BACKGROUND: A randomized, double-blind, placebo-controlled, multicenter trial (EPO-2, N = 1,302) in anemic critically ill patients demonstrated a 29-day survival benefit in the trauma subgroup receiving epoetin alfa (mortality 8.9% vs. 4.1%). A second similarly designed trial (EPO-3, N = 1,460) confirmed this survival benefit in the epoetin alfa-treated trauma cohort (mortality 6.7% vs. 3.5%). This analysis presents trauma cohort data from both trials for evaluation of the impact of baseline factors including trauma-specific variables on outcomes. METHODS: Patients received 40,000 U epoetin alfa or placebo weekly, for a total of 4 (EPO-2) or 3 (EPO-3) doses, starting on ICU day 3. Kaplan-Meier survival curves for the two groups were compared using the log-rank test. Univariate and multivariate Cox proportional hazard regression methods were used to evaluate relationship between baseline factors and mortality. RESULTS: Demographic and trauma variables at baseline were comparable. Mortality was consistently reduced by approximately 50% in both studies (EPO-2--day 29 unadjusted HR: 0.46, 95% CI: 0.24-0.89; EPO-3--day 29 unadjusted HR: 0.51, 95% CI: 0.27-0.98.). Adjusting for baseline and trauma variables had minimal effect on hazard ratios for mortality at day 29 (EPO-2--day 29 adjusted HR: 0.50, 95% CI: 0.26-0.97; EPO-3--day 29 adjusted HR: 0.38, 95% CI: 0.19-0.74) and day 140 (EPO-3--adjusted HR: 0.39, 95% CI: 0.21-0.72). In EPO-3, there appeared to be an increase in clinically relevant thrombovascular events in the epoetin alfa treated group (16.4% vs. 12.5%, RR: 1.3, 95% CI: 0.93-1.85) but not in EPO-2 (11.1% vs. 13.3%, RR: 0.84, 95% CI: 0.56-1.28). CONCLUSION: Epoetin alfa demonstrated a survival advantage in both of the critically ill trauma patient cohorts of two prospective, randomized clinical trials, which was not affected by baseline factors including trauma-specific variables. A definitive study in trauma subjects is warranted.