Background: We conducted a retrospective evaluation of the overall safety of drotrecogin alfa (activated) in surgical patients with severe sepsis enrolled in PROWESS. Methods: A blinded Surgical Evaluation Committee (SEC) verified surgical patients as having undergone a significant operative procedure within 30 days prior to enrollment. Serious and treatment-emergent bleeding events, both during the study drug infusion period (120 h) and the entire 28-day study period were analyzed by surgical status and by treatment assignment. Statistical analysis was performed using Fisher's exact test. Results: Serious bleeding rates during infusion in the surgical patients were 3.1% (7/228) and 0% (0/246) in the drotrecogin alfa (activated) and placebo groups, respectively (p = 0.006). Treatment-emergent bleeding rates during infusion in the surgical patients were 16.7% (38/228) and 7.7% (19/246) in the drotrecogin alfa (activated) and placebo groups, respectively (p = 0.003). None of the treatment-emergent bleeding events was fatal. Of seven drotrecogin alfa (activated) serious bleeding events, six were procedure-related. The serious bleeding rates within each treatment group were statistically indistinguishable between the medical and surgical patients. However, the medical patients had numerically higher treatment-emergent bleeding rates than the surgical patients within each treatment group. Despite this observation, overall surgical patients received more transfusions of red blood cells, of platelets, and of fresh frozen plasma than their medical counterparts. Conclusions: Although treatment of surgical patients with drotrecogin alfa (activated) for severe sepsis is associated with a higher incidence of serious bleeding and subsequent treatment-emergent bleeding events, the magnitude of this increase is small and clinically acceptable.