Determination of placebo effect in irritable bowel syndrome

Background and objective: The determinants of the placebo effect are not well established, Goal of this study was to explore likely predictive factors in an already published data set. Methods: We re-analysed data from a study in 120 patients with the irritable bowel syndrome (IBS) that were randomly assigned to three arms of the study to receive (double-blind) either a drug (mebeverin) (n =40) or placebo (n =40), or (in an open trial) dietary treatment (fibre) (n =40) for up to 16 week. Treatment was conducted by 3 different doctors (A, B, Q with 44, 27, and 18 patients, resp. A fourth group (n=31) was treated by different varying physicians. Symptoms were assessed every 4 weeks, and the degree of patient compliance and the number of drop-outs, the number of patients improved/not improved (in %), symptom severity (Kruis Score) at enrolment, and age and gender as covariates were included into the analysis. Results: Drop-out rate was 30% for placebo, 30% for mebeverin, and 15% for the diet. For the patients remaining in the study, average compliance was 75% with placebo, but 89% for the drug and 82% for the diet. Response rates were 39% for placebo, but 20% for the drug; response rate for the diet (open trial) was 43% under all doctors. Response rates for drug and placebo combined were 32% for doctor A (female,43 years), but 19% for doctors B and C together (both males, 32 and 40 years)): this effect was not significant. Placebo responders were more often women (47%) than men (28%), while age effects were only found with dietary treatment: responders were younger. Placebo responders had an overall lower Kruis Score than non-responders (45 vs 52 points), but this was also true for drug (52 vs. 62 points) and diet responders (56 vs 68 points). Conclusion: The major factors contributing to the placebo response are the treating physician (gender, training), and the patients ' gender (female). Patients with lower Kruis score (more likely non-functionally disordered) may be prone to higher (placebo) response rates.