Poor sleep quality is associated with altered cortisol levels in older adults with high anxiety levels: a moderation analysis of Midlife in United States study

Objectives While poor sleep predicts cortisol levels in older adults, the moderating influence of anxiety remains under-examined. This study evaluates whether anxiety alters the association between sleep quality and 12-hour creatinine-adjusted urinary cortisol in community-dwelling older adults. Methods Utilizing biomarker data from the MIDUS 3 study (2017–2022; N = 747, mean age = 66), we performed multiple linear regression-based moderation analysis with bias-corrected bootstrapping and heteroskedasticity-robust standard errors. Anxiety was measured via the STAI scale and sleep quality via the PSQI scale. The adjusted model controlled for mean-centered covariates: BMI, age, LDL, depression (CESD scale), smoking history, sex, and HbA1c. Results In the simple moderation model, anxiety significantly moderated the sleep-cortisol relationship (Beta = -0.022, 95% CI: [-0.036, -0.007]). Specifically, poorer sleep (higher PSQI score) was associated with lower cortisol levels only among individuals with higher relative anxiety score (Mean +1 SD; Beta= -0.34; 95% CI: [-0.57, -0.104]). Neither sleep quality nor anxiety exhibited significant main effects. In the adjusted model, the moderation effect remained significant. Significant covariate associations included age (Beta = 0.118, 95% CI: [0.038, 0.203]), BMI (Beta = -0.222, 95% CI: [-0.313, -0.129]), and LDL (Beta = 0.028, 95% CI: [0.006, 0.048]). Conclusion Anxiety significantly moderates the link between sleep quality and cortisol in community-dwelling older adults. These findings suggest that sleep-related endocrine responses may vary according to anxiety levels, highlighting the necessity of assessing anxiety when evaluating stress biomarkers in aging populations.