January - March 2006, Volume 2, Issue 1

Gheorghiu ML, Galoiu S, Caragheorgheopol A, Coculescu M

Limits of dexamethasone supression tests in the diagnosis of Cushing's syndrome

Acta Endo (Buc) 2006, 2 (1): 45-61
doi: 10.4183/aeb.2006.45

Introduction: The dexamethasone suppression tests (DST) in the diagnosis of Cushing's syndrome (CS) give frequently equivocal results. Our study evaluated the precision of DST in the diagnosis of CS. Patients and methods: 223 patients (15 - 77 years, 130 F / 93 M) were studied for putative CS by morning and midnight serum cortisol, urinary free cortisol (UFC) and 17 hydroxycorticosteroids (17OHCS) levels at baseline and after DST as follows: 1 mg overnight (oDST), 0.5mg q.d., 2 days (LDDST) and 2 mg q.d. 2 days (HDDST). Since not all cases were evaluated by all tests, statistical analysis used available results. Results: 79 patients had CS (47 pituitary, 3 ectopic, 21 adrenal adenoma and 8 adrenal carcinoma), 45 had adrenal tumor without all the criteria for CS (NCT) and 99 were controls. All patients with CS had abnormal cortisol biorhythm, but also 31.8% patients with NCT and 50% evaluated controls. The best basal screening test for CS is UFC, with a cut-off value of 100 g/24 h. For DST with a serum cortisol cutoff level (CCL) of 5 ?g/dl, oDST correctly diagnosed all CS, while lowering the CCL at 1.8 ?g/dl increased the false positive rate to 6.8%. At LDDST, the serum CCL of 5 ?g/dl correctly identified all patients with CS, but was 2.5 % false positive in controls. A significant correlation of serum cortisol values after oDST or LDDST with basal 17OHCS and UFC was found (r=0.6, p<0.001), suggesting that patients with mild CS are more prone to test as false negatives. The classical criterion of 50% suppression for UFC after LDDST correctly identified 12/14 CS patients and all 8 controls. Better sensitivity (Sn) had an UFC cutoff level of 10?g/24h, p<0.001. A 50% suppression of 17OHCS identified 45/61 CS patients and excluded the disease in 30/37 patients. In HDDST, serum cortisol suppression by > 50% diagnosed Cushing's disease (CD) with 85 % Sn and 57% specificity (Sp). The best HDDST accuracy had UFC (83%), with 27% false negative results (3/11 patients) for a cut-off value of\r\n50% from baseline. 17OHCS were suppressed by 50% in 19/35 patients with CD (54%) and in 3/33 (9%) patients with other causes of CS. NCT patients had higher basal values of UFC and 17OHCS than controls (p<0.01) and up to 50% had an abnormal biorhythm, suggesting a degree of hypercortisolism, with an inadequate oDST or LDDST . Therefore, only 6/45 NCT fulfilled the criteria for subclinical CS.\r\nConclusion: The best screening test was oDST (100% Sn and Sp), followed by UFC (100% Sn and 92% Sp). LDDST with serum CCL at 5 ?g/dl had 100% Sn and 97% Sp. HDDST identified CD by UFC with 73% Sn and 92% Sp and by serum cortisol with 85% Sn and 57% Sp. For NCT, the standard tests identified SCS in 13%. However, a thorough evaluation including multiple tests should be undertaken for the positive and differential diagnosis of Cushing's syndrome.

Keywords: Cushing?s syndrome, dexamethasone, cortisol, adrenal tumor

Correspondence: Monica Gheorghiu ?C. I. Parhon? Institute of Endocrinology, Bd. Aviatorilor 34-36, Bucharest, Romania, fax: +4021.3198718, e-mail:monicagheorghiu@yahoo.com