Abstract P172: Per- and Polyfluoroalkyl Substances (PFASs), Apolipoprotein and the Risk of Coronary Heart Disease: A Case-Control Pilot Study

The association of PFASs with blood lipids, lipoproteins and apolipoproteins, and coronary heart disease (CHD) risk are inconsistent in previous studies. The aim of current study is to examine the potential association between plasma PFASs, blood lipoproteins and their subspecies defined by apolipoproteins, and whether PFASs are associated with CHD risk.A case-control study was conducted in 100 men and 102 women aged 45-70 years. We leveraged 101 participants from HPFS and NHS who were initially free of cardiovascular disease and developed a fatal or non-fatal myocardial infarction during follow-up, and randomly matched controls in a 1:1 ratio for age, smoking status, and date of blood draw. Seven plasma PFASs measurements were examined at baseline, including perfluorohexane sulfonic acid (PFHxS), perfluorooctanoic acid (PFOA), total perfluorooctane sulfonic acid (PFOS), branched PFOS (brPFOS), linear PFOS (nPFOS), perfluorononanoic acid (PFNA), and perfluorodecanoic acid (PFDA), with all mean CVs <20%. Pearson correlation was applied to study the relationship between PFASs, lipoproteins and subspecies; Conditional logistic regression model was conducted to test the association between PFASs and CHD risk; a mediating effect model was used to assess how blood lipids and apolipoprotein subspecies mediate the association between PFASs and CHD risk.After multivariable adjustments, we did not find significant associations between PFAS and lipids, except for PFHxS and high-density lipoprotein cholesterol (HDL-C) with a significant positive association (r=0.19, p=0.046). For subspecies with apolipoprotein (apo) C-III, we observed non-significant inverse associations between plasma PFAS concentrations and the subspecies of HDL-C, low-density lipoprotein cholesterol (LDL-C) and VLDL-C that contain only apoC-III. For those without ApoC-III, the directions of most associations are positive. Strong associations were demonstrated between PFNA, PFDA and HDL-C with ApoE (r=0.42, p=0.00004; r=0.27, p=0.004, respectively). In addition, elevated PFOS concentrations were associated with higher relative risk of CHD (HR=1.04, p=0.02), especially for branched PFOS (HR=1.18, p=0.003), and these association were largely dose dependent (PFOS: p for linearity = 0.009; brPFOS: 0.0007). Mediation analysis illustrated no mediating effect regarding blood lipids and apolipoprotein subspecies on the association of interest, except for HDL-C on the association between PFOS, nPFOS and CHD risk (%mediated=25%, p=0.03; %mediated=36%, p=0.03, separately).Our results suggest that there is a linear positive dose-response relationship between PFOS and CHD risk, but the association is probably not mediated by subspecies with/without Apo C-III or Apo E.