Abstract
BACKGROUND: An accurate and comprehensive anthropometric measure for predicting type 2 diabetes mellitus (T2DM) has not yet been depicted. METHODS: A total of 8450 nondiabetic participants were recruited during 2000-2010 in Taiwan. The cohort was followed up to the end of 2013, over an average of 8.87 years. At recruitment, participants completed a questionnaire related to basic demographics, lifestyle variables, personal disease history, and family disease history. 3D body surface scanning was used to obtain 35 anatomical measurements. A Cox proportional hazard model was used to conduct multivariable analyses. RESULTS: A total of 2068 T2DM cases at an incidence rate of 27.59 × 10(-3) (year(-1)) were identified during the follow-up period. Multivariable-adjusted hazard ratios (HRs) demonstrated that neck circumference (NC) (HR = 1.048; 95% CI = 1.033-1.064), waist width (WW) (HR = 1.061; 95% CI = 1.040-1.081), and left thigh circumference (TC) (HR = 0.984; 95% CI = 0.972-0.995) were significant predictors of the occurrence of T2DM. While dividing body measurement into median high/low groups, an increased risk of T2DM was observed among participants with a larger NC and smaller TC (HR = 1.375; 95% CI = 1.180-1.601) and a larger WW and smaller TC (HR = 1.278; 95% CI = 1.085-1.505) relative to other participants. CONCLUSIONS: This study suggests that as well as using traditional waist and TC measurements, NC can be used as an indicator to provide an early prediction of developing T2DM, while providing clues for future mechanistic investigations of T2DM.