A comparison of the naloxone test with the insulin stress test in patientsfollowing transsphenoidal surgery

OBJECTIVE To study the reliability of the naloxone test in assessing the hy pothalamic-pituitary-adrenal (H PA) axis in patients following transsphenoi dal pituitary surgery, by comparison with the insulin stress test (IST). Al so, to establish a normal range for the response to naloxone in healthy con trols. DESIGN All patients had the IST performed, soluble insulin 0.1-0.3 U/kg (us ually 0.15 U/kg) being administered intravenously, followed by venous sampl ing for plasma glucose and serum cortisol at 0, 15, 30, 45, 60, 75 and 90 m inutes. Naloxone (125 mu g/kg body weight) was similarly administered to al l patients 6 days later, with sampling at -15, 0, 15, 30, 45, 60, 90 and 12 0 minutes for plasma glucose, serum cortisol and plasma ACTH. The naloxone test was also performed on control subjects using an identical protocol. SUBJECTS Twenty patients (11F/9M) aged 48.8 +/- 2.8 years (mean +/- SE) 4-6 weeks following transsphenoidal pituitary surgery. Twelve normal healthy c ontrol subjects (6F/6M) aged 31.5 +/- 3.0 years, MEASUREMENTS Serum cortisol was measured by radioimmunoassay and plasma ACT H by immunoradiometric assay. RESULTS Adequate hypoglycaemia (less than or equal to 2.0 mmol/l) was achie ved in all patients. Peak cortisol was >550 nmol/l in 17 subjects (range 57 2-867 nmol/l) and a subnormal response observed in 3 (peak cortisol 163-498 nmol/l). In the 17 patients with a normal cortisol response to hypoglycaem ia the response to naloxone was extremely variable, with serum cortisol fal ling in 4 and an increment of less than or equal to 100 nmol/l in 6 others. In the 3 patients who failed the IST, two responded to naloxone and one di d not. In normal controls, 3 of the 12 had little or no response to naloxon e, CONCLUSIONS As the naloxone test fails to produce a response in all normal subjects it is impossible to define a 'normal' response. The numerous discr epancies with the IST in patients are further evidence of the test's lack o f reliability in assessing the HPA-axis and it is consequently not a viable alternative to traditional dynamic methods such as the IST.