OBJECTIVE: We aimed to characterize the pharmacokinetics and pharmacodynamics of drotrecogin alfa (activated) (recombinant human activated protein C) in patients with severe sepsis. METHODS: Patients (N = 1690) in a randomized, double-blind, placebo-controlled phase 3 trial received a 96-hour infusion of placebo (n = 840) or drotrecogin alfa (activated) (n = 850), 24 microg x kg(-1) x h(-1). Plasma samples from 680 patients were collected for pharmacokinetic assessment. Pharmacodynamic effects on activated partial thromboplastin time, D-dimer, protein C, and interleukin 6 were analyzed by drotrecogin alfa (activated) steady-state plasma concentration (C(ss)) quartile. RESULTS: Transient endogenous activated protein C concentrations above 10 ng/mL were observed in 11 placebo-treated patients (3.3%). In drotrecogin alfa (activated)-treated patients, the median C(ss) was 44.9 ng/mL and the median plasma clearance (CL(p)) was 40.1 L/h. C(ss) was reached within 2 hours after the infusion was started. Plasma concentrations were below the assay quantitation limit of 10 ng/mL within 2 hours after the infusion was stopped in 92% of patients. CL(p) increased with increasing body weight, so infusion rates should be based on predose body weight. Mean CL(p) associated with age, sex, or baseline hepatic or renal function differed by less than 30% from the mean CL(p) in all patients and resided within the interquartile range of CL(p) in all patients. Dose adjustment is not required on the basis of these factors alone or in combination. No correlation was detected between C(ss) quartile and bleeding risk or the magnitudes of effect on biomarkers of coagulopathy (D-dimers and protein C) and inflammation (interleukin 6). CONCLUSIONS: Plasma concentrations of drotrecogin alfa (activated) attain steady state rapidly after the infusion is started and decline rapidly after the infusion is stopped. The infusion rate should be based on predose body weight and not on any other demographic or baseline clinical covariate.