Abstract P338: Diurnal Salivary Cortisol is Associated with Increased Glycemia and Insulin Resistance: The Multiethnic Study of Atherosclerosis

Introduction: Hypercortisolism is associated with insulin resistance (IR) and diabetes mellitus (DM); however, to our knowledge prior studies have not examined the association of diurnal cortisol curve features with measures of glycemia or IR in a population-based setting. We hypothesized that cortisol curve features would be positively associated with (1) glycemia in those with and without DM and (2) IR, in non-diabetic subjects. Methods: We examined the cross-sectional association of cortisol curve features with glycemia and IR in 1,069 ethnically diverse men and women (mean age 70 ± 9) in the Multi-Ethnic Study of Atherosclerosis Stress II Study. Salivary cortisol measures were collected over 2 days with 8 time points measured per day. Cortisol curve features included log-transformed wake-up cortisol, cortisol awakening response (CAR; awakening to 30 minutes post-awakening), early decline slope (30 minutes to 2 hours post-awakening), late decline slope (2 hours post-awakening to bedtime), overall decline slope, bedtime cortisol, and total area under the curve (AUC). DM was defined based on fasting glucose (≥ 126 mg/dL), hypoglycemic medication use, and/or physician diagnosis. Glycemia was assessed via hemoglobin A1c (HbA1c) and IR via the homeostasis model assessment of IR (HOMA-IR). We used linear regression models to examine the cross-sectional association of log-transformed cortisol curve features with HbA1c, stratified by DM status, and with HOMA-IR in non-diabetic participants. Adjustments were made for socio-demographics, adiposity, depressive symptoms, medications and cigarette smoking. Results: The association of cortisol curve features with HbA1c varied by DM status in unadjusted and adjusted models. Following multivariable adjustment, cortisol curve features were not associated with HbA1c among non-diabetic subjects; however, among those with DM, early decline slope, overall decline slope, bedtime cortisol, and total cortisol AUC were significantly and positively associated with a 5.7% (95% CI, 2.3,9.3), 38.8% (95% CI, 4.4, 84.4), 3.1% (95% CI 1.1,5.2), and 5.3% (95% CI 2.3,8.5) higher HbA1c, respectively. Wake-up cortisol, CAR, and late decline slope were not associated with HbA1c. Among those without DM, wake-up cortisol and AUC were inversely associated with a 9.4% lower (95% CI, -14.3, -4.2) and 11% lower (95% CI, -17.2, -4.3) HOMA-IR, respectively, in the unadjusted model; however, these results were attenuated and became non-significant following multivariable adjustment (-1.6%; 95% CI, -6.4, 3.4 for wake-up cortisol and -2.8%; 95% CI, -9.2, 4.2 for AUC). Conclusion: Among those with DM, cortisol curve parameters suggestive of higher cortisol exposure were associated with higher HbA1c. These findings suggest a role for hypothalamic-pituitary-adrenal axis activation in hyperglycemia associated with DM.