Depression is linked to inflammation. But the link is small.

Meta-analysis of observational studies. 423,421 studies. 124 papers; databases searched from inception to January 2023.

Many studies, added up. People were watched, not told what to do. A clue, not proof.

Who: adults 18+ in community and clinical samples with blood inflammatory markers and a depression measure reported by sex; 53% female (98 papers with women, n = 222,507; 85 with men, n = 200,914)

• Overall linksmalla bit more in men
• CRP link in womennonemen: small link
• Later inflammationnoneafter depression

More to know, from our read of the paper:

• Observational studies only: linked, not proven. 92 of 124 papers were cross-sectional (one point in time), so they can't say which came first
• Effects are small (Cohen's d 0.06 in women, 0.14 in men for all markers combined). Say 'small' or 'slightly higher'; never turn d into percentages or 'X times'
• What survives Bonferroni: all markers combined in women (p 0.003) and men (p < 0.001), CRP in men (p < 0.001). Sex difference for all markers p = 0.049 and for CRP p = 0.002 (uncorrected meta-regression)
• E-values mostly ~1.2-2.0: modest unmeasured confounding could explain the associations
• Adjusted estimates smaller than unadjusted; some studies adjusted for possible mediators (BMI, sleep, activity), so the authors call the estimates possibly conservative (true male all-marker effect likely between 0.05 and 0.31)

This meta-analysis pooled 124 studies with 423,421 adults (53% women), looking at women and men separately. Depression went with slightly higher inflammation in the blood in both sexes, and the link was a bit stronger in men. The marker CRP was linked to depression in men, but not clearly in women. Most studies measured everything at one point in time, and in the long-term studies depression did not predict later inflammation: linked, not proven.

We read the abstract and the full text of this paper and checked every number on this page against it (17 facts, checked on 2026-10-29). The study design is named above. Cohort and other observational studies are described as linked, not proven. Found an error? Email theis@onestudyaday.com.