The objective of this study was to compare the continuous infusion of cefepime with the intermittent infusion in patients with sepsis caused by Gram-negative bacilli (GNB). Randomized 1:1 multicenter double-blinded placebo-controlled study with allocation concealment; multicenter study in the intensive care units of Colombia. Patients with sepsis, severe sepsis or septic shock, and GNB-suspected bacteremia. Cefepime was administered for 7 to 14 days over 30 m intermittently every 8 h over 24 h plus continuous saline solution (0.9%) (G1) or 3 g administered continuously plus saline solution every 8 h (0.9%) (G2). The percentage of clinical response at 3, 7, and 14 days, relapse at 28 days, and mortality at discharge were measured. The recruitment was stopped at the suggestion of the Institutional Review Board (IRB) following an FDA alert about cefepime. Thirty-two patients were randomized; 25 received the intervention, and GNB bacteremia was confirmed in 16 (9 G1 and 7 G2). Favorable clinical response in days 3, 7, and 14 was 88.8%, 88.8%, and 77.8% (G1) and was similar for G2 (85.7%). There were no relapses or deaths in G2, while in G1, one relapse and two deaths were observed. The results of this study support the use of cefepime for the treatment of Gram-negative infections in critically ill patients, but we could not demonstrate differences between continuous or intermittent administration because of the small sample size, given the early suspension of the study.