Evening salivary cortisone for evaluation of hypercortisolaemia: evaluation and implications

Cortisol secretion follows a circadian rhythm, and loss of this diurnal variation with sustained hypercortisolemia is strongly suggestive of Cushing’s syndrome [1]. Cushing’s syndrome is associated with an elevated cardiovascular event rate [2] as is the diagnosis of mild autonomous cortisol secretion [3]. While serum cortisol day curves are informative, they are health care resource heavy and not convenient for patients in routine clinical practice. Salivary cortisone, a stable and noninvasive biomarker reflecting circulating free cortisol, may offer a convenient and patient-friendly alternative for assessing cortisol dynamics outside hospital settings, because there is a significant fast conversion of cortisol to cortisone in salivary glands by 11-beta hydroxysteroid dehydrogenase II, leading to a high concentration of cortisone in saliva [4], which mirrors serum cortisol. Salivary cortisone shows a close relation to serum cortisol [5,6].

We previously demonstrated the utility of salivary cortisone measurement both as an alternative to post-dexamethasone morning serum cortisol and of waking salivary cortisone as an alternative to the corticotropin stimulation test [7,8]. The current study evaluated the performance of evening salivary cortisone in patients investigated for suspected Cushing’s syndrome or mild autonomous cortisol secretion (MACS).

We report the results of a service evaluation. The study was a prospective screening accuracy study and aimed to evaluate the performance of evening salivary cortisone levels in the diagnosis of Cushing’s syndrome or MACS. The study period was 6 months. The correlation between evening salivary cortisone and 48-h low dose dexamethasone suppression test (LDDST) serum cortisol was also evaluated. Thirty-two consecutive adult patients attending a single tertiary endocrine centre in the UK were enrolled for the study. For evaluation of suspected Cushing’s syndrome or MACS, all patients provided one or more late-night salivary samples between 2100 and 0000 h. Salivary cortisone concentrations were quantified by liquid chromatography–tandem mass spectrometry (LC-MS/MS), and mean values were calculated where multiple measurements were available. Patients underwent the 48-h LDDST as the reference standard [9]. Index and reference test results were analyzed without blinding. Data analysis was conducted on SPSS ver. 25 (IBM, Chicago). Comparison of means was performed using analysis of variance. Diagnostic accuracy of salivary cortisone was estimated by receiver operating characteristic (ROC) analysis. Optimal thresholds were identified using Youden’s J statistic (J = sensitivity + specificity − 1).

The mean ± SD age of patients was 55.4 ± 14.7 years and 63% were female. Of the 32 patients recruited, 8 were diagnosed with Cushing’s syndrome and 11 with MACS, whilst the remaining 13 had normal adrenocortical status. All 8 patients with Cushing’s had undergone an LDDST. Evening salivary cortisone differed significantly across diagnostic groups (mean ± SD: Cushing’s 36.6 ± 18.5 nmol/L; MACS 10.1 ± 3.9 nmol/L; normal 4.9 ± 2.2 nmol/L; F (2,29), P < 0.001; Fig. 1a). Tukey least significance difference post hoc analysis revealed significantly greater mean salivary cortisone levels in Cushing’s patients compared with both MACS and normal patients (P < 0.001). For Cushing’s syndrome, the ROC area under the curve (AUC) was 0.974 (95% CI: 0.894–1.000) and the optimal threshold of salivary cortisone for Cushing’s syndrome diagnosis was determined as 17.4 nmol/L (Table 1).

Fig. 1.

Diagnostic performance of evening salivary cortisone. (a) Salivary cortisone levels by diagnostic group. (b) Correlation between salivary cortisone and 48-h LDDST serum cortisol. LDDST, low dose dexamethasone suppression test.

Table 1.

Diagnostic accuracy of evening salivary cortisone

Diagnosis	Best threshold (nmol/L)	Sensitivity	Specificity	Youden’s J
Cushing’s syndrome	17.4	0.875	0.958	0.833
MACS	7.0	0.933	0.917	0.850

MACS, mild autonomous cortisol secretion.

The positive predictive value (PPV) and negative predictive value (NPV) of evening salivary cortisone in screening for Cushing’s were 87.5 and 95.8%, respectively.

For MACS, sensitivity was 0.933 and specificity was 0.917. AUC was 0.983 (95% CI: 0.925–1.000), with an optimal salivary cortisone threshold of 7.0 nmol/L. The PPV and NPV of salivary cortisone for screening for MACS were 83.3 and 91.7%, respectively.

Salivary cortisone correlated strongly with post-LDDST serum cortisol (Pearson r = 0.86, P < 0.001; Fig. 1b). No adverse events were reported.

This prospective single-centre study demonstrates the potential utility of evening salivary cortisone for detecting hypercortisolism. The test performance was comparable to the 48-h LDDST, with AUCs > 0.97 for both Cushing’s syndrome and MACS.

The results are in keeping with previous reports [10,11] and support salivary cortisone as a noninvasive, convenient screening tool, particularly suited for outpatient or telemedicine-based endocrine assessment. In particular, Efthymiadis et al. [1] published the first study to compare late-night salivary cortisol, late‐night salivary cortisone, urinary free cortisol, overnight dexamathasone suppression test (ONDST), and to ascertain which had the best diagnostic capability for identifying Cushing’s syndrome over Cushing’s disease or mild autonomous cortisol secretion. They reported that late-night salivary cortisol offered comparable performance, superior to urinary free cortisol in screening for hypercortisolism. The paper proposed a simple screening algorithm based on late-night salivary cortisone in combination with ONDST for screening moderate/high clinical pre‐test probability patients and ONDST for adrenal incidentalomata.

The diagnosis of new or recurrent Cushing’s syndrome requires careful evaluation of clinical findings with biochemical confirmation. Since sensitivity should be maximized in screening tests, the excellent sensitivity and NPV of late evening salivary cortisone make it a potentially valuable screening tool.

We accept that not all endocrinology services at present have access to salivary cortisone estimation. However, given the considerable advantages inherent in this approach, it is anticipated that over time, access to the LC-MS assay will be much wider in health care settings across the world.

Regarding limitations, blinding was not performed and sample size was modest. However, the findings are consistent with prior data validating late-night salivary biomarkers. Larger multicentre studies should confirm optimal diagnostic thresholds and evaluate cost-effectiveness.

Our study was strengthened by its prospective design and recruitment of consecutive patients. We also used LC-MS/MS rather than an enzyme immunoassay for a more precise quantification of salivary cortisone levels. Our study was limited by its moderate sample size and its conduct in a specialist centre as opposed to a primary care setting where screening would be most appropriate in a real-world setting.

We suggest late evening salivary cortisone as a practical and easily applied alternative to the LDDST in the diagnostic workup for suspected hypercortisolism/Cushing’s syndrome so that these individuals can be identified for definitive treatment and cardiovascular risk reduction measures [12].

Acknowledgements

Conflicts of interest

There are no conflicts of interest.