Systematic review and meta-analysis of augmentation and combination treatments for early-stage treatment-resistant depression

Abstract

Background: Major depressive disorder (MDD) is a highly burdensome health condition, for which there are numerous accepted pharmacological and psychological interventions. Adjunctive treatment (augmentation/combination) is recommended for the ~50% of MDD patients who do not adequately respond to first-line treatment. We aimed to evaluate the current evidence for concomitant approaches for people with early-stage treatment-resistant depression (TRD; defined below). Methods: We systematically searched Medline and Institute for Scientific Information Web of Science to identify randomised controlled trials of adjunctive treatment of ⩾10 adults with MDD who had not responded to ⩾1 adequate antidepressant. The cochrane risk of bias (RoB) tool was used to assess study quality. Pre-post treatment meta-analyses were performed, allowing for comparison across heterogeneous study designs independent of comparator interventions. Results: In total, 115 trials investigating 48 treatments were synthesised. The mean intervention duration was 9 weeks (range 5 days to 18 months) with most studies assessed to have low ( n = 57) or moderate ( n = 51) RoB. The highest effect sizes (ESs) were from cognitive behavioural therapy (ES = 1.58, 95% confidence interval (CI): 1.09–2.07), (es)ketamine (ES = 1.48, 95% CI: 1.23–1.73) and risperidone (ES = 1.42, 95% CI: 1.29–1.61). Only aripiprazole and lithium were examined in ⩾10 studies. Pill placebo (ES = 0.89, 95% CI: 0.81–0.98) had a not inconsiderable ES, and only six treatments’ 95% CIs did not overlap with pill placebo’s (aripiprazole, (es)ketamine, mirtazapine, olanzapine, quetiapine and risperidone). We report marked heterogeneity between studies for almost all analyses. Conclusions: Our findings support cautious optimism for several augmentation strategies; although considering the high prevalence of TRD, evidence remains inadequate for each treatment option.