Aims/hypothesis. We aimed to evaluate how an aetiology-based classification, as recommended in the ADA and WHO guidelines for classification of diabetes mellitus, matches clinical judgement in the Diabetes Incidence Study in Sweden (DISS), a study covering incident cases of diabetic patients aged 15 to 34 years. Methods. During a 1-year period (1998), blood samples were taken at diagnosis and 4 months (median) thereafter. Patients were classified according to clinical judgement by the reporting physicians and assessments of islet antibodies (ICA, GADA, and IA-2A) and plasma C-peptide. Results. In 1998, 422 patients were registered in DISS. Among the 313 patients participating in the follow-up, most with clinical Type 1 diabetes (185/218, 85%, 95% CI 79-89%) were islet antibody positive (ab+) at diagnosis. In addition, 14 out of 58 (24%, 14-37%) with clinical Type 2 diabetes and 21 out of 37 (57%, 40-73%) with unclassifiable diabetes were antibody positive at diagnosis. Further to this, 4 out of 33 (12%, 3-28%) were antibody negative with clinical Type 1 diabetes and 4 out of 44 (9%, 3-22%) with Type 2 had converted to antibody positive at follow-up. Among those who were constantly antibody negative, 10 out of 29 (34%, 18-54%) with clinical Type 1 and 1 out of 16 (6%, 0-30%) with unclassifiable diabetes had fasting plasma C-peptide concentrations below the normal range (<0.25 nmol/l) at follow-up. Conclusion/interpretation. Most young adults with clinical Type 1 diabetes (199/218, 91%) had objective Type 1 (ab+ at diagnosis/follow-up and/or low fasting plasma C-peptide concentrations at follow-up), as did one third (18/58, 31%) with clinical Type 2 diabetes and more than half (22/37, 59%) with unclassifiable diabetes. About 10% of those who were antibody negative converted to antibody positive. Our study underlines that a classification considering aetiology is superior to clinical judgement.
2003. Vol. 46, no 2, 173-181 p.