 50 patients admitted to our General ICU, 8 patients were diagnosed with VAP due to PDR-Kp. After clinical suspicion, VAP diagnosis was performed from microbiological and clinical tools, namely positive tracheal aspirate cultures (≥ 105 cfu/ml) and CPIS score ≥ 6. Kp was considered PDR if microbiological tests revealed resistance to cephalosporins, carbapenems, monobactams, quinolones, sulfamethoxazole-trimethoprim, aminoglycosides, colistin, tigecycline, fosfomycin and ceftazidime/avibactam. If patients were haemodynamic stable, with no need of vasopressors, combination therapy CZA-MEM was introduced. Primary outcome was the resolution of pneumonia and secondary outcome was 30-day mortality.

Results: Table 1 tabulates clinical characteristics of the population enrolled. Among 8 patients (3 female) with PDR-Kp associated VAP, who received the combination of CZA-MEM, all showed signs of clinical resolution of pneumonia. CPIS score reduction after 8–10 days of therapy was statistically positive (p < 0.001). Tracheal aspirate cultures, who were subsequently obtained, showed microbiological eradication of the specific strain. All patients tolerated therapy with no adverse events. 30-day mortality was 25% (2/8).

Conclusions: In the era of multidrug resistance the synergistic activity of MEM plus CZA against PDR-Kp can be an efficient therapeutic option in critically ill patients.


  1. 1

    Michail S et al. Antimicrob Agents Chemother 2019;63:e00779–19.

Table 1 (abstract P152). Clinical characteristics of patients with VAP due to PDR/Kp




Age (years)



Length of ICU stay (days)



APACHE II on admission



SOFA score on admission



CHARLSON comorbidity index




Model-informed precision dosing of beta-lactam antibiotics and ciprofloxacin in critically ill patients

TMJ Ewoldt1, A Abdulla2, WJR Rietdijk2, AE Muller3, BCM De Winter2, NGM Hunfeld1, D Gommers1, H Endeman1, BCP Koch2

1Erasmus University Medical Center, Intensive Care Medicine, Rotterdam, Netherlands, 2Erasmus University Medical Center, Hospital Pharmacy, Rotterdam, Netherlands, 3Erasmus University Medical Center, Microbiology, Rotterdam, Netherlands

Critical Care 2023, 27(S1): P153

Introduction: Individualising drug dosing using model-informed precision dosing (MIPD) of beta-lactam antibiotics and ciprofloxacin has been proposed as an alternative to standard dosing to optimise antibiotic efficacy in critically ill patients. However, randomised clinical trials (RCT) on clinical outcomes have been lacking.

Methods: This multicentre RCT, including patients admitted to the ICU who were treated with antibiotics, was conducted in eight hospitals in the Netherlands [1]. Patients were randomised to MIPD with dose and interval adjustments based on monitoring serum drug levels (therapeutic drug monitoring) combined with pharmacometric modelling of beta-lactam antibiotics and ciprofloxacin (Fig. 1). The primary outcome was ICU length of stay (LOS). Secondary outcomes were ICU mortality, hospital mortality, 28-day, and 6-month mortality, delta sequential organ failure assessment (SOFA) score, adverse events, and target attainment.

figure be

Fig. 1 (abstract P153). Study flow and dose advice given

Results: In total, 388 (MIPD n = 189; standard dosing n = 199) patients were analysed (median age 64 [IQR 55–71]). In the intervention group a dose adjustment was recommended for 71 (37.6%), 27 (23.1%), and 8 (13.3%) patients at T1, T3, and T5, respectively. We found no significant differences in ICU LOS between MIPD compared to standard dosing (10 MIPD vs 8 standard dosing; IRR = 1.16; 95% CI 0.96, 1.41; p = 0.13). There was no significant difference in target attainment before intervention at day 1 (T1) (55.6% MIPD vs 60.9% standard dosing; p = 0.24) or at day 3 (T3) (59.5% vs 60.4%; p = 0.84). There were no significant differences in other outcomes.

Conclusions: We could not show a beneficial effect of MIPD of beta-lactam antibiotics and ciprofloxacin on ICU LOS in all critically ill patients. Maximal dosing according to the summary of product characteristics was often insufficient to reach target attainment. Our data highlights the need to identify other approaches to dose optimization, including earlier optimisation and patient selection.


  1. 1.

    Ewoldt TMJ et al. Intensive Care Med. 2022;48:1760–1771.


Beta-lactam concentrations in patients with sepsis treated in the ICU

E Kryss1, K Liljedahl Prytz2, M Sundqvist3, J Oxelbark4, K Nilsson5, J Källman6, J Savilampi5

1Institution of Medical Sciences, Örebro, Sweden, 2Institution of Medical Sciences, Örebro Univerisity, Örebro, Sweden, 3Institution of Medical Sciences, Clinical Microbiology, Örebro, Sweden, 4Institution of Medical Sciences, Clinical Chemistry, Örebro, Sweden, 5Institution of Medical Sciences, Anaesthesiology and Intensive Care, Örebro, Sweden, 6Institution of Medical Sciences, Infectious Diseases, Örebro, Sweden

Critical Care 2023, 27(S1): P154

Introduction: Sepsis pathophysiology may lower plasma concentrations of antibiotics thus reducing antibacterial effects. Optimal dosing of beta-lactam antibiotics in septic patients remains challenging [1]. Time above minimal inhibitory concentration (MIC) is the determinant of beta-lactam antibiotic efficacy. Plasma concentration from 1 to 4 × MIC for 100% of the dosing interval has been suggested as target level. Repeated measurement of plasma concentrations are proposed to avoid insufficient concentrations of beta-lactams in critically ill patients. The aim was to describe plasma concentration achievement of beta-lactams in septic patients during the first 48 h after admittance to the ICU.

Methods: A prospective, observational study included adults with community-acquired sepsis admitted from the emergency department to the ICU. Antibiotic plasma concentration measurements during administration of piperacillin-tazobactam [PTZ, 4 g× 3–4], cefotaxime [CTX, 2 g× 3] and benzyl-penicillin [PEN, 3 g× 3]) were collected at 2–7 occasions/patient and analysed using liquid chromatography-mass-spectrometry. To evaluate time above MIC, theoretical MICs representing the EUCAST I breakpoints were used for bacteria common in community acquired pathogens.

Results: Out of the thirty-two patients included, 56% were men (18/32), with median age of 71 (min–max 28–85) and with median SOFA score of 7 (min–max 3–13). They were either treated with PTZ (n = 23), PEN (n = 4) or CTX (n = 11). Seven patients had an altered antibiotic regimen during the 48 h, resulting in a total of 169 plasma concentration measurements. The percentage of plasma concentrations above MIC × 1/ × 4 during the 48 h was 97/89% for PTZ (n = 113), 91/73% for PEN (n = 11) and 98/76% for CTX (n = 45) (Fig. 1).

Conclusions: In septic patients receiving standard dosing of CTX, PTZ or PEN, plasma concentrations had a large variation but most were above effective concentrations.


  1. 1.

    Petersson J et al. Acta Anaesthesiologica 2016;60:1425–1436.

figure bf

Fig. 1 (abstract P154). Plasma concentrations of beta-lactams (A: PTZ [n = 113], B: CTX [n = 45] and C: PEN [n = 11]), related to time from antibiotic dose to blood sampling. Concentrations were compared to the clinical susceptibility breakpoints for Enterobacteriaceae (PTZ 8 mg/l and CTX 2 mg/l) and S. pneumoniae (PEN 2 mg/l). These breakpoints are the highest MICs for pathogens considered susceptible if increased exposure for these antibiotics. Target plasma concentration levels are shown in graphs as MICx1 and MICx4


A photo finish of machine learning models for predicting beta-lactam target non-attainment in ICU patients

A Wieringa1, TMJ Ewoldt2, R Gangapersad3, M Gijsen4, A Abdulla3, BCP Koch3

1Isala, Dept. of Clinical Pharmacy, Zwolle, Netherlands, 2Erasmus University Medical Center, Department of Hospital Pharmacy and Intensive Care, Rotterdam, Netherlands, 3Erasmus University Medical Center, Department of Hospital Pharmacy, Rotterdam, Netherlands, 4KU Leuven and UZ Leuven, Department of Pharmaceutical and Pharmacological Sciences, Leuven, Belgium

Critical Care 2023, 27(S1): P155

Introduction: In the ICU, the infection-related mortality is high (30%) and beta-lactam target non-attainment (TNA) occurs in 37% of ICU patients [1, 2]. The recommended pharmacodynamic target for beta-lactam antibiotics is 100% of the time above the minimal inhibitory concentration [1]. Antibiotic dose optimization can be performed with therapeutic drug monitoring (TDM). In this study we therefore developed prediction models for TNA in ICU patients receiving beta-lactams.

Methods: Data of the Dolphin and Expat studies were used to train and internally validate three machine learning models to predict TNA on the day after antibiotic initiation [2]. After excluding patients on dialysis, 378 patients remained. For covariate selection, the Boruta algorithm was used in the training set and afterwards manually verified by adding and removing covariates. A random forest (RF), logistic regression (LR) and naïve bayes (NB) model were developed and externally validated with data from UZ Leuven and Amsterdam UMC (150 patients). Model performance was evaluated using a 1000-times cross-validation. The main performance metric used was the area under the receiver operator curve (AUROC).

Results: The covariates age, sex, serum creatinine and type of antibiotic were found to be TNA predictors. In the internal validation, the NB, LR and RF models showed an AUROC of 0.82 [0.74, 0.91], 0.81 [0.76, 0.91] and 0.81 [0.75, 0.91], respectively. In the external validation, the models showed an AUROC of 0.75 [0.67, 0.82], 0.80 [0.73, 0.87] and 0.79 [0.72, 0.86], respectively (Fig. 1 including other metrics).

Conclusions: All models showed good performance in both validation sets using four readily available covariates. These models can therefore be easily implemented in daily clinical practice to identify patients at risk of TNA, in which TDM of beta-lactams is indicated, hence allowing more efficient use of limited hospital resources.


  1. 1.

    Abdul-Aziz MH et al. Intensive Care Med 2020;46:1127–1153.

  2. 2.

    Abdulla A et al. Crit Care 2020;24:558–569.

figure bg

Fig. 1 (abstract P155). Performance metrics for the target non-attainment models in the internal validation set (A) and the external validation set (B). NB: naïve bayes; LR: logistic regression; RF: random forest; AUC: Area under the receiver operator curve; PPV: positive predictive value; NPV negative predictive value; Red line: The chosen threshold for each metric


How many vancomycin levels are required for PK/PD dosing recommendation using Bayesian dose-optimizing software in critically ill patients?

ND Dreyse, P Baeza, N Salazar, R Lopez

Clínica Alemana de Santiago, Paciente Crítico, Santiago, Chile

Critical Care 2023, 27(S1): P156

Introduction: The aim of this study was to determine of the quantity of vancomycin levels be required for PK/PD dosing recommendation using Bayesian dose-optimizing software with more accuracy and less bias. The trearment guidelines recommend an individualized target of AUC/MIC 400–600, assuming MIC 1 mg/l. However, most programs performing Bayesian estimation require only one vancomycin level. Currently no data suggest how many vancomycin levels should be collected for the correct estimation AUC/MIC in critically ill patients.

Methods: In this prospective study performed in the ICU of Clinica Alemana de Santiago, patients receiving vancomycin as part their care was enrrolled. Vancomycin levels were taken at the end-infusion (peak), 120 min following the end-infusion (beta) and immediately before the next dose (trough). The AUCREF was calculated by linear-long trapezoidal rule using all vancomycin levels taken. The Bayesian software used was PrecisePK™, PreciseRX, San Diego, USA. Five different AUCs were obtained by PrecisePK™ as follow: 1.- peak, beta and trough; 2.- beta and trough; 3.- peak and trough; 4.- trough; 5.- AUC obtained PrecisePK™ without vancomycin level. Accuracy was defined as the median ratio of the estimated AUC to AUCREF. Bias was defined as the median of the absolute value of the percentage difference of the estimated AUC from AUCREF. Then each AUC obtained by PrecisePK™ was comapred with AUCREF using Mann–Whitney U Test. p value was fixed at < 0.05.

Results: Thirty-six patients were inclued. In general, no significant differences were observed between five ways to approach estimated AUC. However, the most accurate and the least biased estimated AUC was using peak and trough.

Conclusions: The findings of this study showed that the critically ill patients Bayesian dose-optimizing program using peak and trough produce estimated AUC more accurate and with less bias. These results give an empirical basis to support the recommendations of current clinical guidelines.


Aminoglycoside dosing in intensive care: are we getting it right?

F Mohideen, A Collins, L Coslett, A Conway Morris, D Sapsford

Addenbrooke’s Hospital, John V Farman Intensive Care Unit, Cambridge, UK

Critical Care 2023, 27(S1): P157

Introduction: Aminoglycosides (AMGs) are extensively used in intensive care units (ICUs) to treat gram-negative bacterial infections, yet there is much variation in practice to ensure optimal dosing. Adverse effects are often associated with sustained trough concentrations and include neuromuscular weakness, ototoxicity and nephrotoxicity, with some studies reporting the latter in > 58% of all ICU patients receiving AMGs [1]. Therefore, we investigated the AMG prescribing practices in the ICU of a large tertiary centre in the UK.

Methods: All ICU patients prescribed AMGs were identified from March 2021 to March 2022, excluding indications for infective endocarditis and medical or surgical prophylaxis. A retrospective review of patients’ clinical notes alongside the hospital’s AMG guideline was performed to assess prescribing and protocol compliance as a registered service evaluation.

Results: 228 patients were identified with 136 receiving gentamicin (60%) and 92 receiving amikacin (40%). 159 patients received a single dose, where 94 (59%) of these did not have a subsequent trough level taken in case of requirement for further doses. 69 patients (30%) received multiple doses. Prescription errors were identified in 28 of these (41%). Errors included subsequent doses being given too early (< 24 h) in 43% (12/28), without a confirmed trough level in 21% (6/28) or a combination of both in 21% (6/28). Moreover, 14% of these patients (4/28) received a subsequent dose despite a high trough level.

Conclusions: There was a risk of significant harm for a small number of patients receiving multiple doses. ICU-specific AMG guidelines need to be developed in line with current evidence and a safety alert will be added onto the prescribing system. Re-audit following these interventions is planned alongside engagement of ICUs from other centres to promote collaborative learning and service improvement.


  1. 1

    Oliveira JF et al. Antimicrob Agents Chemother 2009;53:2887–91.


Concentration target attainment after therapeutic drug monitoring of linezolid in critically ill patients

J Langui1, P De Cock1, N Verougstraete2, J Boelens2, A Somers1, J De Waele3, B Claus1

1Ghent University Hospital, Pharmacy Department, Ghent, Belgium, 2Ghent University Hospital, Department of Laboratory Medicine, Ghent, Belgium, 3Ghent University Hospital, Department of Intensive Care Medicine, Ghent, Belgium

Critical Care 2023, 27(S1): P158

Introductions: Linezolid (LZD) is often used to treat severe Gram-positive infections. In recent years, pharmacokinetic variability of LZD has been described in critically ill patients. The aim of this study was to analyse LZD exposure in critically ill patients monitored with TDM.

Methods: A retrospective observational study of all adult intensive care unit (ICU) patients receiving LZD for at least three days was conducted between August 2021 and February 2022 (Ethics Committee BC-11209). According to the institutional protocol at the ICU of the Ghent University Hospital, TDM of LZD is recommended in patients with augmented renal clearance, renal/hepatic impairment, low baseline platelet count, sepsis/septic shock and/or expected treatment duration > 14 days [1]. Decision to monitor is at the discretion of the treating physician. Target attainment, defined as a trough concentration between 2 and 7 mg/L, was the primary endpoint and TDM-guided dose adjustments were descriptively analysed.

Results: Of the 47 patients treated with LZD during the study period, TDM was performed in 14 patients (29.8%), all meeting the criteria. Noteworthy, 90.9% of the patients without monitoring were eligible for TDM as well. Intra-abdominal infections were the most common infection (42.6%). Median treatment duration of LZD in ICU was 5.7 [3.7–8.5] days for all patients. Therapeutic range was reached at first measurement in 9 of 14 (65.0%) patients. After TDM-guided dose adjustments, a total of 11/14 (78.6%) patients were within therapeutic range after a median of 2.6 [2.5–4.5] days.

Conclusions: Two thirds of the TDM patients reached therapeutic concentrations at first measurement and an extra 13.6% patients after dose adjustment. More patient data are needed to refine the criteria for monitoring critically ill patients most at risk for target non-attainment and to draw more robust conclusions.


  1. 1.

    Rao GG et al. Ther Drug Monit. 2020;42:83–92.


Meropenem model-informed precision dosing (MIPD) in the treatment of critically ill patients: which population pharmacokinetic model to choose?

L Li1, S D.T. Sassen1, T M.J. Ewoldt2, A Abdulla1, N G.M. Hunfeld2, A E. Muller3, B C.M. de Winter1, H Endeman2, B C.P. Koch1

1Erasmus Medical Center, Pharmacy, Rotterdam, Netherlands, 2Erasmus Medical Center, Intensive Care, Rotterdam, Netherlands, 3Erasmus Medical Center, Medical Microbiology and Infectious Diseases, Rotterdam, Netherlands

Critical Care 2023, 27(S1): P159

Introductions: Population pharmacokinetic (PopPK) studies of meropenem in intensive care unit (ICU) patients have shown large differences in estimated pharmacokinetic (PK) parameters, making it difficult to select an appropriate model for clinical use. We performed external validation using real-world ICU patient data to assess the suitability of published meropenem models for clinical application in initial dosing and therapeutic drug monitoring (TDM) based dose adjustments.

Methods: All these models were replicated in NONMEM® 7.4. The predictability of the models was evaluated using a dataset collected from 20 ICU patients. The ability of the models to estimate the starting dose and the ability to adjust towards the target concentration based on previous concentrations was evaluated using prediction-based and simulation-based diagnostics.

Results: 18 parametric PopPK models were included for assessment. For initial dosing prediction evaluation with the patients’ characteristics only, neither the GOF plot nor VPC plot fit well, except for 3 models. The predictive power of all models was low, with prediction error (PE)% of the interquartile range (IQR) exceeds ± 30% threshold (Fig. 1). For TDM-based dose adjustment assessment, the model fit improved after accounting for measured concentrations in the GOF plot. When incorporating 1–3 concentrations into five selected models, two models had an IQR PE% within ± 30%, when 2 or 3 concentrations were utilized.

Conclusions: Pharmacokinetic parameters from controlled ICU trial should not be used to guide initial dosing in real-world patients. On the other hand, certain models are suitable for TDM based dose adjustment with at least two or more samples incorporated.

figure bh

Fig. 1 (abstract P159). Prediction error of 5 models with prior 0–3 concentrations


COVID-19 patients are at greater risk of developing fungal infections with prolonged course of steroids and invasive ventilation in intensive care unit (ICU)

J Ainsworth, P Sewell, S Eggert, S Pillai

Morriston Hospital, Ed Major Intensive Care Unit, Swansea, UK

Critical Care 2023, 27(S1): P160

Introduction: COVID-19 pandemic infection has affected over 650 million people with over 6 million deaths. Critically unwell patients are at increased risk of developing invasive fungal infections [1]. The aim of this study was to identify the number of patients admitted to ICU with COVID-19 who developed fungal infections and to compare these patients (fungal group) with those without fungal infections (non-fungal group) to investigate which factors may have contributed to increased risk of infection.

Methods: Retrospective study undertaken in a tertiary teaching hospital ICU. 174 patients admitted with severe COVID-19 infection during March 2020 until May 2021 were included.

Results: 81(47%) patients developed fungal infections of which 94% had Candida and 6% had Aspergillus infection. Age and smoking history did not appear to be a contributing factor. The non-fungal group had significantly higher body mass index (33 ± 8 vs 31 ± 7, p = 0.01). ICU length of stay [23(1–116) vs 8(1–60), p < 0.001], hospital length of stay [30(3–183) vs 15(1–174) ± 7, p < 0.001], steroid days [10(1–116) vs 4(0–28), p = 0.02] and ventilation days [18(0–120) vs 2(0–55), p < 0.001] were significantly higher in the fungal group. The mortality rate in both groups were similar (51% vs 51.6%).

Conclusions: Fungal infections are extremely common in COVID-19 patients admitted to ICU, seen in almost half of patients in this cohort (47%). Longer treatment with corticosteroids appears to increase the risk of developing fungal infections. Increased length of ICU stay, and a greater length of mechanical ventilation significantly increase the risk of fungal infections in COVID-19 patients in intensive care. Fungal infection, however was not associated with increase in mortality.


  1. 1.

    Coşkun AS. et al. Polish J Microbiol 2021;70:395–400.


Candida colonization and mortality in an ICU: is candida colonization a poor prognosis factor in the critical patient?

RV Jorge1, A Carvalho2, A Fernandes3, J Patrício4, P Patrício4, T Nascimento5

1Hospital Beatriz Ângelo, Intensiva Medicine Department, Loures, Portugal, 2Hospital da Sra. da Oliveira, Guimarães, Portugal, 3Hospital Beatriz Ângelo, Internal Medicina Department, Loures, Portugal, 4Hospital Beatriz Ângelo, Loures, Portugal, 5Instituto de Higiene e Medicina Tropical/Universidade Nova de Lisboa(IHMT/UNL)/Instituto Universitário Egas Moniz, Lisboa, Portugal

Critical Care 2023, 27(S1): P161

Introduction: Candida colonization in hospitalized patients may predispose to the establishment of severe infections namely in the critical patient. The aim of this prospective study was to evaluate if the patients admitted in the ICU who are Candida colonized have higher mortality associated rates.

Methods: A total of 497 patients (177 female, 320 male; age range: 12–92 years, mean age: 64) who were in ICU of Beatriz Ângelo Hospital (Lisbon, Portugal) between September, 2021 and September, 2022 were included in the study. After admission and consent, axillary and groin region cultures have been collected in D1, D5 and D8 of staying. Isolation and identification of Candida strains were performed by using conventional mycological methods. Patient records were used to describe and compare the demographic, health indicators and risk factors of the screened. The primary end point was 3, 6 and 12 months all-cause mortality, when available. Multiple logistic regression analysis was then applied.

Results: In our study, yeast growth was detected in 137 patients (27.6%) and C. albicans (72.3%) and C. parapsilosis (43.1%) were the most common. 21.9% were mixed cultures and C. glabrata was isolated in 13.9% of the positive cultures. Our results have indicated a statistically significant relationship between colonization by C. albicans (odds ratio 1.864; 95% CI 1.133–3.068) and C. glabrata (odds ratio 3.627; 95% CI 1.312–10.022) with 3-month mortality. Also, patients with anemia and mechanic ventilation are more likely to be candida colonized. however, the correlation between candida colonization and the presence of other underlying diseases, invasive procedures, parenteral nutrition, haemodialysis, surgery and use of extended-spectrum antibiotics was not statistically significant.

Conclusions: These findings could imply candida colonization as a poor prognosis factor in the critical patient. Further analysis and high-scale studies are needed to support our data.




Characterization of secondary bacterial and fungal infections in ICU hospitalized patients with COVID-19: retrospective observational study

AR Rutins1, EK Kalnins2, SK Kazune3, ZS Smatcenko4

1 iga Stradins University, Faculty of Residency, Riga, Latvia, 2University of Latvia, University of Latvia- Residency Studies, Riga, Latvia, 3Riga Stradins University, Riga, Latvia, 4Riga East University Hospital, Toxicology and Sepsis Department, Riga, Latvia

Critical Care 2023, 27(S1): P163

Introduction: We aimed to describe the incidence, risk factors, and clinical outcomes of bacterial and fungal co-infections and superinfections in intensive care patients with COVID-19 in a retrospective observational study.

Methods: A retrospective cohort of intensive care patients with confirmed SARS-CoV-2 by PCR was analysed from January to March 2021. This was contrasted to a control group of influenza-positive patients admitted during 2012–2022. Patient demographics, microbiology and clinical outcomes were analysed.

Results: A total of 70 patients with confirmed SARS-CoV-2 were included; 6 (8.6%) of 70 had early bacterial isolates identified rising to 42 (60%) of 70 throughout admission. Blood cultures, respiratory samples, and urinary samples were obtained from 66 (94.3%), 18 (25.7%) and 61 (87.1%) COVID-19 patients. Positive blood culture was identified in 13 patients (18.6%). Bacteraemia resulting from respiratory infection was confirmed in 3 cases (all ventilator-associated). Line-related bacteraemia was identified in 9 patients (6 Acinetobacter baumannii, 4 Enterococcus spp. and 1 Pseudomonas aeruginosa and 1 Micrococcus lylae). No concomitant pneumococcal, Legionella or influenza co-infection was detected. Invasive fungal infections with Aspergillus spp. were identified in 2 cases. Pneumococcal coinfections (7/68; 10.3%) were identified in the control group of confirmed influenza infection; clinically relevant bacteraemias (6/68; 8.8%), positive respiratory cultures (15/68; 22.1%). The rate of hospital-acquired infections was 51.4% for COVID-19 and 27.9% for influenza. Longer intensive care stay, type 2 diabetes, obesity and hematologic diseases were independent risk factors for superinfections in the COVID-19 cohort.

Conclusions: Respiratory coinfections occurred in influenza but not in COVID-19 patients. The rate of hospital-acquired infections (51.4% for COVID-19; 27.9% for influenza) was unexpectedly high in both groups.


Efficacy and safety of rezafungin in critically ill patients with candidaemia and/or IC: Outcomes of patients treated in the ICU from an integrated analysis of Phase 2 and Phase 3 trials

P.M. Honoré1, M. Kollef2, G.R. Thompson3, A. Soriano4, O.A. Cornely5, C Dignani6, N Manamley7, T Sandison8, PG Pappas9, M Bassetti10, On behalf of the ReSTORE trial investigators

1UCLouvain Medical School, Intensive Care Unit Department, Brussels, Belgium, 2Washington University, St. Louis, USA, 3University of California Davis Medical Center, Davis, USA, 4Hospital Clínic de Barcelona, IDIBAPS, University of Barcelona, Barcelona, Spain, 5University of Cologne, Cologne, Germany, 6PSI-CRO, Durham, NC, USA, 7Mundipharma Research Ltd, Cambridge, UK, 8Cidara Therapeutics, Inc., San Diego, CA,, USA, 9University of Alabama at Birmingham, Birmingham, AL, USA, 10University of Genoa, Genoa, Italy

Critical Care 2023, 27(S1): P164

Introduction: Critically ill patients are susceptible to invasive candidiasis (IC), with approximately one-third of all candidaemia (C) cases occurring in the intensive care unit (ICU) [1]. This post-hoc, integrated analysis of STRIVE (Phase 2; NCT02734862) and ReSTORE (Phase 3; NCT03667690) investigated the efficacy and safety of rezafungin (RZF), a novel, once-weekly (QW) echinocandin, in a subgroup of patients with C and/or IC in the ICU.

Methods: Patients received RZF QW (400 mg in Week 1, then 200 mg) or caspofungin (CAS) once daily (70 mg on Day [D] 1, then 50 mg with optional oral fluconazole stepdown) for 2‒4 weeks. The primary endpoint (U.S. FDA) was all-cause mortality (ACM) at D30. Other endpoints included mycological eradication at D5 and D14 (secondary) and time to negative blood culture (TTNBC; exploratory). Safety was evaluated by adverse events (AEs).

Results: Baseline patient characteristics were generally similar between the RZF (n = 55) and CAS (n = 71) arms; 32.1% (17) and 29.6% (21) of patients receiving RZF and CAS, respectively had an acute physiologic assessment and chronic health evaluation II score ≥ 20. ACM at D30 was 36.4% (20/55) with RZF and 26.8% (19/71) with CAS. Through Day 30, 4 (7.3%) deaths were attributable to C/IC in the RZF arm and 6 (8.5%) in the CAS arm. Mycological eradication at D5 and D14 and the proportion of patients with negative blood culture at 24 h and 48 h are shown in Table 1. Median TTNBC was 18 h with RZF versus 38 h with CAS (stratified log-rank P = 0.003; not adjusted for multiplicity). Safety profiles were similar between treatment arms; the incidence of treatment-emergent AEs was 90.9% (50/55) with RZF and 81.7% (58/71) with CAS.

Conclusions: RZF is efficacious and well tolerated in critically ill patients with C and/or IC, with high rates of mycological eradication observed early in treatment. This integrated dataset provides additional insights into RZF efficacy and safety in the critical care population.


  1. 1.

    Logan C et al. Intensive Care Med 2020;46:2001‒2014.

Table 1 (abstract P164). Efficacy outcomes in RZF and CAS arms

Efficacy outcomes of patients

RZF (n = 55) 400/200 mg QW

CAS (n = 71) 70/50 mg once daily

Negative blood culture; 24 h, n/N (%)

25/41 (61.0)

22/52 (42.3)

Negative blood culture; 48 h, n/N (%)

30/40 (75.0)

29/52 (55.8)

Mycological eradication; D5, n/N (%)

40/55 (72.7)

41/71 (57.7)

Mycological eradication; D14, n/N (%)

37/55 (67.3)

45/71 (63.4)


Renal safety of liposomal amphotericin B in critically ill mechanically ventilated patients with SARS-CoV-2 pneumonia

Á Estella1, I Blanco1, M Gracia Romero2, A Garrino1, M Recuerda Nuñez1

1University Hospital of Jerez, University of Cadiz, INIBiCA, Intensive Care Unit University Hospital of Jerez, Jerez de la Frontera, Spain, 2University Hospital of Jerez. University of Cadiz, INIBiCA, Intensive Care Unit, Anatomy Department, Jerez de la Frontera, Spain

Critical Care 2023, 27(S1): P165

Introduction: Liposomal amphotericin B (L-AmB) represent a good treatment strategy for critically ill patients according to its unique pharmacological characteristics and a relatively broad spectrum of action. The aim of the present study is to asses the impact on renal function of L-AmB during the first days of ICU admission in critically ill patients.

Methods: Retrospective, single-center case series of patients with SARS-CoV-2 pneumonia admitted in ICU. Setting: 19-bed medical-surgical ICU of a community hospital. Time of study: 2 years. Study variables: APACHE II and SOFA at admission,clinical characteristics, oliguria and creatinine level at admission and 72 h after L-AmB treatment were recorded. Oliguria was defined as urinary output less than 400 ml per day or less than 20 ml per hour. Two groups of patients were selected according to whether or not they received anticipated antifungal treatment pending microbiologic confirmation or discarding of aspergillosis; dosage of L-AmB was 3 mg/kg/d. Statistical analysis: Data were analyzed by SPSS 18 and quantitative variables were expressed as a mean ± standard deviation.

Results: 160 patients were included, 102 who received 3 days of anticipated treatment with L-AmB at ICU admission or at the beginning of mechanical ventilation were compared with patients without this treatment. There were not differences in age, median 65 [57–71] years, gender with 28% female and BMI (kg/m2), 30,4 [26,6–33,2]. APACHE II at admission was higher in the treated group of patients 17 [12–23] vs 12 [9–14]. SOFA was 7 [4–8] in the treated group of patients vs 6 [3–8]. There were not differences in urinary output between groups during the three first days of ICU stay. Table 1 shows creatinine levels.

Conclusions: According to our retrospective analysis, L-AmB is safe in the first days of treatment in critically ill patients admitted in ICU requiring mechanical ventilation.

Table 1 (abstract P165). Creatinine levels according use of liposomal amphotericin B at ICU admission and at third day of treatment


Overall (n = 160)

Not anticipated antitifungal treatment (n:58)

Anticipated antitifungal treatment (n = 102)


Admission to ICU: Creatine (mg/dl)

0.9 [0.7–1.2]

0.9 [0.7–1.2]

0.8 [0.7–1.1]


Third day to admission to ICU Creatine (mg/dl)

0.9 [0.7–1.2]

0.9 [0.7–1.3]

0.9 [0.7–1.2]



The immunological aspects of the efficiency of prevention of sepsis with inhaled nitrogen oxide in newborns on mechanical ventilation

M Pukhtinskaya

State Medical University, Rostov-on-Don, Department of Anesthesiology and Critical Care Medicine, Rostov on Don, Russian Federation

Critical Care 2023, 27(S1): P166

Introduction: It is known, that the molecule NO—is fundamental factor of anti-infectious resistance of an organism. Research objective: evaluate the efficacy and safety of inhalation nitrogen oxide (iNO) to prevent sepsis.

Methods: The controlled, randomized, blind clinical trial included 97 full-term newborns admitted for mechanical ventilation; no clinical signs of bacterial infection were diagnosed. Group I (n = 44) received iNO (10 ppm, 48 h). Group II (n = 53) did not receive iNO. On the 1, 5, 20 days the plasma concentration of IL-1β, IL-6, IL-8, TNF-α, G-CSF, sFasL, FGF, NO was determined by capture ELISA; CD3+CD19, CD3+CD19+, CD3+CD4+, CD3+CD8+, CD3+CD69+, CD3+CD71+, CD3+CD95+, HLA-DR+, CD34+, CD14+, CD3+CD56+, AnnexinV-FITC+PI, AnnexinV-FITC+PI+—immunophenotype analysis. Statistical programs RStudio Desktop; Kaplan–Meier, Fisher-Irwin method bilateral alternative at 5% significance level.

Results: Development of sepsis: I group—4 patients; II—13 (p1=0.04; p2 = 0.005); fatal outcomes: I—6 patients, II—10 (p1 = 0.37; p2 = 0.59). Median of the duration of mechanical ventilation: I—5 days, II—10 (p = 0.00007); hospitalizations: I—11 days, II—15 (p = 0.026). Comparative dynamics of indicators of I group relative to II: a decrease (p < 0.05) the concentration of TNF-α, IL-8, IL-6; CD3+CD69+; CD3+CD95+, lymphocytes in apoptosis; increase (p < 0.05) G-CSF, sFasL, FGF, NO; CD14+, CD3+CD19.

Conclusions: iNO as a part of intensive care decrease in frequency of development of sepsis, the duration of mechanical ventilation, hospitalization; provides a tendency to decrease in frequency of lethal outcomes; reduces cytokine aggression; inhibits apoptosis of lymphocytes, activates the monocyte-macrophage immunity and proliferative processes. It is relevant to continue a research.


Left and right ventricular dysfunction in patients with septic shock in relation to seven day and hospital mortality

S Kim1, YJ Kim2, YH Kim3, JH Kim1

1Korea University Ansan Hospital, Critical Care Medicine, Ansan, South Korea, 2Korea University Ansan Hospital, Pulmonology, Internal Medicine, Ansan, South Korea, 3Korea University Ansan Hospital, Cardiology, Internal Medicine, Ansan, South Korea

Critical Care 2023, 27(S1): P167

Introduction: The prognostic implication of septic cardiomyopathy has not been clearly delivered with the complex hemodynamic change during septic shock. Herein, we evaluated the serial change of left (LV) and right ventricular (RV) function in patients with septic shock and their clinical impact on seven-day and hospital mortality.

Methods: Patients were performed transthoracic echocardiography within 48 h from the diagnosis of septic shock and after seven days from the initial evaluation. LV systolic function was evaluated with global longitudinal strain (GLS) and RV function was evaluated with tricuspid annular posterior systolic excursion (TAPSE).

Results: A total of 162 patients (men 83, 51.5%; 70.7 ± 13.4 years; APACHE II 30.6 ± 9.2) was enrolled. Their baseline GLS was-14.9 ± 5.2% and TAPSE was 16.9 ± 5.5 mm, and they were improved in follow-up study (GLS-17.1 ± 4.5%, TAPSE 19.9 ± 5.0 mm). Seven-day and hospital mortality was 24 (14.9%) and 64 (39.8%). Seven-day mortality was significantly related with GLS > -16% (odds ratio [OR] 5.938, 95% confidence interval [CI] 1.031–34.190, p = 0.046) and APACHE II score (p = 0.004). Among seven-day survivors, hospital mortality of seven-day survivors were related with TAPSE < 16 mm at follow-up evaluation (OR 8.855, 95% CI 1.524–51.446, p = 0.015) and follow-up SOFA score (OR 1.352, 95% CI 1.083–1.687, p = 0.008) but not with follow-up GLS > − 16% (p = 0.506) or initial GLS > − 16% (p = 0.177) (Fig. 1).

Conclusions: LV and RV dysfunction was common in septic shock, but shortly began to improve. Seven-day mortality of septic shock patients was related with GLS, whereas hospital mortality of 7 day survivors were related with TAPSE but not with GLS. However, for the interpretation of the impact of cardiac dysfunction on the prognosis of septicshock, cautious approach is in need considering the stage of septic shock and cut-off values for the parameters.

figure bi

Fig. 1 (abstract P167). Kaplan–Meier curve for the hospital mortality in seven-day survivors, by global longitudinal strain < -16% and ≥ -16% (A) and by tricuspid annular posterior systolic excursion > 16 mm and ≤ 16 mm (B)


Longitudinally monitored hemodynamic responsiveness characteristics in non-invasively monitored shock patients predict hospital mortality

J Sahatjian1, I Douglas2, M Abs1, T Hiller3, D Tran3, A Garcia3, J Oakes3, C Rosenthal3, D Hansell4

1Baxter Healthcare, Deerfield, USA, 2Denver Health Hospital, Critical Care, Denver, USA, 3Denver Health Hospital, Denver, USA, 4Massachusetts General Hospital, Boston, USA

Critical Care 2023, 27(S1): P168

Introduction: Non-invasive hemodynamic monitoring of cardiac output (CO) and stroke volume (SV), including dynamic measures of fluid responsiveness, has the potential to guide fluid and vasopressor (VP) management in patients with circulatory shock. The goal of this study was to explore the variables that may influence SV and CO improvement and mortality in shock patients when monitored during usual care resuscitation.

Methods: The Starling Registry trends CO and SV over time as related to patient outcome (NCT04648293). Patients were managed at a single community academic medical ICU using dynamic assessments of fluid responsiveness through passive leg raise (PLR) to guide the administration of fluid and VP. Patients exhibiting overall improvement in SV and CO (Responders) from initial presentation through hours of monitoring were compared to those who did not improve (Non-Responders).

Results: 31 patients received hemodynamic monitoring during their ICU stay. 45% were female; average age was 58 years. 26% of patients had sepsis; 35% had known LV/RV dysfunction. 16 patients (52%) were classified as responders and received a total fluid volume of 500 ml (± 273.86). 15 patients (48%) were non-responders and received 1232 ml of fluid (± 635.37), (Δ732 ml, p = 0.040). Both groups were matched with regard to age, gender, and co-morbidities. 100% of patients that had VP started in response to fluid challenge exhibited an improved SV. Responders received longer VP administration (4.3 days ± 4.51 vs 1.7 days ± 1.72, p = 0.041). Notably, responders experienced significantly lower mortality (6%) compared with non-responders (33%, p = 0.056).

Conclusions: Routine incorporation of non-invasive hemodynamic monitoring to inform a personalized approach to shock resuscitation yields novel insights regarding circulatory responsiveness to fluid and VP administration. Trending SV improvement identified non-responder patients at high risk for in hospital death, who receive higher fluid volumes but shorter durations of VP administration.


Early versus late administration of intravenous immunoglobulins as adjunctive treatment in patients with septic shock. A retrospective observational study

M Benlabed1, S Benlabed2, R Gaudy3, K Benlabed4, A Ladjouze5, S Aissaoui6

1Lille University, Anesthesiology, Lille, France, 2Free University of Brussels, Intensive Care Unit, Brussels, Belgium, 3Lille University, Emergency Department, Lille, France, 4Algiers University, Pharmacology, Algiers, Algeria, 5Algiers University, Anesthesiology, Algiers, Algeria, 6Algiers University, Intensive Care Unit, Algiers, Algeria

Critical Care 2023, 27(S1): P169

Introduction: The efficacy of intravenous immunoglobulins (IVIG) as adjunctive therapy in sepsis is controversial. Some authors emphasized,in randomized controlled trials, the importance of early administration of IVIG in patients with septic shock.So, the objective of our study was to test, in these patients, the impact of timing of IVIG administration( early versus late) on the outcome in terms of morbidity and mortality.

Methods: In this retrospective observational study,we collected and analyzed between 2018 and 2021,the data of 2 groups of 20 patients in septic shock according to definition of Sepsis 3 criteria,admitted in university hospital ICU..The patients were 62 + -10 years old, mechanically ventilated, with the same baseline charasteristics and severity scores on admission.All the patients received empirical antibiotics and norepinephrine in the first hour of admission in ICU.The patients were divided into 2 groups according to the timing of IVIG administration: the first group receiving IVIG early on the first hour of admission (Early IVIG group) and the second group receiving IVIG 48 h after admission on ICU(Late IVIG group). IVIG was infused continuously at a dose of 0.5 g/kg with a flow of 3 ml/kg/h during 3 days in the 2 groups. We recorded in the 2 groups: SOFA score day 1 and day 4, blood lactate every day, C-reactive protein (CRP), mean arterial pressure (MAP), P/F ratio day 1 and day 3, time on vaspressors, time on mechanical ventilation, ICU stay, hospital length of stay (LOS) and 28 day mortality.

Results: Statistic analysis used Mann Whitney test and results expressed as Mean ± standard deviation. Results are summarized in Fig. 1. We observed a significant reduction of morbidity and 28 day mortality in the Early IVIG group compared to Late IVIG group respectively 20% versus 40%.

Conclusions: Our study confirm previous works [1] showing the beneficial effect of early administration of IVIG in sepsis in terms of morbidity and mortality. A large RCT is needed.


  1. 1.

    Rodriguez A et al. Shock 2005;23: 298–304.

figure bj

Fig. 1 (abstract P169). Variables in relation with outcome in early IVIG and late IVIG groups


Secondary infections in critically ill patients with COVID-19: effects of intravenous IgM-enriched immunoglobulin therapy

E Vitali1, A Salvucci1, A Damia Paciarini2, E Casarotta3, R Antolini1, G Giaccaglia1, C Scorcella2, R Domizi2, E Damiani1, A Donati1

1Università Politecnica delle Marche, Ancona, Italy, 2AOU delle Marche, Ancona, Italy, 3Università Politecnica delle Marche, Biomedical Sciences, Ancona, Italy

Critical Care 2023, 27(S1): P170

Introduction: Critically ill patients with severe COVID-19 have an increased risk of bacterial and fungal superinfections due to a dysregulated immune response characterized by lymphopenia and low immunoglobulins levels. The intravenous immunoglobulins are involved in pathogen/toxin scavenging and inhibition of inflammatory mediators gene transcription with anti-apoptotic effects on immune system cells. This research aimed to describe the effects of intravenous IgM-enriched immunoglobulins in COVID-19 patients with sepsis due to secondary infections and low IgM levels.

Methods: We performed an observational retrospective study, including patients admitted to our intensive care unit (ICU) between March 2020 and February 2021 with severe COVID-19 and sepsis due to a superinfection (known or suspected) treated with intravenous IgM-enriched immunoglobulins. We collected demographic data and comorbidities. We noted hemodinamic data, antimicrobial and adiuvant therapies, laboratory results at ICU admission (T0), at the beginning (T1) and at the end (T2) of the IgM-enriched immunoglobulins infusion and at ICU discharge (T3).

Results: In our cohort of 36 patients (Table 1) the prevalence of documented secondary infections was 83%. We observed a significant reduction of leukocytes from T0 to T3 (10.4 [8.3–14.5] × 103/mmc vs 7.1 [4.8–11.2] × 103/mmc, p < 0.01) and the SOFA score from T0 to T2 (7 [6–19] vs 5 [3–7], p < 0.01) and from T0 to T3 (7 [6–10] vs 4 [2–9], p < 0.01); from T1 to T2 (7 [6–9] vs 5 [3–7], p < 0.01) and from T1 to T3 (7 [6–9] vs 4 [2–9], p < 0.01). Cardiovascular SOFA showed a statistically significant reduction from T1 to T2 (4 [3–4] vs 0 [0–3], p < 0.01).

Conclusions: The IgM-enriched immunoglobulins could improve organ function, as evidenced by the reduction of the SOFA score. Although the latest Surviving Sepsis Campaign guidelines suggest against using of IgM-enriched immunoglobulins, our study supports its use as an adjunctive therapy in COVID-19 patients with septic shock.

Table 1 (abstract P170). Summary of demographic and clinical characteristics at ICU admission


Patients, n (%)


31 (86)

Age, years

64 + 11


19 (53)


17 (47)

Cardiovascular disease

5 (14)


Continuous renal replacement therapy with oXiris filter in critically ill COVID-19 patients

R Bounab1, N Heming1, V Maxime1, E Kuperminc1, M Carlos1, P Moine1, S Chevret2, D Annane1

1Raymond Poincaré Hospital (AP HP), Critical Care Unit, GARCHES, France, 2Saint-Louis Hospital, Biostatistics and Clinical Epidemiology, Paris, France

Critical Care 2023, 27(S1): P171

Introduction: COVID-19 causes a major inflammatory response, which may progress to shock and multiple organ failure. We explored whether continuous renal replacement therapy (CRRT) using adsorption membrane (oXiris®) could reduce the inflammatory response in critically ill COVID-19 patients with acute renal failure (ARF) [1, 2].

Methods: Case–control study including 24 critically ill COVID-2019 patients requiring RRT using an oXiris filter. We measured cytokines before and during treatment as well as relevant clinical endpoints. The control group was selected among COVID-19 patients included into our ongoing RECORDS trial (NCT04280497) who received RRT without adsorbing filters.

Results: 24 severe COVID-19 patients, admitted to the intensive care unit (ICU) and treated with CRRT using the oXiris filter between March and April 2020 (20 males and 4 females); median age 67. The average time from COVID-19 symptoms to initiation of oXiris treatment was 18 ± 7 days, and from ICU admission to initiation of oXiris treatment 9.5 ± 7.8 days and from ARF to oXiris treatment was 3 ± 5 days. The average length of treatment was 152.8 ± 92.3 h. Treatment was associated with cytokine decreases for IL-1beta (p = 0.00022), MCP-1 (p = 0.03), and MIP-1 alpha (p = 0.03). The SOFA scores also showed a reduction over 48 h of therapy without reaching statistical significance. Our study found no significant effect of hemodynamic status. The average ICU stay length was 14 ± 5 days and the mortality rate was 79% in the Oxiris group. We compared the mortality across the two matched groups, there was no evidence of any difference in mortality (Fig. 1).

Conclusions: In our study, CRRT using the oXiris filter seemed to effectively remove IL-1 beta, MCP-1, and MIP-1 alpha in COVID-19 patients. These exploratory results should be confirmed in a randomized controlled study.


  1. 1.

    Broman ME et al. PLoS One 2019;14:e0220444.

  2. 2.

    Stockmann H et al. Crit Care Med 2022;50: 964–976.

figure bk

Fig. 1 (abstract P171). Cumulative incidence of death before (left) and after (right) matching


A novel haemofiltration system to selectively remove IL-6 from the bloodstream: preclinical efficacy and safety in healthy volunteers

N Waalders1, D Van Lier1, L Moran2, M Cheppallan2, D Gold2, W Twiggers2, C Blanco-Andujar2, G Frodsham2, M Kox1, P Pickkers1

1Radboudumc, Department of Intensive Care Medicine, Nijmegen, Netherlands, 2MediSieve Limited, MediSieve Limited, London, UK

Critical Care 2023, 27(S1): P172

Introduction: Extracorporeal blood purification therapies have not yet demonstrated to selectively remove cytokines from the blood or to improve outcome of septic patients. A novel method is represented by the MediSieve Magnetic Haemofiltration System (MMHS), which aims to selectively remove interleukin (IL)-6 from the bloodstream by infusing magnetic beads coupled to anti-IL-6 antibodies into an extracorporeal circulation system containing a filter placed inside a magnet (Fig. 1A). We assessed the efficacy of MMHS to remove IL-6 in a benchtop experiment and its safety in humans.

Methods: Four human plasma samples were spiked with a mixture of several cytokines and passed through the MMHS at a flow rate of 120 ml/min. Magnetic beads coupled with anti-IL-6 antibodies were infused into the system and cytokine concentrations were determined at several timepoints. Safety was evaluated by performing haemoperfusion using the MMHS in six healthy male and female volunteers (both n = 3) for five hours with a flow rate of 120 mL/min. Vital signs were monitored continuously and blood samples were serially obtained.

Results: The MMHS rapidly and selectively removed IL-6 from human plasma, without affecting concentrations of other cytokines (Fig. 1B). No serious adverse events occurred during haemoperfusion in healthy volunteers, who did not experience any symptoms either. Several mild adverse events were observed, including a small haematoma around the canulation site and transient, clinically irrelevant increases in iron, chromium and manganese. Haemofiltration using the MMHS did not cause relevant changes in vital signs (Fig. 1C, D), thrombocytes, leukocytes (Fig. 1E, F), or increases in plasma cytokine concentrations.

Conclusions: The MMHS selectively removes IL-6 from human plasma and is well-tolerated and safe in healthy volunteers. In future studies, we will use the experimental endotoxemia model to assess the efficacy of the MMHS in removing IL-6 from the circulation in humans in vivo.

figure bl

Fig. 1 (abstract P172). A Principle of the MediSieve Magnetic Haemofiltration System (MMHS). Blood passes through an extracorporeal circuit containing the MediSieve Filter placed inside the MediSieve Magnet. Magnetic beads coupled to an antibody directed against a substance or cell type of choice are infused into this extracorporeal circuit before the Filter. Of note, bead infusion was not used in the current human safety study. B Cytokine removal by MMHS in the benchtop experiment. Data are presented as median


Cytosorb hemoperfusion markedly attenuates circulating cytokine concentrations during systemic inflammation in humans in vivo

A Jansen1, N Waalders1, DPT Van Lier1, EN Deliargyris2, J Scheier3, M Kox1, RP Pickkers1

1Radboud University Medical Center, Department of Intensive Care Medicine, Nijmegen, Netherlands, 2CytoSorbents Corporation, Princeton, USA, 3CytoSorbents Europe, Berlin, Germany

Critical Care 2023, 27(S1): P173

Introduction: Although the CytoSorb hemoadsorption device has been demonstrated to be capable of clearing inflammatory cytokines, it has not yet been shown to attenuate plasma cytokine concentrations in humans in vivo. We investigated the effects of CytoSorb hemoperfusion on plasma levels of various cytokines using the repeated human experimental endotoxemia model, a highly standardized and reproducible human in vivo model of systemic inflammation and immunological tolerance induced by administration of bacterial lipopolysaccharide (LPS).

Methods: Twenty-four healthy male volunteers (age 18–35) were intravenously challenged with LPS (a bolus of 1 ng/kg followed by continuous infusion of 0.5 ng/kg/hr for three hours) twice: on day 0 to quantify the initial cytokine response and on day 7 to quantify the degree of endotoxin tolerance. Subjects were randomized to receive either CytoSorb hemoperfusion for 6 h during the first LPS challenge (CytoSorb group), or no intervention (control group). Blood samples were serially obtained to determine cytokine plasma concentrations and clearance rates.

Results: LPS administration led to a profound increase in cytokine plasma concentrations during both LPS challenge days (Fig. 1). Compared to the control group, significantly lower plasma levels of tumor necrosis factor (TNF, −58%, p < 0.0001), interleukin (IL)-6 (−71%, p = 0.003), IL-8 (−48%, p = 0.02) and IL-10 (−26%, p = 0.03) were observed in the CytoSorb group during the first LPS challenge. Peak clearance rates ranged from a median of 75.1 [70.7–87.1] ml/min for TNF, to 32.5 [28.0–44.2] ml/min for IL-6. The degree of endotoxin tolerance apparent during the second LPS challenge was not affected by previous CytoSorb use.

Conclusions: CytoSorb hemoperfusion effectively attenuates circulating cytokine concentrations during systemic inflammation in humans in vivo, whereas it does not affect tolerance induction. Therefore, CytoSorb therapy may be of benefit in conditions characterized by excessive cytokine release.

figure bm

Fig. 1 (abstract P173). Plasma cytokine concentrations during the first (day 0) and second (day 7) LPS challenge. Data displayed as median and interquartile range. P values were computed by repeated measures two-way analysis of variance (time*group interaction term) on log-transformed data


Cytokine hemoadsorption with CytoSorb® in post-cardiac arrest syndrome, a pilot randomized controlled trial

C Monard, N Bianchi, E Poli, M Altarelli, A Debonneville, M Oddo, L Liaudet, A Schneider

Centre Hospitalier Universitaire Vaudois, Service de Médecine Intensive Adulte, Lausanne, Switzerland

Critical Care 2023, 27(S1): P174

Introduction: Hemoadsorption (HA) might mitigate the systemic inflammatory response associated with post-cardiac arrest syndrome (PCAS) and improve outcomes. We investigated the feasibility, safety and efficacy of HA with CytoSorb® in cardiac arrest (CA) survivors at risk of PCAS.

Methods: In this pilot randomized controlled trial, we included patients admitted in ICU following CA, who required norepinephrine (> 0.2 µg/kg/min), and/or had serum lactate > 6 mmol/l and/or a time-to-return of spontaneous circulation (ROSC) > 25 min. Those requiring ECMO or renal replacement therapy were excluded. Eligible patients were randomly allocated to receive standard-of-care (SOC) or SOC plus HA. HA was performed as standalone therapy for 24 h, using CytoSorb® and regional heparin-protamine anticoagulation. We collected feasibility, safety, clinical data and serial plasma cytokines levels over 72 h of randomization.

Results: We enrolled 21 patients, 16 (76%) had out-of-hospital CA and median (IQR) time-to-ROSC was 30 (20, 45) minutes. Eleven received SOC and 10 HA, which was initiated in all 10 patients within 18 (11, 23) hours after ICU admission, and for a duration of 21 (14, 24) hours. The intervention was well tolerated except for a trend for higher rate of aPTT elevation (5 (50%) vs 2 (18%) p = 0.18) and mild (< 150 g/l) thrombocytopenia (5 (50%) vs 2 (18%) p = 0.18). Interleukin (IL)-6 plasma levels at randomization were low (< 100 pg/ml) in 10 (48%) patients and elevated (> 1000 pg/ml) in 6 (29%). Median relative reduction of IL-6 at 48 h was 75% (60, 94) in the HA group versus 5% (-47, 70) in the SOC group (p = 0.06). A similar decrease was observed in all tested cytokines (Table 1).

Conclusions: In CA survivors at risk of PCAS, HA was feasible, safe and was associated with a higher reduction in cytokine plasma levels. Future trials are needed to further define the role of HA after CA. Those studies should include cytokine assessment to enrich study population.

Acknowledgement: These data have been published in full in Crit Care. 2023;27:36.

Table 1 (abstract P174). Relative reduction (%) in cytokine blood levels within 48 h post randomization


Control, n = 10

Hemoadsorption, n = 8

Total, n = 18

p value


5 (− 47, 7)

75 (60, 94)

60 (− 13, 86)



72 (13, 90)

90 (68, 98)

83 (33, 93)



62 (41, 76)

91 (35, 99)

75 (41, 92)


Tumor necrosis factor-α

60 (5, 100)

100 (96, 100)

98 (25, 100)



20 (2, 100)

100 (89, 100)

94 (4, 100)



77 (35, 90)

82 (63, 100)

77 (63, 100)



38 (− 14, 93)

84 (72, 97)

75 (16, 97)


  1. Data are presented as median (interquartile range), and compared between groups with Fisher exact test. Relative reduction = (value at randomization-value at 48 h)/value at randomization. One patient received hemoadsoprtion for < 12 h and was excluded from efficacy analysis. One patient in each group died within 48 h, reduction at 48 h could be evaluated in 10 controls and 8 hemoadsorption patients


Italian retrospective multicentric observational study on the use of CytoSorb® in critically ill paediatric patients (CYTOPED Study)

S Bianzina1, A Tessari2, M Agulli3, G Bregant4, P Lonardi5, S Scollo6, A Moscatelli1, C Serpe7, G Longo8, G Bottari7

1Neonatal and Pediatric Intensive Care Unit, Emergency Department, IRCCS G. Gaslini Hospital, Genoa, Italy, 2Pediatric Intensive Care Unit, Department of Women and Children’s Health, University Hospital of Padua, Padua, Italy, 3Anaesthesiology and Intensive Care, Cardiothoracic and Vascular Department, IRCCS University Hospital of Bologna, Bologna, Italy, 4Pediatric Intensive Care Unit, Children Hospital Burlo Garofalo, Trieste, Italy, 5Pediatric Nephrology Unit, AOU Città della Salute e della Scienza di Torino, Torino, Italy, 6Mediterranean Pediatric Cardiac Center, Taormina – San Vincenzo Hospital, Taormina, Italy, 7Pediatric Emergency Department, Pediatric Intensive Care Unit, Children Hospital Bambino Gesù, IRCCS, Rome, Italy, 8Pediatric Nephrology, Dialysis and Transplant Unit, Department of Pediatrics, University Hospital of Padua, Padua, Italy

Critical Care 2023, 27(S1): P175

Introduction: CytoSorb® is a cartridge for the adsorption of inflammatory mediators, bilirubin, myoglobin and other xenobiotics, directly from whole blood. The clinical experience is widely documented in adults, whereas experience in paediatric settings is currently limited to single case reports or monocentric retrospective studies. In order to be able to collect evidences in larger paediatric population, an Italian multicentre network was founded (CYTOPED Study Group).

Methods: Italian multicentric observational study on the use of CytoSorb® in critically ill paediatric patients. The prospective enrollment by Italian children hospitals is ongoing since February 2021. A retrospective analysis was conducted by Italian centres from January 2018 to February 2021.

Results: 50 patients were enrolled. PELOD-2 score at PICU admission was 7.38 (SD 4.29, risk of mortality 10.60%) and before hemoperfusion 8.72 (SD 4.44, risk of mortality 16.46%). Clinical indications to hemoperfusion were sepsis or septic shock (29 patients), liver failure (4 patients), rhabdomyolysis (4 patients), hearth surgery (3 patients) and other indications (11 patients). CytoSorb® has been applied in CKRT (56% of cases), ECMO (32%), hemoperfusion (6%) and cardiac by-pass (6%) (Fig. 1). Median time of hemoperfusion was 3 days and median number of sorbents used was 3. Anticoagulation in extracorporeal circuits has been managed with heparin (71% of cases) and regional anticoagulation with citrate (29%). 8 adverse events were recorded (hemodynamic instability at hemoperfusion starting, haemodilution, circuit clotting and thrombocytopenia).

Conclusions: Our data demonstrates that CytoSorb® therapy is safe and feasible in children. The advancement of our study and the prospective arm of CYTOPED Study will allow to investigate further aspects of this therapy, such as dose, timing and use of antibiotics in conjunction with extracorporeal therapies.

figure bn

Fig. 1 (abstract P175). Use of CytoSorb, treatment indications and adverse events


CytoSorb for apixaban clearance

A Harris, NJ Lees

Harefield Hospital, Intensive Care Unit, Uxbridge, UK

Critical Care 2023, 27(S1): P176

Introduction: Planned cessation of oral anticoagulation (DOAC) prior to elective surgery is straightforward, however challenges arise when urgent treatment is required [1]. Patients taking DOACs have an increased risk of mortality due to bleeding so it is essential to have effective tools to reduce circulating levels prior to emergency surgery [1]. Use of CytoSorb haemoadsorber (Cytosorbents) [2] is EU approved to remove rivaroxaban during on pump cardiac surgery [3], but only a few case reports exist of its effects on apixaban [4,5].

Methods: A patient presented with decompensated heart failure and acute kidney injury (AKI), with a background of rheumatic heart disease, multiple cardiac surgeries and apixaban use for atrial fibrillation. The patient was admitted to cardiac intensive care unit (ICU) and required renal replacement therapy (RRT). At admission, serum apixaban levels were 204 ng/ml, 40% higher than upper limit of normal and the patient was commenced on RRT. After 4 h, apixaban levels continued to rise so the CytoSorb haemoadsorber was added to the filtration system (Prismaflex).

Results: After 3.5 h, apixaban levels were reduced by 77% and the patient experienced no adverse bleeding; this was essential for their heart transplant assessment (Fig. 1). After removal of the CytoSorb at 14 h, the apixaban remained at safe levels and RRT was stopped 4 days later.

Conclusions: Approximately 27% of apixaban clearance is via the kidneys [6], so AKI can lead to bleeding due to dangerously high levels. This case report shows CytoSorb can be used to effectively reduce apixaban levels in a time efficient manner and is a potential therapy for patients undergoing emergency procedures and surgery.

Acknowledgement: Informed consent was obtained from the patient to publish the clinical details.


  1. 1.

    Bolliger D et al. J Cardiothorac Vasc Anesth 2022;36:1645–7

  2. 2.

    Cytosorbents Europe GMBH. CytoSorb at

  3. 3.

    Hassan K et al. Ann Thorac Surg 2019;108:45–51

  4. 4.

    Hassan K et al. J Clin Med 2022;11:5889

  5. 5.

    Tripathi R et al. Eur Hear Journal Acute Cardiovasc Care 2021;10: zuab020.229

  6. 6.

    Byon W et al. Clin Pharmacokinet 2019;58:1265–79

figure bo

Fig. 1 (abstract P176). Graph showing apixaban levels measured at various intervals during ICU admission


Hydrocortisone treatment in septic mice increases cholesterol availability without improving adrenal function and with exacerbated wasting of lean tissue

L De Bruyn, A Téblick, T Van Oudenhove, S Vander Perre, I Derese, L Pauwels, S Derde, G Van den Berghe, L Langouche

KU Leuven, Department of Cellular and Molecular Medicine, Leuven, Belgium

Critical Care 2023, 27(S1): P177

Introduction: Sepsis-induced hypocholesterolemia is associated with poor outcome but the underlying mechanism is unclear. As it has been postulated that low plasma cholesterol could partly be due to an increased conversion to cortisol, we hypothesized that hydrocortisone (HC) treatment, via reduced de novo adrenal corticosterone (CORT) synthesis, might improve cholesterol availability and as such affect the adrenal gland and skeletal muscle.

Methods: In a catheterized, fluid-resuscitated, antibiotics-treated mouse model of prolonged sepsis (5 d), septic mice received either HC (1.2 mg/d) or placebo, whereas healthy mice served as controls (N = 50). Plasma CORT, cholesterol, and adrenocortical cholesterol ester content were quantified. Steroidogenic capacity of ACTH was evaluated in an adrenal explant study. Total body mass loss was measured to evaluate CORT-induced wasting in addition to ex vivo muscle force, myofiber size and mRNA expression of atrogenes (Trim63, Fbox32, Foxo3) and regeneration markers (Myostatin, Myf5, Myogenin).

Results: Plasma CORT was normalized in HC septic mice, whereas the sepsis-induced reduction in plasma HDL- and LDL-cholesterol, and adrenocortical cholesterol ester content was attenuated as compared with placebo septic mice (p < 0.05), but without improving the blunted ACTH-induced CORT secretion (p = 0.5). Total body mass was median 13% further decreased in HC septic mice (p < 0.01) as compared with placebo septic mice, with no additional effect on sepsis-induced loss in muscle mass, force and myofiber size (p > 0.05). The sepsis-induced rise in muscle atrogenes was not further affected by HC treatment (p > 0.05), whereas Myostatin and Myf5/Myogenin expression was respectively increased (p = 0.0003) and decreased (p < 0.05), as compared with placebo septic mice.

Conclusions: Glucocorticoid treatment partly restored the sepsis-induced hypocholesterolemia without improving adrenal function, and at a cost of exacerbated loss of total body mass and suppressed muscle regeneration mechanisms.


Effects of enrichment strategies on outcome of adrecizumab treatment in septic shock: post-hoc analyses of the phase II AdrenOSS-2 trial

DPT van Lier1, A Picod2, G Marx3, PF Laterre4, O Hartmann5, F Azibani6, J Struck7, K Santos8, A Mebazaa2, P Pickkers1

1Radboudumc, Intensive Care, Nijmegen, Netherlands, 2hôpital Lariboisière, Anesthésie-réanimation, Paris, France, 3Universitätsklinikum Aachen, Klinik fur Operative Intensivmedizin und Intermediate Care, Aachen, Germany, 4Cliniques Universitaires Saint-Luc, Unité de Soins Intensifs, Bruxelles, Belgium, 5Sphingotec GmbH, Hennigsdorf, Germany, 6Hôpital Lariboisière, 2. Anesthésie-réanimation, Paris, France, 7Adrenomed AG, Hennigsdorf, Germany, 84TEEN4 Pharmaceuticals GmbH, Hennigsdorf, Germany

Critical Care 2023, 27(S1): P178

Introduction: Adrecizumab, a non-neutralizing antibody of adrenomedullin (ADM) was recently investigated regarding its potential to restore endothelial barrier function in septic shock patients. Circulating dipeptidyl peptidase 3 (cDPP3), a protease involved in the degradation of several cardiovascular mediators, represents another biological pathway associated with septic shock outcome, although unrelated to ADM. Therefore, the prognosis of patients with elevated cDPP3 may not be influenced by adrecizumab. The aim of our study was to evaluate the effects of cDPP3-based population enrichment, as well as time until initiation of treatment, on the treatment efficacy of adrecizumab.

Methods: Post-hoc analysis of AdrenOSS-2, a phase-II, double-blind, randomized, placebo-controlled biomarker-guided trial of adrecizumab.

Results: Compared to the total study cohort (HR for 28-day mortality of 0.84 (95% CI 0.53; 1.31), p = 0.439), therapeutic benefit of adrecizumab tended to be more pronounced in the subgroup of 249 patients with low cDPP3 (< 50 ng/ml); HR of 0.61 (95% CI 0.34; 1.08), p = 0.085) (Fig. 1A/B). Median duration to study drug infusion was 8.5 h. In the subgroup of 129 patients with cDPP3 < 50 ng/ml and an early start of treatment (< 8.5 h after septic shock diagnosis) HR for 28-day mortality versus placebo was 0.49 (95% CI 0.21–1.18), p = 0.094 (Fig. 1C). In multivariate interaction analyses corrected for baseline disease severity, both cDPP3, as well as the cDPP3 * treatment interaction term were associated with a reduced HR for 28-day mortality in the Adrecizumab treated group; p = 0.015 for cDPP3 in univariate analysis, p = 0.025 for the interaction term between cDPP3 and treatment group. In contrast, treatment timing was not significantly associated with 28-day mortality in multivariate interaction analyses.

Conclusions: In septic shock patients, a further post-hoc enrichment strategy based on cDPP3 may indicate that therapeutic efficacy is most pronounced in patients with lower cDPP3 levels.

figure bp

Fig. 1 (abstract P178). Kaplan–Meier of 28-day mortality after adrecizumab/placebo infusion, based on cDPP3 and time until initiation of treatment stratification. Data is displayed for the original ITT analysis (panel A), the cDPP3 < 50 ng/mL subgroup (panel B) and the combined stratification (cDPP3 < 50 ng/mL and time until initiation of treatment < 8.5 h) subgroup (panel C). Both adrecizumab dosing groups were combined. Crosses represent censored patients


Ilofotase alfa exerts renal protective effects in patients with sepsis-associated acute kidney injury

K Liu1, K Doi2, R Bellomo3, M Kraan4, P Pickkers5

1University of California, Division of Nephrology and Critical Care Medicine, San Francisco, USA, 2University of Tokyo Hospital, Emergency and Critical Care Medicine, Tokyo, Japan, 3Austin Health, Intensive Care Research, Melbourne, Australia, 4AM-Pharma B.V., Utrecht, Netherlands, 5Radboudumc, Intensive Care, Nijmegen, Netherlands

Critical Care 2023, 27(S1): P179

Introduction: Sepsis-associated AKI (SA-AKI) is associated with mortality and long-term sequelae, but no pharmacological options are available. Ilofotase alfa is a human recombinant alkaline phosphatase with renoprotective effects that showed improved survival in SA-AKI patients in a Phase 2 study [1]. ‘REVIVAL’, a large global phase 3 trial (NCT04411472), was conducted to confirm the efficacy and safety of Ilofotase alfa in patients with SA-AKI.

Methods: REVIVAL was conducted in SA-AKI patients (on vasopressor and < 48 h of AKI, according to KDIGO definition) recruited from 120 sites in Europe, North-America, Australia, New Zealand, and Japan. The primary endpoint was 28-day all-cause mortality. The key secondary endpoint was Major Adverse Kidney Events on day 90 (MAKE90). A pre-unblinding modified MAKE90 was defined as death up to Day 90, RRT through Day 28 and on Day 90, > 25% drop in eGFR at Day 90, rehospitalization to Day 90.

Results: The trial stopped for futility after 649 inclusions (placebo n = 319; IA n = 330) following a planned interim analysis with an observed 28-day mortality of 27.9% on both arms. No safety concerns were identified. A beneficial effect on the modified MAKE90 was observed (p = 0.031) (Fig. 1), mainly driven by reduction of the number of patients receiving renal replacement therapy (RRT) (28.2% Ilofotase alfa vs 36.4% placebo). This effect was most pronounced in patients with pre-existent renal impairment.

Conclusions: Ilofotase alfa exerts renoprotective effects in SA-AKI. Further research is required to confirm a role for treatment of SA-AKI in patients with pre-existent renal impairment.


  1. 1.

    Pickkers et al. JAMA 2018;320:1998–2009.

figure bq

Fig. 1 (abstract P179). Proportion of patients without a MAKE90 at day 90. p value for the probability difference of ilofotase alfa compared to placebo 0.031. All patient events prior to Day 90 are censored/counted on the date of the event, all patient events following Day 90 are marked on Day 90


Correlation of preoperative albumin value and EVLW shift followed by LUS in patients after esophagectomy

A Šimunić Forić1, M Ilić2, L Štefančić1, A Peršin Beraković1, H Feljan3, M Karaman Ilić4

1Radiochirurgia Zagreb, Anesthesiology and ICU, Sveta Nedelja, Croatia, 2UVA Department of Psychological Methods and Statics, Amsterdam, Netherlands, 3Radiochirurgia Zagreb, Thoracic Surgery, Sveta Nedelja, Croatia, 4Radiochirurgia Zagreb, University of Osijek, School of Dental Medicine, Anesthesiology and ICU, Zagreb, Croatia

Critical Care 2023, 27(S1): P180

Introduction: Pulmonary complications, occurs in 16–67% of post esophagectomy patients. The most frequent is pulmonary edema. Malnutrition is the cause of preoperative low albumin value. The aim of our study was to investigate whether there is a correlation between preoperative albumin values and new-onset pulmonary edema.

Methods: We retrospectively analyzed one year data that included: age, gender, BMI, preoperative albumin, value, intraoperative volume, B-line score (BLS) and lactate value 24 h and 48 h after admission in ICU. For BLS Blue protocol was followed.

Results: The average age of the patients was 61.12 years ± 8.41(for a men 59.31 years ± 7.81 and for the women 69 years ± 8.1). There was no difference in the age of the subjects (p = 0.18). The patients received intraoperativly an average of 8.44 ml/kg/h of volume of which a restrictive approach was used in 5 patients. The highest infused volume was 14.08 ml/kg/h. BLS was independent of intraoperativly infused volume in both measurements (p = 0.823, vrs p = 0.113).Preoperative albumin values were in range between 25 and 47 g/l with an average of 35.9 ± 6.62 g/l. BLSs were between ≤ 5 and ≥ 15. Four patients had BLS > 15 24 h versus two 48 h after admission (Fig. 1). Our data showed that BLS and preoperative albumins are independent values, in both measurements (p = 0.972 vs p = 0.6). Postoperative lactate values were independent of preoperative albumin level in both measurements (p = 0.546 vs p = 0.334).

Conclusions: We have not found evidence to support our hypothesis that preoperative albumin values are in correlation with new-onset pulmonary edema. Despite that, we note an increase in EVLW. Factors that could have contributed to a null result are small sample size and type of sugical procedure that could itself result in an increase of BLS. Further studies need to be done on this matter, as corroborating evidence for this topic.

figure br

Fig. 1 (abstract P180). BLS 24 and 48 h after admission in ICU. ID = patient identification


Colloid administration for cytoreductive surgery plus hyperthermic intraperitoneal chemotherapy (CRS-HIPEC), comparative clinical outcomes from a historical cohort

G Madrid, JD Guerra, JP Caceres

University Hospital Fundacion Santa Fe de Bogota, Bogotá, Colombia

Critical Care 2023, 27(S1): P181

Introduction: Colloids have been associated with increased unfavorable outcomes in patients who underwent major surgery [1]. Their safety profile is under scrutiny due to increased risk of renal failure and allergic reactions, as albumin remains as the mainstay of colloid therapy. We aimed to explore associations between types of colloids administered in CRS-HIPEC and postoperative outcomes.

Methods: A cohort of patients taken to CRS-HIPEC from 2007 to 2020 at a university hospital were analyzed retrospectively. We compared outcomes for two groups: albumin only, and hyroxyethyl starch (HES, Voluven 6% ®) or oxypolygelatine (OPG. Gelifundol 5.5%). HES/OPG were used at our institution from 2007 to 2013. Outcomes were ICU length of stay (LOS), total LOS, ICU readmission, inpatient mortality, time for extubation, need for tracheostomy, time to ambulation and oral feeds, need for transfusion, estimated blood loss, intraoperative crystalloids, and units of red blood cells and plasma transfused in surgery.

Results: 148 patients were included. 98 (60%) patients received albumin, and 50 (30%) HES/OPG. ICU (median 3 [IQR 1], vs 6 [IQR 6], p < 0.01), hospital LOS (16 [IQR 13] vs. 22.5 [IQR 22], p < 0.01), and days for extubation (mean 0.7 [SD 0.78] vs 7.5 [SD 22], p < 0.01) were significantly shorter in patients who received albumin. 4 patients in the HES/OPG group received a tracheostomy compared to 0 in the other group (p = 0.01) (Table 1). Patients who received albumin had less crystalloid and blood product requirements, bled less, and were able to feed and ambulate earlier (p < 0.05).

Conclusions: Patients who received intraoperative albumin during CRS-HIPEC and were admitted to ICU had shorter LOS, days to extubation, and less blood product and fluid requirements when compared to patients who received other types of colloids.


  1. 1.

    Elgendy et al. Indian J Anaesth 2019;63:805–13.

Table 1 (abstract P181). Correlations with Mann–Whitney test and p < 0.05






Mean (SD)

Median (IQR)

Mean (SD)

Median (IQR)

ICU stay—days

3.2 (2.14)

3 (1)

11.2 (20.3)

6 (5)

Days on vent

0.7 (0.78)

1 (1)

7.5 (22)

1 (1)

PRBC units transfused

1.5 (1.7)

1 (2)

5.7 (4.8)

4 (6)

Plasma units transfused

0.5 (1.6)

0 (0)

3.8 (5.3)

1.5 (6)


9080 (4453)

8000 (3800)

14,182 (12,000)

12,000 (9500)

Estimated blood loss—ml

1125 (1025)

800 (100)

2885 (1837)

2000 (2400)


Plasma compared to crystalloid resuscitation does not reduce endothelial and organ injury in a pneumosepsis rat model

DP Van den brink1, DJB Kleinveld1, J Vos 1, JJTH Roelofs2, NC Weber1, JD Van Buul3, NP Juffermans1

1Amsterdam UMC, AMC, Laboratory of Experimental Intensive Care and Anesthesiology, Amsterdam, Netherlands, 2Amsterdam UMC, AMC, Department of Pathology, Amsterdam, Netherlands, 3Sanquin Research and Landsteiner Laboratory, Molecular Cell Biology Laboratory at Department of Molecular Hematology, Amsterdam, Netherlands

Critical Care 2023, 27(S1): P182

Introduction: Endothelial permeability is a hallmark of septic shock leading to tissue edema and organ failure. Primary treatment of sepsis patients consists of resuscitation with crystalloids. However, crystalloids aggravate endothelial permeability, which may be related to low protein content. We hypothesized that resuscitation with plasma compared to crystalloids limits endothelial and organ injury in a pneumosepsis rat model.

Methods: Male Sprague–Dawley rats (n = 11) were intratracheally inoculated with 5 × 108 Streptococcus pneumoniae and randomized to receive no resuscitation (NR) or resuscitation with a fixed volume (8 ml/kg) of either Ringer’s lactate (RL) or rat derived fresh frozen plasma (rFFP). Healthy controls received sham inoculation. Five hours after start of resuscitation, animals were infused with FITC-dextran and sacrificed. Pulmonary FITC-dextran leakage and plasma levels of thrombomodulin and syndecan-1 were measured as markers of endothelial injury. To assess organ injury, pulmonary wet/dry ratios and histopathology injury scores were measured.

Results: Pneumosepsis increased median lactate levels (2.77 vs 1.56 mmol/l, p = 0.001), lung wet/dry ratios (5.6 vs 4.3, p < 0.01) and pulmonary FITC leakage (134 vs 52 µg/mL, p = 0.01) compared to healthy controls (Fig. 1 (abstract P182)). rFFP did not reduce median pulmonary FITC-leakage (58 vs 87 µg/ml, p = 0.39), thrombomodulin (p = 0.70) or syndecan-1 levels (p = 0.98), or pulmonary wet/dry ratios (p = 0.16) compared to animals receiving RL (Fig. 1). rFFP did significantly reduce pulmonary FITC leakage compared to animals receiving no resuscitation (p =  < 0.05).

Conclusions: Plasma resuscitation did not limit endothelial and organ injury compared to crystalloids in a pneumosepsis rat model.

figure bs

Fig. 1 (abstract P182). The effects of rat fresh frozen plasma (rFFP), Ringer’s lactate (RL) and no resuscitation (NR) on endothelial and organ injury


Rapid infusion of Ringer’s lactate solution at different temperatures and the effects on circulation in healthy volunteers: a randomized crossover trial

PB Biesenbach1, M Bitten Mølmer2, E Lobner Svendsen2, G Berg-Beckhoff3, D Teichmann4, LE Lilholm Laugesen1, M Rahbek Kristensen1, M Brabrand2

1Esbjerg Hospital, Department of Emergency Medicine, Esbjerg, Denmark, 2Odense University Hospital, Department of Emergency Medicine, Odense, Denmark, 3University of Southern Denmark, Unit for Health Promotion Research, Odense, Denmark, 4University of Southern Denmark, Department of Health Informatics and Technology, Odense, Denmark

Critical Care 2023, 27(S1): P183

Introduction: Intravenous fluids are routinely administered to treat hypotension. However, increases to blood pressure and other hemodynamic changes are limited in duration and scope. This study investigates if changes to mean arterial pressure (MAP) after fluid bolus (FB) are altered by fluid temperature.

Methods: Eighteen healthy volunteers are randomly allocated to 30 ml/kg Ringer’s lactate at cold (15 °C) or body (37 °C, warm) temperature over 30 min. For 2 h after starting the FB, we measured MAP, systolic blood pressure, diastolic blood pressure and pulse rate continuously. We recorded temperature every 15 min. In a second session, volunteers crossed over.

Results: 15 min after the FB was administered, mean MAP increased more with cold fluids (+ 6.5 mmHg [95% CI 4.8–8.2]) vs warm fluids (+ 0.6 mmHg [95% CI −1.6 to 2.8]; P < 0.001) (Fig. 1). Furthermore, a minimum increase in MAP of ≥ 5 mmHg was observed over longer periods of time after cold FB (51 min [95% CI 38–65]) vs warm FB (11 min [95% CI 0–21]; p < 0.001). Pulse rate decreased similarly in both groups (−8.8 beats/min [95% CI −4.8 to −12.8] versus −5.4 beats/min [95% CI −1.7 to −9.2]; p = 0.35). The temperature decreased more drastically after cold FB (−0.3°Celsius [95% CI −0.1 to −0.4]) vs warm FB (−0.1°C [95% CI 0 to −0.2], p = 0.02).

Conclusions: Intravenous FB at cold temperature leads to a greater increase in MAP compared with body temperature, and the effect is more prolonged. These findings imply that in healthy volunteers, when a cold temperature FB is given, the temperature of the fluid rather than its volume accounts for most of the MAP increase.

figure bt

Fig. 1 (abstract P183). Mean (standard error of the mean) of MAP (mean arterial pressure) during the 30 min of fluid bolus administration and observational period in subjects receiving a warm (dotted line) and cold (solid line) fluid bolus, respectively


Effect of isotonic sodium bicarbonate infusion on perioperative acid–base status among patients undergoing emergency laparotomy surgery for acute abdomen: a randomised controlled trial

A Singh1, RK Chaudhary1, NB Naik1, YR Sakaray2

1Post Graduate Institute of Medical Education & Research Chandigarh, Department of Anaesthesia & Intensive Care, Chandigarh, India, 2Post Graduate Institute of Medical Education & Research Chandigarh, Department of General Surgery, Chandigarh, India

Critical Care 2023, 27(S1): P184

Introduction: Acid–Base disorders are common among surgical patients undergoing emergency laparotomy. Normal anion gap metabolic acidosis develops due to direct loss of sodium bicarbonate from GIT or administration of chloride rich solutions during resuscitation in these patients.In order to treat this, administering sodium bicarbonate makes physiological sense.

Methods: After ethical approval and written informed consent, 90 patients (ages 18 to 60 year) diagnosed with perforated peritonitis and required emergency laparatomy surgery were randomaly allocated in two groups 45 each. Group ISB (isotonic sodium bicarbonate group) received 2 ml/kg/hr 1.3% isotonic bicarbonate and group BSS (balanced salt solution group) received 2 ml/kg/hr Ringer lactate (RL) as maintenance fluid during the intraoperative period. Primary outcome was to compare the difference in base excess (BE) between the two groups at the end of surgery. Secondary outcomes (pH, pCO2, HCO3, lactate levels, postoperative duration of mechanical ventilation, ionotrope requirement, ICU/HDU stay and incidence of acute kidney injury) were observed and compared at different time intervals between the groups.

Results: The BE, median (Q1, Q3) was −7.30(−8.50, −6.30) in the ISS and −4.80(−6.80, −4.10) in the BSS group (p = 0.017) (Fig. 1). The median (Q1, Q3) pH, HCO3 and lactate values [7.3 (7.27, 7.32) vs 7.33 (7.31, 7.35); 18.1 (17.3, 19.3) vs 19.0 (18.1, 20.2); 1.6 (1.3, 2.1) vs 1.3 (1.1, 1.89)] were found significantly different at the end of surgery between group ISS and BSS respectively. The difference in duration of mechanical ventilation, ionotrope requirement and ICU/HDU stay were also found significant between the groups.

Conclusions: Intraoperative administration of isotonic sodium bicarbonate infusion improves perioperative acis-base status among patients undergoing emergency laparotomy.

figure bu

Fig. 1 (abstract P184). The BE values at the end of the surgery among the Study and the Control group


Investigating the incidence of hypophosphatemia after elective on-pump cardiac surgery and the response to phosphate substitution: a retrospective single centre cohort study

T Vansteenkiste1, U Janssens2, S Van de Velde1, W Haerens3, J Heerman1

1AZ Maria Middelares, Intensive Care, Gent, Belgium, 2Vrije Universiteit Brussel, Anesthesiology, Brussel, Belgium, 3AZ Maria Middelares, Data Science, Gent, Belgium

Critical Care 2023, 27(S1): P185

Introduction: The aim of this study was to investigate the incidence of hypophosphatemia after elective on-pump cardiac surgery and the response to phosphate substitution.

Methods: We conducted a retrospective single centre cohort study at AZ Maria Middelares among patients who underwent elective on-pump cardiac surgery from January 1st 2019 until January 1st 2022. Inclusion criteria were: age ≥ 18 years; serum phosphate concentrations available from the day of surgery (pre- and postoperative) and the first postoperative day. Hypophosphatemia was defined as mild (0.64–0.8 mmol/l), moderate (0.32–0.63 mmol/l) or severe (< 0.32 mmol/l). Exclusion criteria were: chronic kidney disease (eGFR < 45 ml/min) and preoperative hypo- or hyperphosphatemia. The primary endpoint was hypophosphatemia immediately after surgery. The secondary endpoint, which was analysed in the subgroup of patients with hypophosphatemia, was persistent hypophosphatemia on day 1.

Results: The incidence of immediate postoperative hypophosphatemia was 62.0% (623/1005), of which most was mild or moderate. Patients with hypophosphatemia immediately after surgery were divided into two groups: phosphate substitution vs. no substitution (Fig. 1). The incidence of persistent hypophosphatemia on day 1 in both groups was low and did not significantly differ from each other (p = 0.1966; Fisher’s Exact Test; two tailed test; α = 0.05).

Conclusions: Hypophosphatemia is common after on-pump cardiac surgery (62.0%). In contrast, persistent hypophosphatemia on the first postoperative day was uncommon (4.0%) despite only 19.3% of patients receiving substitution. In hypophosphatemic patients, substitution of phosphate did not significantly lower the incidence of persistent hypophosphatemia in comparison to those who did not receive substitution.

Acknowledgement: T Vansteenkiste and U Janssens contributed equally as first authors.

figure bv

Fig. 1 (abstract P186). Results


Hypokalemia and hypophosphatemia in the ICU: a retrospective analysis of incidence and correction in a mixed ICU population

KF Bachmann1, B Hess2, M Mändul3, A Bloch2, A Regli4, A Reintam Blaser5

1Inselspital, Bern University Hospital, University of Bern, Department of Intensive Care Medicine, Bern, Switzerland, 2Lucerne Cantonal Hospital, Department of Intensive Care, Lucerne, Switzerland, 3Institute of Genomics, Estonian Genome Center, Tartu, Estonia, 4Fiona Stanley Hospital, Department of Intensive Care, Perth, Australia, 5University of Tartu, Department of Anaesthesiology and Intensive Care, Tartu, Estonia

Critical Care 2023, 27(S1): P186

Introduction: This retrospective analysis aims to describe the prevalence of hypokalemia (HypoK+) and hypophosphatemia (HypoPO43−) and to determine cut-offs for overcorrection.

Methods: Consecutive patients admitted to a mixed ICU from November 2019 to December 2020 were included for up to a maximum of 96 h. Data on electrolyte administration were collected from the minimum below threshold (hypo) up to the end of the study period or, if correction or overcorrection occurred, up to the first value above threshold if substitution occurred. AUC-ROC curves with an index of union were used to define cut-off points between normalization and overcorrection.

Results: 2056 patients were included. Median time of analysis was 5.6 [2.7–11.9] hours for K+ and 15.1 [8.4–24.2] hours for PO43−. 365 (17.8%) patients had HypoK+, of which 38 (10.4%) remained with hypoK+, 238 (65.2%) normalized values and 89 (24.4%) developed hyperkalaemia. 575 (30.0%) patients had HypoPO43−, of which 276 (48.0%) remained hypo, 249 (43.3%) normalized PO43−and 50 (8.7%) developed hyperphosphataemia. 76 (K+ group) and 258 patients (PO43− group) had no substitution. Of these, 70 (92.1%, K+) and 124 (48.1%, PO43−) normalized spontaneously. The remaining patients receiving substitution are presented in Fig. 1. For HypoK+ we identified a cut-off for overcorrection of 30 mmol for all grades (AUC-ROC 0.773 [95% CI 0.692–0.821]), 22 mmol for grade 1 (0.791 [0.704–0.852]) and 35 mmol for grade 2 (0.765 [0.612–0.871]). For HypoPO43− the cut-offs were 45 mmol for all grades (0.753 [0.67–0.816]), 29 mmol for grade 1 (0.718 [0.588–0.822]) and 55 mmol for grade 2 (0.761 [0.629–0.869]).

Conclusions: This study shows that HypoK+ and HypoPO43− are common and were mostly treated successfully, but overcorrection occurred. Cut-offs to reduce overcorrection may be defined using this retrospective dataset.

figure bw

Fig. 1 (abstract P186). Administration of K + and PO43- during the time of analysis, grouped by grade (HypoK + (A) and HypoPO43- (B)) and overcorrection


Detailed analysis of the time course of metabolic derangements in patients admitted to the ICU with diabetic ketoacidosis

MW Van Hasselt1, JE Kootstra2, MWN Nijsten1

1University Medical Center Groningen, Critical Care, Groningen, Netherlands, 2University Medical Center Groningen, Laboratory Medicine, Groningen, Netherlands

Critical Care 2023, 27(S1): P187

Introduction: Current guidelines advise the use of isotonic saline for resuscitation of patients with diabetic ketoacidosis (DKA). Recent evidence shows this may be associated with (iatrogenic) hyperchloremia [1, 2].

Methods: We conducted a single center, retrospective analysis of patients treated for DKA between 2000 and 2021. Selection criteria were plasma glucose > 20 mmol/l and pH < 7.1. Standard of care consisted of fluid resuscitation with NaCl 0.9% and titrated insulin suppletion. Blood glucose measurements were done by full blood gas analysis. Outcome variables were glucose, pH, sodium, chloride and anion gap (calculated as [Na+]–[Cl]–[bic]). Data were analyzed with SPSS v28.

Results: We identified 55 patients who met the inclusion criteria. There were 1081 blood gas samples in the first three days after admission. Forty-two patients (76%) developed hyperchloremia > 107 mmol/l. Within 8 h after initiation of treatment the mean ± SD chloride levels increased from 93 ± 12 to 107 ± 11 mmol/l (p < 0.001) with a maximal mean of 116 ± 8 mmol/l at 22 h (Fig. 1). In contrast, sodium levels only increased by 8 mmol/l over the same period (p < 0.001). After 17 h the pH had risen above 7.30, and after 16 h the mean glucose level fell below 14 mmol/l. The time to resolution of the DKA, defined as pH > 7.3, bicarbonate > 15 mmol/l and glucose < 14 mmol/l, was approximately 20 h. The anion gap was restored after 20 h.

Conclusions: Our detailed retrospective analysis shows that the time to normalization of metabolic derangements is similar to those reported in recent literature. However the use of NaCl 0.9% infusion was associated with a high prevalence of iatrogenic hyperchloremia.


  1. 1.

    Carillo AR et al. Ann Pharmacother 2022;56:998–1006.

  2. 2.

    Van Zyl DG et al. QJM 2012;105:337–43.

figure bx

Fig. 1 (abstract P187). Time courses of key parameters in 55 ICU patients after treatment for diabetic ketoacidosis (mean ± standard error of mean)


Application of accelerometers for identifying movement related artefacts in general intensive care physiological data

I Piper1, K Sandall1, AB Docherty2

1University of Edinburgh, Usher Informatics Insitute, Edinburgh BioQuarter, Edinburgh, UK, 2University of Edinburgh, Centre for Medical Informatics, The Usher Institute, University of Edinburgh, Senior Clinical Fellow and Consultant in Critical Care, Edinburgh BioQuarter, Edinburgh, UK

Critical Care 2023, 27(S1): P188

Introduction: Analysis and interpretation of high frequency physiological data from the bedside monitors is limited by artefact due to patient movement [1]. Addition of real-time accelerometery may identify movement related artefact and improve quantification of adverse events.

Methods: We recruited critically ill patients receiving invasive mechanical ventilation in the Royal Infirmary Edinburgh, Scotland. We placed an accelerometer (Active Insights) on the patient’s chest and recorded high frequency 3-lead electrocardiogram (ECG), invasive arterial blood pressure (BP) and central venous pressure (CVP) waveform resolution data together with accurate annotation of clinical events. Waveform data and annotations were processed and quantified for events that affected a) both accelerometer and physiological signals, b) physiological signals only, c) accelerometer signals only and d) neither type of signal. Time-series and frequency domain features were derived from the raw signals and both mulitinomial regression and random forest machine learning methods were used to study the relationship between clinical events and physiological movement artefact.

Results: 72 h of monitoring data from seven patients were collected. There were 152 clinical events annotated. Tracheal suction (n = 31), patient turns (n = 11) and blood sampling (n = 11) were most common. Addition of accelerometer derived features to models compared with physiological derived features alone significantly improved model fit (p < 0.001) to the event type. Time-series measures of accelerometer and BP variance and the frequency domain feature ECG to BP coherence (Fig. 1) were the most important (p < 0.001) features explaining the event type categorisation.

Conclusions: Accelerometer data enables identification of movement artefact in physiological waveform data. This may enable more accurate analysis of physiological adverse events and their impact on patient outcomes.


  1. 1.

    Donald et al. J Clin Monit Comput. 2019;33:39–51.

figure by

Fig. 1 (abstract P188). Distribution of random forest minimal depth of tree branch by variable (black bar is mean depth). Lower numbers indicate higher importance of variable for categorisation


Polydistrectual resistance index evaluation is an assessment of vascular compliance in patients with septic shock treated with vasopressin

A Barile1, A Recchia2, G Paternoster3, M Copetti4, L Mirabella5, G Cinnella5, A Del Gaudio2

1University of Foggia, Departement of Medical and Surgical Servic,Intensive Care Unit, FOGGIA, Italy, 2IRCCS “Casa Sollievo della Sofferenza“, Anesthesia and Intensive Care 2, San Giovanni Rotondo, Italy, 3Cardiovascular Anaesthesia and ICU, Potenza, Italy, 4IRCCS “Casa Sollievo della Sofferenza“, Unit of Biostatistics, San Giovanni Rotondo, Italy, 5University of Foggia, Department of Anesthesia and Intensive Care, FOGGIA, Italy

Critical Care 2023, 27(S1): P189

Introduction: Surviving Sepsis Campaign recomends using norepinephrine (NE) as the first-line vasopressor to restore mean arterial pressure [1]. If mean arterial pressure remains inadeguate SSC suggests adding vasopressin (VA) [2]. Resistance Index (RI) is a power Doppler ultrasound assessment of vascular compliance to detect organ perfusion.

Methods: Aim of this study is to compare RI in septic shock patients treated with NE (Group1), NE plus VA since the beginning of vasopressor therapy (Group2) and VA plus NE where VA is added if NE dosage was 20 mcg/min (Group3). RI were measured in renal artery (ARE), radial artery (AR), central retinal artery (CRA),superior mesenteric artery (AMS) at three different time points (T0) before vasopressor therapy, (T1) at 1 h, T2 at 24 h and T3 at 48 h.

Results: 48 patients were divided into three groups. 17 patients Group 1; 16 Group 2, 15 Group 3. In Group 1 RI increased from T0 in CRA R[0.90(0.57–1.12)] and ARE L [0.74(0.56–0.92) to T3 in CRA R[0.97(0.97–1.14)] and ARE L [0.96(0.82–1.17)]. In Group 2 RI reduced in AMS, from T0[0.84(0.70,1.02)] to T3[0.75(0.59,0.81)],in CRA R, from T0[0.90(0.57,1.09] to T3[0.79(0.58, 0.87)], in CRA L, from T0[0.91(0.43,1.53)] to T3[0.76(0.58, 0.89]and in ARE L, from T0[0.79(0.58, 0.92)] to T3[0.72(0.59, 0.83)]. In Group 3 RI reduced in AMS, from t0[0.86(0.71,0.93)] to T3 [0.68(0.64,0.81)], in CRA R, from T0 [0.90(0.75,1.12)] to T3 [0.78(0.66,0.88)],in CRA L,from T0[0.96(0.76,1.33)] to T3 [0.96(0.76,1.33)], in ARE L, from T0[0.77(0.66, 0.99)] to T3[0.67(0.61,0.85], in ARE R, from T0[0.82[0.64, 0.90]]to T3[0.70(0.62,0.82)] and in AR R, from T0[1,10(0.81,1.30)] to T3 [0.87(0.64,1.22)].

Conclusions: Resistence Index was significantly reduced in patients treated with early synergic administration of NE and VA (Fig. 1). This strategy optimized multiorgan perfusion.


  1. 1.

    Evans L. Intesnive Care Med 2021;47:1181–1247.

  2. 2.

    Sacha GL. Crit Care Med. 2022;50:614–623.

figure bz

Fig. 1 (abstract P189). Correlations between all the variables analyzed at T0, T2 and T3 (a,b,c). Intensity of Red indicates inverse correlations, Blue indicates direct. There is a greater direct correlation at T3 between the dose of norepinephrine and the Resistance Index. The hypothesis that can be made to explain this is linked to "receptor desensitization", meaning that the dosage of norepinephrine must be increased over time to produce the desired effect (MAP > 65 mmHg), also causing an increase in side effects such as tachyphylaxis, vasoconstriction and therefore the increase in Resistance Indices. IR MES SUP = AMS; IR O right/left = CRA R/L; KIDNEY IR right/left = ARE R/L; IR snuff right/left = AR R/L; NORA DOSE = Ne dose


Microcirculatory response among patients undergoing elective cardiac surgery with cardiopulmonary-bypass

E Favaron1, C Ince2, J Montomoli2, WJ Van Boven3

1Amsterdam UMC, location AMC, Cardiac Surgery, Amsterdam, Netherlands, 2Erasmus MC, Rotterdam, Netherlands, 3Amsterdam UMC, location AMC, Amsterdam, Netherlands

Critical Care 2023, 27(S1): P190

Introduction: Fluid administration secondary to hemodynamics impairment and priming solutions together with the use of cardiopulmonary-bypass(CPB) during cardiac surgery are likely to promote hemodilution and inflammatory responses leading to microcirculatory impairment. We used handheld vital sublingual microscopy to examine microvascular modifications.

Methods: Sublingual microcirculation was assessed using incident dark field technology on 17 elective open cardiac surgery patients before surgery (T0), after induction of general anesthesia (T1), 10 min following the induction of the CPB (T2), after removal of the aortic clamp (T3), and at the end of surgery (T4). Postoperatively measurements were conducted on the day of discharge from the intensive care unit(T5) and three days after (T6). Microtools [1] and manual analysis were performed for the analysis of microcirculatory tissue perfusion [2] and microcirculatory leukocytes [3], respectively. Noninvasive assessment of the macrohemodynamics and of the body fluid composition were also collected during the study.

Results: During surgery, microcirculation showed a general impairment of tissue perfusion when compared to the baseline and in accordance with the macro-hemodynamics(cardiac output and mean arterial pression). In the post-operative days, the microcirculation showed a rapid recover to the baseline levels. Conversely, the inflammatory response at the microcirculatory level peaked at T3 and T4 and persisted also three days after surgery (Fig. 1).

Conclusions: Patients undergoing elective CPB during cardiac surgery showed an impairment of microcirculation in term of tissue perfusion, coherent with hemodynamic modification and characterized by relatively fast recovery. On the contrary, microcirculatory impairments due to tissue oedema and leukocyte activation persisted in the post-operative period.


  1. 1.

    Hilty MP et al. Crit Care Med 2020;48:e864–e875.

  2. 2.

    Hilty MP et al. Curr Opin Crit Care 2020;26:273–280.

  3. 3.

    Uz Z et al. J Appl Physiol 2018;124:1550–1557.

figure ca

Fig. 1 (abstract P190). A Tissue red blood cell perfusion (tRBCp) [2], mean ± sd. B Total microcirculatory leukocyte count (TLC) [3], mean ± sd


Comparison of microcirculatory changes in the first 24 and 48 h between survivors and non-survivors of septic shock

M Briuks1, S Kazune2

1University of Latvia, Student, Riga, Latvia, 2Hospital of Traumatology and Orthopedics, Anesthesiology Reanimatology, Riga, Latvia

Critical Care 2023, 27(S1): P191

Introduction: Microcirculatory dysfunction contributes to organ failure and mortality in patients with septic shock. Microcirculatory perfusion targeted resuscitation has been proposed as a tool to improve mortality in these patients. We hypothesized that microcirculation for non-survivers stays unchanged 24 and 48 h after admission compared to baseline, however, in survivors microcirculation improves in the first 48 h.

Methods: We included 36 adult sepsis patients admitted to the intensive care unit within the previous 24 h who were receiving a noradrenaline infusion. Skin oxygen saturation in the knee area was measured using a hyperspectral imaging-based method at baseline, after 24 and 48 h. The primary outcome was an increase in skin oxygen saturation depending upon survival status at 28 days.

Results: Median skin oxygen saturation changed from 11% (6–26) at baseline to 31 (23–33) % at 48 h in non-survivors and from 30% (23–36) to 28% (23–34) in survivors. There were significant differences in skin oxygen saturation between survivors and non-survivors at baseline (p < 0.01), but this difference was lost after 24 and 48 h. Independent baseline skin oxygen saturation predictors were APACHE II score and mean arterial pressure. Low baseline skin oxygen saturation was not associated with an increase in SOFA score over the first 2 days (r = − 0.38, p = 0.19).

Conclusions: At admission to intensive care, skin oxygen saturation is significantly lower in non-survivors of septic shock. Over the early course of treatment, it increases to survivor levels and is similar in all patients after 48 h of intensive care.


Predictive value of central venous-to-arterial carbon dioxide tension in patients with ARDS treated with veno − venous extracorporeal membrane oxygenation (ECMO)

P Djimafo1, K Kaefer1, C Pierrakos1, D Velissaris2, S Anane1, L Barreto Gutierrez1, S Redant1, A Gallerani1, R Attou1

1CHU Brugmann, ICU, Laeken, Belgium, 2University Hospital of Patras, Internal Medicine Department, Patras, Greece

Critical Care 2023, 27(S1): P192

Introduction: Central venous-to-arterial carbon dioxide tension (PvaCO2) can be useful for monitoring adequacy of tissue perfusion in patients with ARDS supported with veno-venous Extracorporeal Membrane Oxygenation (VV-ECMO). However, in theory, the unavoidable mixing of venous blood with blood after the oxygenator can affect PvaCO2 values by increasing central venous oxygen saturation and substantially decreasing CO2 concentration. This study aimed to evaluate acute changes in PvaCO2 after VV-ECMO installation and determine its association with patient outcomes.

Methods: Retrospectively evaluated coronavirus disease 2019 (COVID–19) ARDS patients with at least one concurrent arterial and central venous blood gas analysis before and after VV-ECMO installation as standard care. The primary outcome was intensive care unit (ICU) mortality at 28 days.

Results: 29 patients were enrolled in the study. All the patients had a 25 F drainage multistage femoral cannula and a 21 F internal jugular infusion cannula. The median distance between the central venous sampling point and the tip of the infusion cannula was 39 [23–73] mm. No statistically significant changes in PvaCO2 were observed 24–48 h after VV-ECMO installation (5 [4–7] mmHg to 6.5 [5–8.2] mmHg, p = 0.12). Hemoglobin concentration decreased 24 to 48 h after VV-ECMO installation (10.7 [9.5–12.7] g/dl to 9.6 [8.8–11.6] g/dl, p < 0.01) but neither central venous (75 [70–81]% to 73 [67–78]%, p = 0.46) nor arterial oxygen saturation (95 [92–97]% to 95 [93–96]%, p = 0.81) changed significantly. Elevated PvaCO2 after VV-ECMO installation had a good predictive value for 28 day ICU mortality (calculated area under the ROC curve 0.81) (Fig. 1 veno-venous).

Conclusions: VV-ECMO support appears to have little effect on the PvaCO2 calculation. PvaCO2 can be used to evaluate patients with ARDS supported with VV-ECMO, as persistently elevated values can be associated with poor outcomes.

figure cb

Fig. 1 (abstract P192). ROC curve for PvaCO2 values calculated before and after ECMO installation in predicting 28-day mortality


Change in perfusion index after 10-s and 20-s recruitment maneuvers to predict fluid responsiveness in mechanically ventilated patients during elective surgery

ST Thertchanakun1, SM Morakul2, ST Thitilerd2, CP Pisitsak2, WA Apinyachon2, WM Mongkolpun3

1Faculty of Medicine Ramathibodi Hospital, Mahidol University, Department of Anesthesiology, Bangkok, Thailand, 2Faculty of Medicine Ramathibodi Hospital, Mahidol University, Bangkok, Thailand, 3Siriraj Piyamaharajkarun Hospital, Siriraj Hospital, Mahidol University, Bangkok, Thailand

Critical Care 2023, 27(S1): P193

Introduction: Excessive volume expansion during surgery worsens outcomes. The perfusion index (PI) can track changes in stroke volume (SV) during recruitment maneuvers (RMs), and thus predict fluid responsiveness. Classic RMs (30 cmH2O continuous positive airway pressure (CPAP) for 30 s) can cause hypotension. We evaluated whether ∆PI after RMs of 10 or 20 s could predict fluid responsiveness.

Methods: Adult patients having elective surgery under general anesthesia with mechanical ventilation and an arterial line were included. After induction and hemodynamic stabilization (T0), a RM (30 cmH2O of CPAP) was performed for 10 s (RM10). Hemodynamic variables, pulse pressure variation (PPV) and PI (index finger) were obtained before and during RM10 and relative changes calculated. After return of SV to T0 values, a 20-s RM (RM20) was performed and the same variables measured. Crystalloid fluid (250 ml over 10 min) was then infused and fluid responsiveness (SV increase ≥ 10%) assessed. SV and PPV were recorded using pulse contour analysis. Areas under the receiver operating characteristic curves (AUCs) of variables (T0) and their changes were calculated.

Results: Thirty-two patients were studied: 18 (53%) were fluid responders (FRs). At T0, only PPV was higher in FRs than in non-FRs. SV, mean arterial pressure and PI decreased during RM10 and RM20, whereas PPV increased. ∆SV (16.8 ± 0.8 vs 36.2 ± 0.7%, p < 0.05) and ∆PPV (36.7 ± 1.5 vs 73.4 ± 1.8%, p < 0.05) were lower after RM10 than after RM20. ∆PI were higher in FRs than in non-FRs for RM10 and RM20 (Table 1 (abstract P193)). PPV at T0 and ∆PI at RM20 had the largest AUCs to predict fluid responsiveness (Table 1), with cutoff values ≥ 7.5% (sensitivity (Sn) 56%, specificity (Sp) 86%) and ≥ 36% (Sn 82%, Sp 64%), respectively. AUCs of ∆PI were similar for RM10 and RM20 (Table 1).

Conclusions: ∆PI during a 10 and a 20 s-RM predicted fluid responsiveness as well as baseline PPV did in surgical patients during the perioperative period.

Table 1 (abstract P193). Changes in hemodynamic parameters and their AUCs for predicting fluid responsiveness

Hemodynamic parameters

Fluid responders (n = 18)

Non-fluid responders (n = 14)

p value


∆PIRM10 (%)

45.0 ± 22.2

26.4 ± 19.7


0.71 [0.52–0.89]

∆PIRM20 (%)

53.2 ± 19.2

31.9 ± 24.0


0.76 [0.59–0.9]

∆SVRM10 (%)

16.3 ± 12.5

10.6 ± 8.4


0.63 [0.43–0.83]

∆SVRM20 (%)

44.4 ± 28.5

26.1 ± 21.6


0.69 [0.52–0.88]


45.2 ± 15.5

24.9 ± 36.9


0.57 [0.36–0.79]


80.3 ± 65.6

64.6 ± 81.16


0.57 [0.35–0.78]

PPV (%) at T0

9.5 ± 4.3

5.8 ± 2.3


0.77 [0.61–0.93]


Intraoperative hypotension when using hypotension prediction index software during major non-cardiac surgery: a European multicenter prospective observational registry (EU-HYPROTECT)

K Kouz1, MI Monge García2, E Cerutti3, G Draisci4, M Sander5, UH Frey6, SJ Davies7, P Bramlage8, B Saugel9

1University Medical Center Hamburg-Eppendorf, Department of Anesthesiology, Hamburg, Germany, 2Unidad de Cuidados Intensivos, Hospital Universitario SAS Jerez, Jerez de la Frontera, Spain, 3Azienda Ospedaliero Universitaria delle Marche, Department of Anesthesia, Transplant and Surgical Intensive Care, Ancona, Italy, 4Fondazione Policlinico Universitario Agostino Gemelli—IRCCS, Department of Emergency, Intensive Care Medicine and Anesthesia, Rome, Italy, 5University Hospital Giessen, Department of Anaesthesiology, Intensive Care Medicine and Pain Medicine, Giessen, Germany, 6Marien Hospital Herne, Department of Anesthesiology, Intensive Care, Pain and Palliative Care, Herne, Germany, 7NHS Foundation Trust, York and Scarborough Teaching Hospitals, York, UK, 8 IPPMed, Institute for Pharmacology and Preventive Medicine, Cloppenburg, Germany, 9 University Medical Center Hamburg-Eppendorf, Hamburg, Germany

Critical Care 2023, 27(S1): P194

Introduction: Intraoperative hypotension is associated with postoperative complications and mortality. Current intraoperative blood pressure management is mainly reactive—with hypotension being treated only after it has occurred. Predictive hemodynamic monitoring may be a promising approach to help clinicians reduce intraoperative hypotension. The Acumen™ Hypotension Prediction Index software (HPI-software) (Edwards Lifesciences; Irvine, CA, USA) was developed to predict hypotension. We built up the European, multicenter, prospective, observational EU-HYPROTECT registry to describe the incidence, duration, and severity of intraoperative hypotension when using HPI-software monitoring in patients having non-cardiac surgery.

Methods: We enrolled 749 patients scheduled for elective major non-cardiac surgery in 12 medical centers in 5 European countries. All patients had intraarterial blood pressure monitoring with an arterial catheter and intraoperative HPI-software monitoring. The primary endpoint of this registry was intraoperative hypotension quantified using the time-weighted average mean arterial pressure (MAP) < 65 mmHg. Secondary endpoints included the proportion of patients with at least one ≥ 1-min-episode of a MAP < 65 mmHg, and the number of ≥ 1-min-episodes of a MAP < 65 mmHg.

Results: Between September 2021 and May 2022, we enrolled 749 patients in the registry and included 702 patients in the final analysis. The median time-weighted average MAP < 65 mmHg was 0.03 (0.00, 0.20) mmHg (Fig. 1). 285 patients (41%) had no ≥ 1-min-episode of a MAP < 65 mmHg—417 patients (59%) had at least one. The median number of ≥ 1-min-episodes of a MAP < 65 mmHg was 1 (0, 3).

Conclusions: The median time-weighted average MAP < 65 mmHg was very low in patients in this registry. This suggests that using HPI-software monitoring may help reduce the duration and severity of intraoperative hypotension in patients having non-cardiac surgery.

figure cc

Fig. 1 (abstract P194). Bar chart illustrating the time-weighted average mean arterial pressure (TWA MAP) < 65 mmHg in patients of this registry and of previously published studies on HPI-software monitoring


Left ventricular function monitored in intensive care using artificial intelligence and transesophageal echocardiography

J Yu1, H Flade1, A Taskén2, EA Berg3, B Grenne3, A Rimehaug1, I Kirkeby-Garstad4, G Kiss2, S Aakhus3

1Norwegian University of Science and Technology/St Olav’s Hospital, Department of Circulation and Medical Imaging/Department of Anesthesia and Intensive Care, Trondheim, Norway, 2Norwegian University of Science and Technology, Department of Computer Science, Trondheim, Norway, 3Norwegian University of Science and Technology/St Olav’s Hospital, Department of Circulation and Medical Imaging/Clinic of Cardiology, Trondheim, Norway, 4Norwegian University of Science and Technology, Department of Circulation and Medical Imaging, Trondheim, Norway

Critical Care 2023, 27(S1): P195

Introduction: Left ventricular (LV) function is usually not assessable by standard hemodynamic monitors. To automatically monitor LV function, we have developed a method for measuring mitral annular plane systolic excursion (MAPSE) by combining transesophageal echocardiography with artificial intelligence: AutoMAPSE (Fig. 1A). Our objective is to assess the feasibility of AutoMAPSE in intensive care, and to compare its bias and least significant change (LSC) against manual measurements.

Methods: We obtained midesophageal two- and four-chamber views of three beats in 40 intensive care patients. Off-line, we measured MAPSE manually and with AutoMAPSE in four LV walls in all available beats. We defined AutoMAPSE as feasible when it was obtainable from at least one beat in one wall. To assess bias and limits of agreement (LOA), we used Bland–Altman analysis. We calculated LSC for each method, which is the least difference between two measurements required before a change is considered real.

Results: AutoMAPSE had a feasibility of 96.2%. Bland–Altman analysis showed that AutoMAPSE had a low bias (Fig. 1B), ranging from −0.2 mm (anterior wall, LOA:  −3.4 to 3.1 mm) to −1.1 mm (septal wall, LOA:  −4.1 to 2.4 mm). AutoMAPSE had higher LSC than manual measurements when comparing a single measurement (5.4 vs 4.7 mm). When AutoMAPSE is set to measure more beats than manual, its LSC is reduced (LSC for two beats: 3.8 mm).

Conclusions: AutoMAPSE is feasible in intensive care patients. Compared to manual measurements, AutoMAPSE had low bias. By using AutoMAPSE to measure more beats, its LSC is reduced.

figure cd

Fig. 1 (P195). A Left ventricular function monitored using AutoMAPSE. Red dots show the detection of mitral annuli in a midesophageal four-chamber view. B Bland–Altman plots of two methods. Red line, bias. Dashed lines, 95% limits of agreement


Vascular access for all

D Cloete, G Barker

John Radcliffe Hospital, Critical Care, Oxford, UK

Critical Care 2023, 27(S1): P196

Introduction: Peripheral (PVC) and central (CVC) venous cannulation are key components of care for many ED and ICU patients. PVC is typically obtained using direct visualization and physical palpation; and performing CVC under ultrasound-guidance (compared to the landmark technique) is the modern gold standard. Ultrasound improves success rates, particularly in patients with difficult IV access. Appropriate training is pertinent for all ED/ICU practitioners, and procedural proficiency is expected. Many departments have ultrasound machines, yet the vast majority of doctors are engaging in self-directed learning in this regard. We evaluated the training received by doctors in ultrasound-guided (U/S) PVC and CVC.

Methods: An electronic survey was disseminated to all ranks of practitioner from several hospital departments, and reviewed levels of training, competence and procedural preferences in both U/S-PVC and CVC.

Results: A total of 148 eSurvey responses were received (response rate of 52%). Regarding U/S-PVC, 68% of respondents had never been taught with 65% achieving this previously, and 12% attempting to do so but failing. For CVC, 61% have inserted them. Of these, 84% always use ultrasound and 16% elect to use either ultrasound or an anatomical landmark-based technique. 55% have undergone formal training. Despite this, 10,2%, 9,6% and 11.3% of these individuals have not received training in the procedure itself, sterility during the procedure or post CVC protocols, respectively. When reviewing operator preference for specific U/S-CVC sites, participants expressed varying levels of confidence (see Fig. 1). Respondents wishing to receive formal, comprehensive training on U/S-PVC and U/S-CVC totaled 72% and 76%, respectively.

Conclusions: U/S-PVC & CVC ability varied significantly in a large academic hospital. A large training gap exists, particularly for those previously taught U/S-CVC to ensure universal, comprehensive proficiency. This dataset will be used to launch an U/S-vascular access course in the region.

figure ce

Fig. 1 (abstract P196). Confidence levels of doctors performing central venous cannulation


Evaluation of the image quality and validity of handheld echocardiography for stroke volume and left ventricular ejection fraction quantification: a method comparison study

FM de Raat1, J Van Houte2, LJ Montenij2, S Bouwmeester3, SEA Felix3, EC De Boer1, P Bingley4, P Houthuizen3, RA Bouwman5

1University of Eindhoven, Electrical engineering, Eindhoven, Netherlands, 2Catharina Hospital, Anesthesiology and Intensive Care, Eindhoven, Netherlands, 3Catharina hospital, Cardiology, Eindhoven, Netherlands, 4Philips, Patient Care and Monitoring, Eindhoven, Netherlands, 5Catharina Hospital, Anesthesiology, Eindhoven, Netherlands

Critical Care 2023, 27(S1): P197

Introduction: Bedside quantification of stroke volume (SV) and left ventricular ejection fraction (LVEF) is valuable in hemodynamically compromised patients. Miniaturized, battery operated handheld ultrasound (HAND) devices are now available for clinical use. However, the performance level of HAND devices for automated quantified cardiac assessment is yet unknown. The aim of this study was to compare the validity of HAND measurements with Standard Echocardiography (SE) and three-dimensional echocardiography (3DE).

Methods: Thirty-six patients were scanned with HAND, SE and 3DE. The image quality of HAND and SE apical four and two chamber views were evaluated by scoring segmental endocardial border delineation (2 = good, 1 = poor, 0 = invisible). LVEF and SV of HAND was evaluated against SE and 3DE using correlation and Bland–Altman analysis. HAND and SE datasets were analysed automatically with Auto Strain. The 3DE dataset was analysed automatically with the Dynamic Heart Model.

Results: The correlation, bias, and limits of agreement (LOA) between HAND and SE were 0.68 [0.46:0.83], 1.60% [−2.18:5.38], and 8.84% [−9.79:12.99] for LVEF, and 0.91 [0.84:0.96], 1.32 ml [−0.36:4.01], 15.54 ml [−18.70:21.35] for SV (Fig. 1), respectively. Correlation, bias, and LOA between HAND and 3DE were 0.55 [0.6:0.74], −0.56% [−2.27:1.1], and 9.88% [−13.29:12.17] for LVEF, and 0.79 [0.62:0.89], 6.78 ml [2.34:11.21], 12.14 ml [−26.32:39.87] for SV, respectively. The image quality scores were 9.42 ± 2.0 for the apical four chamber views of the HAND dataset and 10.49 ± 1.7 for the SE dataset and (p < 0.001).

Conclusions: Clinically acceptable accuracy, precision, and image quality was demonstrated for HAND measurements compared to SE. Also, LVEF quantification with HAND showed a clinically acceptable accuracy and precision compared to 3DE.

figure cf

Fig. 1 (abstract P197). Upper panel: Correlation analysis of the HAND and SE data for (A) LVEF, and (B) SV. Lower panel: Bland–Altman plots of HAND versus SE for (C) LVEF, and (D) SV. LVEF = left ventricle ejection fraction; SV = stroke volume; R = correlation coefficient, R2 = Regression coefficient


Smartphone-based assessment of left ventricular ejection fraction using a silicon chip ultrasound pocket probe and a machine learning algorithm

J Leote1, FA Gonzalez1, J Bacariza1, R Varudo1, C Martins1, RM Meireles1, A Fernandes1, F Michard2

1Hospital Garcia de Orta, EPE., Critical Care, Almada, Portugal, 2MiCo, Denens, Switzerland

Critical Care 2023, 27(S1): P198

Introduction: Our study aims to compare a smartphone-based assessment of left ventricular ejection fraction with a silicon-chip ultrasound (US) probe and a machine learning algorithm (LVEFSMART) to reference manual measurements (LVEFREF) done with a high-end US.

Methods: We studied critically ill patients requiring an echocardiography. LVEFSMART measurements were made with a smartphone algorithm (Caption AI, Caption Health, USA) connected to a US probe (IQ, Butterfly Inc., USA). LVEFSMART data was compared with LVEFREF done by two experts using a conventional US (Venue, GE Healthcare, USA). Measurements were done in triplicate from a 4-chamber apical view and intra-operator reproducibility was also calculated.

Results: We enrolled 95 patients (mean age 60 ± 17 years), 34% were mechanically ventilated and 31% received continuous vasopressor support. LVEFREF ranged from 26 to 80% (mean 54 ± 12%) and the reproducibility was 9 ± 6%. Whereas LVEFSMART ranged from 28 to 79% (mean 54 ± 12%) with the reproducibility of 8 ± 6%. A fair relationship (r = 0.75, p < 0.001) between LVEFREF and LVEFSMART. The average difference (bias) between LVEFSMART and LVEFREF was 0 ± 8% (95% agreement limits of −17 to +16%). Thirty patients had a LVEFREF < 50% considered as left systolic dysfunction. LVEFSMART detected dysfunction with a sensitivity of 70% and specificity of 89%.The algorithm measured LVEFSMART automatically in 45 patients (fully-automatic measurements). In the remaining 50 patients, the algorithm enabled automatic left ventricular border detection, but end-diastolic and end-systolic images were selected by the operator (semi-automatic measurements). The agreement between LVEFSMART and LVEFREF was better for fully automatic than for semi-automatic measurements (Fig. 1).

Conclusions: The accuracy of LVEFSMART showed no bias and a high specificity to detect left ventricular dysfunction. However, sensitivity was low with a wide range of agreement limits, particularly for semi-automatic measurements.

figure cg

Fig. 1 (abstract P198). Reproducibility, correlation, bias, and agreement between reference and semi-automatic measurements


The evaluation of cardiac functions in deep Trendelenburg position during robotic-assisted laparoscopic prostatectomy

E Kilinc1, S Aktas Yildirim1, H Ulugol1, E Eroglu Buyukoner2, B Gucyetmez1, F Toraman1

1Acibadem Mehmet Ali Aydinlar University School of Medicine, Department of Anesthesiology and Reanimation, Istanbul, Turkey, 2Acibadem Mehmet Ali Aydinlar University School of Medicine, Department of Cardiology, Istanbul, Turkey

Critical Care 2023, 27(S1): P199

Introduction: In this study, we investigated how to affect invasive cardiac parameters achieved on invasive artery monitoring by using Pressure Recording Analytical Method (PRAM) in deep Trendelenburg (dTD) position during robotic-assisted laparoscopic prostatectomy (RALP). In RALP, dTD position (25°–45°) is performed following pneumoperitoneum [1]. In dTD position, although protection of cardiac output is expected caused by increased venous return, a decrease in hemodynamic parameters can be seen because of gravity and increased intrathoracic pressure [2].

Methods: After ethical approval, this study was designed as a prospective observational study. Thirty patients with malignant prostate cancer who were > 18 years old, and had no chronic heart failure, valvular disease, arrhythmia, and myocardial ischemia in the last 3 months were included in the study. All invasive cardiac parameters and longitudinal strain (SL) achieved transesophageal echocardiography were recorded in pre-dTD (T3) and 10th minute of dTD (T4). Delta values were calculated for CCE and SL (values at T4 minus values at T3). The estimated power was calculated as 1.0 in accordance with cardiac cycle efficiency (CCE) values at T3 and T4 (Effect size: 0.85 standard deviations of the mean difference: 0.22, alfa:0.05).

Results: At T4, heart rate, pulse pressure variation, CCE, dP/dt and SL were significantly lower than those at T3 (p = 0.009, p < 0.001, p < 0.001, and p < 0.001 respectively). There was a positive correlation between delta-CCE and delta-SL (r2 = 0.36, p < 0.001).

Conclusions: Cardiac workload is increased in dTD position and using traditional monitoring is not enough to evaluate cardiac function in this position. CCE can be used to evaluate changed cardiac workload and changes in CCE are compatible with SL. Therefore, we suggest monitoring CCE in especially risky positions during surgeries.


  1. 1.

    Ficarra V et al. Eur Urol 2009;55:1037–63.

  2. 2.

    Lestar M et al. Anesth Analg. 2011;5:1069–1075.


Machine learning-driven early detection of bleeding events in a human bleeding model

A Dey1, J Kim2, X Li3, M Hravnak4, MR Pinsky5, G Clermont5, A Dubrawski3, JH Yoon2

1Georgia Institute of Technology, Department of Biomedical Engineering, Atlanta, USA, 2University of Pittsburgh, Pulmonary and Critical Care Medicine, Pittsburgh, USA, 3Carnegie Mellon University, School of Computer Science, Pittsburgh, USA, 4University of Pittsburgh, School of Nursing, Pittsburgh, USA, 5University of Pittsburgh, Critical Care Medicine, Pittsburgh, USA

Critical Care 2023, 27(S1): P200

Introduction: Unanticipated physiologic instabilities from hemorrhage are dangerous. With a simulated human bleeding model on whole blood donor volunteers, we sought to identify early signatures of bleeding using a supervised machine learning model.

Methods: 51 adult human subjects donated whole blood (mean age 49 ± 18 years; 55% male). On average, subjects pre-bleed hemoglobin was 14.3 g/dl, and 495 ml of blood was donated during 7 min 3 s. Each individual was instructed to stand and sit before and after donation to obtain orthostatic blood pressures. Noninvasive vital signs including electrocardiogram (250 Hz) and photoplethysmography (125 Hz) were obtained from a portable MP50 monitor, and non-invasive arterial waveform (100 Hz) with cardiac indices were acquired from a LiDCOrapid device. During donation, blood volume loss was documented each minute. Data were curated and 20 s moving time windows were extracted every 10 s. From these time windows, statistical, time-series, and morphologic features were computed. A random forest classifier was trained to identify bleeding events using these features, which were normalized by their non-bleeding baseline period before any orthostatic changes.

Results: Our model, evaluated using a leave-one-patient-out cross validation scheme, demonstrated high performance in detecting bleeding using only non-invasive vital sign features: the mean AUROC was 0.955, and the TPR and TNR at clinically relevant low error regions (1% FPR & 1% FNR) were 55% and 46% (Fig. 1). Bleeding was then subclassified across the bleeding epoch progression as early (≤ 100 ml), mid (100–250 ml), and late (> 250 ml) and analyzed. Truncated ROC curves in Fig. 1 show that the models were able to detect even mild bleeding, showing the robustness of the model in very subtle physiologic changes in bleeding human.

Conclusions: Using a random forest model, physiologic evidence of bleeding could be identified at a very early course of bleeding in human subjects.

figure ch

Fig. 1 (abstract P200). Truncated receiver operating characteristic curves to detect bleeding events


An experimental study via novel dynamic lower body negative pressure model to analyse simulated hemorrhage using arterial blood pressure

N Nesaragi1, LØ Høiseth2, HA Qadir3, LA Rosseland4, PS Halvorsen5, I Balasingham6

1Oslo University Hospital, The Intervention Centre, Oslo, Norway, 2University of Oslo, Division of Emergencies and Critical Care, Department of Anesthesiology, Oslo, Norway, 3Oslo University Hospital, Oslo, Norway, 4University of Oslo, Oslo, Norway, 5University of Oslo, Institute of Clinical Medicine, Faculty of Medicine, Oslo, Norway, 6Norwegian University of Science and Technology, Department of Electronic Systems, Trondheim, Norway

Critical Care 2023, 27(S1): P201

Introduction: Bleeding is a life-threatening complication in trauma and surgery. Early detection of bleeding is mandatory to improve outcomes but is difficult due to physiological responses and unreliable vital signs monitoring. We explored the discriminative ability of a novel deep learning (DL) framework on arterial waveforms to classify levels of ongoing hypovolemia in a new dynamic version of the lower body negative pressure (LBNP) model.

Methods: Healthy subjects (n = 23) with invasive radial arterial pressure had three runs of LBNP (69 trials). The LBNP levels (range 0 to −60 mmHg) in each run were randomized in order to; 1) circumvent the time dependency of measurements, and 2) mimic both bleeding and fluid resuscitation (dynamic model). A supervised DL-based framework for ternary classification was realized by segmenting the underlying invasive signal and labeling segments with corresponding LBNP target levels. The DL framework had two inputs that were trained with respective time–frequency representations extracted on waveform segments to classify each of them into blood volume loss: Class 1 (mild); Class 2 (moderate); or Class 3 (severe). At the outset, the latent space derived from the unified DL model via late fusion among both inputs assisted in enhancing classification performance.

Results: The experimental study was performed using threefold cross-validation based on a patient-wise stratification scheme. i.e., within each fold, the model is trained and developed using a unique 70% (16 subjects) of total data, and the remaining 30% (7 subjects) data is considered for validation. In a threefold cross-validation setup with stratified subjects, the experimental findings demonstrated AUROC: 0.84, AUPRC: 0.73, and F1-score: 69.12% (Fig. 1).

Conclusions: Our proposed DL algorithm on arterial waveforms enabled to capture accurately the complex interplay in physiological responses related to both bleeding and fluid resuscitation.

figure ci

Fig. 1 (abstract P201). A supervised deep learning (DL)-based framework for ternary classification of the entire dynamic LBNP trajectory depicting simulated hemorrhage is realized by segmenting the ABP and labeling segments with corresponding LBNP target levels. a The proposed DL model is trained with time frequency representations extracted on waveform segments to classify each of them into blood volume loss: Class 1 (mild); Class 2 (moderate); or Class 3 (severe). b Average area under receiver-operating characteristic curves (AUC) displaying the ability of the proposed unified model to perform the desired classification in a threefold cross-validation setup. c Sample distribution of waveform segments for model training and validation in threefold cross-validation


Continuous quantitative analysis of urine in ICU patients

N Postma1, S Struiken1, L De Jager1, M Onrust1, J Kootstra 2, C Cordeiro3, M Nijsten4

1UMC Groningen, Dpt of Critical Care, Groningen, Netherlands, 2UMC Groningen, Dpt of Clinical Chemistry, Groningen, Netherlands, 3Health Hub, Paramedir, Roden, Netherlands, 4UMC Groningen, ICU, Groningen, Netherlands

Critical Care 2023, 27(S1): P202

Introduction: With advanced sensor technology, it is in principle feasible to perform continuous monitoring of urinary electrolytes and metabolites in ICU patients with a urinary catheter. An open question is whether frequent urinary measurements in ICU patients will detect relevant changes. Thus we examined if semi-continuous quantitative monitoring of urine might be useful in the ICU.

Methods: Patients with a range of indications for ICU-admission and different length of stays in the ICU were included. For each patient, 20 urine samples were collected every 15 min for 5 h. In addition to urinary volume, samples were analyzed at the clinical chemistry laboratory for 10 parameters: diuresis, creatinine, urea, glucose, natrium, chloride, kalium, phosphate, total protein and albumin as well as the anion gap. Data were analyzed in R studio and examined for bivariate correlations, specific patterns and rapid changes of these parameters. For each patient and parameter, the ratio between the median value and highest/lowest value was determined. The frequency distributions of these ratios are shown to illustrate the dynamics.

Results: In 209 patients 3892 samples were analyzed. Remarkably rapid and pronounced changes were seen in all parameters (Fig. 1). The independent behavior of the studied parameters was underscored by limited bivariate correlations that ranged from R = −0.19 (creatinine and sodium) to R =  + 0.84 (albumin and total protein). Sudden deviations of certain parameters could be frequently associated with documented events.

Conclusions: The concentrations of the studied parameters could change rapidly and independently within the 5 h observation period. High time-resolution monitoring of urine electrolytes and metabolites may provide a novel near real-time and non-invasive window on the complex pathophysiology of ICU patients.

figure cj

Fig. 1 (abstract P202). Ratio between most extreme value and median value within 8 h study period. Na: sodium; K: potassium; Cl: chloride; P: phosphate; glu: glucose; TP: total protein


Bioimpedance analysis in major surgeries

E Kyosebekirov, I Minev, CH Stefanov, D Kazakov, E Mitkovski, S Nikolova, G Pavlov

Medical University, Department of Anaesthesiology and Intensive Care Medicine, Plovdiv, Bulgaria

Critical Care 2023, 27(S1): P203

Introduction: The aim of this study is to assess the feasibility of bioimpedance analysis (BIA) in the assessment of fluid status and the effect of fluid therapy in major surgeries.

Methods: Observational, prospective study conducted in the period April 2021-September 2022 amongst patients hospitalized for elective major surgical interventions in University hospital St. George-Plovdiv, Bulgaria. BIA measurements were performed 1 h before and after general anesthesia. Intraoperatively standard monitoring was applied. Perspiratio insensibilis were accepted as 0.5 ml/kg/h and net balance was calculated. BodyStat Multiscan 5000 and manufacturer’s electrodes were used. Overhydration (OHY), total body water (TBW), extracellular water (ECW), and intracellular water (ICW) were the analyzed parameters.

Results: 40 patients aged 59.15 ± 9.68 years with BMI 27.1 ± 4.3 kg/cm2 and duration of the procedure 189.75 ± 87.63 min were involved in the study. There were 18 laparotomies, 15, thoracotomies and 7 craniotomies. Calculated net balance was 1344.64 ± 533.02 ml. TBW, ECW, ICW and OHY increased from the respective 44.84 ± 10.38 l, 15.08 ± 3.35 l, 29.66 ± 7.29 l, -0.77 ± 0.84 to 46.38 ± 10.57 l, 16.62 ± 3.52 l, 29.76 ± 7.3 l, 1.28 ± 0.75. There was a correlation between net balance and the increase in TBW (r = 0.41, p < 0.05), ECW (r = 0.47, p < 0.05) and OHY(r = 0.33, p < 0.05), whereas ICW increase was insignificant (p = 0.24).

Conclusions: BIA gives promising results in the evaluation of TBW and body fluid distribution in the perioperative period and could be a useful tool in the complex fluid therapy management. Although relatively old, the method is gaining increasing popularity and wider application amongst patients.


Fluid balance association with mechanical ventilation time in cardiopulmonary bypass

CD Del Angel Argueta1, OI Aguilera Olvera2, GA Aguirre Gomez2, GL Arcos Lopez2, GL Velazquez Estrada2, JC Muñoz Chaves2, JA Villalobos Silva2

1Hospital Regional De Alta Especialidad, Anestesiology, Ciudad Victoria, Mexico, 2Hospital Regional De Alta Especialidad, Critical Care, Ciudad Victoria, Mexico

Critical Care 2023, 27(S1): P204

Introduction: In patients undergoing cardiac surgery, positive fluid balances are common and are associated with longer mechanical ventilation, renal failure, and mortality, causing tissue edema, organ dysfunction, and coagulopathy, it is associated with an increased risk of post-operative complications [1, 2].

Methods: We performed a retrospective cohort study to analyze the association between positive fluid balances during the perioperative stage of cardiopulmonary bypass surgery and mechanical ventilation duration in a high specialty hospital of northeast Mexico during 2016–2022, we included patient records and collected data demographics, clinical variables. We use percentages, means and standard deviation accordingly, p value (< 0.05), odds ratio (CI 95%) for mechanical ventilation time (< 12 h), AKI, LOS ICU, delta lactate and urgent reintervention.

Results: We analyzed 60 patient records who were intervened in the Regional High Specialty Hospital, 66.7% were women, the mean age 53 years (SD 14.1), BMI mean 27.3 kg/m2 (SD 4.7), NYHA class III in 48.3%, 50% was coronary heart disease and 30% for valvular heart disease, EuroScore II mean 1.56% (SD 2.16), with preoperative creatinine of 0.8 mg/dl (SD 0.3 mg/dl), during the procedure the lactate maximum value 4.13 (SD 2.1 mmol/l), odds ratio (OR) for mechanical ventilation less than 12 h favoring positive fluid balance less than 2000 ml was 2.73 (CI 95% 0.35–21.17, p = 0.32) (Fig. 1), for AKI OR 0.23 (CI 95% 0.06–0.84, p = 0.21), ICU LOS OR 1.40 (CI 95% 0.42–4.66, p = 0.57), delta lactate 4 mmol/L OR 3.62 (CI 95% 0.72–18.06, p = 0.10) and urgent reintervention OR 0.78 (CI 95% 0.66–9.21, p = 0.84).

Conclusions: In our study we found that there is no association between fluid balance and the duration of mechanical ventilation, instead a correlation was observed to present acute kidney injury, however more studies with a larger sample are needed for conclusions.


  1. 1.

    Romagnoli S et al. J Cardiothorac Vasc Anesth 2016;30:1076–84.

  2. 2.

    Koc V et al. Crit Care Res Pract. 2020;2020:4,836,862.

figure ck

Fig. 1 (abstract P204). Boxplot of perioperative fluid balance in mechanical ventilation time


Correlation between the initial fluid balance of patients with preeclampsia with severity data and its complications in the intensive care unit

JA Garza Carrión, C Zarazúa Sosa, OI Aguilera Olvera, A Zárate Gracia, JA Villalobos Silva

General Hospital "Dr. Norberto Treviño Zapata", Adult Intensive Care Unite, Ciudad Victoria, Mexico

Critical Care 2023, 27(S1): P205

Introduction: Preeclampsia is one of the most frequent diagnoses for admission to the ICU. We will analyze the relationship between the initial fluid balance and the complications expressed in patients admitted to the ICU of an hospital in the Northeast of Mexico from January 2020 to October 2022.

Methods: In a retrospective cohort study we analized the association between fluid balance at 24, 48 and 72 h from ICU admission at an hospital in northeast of Mexico, included obstetric patients with complications such as preeclampsia-eclampsia, HELLP syndrome, fluid overload, gestational hypertension. We analyzed categorical variables with percentages, numerical variables with mean and standard deviation and obtained odds ratio for complication events.

Results: We analyzed data from 66 patients admitted to ICU during 2020–2022, mean age group were 21–29 years old (48.5%), with obstetric history of 2–3 pregnancies (53%), comorbidities such as previous hypertension and DM2 were 21.2 and 9.1%, most affected organ were kidney, brain and liver (63.6%), complications like pregnancy arterial hypertension (42.4%), eclampsia (9.1%), fluid overload (7.6%), HELLP syndrome (1.5%) and liver rupture (1.5%) were registered. Fluid balance at 24 h admission was 1148.95 ml mean value (sd 819.524 ml), 48 h -187.16 ml (sd 1097.167 ml), 72 h -1537 ml (sd 1302.491). Association for 24 h fluid balance and eclampsia was OR 0.82 (95% CI 0.72–0.934), fluid overload OR 8.438 (95% CI 1.028–69.247), arterial hypertension OR 0.367 (95% CI 0.074–1.829), liver rupture OR 1.067 (95% CI 0.94–1.211), HELLP syndrome OR 0.951 (95% CI 0224–4.047).

Conclusions: In this study we found association between positive fluid balances at 24 and 48 h in obstetric patients for fluid overload and for eclampsia at 72 h (Fig. 1). First 24 h there is a tendency for more positive fluid balances that gradually becomes negative when ICU arrival.

figure cl

Fig. 1 (abstract P205). Comparison between fluid balance at 24 h, 48 h and 72 h of patients with severe preeclampsia in ICU


The outcomes of PICC insertion in pediatric patients at tertiary-care, university hospital

N Wongchompoo

Siriraj Hospital, Department of Anesthesiology, Bangkok, Thailand

Critical Care 2023, 27(S1): P206

Introduction: Peripherally-inserted central catheter (PICC) is widely used for intermediate- and long-term venous access. Venipuncture and catheterization in pediatric patients were challenging. Children's vein may become damaged by frequent and painful needle insertions. In Thailand, PICC was not yet prevailing even in adult patients and limited of the published works in pediatric patients. This study aims to demonstrate the outcome of PICC insertion in pediatric patients by the Anesthesia Line Service Team (ALiST) of a tertiary-care, university hospital in Thailand.

Methods: This is a retrospective, descriptive study between January 2018 and December 2021. The inclusion criteria were pediatric patients (age < 15 years), body weight more than 5 kg and no history of vein abnormality. The primary outcome is the success rate of insertion. The characteristic of patients, catheter, reason of removal and complications were also reported.

Results: A total of 1850 PICCs were inserted during the study period and 149 PICC were inserted in pediatric patients. There were 63 boys and girls equally. The median age of patients was 5.47 years (ranging from 3 months to 15 years). The median height was 106.06 cm, while median weight was 20.10. The successful insertion rate was 99.21%. All of them were inserted under ultrasound-guided, with or without real-time fluoroscopy. No complications during insertion were noted. The average usage days were 66.48 days (4–402). Four Fr, single lumen catheter was the most common PICC used (38.1%), followed by 3Fr, single lumen (32.5%) and 5Fr, double lumen (29.4%). Reason for removal of PICC lines were completion of therapy (50.86%), catheter malfunction (25.86%), infection (6.9%), and accidental removal (2.59%).

Conclusions: This is the first study about PICC lines insertion in pediatric patient in university hospital of Thailand. Our study showed the successes rate of 99.41%. PICC in pediatric patients are safe and low complication.


Intracavitary electrocardiogram guided positioning of central venous catheter in critically ill patients

LC Ciatti, CB Biuzzi, SF Finetto, SS Scolletta

Azienda ospedaliera universitaria Senese, Dipartimento Anestesia e Rianimazione, Siena, Italy

Critical Care 2023, 27(S1): P207

Introduction: Our research group conducted this study with a primary objective to determine if the ECG guided technique (ECGgt) could reduce the number of post-procedural chest x-rays (CXR). In critically ill patients central venous catheter (CVC) placement is a commonly performed procedure with potential serious complications [1]. Various techniques to ascertain CVC tip placement (CVCtp) include anatomical landmark-guided methods, CXR and right atrium (RA) ECG.

Methods: All 128 patients enrolled in the study were hospitalized in Siena ICU, ‘Anestesia e Rianimazione’ department. Data were collected from 1° January 2022 to 31° June 2022 and retrospectively compared to data of 2021. The ECGgt included the metallic wire technique (CVC Braun Certofix + ECG adapter kit) and the fluid column technique (ECG VygoCard Kit). With ECGgt an accurate analysis of the P-shape at the ECG monitor was required. By monitoring the P-wave, we were able to determine CVCtp as it travelled through SVC-RA and RA.

Results: Out of the 128 patients 106 patients underwent positioning of a CVC with ECGgt (82.8%), for 22 the procedure was contraindicated. Of the 106 CVC, 80 were PICC, 26 were CICC. Intra-procedural complication rate was 1.8%. 80 patients received CXR after ECGgt and correct positioning was confirmed in 100% of cases. Data of 2022 showed that ECGgt had a 49% increase (34,1% in 2021 and 84% in 2022). Conversely a decrease in CXR was shown (Fig. 1).

Conclusions: Our study shows that ECGgt is a safe, reliable technique and a potential valid alternative to CXR. The advantages of ECGgt are the high feasibility, the catheter being operable just after the procedure, the optimisation of resources in the ICU and patients/health care staff x-rays exposure being avoided.


  1. 1.

    Soldati G. et al. Chest 2008;133;204–21.

figure cm

Fig. 1 (abstract P207). Use of CXR


The impact of central venous catheter (CVC) insertion on intravenous fluid burden in high-risk colorectal surgery: an observational study

B Milne1, P Beddoes1, A Davidson1, H Carter1, D Melia2

1Whipps Cross University Hospital, London, UK, 2Whipps Cross University Hospital, Dept of Anaesthesia and Critical Care, London, UK

Critical Care 2023, 27(S1): P208

Introduction: Perioperative fluid strategy remains controversial [1]. We sought to address whether placement of a CVC influences this. There have been conflicting results from trials of goal-directed therapy [2–3], but avoidance of aggressive fluid loading appears to be associated with better outcome [4].

Methods: The study was granted Health Research Authority approval (London Central). All adult patients presenting to our institution for elective major colorectal resection, with planned postoperative critical care, were included between May 2019 and March 2020. Primary outcome measure was the difference in mean fluid volume infused between the two groups (‘CVC Inserted’ and ‘No CVC’). Student's T-Test was used to compare means between the two groups. Where this was not appropriate, differences in median values were calculated using non-parametric methods (i.e. Mann Whitney U Test).

Results: 40 patients were recruited (17 'CVC Inserted' and 23 ‘No CVC’). There were significantly lower median volumes of both crystalloid and total intravenous fluid volume (1800 ml vs. 3000 ml; p value < 0.0001, 95% CI 750–1950 ml and 750–1800 ml respectively) infused in the ‘CVC Inserted’ group. Postoperative fluid volumes demonstrated a trend for reduced fluid volume in the ‘CVC Inserted’ group, however this did not reach statistical significance for either total crystalloid (460 ml vs. 980 ml; p value = 0.0633, 95% CI 0–940 ml) or weight and time adjusted fluid volume (0.012 ml/kg/min vs. 0.015 ml/kg/min; p value = 0.0880, 95% CI − 0.001 to 0.015 ml/kg/min).

Conclusions: An average of 40% less fluids were given to the ‘CVC inserted group’ in our study across the perioperative period. The potential reasons for this are numerous and further research is required to investigate this further.


  1. 1.

    Miller TE et al. Anesthesiology 2019;130:825–832.

  2. 2.

    Pearse RM et al. JAMA 2014;311:2181‐2190.

  3. 3.

    Lewis SR et al. Cochrane Database of Systematic Reviews 2016;3:CD003004.

  4. 4.

    Lobo DN et al. Lancet 2002;359:1812‐1818.


Effects of fluid bolus on left ventricular myocardial performance index and global longitudinal strain in critically ill patients: a pilot study

C Pierrakos1, K Kaefer1, S Anane1, A Gallerani1, L Barreto Gutierrez1, P Djimafo1, S Redant1, D Velissaris2, R Attou1

1CHU Brugmann, ICU, Laeken, Belgium, 2University Hospital of Patras, Internal Medicine Department, Patras, Greece

Critical Care 2023, 27(S1): P209

Introduction: Left ventricular myocardial performance index (LV − MPI) may be an attractive method for monitoring changes in LV systolic and diastolic function during fluid bolus (FB). However, it is not clear whether FB induced changes in LV preload can affect LV − MPI calculation. In this study, we evaluated changes in LV − MPI during FB.

Methods: This is a post-hoc analysis of a prospective observational study that evaluated patients who received a comprehensive transthoracic echocardiogram examination before and after FB. All patients received 4 mL/kg crystalloids at a rate of 1000 mL/h as local standard care. Using the tissue Doppler technique, LV − MPI was calculated as the ratio between the sum of the iso-volumetric contraction and relaxation time to the ejection time. Patients were considered FB responders when an increase in cardiac index (CI) of > 10% was observed.

Results: 24 patients were included in the analysis. No significant correlation was observed between the changes in central venous pressure (CVP) during FB and LV − MPI (r = 0.12, p = 0.67). Thirteen patients did not have any significant increase in CI after FB (2.4 ml/min/m2 to 2.3 ml/min/m2, P = 0.83) despite a significant increase in CVP (7.5 [2.5–12] mmHg to 10 [8.1–15.2] mmHg, P = 0.04). In these patients, no significant change in LV − MPI was observed (0.47 [0.43–0.61] to 0.48 [0.41–0.68], p = 0.68) (Fig. 1). In 11 patients, FB increased CI (1.86 [1.41–2.77] ml/min/m2 to 2.27 ml/min/m2 [1.82–3.33], p < 0.01) without statistically significant changes in central venous pressure (3.1 [1.0–8.3] mmHg to 4.0 [2.0–9.5], p = 0.44). In these patients, a trend for decreasing LV − MPI was observed (0.47 [0.36–0.48] to 0.34 [0.28–0.41], p = 0.17) (Fig. 1).

Conclusions: Fluid bolus (FB) induced acute preload changes which did not affect LV − MPI. Decreases in LV − MPI observed during FB may be associated with heart function improvement.

figure cn

Fig. 1 (abstract P209). Changes in left ventricular myocardial performance index (LV − MPI) in fluid bolus (FB) non-responders and responders


Monitoring of fluid responsiveness using mitochondrial oxygenation after cardiac surgery

B Hilderink1, R F Crane1, B Van den Bogaard1, J Pillay2, N.P. Juffermans3

1OLVG, Intensive Care, Amsterdam, Netherlands, 2University Medical Center Groningen, Intensive Care, Groningen, Netherlands, 3Erasmus University Medical Center, Laboratory of Intensive Care, Rotterdam, Netherlands

Critical Care 2023, 27(S1): P210

Introduction: Fluid responsiveness is conventionally defined at the level of the heart. The aim of volume expansion is to improve organ perfusion and oxygenation by increasing cardiac output. However, after cardiac surgery, there is marked dissociation between cardiac output and tissue oxygenation due to the deleterious effects of cardiopulmonary bypass on the microcirculation. We investigated the utility of mitochondrial oxygen tension (mitoPO2) to guide fluid therapy in patients after cardiac surgery by measurements of cardiac output in parallel with mitoPO2. MitoPO2 is a novel bed-side quantitative marker which measures the balance between oxygen delivery and consumption in the skin.

Methods: Patients undergoing cardiac surgery with cardiopulmonary bypass were included. Cardiac output was measured using non-invasive finger cuff pulse contour analysis (Nexfin, BMEYE ®, Amsterdam). Fluid responsiveness was defined as an increase in CO of > 10% after a fluid bolus of > 500 ml in 1 h. MitoPO2 was measured before and 15 min after the administration of a fluid bolus using the COMET device (Photonics healthcare, Utrecht, The Netherlands).

Results: 44 patients after cardiac surgery were included in whom 46 fluid boluses were investigated. Of these, 13 (28%) were fluid responders. In fluid responders, mitoPO2 increased by 7.6 (1.4–9.7) mmHg whereas in fluid non-responders, mitoPO2 decreased by 4.8 mmHg (p = 0.004). Mitochondrial oxygen consumption (mitoVO2) increased in fluid responders 0.52 (−0.15 to 1.5) mmHg/s compared to non-responders (p = 0.02). Mean arterial pressure, core-toe temperature gradient, lactate did not change when compared to fluid non-responders. The AUROC of mitoPO2 to monitor fluid responsiveness was 0.78 (p < 0.05).

Conclusions: Changes in mitoPO2 and mitoVO2 are associated with fluid responsiveness and can potentially be used to guide fluid resuscitation after cardiac surgery.


The increase in cardiac output induced by a decrease in positive end-expiratory pressure reliably detects volume responsiveness: the PEEP-test study

C Lai1, R Shi1, A Beurton2, F Moretto1, N Fage1, S Ayed 1, M Dres2, JL Teboul1, X Monnet1

1AP-HP, Service de Médecine Intensive-Réanimation, Hôpital de Bicêtre, DMU 4 CORREVE, Inserm UMR S_999, FHU SEPSIS, CARMAS, Université Paris-Saclay, Le Kremlin-Bicêtre, France, 2AP-HP, Groupe Hospitalier Universitaire APHP-Sorbonne Université, site Pitié-Salpêtrière, Service de Pneumologie, Médecine intensive Réanimation (Département R3S), Paris, France

Critical Care 2023, 27(S1): P211

Introduction: In patients on mechanical ventilation, positive end-expiratory pressure (PEEP) can decrease cardiac output through a decrease in cardiac preload and/or an increase in right ventricular afterload. We hypothesized that a transient decrease in PEEP (PEEP-test) may be used as a test to detect volume responsiveness.

Methods: Mechanically ventilated patients with PEEP ≥ 10 cmH2O (“high level”) and without spontaneous breathing were prospectively included. Volume responsiveness was assessed by a positive passive leg raising (PLR) test, defined as an increase in cardiac index (CI) during PLR ≥ 10%. The PEEP-test consisted in reducing PEEP from the high level to 5 cmH2O for one minute. Changes in CI were monitored with pulse-contour-derived CI (PiCCO2).

Results: We included 64 patients among whom 31 were volume responsive. The increase in CI during PLR was 14%(11–16%). The PEEP at baseline was 12(10–15) cmH2O and the PEEP-test resulted in a median decrease in PEEP of 7(5–10) cmH2O, without difference between volume responsive and unresponsive patients. Among volume responsive patients, the PEEP-test induced a significant increase in CI of 16%(12–20%) (from 2.4 ± 0.7 to 2.9 ± 0.9 L/min/m2, p < 0.001) compared to volume unresponsive patients. In volume unresponsive patients, PLR and the PEEP-test increased CI by 2%(1–5%) and 6%(3–8%). Volume responsiveness was predicted by an increase in CI > 8.6% during the PEEP-test with a sensitivity of 96.8%(95% confidence interval (95% CI) 83.3–99.9%) and a specificity of 84.9% (95% CI 68.1–94.9%). The area under the receiver operating characteristic curve of the PEEP-test for detecting volume responsiveness was 0.94 (95% CI 0.85–0.98) (p < 0.001 vs. 0.5). Spearman’s correlation coefficient between the changes in CI induced by PLR and the PEEP-test was 0.76 (95% CI 0.63–0.85, p < 0.001).

Conclusions: A CI increase > 8.6% during a PEEP-test, which consists in reducing PEEP to 5 cmH2O, reliably detects volume responsiveness in mechanically ventilated patients with a PEEP ≥ 10 cmH2O.


Systolic volume variation and pulse pressure variation as fluid responsiveness predictors in patients in septic shock: assessment of limitations in a prospective study

T Petrucci1, A Costa2, D Silva1, G Nobre de Jesus1, J Ribeiro1

1University Hospital Santa Maria, Intensive Care Department, Lisbon, Portugal, 2Faculdade de Medicina de Lisboa, Clínica Universitária de Medicina Intensiva, Lisboa, Portugal

Critical Care 2023, 27(S1): P212

Introduction: Fluid resuscitation is a key stone in septic shock (SC) treatment. Nonetheless, assessment of fluid responsiveness (FR) has many confounders. Systolic volume variation (SVV) and pulse pressure variation (PPV) are validated as surrogate markers for FR, but some conditions limit their application: low tidal volume (VT) (< 8 ml/kg), right heart (RH) disfunction, high intra-abdominal pressure (IAP), low compliance, dysrhythmia, and patient triggered respiratory cycles. We conducted a prospective pragmatic study with the aim of characterizing the prevalence of these limitations in a cohort of patients with SC.

Methods: Patients were included in the first 24 h after SC recognition, identified by hyperlactacidemia and vasopressor dependence associated with presumed or declared infection. Hemodynamic (HD) data was gathered whenever a FR assessment was performed. Demographic, clinical and laboratory data were also collected.

Results: Twenty-three patients were included (Table 1), of which 20 had a minimally invasive HD monitoring device. There were limitations in 76.2% of the patients: 9 had a VT < 8 ml/kg, 9 had RH disfunction, 3 had high IAP, 2 had low compliance, 6 had dysrhythmia and 10 patients had patient triggered respiratory cycles. About 56.2% of the patients only had one limitation, 25% had two limitations and 18.8% had three. ICU mortality was 45%. Table 1 summarizes other important data collected.

Conclusions: In this preliminary study, the use of SVV or PPV to assess fluid responsiveness in patients with septic shock had many limitations. Most were related to invasive mechanical ventilation or right heart dysfunction. Albeit SVV and PPV are validated variables to predict FR in patients in SC, their applicability is limited, and our data suggests they may be of little use in a large percentage of ICU population. Intensivists should be aware of these conditions when they choose minimally invasive HD monitoring to guide fluid therapy in SC.

Table 1 (abstract P212). Additional data


Mean value

Age (years)

66.3 ± 10.2


61.7 ± 15.6

SOFA at ICU admission

10.3 ± 6.6

Initial lactate (mmol/l)

4.7 ± 3.1


Heart rate variability as a surrogate for fluid responsiveness in patients in septic shock

A Costa1, T Petrucci2, S Mendes Fernandes1, D Silva2, G Nobre de Jesus3, JM Ribeiro2

1Faculdade de Medicina de Lisboa, Clínica Universitária de Medicina Intensiva, Lisboa, Portugal, 2University Hospital Santa Maria, Intensive Care Department, Lisbon, Portugal, 3Faculdade de Medicina de Lisboa, Intensive Care Department, Lisboa, Portugal

Critical Care 2023, 27(S1): P213

Introduction: Fluid resuscitation is paramount in septic shock (SC) treatment and should be performed when hemodynamic benefits likely outweigh the risks of fluid overload. Usually, benefits are based on fluid responsiveness (FR) assessment. Defining FR is essential to avoid iatrogenesis. We hypothesized that heart rate (HR) variation, evaluated before and after a fluid challenge, might be a useful tool to validate fluid responsiveness.

Methods: The study was performed in a large urban, university hospital. Patients were included in the first 24 h after SC recognition, identified by hyperlactacidemia and vasopressor dependence associated with presumed or declared infection. Whenever clinical judgement recommended, fluid challenge was performed (500 mL crystalloid infusion in 30 min), without other interventions. FR data was evaluated with echocardiography and considered to be present when there was a 10% increase in cardiac output (CO) or left ventricle outflow tract-velocity time integral (VTI). An average of three VTI and peak velocity measures was used. Demographic, clinical and laboratory data, and other hemodynamic variables were collected, including HR prior and after FR.

Results: Preliminary assessment of data was performed in 23 patients. Three of which were excluded due to poor acoustic TE window. ICU mortality was 45%. All patients were under noradrenaline (NA); 19 were under invasive mechanical ventilation, 2 with non-invasive ventilation and 2 were spontaneously breathing. Table 1 summarizes other important data collected. Using the aforementioned FR definition, we identified 12 responsive and 8 nonresponsive patients. Mean HR variation was -3% (± 9.6; −18 to 22%). In responsive patients, there was no correlation between ΔVTI% and ΔHR% (p = 0.235) nor between ΔVTI% and ΔHR% if initial HR > 100 (p = 0.114).

Conclusions: In this pragmatic study, preliminary data suggests that ΔHR does not correlate with ΔVTI and cannot predict fluid responsiveness.

Table 1 (abstract P213). Additional data


Mean value

Age (years)

66.3 (± 10.2)


61.7 (± 15.6)

SOFA at admission

10.3 (± 6.6)


0.54 mcg/kg/min (± 0.76)

Initial lactate

4.73 mg/dl (± 28.6)

Initial HR

88.8 (± 22.2)

Final HR

85.2 (± 19.4)


Mottling in septic shock: skin colour matters!

S Jog, L Vikram

Deenanath Mangeshkar Hospital and Research Center, Intensive Care Medicine, Pune, India

Critical Care 2023, 27(S1): P214

Introduction: Skin mottling as a peripheral perfusion marker in septic shock is correlated well with severity and outcome [1–3]. Skin mottling is well studied in Caucasian population and its validity as a clinical sign in dark skinned ethnic population is not known. Our aim was to study the mottling in septic shock in the Indian ethnic population who has different skin colour as compared to Caucasian population.

Methods: We conducted a prospective observational study of consecutive patients with skin colour category 21–34 on von Luschan scale or Fitzpatrick type IV and V who had septic shock needing high dose of norepinephrine ≥ 0.2 mcg/kg/min after fluid optimisation.The study was conducted in a mixed medical—surgical ICU over a span of 12 months. Two experts (a dermatologist and a cosmetic surgeon) independently classified the skin type and scored mottling in our patients. We recorded the demographics, hemodynamic variables, mottling score and observed for the incidence of mottling and its correlation with predictors of severity of septic shock.

Results: We included 108 patients. Mean age was 61.3 years. Mean SOFA score and APACHE II score at enrolment was 10.3 and 21.9 respectively. Overall mortality was 72%. Incidence of mottling was 22/86 (20.3%). Development of mottling was significantly associated with APACHE II score > 22, SOFA score > 10 and norepinephrine dose > 0.3 mcg/kg/min (Table 1). Occurrence of mottling irrespective of its grade was strongly associated with mortality 19/22 (86.3%) (p = 0.028).

Conclusions: Incidence of mottling in septic shock is much less in Indian ethnicity with brown skin colour as compared to Caucasian population. Occurrence of mottling in this setting is associated with very high mortality.


  1. 1.

    Ait-Oufella H et al. Intensive Care Med. 2011;37:801–7.

  2. 2.

    Ferraris A et al. Indian J Crit Care Med. 2020;24:672–676.

  3. 3.

    Dumas G et al. Crit Care. 2019;23:211.

Table 1 (abstract P214). Variables associated with development of mottling


Yes (n = 22) Mean ± sd

No (n = 86) Mean = /−sd


p value


58.5 ± 15.41

62.06 ± 15.68

− 0.953



65.96 ± 11.01

66.76 ± 10.07

− 0.325


SOFA score

11.59 ± 3.13

9.93 ± 2.94




25.05 ± 7.07

21.15 ± 7.1



Norepinephrine dose

0.37 ± 0.12

0.29 ± 0.12



Duration of vasopressor support

6.32 ± 6.07

6.13 ± 4.26




Assessing diuresis response in stable cardiogenic shock

SA Narra, H Batool, S Isha, S Jonna, K Singh, A Jena, D Sanghavi, A Bhattacharyya

Mayo Clinic, Critical Care Medicine, Jacksonville, USA

Critical Care 2023, 27(S1): P215

Introduction: Optimizing preload via diuresis and dialysis is one of the prime modalities for the management of cardiogenic shock. There is little consensus in the ICU on how to monitor patients’ readiness for fluid removal beyond the initial decompensated phase of cardiogenic shock. Echocardiography-based tools help in assessing hemodynamic status. We look at various Echo and Lab parameters to see if fluid removal can be dictated by them.

Methods: A retrospective single-center observational study was performed on 69 patients diagnosed with cardiogenic shock and on mechanical circulatory support. We collected echocardiographic parameters, lab values, vital signs, medications, and outcomes for 24 h from the day that echocardiograms were performed. This data was analyzed to study the association with bad outcomes. A bad outcome was defined by—fluid administration after diuresis, or an increase in vasopressor requirements. We used R 4.2.2 for statistical analysis and chi-square test or Fischer’s exact test used to calculate p values.

Results: Most of the patients in our cohort were males (78.3%) and of white (58%) ethnicity. 13% of them were on IABP, and 87% were on Impella devices, with Impella 5.5 being the most common (68.3%). 80% of our total cohort survived the hospital stay. Our analysis of 242 serial observations revealed that patients in the bad outcome category had more negative fluid balance [(− 995 ± 2040 vs. − 96 ± 1760); p < 0.001], and a higher prevalence (28.2% vs. 18.6%, p = 0.011) of reduced IVC collapsibility (< 50%) than the better outcome group. Relevant results are demonstrated in Table 1.

Conclusions: Beyond initial fluid optimization, echo-guided determination of volume status with euvolemic states is associated with better outcomes. Aggressive attempts to obtain a negative fluid balance is associated with poor short-term outcomes. Further prospective trials are required to determine causality.

Table 1 (abstract P215). Results


Bad outcome group; Mean ± SD/%

Good outcome group; Mean ± SD/%

p value

IVC collapsibility reduced (< 50%)




RA pressure, mmHg

11.90 ± 5.80

12.70 ± 5.19


RV pressure, mmHg

46.00 ± 12.50

46.80 ± 13.20


Cardiac index, l/min/m2

2.59 ± 1.11

2.19 ± 0.98


Input–output, mL

−995 ± 2040

−96 ± 1760

 < 0.001

BUN/Cr ratio

23.30 ± 19.80

20.20 ± 10.80


Bicarbonate, mmol/l

25.40 ± 9.22

24.10 ± 3.85


  1. Bad outcome group = fluid administration after diuresis, or an increase in vasopressor requirements


Diastolic shock index in sepsis and septic shock: A predictive factor of mortality

W Bahria, F Hamdi, W Demni, B Fradj, S Hayouni, S Drira, M Bouraoui, N Nouira

Mongi Slim Academic Hospital, Emergency Department, La Marsa, Tunisia

Critical Care 2023, 27(S1): P216

Introduction: Hypotension results from an interaction between vasodilatation, myocardial dysfunction and blood flowdisturbance in patients with sepsis and septic shock. In this cases, the diastolic blood pressure (DBP) and theheart rate (HR) may be usefull to reflect the hypotension conditions. We aimed to evaluate The relationshipbetween the diastolic shock index (DSI) and mortality in the ED.

Methods: This observational prospective and analytical study was conducted over three years in an universal hospitalincluding patients admitted to the ED for sepsis and septic shock. DSI defined as a ratio of HR to DBP (HR/DBP)and SSI defined as a ratio of HR to SBP (HR/DBP) were calculated for all patients on admission. A cutt-off pointof SSI > 0.9 and DSI > 1.4 were used. The primary end-poind was the intra-hospital mortality. ROC curves were used to identify variables predicting mortality.

Results: Two hundred and four patients were included in our study. One hundred and thirty three patients (65%) were in sepsis and 71 patients (35%) were in septic shock. The mean age was 64 years old [17–94] with a sex ratio of 1.08. SSI > 0.9 and the DSI > 1.4 are significantly associated with mortality in the ED with respectively (p = 0.000; OR = 2.3; 95% CI [1.57–3.59]) and (p = 0.004; OR = 1.8; 95% CI [1.18–3.00]). The SSI and DSI area underthe curve for in-hospital mortality were respectively (p = 0.000; OR = 0.725; 95% CI [0.64–0.80]) and (p = 0.000; OR = 0.68; 95% CI [0.60–0.76]).

Conclusions: DSI and SSI calculated on admission to the ED might identify patients with sepsis or septic shock at high risk ofdeath. In this study, DSI didn’t show a significant superiority to SSI for predicting mortality. This funding deserves more research in multicentre studies to be useful.


Predictive value of intra-hospital mortality of "age-shock index" and "age-modified shock index" in ST-segment elevation myocardial infarction patients (STEMI)

Y Walha1, D Hamdi1, S Drira1, F Lazez1, B Fradj1, M Boussen1, F Azaiez2, N Nouira1

1Mongi Slim Academic Hospital, Emergency Department, La Mara, Tunisia, 2Mongi Slim Academic Hospital, Cardiology Department, La Mara, Tunisia

Critical Care 2023, 27(S1): P217

Introduction: Early recognition of patients at high risk of intra-hospital death, will allow better, faster and more rigorous therapeutic management. The association between A-SI (age-shock index) or A-mSI (age-modified shock index) and short-term mortality in STEMI patients has not yet been sufficiently examined. The aim of this study was to prove the predictive values of A-SI and A-mSI for intra-hospital mortality in patients with myocardial infarction.

Methods: Two hundred STEMI patients admitted to the emergency department between January 2022 and November 2022 were analyzed in a retrospective cohort study. A-SI and A-mSI were calculated for all patients, the primary outcome was intra-hospital mortality.

Results: During the study period, 200 patients with STEMI were admitted to the emergency room. The mean age was 59 ± 12 years. The sex ratio was 4.52. Only 20% of patients consulted within the hour, the median consultation time was 3 h with extremes of up to 48 h, the rates of thrombolysis and primary ATC were respectively 63.5% and 22.5%. For other patients, 13% consulted late with STEMI semi recent and two cases of myocardial infarction spontaneously reperfused were hospitalized in the cardiology intensive care unit awaiting angioplasty. Intra-hospital mortality was 8%. The predictive value of SI and mSI mortality were tested statistically but no significant results were found. However by testing A-SI and A-mSI, the following significant results were found: A-SI (p = 0.01, odds ratio (OR) = 12.29, area under the curve (ROC-AUC) = 0.68, p = 0.018, CI[0.53–0.83]); A-MSI (OR = 16.13, p = 0.03, ROC-AUC = 0.70, p = 0.009), CI[0.54–0.85]).

Conclusions: Both A-SI and A-mSI could predict in-hospital mortality of patients presenting to the emergency department with a myocardial infarction.


The prognostic value of reverse shock index in patients with sepsis at the emergency department

W Demni, F Hamdi, W Bahria, Y Zouaghi, K Jemai, F Lazzez, Z Farhani, N Nouira

Mongi Slim Academic Hospital, Emergency Department, La Marsa, Tunisia

Critical Care 2023, 27(S1): P218

Introduction: The “reverse shock index” (rSI) defined by the ratio of systolic blood pressure (SBP) to heart rate, can be quickly calculated to identify the prognosis of trauma patients mainly. In this study, we sought to evaluate the relationship between the rSI and in-hospital mortality in patients with sepsis at the emergency department (ED).

Methods: This was a prospective and analytical study, conducted over 9 months (February-October 2022). We included all patients with sepsis aged over 18 years and admitted to the ED. RSI and quick sequential organ fail assessment (qSOFA) were calculated at admission. We have determined the predictive factors of in-hospital mortality in uni and multivariate analysis. The area under the ROC curve of the rSI was analyzed.

Results: We enrolled 185 patients. The average age was 64 ± 16 years. The Sex ratio was 1.17. In-hospital mortality was 31.4%. On admission: 33% of the patients had a GCS < 15, 41% had SBP ≤ 100 mmHg, and 44% had a breathing rate ≥ 22 cpm. Half of the patients studied had a qSOFA ≥ 2, 69% were in sepsis and 31% were in septic shock. Predictive factors of in-hospital mortality in univariate analysis were: GCS < 15 (p < 0.05;OR = 2.85;95% CI = [1.86–4.36]), SPB ≤ 100 mmHg (p < 0.05; OR = 2.59;95% CI = [1.66–4.04]), qSOFA ≥ 2 (p < 0.05;OR = 2.30; 95% CI = [1.40–3.77]), and rSI < 1 (p < 0.05;OR = 2.70; 95% CI = [1.75–4.16]). In multivariate analysis, predictive factors of in-hospital mortality were: GCS < 15 (p < 0.002;OR = 4.7;95% CI = [1.75–12]), qSOFA ≥ 2 (p < 0.01;OR = 3.38;95% CI = [1.69–6.74]), and rSI < 1 (p < 0.001;OR = 4.8;95% CI = [2.43–9.58]). ROC curve analysis showed that the area under the curve of rSI < 1 was 0.68 (p < 0.05;95% CI = [0.59–0.77]).

Conclusions: Our study shows that the rSI calculated early in the ED appears to be a predictive factor of in-hospital mortality in patients with sepsis. The usefulness of rSI in directing therapeutic procedures in the management of sepsis should be examined in future studies.


Angiotensin II for the treatment of vasodilatory shock in patients with acute kidney injury

A Zarbock1, L Forni2, J Hästbacka3, E Korneva 4, S Herzig4, G Landoni5, P Pickkers6, R Bellomo7

1University Hospital Münster, Department of Anesthesiology and Critical Care, Münster, Germany, 2Royal Surrey Hospital & Faculty of Health Sciences, University of Surrey, Department of Critical Care, Guildford, UK, 3Helsinki University Hospital and University of Helsinki, Department of Perioperative, Intensive Care and Pain Medicine, Helsinki, Finland, 4Paion AG, Development and Regulatory Affairs, Aachen, Germany, 5IRCCS San Raffaele Hospital and University, Department of Anesthesia and Intensive Care, Milan, Italy, 6Radboud University Medical Center, Department of Intensive Care Medicine, Nijmegen, Netherlands, 7Intensive Care Medicine, The University of Melbourne, Melbourne, Australia

Critical Care 2023, 27(S1): P219

Introduction: Angiotensin II may improve patient outcomes in patients with vasodilatory shock associated with acute kidney injury (AKI) [1]. In preparation for a large phase 4 randomised controlled trial we aimed to identify an optimal target population for this vasopressor.

Methods: Post-hoc analysis of the data from the Phase 3 Angiotensin II for the Treatment of High-Output Shock trial was performed in a sub-population of patients with AKI to evaluate the effect of angiotensin II on mortality and renal-replacement therapy (RRT) utilisation. AKI was defined using the Kidney Disease Improving Global Outcomes (KDIGO) criteria. For comparisons between treatment arms, Fisher’s exact test or the chi-square test or the two-sample Wilcoxon rank-sum test was applied.

Results: From a total of 321 patients with distributive shock, 203 patients (63%) had AKI (stage 1: 25 patients (12.3%), stage 2: 42 patients (20.7%), stage 3: 136 patients (67%)). Among AKI patients 98 received angiotensin II and 105 received placebo. For combined AKI stages 2 and 3, the 28-day mortality rate was 51% (95% CI 40–62%) with angiotensin II vs 65% (95% CI 54–75%) with placebo (p = 0.068); this outcome became significant in patients with AKI stage 3 (48% (95% CI 35–61%) with angiotensin II vs. 67% (95% CI 55–78%) with placebo (p = 0.025)). No statistically significant difference was observed for the patients with AKI stage 1 or 2 or AKI overall. In the subgroups of the patients with AKI stage 3, stages 2 and 3, and AKI overall, the mean number of RRT free days up to Day 7 was significantly higher in the angiotensin II group; the same trend was observed for RRT duration (Fig. 1).

Conclusions: In the trial the majority of the patients had AKI at baseline. Angiotensin II treatment resulted in decreased RRT utilization in patients with AKI stages 2 and 3 and in reduced mortality in patients with AKI stage 3. These results introduce a target population for future trials.


  1. 1.

    Tumlin J et al. Crit Care Med 2018;46:949–957.

figure co

Fig. 1 (abstract P219). Mean number of RRT-free days and RRT-duration (days) up to Day 7 with 95% confidence intervals. Patients who died up to Day 7 were counted as patient on RRT for 7 days. Subgroup 1—AKI, stage 1 at baseline; 2—AKI, stage 2 at baseline; 3—AKI, stage 3 at baseline; 4—AKI stage 2 and 3 at baseline; 5—AKI overall at baseline. *p value


Rates and complications of peripheral vasopressor use in the CLOVERS trial

ES Munroe1, IN Co2, IS Douglas3, RC Hyzy1, K Nelson1, PK Park4, TW Rice5, S Seelye1, NI Shapiro6, HC Prescott1

1University of Michigan, Department of Internal Medicine, Ann Arbor, USA, 2University of Michigan, Department of Emergency Medicine, Ann Arbor, USA, 3Denver Health Medical Center and University of Colorado, Pulmonary Science and Critical Care Medicine, Denver, USA, 4University of Michigan, Department of Surgery, Ann Arbor, USA, 5Vanderbilt University, Division of Allergy, Pulmonary, and Critical Care Medicine, Nashville, USA, 6Beth Israel Deaconess Medical Center, Department of Emergency Medicine, Boston, USA

Critical Care 2023, 27(S1): P220

Introduction: Vasopressors are traditionally administered via central venous catheters (CVCs) given concern for extravasation and resultant tissue injury with peripheral administration. However, vasopressor infusion via peripheral IV may be safe and obviate the need for CVC placement in some patients. Little is known about how peripheral vasopressors are used in practice or how safety data from existing small, single-center studies extrapolate to routine care.

Methods: We conducted a post hoc analysis of peripheral vasopressor use in the Crystalloid Liberal or Vasopressors Early Resuscitation in Sepsis (CLOVERS) trial, which enrolled 1563 patients with early sepsis-induced hypotension in 60 US hospitals (3/2018–2/2022). Trial protocol allowed vasopressor administration through central access or a large peripheral IV, per discretion of the treating team, with vasopressor monitoring guided by local hospital protocols. We analyzed peripheral vasopressor use and assessed complications among patients treated with peripheral vasopressors and with CVC placement. Complications were graded on a 5-point scale from no clinical effect (1) to death (5).

Results: Of 1563 patients enrolled in CLOVERS, 750 (48.0%) received vasopressors, including 500 (32.0%) who received vasopressors peripherally. 363 patients (23.2%) had a CVC placed within 72 h of study enrollment, including 234 (64.5%) who received peripheral vasopressors prior to CVC placement and 129 (35.5%) who did not. Use of peripheral vasopressors did not vary over the study period (p = 0.39 for trend). Complications occurred in 3/500 patients who received peripheral vasopressors (0.6%; 0 Grade 3 +), and in 14/363 patients who had a CVC placed (3.9%; 42.9% Grade 3 +).

Conclusions: In this large, multicenter trial, peripheral vasopressor use was common and complications were rare and low-grade, suggesting peripheral vasopressors can likely be used safely in clinical practice.

Acknowledgement: Submitted on behalf of the NHLBI PETAL Network.


Midodrine for the early liberation from vasopressor support in the ICU – LIBERATE

D Opgenorth1, S Kilcommons1, K Fiest2, J Kutsogiannis1, E Macintyre1, J Senaratne1, J Slemko1, X Wang3, SM Bagshaw1, OG Rewa1

1University of Alberta, Department of Critical Care Medicine, Edmonton, Canada, 2University of Calgary, Cumming School of Medicine, Calgary, Canada, 3Alberta Health Services, Health Services Statistical and Analytic Methods, Edmonton, Canada

Critical Care 2023, 27(S1): P221

Introduction: A pilot study to determine the feasibility of conducting an adequately powered clinical trial to evaluate the efficacy of the implementation of an oral vasoactive agent (midodrine) for early liberation of IV vasopressor support in ICU patients.

Methods: This was a single centre, concealed-allocation parallel-group blinded pilot RCT. Patients were randomly assigned in a 1:1 ratio to midodrine (enteral, 10 mg every 8 h) or placebo for the duration of their IV vasopressor therapy and 24 h following the discontinuation of their IV vasopressor therapy. The investigation product (IP) was administered orally, or by ng tube. Primary outcomes were feasibility of recruitment and duration of IV vasopressor support. Secondary outcomes included ICU length of stay, 90 day all-cause mortality, rates of ICU re-admission, rates of re-initiation of IV vasopressors and safety endpoints.

Results: 20 patients were enrolled into the study. One patient withdrew consent prior to data collection. 11 patients were assigned to midodrine and 9 to placebo. Median age was 60 (49.5–64.5) and 56.5 (43.2–68.2) respectively. 54.5% of the midodrine group were female vs 12.5% in the placebo group. Median APACHE II scores were 24.0 (19.5–27.5) and 17.5 (11.5–27.2). Treatment interventions measured included invasive mechanical ventilation (72.7% and 75.0%), CRRT (9.1% and 12.5%), IV hydrocortisone (36.4% and 37.5%) and blood products (45.4% and 37.5%). Outcomes measured included total duration of IV vasopressor support (hours) (48.6 (20.2–120.5) and 37.5 (21.4–81.0), vasopressors re-initiated over 24 h after ICU discharge (36.4% and 25.0%), ICU mortality (9.1% and 37.5%), hospital mortality (18.2% and 37.5%) and ICU length of stay 6.0 (2.5–14.8) and 4.2 (3.2–6.2) days. No adverse events related to the IP were identified.

Conclusions: This pilot study provides evidence that recruitment into a larger multi-site RCT is feasible to evaluate the role of midoddrine in earlier liberation from IV vasopressor therapy.


Effects of levosimendan on cardiovascular performance in critically ill patients

TY Kim, EM Bowcock, CF Duncan, RA Rowley, SR Orde

Nepean Hospital, Intensive Care Unit, Kingswood, Australia

Critical Care 2023, 27(S1): P222

Introduction: Cardiac dysfunction is a frequent complication in critically ill patients. Levosimendan is a unique inodilator that does not increase intracellular calcium concentration while relaxing vascular smooth muscle, hence, reducing preload and afterload of left ventricle (LV) and right ventricle (RV). There is still a lack of data assessing the trend of cardiovascular effects of levosimendan in critically ill patients.

Methods: Single centre, retrospective parallel analysis of 49 patients with non-septic cardiomyopathy (NC) and septic cardiomyopathy (SC). Echocardiogram was performed median time of 1 day prior to levosimendan infusion and median time of 3 days after levosimendan infusion. Physiological markers were collected 24 h before and after therapy. Echocardiographic parameters of LV and RV function were measured by cardiac sonographers and reported by intensivists with advanced echocardiography qualifications. These values were compared between NC and SC.

Results: Prior to the commencement of levosimendan infusion, patients received median amount of 46 ml/kg fluid resuscitation and 40% were on at least two cardiovascular support agents. pH and lactate improved after levosimendan therapy (7.3 vs 7.4, p < 0.01; 2.2 vs 1.3 mmol/l, p 0.01). LV ejection fraction, LV outflow tract velocity time integral (LVOT-VTI) and RVOT-VTI were higher post-therapy (31 vs 44%, p < 0.01; 12 vs 15 cm, p < 0.01; 7 vs 10 cm, p < 0.01). Tricuspid annular plane systolic excursion increased following therapy (15 vs 17 mm, p < 0.01). RVOT-VTI/pulmonary artery systolic pressure also improved (0.20 vs 0.25 cm/mmHg, p 0.03). Improvement in LV systolic function were more prominent in NC than SC.

Conclusions: Early use of levosimendan infusion may be helpful in improving LV and RV systolic function in critically ill patients, especially in NC. It may also increase pulmonary artery compliance. Further prospective echocardiographic studies are required to assess its therapeutic benefit amongst pre-defined cardiac dysfunction phenotypes.


Predictors for failure of early recovery after totally endoscopic aortic valve replacement (TEAVR)

S Lambrecht1, J Dubois2, L Heremans1, D Vranken1, I Gruyters1, E Geerts1, I De Pauw1, L Geebelen1, B Stessel1, J Vandenbrande1

1Jessa, Anesthesiology & Intensive Care, Hasselt, Belgium, 2Jessa, Critical Care, Hasselt, Belgium

Critical Care 2023, 27(S1): P223

Introduction: We investigated the association between perioperative variables and failure for Enhanced Recovery After Cardiac Surgery (ERACS). ERACS is gaining importance since it has been shown to be safe and cost-reducing. Identification of risk factors for ERACS Failure (EFa) is useful in the optimization of individual patients on one hand and in the further expansion of ERACS on the other.

Methods: We performed a retrospective analysis on our TEAVR database. All patients with EuroSCORE II ≤ 3 are enrolled in our ERACS pathway. Postoperatively, patients are admitted to the Intensive Care Department (ICU) while ventilated under dexmedetomidine sedation. Analgesia consists of the routine administration of acetaminophen and ketorolac (unless contra-indications) and supplemental opioids as needed. The primary purpose is to achieve extubation within 6 h after surgery, as stated in the ERACS guidelines. Hence, we defined EFa as intubation for more than 6 h postoperatively and/or ICU length of stay for more than 1 day. Based on the literature, 6 variables (Table 1) were analyzed as possible predictors for EFa. First, a univariate regression analysis was performed. Variables with p value < 0,1 were taken into the multivariate regression analysis.

Results: Between 10/2019 and 12/2021, 176 TEAVR patients were included in the ERACS pathway. The ERACS trajectory was unsuccesful in 55/176 (31%) patients: 20/55 (36%) were intubated ≥ 6 h postoperatively, 40/55 (73%) were not discharged on postoperative day 1 and 5/55 (9%) had a combination of both. After multivariate regression analysis, the need for revision surgery < 24 h appeared to be the main risk factor for unsuccessful ERACS after TEAVR, whereas a higher EuroSCORE II showed no significant association with EFa. Preoperative use of RAAS inhibition did not appear to be a risk factor for failure of ERACS.

Conclusions: Revision surgery within 24 h was the only identifiable risk factor for failure of the ERACS trajectory in our TEAVR patient population.

Table 1 (abstract P223). Results of the univariate and multivariate regression analyses on potential risk factors for failure of ERACS after TEAVR

Patients n = 176

Odds ratio [95% CI] univariate

p value univariate

Odds ratio [95% CI] multivariate

p value multivariate

Age (years)

1.00 [0.97, 1.03]




1.42 [0.95, 2.13]


1.43 [0.93, 2.19]


Revision surgery < 24 h

5.95 [1.12, 31.70]


6.02 [1.10, 33.10]


Coronary artery disease

0.83 [0.28, 2.46]



RAAS inhibitors

0.72 [0.37, 1.41]



BMI (kg/m2)

1.04 [0.97, 1.10]



Chronic kidney disease (eGFR < 50 ml/min)

1.78 [0.70, 4.51]


  1. Variables with univariate p value < 0.1 were taken into the multivariate regression analysis


Time for restoration of radial arterial pressure after a transient arm vascular occlusion correlates with outcome in critically ill patients: preliminary results of an observational study

C Lai, R Shi, JL Teboul, F Moretto, L Guérin, T Pham, X Monnet

AP-HP, Service de médecine intensive-réanimation, Hôpital de Bicêtre, DMU 4 CORREVE, Inserm UMR S_999, FHU SEPSIS, CARMAS, Université Paris-Saclay, Le Kremlin-Bicêtre, France

Critical Care 2023, 27(S1): P224

Introduction: Septic shock is characterized by an impairment of vascular reactivity as assessed by the resaturation slope of the muscular tissue oxygen saturation after a transient vascular occlusion (TVO). Another way to assess vascular reactivity might be to observe the time of the increase in radial arterial pressure after a TVO of the arm.

Methods: In mechanically ventilated patients equipped with a radial artery catheter, a brachial cuff was rapidly inflated to induce a transient arterial stop-flow. Arterial pressure decreased to the mean systemic pressure (Pmsarm). After a 60-s occlusion, the cuff was deflated and time to return from Pmsarm to baseline arterial pressure was measured (Trevasc).

Results: We included 59 patients, among whom 45 (76%) where in shock, including 22 (37%) with septic shock. Twenty-three (39%) had acute respiratory distress syndrome and 12 (20%) were intubated for neurological impairment. The dose of norepinephrine was 0.46 (0.19–0.93) μg/kg/min in patients with shock. Measurements were obtained 3 (1–5) days after onset of mechanical ventilation or vasopressors infusion. Mean arterial pressure was 83 ± 12 mmHg, Pmsarm was 28 ± 9 mmHg, central venous pressure was 13 ± 4 mmHg and the (Pmsarm-CVP) gradient was 15 ± 8 mmHg, with no difference between patients with or without shock. In the whole population, Trevasc was 26 (20–37) sec, with no difference between patients with or without shock (28(23–38) sec vs. 22(18–33) sec, respectively, p = 0.14). Among shock patients, Trevasc was longer in patients with septic shock than in the other ones (37 (29–44) vs. 23 (19–27) sec, respectively, p < 0.001). Trevasc was increased in the 24 (41%) patients who died compared to survivors (35 ± 19 vs. 27 ± 11 s, respectively, p = 0.03).

Conclusions: Trevasc was increased in patients with septic shock compared to patients with non-septic shock or without shock. Mortality was associated with higher Trevasc. This new variable may indicate vasoreactivity and be associated with mortality.


Adding an extension piece to the end of the purge side arm of the Impella device can prevent the incidence of the Impella device failure rate secondary to the cassette breaking and positively impact patient outcome

P Satashia1, J Blasavage1, A White1, S Isha1, S Paghdar1, S Kiley1, P Moreno Franco1, P Patel2, S Chaudhary1, D Sanghavi1

1Mayo Clinic Florida, Critical Care Medicine, Jacksonville, USA, 2Mayo Clinic Florida, Advanced Heart Failure and Heart Transplantation, Jacksonville, USA

Critical Care 2023, 27(S1): P225

Introduction: Impella 5.5® with Smart Assist is a type of minimally invasive left ventricular assist device (LVAD) approved by the Food and Drug Administration (FDA) for the treatment of ongoing cardiogenic shock for up to 14 days. The Impella® intends to reduce ventricular workload and provide the circulatory support necessary for myocardial recovery. The replacement of the purge tube led to the breakage of the luer lock connector leading to Impella device failure, and ultimately needed device replacements. The practice of adding the extension piece to the purge tube sidearm was implemented to avoid such incidents.

Methods: A retrospective chart review of ICU patients was done at a tertiary care center from August 2018 to August 2022 to assess the differences in patient outcomes related to Impella® Device utilization before and after the implementation of the extension piece to the purge tube sidearm. Among patients reviewed, a total of 18 were included in our review, with 5 not having the purge tube side arm extension added, while 13 patients had the extension.

Results: There was no significant difference in patient health outcomes between the two study groups. Additionally, there were no instances of device failure requiring explanation without the extension tubing. However, there were no cases of the purge cassette cracking with the addition of the extension tubing added.

Conclusions: The addition of extension tubing to the purge cassette of the Impella® Device did not impact patient health outcomes or affect the incidence of device failure. There was a complete reduction in the incidence of the purge cassette cracking, which could reduce the potential for infection or device failure over a long period of mechanical support. There is a need for long-term prospective studies to confirm the results.


Prehospital advanced life support by physician versus emergency medical service personnel after pediatric out-of-hospital cardiac arrest: results from a Japanese registry-based study

T Fukuda1, K Matsui2, M Amaki2, H Sekiguchi3

1Toranomon Hospital, Department of Emergency and Critical Care Medicine, Tokyo, Japan, 2Toranomon Hospital, Tokyo, Japan, 3University of the Ryukyus, Okinawa, Japan

Critical Care 2023, 27(S1): P226

Introduction: Controversy remains as to whether advanced life support (ALS) should be provided for pediatric out-of-hospital cardiac arrest (OHCA) in the prehospital setting. We sought to assess whether prehospital ALS by physician would be associated with an increased chance of neurologically favorable survival after pediatric OHCA compared with prehospital ALS by emergency medical service (EMS) personnel.

Methods: We conducted a nationwide population-based observational study using the Japanese government-led registry data of OHCA. This study included pediatric patients (1–17 years) who experienced OHCA in Japan from January 1, 2011, to December 31, 2019. The primary outcome was 1-month neurologically favorable survival with the Glasgow-Pittsburgh cerebral performance category score of 1 or 2.

Results: A total of 2532 patients were included (mean [SD] age, 11.3 [5.5] years; 65.5% male); 1054 received prehospital ALS by physician, and 1478 received prehospital ALS by EMS personnel. Among these patients, 187 (7.4%) survived with a favorable neurological status 1 month after OHCA, including 127/1054 (12.0%) for ALS by physician, and 60/1478 (4.1%) for ALS by EMS personnel. 757 patients receiving prehospital ALS by physician were propensity-matched with 757 patients receiving prehospital ALS by EMS personnel. In the propensity score-matched cohort, prehospital ALS by physician was associated with an increased chance of neurologically favorable survival compared with prehospital ALS by EMS personnel (8.9% vs. 4.1%; RR 2.16; 95% CI 1.43–3.27; p = 0.0002).

Conclusions: Among pediatric OHCA patients, prehospital ALS by physician was associated with an increased chance of neurologically favorable survival compared with ALS by EMS personnel in the prehospital setting.


AED delivery at night – can drones do the job? A feasibility study of unmanned aerial systems to transport automated external defibrillators during night-time

SS Scholz1, D Wähnert2, G Jansen3, O Sauzet4, E Latka5, S Rehberg1, KC Thies1

1Protestant Hospital of Bethel Foundation, University Hospital OWL, Department of Anaesthesiology, Intensive Care, Emergeny Medicine, Transfusion Medicine and Pain Therapy, Bielefeld, Germany, 2Protestant Hospital of Bethel Foundation, University Hospital OWL, Department of Orthopaedics and Trauma Surgery, Bielefeld, Germany, 3Klinikum Bielefeld, Department of Anaesthesiology, Intensive Care, Emergency Medicine and Pain Therapy, Bielefeld, Germany, 4University Bielefeld, Epidemiology and International Public Health, Bielefeld School of Public Health, Bielefeld, Germany, 5Studieninstitut für kommunale Verwaltung Westfalen-Lippe, Fachbereich Medizin und Rettungswesen, Bielefeld, Germany

Critical Care 2023, 27(S1): P227

Introduction: In their recent guidelines the European Resuscitation Council have recommended the use of Unmanned Aerial systems (UAS) to overcome the notorious shortage of AED. Exploiting the full potential of airborne AED delivery would mandate 24 h UAS operability. However, current systems have not been evaluated for nighttime use [1, 2]. The primary goal of our study was to evaluate the feasibility of night-time AED delivery by UAS. The secondary goal was to obtain and compare operational and safety data of night versus day operations.

Methods: We scheduled two (one day, one night) flights each to ten different locations to assess the feasibility of AED delivery by UAS during night-time. We also compared operational data and safety data of night versus day missions.

Results: All missions were completed without safety incident. The flights were performed automatically without interventions, apart from manually choosing the landing site and correcting the descent. Flight distances ranged from 910 to 6960 m, corresponding mission times from alert to AED release between 3:48 min and 11:20 min. Night missions (Tm:night = 7:26 ± 2:29 min) did not take longer than day missions (Tm:day = 7:59 ± 2:27 min). Despite inferior visibility of the target site at night, statistical analysis suggested non-inferiority of night landings versus day landings (p = 0.9).

Conclusions: Our results demonstrate the feasibility of UAS supported AED delivery during nighttime. Operational and safety data indicate no major differences between day- and night-time use. Future research should focus on integration of drone technology into the chain of survival.


  1. 1.

    Baumgarten MC et al. Resuscitation 2022;172:139–145.

  2. 2.

    Schierbeck S et al. Eur Heart J 2022;43:1478–1487.


Reliability and limits of bag valve mask (BVM): a comparative bench study

A Broc1, F Morin2, L Polard1, A Drouet3, H Schmit4, L Lahmaut5, M Taillantou-Candau6, F Beloncle6, JC Richard6, D Savary2

1Air Liquide Medical Systems, Med2Lab, Antony Cedex, France, 2University Hospital of Angers, Emergency Department, Angers, France, 3CHUV Centre Hospitalier Universitaire Vaudois, Emergency Department, Lausanne, Switzerland, 4Centre Hospitalier Annecy Genevois, Emergency Department, Annecy, France, 5SAMU de Paris and Intensive Care Unit, Necker Hospital, Assistance Publique-Hopitaux de Paris (APHP), Emergency Department, Paris, France, 6University Hospital of Angers, Medical ICU, Angers, France

Critical Care 2023, 27(S1): P228

Introduction: BVM is the most common device used for pre-oxygenation and ventilation during resuscitation. The objective is to compare the performances of 9 BVM in simulated pre-oxygenation and cardiopulmonary resuscitation (CPR) conditions in terms of inspired fraction of oxygen (FiO2), insufflated volume (VTi) and rebreathing.

Methods: 9 BVM were evaluated according to a specific bench protocol.

  1. 1.

    Two levels of spontaneous breathing (occurring during pre-oxygenation) were simulated on an ASL test lung at ZEEP and PEEP 5 with 15 L/min O2 for each BVM tested. Actual FiO2 was