Reference | 1. Findings on diagnosis | 2. Findings on transfusion | 3. Findings on mortality |
---|---|---|---|
Kaufmann 1997 [9] | 1. Of 69 patients, 45 were hypercoagulable (mean ISS 13.1) and seven were hypocoagulable (mean ISS, 28.6) by TEG® . Only one was hypocoagulable by elevated PT/aPTT, and two were hypercoagulable by elevated PLT | ||
2. Only ISS (P < 0.001) and TEG® (P < 0.05) predicted transfusion within the first 24 h after injury. Six of the seven hypocoagulable patients received blood within the first 24 hours | |||
3. None | |||
Watts 1998 [10] | 1. Hypothermic patients (34°C) presented significantly lower TEG® α-angle, K and MA values (P < 0.001) even though platelet count, PT, and aPTT were within normal range, and correlated with fluid and blood transfusion. | ||
2. None | |||
3. None | |||
Schreiber 2005 [11] | 1. INR and aPTT failed to detect early hypercoagulability, showing that TEG® is more sensitive. Women are more hypercoagulable than men within the first 24 hours. | ||
2. None | |||
3. None | |||
Rugeri 2007 [12] | 1. Significant correlation between PT - A15-EXTEM, between aPTT - CFT-INTEM, between fibrinogen - A10-FIBTEM and between PLT - A15-INTEM. A cut off value of A15-EXTEM at 32 mm and A10-FIBTEM at 5 mm presented a good sensitivity (87 and 91%) and specificity (100 and 85%) to detect PT >1.5 and a fibrinogen less than 1 g/L, respectively. | ||
2. None | |||
3. None | |||
Nekdulov 2007 [13] | 1. TBI patients had a lower PLT count (180 ± 68 × 109; mean ± SD) and longer bleeding time (674 ± 230 sec) than healthy controls (256 ± 43 × 109, p < .01) and (320 ± 95 sec, p < .005) respectively. TEG® -PM showed reduced PLT response to AA and ADP (0-86%, mean 22%) compared to healthy controls (57-89%, mean 73%). | ||
2. None | |||
3. None | |||
Levrat 2008 [14] | 1. MCF showed the best correlation with the ELT when compared with amplitude and CLI. HF patients also had greater ROTEM® abnormalities, lower INR, lower fibrinogen levels and were more severely injured (↑ ISS) than the control group (all p < .05). | ||
2. None | |||
3. Patients with hyperfibrinolysis had higher mortality rate (100%, CI: 48-100% vs. 11% CI: 5-20%) | |||
Park 2008 [57] | 1. None | ||
2. None | |||
3. Multiple logistic regression analysis identified MA as an independent risk factor for death, AUC ROC 0.961 (95% CI, 0.891, 1.000) | |||
Plotkin 2008 [15] | 1. Increased K time, reduced α-angle and decreased MA demonstrated hypocoagulation. | ||
2. INR, PT and aPTT did not correlate with the use of blood products (r = .57, p < .01). MA correlated with blood product use as well as PLT count. Patients with reduced MA used more blood products and had reduced PLT counts and hematocrit. | |||
3. None | |||
Carroll 2009 [16] | 1. TEG® parameters did not change significantly from the ED sampled to OR samples. | ||
2. Abnormal MA-ADP at 30 min correlated with the need for transfusion (p = .004). | |||
3. R and MA correlated importantly with fatality (both p < .001). HF was an independent predictor of fatality (p = .001 by chi square testing). | |||
Jeger 2009 [17] | 1. Strong correlations between the values of K, alpha angle and MA (p < 0.01). Moderate correlation between K and both INR and PLT count and between MA and both INR and PLT count (p < 0.05). There was decrease in the time for TEG® results with r-TEG® . | ||
2. None | |||
3. None | |||
Kashuk 2009 [49] | 1. None | ||
2. Lab tests triggers result in blood product administration in 73.1% compared with 53.9% based on r-TEG® thresholds (p = .03). FFP administration guided by INR triggers would have been higher (61.5% by INR triggers versus 26.9% by r-TEG® -ACT triggers, p = .003). | |||
3. None | |||
Kashuk 2009 [18] | 1. 67% of patients were hypercoagulable by r-TEG® . 19% of the hypercoagulable group suffered a TE, and 12% had TE predicted by prior r-TEG® . No patients with normal coagulability by r-TEG® had an event (p < .001). G value was the strongest predictor of TE after controlling for thromboprophylaxis (OR: 1.25, 95% CI: 1.12-1.39). For every 1 dyne/cm2 increase in G, the odds ratio of a TE increased by 25%. | ||
2. None | |||
3. None | |||
Park 2009 [19] | 1. PT and aPTT were prolonged compared with controls (p < .05). All other parameters showed hypercoagulability (low protein C, high fibrinogen level and low TAT levels). MA and α-angle were also higher compared with controls (p < .05). PT and aPTT in this population were increased and did not detect hypercoagulability, which was demonstrated by TEG® . | ||
2. None | |||
3. None. | |||
Schöchl 2009 [58] | 1. None | ||
2. None. | |||
3. Prolonged CFT and lower PLT contribution to MCF were associated with increased mortality (p = .042 and p = .026 respectively). The observed mortality was higher than the expected mortality as per TRISS (88 vs. 70%, p = .039). | |||
Doran 2010 [20] | 1. MCF was abnormal in all MTP cases. A10 is subsequently associated with an abnormal MCF. 64% of all patients were coagulopathic by TEM trace and only 10% had abnormal lab tests (p = .0005). | ||
2. None | |||
3. None | |||
Kashuk 2010 [21] | 1. 34% of injured patients requiring MT had PF (ANOVA, p < .0001). PF occurred early (median 58 minutes). Every 1 unit drop in G increased the risk of PF by 30% | ||
2. None | |||
3. The risk of death correlated significantly with PF (p = .026) and every 1 unit drop in G increased risk of death by 10%. | |||
Leemann 2010 [22] | 1. MT patients had significantly altered ROTEM® values on admission compared with non-MT patients. An increase in the CFT (p = .001), a shortening of the MCF (p < .001), and a shortening of the amplitude at all time-points (10/20/30 minutes) were observed in MT patients. | ||
2. Variables independently associated with MT included a hemoglobin level <10 g/dL and an abnormal MCF value (AUC ROC 0.831 [95% CI: 0.719–0.942). | |||
3. None | |||
Schochl 2010 [59] | 1. None | ||
2. None | |||
3. The difference in mortality, after excluded patients with TBI, was 14% observed versus 27.8% predicted by TRISS and 24.3% predicted by RISC. The study shows a favorable survival rate. | |||
Schochl 2011 [23] | 1. ROTEM® analysis revealed shorter clotting times in EXTEM and INTEM (p < .001), shorter CFT in EXTEM and INTEM (p < .0001), and higher MCF in EXTEM, INTEM, and FIBTEM (p < .01) in survivors compared with non-survivors, in severe isolated TBI. | ||
2. According to the degree of coagulopathy, non-survivors received more RBC (p = .016), fibrinogen concentrate (p = .01), and PCC (p < .001) within 24 h of arrival in the ED. | |||
3. Logistic regression analysis revealed EXTEM with cytochalasin D (FIBTEM) MCF and aPTT to have the best predictive value for mortality. | |||
Watters 2010 [24] | 1. Cloth strength baseline and at follow up were elevated in the splenectomy group and not in the control group (p < .01). Platelet count, fibrinogen, aPTT were also elevated in the splenectomy group. In this population TEG® and RSCT were able to diagnose hypercoagulability together. | ||
2. None | |||
3. None | |||
Cotton 2011 [25] | 1. Early r-TEG® values (ACT, k-time, and r-value) were available within 5 min. Late r-TEG® values (MA and α-angle) within 15 min, and RSCTs within 48 min (p < .001). ACT, r-value, and k-time showed strong correlation with PT, INR and aPTT whereas and α-angle correlated with platelet count (both p < .001). | ||
2. Linear regression demonstrated that ACT predicted RBC, plasma and PLT transfusions within the first 2 h of arrival. Controlling for all demographics and ED vitals, ACT > 128 predicted MT in the first 6 h. In addition, ACT < 105 predicted patients who did not receive any transfusions in the first 24 h. | |||
3. None | |||
Davenport 2011 [26] | 1. CFT, α, A5 and MCF are significantly different in the group with coagulopathy. A5 ≤ 35 mm detects great percentage of patients with coagulopathy with lower false positive rates than PT (detected 77% of ATC, with 13% false positive). | ||
2. Patients with A5 ≤ 35 mm were more likely to receive RBC (46% vs. 17%, p < .001) and FFP (37% vs. 11%, p < .001) transfusions. A5 identified patients who would require MT (rate of 71% vs. 43% for INR > 1.2, p < .001). | |||
3. None | |||
Davenport 2011 [50] | 1. None | ||
2. Coagulation profile deteriorated with low FFP:PRBC ratios <1:2. Maximal hemostatic effect was observed in the 1:2 to 3:4 groups: 12% decrease in PT (p = .006), 56% decrease CT (p = .047), and 38% increase in MCF (p = .024). Transfusion with ≥1:1 ratio did not confer any additional improvement. There was a marked variability in response to FFP, and hemostatic function deteriorated in some patients exposed to 1:1 ratios. The beneficial effects of plasma were confined to patients with coagulopathy. | |||
3. None | |||
Differding 2011 [27] | 1. R increased (p < .001) and α-angle decreased (p < .01) in both groups (patients and controls) as T°C decreased. Between groups, R, α-angle, and MA were significantly different at each T°C (p < .01), with patients being more hypercoagulable. R and α-angle were more affected by T°C in controls compared with patients (p < .02). Temperature did not alter coagulability in the range studied in trauma patients while in the controls it did change. | ||
2. None | |||
3. None | |||
Jansen 2013 [28] | 1. Repeated ROTEM® tests on samples stored at 37°C for a median of 51 minutes, show improved MCF (22 mm vs. 54 mm, p < .001) and α-angle (30.5 vs. 59.5°, p = .004) when compared to analysis at the moment of venipuncture. | ||
2. None | |||
3. None. | |||
Nystrup 2011 [29] | 1. Patients with a reduced MA (<50 mm) evaluated by TEG® , presented with a higher ISS - 27 (95% CI, 20–34) vs. 19 (95% CI, 17–22), than the rest of the cohort. | ||
2. MA correlated with the amount of RBC (p = .01), FFP (p = .04) and PLT (p = .03) transfused during the first 24 h of admission. | |||
3. Patients with ↓ MA demonstrated ↑ 30-day mortality (47% vs. 10%, p < .001). By logistic regression ↓MA was an independent predictor of ↑ mortality after adjusting for age and ISS. | |||
Ostrowski 2011 [30] | 1. Fibrinogen and PLT count were associated independently with clot strength in patients with ISS ≤ 26 whereas only fibrinogen was associated independently with clot strength in patients with ISS > 26. In patients with ISS > 26, adrenaline and sCD40L were independently negatively associated with clot strength. | ||
2. None | |||
3. None | |||
Schöchl 2011 [51] | 1. None | ||
2. RBC transfusion was avoided in 29% of patients in the fibrinogen-PCC group compared with only 3% in the FFP group (p < .001). Transfusion of PLT was avoided in 91% of patients in the fibrinogen-PCC group, compared with 56% in the FFP group (p < .001). | |||
3. Mortality was comparable between groups: 7.5% in the fibrinogen-PCC group and 10.0% in the FFP group (p = .69). | |||
Schöchl 2011 [31] | 1. EXTEM and INTEM CT and CFT were significantly prolonged and MCF was significantly lower in the MT group versus the non-MT group (p < .0001 for all comparisons). | ||
2. Of patients admitted with FIBTEM MCF 0 to 3 mm, 85% received MT. The best predictive values for MT were provided by hemoglobin and Quick value (AUC ROC: 0.87 for both parameters). Similarly high predictive values were observed for FIBTEM MCF (0.84) and FIBTEM A10 0.83). | |||
3. None | |||
Tauber 2011 [32] | 1. In patients with or without TBI, the prevalence of low fibrinogen, impaired fibrin polymerization and reduced MCF was 26%, 30%, and 22%, respectively, and thus higher than the prolonged INR (14%). All patients showed ↑ F1 + 2 and TAT and low AT levels, indicating ↑ thrombin formation. | ||
2. MCF FIBTEM correlated with RBC transfusion (OR 0.92, 95% CI 0.87–0.98). | |||
3. ROTEM® parameters correlated with RSCTs and with mortality (FIBTEM and EXTEM MCF (p = .006 and p = .001 respectively). EXTEM MCF was independently associated with early mortality. HF ↑ fatality rates and occurred as frequently in isolated TBI as in polytrauma. | |||
Theusinger 2011 [60] | 1. None | ||
2. None | |||
3. Mortality in the trauma HF group (77% – 12%) as diagnosed by ROTEM® was significantly higher than in the nontrauma HF group (41% – 10%, 95% CI 5%–67%) and the matched trauma group (33% – 10%, 95% CI 13%–74%). HF is significantly (p = .017) associated with mortality in trauma patients. | |||
Cotton 2012 [33] | 1. Controlling for ISS and BD on arrival, pre-hospital fluid was associated with a significant ↑in likelihood of HF. Each additional liter of crystalloid was associated with a 15% ↑ OR of HF. The in vitro model found that hemodilution to 15% of baseline and TF + t-PA was required to achieve an LY30 of 50%. | ||
2. None | |||
3. Compared with patients without HF, the HF group had higher mortality (76% vs. 10%); all p < .001. | |||
Cotton 2012 [34] | 1. The PE group had admission higher MA (66 vs. 63, p = .05) and higher ISS (median, 31 vs. 19, p = .002). When controlling for gender, race, age, and ISS, elevated MA at admission was an independent predictor of PE with an OR of 3.5 for MA > 65 and 5.8 for MA > 72. | ||
2. None | |||
3. None | |||
Davis 2013 [61] | 1. None | ||
2. None | |||
3. Median ADP inhibition of platelet function, as measured by TEG® platelet-mapping analysis, was significantly greater in TBI non-survivors (91.7%) compared to survivors (48.2%) (p = .035). | |||
Holcomb 2012 [35] | 1. Overall, r-TEG® correlated with RSCTs, and could replace RSCTs on admission. | ||
2. ACT-predicted RBC transfusion, and the α-angle predicted massive RBC transfusion better than PT, aPTT or INR (p < .001). The α-angle was superior to fibrinogen for predicting FFP transfusion (p < .001); MA was superior to PLT count for predicting PLT transfusion (p < .001); and LY-30 documented fibrinolysis. These correlations improved for transfused, shocked or TBI patients. | |||
3. None | |||
Ives 2012 [36] | 1. By the 6-h sampling, 8 (61.5%) of the HF patients (detected by TEG® parameters) had died from hemorrhage. Survivors at this point demonstrated correction of coagulopathy. | ||
2. Compared with patients without HF, patients with HF had a greater need for MT (76.9% vs. 8.7%; adjusted OR = 19.1; 95% CI, 3.6 - 101.3) | |||
3. On LR, HF was a strong predictor of early mortality (OR = 25.0; 95% CI, 2.8- 221.4), predicting 53% of early deaths. Patients with HF had ↑ early mortality (69.2% vs. 1.9%; adjusted OR = 55.8; 95% CI, 7.2-432.3) and in-hospital mortality (92.3% vs. 9.5%; adjusted OR = 55.5; 95% CI, 4.8 - 649.7). | |||
Jeger 2012 [52] | 1. None | ||
2. RSCTs correlate moderately with r-TEG® parameters (R: 0.44–0.61). Kaolin and r-TEG® were more sensitive than RSCTs and the r-TEG® α-angle was the parameter with the greatest sensitivity (84%) and validity (77%) at a cut-off of 74.7 degrees. When the r-TEG® α-angle was combined with HR >75 bpm, or hematocrit < 41%, sensitivity (84%, 88%) and specificity (75%, 73%) were improved. Cut-off points for transfusion can be determined with r-TEG® α angle and can provide better sensitivity than RSCTs. | |||
3. None | |||
Kashuk 2012 [37] | 1. INR at 6 h did not discriminate between survivors and non survivors (p = .10). | ||
2. In r = TEG® -guided transfusion, patients with a MRTG > 9.2 received significantly less components of RBCs, FFP, and Cryo (p = .048, p = .03, and p = .04, respectively | |||
3. r-TEG® G value was associated with survival as was MRTG and TG (p = .03). | |||
Kunio 2012 [62] | 1. None | ||
2. None | |||
3. In TBI patients, prolonged R time (>9 min) or reduced MA (<55 mm) as evaluated by TEG® , are associated with greater mortality (50% vs. 11.7% and 33.3% vs. 9.8%, respectively; p = .04). | |||
Kutcher 2012 [38] | 1. Patients with HF diagnosed by ROTEM® had lower T°C, pH, PLT count and higher INR, aPTT and D-dimer. The presence of hypothermia (temperature < 36.0°C), acidosis (pH < 7.2), relative coagulopathy (INR > 1.3 or aPTT > 30), or relative low PLT count (<200) identified HF by ROTEM® with 100% sensitivity and 55.4% specificity (AUC, 0.77). | ||
2. None | |||
3. HF as detected by ROTEM® was associated with MODS (63.2% vs. 24.6%, p = .004) and mortality (52.2% vs. 12.9%, p < .001). | |||
Nascimento 2012 [39] | 1. For detection of coagulopathy, overall, TEG® -R performed worse than INR. TEG® -R had a sensitivity of 33% (95% CI, 16%-55%), specificity of 95% (95% CI, 91%-98%), PPV of 47% (95% CI, 23%-72%), and NPV of 92% (95% CI, 87%-95%). An INR of 1.5 or greater had a sensitivity of 67% (95% CI, 45%-84%), specificity of 98% (95% CI, 96%-99.7%), PPV of 84% (95% CI, 60%-97%), and NPV of 96% (95% CI, 92%-98%). An INR of 1.3 or greater also had better sensitivity, PPV, and NPV, than TEG® . | ||
2. None | |||
3. None | |||
Ostrowski 2012 [53] | 1. None | ||
2. Patients considered coagulopathic (“endogenous heparinization”) based on TEG® parameters (R, K, α-angle and MA) received more RBC (10 vs. 0), FFP (7 vs. 0) and platelet (3 vs. 0) in the first 24 hours (p < .05). | |||
3. These patients showed a tendency towards higher 30-day mortality (50% vs. 16%, p = .15). | |||
Pezold 2012 [54] | 1. None | ||
2. INR, ISS, and G were predictors of MT. The predictive power for outcome MT did not differ among INR (adjusted AUC ROC = .92), aPTT (AUC ROC = .90, p = .41), or G (AUC ROC = .89, p = .39). | |||
3. 21% of patients died of MT-related complications. Age, ISS, SBP, and G were associated with MT-death. For outcome MT-death, G had the greatest adjusted AUC ROC (0.93) compared with the AUC ROC for BD (0.87, p = .05), INR (0.88, p = .11), and PTT (0.89; p = .19). | |||
Raza 2013 [55] | 1. None | ||
2. Patients with moderate and severe fibrinolytic activity, based on plasmin-antiplasmin complex levels and ROTEM® ML > 15%, required more transfusions: RBC (2.0 and 6.5 units, respectively), FFP (1 and 2.9 units, respectively), platelets (0.2 and 0.7 units, respectively) and cryoprecipitate (0.2 and 0.6 units, respectively) (p < .05 for all comparisons). | |||
3. Similarly, patients with moderate and severe fibrinolytic activity, had significantly greater 28-day mortality (12.1% and 40% respectively, p < .05). | |||
Rourke 2012 [40] | 1. ROTEM® parameters correlated with fibrinogen level, and ex vivo fibrinogen administration reversed coagulopathy by ROTEM® | ||
2. None | |||
3. Fibrinogen level was an independent predictor of mortality at 24 h and 28 days (p < .001). Hypofibrinogenemia can be detected early by ROTEM® and administration of cryo or fibrinogen concentrate can improve survival. | |||
Wohlauer 2012 [41] | 1. In trauma patients, median ADP inhibition of platelet function was 86.1% vs. 4.2% and impaired platelet function in response to AA was 44.9% vs. 0.5% when compared to healthy volunteers (p < .0001). | ||
2. ADP inhibition correlated with the RBC transfusion within the first 6 hours, 59.6% (0 RBC) vs. 96.1% (>1 RBC) (Wilcoxon p = .025). | |||
3. None. | |||
Woolley 2013 [42] | 1. 51% of all 48 patients were coagulopathic. EXTEM MCF < 40 mm and interim EXTEM A5 and A10 predicted coagulopathy with sensitivities/specificities of 96%/58% (A5) and 100%/ 70% (A10). In addition, statistical comparison of clotting domains between normal volunteers and trauma patients suggests a difference in clot strengths due to a difference in PLT function rather than PLT number (mean 142,000/mm3). | ||
2. None | |||
3. None | |||
Chapman 2013 [43] | 1. Both G and MA values initially normal, crossed to the hypercoagulable range at 48 hours. G values rose from 7.4 ± 0.5 Kd/cs to 15.1 ± 1.9 Kd/cs (p < .01), and MA from 57.6 mm to 74.5 mm (p = .01). | ||
2. None | |||
3. None | |||
Chapman 2013 [56] | 1. None | ||
2. In the general trauma population, LY30 of greater than 3% was associated with MT in 16.7% of the patients vs. 2.1% of those with LY30 < 3% (p = .006). | |||
3. Similarly, LY30 ≥ 3% was associated with all-cause mortality of 20.8% vs. 4.7% (p = .011). | |||
Harr 2013 [44] | 1. Functional Fibrinogen Levels (FF) significantly correlated with von Clauss fibrinogen levels (R2 = 0.87) and MA (R2 = 0.80). The mean fibrinogen contribution to MA was 30%; however, there was a direct linear relationship with fibrinogen level and% fibrinogen contribution to MA (R2 = 0.83). The addition of fibrinogen concentrate in in vitro studies increased MA (60.44 ± 1.48 to 68.12 ± 1.39) and % fibrinogen contribution to MA (23.8 ± 1.8% to 37.7 ± 2.5%). | ||
2. None | |||
3. None | |||
Johansson 2013 [45] | 1. TEG® FF MA and G were lower in the hypocoagulable and significantly higher in hypercoagulable patients compared to patients with normal kaolin TEG® MA. By r-TEG® , R time, angle, MA, and G were reduced in hypocoagulable patients. LY30 was significantly increased in hypocoagulable patients by both TEG® and r-TEG® | ||
2. Of the investigated TEG® , FF, and r-TEG® variables, MA, G, and LY30 were univariate predictors of MT whereas none were independent predictors of MT at 6 or 24 h | |||
3. Nonsurvivors had significantly lower TEG® MA and lower FF MA and G compared to survivors. Further, r-TEG® angle and LY30 were lower in nonsurvivors. | |||
Lee 2013 [46] | 1. There was a strong correlation between the r-TEG® and TEG® MA, which represents platelet function (R = .80). There was a moderate correlation between the G (R = .70) the overall clot strength, k (R = .66) speed of clot formation, and α-angle (R = .38), which reflects the degree of fibrin cross-linking. Lysis at 30 minutes correlated poorly (R = .19). | ||
2. None | |||
3. None | |||
Tapia 2013 [63] | 1. None | ||
2. None | |||
3. TEG® -directed resuscitation is superior to MTP in MT penetrating trauma receiving ≥10U RBC. TEG® -directed resuscitation is equivalent to standardized MTP for all patients receiving ≥6U RBC and is also equivalent to standardized MTP for blunt trauma receiving ≥10U RBC. MTP worsened mortality in penetrating trauma receiving ≥10U RBC, indicating a continued need for TEG® -directed therapy. | |||
Kornblith 2014 [47] | 1. Coagulopathic patients (INR ≥ 1.3) had lower admission MA FF than non-coagulopathic patients (24.7% vs. 31.2%, p < .05). %MA PLT was higher than MA FF at all-time points, decreased over time, and stabilized at 72 h (69.4% at 0 h, 56.2% at 72 h). In contrast, MA FF increased over time and stabilized at 72 hours (30.6% at 0 h, 43.8% at 72 h). | ||
2. Patients requiring FFP had a significantly lower admission MA FF (26.6% vs. 30.6%, p < .05). | |||
3. Higher admission MA FF was predictive of reduced mortality (hazard ratio, 0.815, p < .001). | |||
Branco 2014 [48] | 1. 26.3% were hypercoagulable, 55.9% had a normal TEG® profile, and 17.8% were hypocoagulable. | ||
2. After adjustment, hypercoagulable patients were less likely to require uncross-matched blood (adjusted p = .004) and less total blood products, in particular, plasma at 6 h (adjusted p < .001) and 24 h (adjusted p < .001). | |||
3. Hypercoagulable patients had lower 24 h mortality (0.0% vs. 5.5% vs. 27.8%, adjusted p < .001) and 7-day mortality (0.0% vs. 5.5% vs. 36.1%, adjusted p < .001). Bleeding-related deaths were less likely in the hypercoagulable group (0.0% vs. 1.8% vs. 25.0%, adjusted p < .001). |