Day 5 morning serum cortisol predicts hypothalamic-pituitary-adrenal function after transsphenoidal surgery for pituitary tumors.

BACKGROUND Adrenal insufficiency is a complication of transsphenoidal surgery (TSS) for pituitary adenoma, and correct identification of patients requiring glucocorticoid replacement is important. Controversy exists over which early postoperative 9 AM cortisol concentration reliably predicts hypothalamic-pituitary-adrenal (HPA) axis reserve, as defined by the insulin tolerance test (ITT). METHODS Data were reviewed for 36 patients undergoing TSS followed by day 5 postoperative 9 AM cortisol measurement and ITT 6 weeks postsurgery. All patients received postoperative glucocorticoid replacement, which was discontinued if the 9 AM serum cortisol was >300 nmol/L. RESULTS Of 23 patients who failed the ITT (peak cortisol <500 nmol/L), 20 also had a day 5, 9 AM serum cortisol <300 nmol/L. Nine of 13 patients who passed the ITT had a day 5, 9 AM cortisol >300 nmol/L. The cutoff cortisol concentration of 300 nmol/L had 86.9% (66.4%-97.2%) diagnostic sensitivity, 69.2% (38.6%-90.9%) diagnostic specificity, and 83.3% (61.8%-94.5%) positive predictive value (PPV) for detecting secondary adrenal insufficiency. Increasing the cutoff to 392 nmol/L resulted in 100% (85.2%-100%) sensitivity, 46.1% (19.2%-74.9%) specificity, and 76.6% (57.3%-89.4%) PPV. Decreasing the cutoff to 111 nmol/L resulted in 100% (75.3%-100%) specificity and 100% (67.9%-100%) PPV, although sensitivity was 47.8% (26.8%-69.4%). CONCLUSIONS A day 5 post-TSS 9 AM serum cortisol <111 nmol/L reliably detects secondary adrenal insufficiency, and concentrations >392 nmol/L support intact HPA function. Because concentrations of 111-392 nmol/L are poorly predictive of HPA function, glucocorticoid replacement should continue in such cases until definitive testing is performed using an ITT.