Long-term remission rates after pituitary surgery for Cushing's disease: the need for long-term surveillance.

OBJECTIVE: There have been a few reports on long-term remission rates after apparent early remission following pituitary surgery in the management of Cushing's disease. An undetectable postoperative serum cortisol has been regarded as the result most likely to predict long-term remission. Our objective was to assess the relapse rates in patients who underwent transsphenoidal surgery in order to determine whether undetectable cortisol following surgery was predictive of long-term remission and whether it was possible to have long-term remission when early morning cortisol was measurable but not grossly elevated. Endocrinological factors associated with late relapse were also studied. PATIENTS: We reviewed the long-term outcome in 63 patients who had pituitary surgery for the treatment of Cushing's disease between 1979 and 2000. MEASUREMENTS: Case notes were reviewed and the current clinical and biochemical status assessed. Our usual practice was that early after the operation, an 08:00 h serum cortisol was measured 24 h after the last dose of hydrocortisone. This was followed by a formal low-dose dexamethasone suppression test. Current clinical status and recent 24-h urinary free cortisol values were used as an index of activity of the Cushing's disease. If there was evidence suggesting relapse, a low-dose dexamethasone suppression test was performed. In many patients, sequential collections of early morning urine specimens for urinary cortisol to creatinine ratio were also performed in an attempt to diagnose cyclical and intermittent forms of recurrent hypercortisolism. We did this if there was conflicting endocrine data, or if patients were slow to lose abnormal clinical features.
RESULTS: Mean age at diagnosis was 40.3 years (range 14-70 years). Mean follow-up up time was 9.6 years (range 1-21 years). Forty-five patients (9 males/36 females) achieved apparent remission immediately after surgery and were subsequently studied long term. Of these 45 patients, four have subsequently died while in remission from hypercortisolism. Ten of the remaining 41 patients have relapsed. Of those 10, six demonstrated definite cyclical cortisol secretion. Two of the 10 had undetectable basal serum cortisol levels in the immediate postoperative period. Thirty-one patients are still alive and in remission. Fourteen (45%) of the 31 who remained in remission had detectable serum cortisol levels (> 50 nmol/l) immediately postoperatively, and remain in remission after a mean of 8.8 years. Our relapse rate was therefore 10/45 (22%), after a mean follow-up time of 9.6 years, with mean time to relapse 5.3 years.
CONCLUSIONS: The overall remission rate of 56% (35/63) at 9.6 years follow-up is disappointing and merits some re-appraisal of the widely accepted principle that pituitary surgery must be the initial treatment of choice in pituitary-dependent Cushing's syndrome. Following pituitary surgery, careful ongoing expert endocrine assessment is mandatory as the incidence of relapse increases with time and also with increasing rigour of the endocrine evaluation. A significant number of our patients were shown to have relapsed with a cyclical form of hypercortisolism.