This was a prospective, randomized study of 123 patients with penetrating abdominal injuries. The patients received ceftriaxone or cefoxitin for 24 h (in the presence of colonic injury, 48 h). The overall incidence of abdominal sepsis was 7.3 per cent (ceftriaxone 5 per cent, cefoxitin 9.5 per cent, P greater than 0.05). Colonic injury was the most important risk factor for the development of septic complications. Other factors, such as the weapon used, a prehospital time longer than 4 h, shock on admission, multiple organ injuries, and small bowel perforation, did not influence the incidence of sepsis.