Undetectable postoperative cortisol does not always predict long-term remission in Cushing's disease: a single centre audit

Objective An undetectable postoperative serum cortisol has been regarded as a definition of cure in Cushing's disease. However, we noted disease recurrence amongst patients with Cushing's disease despite undetectable postoperative cortisol levels, and this led us to audit our data. We have also previously assessed surgical outcome for acromegaly and microprolactinoma for a single surgeon. The alms of this study were twofold: (1) to investigate the treatment and surgical outcome of patients with Cushing's disease. In particular, we wished to compare the data with outcome for other pituitary tumours in our centre; and (ii) to determine whether undetectable cortisol following surgery is predictive of long-term cure for Cushing's disease. Patients and methods We performed a retrospective audit of 97 patients; mean age 39.1 (range: 14-82) years, 78/97 (80.4%) female, mean follow-up 92 months (range: 6 months to 29 years), with Cushing's disease seen in our unit between 1969 and 1998. We documented diagnostic investigation, immediate surgical outcome and disease recurrence in these patients. Results All patients had elevated urinary free cortisol (mean 1270.6 nmol/l, range: 327-3245 nmol/l). In total, 95.5% of patients did not suppress with low-dose dexamethasone suppression testing. Hypokalaemia (K<3.2 mmol/l) was present in 15.6% of patients; 17.5% of patients did not show cortisol suppression with high-dose dexamethasone and 15.8% of patients did not show an ACTH rise of >50% following corticotrophic releasing hormone (CRH) administration. There was no significant (>3) gradient in ACTH or cortisol following CRH during inferior petrosal sinus sampling in 27.3% of patients who had the test. A pituitary tumour was demonstrated on imaging in 55.8% of patients; 10.3% were macroadenomas. Mortality rate following transsphenoidal surgery was 1%. Following surgery, the immediate postoperative remission rate (undetectable postoperative cortisol) was 68.5%. However, 11.5% of these patients developed disease recurrence during a mean follow-up period of 36.3 months. Considering microadenomas, Cushing's disease patients had an immediate postoperative remission rate of 63.2% which is significantly lower (P<0.05)compared to a remission rate of 91.1% in acromegaly. Additionally, new postoperative gonadotrophin deficiency (13.9%) and TSH deficiency (25.8%) was higher in patients with Cushing's disease compared to patients with acromegaly or microprolactinoma. Immediate postoperative remission rates improved from 50% in the first decade of a surgeon's career to consistently above 60% in the second and third decades, demonstrating a trend which may be attributed to surgical experience. Conclusions (i) Despite strict criteria for immediate postoperative remission and recurrence, undetectable postoperative cortisol is not always predictive of long-term remission. (ii) Despite an aggressive surgical approach, immediate postoperative remission rates for Cushing's disease are lower compared to other microadenomas. The development of new pituitary hormonal deficiency following surgery is also commoner than that seen amongst other microadenomas. These data have important implications for the follow-up of patients with Cushing's disease.