番茄社区

Pharmacological trials in irritable bowel syndrome (IBS) should include a run-in period of at least 2 weeks and dosing of no more than once or twice a day in order to minimize the pooled placebo response rate, a new meta-analysis suggested.

More than a quarter of IBS patients had a placebo response on the measure of global improvement, with multiple moderators associated with the response. Identifying these moderators could improve the design of IBS drug trials, said Michelle Bosman, MD, of Maastricht University Medical Center in the Netherlands, and colleagues.

In 70 articles describing 73 randomized controlled trials, the pooled placebo response rate reached 27.3% (95% CI 24.3%-30.9%) using the global improvement endpoint and 34.4% (95% CI 31.2-37.8%) using the abdominal pain endpoint, they reported online in the .

However, the rate dropped to 17.9% (95% CI 15.2%-21.0%) when using the composite FDA endpoint responder definition.

Bosman and colleagues conducted the meta-analysis in light of the introduction of new FDA trial endpoints and diagnostic criteria (Rome III) that have altered the landscape for IBS trials.

The included randomized controlled trials were published from April 1959 to April 2020, and all compared an active pharmacotherapeutic agent with placebo in adults 18 and older. Eligible studies also had a dichotomous outcome of response to therapy in terms of global improvement or improvement in abdominal pain.

The main outcome of the meta-analysis was the magnitude of the pooled placebo response rate for global improvement, abdominal pain, and . When the composite FDA endpoint was subdivided, the pooled placebo response rate was 35.1% using the abdominal pain responder definition and 28.3% using the bowel symptom responder definition.

Studies published before 2006 (introduction of Rome III criteria), as well as those done in Europe with a parallel design, a run-in period of 2 weeks or less, dosing three times a day or more, and a smaller control group, were significantly associated with an increased pooled placebo response rate.

of placebo response rate included treatment expectations, as well as behavioral, psychobiological, and contextual factors such as conditioning and the patient-clinician relationship, the authors explained.

A higher pooled placebo response rate has also been associated with an unbalanced randomization in multiple neurological and psychiatric conditions and a higher number of study visits, but the current analysis did not observe this association, they noted.

The prevalence of the in IBS trials has garnered increasing interest in recent years. A of 73 trials reported a pooled placebo response rate of 37.5% (95% CI 34.4%-40.6%), widely ranging from 0% to 91.7%.

Although adequate assessment of the determinants of the response might be difficult, the authors acknowledged, they nevertheless called for more accurate and detailed reporting of these moderators in future trials. In addition, assessing possible unblinding might also provide insight into treatment expectations, and this can be done in line with FDA recommendations by asking patients at the end of a trial to identify the arm they believe they participated in.

The high placebo response rate associated with straightforward clinical response questions poses barriers to innovation in IBS care, noted Eric Shah, MD, MBA, of Dartmouth-Hitchcock Medical Center in Lebanon, New Hampshire, in an . "These findings are encouraging in suggesting that careful attention to patient selection and clinical trial endpoints have put a substantial dent in this longstanding problem," he wrote.

While an adequate run-in period and less frequent dosing as recommended by the authors might mitigate the problem, more research is needed, Shah added. "As the authors elegantly remark, there remain no trials informing comparative placebo responses between patients meeting the current Rome IV diagnostic criteria for IBS versus the previous Rome III criteria."

The placebo response rate might plausibly be higher in patients meeting Rome IV criteria, he continued, "because these criteria represent a more severe subset of patients who are less likely to lose their IBS diagnosis over the course of a clinical trial compared with patients meeting Rome III criteria (potentially risking a non-specific regression to the mean effect)."

Shah stressed the importance of distinguishing the need to reduce placebo rates for research and regulatory purposes from the need to communicate information on expected treatment outcomes to patients in shared decision making. "Clinicians should understand that the most pertinent question in clinical care is not whether the patient has met the FDA composite responder endpoint, but ultimately, do they feel better?" he wrote. "Thankfully, with greater understanding of placebo response to inform rigorous IBS clinical trial design, we are getting closer to answering yes to that question than ever before."

Among the several limitations to the meta-analysis, Bosman and co-authors noted that 17% of articles had an unclear risk of bias, which might have influenced results. In addition, variability in patient behavior during the trial period and the clinical evolution of the disorder could not be extracted from the trials, but might influence the placebo response.

Furthermore, some response moderators, particularly in older trials, have been inadequately reported. There was also substantial inter-study heterogeneity in trial design, endpoints, and reporting.

Comment