GOAL: To identify patients likely to fail metronidazole as initial treatment of C. difficile infection. BACKGROUND: For moderate to severe Clostridium difficile-associated diarrhea, metronidazole is the drug of choice for treatment. Oral vancomycin is given to patients who fail metronidazole or have intolerable side effects. STUDY: Retrospective review identified all patients treated for C. difficile-associated diarrhea during hospitalization from January 2000 to September 2001. C difficile was documented by a positive toxin assay or pseudomembranes on colonoscopy. Metronidazole failure was defined as persistent symptoms of C. difficile-associated diarrhea after 5 days of uninterrupted therapy. Response was defined as improvement in symptoms at day 5 of therapy including reduction of diarrhea to <or=2 bowel movements per day. RESULTS: 119 C. difficile-associated diarrhea patients were identified, and 99 met inclusion criteria. There were 61 (62%) metronidazole responders and 38 (38%) treatment failures. Albumin <2.5g/l and intensive care unit stay at/prior to diagnosis were the only variables associated with treatment failure. The odds ratios for treatment failure were 11.7 (95% confidence interval: 4.0-31.6) and 4.1 (95% confidence interval: 1.3-12.2), respectively. When considering these 2 variables together (low albumin, intensive care unit care), the area under the receiver operating characteristic curve was 0.80 for predicting treatment failure. CONCLUSIONS: Albumin level <2.5g/l and intensive care unit stay were predictors of failure of metronidazole therapy for C. difficile-associated diarrhea. These patients may benefit from oral vancomycin therapy at outset.
GOAL: To identify patients likely to fail metronidazole as initial treatment of C. difficileinfection. BACKGROUND: For moderate to severe Clostridium difficile-associated diarrhea, metronidazole is the drug of choice for treatment. Oral vancomycin is given to patients who fail metronidazole or have intolerable side effects. STUDY: Retrospective review identified all patients treated for C. difficile-associated diarrhea during hospitalization from January 2000 to September 2001. C difficile was documented by a positive toxin assay or pseudomembranes on colonoscopy. Metronidazole failure was defined as persistent symptoms of C. difficile-associated diarrhea after 5 days of uninterrupted therapy. Response was defined as improvement in symptoms at day 5 of therapy including reduction of diarrhea to <or=2 bowel movements per day. RESULTS: 119 C. difficile-associated diarrheapatients were identified, and 99 met inclusion criteria. There were 61 (62%) metronidazole responders and 38 (38%) treatment failures. Albumin <2.5g/l and intensive care unit stay at/prior to diagnosis were the only variables associated with treatment failure. The odds ratios for treatment failure were 11.7 (95% confidence interval: 4.0-31.6) and 4.1 (95% confidence interval: 1.3-12.2), respectively. When considering these 2 variables together (low albumin, intensive care unit care), the area under the receiver operating characteristic curve was 0.80 for predicting treatment failure. CONCLUSIONS: Albumin level <2.5g/l and intensive care unit stay were predictors of failure of metronidazole therapy for C. difficile-associated diarrhea. These patients may benefit from oral vancomycin therapy at outset.