Morning Salivary Cortisone Versus Serum Cortisol in the Overnight Dexamethasone Suppression Test (ODNST): Evaluation in a Clinical Setting.

INTRODUCTION

Saliva hormone measurement is increasingly being applied in everyday clinical practice. In relation to salivary cortisone measurement, there is a particular advantage, with minimal chance of cross-reaction with prescribed glucocorticoids and greater convenience. We evaluated the utility of salivary cortisone measurement in patients undergoing an overnight dexamethasone suppression test (ONDST).

METHODS

Individuals undergoing an ONDST had parallel measurement of serum cortisol and salivary cortisone at 0900 following midnight dexamethasone (1 mg). Salivary cortisone was measured by electrospray positive liquid chromatography tandem mass spectrometry. The threshold for adequate suppression of salivary cortisone was< 2.7nmol/L; serum cortisol was< 50nmol/L.

RESULTS

Results for 34 individuals which included 21% men (mean age 49.4 years) and 79% women (mean age 56.7 years) were analysed. Serum cortisol did not suppress in 22/34 cases. Salivary cortisone did not suppress in two of the cases where cortisol did suppress. We found a strong correlation between 0900 salivary cortisone and serum cortisol after 1 mg ONDST (r2 = 0.65, p = 0.009). When performance of post-dexamethasone salivary cortisone (< 2.7nmol/L) alone in relation to suppression of serum cortisol (< 50nmol/L) was analysed all but 2 individuals were correctly classified. They had values for post dexamethasone salivary cortisone/serum cortisol of respectively 5.9/23 nmol/L (presented with unexplained fatigue, case 25) and 7/32 nmol/L (investigated for cyclical Cushing's Syndrome that was excluded, case 29). Agreement was 94.1%, kappa 0.87, p < 0.0001. The sensitivity of salivary cortisone for potential Cushing's syndrome as indicated by the post-dexamethasone 0900 serum cortisol was 100% (all cases of potential cortisol excess (0900 cortisol > 50nmol/L) were picked up) and specificity of 84.6% with a positive predictive value of 90.5% of salivary cortisone (using serum cortisol as the standard) and negative predictive value of 100% in relation to ruling out cortisol excess.

CONCLUSION

We have provided further evidence that ONDST salivary cortisone has potential to be the first-line test for suspected Cushing's syndrome, not requiring venepuncture or attendance at hospital, with 100% sensitivity and reasonable specificity. Application of the salivary cortisone test has the potential for significant savings of money and time in this and other contexts.