Purpose: To evaluate the efficacy of pharmacologic agents for the irritable bowel syndrome. Data Sources: Electronic literature search of MEDLINE (1966 to 1999), EMBASE (1980 to 1999), PsyclNFO (1967 to 1999), and the Cochrane controlled trials registry and a manual search of references from bibliographies of identified articles. Study Selection: Randomized, double-blind, placebo-controlled, parallel, or crossover trials of a pharmacologic intervention for adult patients that reported outcomes of improvement in global or irritable bowel-specific symptoms. Data Extraction: Qualitative and quantitative data reported on study groups, interventions, treatment outcomes, and trial methodologic characteristics. Data Synthesis: 70 studies met the inclusion criteria. The most common medication classes were smooth-muscle relaxants (16 trials), bulking agents (13 trials), prokinetic agents (6 trials), psychotropic agents (7 trials), and loperamide (4 trials). The strongest evidence for efficacy was shown for smooth-muscle relaxants in patients with abdominal pain as the predominant symptom. Loperamide seems to reduce diarrhea but does not relieve abdominal pain. Although psychotropic agents were shown to produce global improvement, the evidence is based on a small number of studies of suboptimal quality. Psychotropic drugs, 5-hydroxytryptamine (5-HT)-receptor antagonists, peppermint oil, and Chinese herbal medicine require further study. Conclusion: Smooth-muscle relaxants are beneficial when abdominal pain is the predominant symptom. In contrast, the efficacy of bulking agents has not been established. Loperamide is effective for diarrhea. Evidence for use of psychotropic agents is inconclusive; more high-quality trials of longer duration are needed. Evidence for the efficacy of 5-HT-receptor antagonists seems favorable, although more studies are needed.
|Original language||English (US)|
|Number of pages||12|
|Journal||Annals of internal medicine|
|State||Published - Jul 18 2000|
ASJC Scopus subject areas
- Internal Medicine