Age-adjusted analysis of insulin responses during normal and abnormal glucose tolerance tests in children and adolescents.

This report analyzes age-specific glucose (PG) and immunoassayable insulin (IRI) responses during oral glucose tolerance testing (OGTT) and examines test results in children and adolescents with OGTT abnormality using the age-appropriate control data. Controls' (n = 93) and patients' (n = 63) results were compared on the basis of statural age (SA) at the time of testing. Control tests (n = 101) showed significant positive correlation of fasting and four-hour postingestion PG with SA (p less than 0.001), but mean area under the PG curves did not vary between the SA groups (I--18.69 months, II--70-131 months, III--132+ months). The absence of differences of other sampling times permits uniform diagnostic criteria for this age group. IRI was positively correlated with SA at all testing times, and mean levels differed significantly between each SA group at every sampling time; the mean areas under the IRI curve also differed significantly between SA groups as did the mean ratios of IRI area to PG area (group I--0.2639 +/- 0.0175 S.E.M., group II--0.3864 +/- 0.0235, group III--0.6262 +/-0.0491). Patient tests (n = 110) were separated into normal (N), borderline (B), and chemical-diabetic (C) for each SA group. IRI means were above control data for each test type in each SA group at all sampling times; one fourth of these differences were significant. IRI responses also increased within each SA group from N to B to C tests. Mean IRI areas and IRI area to PG are mean ratios were higher than in controls, and this difference was greatest with the most abnormal (C) test type in each SA group. A subgroup of three patients who had low IRI responses from the outset and developed overt diabetes in one to three years was excluded from the analysis. In contrast to apparent relative insulin inefficiency with normal maturation and with chemical diabetes, they had exceptional responsiveness to their low IRI levels. Variable involvement of alpha as well as beta cells in the pathophysiology of diabetes is suggested as one explanation for these paradoxic observations. Changing receptor affinity might also be implicated.