Abstract P076: Regression of Prediabetes and Diabetes is Associated with Mild Weight and Waist Loss - ELSA-Brasil

Abstract

Introduction: The prevalence of diabetes has risen dramatically in the last decades. Prediabetes, or intermediate hyperglycemia, represents an early stage of diabetes development. Structured diabetes prevention programs can lead to regression of prediabetes, but observational studies may reflect better a real-world scenario for public health planning. Hypothesis: We assessed the hypothesis that regression of prediabetes to normoglycemia (and from newly diagnosed diabetes to prediabetes) is associated with weight loss and with decrease in waist circumference in a well-characterized, nonintervention contemporary cohort, ELSA-Brasil. Methods: We enrolled 15105 civil servants (34-75 years) in 6 capital cities in Brazil from 2008-10 and followed those without previously known diabetes for an average of 3.7 years. We performed a standardized oral glucose tolerance test and measured fasting and 2h plasma glucose as well as HbA1c in a central laboratory. Diabetes and prediabetes were defined using American Diabetes Association criteria at each visit. We measured weight and waist circumference with standardized protocols and obtained sample characteristics by questionnaire. Those with abnormal glycemia were instructed to seek medical attention. Results: Of the 6301 participants with prediabetes at baseline not receiving anti-diabetic medication at either visit, 1671 (27%) had regressed to normal values at follow-up. In Poisson regression analyses adjusting for age, sex, race, family history of diabetes and baseline BMI, glycemia and physical activity, in comparison with those with no change or weight gain, those with weight loss ≤2.05kg had a 30% greater probability of returning to normoglycemia (RR=1.30; 95%CI 1.17-1.43) and those with greater weight loss had a 59% greater probability (RR=1.59; 1.44-1.75). Findings were similar for a decreased waist circumference in comparison to no change or a gain (≤2.6cm loss: RR=1.18; 1.05-1.31; >2.6cm loss: RR=1.52; 1.37-1.68). Of the 661 participants newly ascertained with diabetes at baseline and not receiving medication at either visit, 375 (57%) had non-diabetic values at follow-up. Regression among these newly ascertained cases was also greater among those with decreased weight or waist circumference, though only statistically significant for those with a greater decrease in weight (>2.7kg loss: RR=1.39; 1.19-1.62) or waist circumference (>2.6cm loss: RR=1.35; 1.14-1.61) Conclusion: Small decreases in weight or waist circumference in a free-living population not receiving structured interventions to prevent diabetes can lead to important improvement in glycemia, with impact on the prevalence of prediabetes and newly ascertained diabetes.