(P22-025-20) Flavonoid Intake Is Inversely Associated with Hemoglobin A1c in Men but Not Women in a Diverse Urban Population

Objectives: Identify relationships between usual flavonoid intake and hemoglobin A1c (HbA1c), a biomarker of diabetes and prediabetes, among African American (AA) and white (W) adults in an economically diverse urban population.

Methods: The Healthy Aging in Neighborhoods of Diversity across the Life Span (HANDLS) study is a fixed cohort of adults 30-64 years of age (at baseline) residing in Baltimore City [n=3,720].  The design is a four way factorial cross of race (AA, W), household poverty status, age, and sex. 3,418 individuals completed at least one dietary recall.  Flavonoid intakes were estimated using the publicly available database, Flavonoid Values for USDA Survey Foods and Beverages 2007-2010. The National Cancer Institute (NCI) Method was employed to estimate usual intake distributions of total flavonoids and six flavonoid classes and then to calculate predicted values for regression calibration. This latter step allowed the use of logistic regression to examine associations between usual flavonoid intake and prevalence of HbA1c values at or above 5.7% while adjusting for confounding variables. The 25^th and 75^th percentiles were selected as points of reference. Significance was determined with Wald chi-square tests.

Results: A total of 1,837 individuals (828 males and 1,009 females) met criteria for inclusion in the analytic sample. In both AA and W men, usual intake of total flavonoids, anthocyanidins, flavan-3-ols, and flavonols was inversely related to levels of HbA1c ≥ 5.7% (p< 0.05). The likelihood of HbA1c values that met or exceeded this threshold among men at the 75^th percentile of total flavonoid intake was more than 30% lower than that of men at the 25^th percentile (AA: OR = 0.67, 95% CI: 0.48, 0.95; W: OR = 0.66, 95% CI: 0.47, 0.95). No relationships between usual flavonoid/flavonoid class intake and HbA1c were observed in women.

Conclusions: Within the selected percentile range, higher usual intake of total flavonoids and several flavonoid classes was associated with lower prevalence of HbA1c values indicative of diabetes/prediabetes among men but not women in this at-risk population, supporting other research that has found beneficial relationships between flavonoid intake and diabetes risk.

Funding Sources: This work is supported by the Intramural Research Program, NIA, NIH, and ARS, USDA.