In a prospective randomized clinical trial, 103 patients undergoing elective colorectal surgery received either cefoxitin or a combination of metronidazole and gentamicin. Six of 52 patients in the cefoxitin group (11.5 per cent) and six of 48 patients in the metronidazole/gentamicin group (12.5 per cent) developed serious wound infections. Two patients (3.8 per cent) in the cefoxitin group and one patient (2.1 per cent) in the metronidazole/gentamicin group developed deep sepsis. These results suggest that cefoxitin is as effective as the combination of metronidazole and gentamicin for prophylaxis against serious postoperative septic complications.