Dropout is common and affects the statistical power and randomization balance of randomised controlled trials (RCTs). To estimate the dropout rate in RCTs of metabolic dysfunction-associated steatohepatitis (MASH) and to examine factors associated with dropout in placebo-treated participants. PubMed and Cochrane databases were searched for phase 2-4 MASH RCTs with placebo arms through November 24, 2024. Dropout was defined as the attrition of patients included in the intention-to-treat analysis but did not complete treatment. RCTs were qualitatively reviewed to assess the expected and observed dropouts. Generalised linear mixed model was used to estimate pooled dropout rates. Sixty RCTs with 3230 placebo-treated participants with MASH were analysed. Thirty-three RCTs reported the dropout rate used to estimate the effect size. Of these, 60.6%, 36.4%, and 3.0% had an expected dropout rate that was higher, lower, and similar, respectively, than the observed dropout rate in the placebo arm. Overall, the dropout rate was 11.06% (95% confidence interval [CI] 9.07 to 13.42), with a higher rate in phase 3-4 trials than in phase 2 trials. The corresponding rates due to adverse events, loss to follow-up and patient choice were 2.41% (95% CI 1.67 to 3.48), 1.79% (95% CI 1.06 to 2.99) and 4.06% (95% CI 2.97 to 5.53), respectively. Meta-regression determined that the dropout rate increased with longer treatment duration. Placebo dropout in MASH RCTs is significant, mainly due to patient choice. Factors such as trial phase and treatment duration should be considered when calculating sample size in future clinical trials.
Read full abstract