Persistent Organic Pollutants, Circulating Sex Hormones, and Coronary Heart Disease

Abstract

Introduction: Fish tend to accumulate large quantities of environmental pollutants such as polychlorinated biphenyls (PCBs), Dichlorodiphenyldichloroethylene (DDE), dioxins, furans, and methyl mercury. Many environmental pollutants have been associated with increased risk of CHD, and since fish have generally been accepted as a healthy source of protein and n-3 PUFAs, it is important to weigh the risks and benefits of fish consumption and potentially modify dietary recommendations to prevent exposure to environmental pollutants. However, the mechanisms through which POPs may increase risk of CHD are unclear. One plausible mechanism through which PCBs may influence risk of CHD is modification of steroid sex hormone metabolism. This study aimed to (1) explore longitudinal relationships between fish consumption, POPs, and CHD-related outcomes in the GLSCF Consumer Cohort, (2) explore longitudinal relationships between POPs and endogenous steroid hormones in the GLSCF Consumer Cohort, and (3) explore cross-sectional relationships between POPs and endogenous steroid sex hormones in using data from the NHANES. Methods: For statistical analyses, PCB congeners were summed over all congeners and grouped according to their proposed metabolic pathways (1) for the following four PCB groupings: (1) Total PCBs, (2) MC-type PCBs, (3) PB-type PCBs, and (4) Mixed-type PCBs. DDE was included as a sum of all isoforms only in analysis of data from the GLSCF Consumer Cohort. Proportional hazards regression was used to estimate longitudinal associations of PCB groupings and DDE with time to diagnosis of CHD, MI, and angina using data from the GLSCF Consumer Cohort (n = 819). Likelihood-based longitudinal mixed models were used to estimate associations of PCB groupings and DDE with total testosterone, SHBG, SHBG-bound testosterone, and free testosterone using data from the GLSCF Consumer Cohort (n = 466). Linear regression models were used to estimate associations of PCB groupings with total, free and bound testosterone and estradiol as well as SHBG and 3-α androstenedione glucuronide (3α-ADG) using data from NHANES 1999-2002 (n = 595). Results: After adjusting for known and suspected confounders as well as comorbid hypertension, a 2-fold increase in PCBs was associated with increased risk of CHD for PB-type PCBs (HR = 1.72; 95% CI: 1.06-2.81; p=0.0294). After excluding participants with liver disease, a 2-fold increase in PB-type PCBs was associated with a 17% decrease in SHBG (p=0.03), an 18% decrease in SHBG-bound testosterone (p=0.02), and a 10% increase in free testosterone (p=0.07). We found a similar significant inverse association of DDE with SHBG-bound testosterone (-23%, p=0.3), but limited evidence for an impact of DDE on SHBG (-8%, p=0.14) and free testosterone (3%, p=0.45). After adjusting for known and potential confounders, a 50% increase in PB-type inducer PCBs were associated with an 8% increase in SHBG (95% CI: 2.2, 14.1; p=0.009). A 50% increase in most PCB groupings were associated with a decrease in 3α-ADG, but associations were strongest in males age 20-39 years and near null among males age 60-85 years. Among males age 20-39 years, decrease in 3α-ADG was 18.2% (95% CI: -28.4, -7.9; p=0.001) for a 50% increase in PB-type inducer PCBs. After standardizing 3α-ADG concentrations to total testosterone concentrations using the 3α-ADG to total testosterone ratio, these associations become more consistent across age groups. Conclusions: This study provides evidence that PCBs, and particularly PCB congeners that are known PB-type inducers of CYP2A and CYP2B, may increase risk of CHD and the increase in risk of CHD may, at least in part, be a result of decreased androgenicity in men. The difference in results according to PCB grouping method highlights the importance of developing new methods for assessing the impact of multiple exposures to environmental pollutants in the presence of highly correlated exposure mixtures.