Decrease in nosocomial Clostridium difficile-associated diarrhea by restricting clindamycin use

Objective: To report the investigation and effective control of a nosocomial epidemic of Clostridium difficile-associated diarrhea. Design: Concurrent surveillance for identification of new nosocomial cases, retrospective case-control analysis, and hospital formulary control of antibiotic use. Setting: University-affiliated Veterans Affairs Medical Center located in southwestern United States. Patients: Hospitalized patients who developed diarrhea submitted stool specimens for cytotoxin assay. Patients who were positive for cytotoxin were compared with control patients without infection. Measurements: Isolates of C. difficile were typed by restriction endonuclease analysis. Antimicrobial agent use from hospital pharmacy records and selected patient data from chart review were correlated with frequency of specific laboratory abnormalities. Results: For 13 months, the monthly incidence of C. difficile infection averaged more than five times that for the previous 21 months. Stool specimens from 34 patients (59%) contained a single strain (restriction enzyme analysis type J7). Clindamycin was statistically associated with the epidemic as shown by the following: clindamycin use at our center compared with national normal values, clindamycin use for years before compared with during the epidemic, monthly use of clindamycin compared with monthly frequency of infection, frequency of infection in patients receiving clindamycin compared with that in patients receiving other antimicrobial agents, and amount of clindamycin used by infected patients compared with that used by control patients. Restricting clindamycin use led to a prompt decrease in infection rate and the type J7 organisms. Conclusion: A nosocomial epidemic of C. difficile diarrhea was controlled by analysis of antibiotic use patterns and by subsequent restriction of clindamycin.