BACKGROUND: Several epidemiological cross-sectional studies have found positive associations between serum concentrations of lipids and perfluorooctanoic acid (PFOA, or C8). A longitudinal study should be less susceptible to biases from uncontrolled confounding or reverse causality. METHODS: We investigated the association between within-individual changes in serum PFOA and perfluorooctanesulfonic acid (PFOS) and changes in serum lipid levels (low-density lipoprotein [LDL] cholesterol, high-density lipoprotein cholesterol, total cholesterol, and triglycerides) over a 4.4-year period. The study population consisted of 560 adults living in parts of Ohio and West Virginia where public drinking water had been contaminated with PFOA. They had participated in a cross-sectional study in 2005-2006, and were followed up in 2010, by which time exposure to PFOA had been substantially reduced. RESULTS: Overall serum concentrations of PFOA and PFOS fell by half from initial geometric means of 74.8 and 18.5 ng/mL, respectively, with little corresponding change in LDL cholesterol (mean increase 1.8%, standard deviation 26.6%). However, there was a tendency for people with greater declines in serum PFOA or PFOS to have greater LDL decrease. For a person whose serum PFOA fell by half, the predicted fall in LDL cholesterol was 3.6% (95% confidence interval = 1.5-5.7%). The association with a decline in PFOS was even stronger, with a 5% decrease in LDL (2.5-7.4%). CONCLUSIONS: Our findings from this longitudinal study support previous evidence from cross-sectional studies of positive associations between PFOA and PFOS in serum and LDL cholesterol.