ORIGINAL RESEARCH: Opposing Effects of Waist and Extremity Circumferences on the Risk of Incident Type 2 Diabetes

Abstract

ABSTRACT OBJECTIVE: To investigate the relationship between anthropometric circumferences and incident type 2 diabetes in Canadian adults. METHODS: The sample included 516 men and 579 women from the Physical Activity Longitudinal Study. Waist, hip, thigh and upper arm circumferences were used to predict incident diabetes over 15.5 years of follow-up. The odds of developing diabetes associated with 1 standard deviation of each anthropometric measure was calculated using sex-specific logistic regression, adjusted for age and parental history of diabetes. RESULTS: There were 56 new cases of diabetes, yielding a 15.5 year cumulative incidence of 5.1%. With the exception of arm and thigh circumferences in men, all of the anthro-pometric variables were positive predictors of diabetes. However, after adjustment for body mass index (BMI) and waist circumference (WC), the nature of the associations between hip, thigh and arm circumferences and diabetes reversed, such that they became inversely associated with diabetes development. The positive association between WC and incident diabetes was independent of all other circumference measures but was attenuated by BMI. CONCLUSIONS: Larger extremity circumferences are associated with a lower risk of incident diabetes after adjustment for overall and abdominal adiposity, highlighting the potential value of these measures above and beyond the routine measurement of BMI and WC.