The efficacy of the standard high dose ACTH stimulation test (HDT), using a pharmacological 250-microg dose of synthetic ACTH-(1-24), in the diagnosis of central hypoadrenalism is controversial. The insulin hypoglycemia test is widely regarded as the gold standard dynamic stimulation test of the hypothalamo-pituitary-adrenal (HPA) axis that provides the most reliable assessment of HPA axis integrity and reserve. Alternatively, a prolonged infusion of ACTH causes a continuing rise in plasma cortisol levels that may predict the adrenals' capacity to respond to severe ongoing stress. In nine normal subjects, we compared plasma ACTH and cortisol levels produced by three i.v. bolus low doses of ACTH-(1-24) (0.1, 0.5, and 1.0 microg/1.73 m2; LDTs) with those stimulated by hypoglycemia (0.15 U/kg insulin) and with the cortisol response to a standard 250-microg dose of ACTH-(1-24). The normal cortisol response to an 8-h ACTH-(1-24) infusion (250 microg at a constant rate over 8 h) was determined using three modern cortisol assays: a high pressure liquid chromatography method (HPLC), a fluorescence polarization immunoassay (FPIA), and a standard RIA. In the LDTs, stepwise increases in mean peak plasma ACTH were observed (12.4 +/- 2.0, 48.2 +/- 7.2, 120.2 +/- 15.5 pmol/L for the 0.1-, 0.5-, and 1.0-microg LDTs, respectively; P values all <0.0022 when comparing peak values between tests). The peak plasma ACTH level after insulin-induced hypoglycemia was significantly lower than that produced in the 1.0-microg LDT (69.6 +/- 9.3 vs. 120.2 +/- 15.5 pmol/L; P < 0.0002), but was higher than that obtained during the 0.5-microg LDT (69.6 +/- 9.3 vs. 48.2 +/- 7.2 pmol/L; P < 0.02). In the LDTs, statistically different, dose-dependent increases in peak cortisol concentration occurred (355 +/- 16, 432 +/- 13, and 482 +/- 23 nmol/L; greatest P value is 0.0283 for comparisons between all tests). The peak cortisol levels achieved during the LDTs were very different from those during the HDT (mean peak cortisol, 580 +/- 27 nmol/L; all P values <0.00009. However, the mean 30 min response in the 1.0-microg LDT did not differ from that in the HDT (471 +/- 22 vs. 492 +/- 22 nmol/L; P = 0.2). In the 8-h ACTH infusion test, plasma cortisol concentrations progressively increased, reaching peak levels much higher than those in the HDT [995 +/- 50 vs. 580 +/- 27 nmol/L (HPLC) and 1326 +/- 100 vs 759 +/- 31 nmol/L (FPIA)]. Significant differences in the basal, 1 h, and peak cortisol levels as determined by the three different assay methods (HPLC, FPIA, and RIA) were observed in the 8-h infusion tests. Similarly, in the HDTs there were significant differences in the mean 30 and 60 min cortisol levels as measured by HPLC compared with those determined by FPIA. We conclude that up to 30 min postinjection, 1.0 microg/1.73 m2 ACTH-(1-24) stimulates maximal adrenocortical secretion. Similar lower normal limits at 30 min may be applied in the 1.0-microg LDT and the HDT, but not when lower doses of ACTH-(1-24) are administered. The peak plasma ACTH level produced in the 1.0-microg LDT is higher than in the insulin hypoglycemia test, but is of the same order of magnitude. The peak cortisol concentration obtained during an 8-h synthetic ACTH-(1-24) infusion is considerably higher than that stimulated by a standard bolus 250-microg dose, potentially providing a means of evaluating the adrenocortical capacity to maintain maximal cortisol secretion. Appropriate interpretation of any of these tests of HPA axis function relies on the accurate determination of normal response ranges, which may vary significantly depending on the cortisol assay used.