Associations between specific volatile organic chemical exposures and cardiovascular disease risks: insights from NHANES

Abstract

Introduction: An increasing body of research has demonstrated a correlation between pollutants from the environment and the development of cardiovascular diseases (CVD). However, the impact of volatile organic chemicals (VOC) on CVD remains unknown and needs further investigation.

Objectives: This study assessed whether exposure to VOC was associated with CVD in the general population.

Methods: A cross-sectional analysis was conducted utilizing data from five survey cycles (2005-2006, 2011-2012, 2013-2014, 2015-2016, and 2017-2018) of the National Health and Nutrition Examination Survey (NHANES) program. We analyzed the association between urinary VOC metabolites (VOCs) and participants by multiple logistic regression models, further Bayesian Kernel Machine Regression (BKMR) models and Weighted Quantile Sum (WQS) regression were performed for mixture exposure analysis.

Results: Total VOCs were found to be positively linked with CVD in multivariable-adjusted models (p for trend = 0.025), independent of established CVD risk variables, such as hypertension, diabetes, drinking and smoking, and total cholesterol levels. Compared with the reference quartile of total VOCs levels, the multivariable-adjusted odds ratios in increasing quartiles were 1.01 [95% confidence interval (CI): 0.78-1.31], 1.26 (95% CI: 1.05-1.21) and 1.75 (95% CI: 1.36-1.64) for total CVD. Similar positive associations were found when considering individual VOCs, including AAMA, CEMA, CYMA, 2HPMA, 3HPMA, IPM3 and MHBMA3 (acrolein, acrylamide, acrylonitrile, propylene oxide, isoprene, and 1,3-butadiene). In BKMR analysis, the overall effect of a mixture is significantly related to VOCs when all chemicals reach or exceed the 75th percentile. Moreover, in the WQS models, the most influential VOCs were found to be CEMA (40.30%), DHBMA (21.00%), and AMCC (19.70%).

Conclusion: The results of our study indicated that VOC was all found to have a significant association with CVD when comparing results from different models. These findings hold significant potential for public health implications and offer valuable insights for future research directions.

Keywords: Bayesian kernel machine regression; NHANES; cardiovascular diseases; environment pollution; volatile organic chemicals.

Figure 1

Flowchart of the study.

Figure 2

Spearman’s rank correlation coefficients of VOCs measured in NHANES.

Figure 3

Weighted values of urinary VOCs for CVD in Weighted quantile sum models. Models were adjusted for covariates: age, sex, race, NHANES cycles, family PIR, education levels, physical activity levels, drinking or smoking status, body mass index, diabetes, hypertension, serum total cholesterol, urine creatinine and urine albumin.

Figure 4

The links between urinary VOCs and the estimation of CVD risk using Bayesian Kernel Machine Regression (BKMR). (A) Exposure-response functions for each VOCs with the other VOCs fixed at the median. (B) Combined effects of urinary VOCs mixture on CVD risk. This graph depicted the estimated variance in CVD risk along with the corresponding 95% confidence interval when VOCs concentrations were maintained at specific percentiles in contrast to their respective medians. Models were adjusted for covariates: age, sex, race, NHANES cycles, family poverty income ratio, education levels, physical activity levels, drinking or smoking status, body mass index, diabetes, hypertension, serum total cholesterol, urine creatinine and urine albumin.

Figure 5

Statistically significant VOC and presence of individual CVDs in NHANES. (A) Congestive heart failure; (B) Coronary artery disease; (C) Angina; (D) Heart attack; (E) Stroke. Models were adjusted for covariates: age, sex, race, NHANES cycles, family PIR, education levels, physical activity levels, drinking or smoking status, body mass index, diabetes, hypertension, serum total cholesterol, urine creatinine and urine albumin.