Abstract P152: Do Self-reported Race, Socioeconomic Status And Genetic Ancestry Influence Fasting Glucose? Preliminary Results From the Ongoing Bach Prediabetes Study

Abstract

Background: Race/ethnic disparities in type 2 diabetes (T2D) may have both biophysical and social determinants that likely arise before T2D diagnosis. We examined the association of self-reported (SR) race/ethnicity, socioeconomic status (SES) and genetic ancestry with levels of fasting glucose (FG) in individuals not known to have T2D. Methods: The Boston Area Community Health (BACH) Prediabetes study is an ongoing epidemiologic survey of community-dwelling residents of inner-city Boston. A stratified random sample of 3,300 subjects are being drawn from the parent BACH cohort study, recruiting approximately equal numbers in pre-specified age (30–79 years), sex and race/ethnicity groups. Fasting blood is collected, including FG, other health and SES information, during an early AM in-home interview. SES is measured as a standardized composite score of income and education. Genetic ancestry informative markers (AIMs, n=63) that discriminate European, African and Native American continental ancestry are currently available on a random subset. We compared FG levels across groups using ANOVA, testing independent associations of SR race/ethnicity SES using multiple linear regression with significant P=< 0.05. Results: Of 2,933 participants recruited, 63% were women and 70% were older than 50 years. Black, Hispanic and white SR race/ethnicity comprised 32%, 33% and 35% of the sample, respectively. The majority had low SES (57%), 35% had moderate SES and 8% had high SES. FG varied significantly by SR race/ethnicity (mean±SD, mg/dL): 113±37 for white, 119±46 for Hispanic and 121±49 for black subjects (P<0.001). FG was inversely associated with SES (low 122±49, medium 113±39, high 106±23, P<0.001). In joint SR race/ethnicity and SES models adjusting for age and sex, both main effects were attenuated but remained associated with FG (SR race/ethnicity P=0.01; SES P<0.001). For the subsample with AIMs data (n=373), predominant AIMs ancestry, defined as the ancestry most likely as predicted by AIMs, was closely linked to SR race/ethnicity. European ancestry was predominant in 100% of 110 SR white respondents, while African ancestry was predominant in 95% of 129 SR black respondents. In 134 SR Hispanics, 43% were predominantly European, 41% African and 16% Native American. FG varied with predominant AIMs ancestry (European 115±45 mg/dL, African 121±53 and Native American 127±67), although not significantly so (P=0.33), likely due to the limited sample size. Conclusions: Both self-reported race/ethnicity and SES are associated with FG in community-based individuals without known T2D, and they share explanatory information reflecting a common effect. A larger sample with AIMs will help disentangle the degree to which FG variation is socially patterned in the community versus a biophysical phenomenon determined in part by genetic ancestry.