Cause-specific years of life lost attributable to non-optimal body mass index by county, sex, race, and ethnicity in the USA, 2000–2019: a systematic analysis of health disparities

Abstract Background Over 334,000 deaths in 2021 in the USA were attributed to non-optimal body mass index ([BMI] greater than 20 to 21 kg/m2), with elevated mortality among American Indian and Alaska Native (AIAN), Black, and Latino populations. Disparities in attributable mortality by race, ethnicity, and county are poorly understood. This analysis examined variation by race and/or ethnicity in obesity prevalence (BMI ≥ 30) and years of life lost (YLLs) attributable to non-optimal BMI in 3110 US counties from 2000 to 2019. Methods Using survey data from the Behavioral Risk Factor Surveillance System (BRFSS), Gallup Daily, and National Health and Nutrition Examination Survey (NHANES), we estimated obesity prevalence annually, stratified by county, age, sex, and five mutually exclusive racial and/or ethnic populations (AIAN, Asian or Pacific Islander [Asian], Black, Latino or Hispanic [Latino], and White). We calculated population attributable fractions (PAFs) and estimated YLLs attributable to non-optimal BMI for 27 causes of death (focusing on ischemic heart disease [IHD], colorectal cancer, and diabetes) using cause-specific YLL estimates from a previous analysis. Results Age-standardized obesity prevalence increased by 12.3 percentage points (95% uncertainty interval 11.9–12.8) to 40.2% (40.0–40.6) in the USA from 2000 to 2019 and was highest in the Black population, followed by the AIAN, Latino, White, and Asian populations. In 2019, the Black population had the highest rates of IHD and colorectal cancer YLLs attributable to non-optimal BMI, followed by the AIAN, White, Latino, and Asian populations. The AIAN population had the highest attributable YLL rate for diabetes in 2019, followed by the Black, Latino, White, and Asian populations. All racial and/or ethnic populations had statistically significant reductions in IHD and diabetes YLL rates attributable to non-optimal BMI from 2000 to 2019, with declines in total YLL rates for these causes more than offsetting increases in obesity prevalence and PAFs. Relative disparities among counties were two to four times as large for attributable YLL rates as for obesity prevalence. Conclusions Racial and/or ethnic disparities in obesity prevalence are substantial, but disparities in YLLs attributable to non-optimal BMI are larger because they are compounded by disparities in YLL rates.