Thrombosis, bleeding, and mortality in patients with sepsis-induced coagulopathy: Analysis of a prospective cohort.

BACKGROUND

Sepsis-induced coagulopathy (SIC) is an early phase of disseminated intravascular coagulation and a candidate marker for risk stratification. Although SIC is linked to higher mortality, its value for predicting thrombotic and hemorrhagic events remains uncertain. We assessed the prevalence of SIC and its association with hemostatic complications and mortality in sepsis.

METHODS

In a prospective cohort of 389 adults with sepsis admitted to an intermediate care unit, SIC was defined by International Society of Thrombosis and Haemostasis criteria (score ≥ 4). Primary outcomes were 30-day venous thromboembolism, arterial thrombosis, and bleeding; all-cause 30-day mortality was secondary. Predictive performance of the SIC score was evaluated with receiver operating characteristic analysis, bootstrap resampling, and Monte Carlo simulation.

RESULTS

SIC was present in 33.4 % of patients. Thirty-day mortality was 27.1 % in SIC-positive patients versus 13.1 % in SIC-negative patients (p = 0.001), and SIC remained independently associated with death (adjusted OR 1.43; 95 % CI 1.13-1.80; p = 0.003). SIC positivity was not associated with overall thrombotic events: 42.9 % (12/28) of patients with thrombosis and 32.7 % (118/361) without thrombosis had SIC (p = 0.301). Discrimination for thrombotic and hemorrhagic events was poor (AUROC 0.573 and 0.576, respectively), with further decline after resampling; simulation analyses confirmed limited predictive capacity for either complication.

CONCLUSIONS

In this cohort, SIC was associated with higher mortality but not with thrombotic or hemorrhagic events. This association likely reflects overall severity of illness rather than clinically overt vascular complications. These findings do not support using SIC alone to guide anticoagulation or transfusion decisions and support the development of outcome-specific risk models, potentially integrating dynamic clinical variables and serial laboratory trajectories.