Abstract
Aims: The aim was to assess trends in incidence of pediatric type 1 diabetes (T1D) using data recorded by the population-based Czech Childhood Diabetes Register over 1989-2009. Methods: New cases of childhood-onset T1D aged 0-14.9 yr were recorded using the EURODIAB protocol by two independent sources with the combined estimated completeness of 98.6%.
The incidence was modeled by Poisson regression, and the effects of age and calendar time on incidence were assessed using piecewise linear functions. Results: A total of 5155 cases was ascertained over 1989-2009 from an average pediatric population of 1.76 million.
Two points of change in the incidence trend were identified by the modeling: in 1995 the incidence accelerated, while in 2001 the growth in incidence significantly slowed down in all ages up to 10 yr. In the youngest age category, 0-4 yr at onset, the rapid average annual rise of 15% over 1996-2001 suddenly changed into stagnation over 2002-2009.
Conclusions: Our data contribute to the notion that long- and intermediate-term predictions from the past incidence developments of incidence are difficult, as abrupt changes in the trend can occur. Caution should be exercised against too far-reaching incidence predictions, even if the population has experienced a previous history of a very fast rise in T1D incidence.