Date of Award

January 2016

Document Type

Open Access Thesis

Degree Name

Medical Doctor (MD)



First Advisor

Andres Martin

Second Advisor

Kathleen Merikangas


Objective: The aims of this study were to 1) examine salivary cortisol distributions in a nationally representative sample of U.S. adolescents; and to 2) evaluate whether youth with mood and anxiety disorders are associated with differential salivary cortisol response to an acute mild stressor.

Background: An association between mood disorders and hypothalamic-pituitary-adrenal (HPA) axis dysfunction, as indicated by elevated levels of cortisol, has been well established [1, 2]. Although studies have suggested that these differences emerge early in the development of mood disorders in children and adolescents [3] the samples have generally been very small and primarily derived from clinical settings [3-5].

Methods: The National Comorbidity Survey-Adolescent Supplement (NCS-A) is a nationally representative face-to-face interview of adolescents aged 13-18 years in the continental U.S. [6]. Adolescents were administered a modified version of the World Health Organization Composite International Diagnostic Interview Version 3.0 and saliva was collected pre- and post- interview on a subset of 1,755 adolescents. Quantile regression models were used to assess pre- interview cortisol and rate of change in salivary cortisol level across entire quantile distributions adjusted for time of day cortisol sample collected, sex, age, and additionally pre-interview cortisol level when rate of change in cortisol was used as the outcome variable. For descriptive purpose, traditional linear regression analyses were performed in parallel.

Results: Pre-interview cortisol levels were higher in the morning and decreased over the time of day. The magnitude of rate of change in cortisol measures was greater in the earlier time categories than that in the later time categories. After adjustment for time of cortisol sample collected, pre-interview cortisol levels were significantly lower across quantile distributions in the younger age groups (13-14y, 15-16y) compared to the oldest (17-18y), and were higher among adolescents with lower family income around the lower quantile range. There were no sex differences in pre-interview cortisol levels, but the rate of change was significantly smaller at the lower end of quantile distribution among females than that in males. Across the majority of quantile distributions, adolescents with earlier weekday bedtime and mid-sleep time were significantly associated with lower pre-interview cortisol level as well as rate of change compared to those with later bedtime and delayed mid-time sleep. Although suicidality was not associated with pre-interview cortisol level, there was a greater rate of change in cortisol among adolescents with suicidal behavior. Anxiety status was not associated with pre-interview cortisol level, while adolescents with lifetime mood disorders had somewhat lower pre-interview cortisol level at the lower end of quantile distribution. With respect to rate of change in cortisol level, adolescents with either lifetime or 12-month anxiety disorders showed a greater rate of change than those without disorders at the lower quantile distributions. Those with severe anxiety and those with 12-month mood disorders had a greater rate of changes in cortisol levels than controls.

Conclusions: Our findings reveal that adolescents with mood and/or anxiety disorders have a greater change in salivary cortisol before and after a diagnostic interview as an acute mild stressor compared to their non-anxious or depressed counterparts. These findings provide biologic evidence for greater stress reactivity in adolescents with anxiety and depression in non-clinical samples.

Disclosures: Nothing to disclose.


This is an Open Access Thesis.