Objectives: To assess the safety and effect on coagulopathy of a range of doses of recombinant human activated protein C (rhAPC). To determine an effective dose and duration of rhAPC for use in future clinical trials. Design: Double-blind, randomized, placebo-controlled, multicenter, dose-ranging (sequential), phase II clinical trial. Setting: Forty community or academic medical institutions in United States and Canada. Patients: One hundred thirty-one adult patients with severe sepsis. Interventions: Intravenous infusion of rhAPC (12, 18, 24, or 30 μg/kg/hr) or placebo for 48 or 96 hrs. Measurements and Main Results: No significant differences in incidence of serious bleeding events (4% rhAPC, 5% placebo, p > .999) or incidence of serious adverse events (39% rhAPC, 46% placebo, p = 0.422) between rhAPC- and placebo-treated patients were observed. One of 53 rhAPC-treated patients with suitable immunogenicity samples had a low level, transient, non-neutralizing anti-APC antibody response not associated with any clinical adverse event. Significant dose-dependent decreases in both D-dimer (p <0.001) and end of infusion interleukin 6 levels (p = .021) were demonstrated. No statistically significant effects on fibrinogen or platelet counts were observed. A nonstatistically significant 15% relative risk reduction in 28-day all-cause mortality was observed between rhAPC- and placebo-treated patients. Conclusions: rhAPC was safe and well-tolerated and demonstrated a dose-dependent reduction in D-dimer and interleukin 6 levels relative to placebo. The dose of 24 μg/kg/hr for 96 hrs was selected for use in future clinical studies.