BACKGROUND:

There have been recent reports of frequent treatment failure associated with the use of metronidazole for treatment of Clostridium difficile-associated disease. We tested the hypothesis that treatment failure with metronidazole is associated with a suboptimal microbiological response in comparison with that of vancomycin.

METHODS:

We conducted a 9-month prospective observational study of patients with C. difficile-associated disease. Cox proportional hazards models were used to compare metronidazole-treated and vancomycin-treated patients in terms of time to resolution of diarrhea and time to reduction of C. difficile in stool to an undetectable level.

RESULTS:

Of 52 study patients with C. difficile-associated disease, 34 (65%) received initial therapy with oral metronidazole, and 18 (35%) received initial therapy with oral vancomycin. Diarrhea resolved in >90% of patients who completed 10 days of treatment with either agent. However, vancomycin-treated patients were more likely to develop undetectable levels of C. difficile (adjusted hazard ratio, 3.99; 95% confidence interval, 1.41-11.3;P = .009) and to have resolution of diarrhea (adjusted hazard ratio, 4.17; 95% confidence interval, 1.53-11.40;P = .005) during the first 5 days of therapy. Ten metronidazole-treated patients (29%) had their treatment changed to oral vancomycin because of persistent symptoms. Seven (70%) of these 10 patients had <1 log reduction in C.difficile concentration; however, only 4 had completed > or = 6 days of metronidazole treatment at the time of the treatment change.

CONCLUSION:

In an observational study with a limited number of subjects, a majority of patients with C. difficile-associated disease responded to therapy with metronidazole or vancomycin. Failure with metronidazole treatment may be attributable to a slower and less consistent microbiological response than that with oral vancomycin treatment.