Comparison of one week 0900 h serum cortisol, low and standard dose Synacthen tests with a 4 to 6 week insulin hypoglycaemia test after pituitary surgery in assessing HPA axis

OBJECTIVE To compare the use of 0900 h serum cortisol and both low and stan dard dose Synacthen tests, one week after pituitary surgery with an insulin hypoglycaemia test performed 4-6 weeks after surgery in assessing the inte grity of the hypothalamic-pituitary-adrenal (HPA) axis. DESIGN 0900 h basal serum cortisol was measured on days 6 and 7 after pitui tary surgery (24 h off replacement hydrocortisone) followed by a low dose S ynacthen test (1 mu g intravenously) on day 6 and a standard dose Synacthen test (250 mu g intramuscularly) on day 7. Three to 5 weeks later an insuli n hypoglycaemia test was performed on all patients. Both low and standard d ose Synacthen tests were performed on control subjects using an identical p rotocol. SUBJECTS Forty-two patients (21 male, 21 female), median age 49 years (rang e 23-73) who had pituitary surgery (Cushing's disease excluded). One patien t had undergone repeat surgery for residual tumour and was studied followin g each operation. Sixteen healthy normal volunteers, median age 37 years (r ange 21-55). MEASUREMENTS Serum cortisol measured by radioimmunoassay. RESULTS Two standard deviations below the mean serum cortisol (logarithmic transformation) in the normal volunteers 30 minutes after low dose Synacthe n (1 mu g) was 496 nmol/l and after standard dose Synacthen (250 mu g) was 504 nmol/l. A normal response was therefore taken as serum cortisol > 500 n mol/l at 30 minutes in both tests (using 496 and 504 nmol/l did not alter r esults). 0900 h serum cortisols 1 week after surgery were > 250 nmol/l in 3 1 patients and 29 of these had a normal response to hypoglycaemia (peak ser um cortisol > 550 nmol/l). Of the remaining two patients, one had 0900 h se rum cortisol on day 6 and 7 after surgery of 405 and 441 nmol/l with a peak serum cortisol response to hypoglycaemia of 451 nmol/l; the other patient had 0900 h serum cortisols of 416 and 251 nmol/l and a peak cortisol respon se to hypoglycaemia of 498 nmol/l. All eight patients who had a 0900 h seru m cortisol < 100 nmol/l failed a subsequent insulin hypoglycaemia test. Sev en discrepancies were noted between the low dose Synacthen test and the ins ulin hypoglycaemia test in the 41 patients who had both tests. In six of th ese, a subnormal response to low dose Synacthen was followed by a normal re sponse to hypoglycaemia. Three discrepancies were noted between the standar d dose Synacthen test and the insulin hypoglycaemia test in the 40 patients who had both tests. In all three cases a normal response to Synacthen was followed by a subnormal response to hypoglycaemia. CONCLUSIONS A 0900 h serum cortisol < 100 nmol/l (24 h off replacement hydr ocortisone) indicated ACTH deficiency and need for lifelong steroid replace ment. A 0900 h serum cortisol > 450 nmol/l one week after pituitary surgery is highly suggestive of a normal cortisol response to hypoglycaemia. A 090 0 h serum cortisol between 250 and 450 nmol/l one week after pituitary surg ery permits safe withdrawal of steroid therapy pending an insulin hypoglyca emia test 1 month after surgery. Patients with 0900 h serum cortisol betwee n 100 and 250 nmol/l should continue replacement steroids until definitive testing. Low dose and standard dose Synacthen tests 1 week after pituitary surgery are unreliable and should not be used.