Metronidazole may inhibit intestinal colonization withClostridium difficile

PURPOSE: Antibiotics suppress normal gut flora, allowing overgrowth of acquired or nativeClostridium difficile, with release of toxins that cause mucosal inflammation. Oral metronidazole is used to treat antibiotic-associated colitis (pseudomembranous colitis). This study was designed to determine whether oral metronidazole, as part of preoperative bowel preparation, prevents or decreases incidence of antibiotic-associated colitis after elective colonic and rectal procedures. METHODS: Eighty-two patients (40 men) were prospectively, randomly assigned to receive one of two oral antibiotic regimens before colorectal surgery. All patients underwent mechanical bowel preparation with polyethylene glycol-electrolyte lavage solution before administration of oral antibiotics. Group 1 (n=42) patients received three doses (1 g/dose) of neomycin and erythromycin. Group 2 (n=40) patients received three doses (1 g/dose) of neomycin and metronidazole. Both groups received one preoperative and three postoperative doses of intravenous cefotetan (2 g/dose). Both groups had stool samples tested forC. difficile toxin in the preoperative and postoperative periods by enzyme-linked immunoabsorbent assay or by tissue culture cytotoxicity. Patients with preoperative stool studies positive forC. difficile were excluded from the study. RESULTS: Treatment groups were not different for age, gender, or surgical procedure. Mean age ±1 standard deviation was 67.6±13.6 (range, 34–94) years in Group 1 and 62.1±13.5 (range, 35–84) years in Group 2 (P=0.069). Mean length of hospital stay ±1 standard deviation was 9.76±4.9 (range, 4–28) days for Group 1 and 8.05±2.6 (range, 3–14) days for Group 2 (P=0.053). Five patients in Group 1 (neomycin and erythromycin) and one patient in Group 2 (neomycin and metronidazole) had positive stool studies forC. difficile. Relative risk of colonization withC. difficile in Group 1 was 4.76 times that in Group 2 (95 percent confidence interval, 0.581, 39). This difference was not statistically significant (P=0.202). There were no significant differences inC. difficile colonization rates with respect to age, length of stay, or gender. CONCLUSIONS: This study suggests that there may be a clinical association between use of metronidazole preoperatively and inhibition of intestinal colonization byC. difficile in this patient population undergoing colonic and rectal surgery.Supported, in part, by the Department of Research at St. Joseph Mercy Hospital, Ann Arbor, Michigan.Poster presentation at the meeting of The American Society of Colon and Rectal Surgeons, Philadelphia, Pennsylvania, June 22 to 26, 1997.