Endocrine Abstracts

Background: Endogenous Cushing’s syndrome (CS) is caused by overproduction of cortisol. Current first-line screening tests for CS can produce false positive results due to medication use and stress, rely heavily on patient adherence to sampling instructions, and only measure short-term cortisol exposure which limits sensitivity. In general multiple tests are required to establish a diagnosis. Hair cortisol concentrations (HCC) offer a minimally invasive way to measure long-term cortisol exposure over months of time, and are a promising additional screening tool for CS. In addition, HCC can be used to create retrospective timelines of cortisol exposure.

Methods: We collected scalp hair from patients suspected of CS in a single academic medical center, and in healthy controls. HCC were measured using ELISA. Receiver operator curve (ROC) analysis was used to determine the optimal cut-off for CS.

Results: HCC levels were available in 35 patients with CS, 20 patients in whom the diagnosis CS could be rejected during diagnostic work-up (non-CS patients), and 174 healthy controls. CS patients had higher HCC than non-CS patients and healthy controls (157.6 vs 9.7 and 8.4 pg/mg, respectively, P<0.001). At a cut-off of 31.1 pg/mg, HCC could differentiate between CS patients and healthy controls with a sensitivity of 94% and a specificity of 90%. With non-CS patients as a reference, sensitivity and specificity remained similar (94 and 90% respectively). As expected, HCC was higher in ectopic ACTH secretion than in pituitary or adrenal CS (661.1 vs 105.9 and 123.0 pg/mg, respectively, P<0.05). Retrospective timelines of HCC corresponded well with clinical course in CS.

Conclusion: Analysis of cortisol in a single scalp hair sample offers diagnostic accuracy for CS similar to current first line tests and a collection procedure which is highly convenient for clinical practice, and therefore seems a valuable tool in the diagnosis of CS.

Disclosure: This work was supported by an Erasmus MC fellowship grant and The Netherlands Organization for Scientific Research (NWO) grant number 916.96.069.