We thank Dr. Wilkin (1) for his valuable comments. Overall, we (2) did not observe the hypothesized association between increasing BMI and younger age at onset of diabetes among U.S. youth with autoimmune diabetes. Our results were similar to a report from Birmingham, U.K. (3), and one from Philadelphia (4). They were in contrast to three previous European studies (5,6,7). However, we did observe the inverse association among youth with low residual insulin secretion at diagnosis (fasting C-peptide [FCP] <0.5 ng/ml). We hypothesize that obesity is “accelerating” the onset of the disease at a later stage in the natural history of the diabetes process, after substantial autoimmune destruction of β-cells has occurred.
The discrepancy between our results and those of the European studies may be due to several factors. It is possible that youth in Europe are diagnosed at a later stage in their natural history, when most have low residual FCP. This cannot be addressed, since these studies did not measure FCP. In the U.S., diagnosis may occur at an earlier stage in the natural history, before complete β-cell destruction occurs. Evidence for this exists, since a lower proportion of cases now present in diabetic ketoacidosis than previously reported (8,9). In these youth, acceleration, as assessed using age at clinical diagnosis, may be impossible to document since the time of true disease onset is uncertain.
Dr. Wilkin suggests that the two groups of youth in our Fig. 2 have different HLA genes. According to Dr. Wilkin, a higher proportion of high-risk HLA genes would trigger a more intense insulin resistance–induced autoimmune destruction. This hypothesis is testable in longitudinal studies starting before the onset of autoimmunity. We have collected HLA genotype data in SEARCH; however, none of the previous cross-sectional reports, including the current one, are able to address the impact of HLA genotype in acceleration if the true age of disease onset is uncertain. Even if Dr. Wilkin is correct in predicting that the low-FCP group has more high-risk HLA genes, then this would not provide evidence that insulin resistance triggers the autoimmune process. The two groups in Fig. 2 had identical autoantibody numbers and titers. If insulin resistance triggers an autoimmune response, then part of the acceleration induced by insulin resistance could be mediated through an increased autoimmune destruction. We did not find evidence for such mediation. No important associations between BMI and diabetes autoantibodies were noted. Our findings do not support the hypothesis that insulin resistance triggers the autoimmune process.
There is increasing evidence that obesity and insulin resistance accelerate the onset of type 1 diabetes. The increasing prevalence of childhood obesity (10) may substantially account for the younger age at diagnosis of type 1 diabetes. However, only longitudinal studies starting before the onset of autoimmunity will be able to address the mechanism(s) by which and critical period(s) when increased insulin resistance may accelerate the type 1 diabetes process.
