Rates of and Factors Associated With Placebo Response in Trials of Pharmacotherapies for Nonalcoholic Steatohepatitis: Systematic Review and Meta-analysis

BACKGROUND & AIMS: It is important to know the extent of the placebo effect in designing randomized controlled trials for patients with nonalcoholic steatohepatitis (NASH), to accurately calculate sample size and define treatment endpoints. METHODS: We performed a systematic review and meta-analysis of the placebo groups from randomized controlled trials of adults with NASH that provided histologic and/or magnetic resonance image-based assessments. We identified trials through a comprehensive search of MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials, and Scopus, from each database's inception through January 2, 2018. RESULTS: We identified 39 randomized controlled trials, comprising 1463 patients who received placebo. Histologic assessment data (the nonalcoholic fatty liver disease activity scores, NAS) were available from 956 patients; magnetic resonance spectroscopy data were available from 295 patients and magnetic resonance proton density fat fraction measurements from 61 patients. Overall, 25% of patients given placebo had an improvement in NAS by 2 or more points (95% CI, 21%-29%) with a small amount of heterogeneity (I-2 = 27%). There were improvements by at least 1 point in steatosis scores of 33% +/- 3% of patients, in hepatocyte ballooning scores of 30% +/- 3% of patients, in lobular inflammation scores of 32% +/- 3% of patients, and in fibrosis scores of 21% +/- 3% of patients, with a moderate amount of heterogeneity among trials (I-2 range, 51%-63%). Patients given placebo had a statistically significant improvement in NAS (by 0.72 +/- 0.19), with a large amount of heterogeneity (I-2 = 96%). Univariate and multivariate meta-regression showed that trials with a higher baseline NAS, those conducted in South America, and those in which patients had a decrease in body mass index, were associated with greater improvements in NAS among patients given placebo. Patients given placebo had significant reductions in intrahepatic triglyceride, measured by magnetic resonance spectroscopy (by 1.45% +/- 0.54%) with moderate heterogeneity (I-2 = 40%), and in magnetic resonance proton density fat fraction (by 2.43 +/- 0.89), without heterogeneity (I-2 = 0). Mean serum levels of alanine and aspartate aminotransferases decreased significantly (by 11.7 +/- 3.8 U/L and 5.9 +/- 2.1 U/L, respectively; P < .01 for both). CONCLUSIONS: In a meta-analysis of randomized controlled trials of NASH, patients given placebo have significant histologic, radiologic, and biochemical responses. The placebo response should be considered in designing trials of agents for treatment of NASH.