Lifetime affective episode burden and peripheral inflammation in bipolar and unipolar depression: an exploratory cross-sectional study
Background Lifetime episode count and illness duration may convey different information about peripheral inflammation, but estimates from small cross-sectional samples require cautious interpretation. Methods This exploratory analysis examined adult inpatients with bipolar or unipolar depression recruited in Shenzhen, China, during 2020–2022. Primary log-linear models jointly included log 2 (episode count + 1) and log 2 (illness duration in years + 1), adjusting for age, sex, body mass index, diagnosis, and 17-item Hamilton Depression Rating Scale (HAMD-17) total. HC3 standard errors and Benjamini–Hochberg adjustment across six burden coefficients were used. Results Primary analyses included 63 patients for IL-6 and TNF-α and 57 for IL-1β. Per doubling of episode count plus one, conditional associations were observed for TNF-α (geometric-mean ratio [GMR], 2.238; 95% confidence interval [CI], 1.424–3.517; q = 0.004) and IL-6 (GMR, 3.591; 95% CI, 1.305–9.882; q = 0.029), whereas the IL-1β estimate remained inconclusive (GMR, 1.767; 95% CI, 0.815–3.831; q = 0.175). Duration was inversely associated with TNF-α at fixed episode count (GMR, 0.622; 95% CI, 0.442–0.877; q = 0.023). The TNF-α episode association was nonsignificant without duration adjustment and did not meet multiplicity-adjusted significance after excluding annotated records (n = 47; q = 0.130). A model with fewer covariates supported the association within the original sample, but its estimate was less precise in all 75 age-eligible patients (GMR, 1.815; 95% CI, 0.950–3.467; q = 0.106). Excluded eligible patients differed in depressive severity and inflammatory-marker distributions. Conclusion The associations did not consistently meet multiple-testing-corrected significance thresholds in sensitivity analyses and should be interpreted strictly as hypothesis-generating, requiring confirmation in larger longitudinal cohorts. Missing treatment and assay-quality information further limits interpretation. The results do not establish causality, neuroprogression, protective effects of longer illness, or predictive utility.
Authors
- Xi Yang
- Chao Wan (ORCID: https://orcid.org/0000-0001-8643-6662)
- Dongsheng Han
- Haichen Yang
- Yun Su
Institutions
- Shanghai Mental Health Center (CN)
- Shenzhen KangNing Hospital (CN)
Publication Details
- Journal
- Frontiers in Psychiatry
- Published
- 2026-11-19
- DOI
- https://doi.org/10.3389/fpsyt.2026.1994087
- Primary Topic
- Tryptophan and brain disorders
- Type
- article
- Field-Weighted Citation Impact
- 0.00