Rationale: Continuous infusion of ?-lactam antibiotics may improve outcomes because of time-dependent antibacterial activity compared with intermittent dosing. Objectives: To evaluate the efficacy of continuous versus intermittent infusion in patients with severe sepsis. Methods: We conducted a randomized controlled trial in 25 intensive care units (ICUs). Participants commenced on piperacillin-tazobactam, ticarcillin-clavulanate, or meropenem were randomized to receive the prescribed antibiotic via continuous or 30-minute intermittent infusion for the remainder of the treatment course or until ICU discharge. The primary outcome was the number of alive ICU-free days at Day 28. Secondary outcomes were 90-day survival, clinical cure 14 days post antibiotic cessation, alive organ failure-free days at Day 14, and duration of bacteremia. Measurements and Main Results: We enrolled 432 eligible participants with a median age of 64 years and an Acute Physiology and Chronic Health Evaluation II score of 20. There was no difference in ICU-free days: 18 days (interquartile range, 2-24) and 20 days (interquartile range, 3-24) in the continuous and intermittent groups (P = 0.38). There was no difference in 90-day survival: 74.3 (156 of 210) and 72.5 (158 of 218); hazard ratio, 0.91 (95 confidence interval, 0.63-1.31; P = 0.61). Clinical cure was 52.4 (111 of 212) and 49.5 (109 of 220); odds ratio, 1.12 (95 confidence interval, 0.77-1.63; P = 0.56). There was no difference in organ failure-free days (6 d; P = 0.27) and duration of bacteremia (0 d; P = 0.24). Conclusions: In critically ill patients with severe sepsis, there was no difference in outcomes between ?-lactam antibiotic administration by continuous and intermittent infusion.