[Skip to Navigation]
May 2006

Implications for Primary Care of Diabetes and Impaired Fasting Glucose Prevalence in Adolescents

Arch Pediatr Adolesc Med. 2006;160(5):550-552. doi:10.1001/archpedi.160.5.550

In his article in this issue, Duncan1 estimates the prevalence for type 1 diabetes mellitus (T1DM) and type 2 diabetes mellitus (T2DM) from self-reports by 4370 adolescents aged 12 to 19 years in the National Health and Nutrition Examination Survey, and frequency of impaired fasting glucose in a tested subsample of subjects. A fundamental problem in interpreting the diabetes prevalence data is the case definition of T1DM and of T2DM. During the past several decades, we have gone from classifying the main forms of diabetes by age group (juvenile onset vs maturity onset) to differentiating by treatment (insulin-dependent vs noninsulin-dependent) to the contemporary separation by etiology.2 Type 1 diabetes mellitus is a disease predominantly of autoimmune-mediated insulin deficiency, whereas T2DM results from a combination of insulin resistance and inadequate islet cell capacity to meet the increased insulin demands. The case definition in this study, which lacks validation from medical records, harks back to classification by treatment; those who reported using insulin were considered to have T1DM and those who reported that they were not using insulin were considered to have T2DM. The concern about case definition based on insulin use is exemplified by the recognition that pediatric endocrinologists treat approximately 50% of their patients who have T2DM with insulin, either as monotherapy or in conjunction with oral hypoglycemic agents. Two of 10 subjects in Duncan's study were taking insulin and thus classified as having T1DM, but they were also taking oral agents, making it likely that they had T2DM. Only 1 of the 8 subjects who self-reported diabetes and that they were not taking insulin reported that they were taking an oral hypoglycemic agent. The diagnosis of diabetes is therefore questionable in 7 of these subjects because only about 10% of pediatric patients with T2DM can control their disease for more than a brief period with lifestyle change alone.3,4 These concerns are magnified when the small numbers of purported subjects with diabetes are stratified by ethnicity for each diabetes type, resulting in wide 95% confidence intervals for the estimates.1

Add or change institution