The insulin tolerance test (ITT) has long been used to assess the hypothalamic-pituitary-adrenal axis, but may be hazardous. The standard synthetic ACTH (Synacthen) test has been advocated as a substitute but is sometimes insensitive. In this study the ITT has been compared to a low dose ACTH stimulation test (1 microg) and the standard ACTH stimulation test (250 microg). Twenty-seven subjects were studied, 24 with verified or suspected hypothalamic-pituitary disorders and three on long-term glucocorticoid therapy. Insulin tolerance, low dose ACTH and standard ACTH tests were performed in all patients. The ITT was performed less than 48 hours after the ACTH tests. Synacthen was administered as an intravenous bolus. Serum cortisol values were determined by radioimmunoassay. The peak cortisol value during ITT was compared to the cortisol levels during the ACTH tests. There was a highly significant correlation between peak cortisol values during ITT and cortisol levels after 20-60 minutes in the low dose ACTH test (r(s) = 0.91-0.93; P < 0.0001) and after 30 and 60 minutes in the standard ACTH test (r(s) = 0.85 and 0.89 respectively; P < 0.0001). Four patients showed discrepancies between the three tests. The 1-microg ACTH test follows the ITT more closely and may be more sensitive than the standard ACTH test in detecting more subtle insufficiency of the hypothalamic-pituitary-adrenal axis. The standard ACTH test and the insulin tolerance test may thus be replaced by the 1-microg ACTH test in screening for secondary cortisol insufficiency. We recommend that serum cortisol is measured before and 30 and 40 minutes after the ACTH injection.