Body Weight Gain and T1DM - Accelerator Hypothesis
© 2005 Prime Health Consultants, Inc.
Betts and colleagues presented anthropometric data on 168 children diagnosed with type 1 diabetes (T1DM) between the years 1980 and 2002 in an attempt to provide data to support the “Accelerator Hypothesis”. This hypothesis, originally proposed in 2001 by Wilkin (co-author of the current study), postulates that T1DM and T2DM are a single disease rather that 2 distinct disorders with a common final biochemical abnormality - hyperglycemia. The argument is that the 2 types of diabetes differ only by the rate of beta cell loss and the accelerator responsible. Three accelerators are proposed: a potential for B-cell death, insulin resistance, and a genetic predisposition to develop beta cell autoimmunity. The researchers examine the relationship between body mass index (BMI) with associated visceral obesity and the earlier onset and rising incidence of T1DM in children.
The study population included all children under 16 years of age diagnosed with DM between 1980 and 2002 in Southampton, UK for whom records of height, weight, age of diagnosis, duration of disease, and the presence of ketoacidosis were available. Height and weight at diagnosis were those recorded at the first outpatient visit after treatment began (<3 weeks from presentation). It is assumed that these data reflected pre-onset measurements since weight loss from dehydration, etc. had been restored at that time. Height and weight were also recorded 6, 12, and 24 months post-diagnosis. Pre-diagnosis heights and weights were measured at 3 years of age and at school entry were abstracted from a child health community database. BMI was calculated at each visit and converted to SDS scores using 1990 UK standards. Waist circumference, a surrogate indicator of insulin resistance, was measured in each child at their outpatient visits. (Waist circumference SDS charts for children have been published in the UK.) A local control group of 254 primary school children was available for anthropometric measurements.
The study group included 87 boys and 81 girls, with similar ages at diagnosis (1.1 to 15.7 years). Mean height, weight and BMI SDS were not different from population means and specifically not different from the comparison group. However, there were significant positive correlations between weight SDS and BMI SDS, and the year of diagnosis. Those diagnosed more recently tended to be heavier. Substantial weight SDS gain and BMI SDS increase occurred during the first 6 months after diagnosis, but no further significant changes in these parameters were observed over the next 18 months. There were significant positive correlations between weight SDS and BMI SDS at 6 months and the year of diagnosis; significant inverse correlations were noted between weight SDS and BMI SDS and age at diagnosis. The heavier the child, the younger the age of onset.
There were 91 of the 111 children who were positive for islet cell autoimmunity. Their anthropometric determinations were not different from those without autoimmunity. Likewise there were no differences between children who presented with ketoacidosis and those who did not. There were no significant differences between birth weights of the study population and those of the comparison group. A subset of younger children (n=62) who had been born more recently had been routinely measured at ages 3 or 5 years, (0.1 to 9.8 years prior to diagnosis). There was a significant inverse relationship between age of diagnosis and pre-diagnostic BMI SDS suggesting that children who were heavier pre-diagnosis developed diabetes earlier. Waist measurements were available for 90 children; 90% had waist circumferences above the 50th centile, but there was no correlation between waist circumference and duration of diabetes.
The authors stated that their data showed that heavier children develop diabetes at an earlier age, and that these children were heavier even in their pre-school years. These data confirmed previous reports.1 They suggested that the increase in the incidence of T1DM seen in the UK and throughout the world, especially in younger children has occurred too rapidly to be explained by genetic changes and must therefore be explained by environmental factors. Early weight gain and increased BMI in young children may pre-dispose them to early insulin resistance and early beta cell destruction. The implications for prevention of T1DM (and T2DM) are obvious. Lifestyle interventions must be encouraged. Of note, the authors report that although there is an increase in the incidence of childhood onset T1DM among several registries, this may be at the expense of adult incidence. Obtaining documentation of a shifting incidence of T1DM may be difficult, as many registries of T1DM do not include information regarding adult onset T1DM
Editor’s Comment: This is a very provocative paper. Indeed many pediatric endocrinologists are reporting an increase in the incidence of T1DM in young children. Some registries however, suggest that there may be a shift in the age of diagnosis but not a total increase in incidence. Of course, as noted by the authors, many registries don’t track adults along with children. If the age of diagnosis is becoming lower, there must be a reason; hence, the Accelerator Hypothesis. Early weight gain and increasing BMI SDS of young children may lead to increased insulin resistance and an earlier destruction of pancreatic beta cells in a person who may be genetically predetermined to develop T1DM. It’s an interesting hypothesis.
In an accompanying editorial by Daneman2 the arguments for and against the Accelerator Hypothesis are logically presented. Daneman suggests that there are 4 areas of the hypothesis that warrant further scrutiny. The first area is a better study of the impact of the epidemic of obesity on childhood diabetes. The other 3 areas relate to the 3 potential accelerators described by the authors. Clearly this manuscript and the editorial should stimulate lots of interest, discussion, and future studies.
William L. Clarke, MD
References - (linked to 
)