Population-Specific Probabilistic Risk Assessment of PFOA for Ovarian Function Using PBTK Translation of Data from...

We assessed the risk of perfluorooctanoic acid (PFOA) to ovarian function by integrating in vivo mouse data with physiologically based toxicokinetic (PBTK) modeling and population-level analysis. Adult mice received PFOA (0.06, 1.15, 22 mg/kg bw/day; 14 days), followed by analysis of ovarian histology and whole-ovary transcriptomics. Effect doses were converted to human equivalent doses (HEDs) and propagated through a PBTK model to estimate ovarian bioactive concentrations (OBCs) in human ovaries. Median histology HEDs were 41 and 501 mg/week, whereas transcriptomic HED was 322 mg/week values above PFOA levels inferred from follicular fluid (FF) (∼0.002-0.0069 μg/mL). Lowest histology OBC was 25.6 μg/mL, and the fifth percentile transcriptomic OBC was 21.3 μg/mL, both exceeding PFOA FF levels. A subset of 21 ovarian genes had OBCs within 100-fold of PFOA FF levels. Clinical conditions such as cirrhosis Child-Pugh (CP) class C (CP-C), end-stage renal disease (ESRD), and morbid obesity double the risk for PFOA-induced ovarian dysfunction. Weibull distributions to gene-level OBCs across clinical populations and extrapolation to low percentiles (p = 0.05-10^-4) showed that cirrhosis CP-C, ESRD, and morbid obesity demonstrated progressively smaller OBC percentiles compared with healthy as p decreased. Other populations showed modest changes or slight increases at the lowest percentiles. Overall, predicted PFOA OBCs remain above currently reported human PFOA FF levels, supporting a population-level margin of safety under current exposure. The gene-level and vulnerable population analyses identify sensitive molecular targets and subgroups in which this margin is reduced, warranting prioritization in future biomonitoring and reproductive risk assessment.