Association of waist circumference and BMI with premature death in young and middle-aged population

Abstract

Introduction: Premature death is a global health indicator, significantly impacted by obesity, especially in young and middle-aged population. Both body mass index (BMI) and waist circumference (WC) assess obesity, with WC specifically indicating central obesity and showing a stronger relationship with mortality. However, despite known associations between BMI and premature death, as well as the well-recognized correlation between WC and adverse health outcomes, the specific relationship between WC and premature death remains unclear. Therefore, focusing on young and middle-aged individuals, this study aimed to reliably estimate independent and combined associations between WC, BMI and premature death, thereby providing causal evidence to support strategies for obesity management.

Methods: This study involved 49,217 subjects aged 18-50 years in the United States from 1999 to 2018 National Health and Nutrition Examination Survey (NHANES). Independent and combined associations between WC and BMI with premature death across sex and age stratum were examined by Cox regression. Survey weighting and inverse probability weighting (IPW) were further considered to control selection and confounding bias. Robustness assessment has been conducted on both NHANES and China Health and Retirement Longitudinal Study (CHARLS) data.

Results: A linear and positive relationship between WC and all-cause premature death was found in both males and females, with adjusted HRs of 1.019 (95%CI = 1.004-1.034) and 1.065 (95%CI = 1.039-1.091), respectively. Nonlinear relationships were found with respect to BMI and all-cause premature death. For females aged 36-50 with a BMI below 28.6 kg/m², the risk of premature death decreased as BMI increased, indicated by adjusted HRs of 0.856 (95%CI = 0.790-0.927). Joint analysis showed among people living with obesity, a larger WC increased premature death risk (HR = 1.924, 95%CI = 1.444-2.564).

Discussion: WC and BMI exhibited prominent associations with premature death in young and middle-aged population. Maintaining an appropriate WC and BMI bears significant implications for preventing premature death.

Keywords: BMI; NHANES; inverse probability weighting; premature death; waist circumference; young and middle-aged people.

Figure 1

DAG for assessing confounding covariates in the relationship between obesity and premature death. The red circle represents confounding variables and the blue circle represents mediating variables.

Figure 2

Plots of all-cause premature mortality HR from multivariate adjusted stratified Cox regression analysis using restricted cubic splines of WC with 4 degrees of freedom. (A) Model for 18–50 years old males. (B) Model for 18–35 years old males. (C) Model for 36–50 years old males. (D) Model for 18–50 years old females. (E) Model for 18–35 years old females. (F) Model for 36–50 years old females. P showed the test for nonlinearity. (B,C,E,F) models were adjusted for marital status, race, birthplace, education level, long working hours, job type, emotional support, financial support, number of close friends, economic level, smoking, alcohol drinking, electronic product use time, sleep hours, moderate activity, vigorous activity, number of restaurant meals, healthy eating, BMI. (A,D) models additionally adjusted for age. The horizontal line in the graph indicated a premature death HR of 1, and the vertical line marked the inflection point.

Figure 3

Plots of all-cause premature mortality HR from multivariate adjusted stratified Cox regression analysis using restricted cubic splines of BMI with 4 degrees of freedom. (A) Model for 18–50 years old males. (B) Model for 18–35 years old males. (C) Model for 36–50 years old males. (D) Model for 18–50 years old females. (E) Model for 18–35 years old females. (F) Model for 36–50 years old females. (B,C,E,F) models were adjusted for marital status, race, birthplace, education level, long working hours, job type, emotional support, financial support, number of close friends, economic level, smoking, alcohol drinking, electronic product use time, sleep hours, moderate activity, vigorous activity, number of restaurant meals, healthy eating, BMI. (A,D) models additionally adjusted for age. The horizontal line in the graph indicated a premature death HR of 1, and the vertical line marked the inflection point.