Abstract
BACKGROUND: The ratio of cortisol to dehydroepiandrosterone sulfate (DHEAS) in the awakening response has emerged as a potential biomarker of stress-related dysregulation in neurodegenerative conditions. Whether this ratio differs between people with dementia (PWD) and family caregivers, and how it varies with age, sex, dementia severity and agitation, remains unclear. METHODS: We analyzed 1093 day-level saliva samples from 58 participants (PWD = 28; caregivers = 30). The primary outcome was the log-transformed awakening response ratio of cortisol to DHEAS. Linear mixed-effects models with a participant random intercept and natural splines for age estimated group contrasts as geometric mean ratios (GMRs) from estimated marginal means. Fixed-effect predictors included age, sex, dementia severity (Global Deterioration Scale, GDS), and agitation (Brief Agitation Rating Scale, BARS). Model comparisons were conducted. Within PWD, mean-centered models tested one interaction at a time (Sex × GDS, Sex × BARS, Age × GDS, Age × BARS). RESULTS: There was no between-group difference in the ratio after accounting for within-participant clustering and age (GMR = 0.97, 95 % CI 0.64-1.46; p = 0.87). Within PWD, interaction models indicated that the association between age and the ratio strengthened with higher dementia severity (β = 0.043, p = 0.04) and greater agitation (β = 0.011, p = 0.006). Marginal R(2) ranged 0.114-0.141; conditional R(2) 0.358-0.376. CONCLUSIONS: Although average ratio did not differ between PWD and caregivers, it increased more steeply with age at higher dementia severity and agitation. These findings highlight the cortisol-to-DHEA(S) awakening response ratio as a non-invasive and clinically relevant biomarker reflecting symptom-linked neuroendocrine heterogeneity in dementia.