Conference abstract

Fasting Plasma Glucose ≥100 mg/dL As a Single Predictor of Diabetes in Africans is Equivalent to the A1C-Modified ARIC Diabetes Prediction Equation: The Africans in America Study

Pan African Medical Journal - Conference Proceedings. 2021:11(10).10 Dec 2021.
doi: 10.11604/pamj-cp.2021.11.10.1037

Contact the corresponding author
Keywords: Diabetes, African-descent populations, prediction Equations, ARIC
Poster

Fasting Plasma Glucose ≥100 mg/dL As a Single Predictor of Diabetes in Africans is Equivalent to the A1C-Modified ARIC Diabetes Prediction Equation: The Africans in America Study

Margrethe F Horlyck-Romanovsky1,&, Regine Mugeni1, Jessica Y Aduwo1, Sara M Briker1, Thomas Hormenu1, Christopher DuBose1, Anne E Sumner1,2

1National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD, United States, 2National Institute of Minority Health and Health Disparities, National Institutes of Health, Bethesda, MD, United States

&Corresponding author

Background: The A1C-modified Atherosclerosis Risk in Communities (ARIC) diabetes prediction equation was optimized in African Americans. Our goal was to test in 422 African-born blacks living in America (65% male; age 38±10y (mean ±SD); BMI 27.5±4.4kg/m2): 1) the ability of this equation to detect undiagnosed diabetes; 2) the contribution of each biochemical variable to the equation; 3) the independent value of each biochemical variable.

Methods: The equation had 5 clinical variables (age, family history, systolic BP, waist circumference and height); and 4 biochemical variables (HDL, TG, A1C and fasting plasma glucose (FPG)). Oral glucose tolerance tests were performed. Diabetes diagnosis required 2h glucose ≥200mg/dL. Area under the receiver operator characteristics curve (AROC) predicted diabetes. Youden Index identified optimal cut points. Three models tested: Model 1: Full equation; Model 2: All 5 clinical variables with 1 &biochemical variable excluded at a time; and Model 3: Each biochemical variable singly.

Results: Diabetes occurred in 6% (27/422). Model 1 had excellent prediction (AROC 0.84, referent). Model 2, excluding A1C, HDL or TG showed no difference (AROC 0.84-0.89, P>0.10). Only when excluding FPG did AROC (0.76, P=0.01) decrease. Model 3, single predictors HDL, TG and A1C all had lower prediction (AROC 0.63-0.73, P≤0.01). FPG as single predictor was equivalent to the full equation (AROC 0.89, P=0.25). Optimal cut point for FPG was 100mg/dL. The full equation identified diabetes in 78% (21/27) and FPG≥100mg/dL in 81% (22/27).

Conclusions: For the detection of diabetes in Africans, FPG≥100mg/dL performed as well as the A1C-modified ARIC equation.