BACKGROUND: Clinical trial design is challenging in irritable bowel syndrome (IBS) due in part to a high placebo effect. We postulated that the mere presence of an adverse event (AE) may unmask patients in clinical trials who are assigned to the active agent, and this may lead to higher reported efficacy. METHODS: We evaluated therapies receiving at least a Grade 1B from the American College of Gastroenterology Task Force for IBS or which passed recent phase III clinical trials. Therapies with AE data derived from less than 50 patients in each study arm were excluded. Statistically significant excess AE were identified, risk difference was calculated for each AE, and incidence of AE in the treatment arm was reported. We examined the relationship of attributable drug benefit, defined as the reciprocal of number-needed-to-treat found in literature, with various measures of AE incidence. KEY RESULTS: Attributable drug benefit correlated significantly with average AE risk difference, calculated as treatment arm AE incidence minus placebo arm AE incidence (R(2) = 0.70, p = 0.039), and also with highest treatment arm AE incidence (R(2) = 0.70, p = 0.038) for each therapy. There were also trends toward correlation with average treatment arm AE incidence (R(2) = 0.54, p = 0.096) and highest AE risk difference (R(2) = 0.63, p = 0.059) for each therapy. CONCLUSIONS & INFERENCES: Our study suggests that higher AE incidence on active therapy is associated with more beneficial patient-reported outcomes in IBS clinical trials. This raises the issue of spontaneous unblinding.
BACKGROUND: Clinical trial design is challenging in irritable bowel syndrome (IBS) due in part to a high placebo effect. We postulated that the mere presence of an adverse event (AE) may unmask patients in clinical trials who are assigned to the active agent, and this may lead to higher reported efficacy. METHODS: We evaluated therapies receiving at least a Grade 1B from the American College of Gastroenterology Task Force for IBS or which passed recent phase III clinical trials. Therapies with AE data derived from less than 50 patients in each study arm were excluded. Statistically significant excess AE were identified, risk difference was calculated for each AE, and incidence of AE in the treatment arm was reported. We examined the relationship of attributable drug benefit, defined as the reciprocal of number-needed-to-treat found in literature, with various measures of AE incidence. KEY RESULTS: Attributable drug benefit correlated significantly with average AE risk difference, calculated as treatment arm AE incidence minus placebo arm AE incidence (R(2) = 0.70, p = 0.039), and also with highest treatment arm AE incidence (R(2) = 0.70, p = 0.038) for each therapy. There were also trends toward correlation with average treatment arm AE incidence (R(2) = 0.54, p = 0.096) and highest AE risk difference (R(2) = 0.63, p = 0.059) for each therapy. CONCLUSIONS & INFERENCES: Our study suggests that higher AE incidence on active therapy is associated with more beneficial patient-reported outcomes in IBS clinical trials. This raises the issue of spontaneous unblinding.
Authors: Chantal Berna; Irving Kirsch; Sean R Zion; Yvonne C Lee; Karin B Jensen; Pamela Sadler; Ted J Kaptchuk; Robert R Edwards Journal: Pain Date: 2017-06 Impact factor: 6.961