Abstract
AbstractObjectivesHypothalamic-pituitary-adrenal (HPA) dysregulation is a postulated risk factor for cardiovascular disease but results from previous studies have been mixed. We examined cross-sectional associations of diurnal patterns of salivary cortisol with 10 subclinical measures of cardiovascular disease (including pulse wave velocity, carotid intima-media thickness, and measures of heart structure and function), and prospective associations of cortisol variation with cardiovascular (CV) and non-cardiovascular (NCV) mortality.MethodsWe included up to 1,263 participants of the MRC National Survey for Health and Development birth cohort, who had participated in serial diurnal salivary cortisol sampling in 2006-2010 (aged 60–64 years). We used multivariable linear and multinomial logistic regression to estimate associations of cortisol awakening response (CAR), slope and area under the curve (AUC) with subclinical cardiovascular measures, and multivariable Cox regression to estimate associations of the three cortisol metrics with all-cause, CV and NCV mortality.ResultsAll associations with subclinical cardiovascular measures were weak and most were compatible with the null hypothesis. Each standard deviation (SD) higher cortisol AUC was associated with a 1.9% (95%-CI 0.3-3.5) higher pulse wave velocity. A shallow cortisol slope (per SD gradient) was associated with a 1.4% lower intima media thickness (95%-CI −2.6, − 0.2), and lower odds of concentric remodelling (OR 0.83, 95%-CI 0.70-0.99). Given the multiple analyses performed, these could be chance findings. AUC, cortisol slope and CAR were not convincingly associated with cardiovascular or total mortality, although the 95% confidence intervals of hazard ratio estimates were wide.ConclusionsThis study provides little evidence for clinically important associations of salivary cortisol and subclinical cardiovascular measures and mortality.
Publisher
Cold Spring Harbor Laboratory