Thromboprophylaxis and Intensive Care Unit Mortality in Critically Ill Patients: A Severity-Adjusted Retrospective Cohort Study

BackgroundAs venous thromboembolism (VTE), including pulmonary embolism (PE) and deep venous thrombosis (DVT), is a serious complication of critical illness, prophylaxis is considered for each patient.In everyday ICU practice, the choice of thromboprophylaxis follows the same patient characteristics that determine whether someone survives, so any observational comparison of outcome is confounded by indication unless illness severity is taken into account.We aimed to describe thromboprophylaxis practice in a tertiary ICU, compare clinically detected DVT and PE between prophylaxis strategies (primary outcome), and examine the association between prophylaxis strategy and ICU mortality after adjustment for comorbidity and day-1 severity of illness (secondary outcome). MethodologyWe conducted a retrospective cohort study in an adult ICU in Oman between January 2021 and June 2022.Of the 912 admissions screened, 859 admissions (739 patients) were analyzed.Four admissions with another anticoagulant were excluded; 855 admissions were grouped by the thromboprophylaxis received: lowmolecular-weight heparin (LMWH; enoxaparin, n = 381), unfractionated heparin (UFH, n = 220), or no pharmacological prophylaxis (n = 254).A day-1 Sequential Organ Failure Assessment (SOFA) score was calculated for all.DVT and PE were compared using the Fisher-Freeman-Halton exact test, with Holmadjusted pairwise Fisher's exact tests.ICU mortality was modeled by logistic regression with LMWH as the reference group, before and after adjustment for comorbidity and SOFA.Landmark, first-admission, and propensity-stratified sensitivity analyses and E-values were used to examine the robustness of the mortality association. ResultsDVT was found in 1.3% of admissions and PE in 2.0%, with no statistically significant difference between groups (Fisher-Freeman-Halton exact test, DVT: p = 0.172; PE: p = 0.099).Median day-1 SOFA was 8 (IQR = 5-13) with LMWH, 14 (8-17) with UFH, and 12 (7-16) with no pharmacological prophylaxis (Kruskal-Wallis H = 67.46,p < 0.001).ICU mortality was 14.2%, 48.6%, and 47.6%, respectively (χ² = 110.05,p < 0.001).Against LMWH, the odds ratio for death with UFH was 5.73 unadjusted, 3.92 after adjustment for comorbidity, and 4.94 (95% CI: 2.73-8.93)once SOFA was added; for no pharmacological prophylaxis, the corresponding figure was 5.16 (95% CI: 3.05-8.72).On the risk scale, adjustment for comorbidity and SOFA reduced the difference between UFH and LMWH from 34.4 to 16.9 percentage points (95% CI: 10.5-24.5).The association persisted within every severity stratum and among admissions still in the ICU at 48 and 72 hours; across analyses, the odds ratio for UFH ranged from 2.91 (propensity-stratified) to 5.59 (72-hour landmark).The comparison with no pharmacological prophylaxis is confounded by design, as 98% of these patients had a documented contraindication. ConclusionsNo statistically significant difference in clinically detected thrombotic events was observed between prophylaxis strategies, although the study was not powered to detect modest differences.ICU mortality was considerably higher among patients given UFH or no pharmacological prophylaxis.This association persisted after adjustment for measured comorbidity and day-1 organ dysfunction, but residual confounding, including by unrecorded treatment-limitation decisions and treatment timing, remains possible.The mortality finding should be regarded as hypothesis-generating.