Pre-diagnostic biomarkers and risk of stress-related disorders: a cohort study based on electronic health records.

Abstract

Stress-related disorders are associated with future somatic health conditions, suggesting these disorders involve systemic mechanisms. Yet, evidence from biomarkers remains fragmented across physiological systems. We investigated whether routinely collected laboratory biomarkers are associated with stress-related disorder risk and analyzed their temporal trends before diagnosis. We conducted a retrospective cohort study using electronic health records from Central Finland, collected between 2010 and 2023. Our analytical sample included 73,909 individuals: 6,758 cases and 67,151 controls matched on sex and birth year. Cases were diagnosed with stress-related disorders: acute stress reaction, posttraumatic stress disorder, adjustment disorder, other/unspecified reactions to severe stress, burnout, or stress not elsewhere classified. At baseline, participants were aged 34-92 years and were followed for an average of 4.6 ± 3.3 years. Ten routine biomarkers were examined: C-reactive protein, hemoglobin, glucose, glycated hemoglobin, triglycerides, high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), creatinine, sodium, and potassium. Temporal trends were visualized using generalized additive models. Cox proportional hazards models, adjusted for comorbidities, prescribed medications, and care visit frequency, were used to estimate associations between the risk of stress-related disorder diagnosis and biomarker levels measured within one year prior. Three biomarkers were associated with the risk of stress-related disorder diagnosis in the multivariable model. Higher hemoglobin (HR 0.98 per g/L, 95% CI, 0.97-0.99) and higher potassium (HR 0.74 per mmol/L, 95% CI, 0.64-0.86) were associated with a reduced risk, while higher LDL-C was associated with an increased risk (HR 1.12 per mmol/L, 95% CI, 1.06-1.18). The multivariable model concordance was 0.67. In time-varying models extended over the full follow-up, only hemoglobin retained a significant association. Hemoglobin, potassium, and LDL-C showed modest but robust associations with stress-related disorder diagnosis. These findings point to physiological domains for future research into the somatic aspects of stress-related disorders.