Exposure to Lead and Coronary Artery Atherosclerosis: A Swedish Cross-Sectional Population-Based Study

Abstract

Background: Lead is an established causal risk factor for coronary heart disease. Atherosclerosis may be the key mediator for this association, but evidence from studies in humans is limited. Our objective was to test the hypothesis that environmental lead exposure is associated with coronary atherosclerosis.

Methods: We used cross-sectional data from the SCAPIS (Swedish Cardiopulmonary Bioimage Study), including 5627 men and women aged 50 to 64 years. Coronary artery calcium score (CACS), measured using computed tomography, was used as a marker of atherosclerosis, and blood lead was used as a biomarker of lead exposure. The prevalence ratio (PR) of positive (>0) and high (≥100) CACSs in relation to blood lead (continuous variable) was modeled using Poisson regression with robust SEs, adjusted for age, sex, smoking, low-density lipoprotein/high-density lipoprotein ratio, waist circumference, heredity for cardiovascular diseases, statin use, diabetes, blood cadmium, low physical activity, and educational level.

Results: Median blood lead was 14.2 μg/L. Positive CACS (prevalence, 41%) was not significantly associated with blood lead (PR per Δ10 μg/L, 1.02 [95% CI, 0.99-1.04]), whereas the association was stronger for high CACS (prevalence, 13%; and PR per Δ10 μg/L, 1.05 [95% CI, 1.00-1.11]). The estimate for high CACS was stronger in men (PR per Δ10 μg/L, 1.07 [95% CI, 1.01-1.13]) than among women (PR per Δ10 μg/L, 1.01 [95% CI, 0.85-1.20]).

Conclusions: Our study, which found that lead is associated with coronary artery calcification in men, provides further evidence that lead is a risk factor for atherosclerosis and coronary heart disease.

Keywords: atherosclerosis; coronary artery disease; environmental exposure; lead.

Figure 1. Forest plot of overall and stratified prevalence ratios (PRs) with 95% CIs for the relationship between blood lead concentration (B‐Pb) and coronary artery calcium score ≥100, in a population‐based Swedish cohort.

PRs are presented for a 10‐μg/L change in B‐Pb. All participants with covariate data according to model 3 (age, sex, smoking status, blood cadmium, diabetes, low‐density lipoprotein/high‐density lipoprotein [LDL/HDL] ratio, statin treatment, waist circumference, low physical activity, heredity for cardiovascular diseases [CVDs], and educational level) were included. The PRs were modeled using B‐Pb as a continuous variable in a Poisson regression model adjusted for covariates in model 3. Stratifications were done for covariates in model 3. For age, LDL/HDL ratio, and waist circumference, the data were split by the median. For blood cadmium concentration, the data were split into 2 groups using the quartile 4 cutoff point. For categorical variables, the data were split by category. P values for interaction were obtained in separate models for each stratification factor.

Figure 2. Adjusted prevalence ratio (PR) (solid line) with 95% CI (dashed lines) for the relationship between blood lead concentration (B‐Pb) (μg/L) and the prevalence of coronary artery calcium score (CACS) >0, in a population‐based Swedish cohort.

Participants with covariate data according to model 3 (age, sex, smoking status, blood cadmium, diabetes, low‐density lipoprotein/high‐density lipoprotein ratio, statin treatment, waist circumference, low physical activity, heredity for cardiovascular disease, and educational level) were included. Observations >50 μg/L (n=48) were removed. The PRs were modeled using restricted cubic splines with 4 knots placed at percentiles 5, 35, 65, and 95 (6.7, 12.0, 17.0, and 30.1 μg/L, respectively) in a Poisson regression model adjusted for covariates in model 3. The reference exposure level was set at percentile 12.5 (8.5 μg/L). The histogram shows the frequency distribution of B‐Pb. The numbers of cases/observations were 2069/5171.

Figure 3. Adjusted prevalence ratio (PR) (solid line) with 95% CI (dashed lines) for the relationship between blood lead concentration (B‐Pb) (μg/L) and the prevalence of coronary artery calcium score (CACS) ≥ 100 in a population‐based Swedish cohort.

Participants with covariate data according to model 3 (age, sex, smoking status, blood cadmium, diabetes, low‐density lipoprotein/high‐density lipoprotein ratio, statin treatment, waist circumference, low physical activity, heredity for cardiovascular disease, and educational level) were included. Observations >50 μg/L (n=48) were removed. The PRs were modeled using restricted cubic splines with 4 knots placed at percentiles 5, 35, 65, and 95 (6.7, 12.0, 17.0, and 30.1 μg/L, respectively) in a Poisson regression model adjusted for covariates in model 3. The reference exposure level was set at percentile 12.5 (8.5 μg/L). The histogram shows the frequency distribution of B‐Pb. The numbers of cases/observations were 619/5171.

Figure 4. Adjusted prevalence ratio (PR) (solid line) with 95% CI (dashed lines) for the relationship between blood lead concentration (B‐Pb) (μg/L) and the prevalence of coronary artery calcium score (CACS) ≥100, among women in a population‐based Swedish cohort.

Participants with covariate data according to model 3 (age, sex, smoking status, blood cadmium, diabetes, low‐density lipoprotein/high‐density lipoprotein ratio, statin treatment, waist circumference, low physical activity, heredity for cardiovascular disease, and educational level) were included. Observations >50 μg/L were removed. The PRs were modeled using restricted cubic splines with 4 knots placed at percentiles 5, 35, 65, and 95 (6.5, 11.9, 16.9, and 29.8 μg/L, respectively) in a Poisson regression model adjusted for covariates in model 3. The reference exposure level was set at percentile 12.5 (8.2 μg/L). The histogram shows the frequency distribution of B‐Pb. The numbers of cases/observations were 162/2690.

Figure 5. Adjusted prevalence ratio (PR) (solid line) with 95% CI (dashed lines) for the relationship between blood lead concentration (B‐Pb) (μg/L) and the prevalence of coronary artery calcium score (CACS) ≥100, among men in a population‐based Swedish cohort.

Participants with covariate data according to model 3 (age, sex, smoking status, blood cadmium, diabetes, low‐density lipoprotein/high‐density lipoprotein ratio, statin treatment, waist circumference, low physical activity, heredity for cardiovascular disease, and educational level) were included. Observations >50 μg/L were removed. The PRs were modeled using restricted cubic splines with 4 knots placed at percentiles 5, 35, 65, and 95 (7.0, 12.2, 17.0, and 30.7 μg/L, respectively) in a Poisson regression model adjusted for covariates in model 3. The reference exposure level was set at percentile 12.5 (8.7 μg/L). The histogram shows the frequency distribution of B‐Pb. The numbers of cases/observations were 457/2481.