Background Pouchitis is a relatively new disease and criteria for diagnosis, classification, and measurement of disease activity were only recently proposed. The previous lack of consensus on these issues has hampered the design and conduct of randomized, double-blind, placebo-controlled trials, and a result medical therapy for pouchitis has been largely empiric. Numerous medical therapies have been reported to be of benefit in uncontrolled trials. Only four small placebo-controlled trials and one small controlled trial of two active agents have been performed, evaluating treatment with metronidazole, oral probiotic bacteria (VSL-3), bismuth carbomer enemas and glutamine versus butyrate suppositories. This review will examine the results from these five controlled trials to determine which of the currently utilized empiric medical therapies for pouchitis can be substantiated with valid data from controlled trials. Objectives To determine the effectiveness of medical therapy (including metronidazole, bismuth carbomer enemas, oral probiotic bacteria, butyrate suppositories, and glutamine suppositories) for inducing a response or maintaining remission in pouchitis. Search strategy Studies were selected using the MEDLINE data base (1966 - December 1997), abstracts from major gastrointestinal meetings and references from published articles and reviews. The Cochrane Controlled Trials Register and the Inflammatory Bowel Disease Review Group Trials Register were also searched. Selection criteria Four randomized controlled trials of medical therapy in adult patients with pouchitis were identified: two placebo controlled trials in active chronic pouchitis; one maintenance of remission trial comparing two active agents in chronic pouchitis; and one placebo-controlled maintenance of remission trial for chronic pouchitis. A single patient "n-of-1" trial for active chronic pouchitis was excluded. Data collection and analysis Data were extracted by three independent observers based on the intention to treat principle. Each study was given a quality score based on predetermined criteria. Extracted data were converted to 2X2 tables (response versus no response and medical therapy versus placebo or medical therapy versus medical therapy) and then synthesized in to a summary statistic using the pooled odds ratio and 95% confidence intervals as described by Cochran and Mantel and Haenszel (the "odds ratio" in MetaView). Main results The odds ratios of inducing a response using oral metronidazole or bismuth carbomer foam enemas compared with placebo in active chronic pouchitis were 26.67 (95% CI 2.31-308.01) and 1.00 (95% CI 0.29-3.48), respectively. The number needed to treat with oral metronidazole compared to placebo to prevent one additional relapse was 2. There was no statistically significant difference between bismuth carbomer foam enemas and placebo for prevention of relapse. The odds ratio of maintaining remission in chronic pouchitis for oral probiotic bacteria (VSL-3) compared with placebo was 205.00 (95% CI 9.89-4247.71), while the number needed to treat to prevent one additional relapse was 2. After discontinuation of suppressive medical therapy for chronic pouchitis, there was no statistically significant difference in the maintenance of symptomatic remission with glutamine suppositories compared to butyrate suppositories (odds ratio=3.00; 95% CI 0.46-19.59). Authors' conclusions The results presented in this review must be interpreted with extreme caution given the small numbers of trials and patients evaluated for any one comparison. Metronidazole appears to be an effective therapy for active chronic pouchitis. Bismuth carbomer foam enemas may not be an effective therapy for chronic active pouchitis. Oral probiotic therapy with VSL-3 appears to be an effective therapy for maintaining remission in patients with chronic pouchitis in remission. There is no evidence of a difference in the maintenance of symptomatic remission in patients with chronic pouchitis treated with glutamine versus butyrate suppositories, and it is unknown whether glutamine and butyrate are equally effective or ineffective. Additional randomized, double-blind, placebo-controlled, dose-ranging clinical trials are needed to determine the efficacy of empiric medical therapies currently being used in patients with pouchitis.