To evaluate the use of serum 1,5-anhydroglucitol (AG) levels in screening for diabetes mellitus, we compared the sensitivity and specificity of HbA1c, fructosamine (FA), and AG in 1620 randomly selected subjects in 11 institutions throughout Japan. Most individuals were receiving diet and/or drug therapy for diabetes. Subjects were separated into four groups based on World Health Organization criteria: nondiabetic control subjects, subjects with impaired glucose tolerance (IGT), patients with diabetes, and patients with other disorders without IGT. The overlap of AG values between each group was less than that of HbA1c or FA values. AG levels were significantly correlated with fasting plasma glucose (r = -0.627), HbA1c (r = -0.629), and FA (r = -0.590) levels. If we took 14 micrograms/ml as the normal lower limit, AG level was highly specific (93.1%), and a decreased AG level indicated diabetes mellitus (84.2% sensitivity). According to the selectivity index (sensitivity value times specificity value), AG determinations were superior to both HbA1c and FA measurements for diabetes screening. When combinations of these tests were used, only AG and HbA1c together were slightly better than AG alone. Thus, together with other advantages of AG, e.g., its wide variance with relatively fair glycemic control and the negligible influence of the sampling conditions, AG level has more potential than HbA1c or FA level as a screening criterion for diabetes.

No abstract

Growth hormone (GH) responses to growth-hormone-releasing hormone (GRH) and thyrotropin-releasing hormone (TRH) were studied in 17 diabetic patients. Ten patients (group 1) had retinopathy corresponding to stage III-V (Scott's classification), and the remaining seven patients (group 2) had no retinopathy despite longer duration of diabetes in comparison with the patients in group 1. There were no differences in age, percent of ideal body weight, and serum HbA1 levels between the two groups. Basal serum GH levels were 1.9 +/- 0.4 ng/ml (mean +/- SEM) in group 1, and not different from the values in group 2 (1.6 +/- 0.7 ng/ml). However, GH responses to synthetic human GRH-44 (1 micrograms/kg body wt, i.v. bolus) were significantly greater in group 1, as judged by the maximal response or integrated GH secretion after the administration of GRH. There were no differences in serum insulin-like growth factor I (IGF-I) levels between group 1 (262 +/- 35 ng/ml) and group 2 (232 +/- 30 ng/ml), and no significant correlation was found between serum IGF-I levels and GH responses to GRH in either of the two groups. Paradoxical GH responses to TRH (500 micrograms, i.v. bolus) were found in only one patient in each group. We have thus demonstrated that GH responses to GRH are more pronounced in diabetic patients with retinopathy than in patients without this complication, although it remains to be determined whether or not greater GH responses to GRH are causally related to the development of diabetic retinopathy.

No abstract