A Two-Step Strategy for Identification of High-Risk Subjects for a Clinical Trial of Prevention of NIDDM

Abstract

OBJECTIVE To evaluate 2-h plasma glucose (2HPG), fasting plasma glucose (FPG), glycosylated hemoglobin (HbA1), a combination of FPG and HbA1 (FPG-HbA1), and other factors as screening tests for identifying high-risk subjects for a clinical trial of prevention of NIDDM and to identify strategies to minimize the total number of oral glucose tolerance tests (OGTTs) required to recruit eligible subjects to the trial. RESEARCH DESIGN AND METHODS One thousand, one hundred and eight nondiabetic Pima Indians aged 25–64 years were followed for up to 5 years, and factors predicting NIDDM, defined by World Health Organization criteria (2HPG ≥ 11.1 mmol/l), were assessed using Cox's proportional hazards analysis. Various threshold values of FPG, HbA1, and FPG-HbA1 were determined, which, when combined with an OGTT, identified subjects with impaired glucose tolerance (IGT) (2HPG ≥ 7.8 and < 11.1 mmol/l) at a number of specified risks for developing NIDDM in 5 years. The value of each of the three tests was then assessed by calculating (for each threshold) the numbers to be screened, the numbers requiring an OGTT, and the sample size of IGT subjects needed to detect a 33% reduction in NIDDM by an experimental intervention at a power of 80%. RESULTS During a median of 4.3 years of follow-up, 91 (8.2%) of the 1.108 nondiabetic subjects developed NIDDM. The estimated 5-year cumulative incidence rate was 13.5%. Each of the variables, 2HPG, FPG, HbA1, FPG-HbA1, BMI, IGT, and systolic (sBP), diastolic (dBP), and mean (MBP) blood pressures, predicted NIDDM (P < 0.05 for each) when controlled for age and sex. In a stepwise proportional hazards analysis model, 2HPG and FPG-HbA1 (P < 0.001 for each) were selected as the best set of predictors of NIDDM and of fasting hyperglycemia (FPG ≥ 7.8 mmol/l). CONCLUSIONS A two-step strategy, in which high-risk individuals are first identified by FPG or FPG-HbA1 and then the OGTT is used to select subjects with IGT, requires fewer OGTTs than when using 2HPG as the initial screening test without substantially increasing the numbers that would need to be screened. Such a strategy also offers the advantage of reducing the necessary sample size and is therefore an effective, efficient, and convenient method of identifying eligible subjects for a clinical trial of prevention of NIDDM.