Abstract
BACKGROUND: Previous studies have shown that US estimates of prediabetes or diabetes differ depending on test type, fasting plasma glucose (FPG) vs hemoglobin A1c (HbA1c). Given age, race, and test differences reported in the literature, we sought to further examine these differences in prediabetes detection using a nationally representative sample. METHODS: Using the National Health and Nutrition Examination Survey (NHANES) 1999-2016, individuals were identified as having prediabetes with an HbA1c of 5.7% to 6.4% or a FPG of 100 to 125 mg/dL. We excluded individuals with measurements in the diabetic range. We ran generalized estimating equation logistic regressions to examine the relationship between age, race, and test type with interactions, controlling for sex and body mass index. We compared the difference in predicted prediabetes prevalence detected by impaired fasting glycemia (IFG) vs HbA1c by race/ethnicity among children and adults separately using adjusted Wald tests. RESULTS: The absolute difference in predicted prediabetes detected by IFG vs HbA1c was 19.9% for white adolescents, 0% for black adolescents, and 20.1% for Hispanic adolescents; 21.4% for white adults, -1.2% for black adults, and 19.2% for Hispanic adults. Using adjusted Wald tests, we found the absolute differences between black vs white and black vs Hispanic individuals to be significant, but, not between Hispanic and white individuals among children and adults separately. CONCLUSIONS: These observations highlight differences in test performance among racial/ethnic groups. Our findings corroborate the need for further studies to determine appropriate HbA1c cutoff levels for diagnosis of prediabetes by age group and race.