Comparative efficacy and acceptability of pharmacological interventions for postpartum depression: a systematic review and network meta-analysis.

BACKGROUND

Pharmacological treatment for postpartum depression (PPD) have mainly involved conventional antidepressants, RCTs have also evaluated rapid-acting glutamatergic agents, neuroactive steroids, hormonal therapies, and botanical preparations. An updated comparative synthesis of these treatments is therefore needed.

OBJECTIVES

To synthesize RCT evidence and compare the efficacy, response, remission, acceptability, and safety of pharmacological interventions for PPD.

DATA COLLECTION AND ANALYSIS: Two reviewers independently extracted data and assessed risk of bias. A frequentist random-effects network meta-analysis was conducted to integrate direct and indirect evidence across interventions.

MAIN

RESULTS

18 trials (N = 1722) were included. Compared with placebo, saffron (SMD -1.07, 95% CI -1.52 to -0.61), fluoxetine (SMD -0.95, 95% CI -1.43 to -0.47), esketamine (SMD -0.63, 95% CI -0.89 to -0.37), brexanolone, and zuranolone showed signals of improvement in depressive symptom severity. Fluoxetine, saffron, and zuranolone were associated with higher response rates, although the estimates for fluoxetine and saffron were imprecise. Brexanolone was the only intervention showing a statistically significant improvement in remission. Based on all-cause discontinuation, brexanolone may have a potentially unfavourable acceptability signal. Zuranolone was associated with an increased incidence of adverse events. Given the sparse networks and mostly low-to-moderate certainty of evidence, SUCRA rankings should be interpreted cautiously as exploratory findings.

CONCLUSIONS

Current RCT evidence suggests that saffron, fluoxetine, esketamine, brexanolone, and zuranolone may improve symptoms of PPD. Brexanolone may have a potentially unfavourable acceptability signal, while zuranolone was associated with increased adverse event incidence. However, the evidence is limited by sparse networks, limited active-comparator evidence, imprecision in some estimates, and clinical heterogeneity, and should therefore be interpreted with caution.