Comparison of efficacy between the serum cortisol and 24 hour urine free cortisol in combined dexamethasone suppression test in the diagnosis of Cushing syndrome

Abstract

To compare the sensitivity and specificity between the 24 hour urine free cortisol (24 h UFC) and serum cortisol in dexamethasone suppression test (DST) in the diagnosis of Cushing syndrome (CS). Combined low dose DST (LDDST) and high dose DST (HDDST) were carried out in 67 cases of CS with surgically confirmed cases in recent 3 years(from January 2011 to November 2015). The serum cortisol and 24 h UFC were collected simultaneously for each subject and the sensitivity and specificity of serum cortisol and 24 h UFC were compared. There were Cushing disease (CD) group (n=53), ectopic adrenocorticotropic hormone (ACTH) syndrome group (n=7) and ACTH-independent Cushing syndrome group (n=7) according to the etiology of hypercorticordism.There were no significant differences among 3 groups in gender and age.The sensitivity of serum cortisol of different cut off points(50, 110, 140 nmol/L and 50% of control)after LDDST was 97.01%, 86.57%, 83.58% and 70.15% respectively.Meanwhile, the sensitivity of cutoff point of 24 h UFC <32 nmol in combined LDDST was 92.54% in the diagnosis of Cushing syndrome.There was no significant differences in two groups between serum cortisol <110 nmol/L and 24 h UFC <32 nmol.However, the sensitivity of serum cortisol <50 nmol/L was significantly higher than 24 h UFC<32 nmol (P<0.05). Furthermore, in combined HDDST, if the suppression rate was ≥50%, the sensitivity of serum cortisol and 24 h UFC in differentiating the etiology of Cushing disease was 60.38% and 90.57%, and the specificity was 91.43% and 96.00% respectively.There were significant differences between serum cortisol and 24 h UFC in both of sensitivity and specificity (both P<0.05). In addition, if the suppression rate of 24 h UFC in HDDST was adjusted to 60.85% according to receiver operating characteristic (ROC) curve, it could have the best levels of sensitivity (92.6%) with the specificity of 85.7%. If the suppression rate of serum cortisol was adjusted to 61.53% in HDDST according to ROC curve, it could have the best sensitivity (64.8%) with the specificity of 78.6% accordingly. In combined LDDST, the serum cortisol <50 nmol/L had a higher sensitivity than the 24 h UFC<32 nmol when they were used as the criteria in determining the diagnosis of CS.In HDDST, the sensitivity and specificity of suppression rate of 24 h UFC ≥50% were better than serum cortisol to differentiate the etiology of CS.