The influence of the fasting blood glucose (FBG) concentration on the beta-cell responsiveness to glucagon was studied twice in 9 insulin-dependent diabetic patients with residual beta-cell function. At a FBG of 7.7 +/- 0.3 mmol/l (mean +/- SEM) all patients displayed a preserved beta-cell function with a plasma C-peptide concentration of 0.25 +/- 0.03 nmol/l 6 min after an i.v. injection of glucagon. In contrast, at a FBG of 3.2 +/- 0.01 mmol/l 4 out of 9 patients would have been classified as not having an endogenous insulin secretion. All the patients had the lowest 6 min C-peptide concentration 0.06 +/- 0.01 nmol/l on the day with the lowest blood glucose concentration. The reproducibility of the glucagon test was assessed by comparing the results from two test days in 12 insulin-dependent, 9 non-insulin-dependent diabetic patients and 6 normal subjects. The 6 min plasma C-peptide concentration, the peak plasma C-peptide concentration, and the area under the plasma C-peptide curves were not different on the two test days in any subgroup. In all diabetic patients and normal subjects, the 6 min plasma C-peptide concentration (r = 0.93, coefficient of variation (CV) = 0.16), the peak plasma C-peptide concentration (r = 0.93, CV = 0.16) and the area under the plasma C-peptide curve (r = 0.94, CV = 0.16) were all significantly correlated. The results show that the prevailing FBG significantly affects the outcome of the glucagon test and confirm its reproducibility.