Endocrine Abstracts

Purpose: According to modern guidelines late night serum cortisol measurement is not considered to be among the first steps for screening of hypercortisolism. But in some cases it may be useful as it allows identifying the loss of normal circadian rhythm of cortisol secretion. This test is not rarely used in hospitalized patients when Cushing’s syndrome is suspected. But the cutt-off point of nighttime serum cortisol level for the diagnosis of hypercortisolism is inconsistent yet for this category of patients.

Aim: To evaluate the nighttime serum cortisol level in hospitalized patients who were examined due to suspicion on Cushing’s syndrome and to determine the cutt-off point of nighttime plasma cortisol level for the diagnosis of hypercortisolism.

Materials and methods: For the purpose of the study central laboratory and clinic database was analyzed. Results of 120 hospitalized patients (38 women and 82 men, mean age 51+1) were included. In the clinic of our center normally nighttime serum cortisol sampling is performed in awake patients at least 48 hours after submission between 2300 h and 2400 h Serum cortisol was measured with enzyme immunoassay (Roche Diagnostic). The results of patients with BMI >35 kg/m²; glycosylated hemoglobin >8%, uncontrolled hypertension, severe cardiovascular disease and other severe conditions were excluded from the analysis. The data on examination for the hypercortisolism confirmation were studied. Cushing’s syndrome had been considered to be established if at least two recommended screening tests were positive. Receiver-operating-characteristic (ROC) curve analysis was used to determine an optimal threshold value of nighttime serum cortisol level for diagnosis of hypercortisolism.

Results: Cushing’s syndrome was excluded in 97 patients. In 23 cases hypercortisolism was confirmed (in 15 patients – Cushing’s disease, in 8 patients – ACTH independent Cushing’s syndrome). According to the ROC curve analysis, the optimal threshold value of nighttime serum cortisol level for diagnosis of hypercortisolism was 169.45 nmol/l. The sensitivity and specificity of the method were 87% and 76.3%, respectively, the probability of incorrect prognosis (P=0.0001).

Conclusion: According to our data the cutt-off point of nighttime serum cortisol level for the diagnosis of hypercortisolism in hospitalized patients was 169.45 nmol/l. It may be used in complex examination, but considering the data on sensitivity and specificity of the method other screening tests are necessary for the confirmation of hypercortisolism in hospitalized patients.