REVIEW OF: Henssler J et al, JAMA Psychiatry 2022;79(4):300–312

STUDY TYPE: Meta-analysis of randomized trials

Guidelines recommend a single antidepressant for nonpsychotic major depressive disorder, but is this the most effective approach? Or should we combine two antidepressants into a synergistic combination where the sum is greater than the parts? This meta-analysis seeks to answer these questions.

The authors identified 39 randomized trials lasting 3–12 weeks that compared dual antidepressant therapy to monotherapy for 6,751 patients with unipolar depression. They excluded maintenance trials. The primary outcome was the standardized mean difference (SMD, also known as effect size) of combination therapy compared to monotherapy. Secondary outcomes included remission rates, response rates, overall dropout rates, and dropout rates due to adverse events. The authors also performed several subgroup analyses to determine which combination of antidepressants worked best. 

Overall, there was a small benefit to combination therapy relative to monotherapy with an SMD of 0.31 (95% confidence interval [CI], 0.19–0.44), which roughly translates to a number needed to treat of 8–10. For the secondary outcomes, there was a 52% increase in odds of remission and a 40% increase in odds of response. Notably, there were no differences in the rates of overall dropouts or dropouts due to adverse events. 

So which combination should we use? Bupropion combinations failed to do better than monotherapy (SMD = 0.10; 95% CI, -0.07–0.27), while combinations involving the presynaptic α2-autoreceptors (specifically, augmentation with mirtazapine 7.5–45 mg/day, trazodone 100 mg/day, and mianserin 30–60 mg/day) did separate from placebo. The α2-autoreceptors were particularly effective when paired with an SSRI, SNRI, or tricyclic antidepressant (SMD = 0.37; 95% CI, 0.19–0.55). The authors suggest this combination as first-line treatment for severe cases of depression and for patients who do not respond to monotherapy.

These results conflict with other meta-analyses, which did not find a significant effect for mirtazapine augmentation, and the difference may lie in how the data were grouped (Zhou X et al, J Clin Psychiatry 2015;76(4):e487–e498). Nearly all of the large studies for the α2-autoreceptors were negative, while the positive results that tilted the balance came almost entirely from small studies involving fewer than 100 subjects. Specifically, all six large trials of mirtazapine augmentation were negative, as was the single large trial of mianserin augmentation. The two trazodone studies used a low dose of 100 mg, which is more in line with its use as a hypnotic than its antidepressant dose range, and neither of the studies tested the augmentation against placebo. 

While these results paint a mixed picture of mirtazapine in depression, there is evidence that mirtazapine augmentation helps associated symptoms of insomnia and anxiety (Rifkin-Zybutz R et al, J Psychopharmacol 2020;34(12):1342–1349).

CARLAT TAKE

Although this analysis finds a ray of hope for antidepressant combinations involving mirtazapine, trazodone, or the non-US antidepressant mianserin, that optimism is tempered by the negative results of large controlled trials. In practice, trazodone and mirtazapine may work best for associated symptoms of insomnia or—in the case of mirtazapine—anxiety.