A Risk Score for Type 1 Diabetes Derived From Autoantibody-Positive Participants in the Diabetes Prevention Trial–Type 1

Open Access
Publisher Website
Google Scholar
Abstract
OBJECTIVE—The accurate prediction of type 1 diabetes is essential for appropriately identifying prevention trial participants. Thus, we have developed a risk score for the prediction of type 1 diabetes. RESEARCH DESIGN AND METHODS—Diabetes Prevention Trial–Type 1 (DPT-1) participants, islet cell autoantibody (ICA)-positive relatives of type 1 diabetic patients (n = 670), were randomly divided into development and validation samples. Risk score values were calculated for the validation sample from development sample model coefficients obtained through forward stepwise proportional hazards regression. RESULTS—A risk score based on a model including log-BMI, age, log-fasting C-peptide, and postchallenge glucose and C-peptide sums from 2-h oral glucose tolerance tests (OGTTs) was derived from the development sample. The baseline risk score strongly predicted type 1 diabetes in the validation sample (χ2 = 82.3, P < 0.001). Its strength of prediction was almost the same (χ2 = 83.3) as a risk score additionally dependent on a decreased first-phase insulin response variable from intravenous glucose tolerance tests (IVGTTs). Biochemical autoantibodies did not contribute significantly to the risk score model. A final type 1 diabetes risk score was then derived from all participants with the same variables as those in the development sample model. The change in the type 1 diabetes risk score from baseline to 1 year was in itself also highly predictive of type 1 diabetes (P < 0.001). CONCLUSIONS—A risk score based on age, BMI, and OGTT indexes, without dependence on IVGTTs or additional autoantibodies, appears to accurately predict type 1 diabetes in ICA-positive relatives.