Reliability of diurnal salivary cortisol metrics: A meta-analysis and investigation in two independent samples

• Published in: Comprehensive Psychoneuroendocrinology (Online) Vol. 16; p. 100191
• Main Authors: Norton, Sara A., Baranger, David AA, Young, Ethan S., Voss, Michaela, Hansen, Isabella, Bondy, Erin, Rodrigues, Merlyn, Paul, Sarah E., Edershile, Elizabeth, Hill, Patrick L., Oltmanns, Thomas F., Simpson, Jeffry, Bogdan, Ryan
• Format: Journal Article
• Language: English
• Published: Elsevier 01-11-2023
• Subjects: Cortisol; Reliability; Stress; Variability
• ISSN: 2666-4976; 2666-4976
• DOI: 10.1016/j.cpnec.2023.100191

Stress-induced dysregulation of diurnal cortisol is a cornerstone of stress-disease theories; however, observed associations between cortisol, stress, and health have been inconsistent. The reliability of diurnal cortisol features may contribute to these equivocal findings. Our meta-analysis (5 diurnal features from 11 studies; total participant n = 3307) and investigation (15 diurnal cortisol features) in 2 independent studies (St. Louis Personality and Aging Network [SPAN] Study, n = 147, ages 61–73; Minnesota Longitudinal Study of Risk and Adaptation [MLSRA] Study, n = 90, age 37) revealed large variability in the day-to-day test-retest reliability of diurnal features derived from salivary cortisol data (i.e., ICC = 0.00–0.75). Collectively, these data indicate that some commonly used diurnal cortisol features have poor reliability that is insufficient for individual differences research (e.g., cortisol awakening response) while others (e.g., area under the curve with respect to ground) have fair-to-good reliability that could support reliable identification of associations in well-powered studies.
•
Stress is among the strongest predictors of the onset and course of many negative mental and physical health outcomes.
•
However, observed associations between cortisol, stress, and health have been inconsistent.
•
In a meta-analysis and 2 independent studies, we find widespread variability in the reliability of diurnal cortisol features.
•
Some cortisol features, such as the CAR, have poor reliability that is insufficient for individual differences research.
•
Area-under-the-curve (AUC
g
) has fair-good reliability and should be prioritized as a measure of cortisol.