High organofluorine concentrations in municipal wastewater affect downstream drinking water supplies for millions of Americans
Edited by William Schlesinger, Cary Institute of Ecosystem Studies, Millbrook, NY; received August 22, 2024; accepted December 9, 2024
Significance
US municipal wastewater facilities are major per- and polyfluoroalkyl substances (PFAS) sources known to affect drinking water quality. Among eight large wastewater treatment facilities with comparable treatment technologies and sizes to those serving 70% of the US population, we found that the six regulated PFAS in drinking water accounted for <10% of extractable organofluorine (EOF) in wastewater influent and effluent. Most (62 to 75%) of the EOF consisted of commonly prescribed fluorinated pharmaceuticals, and the maximum EOF removal efficiency was <25%. Results from a national wastewater dilution model suggest that wastewater PFAS discharges impair drinking water supplies for >20 million Americans, emphasizing the importance of reducing diverse PFAS sources entering wastewater.
Abstract
Wastewater receives per- and polyfluoroalkyl substances (PFAS) from diverse consumer and industrial sources, and discharges are known to be a concern for drinking water quality. The PFAS family includes thousands of potential chemical structures containing organofluorine moieties. Exposures to a few well-studied PFAS, mainly perfluoroalkyl acids (PFAA), have been associated with increased risk of many adverse health outcomes, prompting federal drinking water regulations for six compounds in 2024. Here, we find that the six regulated PFAS (mean = 7 to 8%) and 18 measured PFAA (mean = 11 to 21%) make up only a small fraction of the extractable organofluorine (EOF) in influent and effluent from eight large municipal wastewater treatment facilities. Most of the EOF in influent (75%) and effluent (62%) consists of mono- and polyfluorinated pharmaceuticals. The treatment technology and sizes of the treatment facilities in this study are similar to those serving 70% of the US population. Despite advanced treatment technologies, the maximum EOF removal efficiency among facilities in this work was <25%. Extrapolating our measurements to other large facilities across the United States results in a nationwide EOF discharge estimate of 1.0 to 2.8 million moles F y−1. Using a national model that simulates connections between wastewater discharges and downstream drinking water intakes, we estimate that the sources of drinking water for up to 23 million Americans could be contaminated above regulatory thresholds by wastewater-derived PFAS alone. These results emphasize the importance of further curbing ongoing PFAS sources and additional evaluations of the fate and toxicity of fluorinated pharmaceuticals.
Sign up for PNAS alerts.
Get alerts for new articles, or get an alert when an article is cited.
Fluorine is the 13th most abundant element in the Earth’s crust, but natural organofluorine is extremely rare in the biosphere and is not essential for sustaining life (1). Since the 1940s, humans have synthesized tens of thousands of organofluorine chemicals to take advantage of the high electronegativity and density and low polarizability of fluorine (1, 2). Organofluorine is extensively used in modern commerce for products such as refrigeration, fluoropolymers, pharmaceuticals, agrochemicals, and nonstick and greaseproof coatings. A subset of organofluorine compounds, per- and polyfluoroalkyl substances (PFAS), has garnered intense interest in recent years because they have been associated with numerous adverse effects on the health of humans and wildlife (3, 4). Municipal wastewater treatment facilities, hereon referred to as publicly owned treatment works (POTWs), receive PFAS from diverse domestic and industrial sources and have been statistically associated with impaired drinking water quality across the United States (5, 6).
Water scarcity across many regions of the United States, particularly in the Southwestern parts of the country, means that wastewater reuse and intentional recharge of groundwater supplies are being used as water conservation measures (7). Prior work developed a national model for the dilution of wastewater effluent that enters streams and groundwater and serves as drinking water supplies for public water systems across the United States (known as the “De facto Reuse Incidence in our Nations Consumable Supply model or DRINCS model) (8, 9). This work suggests that 6% of US public water systems receive wastewater that is diluted by less than a factor of 10 before entering drinking water intakes (8), indicating the potential for residual chemical contamination in POTW effluents to affect downstream drinking water quality.
Existing estimates of the magnitude of PFAS and organofluorine discharged to aquatic environments have been extrapolated from manufacturing data rather than direct measurements (10–12). Empirically constrained release inventories for major organofluorine sources such as POTW discharges (6, 13) are therefore critically needed. The complexity of analytical methods used to detect and quantify organofluorine in wastewater has limited present understanding. Most wastewater measurements have focused on a few intensively studied PFAS, particularly perfluoroalkyl acids (PFAA) (14). However, recent work using bulk organofluorine measurements such as extractable organofluorine (EOF) has noted the presence of large quantities of unknown organofluorine (15–17). Further elucidating the composition of unknown organofluorine is important for identifying accumulation of any replacement PFAS used by industry following the phase out of legacy compounds. Prior work on wastewater biosolids suggests that mono- (compounds containing only –CF moieties) and polyfluorinated (compounds containing at least one –CF3 or –CF2– moiety) pharmaceuticals may account for a substantial fraction of the unknown EOF mass (18). SI Appendix summarizes the specific organofluorine chemicals captured by EOF measurements, compared to other bulk organofluorine techniques.
The main objective of this work was to better understand the magnitude and composition of aqueous organofluorine discharged from large US POTWs (defined as those serving >10,000 individuals) (19) and impacts on downstream water quality. We constructed a mass budget for EOF measured in wastewater influent and effluent samples from eight large POTWs with similar treatment technologies and sizes as those serving 70% of the US population. We extended these measurements using an empirical relationship that describes the magnitudes of wastewater EOF released across the country (20). We combined measurements in this study with the DRINCS model to quantify wastewater impacts on downstream drinking water sources. Results of this work provide estimates of the number of drinking water facilities (and their service populations) which would need to mitigate upstream wastewater-derived organofluorine sources and/or implement advanced drinking water treatment to prevent exposures to toxic substances.
Results
PFAA and Their Precursors Account for a Minor Fraction of EOF.
EOF concentrations measured in aqueous effluent from the eight US POTWs included in this work ranged from 19 to 41 nanomolar equivalents as fluorine (nM F) (Fig. 1A and Dataset S1). The measured coefficient of variance in the concentrations of EOF in effluent across facilities (25%) falls within the range for PFAA reported across the United States in previous work (13 to 59%) (14). PFAA concentrations accounted for a small fraction of the EOF in POTW influent (11 ± 8%) and effluent (21 ± 12%) (Fig. 1B and Dataset S2). Four PFAA in aqueous effluent made up an average of >1% of EOF including perfluorobutanoate (PFBA), perfluoropentanoate (PFPeA), perfluorohexanoate (PFHxA), and perfluorobutane sulfonate (PFBS). These compounds are all short-chain PFAA with <6 perfluorinated carbons that have been introduced as replacements for the legacy PFAS, perfluorooctanoate (PFOA), and perfluorooctane sulfonate (PFOS).
Fig. 1.
Total PFAA precursors estimated using an oxidative precursor assay (TOP) also accounted for only a small fraction of EOF in POTW influent (13 ± 5%) and effluent (16 ± 13%) (Fig. 1B and Dataset S2). The TOP assay converts precursors, many of which do not have commercially available analytical standards required for quantification, into quantifiable PFAA (21). The concentrations of PFAA produced by the TOP assay are then used to estimate the original concentrations of precursors using Bayesian inference (TOP + BI) (22, 23). We also directly quantified 16 targeted PFAA precursors with available commercial analytical standards and compared their concentrations to the TOP + BI estimate of total PFAA precursors. We detected eight of the 16 precursors above method detection limits, but their sum accounted for only 1% of EOF in either POTW influent and effluent. Additionally, targeted precursors accounted for <25% of the TOP + BI estimate of total precursors in influent and effluent. Suspect screening for >1,000 additional precursors in these samples only identified six additional structures in the aqueous phase with low detection frequency and peak areas (24). Therefore, it is likely that the exact molecular structures of most PFAA precursors in wastewater influents and effluents remain unknown and the TOP assay is presently the only way to holistically quantify this class of PFAS.
Fluorinated Pharmaceuticals Account for the Majority of EOF.
We further identified 12 fluorinated pharmaceuticals/metabolites that accounted for a mean of 75% of the EOF in wastewater influent and 62% of the EOF in effluent (Dataset S2). Polyfluorinated pharmaceuticals and metabolites that contain –CF3– or –CF2– moieties (celecoxib, flecainide, maraviroc, hydroxyphenylmaraviroc, and sitagliptin) made up the majority of quantified EOF in POTW influent (58 ± 15%) and effluent (53 ± 20%). Monofluorinated pharmaceuticals and metabolites that only contain –CF groups (atorvastatin, hydroxyatorvastatin, citalopram, desmethylcitalopram, diflunisal, pantoprazole sulfide, and rosuvastatin) accounted for a smaller fraction of EOF in POTW influent (17 ± 6%) and effluent (10 ± 9%). Suspect screening for all registered fluorinated pharmaceuticals, known human metabolites, and agrochemicals did not reveal any other compounds in influent or effluent (Datasets S3 and S4) (25–27).
The eight parent pharmaceuticals detected at all eight POTWs in this study are among the most commonly used medications in the United States (28), suggesting that they may be ubiquitously present in municipal wastewater. By contrast, suspect screening did not detect a few highly prescribed fluorinated pharmaceuticals (fluticasone, fluoxetine, paroxetine, ezetimibe, and known human metabolites) in the aqueous phase at any of the study POTWs. We confirmed these results on raw water samples to eliminate the potential for bias related to losses during the extraction procedure. Similarly, a prior survey of major US POTWs infrequently or never detected these compounds in influent or effluent (29). The absence of some frequently prescribed pharmaceuticals could indicate unknown transformation products that are not included in the known human metabolites screening database (27). Alternately, these compounds may have preferential affinity for wastewater biosolids, which should be explored in future work.
EOF Mass Budget in Wastewater.
The sum of targeted PFAS, precursors (TOP + BI), and fluorinated pharmaceuticals explains all of the EOF in aqueous influent and effluent samples in this study, within commonly accepted uncertainty bounds (±30%) (30–32) in all but two samples (SI Appendix, Fig. S1). Our extraction procedure for EOF did not retain trifluoroacetate (TFA), the smallest molecular weight PFAA. In prior work, TFA has been found ubiquitously in surface water and reflects degradation of hydrofluorocarbons and hydrochlorofluorocarbons (33, 34). Concentrations of TFA in wastewater in Europe have been reported to be ~20 nM F (35), indicating that it is an important component of organofluorine in wastewater discharges.
Results of this work showing near quantitative reproduction of the EOF by targeted compounds in aqueous wastewater effluent contrast with work on biosolids in Sweden (18). That work showed 73% of the EOF was not explained by a similar suite of PFAS and the same suite of pharmaceuticals as this study. Additional work identifying organofluorine associated with wastewater biosolids is therefore needed to fully characterize potential releases to terrestrial environments. The fate of biosolids differs from wastewater effluent since biosolids are often intentionally distributed on soils and various landscapes, creating alternate PFAS exposure concerns.
EOF Poorly Removed During Wastewater Treatment.
All eight POTWs in this study had primary (physical screening/settling) and secondary (microbial processing of labile organic matter) treatment. Half of the facilities had advanced tertiary treatments including ozonation, activated carbon filtration, and ultrafiltration. However, we observed a maximum of 24% decline in aqueous-phase EOF compared to influent and no significant differences between aqueous influent and effluent concentrations (two-sided paired t test; P-value = 0.15; Fig. 1C and Dataset S5). Similar to this study, previous work at two wastewater facilities in Austria noted no change in EOF concentrations between influent and effluent (17). In our work, monofluorinated pharmaceuticals were the only organofluorine compounds that exhibited significant removal from the aqueous phase during treatment (two-sided paired t test; P-value = 0.04) whereas PFAA concentrations increased rather than decreased at 75% of the facilities during treatment. Increasing concentrations of PFAA at this study’s POTWs is consistent with previous work showing transformation of precursors into PFAA during wastewater treatment (36, 37).
National Estimates of Aqueous Organofluorine Discharges from POTWs.
The eight POTWs in our work used a range of treatment technologies and served populations characteristic of large POTWs across the US Large POTWs serve >230 million US residents (>90% of the population using centralized wastewater treatment) and 70% of the total population (20). A comparison of aqueous wastewater PFAA concentrations from this work to prior US studies shows similar concentrations and ranges (14, 38). All POTWs fell within one SD of the US average (SI Appendix, Fig. S2), suggesting that the range of concentrations measured at facilities in this study provides a reasonable screening level estimate of EOF concentrations for other large US POTWs, in the absence of other data.
We found a strong (R2 = 0.7) and near-linear relationship between the sizes of populations served by the facilities included in this study and EOF mass discharge (Methods: Eq. 1 and Dataset S6). Using populations served from the US EPA’s 2022 Clean Watersheds Needs Survey (20), we extended this relationship for large POTWs at the national scale. We limited estimates to large POTWs because measurements in this study did not include POTWs serving smaller populations. Releases from smaller POTWs may be very different from larger ones due to their unique service populations and operating conditions, and potential impacts from industrial discharges.
Results of this extrapolation from measured concentrations suggest that the total magnitude of aqueous EOF discharges from large US POTWs is approximately 1.7 million mol (or Mmol) F (90th percentile CI = 1.0 to 2.8 Mmol F). We compared this estimate to the mass of EOF prescribed nationally in fluorinated pharmaceuticals for the same populations. In 2020, prescribed fluorinated pharmaceuticals accounted for ~3.6 Mmol F across the United States (Dataset S7) (28). Scaling that number by the fraction of the population served by large POTWs (70%), and an assumed maximum removal efficiency of 24% from this study, produces a remarkably similar estimate of ~1.9 Mmol F to that based on POTW discharges. These results suggest that virtually all the organofluorine in the US pharmaceutical prescriptions examined in this work ends up in wastewater effluent and much of this enters downgradient hydrological systems.
Large Influence of Wastewater PFAS on Drinking Water Quality.
In 2024, the US Environmental Protection Agency (EPA) finalized federal regulations for six PFAS in drinking water (39). PFAS regulated by EPA include PFOS and PFOA (4 ng L−1 or 0.14 nM F each) and the hazardous mixture of PFBS, perfluorohexane sulfonate (PFHxS), perfluorononanoate (PFNA), and hexafluoropropylene oxide dimer acid (HFPO-DA/GenX) (SI Appendix, Eq. S1). Current federal PFAS drinking water regulations do not include other PFAA, PFAA precursors, or the polyfluorinated pharmaceuticals measured in POTW effluent in this work.
The six regulated PFAS accounted for an average of 8 ± 8% of the quantified EOF (max = 25%) in POTW effluent samples measured in this work (Dataset S2). PFOS and PFOA exceeded federal standards in 63% of the effluents, while the hazardous PFAS mixture standard was not exceeded in any effluent. The greatest exceedance at any facility was observed for PFOA (six times greater than the regulation). At that site, environmental dilution with contaminant-free water or drinking water treatment up to a factor of six would be needed to prevent downstream concentrations that exceed regulatory standards.
We used the DRINCS model to assess the influence of upstream POTW discharges on downstream water quality at drinking water intakes after dilution in natural waters (Methods: Eq. 2) (8, 9). Specifically, we forced POTW discharges in the model using mean PFAS concentrations measured in this work and compared diluted concentrations at almost 7,000 drinking water intakes across the continental United States to federal drinking water standards (Fig. 2). Simulations using the 90th percentile CI around mean PFAS concentration were performed to estimate model uncertainty. We considered both average and low flow conditions to account for variability due to potential droughts and climate change.
Fig. 2.
The model assumes a) wastewater treatment plants are operating at their design capacity, b) no contributions from combined sewer overflows or wet weather by-passes, c) no in-stream losses of chemical tracer in wastewater effluent, and d) complete mixing. The use of design capacity, instead of operationing capacity, represents a conservative estimate of de facto wastewater reuse (8, 9). Field studies using a conservative chemical tracer for wastewater discharges (sucralose) and paired discharge and PFAS measurements support assumptions (b–d) (9, 40). Other field studies that have sampled downgradient transects in natural river systems suggest that simple dilution is the main factor controlling shifts in PFAS concentrations (41). However, additional investigations on how water chemistry (e.g., pH, salinity, particulates) affects downstream PFAS fate and transport would be helpful.
Under average riverine flow conditions, our modeling results suggest that wastewater-derived PFAS is sufficient to result in exceedances of federal regulatory limits at 1% of drinking water intakes serving ~15 million people (Fig. 2). Since conventional drinking water treatment is not effective at reducing PFAS concentrations (42, 43), advanced treatment at these facilities (e.g., granular activated carbon filtration) and upstream source mitigation would be needed to prevent hazardous exposures. Impacted facilities are distributed throughout the country and are particularly prevalent in the Los Angeles region of California, the South, and the mid-Atlantic/Ohio. Low flow conditions (i.e., fall and drought) result in less dilution of PFAS-contaminated wastewater prior to reentering drinking water supplies. Our results suggest that the number of drinking water intakes that exceed the Federal PFAS drinking water standards could triple under low flow conditions and increase the affected population to 23 million people. Climate change is expected to alter total precipitation as well as increase the frequency of drought throughout the United States (44). Since utilities must ensure safe drinking water all year, increasingly common low flow conditions due to frequent drought must be accounted for in future water quality management.
A limitation of this modeling analysis is that we did not consider temporal variability in PFAS discharges and instead relied on PFAS concentrations measured in POTW effluent at a single point in time. Preliminary work suggests that PFAS discharges do not exhibit substantial seasonality (14), although additional research is needed to support these findings. Modeled estimates quantify the impact of wastewater discharges to drinking water intake quality and do not consider other major PFAS sources, such as military bases, airports, and manufacturing facilities (5, 6). These other PFAS sources add to the total contamination in drinking water and are expected to increase both the number of facilities and populations served that exceed regulatory standards for PFAS in source water (45).
Potential Impacts of Wastewater Pharmaceutical Discharges.
Current federal drinking water regulations do not include the most abundant PFAS found in wastewater effluent. By contrast, E.U. regulations specify a “sum of PFAS” < 100 ng L−1 in drinking water for the sum of two subclasses of PFAA, perfluoroalkyl carboxylates with 3 to 12 perfluorinated carbons and perfluoroalkyl sulfonates with 4 to 13 perfluorinated carbons (SI Appendix, Eq. S2) (46). These PFAA account for 21 ± 13% of the quantified EOF in POTW effluents measured here (Dataset S2). The E.U. also considers a “PFAS total” standard for the sum of all PFAS of <500 ng L−1 (SI Appendix, Eq. S3), but whether this regulation includes polyfluorinated pharmaceuticals is ambiguous. Without polyfluorinated pharmaceuticals, “PFAS total” includes 37 ± 21% of the EOF in POTW effluents in this study. With polyfluorinated pharmaceuticals, the fraction of EOF included in the “PFAS total” regulation would increase to 90 ± 9%.
A greater fraction of EOF discharges from POTWs are included in E.U. PFAS regulations compared to equivalent US standards. However, modeling results show that both regulatory approaches result in similar number of impaired facilities and impacted populations (SI Appendix, Fig. S3). This reflects a trade-off between more expansive coverage but higher thresholds in the E.U. versus narrower coverage but lower thresholds in the United States.
Implications.
Chemical regulation in the United States typically considers risks associated with individual toxicants rather than the complex mixtures present in wastewater effluent or the environment (47). This poses a challenge for regulating PFAS, pharmaceuticals, and other organofluorine compounds because there are potentially tens of thousands of these chemicals in commerce (2). Most organofluorine compounds lack analytical standards needed for routine environmental measurements and for evaluating toxicity. Experts have called for a class-based approach for regulating organofluorine, focusing on PFAS, due in part to the extreme persistence of these compounds and their transformation products (48). EOF measurements could potentially support such an approach. One challenge with EOF is that it cannot distinguish among types of organofluorine moieties (i.e., more labile –CF chemistries versus more persistent –CF2– and –CF3 chemistries). Therefore, EOF is useful as a first screen for organofluorine contamination but should be complemented by other analytical methods that can confirm the presence of specific compounds or chemical moieties of concern. Future methodological improvements, such as PFAS-specific extraction procedures, may further enhance the utility of EOF measurements for regulatory use (49).
The utility of EOF measurements in wastewater discharges will depend in part upon the exact definition of PFAS for potential class-based regulation and management approaches. Using a variety of analytical methods, we showed that chemicals with at least one –CF3 and/or –CF2– group made up 90 ± 9% of the quantified EOF in effluent from eight large POTWs (Fig. 1B and Dataset S2). PFAA and PFAA precursors are the subclasses within the PFAS chemical family that have received the most attention by the scientific and regulatory communities. However, they only accounted for 21 ± 13% of quantified EOF in wastewater effluent and their concentrations were not significantly correlated with EOF concentrations (Pearson correlation coefficient = 0.2; P-value = 0.7) (Dataset S1). Instead, polyfluorinated pharmaceuticals accounted for the majority of quantified EOF in POTW effluent. Monitoring and a class-based PFAS regulation for municipal wastewater discharge through the National Pollutant Discharge Elimination System permits under the US Clean Water Act is a priority of the Federal PFAS Strategic Roadmap (50). Thus, clarification on how PFAS are defined is urgently needed to support such oversight of pollutant discharges.
The US Food and Drug Administration’s drug approval process does not consider the environmental persistence and secondary human and ecological exposures to pharmaceuticals (51). Pharmaceuticals are bioactive at doses much lower than they are prescribed and have been shown to alter animal physiology in ecosystems downstream from wastewater discharges (52). Pharmaceuticals that are persistent enough to reenter drinking water supplies, such as the highly recalcitrant organofluorine compounds in this study, could therefore affect otherwise healthy and/or sensitive human populations (i.e., pregnant people, children). Organofluorine pharmaceuticals now make up ~20% of all commercialized pharmaceuticals and have become the dominant chemistry in the past decade (1, 25). The potential public health consequences of ubiquitous low-level exposures to these compounds require urgent consideration.
Future Research Needs.
Data presented here suggest that US POTWs do not effectively remove most EOF prior to effluent discharge, regardless of whether they have secondary or tertiary treatment (Fig. 1C and Dataset S5). Aquatic discharges from POTWs contain elevated levels of PFAS, including PFAA, PFAA precursors, and polyfluorinated pharmaceuticals. These discharges impact receiving water bodies, including downstream drinking water intakes, aquifers, surface waters, and the ocean. Past work has suggested that the surface ocean will be the terminal receptor of PFAA (11, 12, 53). However, the environmental fate of the most abundant organofluorine compounds in this work is largely unknown. Future work should include random sampling of a wider array of diverse wastewater treatment facilities. A better understanding of the fate and distribution of organofluorine pharmaceutical/transformation products as well as further elucidation of exact PFAA precursor structures in water infrastructure is also urgently needed. These steps will help support a broader characterization of ecosystem/human health impacts of organofluorine pharmaceuticals, replacement PFAS, and their mixtures.
Methods
Sampling Protocol.
24-hour flow-weighted composite samples were collected from the influent and effluent of eight large POTWs in the continental United States throughout the summer of 2021 as part of a collaborative study with the Water Research Foundation (14, 24). Each POTW sampled serves metropolitan populations with more than 10,000 customers and had average wastewater flows of greater than 26,000 m3 d−1 (Dataset S6). The identities of the facilities are confidential, and samples were given anonymous numerical identifiers prior to analysis at Harvard University. Full details of the sampling protocol are provided in SI Appendix.
Fluorine Quantification Overview.
Centrifuged subsamples representing the aqueous phase of POTW influent and effluent were analyzed for a) total fluorine using combustion ion chromatography (CIC), b) fluoride using ion chromatography, c) EOF using CIC, d) PFAS using liquid chromatography–tandem mass spectrometry (LC-MS/MS), and e) organofluorine pharmaceuticals and agrochemicals using high-resolution mass spectrometry (HRMS). Total fluorine (which represents both organofluorine and fluoride) and fluoride analyses were performed on subsamples of raw water. To elucidate the organofluorine composition, organofluorine must be separated from fluoride prior to determination on the CIC. We performed this separation on subsamples through solid phase extraction using mixed mode weak anion exchange cartridges that we then split three ways for EOF, PFAS, and organofluorine pharmaceutical/agrochemical analyses. During the extraction phase, inorganic fluoride that could interfere with the organofluorine quantification (54) was completely removed (SI Appendix). Analyses followed previously established methods (22, 23, 55), which are described in detail in SI Appendix.
LC-MS/MS analysis for targeted PFAS quantified 34 analytes including 18 PFAA and 16 precursors (Dataset S8). PFAS analysis on EOF extracts differs from routine targeted PFAS analyses in that internal standards (IS) are not added to samples prior to extractions (30, 31) because IS addition interferes with organofluorine quantification.
HRMS analysis for all-time globally registered small molecular weight (<1,500 g mol−1) organofluorine pharmaceuticals, known human metabolites, and agrochemicals screened for 766 molecules (Datasets S3 and S4) (25, 26). We obtained analytical standards for the compounds that were positively matched to suspect screening lists (eight pharmaceuticals and four metabolites) and subsequently quantified them on the same instrument (Dataset S9).
Total PFAA precursors were estimated from the TOP assay using Bayesian inference (TOP + BI) (22, 23). The TOP assay was performed on separate subsamples and reported in a previous publication that was also part of the collaborative study with the Water Research Foundation (Dataset S10) (24). Estimation of total PFAA precursors from the TOP + BI method, which corrects for mass loss during the procedure, has been shown to more accurately estimate precursor concentrations compared to raw TOP data (22, 55). Details of this method are provided in SI and the Bayesian inference code is publicly available at https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/BWEW5H (56).
National EOF Discharges.
The mass of EOF discharged from each facility included in this study was calculated from measured concentrations and their reported annual flow rates (Dataset S6). Populations served by each facility were also reported. Following prior work (57, 58), we fit a power law relationship between the mass of EOF discharged from each POTW and population (Eq. 1; Fig. 3). This empirical approach uses population as a proxy for both industrial and consumer sources of EOF:
[1]
Fig. 3.
where
Ew is the mass discharge from a POTW (mol F day−1; the product of the concentration and facility flow rate),
β is the baseline per capita mass discharge rate (mol F day−1 person−1),
Pw is the population served by the POTW (person), and
α is a scaling factor (unitless).
We calculated α (0.94 ± 0.07) and β (3 × 10−5 ± 2 × 10−6 mol F day−1 person−1), by log-transforming Eq. 1 and using Bayesian linear regression implemented in PyMC3 version 3.11.5 (59) in Python version 3.9.7. Normal priors were used for α and ln β with means equal to the values determined from ordinary least squares regression and SD equal to 1. The No-U-Turn (NUTs) sampler with four independent walkers was used to sample from the posterior distribution with a step size of 0.9. A burn-in window of 1000 samples was followed by the collection of 5000 samples. The effective sample size for both α and ln β were >4,500. The mean and SD of β was calculated from the expected value and variance of the log-normal distribution of ln β.
The posterior prediction of daily EOF discharges was calculated and the mean and 90th percentile CI are presented throughout the manuscript. The posterior predictive captures all observations within the 90th percentile CI, showing that the model can fully capture the spread of the data (Fig. 3). These results suggest that the service population provides a good proxy indicator of variability among observed EOF mass discharges in large POTWs.
For large POTWs across the country (those serving >10,000 customers), we estimated aqueous municipal wastewater discharges of EOF, PFAS, and fluorinated pharmaceuticals using this simple empirical estimate (Eq. 1) and populations served reported in the 2022 US EPA Clean Watersheds Needs Survey (20). Results were compared to fluorinated pharmaceutical prescription data from the US Department of Health and Human Services’ 2020 Medical Expenditure Panel Survey (Dataset S7) (28). The code to recreate these analyses is available at https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/BWEW5H (56).
De Facto Wastewater Reuse in the Drinking Water Systems Model.
We used the DRINCS model to relate wastewater PFAS discharges measured in this study to downstream drinking water intake quality (8). DRINCS is a proprietary geospatial model that estimates the percentage of de facto wastewater reuse (%DFR) at drinking water intakes according to Eq. 2:
[2]
where is the cumulative design capacity discharge of all upstream wastewater treatment plants in the US EPA’s 2012 Clean Watersheds Needs Survey, is the streamflow at drinking water intake from the US EPA’s NHDPlus Version 2 Database (60). We used long-term annual averages and minima of to evaluate %DFR under different hydrological conditions.
Data, Materials, and Software Availability
Source code and input data have been deposited in the Harvard Dataverse Repository (https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/BWEW5H) (56). All study data are included in the article and/or supporting information.
Acknowledgments
Financial support for this work was provided by the National Institute for Environmental Health Science Superfund Research Program (P42ES027706) and the Water Research Foundation (Project 5031). This study was also supported by contributions from the anonymous participating wastewater treatment facilities. The content is solely the responsibility of the authors and does not necessarily represent the official views of the funders.
Author contributions
B.J.R. designed research; B.J.R., S.V., and J.B. performed research; T.F.W., W.H.-B., R.L., P.W., C.E.S., and E.M.S. contributed new reagents/analytic tools; B.J.R., E.H.P., and M.I. analyzed data; and B.J.R. and E.M.S. wrote the paper.
Competing interests
The authors declare no competing interest.
Supporting Information
Appendix 01 (PDF)
- Download
- 394.43 KB
Dataset S01 (XLSX)
- Download
- 23.33 KB
Dataset S02 (XLSX)
- Download
- 23.99 KB
Dataset S03 (XLSX)
- Download
- 27.48 KB
Dataset S04 (XLSX)
- Download
- 16.73 KB
Dataset S05 (XLSX)
- Download
- 10.33 KB
Dataset S06 (XLSX)
- Download
- 9.79 KB
Dataset S07 (XLSX)
- Download
- 10.65 KB
Dataset S08 (XLSX)
- Download
- 12.56 KB
Dataset S09 (XLSX)
- Download
- 10.21 KB
Dataset S10 (XLSX)
- Download
- 10.33 KB
References
1
J. Han et al., Chemical aspects of human and environmental overload with fluorine. Chem. Rev. 121, 4678–4742 (2021).
2
Z. Wang, J. C. DeWitt, C. P. Higgins, I. T. Cousins, A never-ending story of per- and polyfluoroalkyl substances (PFASs)? Environ. Sci. Technol. 51, 2508–2518 (2017).
3
E. K. Richmond et al., Pharmaceuticals and personal care products (PPCPs) are ecological disrupting compounds (EcoDC). Elementa (Wash. D. C.) 5, 66 (2017).
4
S. E. Fenton et al., Per- and polyfluoroalkyl substance toxicity and human health review: Current state of knowledge and strategies for informing future research. Environ. Toxicol. Chem. 40, 606–630 (2021).
5
X. C. Hu et al., Detection of poly- and perfluoroalkyl substances (PFASs) in U.S. drinking water linked to industrial sites, military fire training areas, and wastewater treatment plants. Environ. Sci. Technol. Lett. 3, 344–350 (2016).
6
J. M. Liddie, L. A. Schaider, E. M. Sunderland, Sociodemographic factors are associated with the abundance of PFAS sources and detection in U.S. community water systems. Environ. Sci. Technol. 57, 7902–7912 (2023).
7
Committee on the Assessment of Water Reuse as an Approach for Meeting Future Water Supply Needs, Water Reuse: Potential for Expanding the Nation’s Water Supply Through Reuse of Municipal Wastewater (The National Academies Press, 2012).
8
J. Rice, P. Westerhoff, Spatial and temporal variation in De Facto wastewater reuse in drinking water systems across the U.S.A. Environ. Sci. Technol. 49, 982–989 (2015).
9
M. Islam et al., Sucralose and predicted De Facto wastewater reuse levels correlate with PFAS levels in surface waters. Environ. Sci. Technol. Lett. 10, 431–438 (2023).
10
W. H. Schlesinger, E. M. Klein, A. Vengosh, Global biogeochemical cycle of fluorine. Global Biogeochem. Cycles 34, e2020GB006722 (2020).
11
Z. Wang, I. T. Cousins, M. Scheringer, R. C. Buck, K. Hungerbühler, Global emission inventories for C4–C14 perfluoroalkyl carboxylic acid (PFCA) homologues from 1951 to 2030, Part I: Production and emissions from quantifiable sources. Environ. Int. 70, 62–75 (2014).
12
Z. Wang, J. M. Boucher, M. Scheringer, I. T. Cousins, K. Hungerbühler, Toward a comprehensive global emission inventory of C 4 –C 10 perfluoroalkanesulfonic acids (PFSAs) and related precursors: Focus on the life cycle of C 8-based products and ongoing industrial transition. Environ. Sci. Technol. 51, 4482–4493 (2017).
13
J. L. Wilkinson et al., Pharmaceutical pollution of the world’s rivers. Proc. Natl. Acad. Sci. U.S.A. 119, e2113947119 (2022).
14
K. A. Thompson et al., Poly- and perfluoroalkyl substances in municipal wastewater treatment plants in the United States: Seasonal patterns and meta-analysis of long-term trends and average concentrations. ACS EST Water 2, 690–700 (2022).
15
A. Kärrman et al., Can determination of extractable organofluorine (EOF) be standardized? First interlaboratory comparisons of EOF and fluorine mass balance in sludge and water matrices. Environ. Sci. Process. Impacts 23, 1458–1465 (2021).
16
R. Aro, P. Carlsson, C. Vogelsang, A. Kärrman, L. W. Y. Yeung, Fluorine mass balance analysis of selected environmental samples from Norway. Chemosphere 283, 131200 (2021).
17
V. Müller, A. Kindness, J. Feldmann, Fluorine mass balance analysis of PFAS in communal waters at a wastewater plant from Austria. Water Res. 244, 120501 (2023).
18
K. M. Spaan, F. Seilitz, M. M. Plassmann, C. A. de Wit, J. P. Benskin, Pharmaceuticals account for a significant proportion of the extractable organic fluorine in municipal wastewater treatment plant sludge. Environ. Sci. Technol. Lett. 10, 328–336 (2023).
19
U.S. Environmental Protection Agency, 2006 Community Water System Survey Volume 1: Overview (U.S. Environmental Protection Agency Office of Water, 2009).
20
K. Marrese et al., 2022 Clean Watersheds Needs Survey Report to Congress (Environmental Protection Agency, U.S., 2024).
21
E. F. Houtz, D. L. Sedlak, Oxidative conversion as a means of detecting precursors to perfluoroalkyl acids in urban runoff. Environ. Sci. Technol. 46, 9342–9349 (2012).
22
B. J. Ruyle et al., Reconstructing the composition of per- and polyfluoroalkyl substances in contemporary aqueous film-forming foams. Environ. Sci. Technol. Lett. 8, 59–65 (2021).
23
B. J. Ruyle et al., Isolating the AFFF signature in coastal watersheds using oxidizable PFAS precursors and unexplained organofluorine. Environ. Sci. Technol. 55, 3686–3695 (2021).
24
C. E. Schaefer et al., Occurrence of quantifiable and semi-quantifiable poly- and perfluoroalkyl substances in united states wastewater treatment plants. Water Res. 233, 119724 (2023).
25
M. Inoue, Y. Sumii, N. Shibata, Contribution of organofluorine compounds to pharmaceuticals. ACS Omega 5, 10633–10640 (2020).
26
Y. Ogawa, E. Tokunaga, O. Kobayashi, K. Hirai, N. Shibata, Current contributions of organofluorine compounds to the agrochemical industry. iScience 23, 101467 (2020).
27
D. S. Wishart et al., DrugBank 5.0: A major update to the DrugBank database for 2018. Nucleic Acids Res. 46, D1074–D1082 (2018).
28
Agency for Healthcare Research and Quality, Medical expenditure panel survey (MEPS) 2013-2020 ClinCalc DrugStates Database version 2022.08. (2023). https://clincalc.com/DrugStats/. Accessed 1 November 2023.
29
M. S. Kostich, A. L. Batt, J. M. Lazorchak, Concentrations of prioritized pharmaceuticals in effluents from 50 large wastewater treatment plants in the US and implications for risk estimation. Environ. Poll. 184, 354–359 (2014).
30
J. A. Shoemaker, D. R. Tettenhorst, Method 537.1 determination of selected per- and polyfluorinated alkyl substances in drinking water by solid phase extraction and liquid chromatography/tandem mass spectrometry (LC/MS/MS). (2018). https://cfpub.epa.gov/si/si_public_record_Report.cfm?dirEntryId=343042&Lab=NERL. Accessed 3 November 2023.
31
L. Rosenblum, S. C. Wendelken, Method 533: Determination of per- and polyfluoroalkyl substances in drinking water by isotope dilution anion exchange solid phase extraction and liquid chromatography/tandem mass spectrometry. (2019). https://www.epa.gov/sites/default/files/2019-12/documents/method-533-815b19020.pdf. Accessed 3 November 2023.
32
E. Furlong et al., Determination of Human-Use Pharmaceuticals in Filtered Water by Direct Aqueous Injection–High-Performance Liquid Chromatography/Tandem Mass Spectrometry (U.S. Geological Survey, 2014).
33
J. Janda, K. Nödler, H.-J. Brauch, C. Zwiener, F. T. Lange, Robust trace analysis of polar (C2–C8) perfluorinated carboxylic acids by liquid chromatography-tandem mass spectrometry: Method development and application to surface water, groundwater and drinking water. Environ. Sci. Pollut. Res. 26, 7326–7336 (2019).
34
H. M. Pickard et al., Ice core record of persistent short-chain fluorinated alkyl acids: Evidence of the impact from global environmental regulations. Geophys. Res. Lett. 47, e2020GL087535 (2020).
35
M. Scheurer et al., Small, mobile, persistent: Trifluoroacetate in the water cycle–Overlooked sources, pathways, and consequences for drinking water supply. Water Res. 126, 460–471 (2017).
36
M. M. Schultz et al., Fluorochemical mass flows in a municipal wastewater treatment facility. Environ. Sci. Technol. 40, 7350–7357 (2006).
37
E. Sinclair, K. Kannan, Mass loading and fate of perfluoroalkyl surfactants in wastewater treatment plants. Environ. Sci. Technol. 40, 1408–1414 (2006).
38
E. H. Antell et al., The total oxidizable precursor (TOP) assay as a forensic tool for per- and polyfluoroalkyl substances (PFAS) source apportionment. ACS EST Water 4, 948–957 (2023), https://doi.org/10.1021/acsestwater.3c00106.
39
U.S. Environmental Protection Agency, Part 141 - National Primary Drinking Water Regulations (Federal Register 141:373–637). Regulations.gov (2023). https://www.regulations.gov/document/EPA-HQ-OW-2022-0114-0027. Accessed 27 June 2023.
40
E. E. Woodward et al., Using a time-of-travel sampling approach to quantify per- and polyfluoroalkyl substances (PFAS) stream loading and source inputs in a mixed-source, Urban Catchment. ACS EST Water 4, 4356–4367 (2024).
41
H. M. Pickard et al., Characterizing the areal extent of PFAS contamination in fish species downgradient of AFFF source zones. Environ. Sci. Technol. 58, 19440–19453 (2024).
42
C. Eschauzier, E. Beerendonk, P. Scholte-Veenendaal, P. De Voogt, Impact of treatment processes on the removal of perfluoroalkyl acids from the drinking water production chain. Environ. Sci. Technol. 46, 1708–1715 (2012).
43
T. D. Appleman et al., Treatment of poly- and perfluoroalkyl substances in U.S. full-scale water treatment systems. Water Res. 51, 246–255 (2014).
44
B. I. Cook, J. S. Mankin, K. J. Anchukaitis, Climate change and drought: From past to future. Curr. Clim. Change Rep. 4, 164–179 (2018).
45
D. Q. Andrews, O. V. Naidenko, Population-wide exposure to per- and polyfluoroalkyl substances from drinking water in the united states. Environ. Sci. Technol. Lett. 7, 931–936 (2020).
46
EUR-Lex - L:2020:435:TOC - EN - EUR-Lex. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=OJ%3AL%3A2020%3A435%3ATOC. Accessed 17 January 2024.
47
A. Kortenkamp, M. Faust, Regulate to reduce chemical mixture risk. Science 361, 224–226 (2018).
48
C. F. Kwiatkowski et al., Scientific basis for managing PFAS as a chemical class. Environ. Sci. Technol. Lett. 7, 532–543 (2020).
49
Y. Pan, D. E. Helbling, Revealing the factors resulting in incomplete recovery of perfluoroalkyl acids (PFAAs) when implementing the adsorbable and extractable organic fluorine methods. Water Res. 244, 120497 (2023).
50
R. Fox, Addressing PFAS Discharges in NPDES permits and through the pretreatment program and monitoring programs. (2022). https://www.epa.gov/system/files/documents/2022-12/NPDES_PFAS_State%20Memo_December_2022.pdf. Accessed 17 November 2023.
51
T. Brodin et al., The urgent need for designing greener drugs. Nat. Sustain. 7, 949–951 (2024), https://doi.org/10.1038/s41893-024-01374-y.
52
A. Bouzas-Monroy, J. L. Wilkinson, M. Melling, A. B. A. Boxall, Assessment of the potential ecotoxicological effects of pharmaceuticals in the world’s rivers. Environ. Toxicol. Chem. 41, 2008–2020 (2022).
53
J. M. Armitage et al., Modeling the global fate and transport of perfluorooctane sulfonate (PFOS) and precursor compounds in relation to temporal trends in wildlife exposure. Environ. Sci. Technol. 43, 9274–9280 (2009).
54
B. J. Ruyle et al., Interlaboratory comparison of extractable organofluorine measurements in groundwater and Eel (anguilla rostrata): Recommendations for methods standardization. Environ. Sci. Technol. 57, 20159–20168 (2023).
55
B. J. Ruyle et al., Centurial persistence of forever chemicals at military fire training sites. Environ. Sci. Technol. 57, 8096–8106 (2023).
56
B. J. Ruyle et al., U.S. Municipal Wastewater Organofluorine. Harvard Dataverse. https://doi.org/10.7910/DVN/BWEW5H. Deposited 20 December 2024.
57
A. Pistocchi, R. Loos, A map of european emissions and concentrations of PFOS and PFOA. Environ. Sci. Technol. 43, 9237–9244 (2009).
58
X. Zhang, Y. Zhang, C. Dassuncao, R. Lohmann, E. M. Sunderland, North atlantic deep water formation inhibits high arctic contamination by continental perfluorooctane sulfonate discharges. Global Biogeochem. Cycles 31, 1332–1343 (2017).
59
J. Salvatier, T. V. Wiecki, C. Fonnesbeck, Probabilistic programming in Python using PyMC3. PeerJ Comput. Sci. 2, e55 (2016).
60
U.S. Environmental Protection Agency, NHDPlus Version 2.0. (2023). https://www.epa.gov/waterdata/nhdplus-national-hydrography-dataset-plus. Accessed 18 December 2023.
Information & Authors
Information
Published in
Classifications
Copyright
Copyright © 2025 the Author(s). Published by PNAS. This open access article is distributed under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND).
Data, Materials, and Software Availability
Source code and input data have been deposited in the Harvard Dataverse Repository (https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/BWEW5H) (56). All study data are included in the article and/or supporting information.
Submission history
Received: August 22, 2024
Accepted: December 9, 2024
Published online: January 6, 2025
Published in issue: January 21, 2025
Keywords
Acknowledgments
Financial support for this work was provided by the National Institute for Environmental Health Science Superfund Research Program (P42ES027706) and the Water Research Foundation (Project 5031). This study was also supported by contributions from the anonymous participating wastewater treatment facilities. The content is solely the responsibility of the authors and does not necessarily represent the official views of the funders.
Author contributions
B.J.R. designed research; B.J.R., S.V., and J.B. performed research; T.F.W., W.H.-B., R.L., P.W., C.E.S., and E.M.S. contributed new reagents/analytic tools; B.J.R., E.H.P., and M.I. analyzed data; and B.J.R. and E.M.S. wrote the paper.
Competing interests
The authors declare no competing interest.
Notes
This article is a PNAS Direct Submission.
Authors
Metrics & Citations
Metrics
Citation statements
Altmetrics
Citations
Cite this article
High organofluorine concentrations in municipal wastewater affect downstream drinking water supplies for millions of Americans, Proc. Natl. Acad. Sci. U.S.A.
122 (3) e2417156122,
https://doi.org/10.1073/pnas.2417156122
(2025).
Copied!
Copying failed.
Export the article citation data by selecting a format from the list below and clicking Export.
Cited by
Loading...
View Options
View options
PDF format
Download this article as a PDF file
DOWNLOAD PDFLogin options
Check if you have access through your login credentials or your institution to get full access on this article.
Personal login Institutional LoginRecommend to a librarian
Recommend PNAS to a LibrarianPurchase options
Purchase this article to access the full text.