Salivary Cortisone Reflects Cortisol Exposure Under Physiological Conditions and After Hydrocortisone

Abstract Context: Measuring serum cortisol to evaluate stress, adrenal disease, and monitor hydrocortisone replacement requires venepuncture. Conversely, salivary measurements are noninvasive. Objective: This study aimed to investigate measurement of salivary cortisol and cortisone as alternatives to serum cortisol. Design and Setting: This was a prospective cross-over study in a clinical research facility. Patients and Methods: Over three periods (Period 1, 24-h physiological cortisol rhythm; Periods 2 and 3, after 20 mg oral and iv hydrocortisone) 14 male volunteers had serum and saliva cortisol and cortisone, serum albumin, cortisol-binding globulin, and free cortisol measured. Data were analyzed for rhythm parameters and correlations. Linear mixed-effects modelling was performed to determine the relationship between serum cortisol and salivary cortisone. Results: Serum cortisol and cortisone showed similar circadian rhythms with large peak:trough ratios (cortisol median ratio, 11). Albumin and cortisol-binding globulin showed minor peak:trough ratios <1.2. When serum cortisol was <74 (SD, 29) nmol/L, salivary cortisol was not detectable but salivary cortisone was always detected. Salivary cortisol post-oral hydrocortisone produced spurious results due to contamination. Under physiological conditions, salivary cortisone correlated strongly with serum cortisol (ρ, 0.91; 95% confidence interval, 0.89–0.93; P < .001). Similarly, following iv or oral hydrocortisone, salivary cortisone correlated strongly with serum cortisol (ρ, 0.91; 95% confidence interval, 0.89–0.92; P < .001). A mixed-effects model showed that in this population 94% of the variation in salivary cortisone could be predicted from serum cortisol. Conclusion: Salivary cortisol is frequently undetectable and contaminated by oral hydrocortisone. In contrast, salivary cortisone reflects serum cortisol and provides a noninvasive alternative to measuring serum cortisol levels.