Diet quality and depression risk: A systematic review and meta-analysis of prospective studies

The association between diet and the onset and treatment of depression is unclear. This study aims to estimate the evidence for the prospective association between diet quality and depression.

Systematic searches in PubMed, Web of Science and Embase up to 2024/01/04 of propective studies on the potential relationship between diet quality and depression identified 21 randomized clinical trials (RCT) and 92 prospective cohorts (PC) (>700,000 participants), that were pooled in random-effects frequentist and Bayesian meta-analyses.

Treatment-RCT yielded anecdotal-to-moderately strong support for the hypothesis that dietary interventions improve depression (d = -0.80, 95 % CI = -1.46 to -0.14), though their internal validity was low. Interventional effects were observed in nonpatient samples (d = -0.17, -0.29 to -0.08). Prevention-RCT yielded strong evidence indicating no association. Data from PC showed that adherence to some dietary patterns/food groups was associated with depression, e.g., adherence to the Mediterranean-diet was associated to fewer self-reported symptoms (OR = 0.91, 95%CI = 0.88-0.96), with anecdotal-to-moderate strength of evidence but no dose-response relationship. For most dietary patterns and food-groups (e.g., fish), data supported the null-hypothesis of no effect. Effect moderation was present: associations were evident when depression was self-reported but not when diagnosed. Postpartum-depression studies yielded evidence showing dose-response associations between fish (OR = 0.84, 95%CI = 0.78-0.91) and dietary DHA/EPA/Omega3 consumption (OR = 0.84, 95%CI = 0.77-0.92) with self-reported depression-scores.

The strength of the current evidence on the hypothesis that dietary quality influences depression outcome is very low, due to reverse causation and low internal and construct validity. The clinical significance of fish/dietary DHA/EPA/Omega3 intake to reduce the risk of postpartum-depression should be further investigated.