Activated protein C and hospital mortality in septic shock: A propensity-matched analysis

Objective: Evidence regarding the efficacy and safety of human recombinant activated protein C in severe sepsis is limited, especially outside of clinical trials. We sought to compare the outcomes of patients with septic shock who received early treatment with activated protein C to those who did not. Design, Setting, and Patients: A retrospective cohort study at 404 U. S. hospitals. We studied 33,749 patients with sepsis who were admitted to intensive care and administered antibiotics and vasopressors within 2 days of admission. Measurements and Main Results: Hospital mortality, intracranial and gastrointestinal hemorrhage, major transfusion. Compared to the entire cohort, the 1576 activated protein C-treated patients included in the matched analysis were younger (mean age, 61 vs. 67), more likely to be white (70% vs. 63%), and had fewer comorbidities. Treated patients were more likely to require mechanical ventilation (77% vs. 48%), to be administered two or more vasopressors (68% vs. 41%), to undergo pulmonary artery catheterization (9% vs. 4%), and to die in the hospital (40.7% vs. 38.1%). In a propensity-matched sample in which all covariates achieved balance, receipt of activated protein C was associated with reduced hospital mortality (40.7% vs. 46.6%; risk ratio, 0.87; 95% confidence interval, 0.80-0.95). This result was insensitive to a hypothetical unmeasured confounder. A similar pattern was observed across groups stratified by age and number of organ-supportive therapies. Four activated protein C-treated patients (0.25%) had hemorrhagic stroke, 107 (6.8%) had gastrointestinal bleeding, and five (0.3%) required major transfusion. Conclusions: Among patients presenting with septic shock, early treatment with activated protein C may be associated with reduced hospital mortality. (Crit Care Med 2010; 38: 1101-1107)