Cut-off values for anthropometric indices in predicting morbidity and mortality in Turkey: a cohort

Abstract

Background: Anthropometric measurements are associated with cardiovascular outcomes and mortality and the association varies by community. This study aims to estimate the incidence of cardiovascular outcomes and all cause mortality and determines their association with the anthropometric measurements. Methods: The data from Turkey Chronic Diseases and Risk Factors Survey were used for the cohort which recruited two 15 years and older adults from each family physicians' list in 2011 in Turkey using simple random sampling (n = 18,477). Family physicians collected the medical history and performed anthropometric measurements. Initially healthy participants for cardiovascular morbidity (coronary heart disease (ICD-10 codes I20.0-I25.0) and stroke (ICD-10 codes I60.0-I69.0) and all participants for all-cause and cardiovascular mortality were followed-up through electronic health records by 2017. Area under the curves (AUCs) were estimated and the adiposity risk was defined as cut-off points determined using Youden Index. The sample was stratified by age-groups (<65, ≥65). Poisson and Cox models were fitted for morbidity and mortality outcomes, respectively. Results: Over six years of follow-up, 5.7% of 15,877 and 1.3% of 16,062 participants developed CHD and stroke respectively. In total, 664 all-cause deaths (4.0%) were observed with 155 caused by CVDs (0.9%). A body shape index (i.e. 0.658 for all-cause mortality among younger men), waist-to-height ratio (i.e. 0.743 for CHD incidence among younger women), and BMI had higher AUCs for the outcomes. Higher anthropometric measurement groups had almost always higher CHD incidence risk in all groups but lower all-cause mortality in older men. Higher body-mass index was associated with lower mortality in younger men. Conclusions: This cohort study in Turkey showed that having higher anthropometric measurements is mostly associated with higher CHD incidence and mostly lower mortality among older men. This association deserves further studies.