Abstract

Salivary and serum cortisol are widely used biomarkers of stress, although they represent different physiological fractions and may not provide equivalent estimates of chronic stress. This study compared their agreement and validity in shift workers using hair cortisol as a reference measure of chronic exposure. Methods: A cross-sectional study included 160 adults, comprising 80 shift workers and 80 daytime workers. Salivary cortisol, serum total and free cortisol, corticosteroid-binding globulin, hair cortisol, and perceived stress were assessed. Agreement and criterion validity were examined using correlation, concordance, Bland-Altman, regression, and receiver operating characteristic analyses. Results: Shift workers showed higher salivary, serum, and hair cortisol concentrations than daytime workers. Salivary and serum free cortisol showed moderate agreement (r = 0.70; concordance coefficient = 0.62) with proportional bias. Salivary cortisol correlated more strongly with hair cortisol (r = 0.592) than serum total cortisol (r = 0.438; p = 0.006 for difference). Adjustment for corticosteroid-binding globulin increased the serum correlation to 0.53. Salivary cortisol remained independently associated with hair cortisol and demonstrated greater discrimination of high chronic stress (AUC = 0.801) than serum total cortisol (AUC = 0.716). Conclusion: Salivary cortisol showed greater validity than serum total cortisol for assessing chronic stress, partly because serum measurements are influenced by binding-globulin variation.

Keywords: Salivary cortisol, Serum cortisol, Hair cortisol, Chronic stress, Corticosteroid binding globulin, Biomarker validity