Impact of hip circumference and height on incident diabetes: results from 6-year follow-up in the Tehran Lipid and Glucose Study

Aim To investigate the effects of height and hip circumference, individually and in combination with waist circumference, on incident diabetes. Methods A total of 1589 men and 2132 women, free of diabetes at baseline aged 20 years participated in a 6-year follow-up examination. The standard 2-h post-challenge plasma glucose test was performed at baseline and during follow-up. The Cox model was used to regress the hazard of diabetes on height and hip circumference individually and in combination with waist circumference. Results Median follow-up was 6 years, with a total of 9433 and 12 607 person-years follow-up among men and women, respectively. The incidence rate of diabetes (95% CIs) was 9.0 (7.6-11.5) and 11.0 (9.6-13.4) per 1000 persons per year in men and women, respectively. Among men, after multiple adjustments, the anthropometric measures were no longer associated with incident diabetes. Among women, the age-adjusted hazard ratio (95% CIs) of diabetes was 0.80 (0.61-1.04) for height, 1.39 (1.21-1.60) for hip circumference and 1.86 (1.59-2.17) for waist circumference. After further adjustment for waist circumference, the hazard ratios were 0.69 (0.53-0.90) for height and 0.76 (0.61-0.96) for hip circumference. These inverse associations resisted multiple adjustments. The hazard ratio for waist circumference increased by 7 and 54% when height and hip circumference were included in the models. Conclusion We observed that height and hip circumference were inversely associated with incident diabetes among women and that these associations could be explored by considering the effect of waist circumference. The risk of diabetes attributable to waist circumference might be underestimated were the effects of height and hip circumference not considered.