Placebo Response Rates in Trials of Licensed Drugs for Irritable Bowel Syndrome With Constipation or Diarrhea: Meta-analysis.

Journal Article (Journal Article;Review;Systematic Review)

BACKGROUND & AIMS: There are several licensed drugs for irritable bowel syndrome (IBS) that have proven efficacy in randomized controlled trials (RCTs), but placebo response rates are high. We conducted a systematic review and meta-analysis of licensed drugs to estimate magnitude of placebo response rate according to Food and Drug Administration (FDA)-recommended endpoints and to assess how this varies with stringency of the endpoint used to define response. METHODS: We searched MEDLINE, EMBASE CLASSIC and EMBASE, and the Cochrane central register of controlled trials (through January 2021) to identify RCTs comparing licensed drugs with placebo in adult IBS patients. Studies assessed efficacy according to at least one of composite response, abdominal pain response, or stool response. Data were extracted as intention-to-treat analyses, with dropouts assumed to be treatment failures and pooled using a random-effects model. RESULTS: There were 17 RCTs of licensed drugs versus placebo in IBS with constipation (4603 patients placebo) and 17 trials in IBS with diarrhea (3908 patients placebo). In IBS with constipation, according to FDA criteria, pooled composite, abdominal pain, and stool response rates with placebo over ≥6 of 12 weeks were 18.9%, 34.6%, and 30.1%, respectively. Evaluating response rates over ≥9 of 12 weeks led to placebo response rates of 4.3% for the composite endpoint, 24.5% for abdominal pain, and 7.7% for stool. In IBS with diarrhea, pooled placebo response rates according to FDA criteria were 16.2% for the composite endpoint, 40.2% for abdominal pain, and 16.2% for stool. Increasing the threshold used to define abdominal pain response from ≥30% improvement to ≥40% or ≥50% led to lower placebo response rates of 34.5% and 23.4%. CONCLUSIONS: Future RCTs should adhere to current FDA-recommended endpoints for IBS because these lead to lower placebo response rates. However, consideration should be given to further refining some of these to better differentiate between active drug and placebo.

Full Text

Duke Authors

Cited Authors

  • Barberio, B; Savarino, EV; Black, CJ; Ford, AC

Published Date

  • May 2022

Published In

Volume / Issue

  • 20 / 5

Start / End Page

  • e923 - e944

PubMed ID

  • 34425274

Electronic International Standard Serial Number (EISSN)

  • 1542-7714

Digital Object Identifier (DOI)

  • 10.1016/j.cgh.2021.08.025


  • eng

Conference Location

  • United States