Article Text

  1. H. Liu,
  2. J. Cabaccan,
  3. H. Raff,
  4. J. Findling,
  5. E. Ryzen
  1. Palo Alto, CA.


Background Although endogenous Cushing's syndrome (CS) is not common, its prevalence may be increased in high-risk populations, including diabetic and obese patients. Late-night salivary cortisol (SC) determination is a simple test and may be ideally suited for screening for CS.

Methods Male veterans with type 2 diabetes mellitus were recruited from endocrine referral clinics. 157 male veterans with diabetes not on exogenous glucocorticoids received two test kits for late-night (2300h) SC outpatient screening of CS. 53 male obese veterans without diabetes served as controls. Participants with persistently positive SC results (≥ 4.3 nmol/l) received additional testing to confirm hypercortisolemia, including 24-hour urinary free cortisol (24h UFC, ≥ 60 mcg/d) and low-dose dexamethasone suppression testing (LDDST, serum cortisol ≥ 1.8 mcg/dl). Participants with positive secondary testing underwent serum ACTH testing and imaging to determine the source of hypercortisolemia.

Results 144 patients and 47 controls returned samples for testing (92% and 89% participation rate, respectively). Diabetic patients and controls were well-matched in terms of age (patients vs controls, 61.9 vs 59.5 years), BMI (34.4 vs 34.5 kg/m2), systolic blood pressure (137 vs 138 mmHg), and history of prior psychiatric illness (33% vs 41%) (p≥0.05 for each variable). The mean initial SC levels were 4.2±5.7 nmol/l for diabetic patients and 3.3±6.8 nmol/l for controls (p = 0.20). Although there was no statistically-significant correlation between SC levels and age, BMI, or HbA1C, 31 diabetic patients and only 4 controls (21.5% and 8.5%, respectively, p = 0.02) required further testing to confirm hypercortisolemia. 25 of 31 diabetic patients with elevated late-night SC results agreed to further testing. 18 patients of the 25 diabetic patients (72%) had normal 24h UFC and LDDST results, suggesting false positive SC results. 3 patients (12%) had serum cortisol levels greater than 1.8 mcg/dl on LDDST and normal 24h UFC levels; all had ACTH levels ≥ 20 pg/ml. 4 patients (16%) are still undergoing evaluation. Of the 3 patients with elevated LDDST results, 2 had unremarkable pituitary MRIs and 1 was unable to tolerate MRI; they are being followed for the development of CS. No cases of CS have been diagnosed to date, although it has not been ruled out in all participants.

Conclusion Late-night SC testing is a simple test that can be utilized in screening for CS. More research, however, is needed to determine which, if any, population should be screened for CS. For our older male veteran population, late-night SC may not be an optimal screening test for CS due to high rates of false positive results.

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.