1458

Background: Exposure to per- and polyfluoroalkyl substances (PFAS) has been linked with kidney cancer and reduced kidney function, but findings for plasma PFAS concentrations and kidney function are inconsistent. We evaluated cross-sectional associations of plasma concentrations of 6 PFAS [perfluorooctane sulfonate, perfluorooctanoate (PFOA), perfluorohexane sulfonate, N-ethyl-perfluorooctane sulfonamido acetate (EtFOSAA), N-methyl-perfluorooctane sulfonamido acetate, and perfluorononanoate] with estimated glomerular filtration rate (eGFR) and albumin-to-creatinine ratio (ACR) among prediabetic adults at risk for kidney disease. Methods: We used baseline data from the Diabetes Prevention Program (enrolled 1996-1999) and applied linear regression to estimate associations between plasma PFAS concentrations (log-2 transformed) with eGFR and ACR. We accounted for sex, age, race, marital status, income, smoking, menopause status, medication use, and diet score in the models, and tested for effect modification by sex and hypertension status. Results: Plasma PFAS concentrations of 925 overweight or obese adults (64.9% female, 57.8% non-Hispanic White) and those among the U.S. population were similar. Mean kidney functions were normal (eGFR: 97.3 ± 15.5 ml/min/1.73m²; ACR: 11.2 ± 30.5). We found no significant associations between PFAS concentrations and eGFR but detected inverse associations of EtFOSAA [β: -3.7 (95% CI: -7.1, -0.3)] and PFOA [(β: -6.4 (95% CI: -12.3, -0.1)] with ACR. We observed potential effect modification by hypertension status (inverse associations among those with hypertension but positive associations among those without hypertension), but none of the effect estimates were statistically significant. Conclusions: We observed weak inverse associations between plasma concentrations of some PFAS with ACR, suggesting higher plasma PFAS concentrations are associated with lower microalbumin ratio in the urine, thus better kidney function. Our main limitation is the potential for reverse causation due to the cross-sectional nature of the data. Longitudinal studies are needed to establish temporality of associations.