Reproducibility of the insulin tolerance test (ITT) for assessment of growth hormone and cortisol secretion in normal and hypopituitary adult men

OBJECTIVE The within subject variability of the insulin tolerance test (ITT) for assessment of growth hormone (GH) status and cortisol reserve has rarely been examined, particularly in patients with hypopituitarism. This becomes important when biochemical criteria are used to determine which adults with hypopituitarism should receive GH and/or cortisol replacement. In the present study we assessed the reproducibility of GH and cortisol responses in repeated ITTs. Baseline insulin-like growth factor 1 (IGF-1) levels were also assessed for reproducibility on each occasion. DESIGN and PATIENTS Three consecutive ITTs were performed in seven normal adult men (ages 22-27 years) and two ITTs in 11 men with hypopituitarism and suspected GH deficiency (ages 23-48 years). MEASUREMENTS Serum GH and IGF-1 were measured by immunoradiometric and cortisol by immunofluorimetric assays. RESULTS In normal men group peak GH responses did not differ between the three tests. There was no correlation between individual peak values. The within subject peak GH variability was between 4.6 and 59.3%, and the overall variability in 21 tests was 35%. The lowest peak GH concentration was 70 mU/l (27 mug/l). All hypopituitary men had severe GH deficiency (all peak GH concentrations < 4 mU/l (1.5 mug/l) in both tests). There was a highly significant correlation between individual peak GH values (r = 0.95, P < 0.0001). Basal IGF-1-values in normal and hypopituitary men were highly correlated between tests (r = 0.98, P < 0.0001). The overall within subject variability of IGF-1-values was 11.9% in normal and 22.7% in hypopituitary men. In normal men group peak cortisol responses were not different between the three tests. There was a good correlation between individual peak cortisol responses in the three ITTs. The within subject peak cortisol variability (median 8.3%; range 0.7-21.5%) was significantly less than that of GH (P < 0.03) in two of three test comparisons. In hypopituitary men the within subject peak cortisol variability (median 41.6%; range 3.5-92.7%) was significantly greater (P < 0.001) than in normal men. All patients were correctly classified as cortisol deficient or normal in both ITTs. CONCLUSION The cortisol response to repeated hypoglycaemia is very reproducible in normal men but the GH response is less so. In hypopituitary men the reproducibility of the GH response is good while that of the cortisol response is poor. However, a single ITT did not misclassify hypopituitary patients who are severely GH and/or ACTH deficient and was therefore adequate for clinical decisions regarding GH and/or cortisol replacement. Nevertheless, it remains possible that a single ITT could misclassify some hypopituitary patients with partial GH or ACTH deficiency.