Long-term patterns in seasonality of insulin-dependent diabetes mellitus diagnosis in Austrian children

J Clin Epidemiol. 1997 Feb;50(2):159-65. doi: 10.1016/s0895-4356(96)00334-4.

Abstract

The analysis of the seasonal pattern of incidence of childhood insulin-dependent diabetes mellitus in Austria was carried out among cases where the child was under the age of 15 when diagnosed between 1979 and 1993. The cases are registered in the nationwide population-based Austrian insulin-dependent diabetes mellitus registry. Seasonal variation was compared between boys and girls and between three 5-year age groups. We also tested whether the seasonal pattern changed over the 15-year observation period. We found a significant seasonal variation among boys aged 10-14 and girls aged 5-14, while in the 0-4 years age group no seasonal pattern could be demonstrated. Two peaks in incidence were identified during a calendar year (February-March and September-October) for girls aged 5-14. For boys aged 10-14, a yearly cycle was found with a peak in January and October. An extension of the Poisson regression model for testing seasonality by Jones et al. was developed to allow for estimation of a time-dependent amplitude of the seasonal component. The annual incidence rate increased by 36% during the observation period, but no significant change in seasonal pattern could be demonstrated.

MeSH terms

  • Adolescent
  • Austria / epidemiology
  • Child
  • Child, Preschool
  • Diabetes Mellitus, Type 1 / epidemiology*
  • Female
  • Humans
  • Incidence
  • Infant
  • Male
  • Models, Statistical
  • Periodicity
  • Poisson Distribution
  • Seasons*