The Diabetes Incidence Study in Sweden (DISS): Ascertainment, risk factors, and problems with classification

Sammanfattning: The Diabetes Incidence Study in Sweden (DISS) registers diabetes in young adults aged 15-34 years in Sweden. This thesis aims to validate the ascertainment, methods in the classification of diabetes, and putative etiological factors such as family background, psychological stress, and BMI, that may affect the incidence of diabetes in young adults, in the DISS registry. The capture-recapture method was used in the evaluation of ascertainment, whereas C-peptide and islet antibodies (islet cell antibodies ?ICA? and glutamic acid decarboxylase antibodies ?GADA?) were used in the classification of diabetes. In a case-control study 1992-1993, autoimmune Type 1 diabetic patients were compared with controls. Body mass index (BMI) was calculated from weight and height in the registration forms 1983-1999. Level of ascertainment in the DISS registry was 0.86 for Type 1 diabetes. Clinical characteristics at diagnosis or C-peptide values 2.5-3 years after diagnosis did not clarify whether the type of diabetes was of autoimmune (Type 1 diabetes) origin or not. Six years after diagnosis of diabetes in the 1987-1988 cohort, 72% of patients at diagnosis considered to have Type 2 or unclassifiable diabetes, were insulin-treated and ICA or GADA at diagnosis were present in 74% of this group of patients compared with only 12% in those still non-insulin treated. The positive predictive values for later insulin treatment were for ICA and GADA 100%, for ICA alone 98%, and for GADA alone 95%. In the case-control study, heredity for diabetes (OR 2.6) and a positive association of being born of Swedish mothers were associated with a development of diabetes. Major stress factors, however, were not associated with the development of diabetes. In addition, BMI at diagnosis but not the incidence of diabetes increased during the years 1983-1999. In conclusion, islet antibodies improved the classification of diabetes. Autoimmune Type 1 diabetic patients were more exposed to heredity for diabetes but no major prediabetic stress factors or environmental factors were detected. BMI at diagnosis but not the incidence of diabetes had increased during 1983-1999. Ascertainment of Type 1 diabetic patients was adequate in DISS.