The likelihood of receiving the drug determines the size of the placebo response in IBS trials

While placebo controlled trials are common in the evaluation of novel drugs, they have been questioned because of the ethical problems of withholding effective treatment in (pain) patients. As alternative, comparator controlled trials ave been proposed. Both however, differ in the likelihood of providing effective treatment: 50% and 100%, resp.

Methods: We investigated whether the likelihood of receiving drug or placebo determines the size of the drug response and the size of the placebo response (DR, PR) in treatment studies of irritable bowel syndrome (IBS).

Results: In 92 published randomized placebo-controlled trials with more than 22.000 patients with IBS, 16 had a drug:placebo ratio of 2:1 or greater while the remaining were powered 1:1. Computing the overall DR and PR, both responses were lower with higher likelihood of receiving the drug (ratio 2:1 or higher), but the difference between DR and PR („therapeutic gain“) decreased. Ten of the 16 studies (4 with a 2:1 ratio, 3 with 3:1, and 3 with 4:1; total n=6484) were compared to 10 studies (n=3782) of similar size (>80 patients per study arm) and design (drug trials with novel compounds) based on a standardized patient definition (according to Rome criteria), and with comparable (dichotomous) endpoints. Again, the placebo response decreased with increasing likelihood of receiving a drug from 37.6% (1:1) to 29.1% (4:1). Similarly, the drug response dropped from 49.8% (1:1) to 42.4% (4:1) (MANOVA: significant decrease for DR (p=.035), but not for PR). Discrimination between DR and PR was best for 1:1 trials. Correlation between the drug response and the placebo response across all 20 trials was significant (Pearsons r=0.781, p<0.001), but was higher in 1:1 trials (n=10; r=0.91, p<0.001) than in trials with a higher drug-to-placebo ratio (n=10, r=0.69, p=0.027).

Conclusion: The higher the likelihood of receiving active treatment (a drug as compared to a placebo), the lower the response rate to both placebo and the drug in IBS trials, at the expense of a poorer discrimination between DR and PR. This is contrary to respective data from depression trials. (Supported by a grant from Volkswagen-Stiftung)