ADA Scientific Sessions 2026 Abstract · Jun 5, 2026

Equity-Sensitive Prediabetes Screening: Discordant Phenotypes across HbA1c, FPG, and OGTT in NHANES 2005–2016

By FRASER R, WALKER RJ, CAMPBELL JA et al.

Introduction and Objective: Prediabetes classification varies by test, shaping risk profiling and equity monitoring. We estimated national prevalence of glycated hemoglobin (HbA1c)-, fasting plasma glucose (FPG)-, and two-hour oral glucose tolerance test (OGTT)-defined prediabetes phenotypes and identified correlates of phenotype membership. Methods: We analyzed National Health and Nutrition Examination Survey 2005-2016 adults aged 18+ with mobile examination center data and complete HbA1c, FPG, and OGTT measures, excluding pregnancy and diabetes. Phenotypes were HbA1c-only, FPG-only, OGTT-only, or concordant (≥2 tests in range), with normoglycemia as the reference category. Prevalence was estimated; multinomial logistic regression produced adjusted relative risk ratios (aRRRs), fully adjusted for demographic, socioeconomic, and cardiometabolic factors. Results: The sample (n=11,091) represented 182.9 million nonpregnant U.S. adults without diagnosed diabetes. Weighted prevalence was 47.2% normoglycemia, 6.8% HbA1c-only, 21.7% FPG-only, 3.7% OGTT-only, and 20.7% concordant. In fully adjusted models, older adults showed the strongest associations: age 75+ (vs 18-54) was associated with HbA1c-only (aRRR 5.23, 95% CI 2.98-9.18), OGTT-only (7.73, 3.09-19.32), and concordant (5.18, 3.09-8.67) prediabetes. Non-Hispanic Black adults (vs non-Hispanic White) had higher risk of HbA1c-only (4.30, 2.61-7.08) and concordant (1.67, 1.10-2.53) prediabetes. Women had lower FPG-only prediabetes (0.65, 0.43-0.98), and antihypertensive medication use was associated with HbA1c-only (2.52, 1.56-4.05) and concordant (2.41, 1.52-3.83) prediabetes. HbA1c-only screening would miss 25.4% (FPG-only & OGTT-only). Conclusion: Discordant phenotypes are common and patterned by age and race/ethnicity. Reliance on a single test may miss substantial dysglycemia and alter follow-up triage screening. Screening workflows should account for test-dependent phenotypes, including HbA1c-only risk among non-Hispanic Black adults. Disclosure: R. Fraser: None. R.J. Walker: None. J.A. Campbell: None. O. Ekwunife: None. L.E. Egede: None. Funding: American Diabetes Association (7-25-ICTSPC-0412)

doi.org/10.2337/db26-2241-pPopulations

AgingResearch.News