Song · JAMA 2018 · retrospective matched cohort study · n=1297756

Association of Stress-Related Disorders With Subsequent Autoimmune Disease.

Cited 442 times in the scientific literature.

Level 3 - non-randomized controlled study

Retrospective population- and sibling-matched cohort study

PubMed 29922828 · doi:10.1001/jama.2018.7028 · record verified 2026-08-29

What was done

Population- and sibling-matched retrospective cohort study using Swedish health registries between 1981 and 2013. The study included 106,464 patients with stress-related disorders (posttraumatic stress disorder, acute stress reaction, adjustment disorder, and other stress reactions), 1,064,640 matched unexposed individuals, and 126,652 full siblings. Cox proportional hazards regression estimated hazard ratios for 41 autoimmune diseases occurring beyond 1 year after stress disorder diagnosis, adjusting for multiple covariates.

What was found

Over a mean follow-up of 10 years, autoimmune disease incidence was 9.1 per 1000 person-years in exposed individuals, 6.0 in unexposed controls, and 6.5 in siblings (absolute rate differences: 3.12 [95% CI, 2.99-3.25] and 2.49 [95% CI, 2.23-2.76] per 1000 person-years compared with unexposed and sibling cohorts, respectively). Stress-related disorders were associated with increased autoimmune disease risk (HR, 1.36 [95% CI, 1.33-1.40]), with consistent results in sibling comparisons. For posttraumatic stress disorder, HRs were 1.46 (95% CI, 1.32-1.61) for any autoimmune disease and 2.29 (95% CI, 1.72-3.04) for multiple (3 or more) autoimmune diseases. Relative risk was higher in younger patients (HR, 1.48 for age 33 or younger vs 1.23 for age 51 or older; P for interaction < .001). Persistent SSRI use for 320 days or more in the first year of PTSD diagnosis was associated with attenuated risk compared to 179 days or less (HR, 1.82 vs 3.64; P for trend = .03).

Why it matters

This study provides large-scale evidence linking stress-related disorders to an elevated risk of subsequent autoimmune disease, independent of shared familial and genetic factors.

Limits

The study is observational and cannot demonstrate direct causation. Registry data captured specialist diagnoses, potentially missing milder cases treated in primary care. Unmeasured lifestyle factors and residual confounding remain possible despite sibling matching.

Cited by