Associations between apolipoprotein B and selected perfluoroalkyl substances among diabetics and nondiabetics

Previous population investigation of perfluoroalkyl substances (PFAS) features associations with lipids in a number of populations; these investigations have seldom included consideration of apolipoproteins. Apolipoprotein B (Apo B) fractions were considered in this descriptive analysis because they are essential to the assembly, transport, and cellular uptake of lipid classes associated with poorer health outcomes, and they are associated with incident and prevalent disease. Regression models stratified by diabetes and lipid lowering medication (LLM) status for data from National Health and Nutrition Examination Survey for 2007-2014 were fitted to interrogate associations between selected PFAS and Apo B for US adults aged ≥ 20 years. Adjusted concentrations of Apo B were positively associated with perfluorooctanoic acid (PFOA β = 0.03878, p < 0.01), perfluorooctane sulfonic acid (PFOS β = .02029, p = 0.02), and perfluorononanoic acid (PFNA β = .01968, p = .03) for nondiabetics who were not taking lipid lowering medications. These associations were not seen among diabetic participants, except for perfluorodecanoic acid (PFDA) in those taking LLMs (β = 0.03831, p = 0.02). We also note that LLMs have an inferred greater impact on Apo B in the diabetics compared to the nondiabetic populations. We have considered several sources of confounding and think the data are most consistent with a weak causal association that PFAS exposure increases Apo B. The rodent toxicology literature also contains evidence that PFAS disrupt fatty acid trafficking including Apo B, although how the specific findings may relate to circulating human Apo B concentrations is unclear. We therefore advocate for attempts to replicate the findings in other populations and to consider additional types of mechanistic studies.