An exploration of factors associated with pre-diabetes in the REasons for Geographic and Racial Differences in Stroke (REGARDS) study cohort
Pre-diabetes is a major contributor to increased morbidity and mortality in the US. Without proper and timely intervention the majority of individuals with pre-diabetes will develop Type 2 diabetes. African Americans (AA) and people living in the Stroke Belt have increased odds of developing pre-diabetes compared to Whites and people living outside the Stroke Belt. While several covariates have been identified as contributors to the increased odds of pre-diabetes there is little consensus on which covariates may be responsible for the disparities of pre-diabetes. The reasons for disparities in pre-diabetes are complex and require further investigation. This study explored the relationship between multiple covariates in order to identify which factors might contribute to disparities in pre-diabetes in AA compared to Whites. A secondary data analysis of 19,889 normoglycemic and pre-diabetes participants from the REGARDS study was used. All participants with fasting blood glucose above 100mg/dL and below 126mg/dL were included in the study. The response variable was pre-diabetes, and covariates of interest were race, region, age, gender, income level, education level, smoking history, alcohol history, intense physical activity history, and illness perception. Univariate and multivariable logistic regression with a set of incremental models was used to identify variables associated with increased odds of pre-diabetes.
The mean age of individuals with pre-diabetes living in or outside the stroke belt was 64 years vs.65 year, as well as mean ages among AAs and Whites were 63 years vs. 65 years. The final analysis population was 37% AA males, 63% AA females, 48% White males, and 52% White females. Individuals with diabetes were excluded from the study. AAs had increased odds of pre-diabetes compared to Whites. Individuals living in the Stroke Belt also had increased odds of pre-diabetes compared to individuals not living in the Stroke Belt. Covariates of interest that increased the odds of pre-diabetes were gender (male), education level, past and current smoking history, heavy alcohol history, and illness perception.