Assessing cortisol from hair samples in a large observational cohort: The Whitehall II study

Hair cortisol concentrations (HCC) have been suggested to reflect long-term integrated cortisol levels, but most evidence of associations with co-variates is from small samples of healthy volunteers. The objective of this study was to describe the collection of hair samples in a large cohort study and report associations of demographic and health measures with HCC. We examined HCC measured from the 3cm hair segment near the scalp in 3507 participants (aged 59-83y) from The Whitehall II occupational cohort study of British civil servants. Hair samples were analysed using a column switching LC-APCI-MS/MS assay. Findings from mutually adjusted linear regression analyses revealed lower HCC in participants who reported use of hair dye [% difference (95%CI); -12.5 (-22.0, -1.9), p value=0.022] and evidence suggestive of differences by length of sample storage and seasonal variation. With regard to demographic variables, HCC was lower in women compared to men [-17.0 (-24.8, -8.4), p value <0.001] and higher in Black compared to other ethnic groups. Prevalent diabetes, use of systemic corticosteroids and cardiovascular medication were independently associated with higher HCC. With regard to health, depressive symptoms were associated with higher HCC [20.0 (8.1, 33.3), p value=0.001] following adjustment for physical disease and medication. We conclude that hair steroid analysis presents significant opportunities for assessing cortisol in large scale cohorts. Demographic factors, sample storage, season of collection and hair characteristics should be considered in future analyses. Health status, both mental and physical, is linked to HCC.