**Background:** Per- and polyfluoroalkyl substances (PFAS) are a large class of synthetic chemicals used extensively in industry and detected in the environment. Human exposure is widespread due to their persistence and bioaccumulation. While toxicity data exist for a few legacy PFAS (e.g., PFOA, PFOS), the health effects of thousands of other PFAS are largely unknown. The US EPA's PFAS Strategic Roadmap calls for evaluating large numbers of PFAS for potential human health effects. This systematic evidence map (SEM) expands on a previous effort that covered ~150 PFAS by characterizing the evidence base for an additional 345 PFAS, identified by the EPA's Center for Computational Toxicology and Exposure (CCTE) for in vitro testing and through interagency discussions. The goal was to catalog available hazard information and critically evaluate studies, not to synthesize data or reach causal conclusions.
**Methods:** Systematic evidence map methods were used, following the IRIS Handbook and published templates. The search included 345 PFAS (303 from the CCTE's PFAS 430 library and 42 additional PFAS identified in interagency discussions). Literature searches were conducted in PubMed, Web of Science, and ToxLine (2019, 2021, 2022 updates). Additional sources included ATSDR toxicological profiles, the PFAS-Tox Database, ECHA registration dossiers, and the EPA CompTox Chemicals Dashboard (ToxValDB). After deduplication, SWIFT-Review filtered references for human health relevance. Title/abstract screening used SWIFT-Active Screener with machine learning until 95% of relevant studies were likely captured. Full-text screening and tagging were performed by two independent reviewers. PECO criteria defined populations (humans and nonhuman mammals), exposures (oral/inhalation routes or biomarkers), comparators (lower/no exposure groups or concurrent controls), and outcomes (all health outcomes). Studies not meeting PECO but containing potentially relevant supplemental material (e.g., in vitro, ADME, nonmammalian studies) were tagged. For included studies, literature inventories captured study design details and health systems assessed. Study evaluation was conducted for nonacute mammalian bioassay studies and all epidemiological studies using HAWC, with domain ratings (Good, Adequate, Deficient, Critically Deficient) and overall confidence (High, Medium, Low, Uninformative). Detailed data extraction of methods and findings was performed for all epidemiological studies and mammalian bioassays with sufficient data. A comprehensive PFAS dashboard was created to compile data from this SEM, the previous SEM, and other EPA assessment products.
**Key Results:** Over 13,000 studies were identified from scientific databases. After screening, 121 mammalian bioassay and 111 epidemiological studies met PECO criteria. Epidemiological evidence was available for 12 PFAS, most commonly perfluorododecanoic acid (83 studies), 2-(N-Methylperfluorooctanesulfonamido)acetate (21 studies), and 2-(N-Ethyl-perfluorooctanesulfonamido)acetate (15 studies). Mammalian bioassay evidence was available for 30 PFAS, with 2H,3H-Decafluoropentane (16 studies), perfluoro-2-methyl-3-pentanone (15 studies), and perfluorododecanoic acid (10 studies) being most studied. Overall, 41 PFAS (approximately 11% of searched chemicals) had evidence across both data streams. In mammalian studies, the most commonly assessed health systems were whole body (e.g., body weight), reproductive, nervous, and hepatic. In epidemiological studies, the most studied health systems were developmental (21 studies), cardiovascular (18), female reproductive (17), metabolic (19), endocrine (13), and immune (13). Study evaluation of mammalian bioassays found that most ECHA summaries (28/37) had 'Uninformative' or 'Low' confidence due to limited reporting. Among non-ECHA studies, about 50% (11/19) were 'Medium' confidence; none were 'High' confidence. For epidemiological studies, 83 (75%) received a 'High'/'Medium' confidence rating for at least one outcome, but 93 (84%) were rated 'Deficient' for study sensitivity, primarily due to limited exposure contrast. The comprehensive PFAS dashboard includes 370 mammalian bioassay studies for 87 PFAS and 631 epidemiological studies for 35 PFAS across multiple EPA projects.
**Clinical Implications:** This evidence map reveals that for the vast majority of PFAS (89% of those searched), no human or animal toxicity data are available, representing a critical data gap for risk assessment. For the few PFAS with data, the evidence is often limited by study quality (e.g., poor reporting in gray literature) or low sensitivity (e.g., low exposure levels in human studies). The findings underscore the need for targeted research to fill these gaps, particularly for health systems like immune function and developmental effects that have been sensitive endpoints for other PFAS. The publicly available comprehensive PFAS dashboard provides a scoping tool for researchers and regulators to prioritize PFAS for further study and risk assessment activities. The lack of data for most PFAS also highlights the importance of using alternative approaches (e.g., high-throughput toxicology, read-across) to inform hazard identification and support public health decision-making.