Abstract
Personalizing treatment for late-life depression requires identifying and integrating information from multiple factors that influence treatment efficacy (moderators). We performed exploratory moderator analyses using data from a multi-site, randomized, placebo-controlled, double-blind trial of aripiprazole augmentation. Patients (n = 159) aged ≥60 years had major depressive disorder that failed to remit with venlafaxine monotherapy. We examined effect sizes of 39 potential moderators of aripiprazole (vs. placebo) augmentation efficacy using the outcome of percentage reduction in depressive symptom after 12 weeks. We then incorporated information from the individually relevant variables in combined moderators. A larger aripiprazole treatment effect was related to: white race, better physical function, better performance on Trail-Making, attention, immediate, and delayed memory tests, greater psychomotor agitation and suicidality symptoms, and a history of adequate antidepressant pharmacotherapy. A smaller aripiprazole treatment effect was observed in patients with: more pain and more work/activity impairment and libido symptoms. Combining information from race and Trail-Making test performance (base combined moderator (Mb*)) produced a larger effect size (Spearman effect size = 0.29 (95% confidence interval (CI): 0.15, 0.42)) than any individual moderator. Adding other individually relevant moderators in the full combined moderator (Mf*) further improved effect size (Spearman effect size = 0.39 (95% CI: 0.25, 0.52)) and identified a sub-group benefiting more from placebo plus continuation venlafaxine monotherapy than adjunctive aripiprazole. Combining moderators can help clinicians personalize depression treatment. We found the majority of our patients benefited from adjunctive aripiprazole, but a smaller subgroup that is identifiable using clinical measures appeared to benefit more from continuation venlafaxine plus placebo.