Islet Cell Antibodies Are Less Predictive of IDDM Among Unaffected Children in the General Population Than in Sibs of Children With Diabetes
- Mikael Knip, MD,
- Jukka Karjalainen, MD,
- Hans K Åkerblom, MD and
- The Childhood Diabetes in Finland Study Group
- Medical School and the Department of Pediatrics, Tampere University Hospital, University of Tampere Tampere
- Department of Pediatrics, University of Oulu Oulu
- Children's Hospital, University of Helsinki Helsinki, Finland
- Address correspondence and reprint requests to Mikael Knip, MD, Medical School, University of Tampere, P.O. Box 607, FIN-33101 Tampere, Finland. E-mail: llmikn{at}uta.fi
Abstract
OBJECTIVE To examine the controversial issue of whether islet cell antibodies (ICAs) have a higher predictive value for progression to clinical IDDM in first-degree relatives of patients with diabetes than in the general population.
RESEARCH DESIGN AND METHODS ICAs were analyzed with standard immunofluorescence in two population-based groups: 765 sibs of children with recent-onset diabetes and 1,212 unaffected Finnish children <20 years of age at initial screening. Those positive for ICAs were additionally tested for antibodies to GAD (GADAs) and the protein tyrosine phosphatase-related 1A-2 antigen (IA-2As). Subsequently, these subjects were observed for the manifestation of clinical IDDM over the next 7 years.
RESULTS The frequency of both detectable ICAs and ICA levels ≥20 Juvenile Diabetes Foundation units (JDF U) was significantly higher among the sibs than in the general population (7.8 vs. 4.1% and 4.8 vs. 2.0%, respectively; P < 0.001). The prevalence of GADAs (37/60 vs. 3/48; P < 0.001) and IA-2As (31/60 vs. 0/48; P < 0.001) was increased among ICA-positive sibs compared with ICA-positive individuals from the background population. Over the next 7 years, 24 sibs (3.1%) and 3 unrelated children positive for ICAs (0.3%) progressed to clinical diabetes. The positive predictive value of ICAs was thus 6% in the general population and 40% among the sibs (P < 0.001), or 13 and 59%, respectively (P < 0.001), with an antibody cutoff level of 20 JDF U. The positive predictive value was related to the number of positive autoantibodies in sibs, which was 57% in those with three antibodies, 50% in those with two antibodies, and only 6% in those with ICAs alone.
CONCLUSIONS These data show that the frequency of multiple autoantibodies is substantially lower in ICA-positive children representing the general population than in ICA-positive sibs of children with IDDM. As a consequence, the predictive value of ICAs for IDDM is higher in sibs of affected children than in the general population. This finding must be taken into account when planning intervention trials aimed at preventing or delaying the manifestation of clinical diabetes in individuals from the general population who test positive for ICAs.
- Received October 6, 1997.
- Revision received June 17, 1998.
- Accepted June 17, 1998.
- Copyright © 1998 by the American Diabetes Association
