Objectives: To compare the clinical and bacteriological outcome of critically ill patients with sepsis treated by ceftriaxone administered as a once-a-day intermittent bolus dose or by 24 h continuous infusion. Patients and methods: We conducted an open-label, randomized controlled pilot study in 57 patients clinically diagnosed with sepsis (suspected/proven infection and systemic inflammatory response syndrome) in a tertiary level intensive care unit. Patients were randomized to receive 2 g of ceftriaxone administered by once-daily intermittent bolus dosing or by 24 h continuous infusion. Clinical and bacteriological outcomes were assessed by blinded clinicians. Results: Fifty-seven patients were enrolled in the study, 50 of whom fulfilled the a priori definition of treatment for 4 or more days. The infusion (n = 29) and bolus groups (n = 28) were similar in terms of demographics, although the median age of those receiving the infusion was younger. Intention-to-treat analysis found no statistically significant differences in the primary outcomes for clinical response (P = 0.17), clinical cure [infusion n = 13/29 versus bolus n = 5/28; adjusted odds ratio (AOR) = 3.74; 95% confidence interval (95% CI) = 1.11-12.57; P = 0.06], bacteriological response (P = 0.41) and bacteriological cure (infusion n = 18/29 versus bolus 14/28; AOR = 1.64; 95% CI = 0.57-4.70; P = 0.52). However, logistic regression in patients that complied with the a priori definitions who received ceftriaxone by continuous infusion (AOR = 22.8; 95% CI = 2.24-232.3; P = 0.008) or patients with a low Acute Physiology and Chronic Health Evaluation (APACHE) II score (AOR = 0.70; 95% CI = 0.54-0.91; P = 0.008) were associated with an improved clinical outcome when age and Sepsis Organ Failure Assessment (SOFA) score at time of study entry were controlled for. Conclusions: This pilot study suggests clinical and bacteriological advantages of continuous infusion of ceftriaxone over bolus administration in critically ill patients in patients requiring 4 or more days of treatment. This sets the scene for a large multicentre double-blind randomized controlled trial to confirm these findings.