BackgroundThe high heterogeneity was existed among the studies of the elderly participants with major depressive disorder (MDD), which may lead to incorrect conclusions in the previous meta-analysis. This study used model based meta-analysis to compare the efficacy of fluoxetine between the elderly and non-elderly participants with MDD and to explain the heterogeneity among the studies. MethodsA comprehensive literature search was conducted in the public databases, involving utilization of fluoxetine for treating MDD in the acute-phase. The time-efficacy model was established based on the changes of the Hamilton Depression Rating Scale (HDRS) score compared to baseline level. The efficacy features and related factors of fluoxetine in the elderly participants were investigated by comparing with the non-elderly population. ResultsSixty-one studies encompassing 4058 participants were included in the analysis. We found the trial design of placebo controlled vs. comparator controlled was a significant impact factor for the efficacy of fluoxetine. The typical decrease rate of HDRS score in the elderly participants was strikingly lower than that of the non-elderly participants at week 8, with 39.9% vs. 49.1% in the placebo controlled trial and 46.5% vs. 57.2% in the comparator controlled trial. LimitationThe efficacy of other antidepressants except fluoxetine in the elderly participants need to be explored in the future study. ConclusionsThe efficacy of fluoxetine in the elderly participants was lower than that of the non-elderly participants. The heterogeneity of the trial design should be distinguished when comparing the efficacy of antidepressants between the elderly and non-elderly MDD participants.