• accelerator hypothesis;
  • BMI;
  • body mass index;
  • type 1 diabetes mellitus;
  • weight gain


Aim of this study was to test Wilkin's ‘accelerator hypothesis’: whether excessive weight gain accelerates the onset of type 1 diabetes.

Subjects and methods

Anthropometric birth data of 1117 children who developed diabetes between 1988 and April 2013 were compared with those of a sex, age, and gestational age matched, contemporary regional control group (n = 54 344). Cases were divided into three manifestation groups (G1:0–4.9 yr, G2:5–9.9 yr, and G3: 10–20 yr). Furthermore, growth data of 540 children with diabetes were compared with controls (n = 134 249) in pre-, peri-, and post-onset intervals (interval: 1–6). Also, correlation of age at onset and body mass index (BMI) standard deviation score (SDS) at this point of time were examined.


Cases had significantly higher SDSs for birth weight when compared with controls (boys: p = 0.007, girls: p = 0.002). Children with early manifestation had the highest mean of birth weight SDS (G1>G2>G3), (p = 0.22, adjusted r2 = 0.001). BMI SDS trend curves of cases are slightly higher compared with those of the healthy controls. This was only significant in years after diagnosis (interval 6, p < 0.000). Cases did not show excessive weight gain in any of the examined intervals before the onset of diagnosis (interval 1–3). One year after diagnosis, we found an inverse correlation between age at diagnosis and BMI SDS at diabetes manifestation. The youngest children at diagnosis (G1) had the lowest BMI SDS at manifestation and vice versa (G1<G2<G3, p = 0.06).


Our data do not support the ‘accelerator hypothesis’. There was no sign of excessive weight gain before manifestation. Discrepant results from other studies could be due to non-age-adjusted controls.