Abstract
Aim: To examine whether the association between overweight and obesity with 2-h plasma glucose (2hPG) and HbA1c levels differs according to educational attainment. Method: Using cross-sectional baseline data from the Australian Obesity, Diabetes, and Lifestyle study in 2000 (n = 8576), we performed multivariable linear regression analysis adjusted for confounding factors and stratified by education. We performed a log-likelihood test to see whether the model including the interaction between education and body mass index (BMI)/waist circumference (WC) predicted the outcome better than the model without the interaction. Analyses were repeated stratified by sex and in never smokers. Results: Increased BMI/WC was associated with increased 2hPG and HbA1c levels. No moderating effect of education on the relationship between BMI with 2hPG and HbA1c was observed in the total population, or in males or females. However, in a subpopulation of never smokers, effect modification by education was observed, particularly when WC was used as the exposure. The association between obesity with 2hPG and HbA1c was stronger in people with a lower level of education. For example, the increase in HbA1C associated with very increased risk WC was 1.1 (95% CI 0.81–1.29) in the secondary only group compared to 0.61 (95% CI 0.25–0.96) in the degree group. Conclusions: To confirm these results, this analysis should be repeated using a longitudinal design in a population of non smokers. If the impact of obesity on the risk of diabetes is worse in those with lower education, obesity trends are likely to cause further inequalities in diabetes than currently expected.