OBJECTIVES: To estimate the clinical effectiveness of augmentation of SSRI antidepressant therapy with either lithium or an AAP in TRD, defined as failure to respond to two or more antidepressants in the current episode of depression. METHODS: Systematic review of CENTRAL, EMBASE, MEDLINE, and PsycINFO was completed in August 2011. Additional data were obtained from manufacturers. Studies were assessed for quality using the Cochrane Risk of Bias Tool. Pairwise meta-analysis and mixed treatment comparison (MTC) were undertaken based on intention-to-treat analyses. RESULTS: Of the 3,721 papers found in the literature search, 12 randomised controlled trials (RCTs) were identified; 10 (SSRI + AAP vs SSRI + placebo/no treatment); 1 (SSRI + AAP vs SSRI + lithium); 1 (SSRI + lithium vs SSRI + placebo). The RCTs included in the primary analyses used fluoxetine as the SSRI and olanzapine as the AAP. Results of the MTC showed a non-significant trend in favour of lithium augmentation for response [lithium odds ratio (OR) 1.29; 95% credible interval (95% CrI): 0.11 to 5.32], mean change in Montgomery-Åsberg Depression Rating Scale (MADRS) score from baseline (mean difference -1.47; 95% CrI: -9.10 to 6.41) and all-cause withdrawals (OR 0.74; 95% CrI: 0.10 to 2.66). CONCLUSIONS: In patients with TRD, there is a lack of direct evidence comparing the clinical effectiveness of augmenting an SSRI with an AAP compared with augmenting with lithium. Augmentation of SSRIs with lithium or AAP is likely to be beneficial in people with TRD. The limited evidence indicates no statistically significant difference between the two augmentation strategies.