OBJECTIVES: To review and summarize the status of diagnosis, epidemiology, infection control, and treatment of Clostridium difficile -associated disease (CDAD). DIAGNOSIS: A case definition of CDAD should include the presence of symptoms (usually diarrhea) and at least one of the following positive tests: endoscopy revealing pseudomembranes, stool cytotoxicity test for toxin B, stool enzyme immunoassay for toxin A or B, or stool culture fo r C difficile (preferably with confirmation of organism toxicity if a direct stool toxin test is negative or not done). Testing of asymptomatic patients, including those who are asympto- matic after treatment, is not recommended other than for epidemiologic purposes. Lower gastrointestinal endoscopy is the only diagnostic test for pseudomembranous colitis, but it is expensive, invasive, and insensitive (51% to 55%) for the diagnosis of CDAD. Stool culture is the most sensitive laboratory test currently in clinical use, but it is not as specific as the cell cytotoxicity assay. EPIDEMIOLOGY: C difficile is the most frequently identified cause of nosocomial diarrhea. The majority o f C difficile infections are acquired nosocomially, and most patients remain asymptomatic following acquisition. Antimi- crobial exposure is the greatest risk factor for patients, especially clindamycin, cephalosporins, and penicillins, although virtually every antimicrobial has been implicated. Cases of CDAD unassociated with prior antimicrobial or antineoplastic use are very rare. Hands of personnel, as well as a variety of environmental sites within institutions, have been found to be contaminated with C difficile, which can persist as spores for many months. Contaminated com- modes, bathing tubs, and electronic thermometers have been implicated as sources of C difficile. Symptomatic and asymptomatic infected patients are the major reservoirs and sources for environmental contamination. Both genotypic and phenotypic typing systems for C difficile are available and have enhanced epidemiologic investigation greatly. INFECTION CONTROL: Successful infection con- trol measures designed to prevent horizontal transmission include the use of gloves in handling body substances and replacement of electronic thermometers with disposable devices. Isolation, cohorting, handwashing, environmental disinfection, and treatment of asymptomatic carriers are recommended practices for which convincing data of effi- cacy are not available. The most successful control measure directed at reduction in symptomatic disease has been antimicrobial restriction. TREATMENT: Treatment of symptomatic (but not asymptomatic) patients with metronidazole or vancomycin for 10 days is effective; metronidazole may be preferred to reduce risk of vancomycin resistance among other organ- isms in hospitals. Recurrence of symptoms occurs in 7% to 20% of patients and is due to both relapse and reinfection. Over 90% of first recurrences can be treated successfully in the same manner as initial cases. Combination treatment with vancomycin plus rifampin or the addition orally of the yeast Saccharomyces boulardii to vancomycin or met- ronidazole treatment has been shown to prevent subsequent diarrhea in patients with recurrent disease (Infect Control Hosp Epidemiol 1995;16:459477).