The postoperative course of patients with subclinical hypercortisolism (SH) is yet to be clarified. The aims are to review the prevalence and predictive factors of postoperative adrenal insufficiency and the time to recover a normal adrenocortical function in patients with SH and Cushing's syndrome (CS).

Using the PubMed database, we conducted a systematic review of the literature, selecting studies published from 1980 to 2013.

Of the 1522 papers screened, 28 were selected (13 retrospective, 14 prospective, and one randomized controlled trial). The prevalence of postoperative adrenal insufficiency was 65.3% in 248 SH subjects and 99.7% in 377 CS patients. Patients with SH were reclassified according to the following diagnostic criteria: subjects defined by pathological dexamethasone test only (DEX), and those defined by the dexamethasone test with one (DEX+1) or two additional criteria (DEX+2); and they were compared with CS patients. The prevalence of adrenal insufficiency was 51.4, 60.6, 91.3, and 99.7%, respectively, with no significant difference between the two latter groups. The test with the best compromise between sensitivity (64%) and specificity (81%) in predicting adrenal insufficiency was the midnight serum cortisol. The time to achieve eucortisolism was lower in SH patients than in CS patients (6.5 vs 11.2 mo; P < .001).

Adrenal insufficiency occurs in about half of the patients with SH if defined only by the pathological dexamethasone test. However, prevalence of adrenal insufficiency and time to recovery are tightly related to the degree of hypercortisolism and diagnostic criteria to define SH, which might help to better define SH for future studies.