Per- and Polyfluoroalkyl Substances in Potential Drinking Water Sources Globally: Distributions, Monitoring Trends, and Risk Assessment

Abstract

Due to widespread industrial applications and increased discharges, concentrations of polyfluoroalkyl substances (PFAS) in potential drinking water sources have risen significantly, putting more people at risk of PFAS exposure. This study aimed to systematically clarify the occurrence characteristics (concentrations, detection frequencies, and temporal trends) of PFAS in global potential drinking water sources over the past decade, assess their oral exposure risks, and identify key PFAS species with high detection frequencies, high contamination levels, or high toxicity risks, thereby providing scientific support for the development of targeted control technologies and management strategies. This study systematically searched and reviewed the relevant literature published between 2014 and 2024 on PFAS levels in global potential drinking water sources, extracting data on PFAS concentrations, detection information, and sampling characteristics. Using the U.S. Environmental Protection Agency (EPA) Reference Dose (RfD) for oral exposure as the Acceptable Daily Intake (ADI), we evaluated the exposure risks of eight specific PFAS via the Risk Quotient for Specific Contaminants (RQ_RSC) model and analyzed the annual detection trends of the top thirty PFAS with the highest detection frequencies. Regarding total PFAS contamination, China, Brazil, Germany, South Africa, and the Danube River Basin exhibited particularly high levels, with China being the most severely contaminated. Risk assessment indicated that 45.6% of global potential drinking water sources were at high risk (RQ_RSC > 1), while 48.4% were at low risk (RQ_RSC < 0.2). Among the evaluated PFAS, PFOA, PFOS, PFDA, and GenX were associated with higher toxicity exposure risks. For the identified key concern PFAS, it is necessary to simplify detection techniques, promote targeted large-scale safe treatment technologies, and explore intelligent monitoring tools to reduce regulatory lag, thereby effectively monitoring, preventing, and controlling PFAS contamination.

1. Introduction

Since the introduction of PFAS in the middle of the 20th century, they have rapidly occupied an important position in key areas such as textiles, firefighting, medical care, and food packaging by their excellent heat resistance and chemical stability [1], which has greatly contributed to the progress of these industries [2]. However, this widespread application has also led to the excessive consumption and environmental pollution of PFAS, whose excellent ability to migrate across geographic boundaries and whose presence has been detected even in polar waters far away from human activities [3], further emphasizing the global nature and severity of PFAS contamination [4].

PFAS are highly stable and resistant to biodegradation due to their extremely stable and difficult-to-break carbon-fluorine (C-F) bonds, and they are persistent and widely distributed in the environment. Since water is the main medium for the collection and transportation of PFAS [5], PFAS have gradually become a widespread pollutant in various potential drinking water sources. PFAS have been frequently detected in potential drinking water sources around the world, even in lakes at the headwaters of rivers [3,6]. The presence of PFAS in potential drinking water sources has not only led to the enrichment of PFAS in aquatic plant and animal organisms, but has also promoted the transport of PFAS to aquatic plants and animals for consumption, as well as to terrestrial plants and animals that consume water resources [7]. Drinking water is a key route of human exposure to the risk of PFAS. Water, on the one hand, enters directly into daily life as drinking water for human use and consumption. Currently, most drinking water plants still use traditional water treatment processes to treat drinking water; however, traditional processes are not effective in removing PFAS, and, after biological treatment, the carbon-carbon (C-C) bond of long-chain PFAS in water may be broken, resulting in an increase in the total concentration of PFAS [8]. Compared to long-chain PFAS, short-chain PFAS have been less studied, and their biotoxicity and health risks need to be evaluated by further in-depth studies. On the other hand, water can also be enriched by plants and animals through agricultural irrigation, fishery farming, etc., and thus enter the human body indirectly as food [9]. When PFAS-contaminated water is used to irrigate farmland, PFAS will migrate from the water to plants. It has been shown that long-chain PFAS are more likely to accumulate in wheat, oats, potatoes, and maize [10], while short-chain PFAS are more abundant in fruits and leaves [11], and the bioconcentration factor of aquatic animals such as fish and shellfish increases with increases in the chain length of PFAS [12], while in aquatic plants such as algae, long-chain PFAS are concentrated in the roots and short-chain PFAS are enriched in the main body parts [13]. These studies have shown that the presence of PFAS in water sources is directly related to the accumulation of PFAS in human bodies and the level of risk.

PFAS in potential drinking water sources enter the human body through direct or indirect routes, leading to their detection in various organs and body fluids [14]. For example, the presence of PFAS has been found in blood, urine, and breast milk. The accumulation of PFAS in the body increases the risk of cardiotoxicity, hypertension, and vascular disease [15]. Due to their protein affinity, PFAS are able to break through the skin-mucosal barrier, blood-brain barrier, and placental barrier to reach the relevant organs and tissues or accumulate in the brain, leading to neurotoxicity [16,17]. For example, long-chain PFAS may penetrate the brain by breaking the blood-brain barrier or bind to specific transporter proteins and pass through the plasma membrane to enter nerve cells, where they can have toxic effects on the central nervous system [18]. In pregnant women, PFAS can be passed to the fetus through the placenta by pathways such as passive diffusion, leading to in utero exposure [19]. In addition, PFAS is biotoxic, with known toxic effects on the liver, kidneys, nervous system, reproductive system, and cardiovascular system, as well as significant effects on the immune system, metabolic abnormalities, and an increased risk of cancer [15,20,21]. Numerous studies have confirmed the adverse effects of PFAS on the human body. Therefore, by analyzing the latest PFAS presence levels and their distribution, we can provide information to support the assessment of exposure risks and the development of effective PFAS control and management strategies.

PFAS exhibit exceptional environmental persistence due to their unique chemical structure, characterized by the extensive substitution of hydrogen atoms with fluorine atoms along the carbon chain. The carbon-fluorine (C-F) bond is one of the strongest covalent bonds in nature, possessing high bond energy and remarkable stability, which renders PFAS as highly resistant to conventional degradation processes such as hydrolysis, photolysis, biodegradation, and chemical oxidation under ambient environmental conditions [22,23]. Consequently, once released into the environment, PFAS can persist for extended periods—ranging from several years to decades—in soil, sediment, surface water, and groundwater systems [24,25]. Moreover, many PFAS compounds are highly water-soluble and mobile; in particular, short-chain PFAS exhibit a greater propensity to migrate via rainfall runoff or groundwater flow, leading to widespread contamination [26]. Their bioaccumulative potential and capacity for long-range environmental transport have been extensively documented, with PFAS residues detected even in remote regions such as polar ice caps and marine organisms in isolated ecosystems [27]. This “forever chemical” behavior, combined with the limited efficacy of conventional water treatment technologies in removing PFAS, significantly exacerbates human exposure risks through contaminated drinking water sources.

With the increasing attention on PFAS, there has been a notable rise in review articles addressing the occurrence, distribution, and exposure of these substances. Most of these studies focus on the occurrence and distribution of specific PFAS in localized regions [28], the presence of PFAS in marine and sewage non-drinking water sources [5,29,30], analyses of PFAS treatment technologies [31,32,33,34,35], summaries of detection methods [36], or are limited to investigations concerning PFOA and PFOS [2,4]. However, a systematic review examining the occurrence and distribution of multiple PFAS in potential drinking water sources remains lacking. In light of this gap, this study aims to (1) perform a comprehensive statistical analysis of the levels and distribution of PFAS in potential drinking water sources globally over the past decade to identify the PFAS species that were detected at high frequencies or that are highly contaminated; (2) analyze the year-to-year trends in the number of detections of the top 30 PFAS with the highest frequencies of detection; and (3) assess the risk of exposure to eight specific PFAS based on the PFAS toxicity criteria, identifying high-risk categories.

2. Methods

2.1. Sources of Information and Methodology

To assess the actual exposure risk of PFAS in potential drinking water sources and provide data support for PFAS management policies, this systematic review was conducted via a structured search of the Web of Science Core Collection database (covering January 2000–December 2024) using the Boolean logic combination of the following keywords: (“water” OR “river” OR “lake” OR “reservoir” OR “drinking water”) AND (“PFAS” OR “PFASs” OR “PFAA” OR “perfluoroalkyl substances” OR “polyfluoroalkyl substances”). The literature screening followed strict inclusion and exclusion criteria: included studies were peer-reviewed original research (cross-sectional, cohort, or surveillance) reporting quantified PFAS concentrations in potential drinking water sources (rivers, lakes, reservoirs, etc.) with complete sampling location, time, and detection method information; excluded studies were those focusing on non-drinking water sources (e.g., seawater, high-salinity brine), reviews, conference abstracts, gray literature, studies lacking key data, or laboratory simulations. The screening process involved three steps: automatic duplicate removal using EndNote X9 (Version J17.3.0.10362), title/abstract screening by two independent researchers, and full-text assessment (disagreements resolved via discussion with a third senior researcher). More importantly, the data were extracted independently by two researchers using a standardized form (covering basic study details, sampling information, PFAS concentrations, and analytical methods) with cross-validation to minimize bias. This study focused on analyzing the occurrence and potential risks of PFASs in water bodies viable for drinking water use.

2.2. Data Treatment

Harmonize units and abbreviations for acquired data (Table S1). When calculating the overall statistical sum of PFAS concentrations, the median concentration provided in the report was used in preference. If the median concentration was not available, the mean concentration was used as a proxy; if there were still no optional data, half of the maximum concentration was used for estimation. For substances that were not detected (ND) or had concentrations below the limit of quantification (LOQ), their concentrations were considered to be zero.

In addition, this study was conducted with due consideration of the possible uncertainties associated with the research methodology, analytical techniques, risk assessment methods, and assessment criteria. For the ease of reference of future researchers, the data used in this study have been collated and presented in detail in Supplementary Table S2.

2.3. Risk Assessment

In order to more objectively assess the health risks associated with human exposure to PFAS-containing water, we conducted a comparative and conversion analysis (Text S1) of the risk quotient (RQ), hazard ratio (HR), and hazard quotient (HQ), which are commonly used in current risk assessment studies, and chose the RQ as the main indicator for the risk assessment of PFAS contamination in the present study [37,38,39]. According to the following equation, based on the measured concentration (C) and drinking water equivalent level (DWEL), the calculation formula is as follows:

$$\mathrm{DWEL}={\displaystyle \frac{\mathrm{A}\mathrm{D}\mathrm{I}\times \mathrm{B}\mathrm{W}}{\mathrm{D}\mathrm{W}\mathrm{I}\times \mathrm{A}\mathrm{B}\times \mathrm{F}\mathrm{O}\mathrm{E}}}$$

$$\mathrm{RQ}={\displaystyle \frac{\mathrm{C}}{\mathrm{D}\mathrm{W}\mathrm{E}\mathrm{L}}}$$

where DWEL: Drinking water equivalent level; ADI: Acceptable daily intake (ng/kg/day); BW: Body weight (kg); HQ: Hazard quotient (HQ = 1); DWI: Drinking water intake (L/day); AB: Gastrointestinal absorption rate (AB = 1); FOE: Frequency of exposure (FOE = 0.96, 350 days/365 days) [38]; and C: Median detected concentration (ng/L).

DWI and BW are taken from the age-specific drinking water intake recommendations for the ATSDR standard age groups (

Table 1. Recommended water intake rates specific to age for ATSDR’s standard age groups.

Age Range (Years)	Drinking Water Intake (L/day)	Body Weight (kg)
0–1	0.595	7.8
1–2	0.245	11.4
2–6	0.337	17.4
6–11	0.455	31.8
11–16	0.562	56.8
16–21	0.722	71.6
Adult (21–78)	1.313	80.0

). In this study, the Reference Dose for Oral Exposure (RfD) proposed by the U.S. Environmental Protection Agency (EPA), based on animal toxicity data, was used as the tolerable daily intake (ADI) (Table S3). If RQ < 0.2, the risk posed is assumed to be negligible; whereas when RQ ≥ 1, there may be adverse effects on human health. For 0.2 < RQ ≤ 1, further research and assessment of risk is required on a case-by-case basis.

Many studies have used a relative source contribution factor (RSC) to determine the percentage of exposure due to drinking water consumption. The RSC factor typically used in PFAS drinking water guidelines does not exceed 20%. Given the presence of areas of higher contamination in the potential drinking water data, the higher RSC value (20%) was chosen for this study, i.e., it was assumed that 20% of daily water intake comes from drinking water and the remaining 80% from other sources [39,40]. The equation for calculating the RQ_RSC is as follows:

$${\mathrm{RQ}}_{\mathrm{RSC}}={\displaystyle \frac{\mathrm{C}}{0.2\mathrm{D}\mathrm{W}\mathrm{E}\mathrm{L}}}$$

3. Results and Discussion

3.1. Concentration and Distribution Characteristics of PFAS in Potential Drinking Water Sources Worldwide

Concentrations of PFAS in potential drinking water sources: This study reviewed 8691 samples from 388 potential drinking water sources and 50 treated water bodies (including tap water from water treatment plants, bottled water, and tap water) from 161 articles included in Web of Science over the past decade (Published after 2014) (Table S2). According to the sample distribution, studies in China (2545 samples) and the United States (2379 samples) are the most comprehensive. Among the investigated bodies, 197 types of PFAS were detected, including PFOA, PFOS, PFHxA, PFBS, PFHpA, PFHxS, PFPeA, PFNA, PFBA, PFDA, PFUnA, PFDoA, 6:2 FTS, PFHpS, GenX, PFOSA, PFPeS, PFTeDA, PFTrDA, 6:2 Cl-PFESA, PFDS, EtFOSAA, MeFOSAA, 8:2 FTS, 4:2 FTS, PFHxDA, PFOcDA, PFEtCHxS, and Adona, as well as HFPO-TA, which were widely detected, accounting for about 88.5% of all PFAS measured in water. Meanwhile, 24 of these 30 PFAS meet the detectable substances in the U.S. drinking water testing standard (which covers a total of 29 PFAS). This also indicates that these PFAS are importantly represented in academic research and practical management. Therefore, the following discussion in this study will focus primarily on these substances and their sum (ΣPFSA).

Figure 1a clearly shows the median concentration of ΣPFSA at a total of 438 detection sites in 34 regions, while Figure 1b presents the ranking of the frequency of detection of 30 PFAS species in each region. In terms of the global distribution of monitoring sites, China and the United States had a significant dominance in the number of PFAS monitoring sites, accounting for 44.7% and 19.7% of the total, respectively. They were closely followed by Korea (3.7% of the total), Vietnam (3.0%), and India (1.7%). Further analysis of the ΣPFSA median concentration data revealed that monitoring sites with concentrations between 20 and 50 ng/L were the most common, accounting for 19.0% of all monitoring sites. Meanwhile, about 80.0% of the monitoring sites had median ΣPFSA concentrations below 100 ng/L. On this basis, a comparison of the distribution of concentrations at monitoring sites in China and the United States shows that when PFAS concentrations are at low levels (<50 ng/L), the proportions of monitoring sites in China and the United States are nearly the same. However, once ΣPFSA concentrations exceeded 50 ng/L, the proportion of monitoring sites in China was significantly higher than in the United States. While this difference may suggest more severe PFAS contamination in some regions of China, a critical caveat is that direct cross-country comparison must account for the following confounding factors: First, China’s monitoring network has expanded rapidly only in the past 5 years, with more sites targeting industrialized or densely populated areas (high-risk hotspots), whereas U.S. monitoring covers a more balanced mix of urban, rural, and remote regions. Second, differences in detection methods (e.g., lower limits of detection in U.S. studies using LC-MS/MS with isotope dilution) may lead to underreporting low concentrations of PFAS in some Chinese studies. Third, the implementation timeline of PFAS restrictions varies: the U.S. banned PFOA/PFOS production in 2002, while China’s restrictions on perfluorinated compounds were not fully enforced until 2021, resulting in a lag in contamination reduction. These factors highlight that concentration disparities alone cannot fully reflect the effectiveness of pollution control, and a more nuanced assessment requires integrating policy timelines, sampling strategies, and analytical methodologies.

Although PFOS and PFOA have been banned in many countries, it is worrying that they are still widely present in potential drinking water sources and can be detected at almost all monitoring sites. A critical analysis of this phenomenon reveals two key issues: first, the environmental persistence of PFAS (half-lives of decades in soil and groundwater) means that historical emissions continue to leach into water sources, even after production bans. Second, non-point source pollution (e.g., atmospheric deposition, wastewater from textile/plating industries, and leachate from landfills) remains poorly regulated in most countries, becoming a persistent source of new contamination. Additionally, some countries have replaced PFOS/PFOA with short-chain alternatives (e.g., PFBS, PFHxA), but the lack of long-term toxicity data for these alternatives raises concerns about “regulatory displacement”—where banning one class of PFAS simply shifts contamination to another unregulated class with unknown risks.

In the monitoring of the new PFOS or its alternatives, water bodies in China, the United States, Germany, Korea, France, and Canada showed high detection rates. This phenomenon on the one hand fully reflects the wide distribution of PFAS in water bodies in these countries, and, on the other hand, it also proves that these countries have reached a high level of PFAS monitoring technology. It is noteworthy that these countries have shown positive developments in the regulation of PFAS and the development of alternatives. These positive actions do not exist in isolation, but rather interact and complement each other to promote the continuous development and wide application of PFAS monitoring technology. Based on the above findings, in order to provide a more in-depth and comprehensive analysis of the contamination status of potential drinking water sources in each region, the following detailed analysis and discussion will be conducted for each region in turn based on the order of ΣPFAS concentration from lowest to highest. At the same time, an overview of the types of PFAS with high concentrations in potential drinking water sources in each region was analyzed in order to accurately understand the characteristics of PFAS contamination in different regions.

3.1.1. Low and Medium Concentrations of ΣPFAS

In 13.24% of potential drinking water sources, the median concentration range of ΣPFAS is from 100 to 500 ng/L. These areas are often designated as potential drinking water sources in regions that are economically disadvantaged or lack robust regulatory frameworks. Furthermore, elevated PFAS levels have been identified in potential drinking water sources across multiple countries. PFAS has been detected in multiple water sources in Australia. Tanya Paige et al. detected 33 and 21 types of PFAS in the rivers of Greater Melbourne [41] and Victoria Melbourne [42], respectively, with a range of total PFAS of 0–526 and 0.58–658.9 ng/L. PFOS was detected at high levels in both locations, with a sudden 6:2 FTS contamination event reaching a maximum of 145 ng/L in Greater Melbourne and PFAS contaminants mainly consisting of PFHxA (186 ng/L maximum) and PFHxS (193 ng/L maximum) in Victoria, Melbourne. The PFAS contamination situation is different in the two locations, with Greater Melbourne detecting a greater variety of PFAS with an average concentration of 0–15 ng/L for individual PFAS, while Victoria, Melbourne, has a more pronounced PFAS contamination with individual PFAS detected at a maximum concentration of 66.8 ng/L, which requires pollutant source tracing for control. In addition to rivers, PFAS are also present in lakes in Australia, such as Albert Park Lake, Swan Canning, and Neusiedl Lake [43,44]. Albert Park Lake and Swan Canning have high levels of PFAS contamination nearby, with average concentrations of Σ₄₆PFAS = 252 ng/L and Σ₁₀PFAS = 329 ng/L, respectively. There is a significant localized pollution of PFDA and PFOS in Albert Park Lake, with maximum concentrations of 156 and 151 ng/L, respectively, while PFOA (75.1 ng/L average) and PFHxS (156.0 ng/L average) are commonly found in Swan Canning.

Hartbeespoort Dam and Roodeplaat Dams are important water resources in South Africa, with Roodeplaat Dams being the source of drinking water for three provinces [45]. However, very high concentrations of Σ₉PFAS were detected in the water bodies of both dams, ranging from 159.38 to 909.86 ng/L for Hartbeespoort Dam and 169.5–614.21 ng/L for Roodeplaat Dams, and the concentrations of Σ₉PFAS on the banks of the dams were higher than in the center. This may be due to the lack of public water points near the river, which has led to a high population living near the river and the disposal of large amounts of waste on the banks of the dams. Similarly, higher concentrations of PFAS were detected in the Vaal River in South Africa, with Σ₈PFAS being approximately 142.2 ng/L (the maximum value of individual PFAS was accumulated to obtain this value) [46]. The detection of high concentrations of PFAS in these important water sources in South Africa makes policy regulation even more important.

Important rivers in Spain include the Guadalquivir River (northeast), the Ebro River (south), the Jucar River (southeast), and the Llobregat River (northeast), covering most of the water systems in the eastern region of Spain. The Guadalquivir River flows through the region of Andalusia and is the main source of water for the area. María Lorenzo and others detected Σ₉PFAS up to 1575.2 ng/L (average value = 260.9 ng/L) in this river, with PFBA contributing the most (maximum value: 742.9 ng/L), which may be caused by sewage treatment plant discharges from Cordoba and Ecija, industrial activities, or military activities [47]. For the Ebro river, María Lorenzo and others also conducted investigations, finding Σ₁₀PFAS maximum value of 496.0 ng/L (average value = 51.9 ng/L), with PFBA still detectable at high concentrations (maximum value = 251.3 ng/L; average value = 35.2 ng/L) [47]. Five years later, Emanuela Pignotti also studied the Ebro River, finding that in surface water in the Ebro Delta area, except for PFPeA (maximum value = 2775 ng/L; average value = 1.2 ng/L), the concentrations of all PFAS were below 9 ng/L [48]. The high concentration of PFPeA may be because it is widely used as a substitute for PFOA and PFOS, so it was detected in large quantities in river water near sewage treatment plants. The Jucar River flows through three provinces and provides water for a million people. The Jucar River flows through three provinces and provides water to millions of people. After testing the Jucar River, Julian Campo and others found that the tributaries of the Cabriel River and Magro River, ΣPFAS were all below 50 ng/L [49]. For the entire river basin, the range of Σ₁₂PFAS was 21.1–1140 ng/L (average value = 240 ng/L), PFBA, PFDA, and PFOS were detected at high concentrations (maximum values were 644, 213, and 128 ng/L), which was directly related to the industrial and urban emissions in Cuenca city. The Llobregat River is one of the main drinking water sources for Barcelona city. The maximum value of Σ₁₃PFAS in the Llobregat River was 3130 ng/L (average value = 450 ng/L), and the Anoia River, a tributary of the Llobregat River, was the most polluted (PFOS = 2710 ng/L), which may be related to the discharge of industrial waste from tanneries and textile factories [50]. According to the analysis by Julian Campo and others, this pollution does not extend downstream along the Llobregat River. Similarly, water bodies in Norway are also in need of attention regarding water pollution. For example, the PFAS concentration range detected in the rivers of Ny-Ålesund and Longyearbyen are 110–1156 ng/L and 112.8–118.6 ng/L, respectively [51], while the ΣPFAS in other surface water and Lake Linnévatne t in Ny-Alesund area are all less than 10 ng/L [51,52].

In various surface water sources in China and the United States, the median concentration of ΣPFAS detected ranges from 100 to 500 ng/L. In the United States, the Banana River in Florida, one of the state’s major water sources, has a PFAS concentration of 110.6 ng/L [53]. The Alabama River in Alabama, which is crucial for agricultural and industrial water use, has a PFAS concentration of 100.0 ng/L [54]. The Coosa River in the southeastern United States, an important ecological and recreational resource, has a PFAS concentration of 191.0 ng/L [54]. The Cape Fear River in North Carolina, one of the state’s major water sources, has a PFAS concentration ranging from 106.1 to 212 ng/L [55,56]. The Las Vegas Wash in the Las Vegas area, a vital water source, has a PFAS concentration ranging from 102.0 to 178.7 ng/L [57,58]. In China, the Mi River in Henan Province, an important source of water for agricultural irrigation, has a PFAS concentration of 252.4 ng/L [59]. The Moshui River in Yunnan Province, a significant local water resource, has PFAS concentrations of 149.7 ng/L and 178.1 ng/L at two different locations [59,60]. The Yangtze River, China’s longest river and of significant importance to economic development and the ecological environment, has a PFAS concentration of 142.5 ng/L [61]. The Jiyu River in Shandong Province, a significant source of water for industrial and agricultural use, has a PFAS concentration of 262.2 ng/L [62]. The Zhimai River in Zhejiang Province, an important ecological and recreational resource, has a PFAS concentration of 302.2 ng/L [63]. In Longgang District, Shenzhen, Guangdong Province, a source of water for industrial and domestic use, the PFAS concentration is 474.3 ng/L [64]. The Maozhou River in Guangxi has a PFAS concentration of 132.2 ng/L [65]. The Qiantang River in Zhejiang Province, an important source of water for agricultural irrigation, has a PFAS concentration ranging from 144 to 367.53 ng/L [66]. In terms of lake water sources, West Lake in Hangzhou, a famous tourist attraction and significant ecological resource, has a PFAS concentration of 122.4 ng/L [67]. Taihu Lake in Jiangsu Province, the third largest freshwater lake in China and of significant importance to local ecology and economy, has a PFAS concentration ranging from 130.5 to 288.93 ng/L [68]. Additionally, the surface water around the areas of Taihu, Nanjing, Shanghai, Jiangsu, and Suzhou has PFAS concentrations of 109.7 ng/L, 102.0 ng/L, 103.5 ng/L, 114.5 ng/L, and 344.5 ng/L, respectively [59,69,70,71]. The PFAS content in surface water sources does not only affect the PFAS content in groundwater and tap water; for example, tap water in North Carolina, USA, has a PFAS concentration of 229 ng/L [72] while groundwater in Cape Cod, MA, has a PFAS concentration of 225.2 ng/L [73]. In China, the groundwater of the Maozhou River has a PFAS concentration of 119.5 ng/L, while groundwater in Nanjing is 104.7 ng/L; tap water in Nanjing, Changshu, and the Taihu Lake area has PFAS concentrations of 188 ng/L, 122.4 ng/L, and 112.2 ng/L, respectively [67,71,74].

3.1.2. High Concentrations of ΣPFAS

Approximately 5.9% of ΣPFAS samples have median concentrations exceeding 500 ng/L. The Pampulha Lake in Brazil, once a drinking water source, is now heavily polluted due to industrial wastewater, with Σ₆PFAS ranging from 5716 to 259,885.2 ng/L [75]. Similarly, firefighting training in Ronneby Sweden has contaminated groundwater with PFAS, primarily PFOS and PFHxS, at concentrations up to 20,000 ng/L [76]. A study by Isabelle J. Neuwald across 13 German drinking water sources found high PFAS concentrations (median: 918.1–1566.8 ng/L), with TFA (median: 900 ng/L) being the most prevalent. Ultra-short-chain PFAS like TFMS and PFPrA were also identified, highlighting the need for more research on these compounds [77]. The Danube River Basin, spanning multiple countries, shows a decrease in PFAS detection from upstream to downstream, with high concentrations in both rivers (5201–81,744 ng/L) and groundwater (5330–233,130 ng/L) [78]. High concentrations of PFAS were detected in the mouths of the Kishon River, Alexander River, and Lachish River (median: 5648.1 ng/L, 283.2 ng/L, 5838.4 ng/L) in Israel, where the rivers’ sources are located upstream and treated wastewater flows downstream, resulting in high PFAS concentrations in the river mouths [79]. Bangalore, India, with its three main valleys, has also seen significant PFAS contamination in water bodies used for irrigation and drinking, with the highest levels found in the Vrishabhavathi River (max: 2315 ng/L) [80]. This global issue calls for more systematic monitoring and regulation to address PFAS pollution in water sources.

China and the United States are the two countries with the highest level of concern for PFAS in water bodies, with detection numbers accounting for 56.7% of the total. In drinking water screening conducted in 16 states in the United States, the range of ΣPFAS in private water samples was 2.3–7135 ng/L, and the range in public water samples was 1.9–234.5 ng/L [81]. Among them, the frequency of detection of ultra-short chain PFPrA in water samples was the highest, and the detection concentrations of PFOA (2.0–2100 ng/L), PFOS (1.9–1700 ng/L), and PFHpA (2.1–1100 ng/L) were the highest. At the same time, heavily polluted water bodies were detected in the United States, including the Lemmon Valley area near a fire training ground (ΣPFAS ranged from 20 to 4754 ng/L), while the concentrations in urban and remote areas were lower, at only 2–15 ng/L [82]. Areas in China with severe PFAS pollution are mainly concentrated in rivers, including the Majia River, the Zhimai River, the Jiaolai River [59], the Xiaoqing River [59,61,63], the Xi River [61], the Daling River [83,84], the Liuxi River [85], and rivers flowing into the Bohai Sea [86]. These rivers are located downstream and in economically developed areas, so they not only bear the pressure of upstream discharge, but also receive PFAS wastewater from economic zones, which inevitably affects the marine fishery and ecology along the coast. The maximum concentration of ΣPFAS detected in these rivers is over 1000 ng/L, with the most polluted being the Xiaoqing River, which flows through five Chinese cities and ultimately flows into the Bohai Sea (2018 maximum concentration of 415,488.0 ng/L, median of 9547.7 ng/L) [61]. With China implementing restrictions on the production and use of some PFAS, the maximum value detected in 2020 was reduced to 51,333.7 ng/L [59], while the maximum value detected in 2021 was 4848.1 ng/L [63]. Although the detected concentration of ΣPFAS has significantly decreased, they still pose a potential threat to humans and aquatic ecosystems.

3.1.3. High Presence Level of PFAS

By synthesizing the mean median concentrations of different types of PFAS in 34 regions, the results of this study show that the contamination of certain PFAS species was particularly significant in China, Sweden, India, Israel, and the Danube River Basin (Figure 2a). High-concentration PFAS species include PFOA, PFOS, PFHxA, PFHxS, PFBS, PFDA, 6:2 FTS, and GenX, with average median concentrations exceeding 100 ng/L. When examining the median concentration of these substances in the largest areas (Figure S1), the areas affected by single-PFAS pollution have expanded to the United States, Australia, Spain, Brazil, and Norway. In addition, the types of PFAS pollutants have increased to include PFHpA, PFNA, PFPeA, PFBA, and PFDoA. By comparing the average median and maximum median, China, the United States, and Sweden show a serious regional PFAS pollution situation. Reasonable regional control measures are expected to alleviate this problem to a large extent. However, the Danube River Basin and Brazil not only face significant regional pollution, but also have very prominent overall PFAS and single-PFAS pollution, which goes beyond the scope that can be solved by relying solely on regional control measures. Therefore, it is urgently needed to implement more comprehensive management strategies and mandatory control measures to effectively deal with PFAS pollution.

To ensure drinking water safety, upstream rivers and less polluted lakes and reservoirs are generally selected as sources. To comprehensively understand actual potential drinking water sources, we conducted a statistical analysis that excluded obviously or highly polluted water bodies, which are unlikely to serve as major drinking sources. To minimize the impact of significant or sudden pollution events on our analysis, we used the median average for each region, as opposed to the mean, which can be greatly influenced by the number of samples. This analysis focused on the remaining areas across different countries, with median values detected by each country averaged to create Figure 2b. Germany, India, and South Africa showed relatively high average median PFAS levels at 532.27 ng/L, 293.49 ng/L, and 268.19 ng/L, respectively. This data offers a preliminary insight into the PFAS content in various countries’ water bodies. The maximum median values are displayed in Figure S2. Notably, China, Germany, and India have significantly higher PFAS content in their water bodies, with levels at 1736.7 ng/L, 1566.85 ng/L, and 1157.5 ng/L, respectively. This indicates that China’s PFAS point source pollution problem is relatively serious, so subsequent monitoring and management should pay more attention to reducing emissions. In order to more effectively inform regulation and reduce potential PFAS contamination in potential drinking water sources, it is necessary to further analyze the focus of PFAS research and examine the spatiotemporal distribution characteristics of its varieties.

3.2. Monitor the Time-Series Changes in PFAS Species

Countries around the world have issued bans on the use and import of PFAS, such as China’s ‘Key List of New Pollutants under Strict Control (2023)’, the United States’ Toxic Substances Control Act (TSCA), the European Union’s Regulation (EU) 2024/573, and Japan’s Chemical Substance Control Law (CSCL). However, given that these policies are still being refined or have not yet taken full effect, and considering the persistence of PFAS, the levels of PFAS in the world’s potential drinking water sources may not have reached their peak. Therefore, research on the spatiotemporal distribution characteristics of PFAS remains highly necessary.

From the reported data from 2014 to September 2024, the time span of water sampling can be traced back to 2010. As shown in Figure 3a, the detection types and frequencies of PFAS have shown an overall upward trend. Although the detection volume has decreased in 2022–2024, this is mainly due to the limitation of the research period and the fact that some studies have not been published. The attention paid to PFAS has increased annually, from 15 PFAS detected 100 times in 2010 to 92 PFAS detected 435 times in 2019, to 74 PFAS detected 1191 times in 2021. The focus of monitoring has also shifted from traditional PFAS such as PFOS and PFOA to emerging areas such as PFAS alternatives, PFAS precursors, and ultra-short-chain PFAS. Figure 3b shows the number of 30 PFAS monitored in different years, showing that, as time goes on, the understanding of PFAS types is becoming more comprehensive. As a substitute for PFOA, GenX was detected in the Cape Fear River in North Carolina in 2013 [55], with a median concentration of 304 ng/L, and has since become a focus of research, with the number of detections increasing from 1 in 2013 to 26 in 2021. After the U.S. EPA set a concentration limit of 10 ng/L for GenX in drinking water in 2024, GenX is expected to become a hot topic in future monitoring work. PFAS precursors refer to substances that can be converted into PFAS under certain conditions. There are many challenges in identifying PFAS precursors with current monitoring techniques. Currently, the content of PFAS precursors is usually estimated by measuring the total organic fluorine (TOF), PFAS concentration, and PFAS concentration after the water body is oxidized. The accurate identification of PFAS precursors not only requires precise instruments, but also rigorous analytical methods. These difficulties have led to relatively fewer related studies. If PFAS precursors can be accurately identified and their conversion to PFAS can be blocked, it will be one of the effective ways to reduce PFAS in water.

Ultra-short-chain PFAS typically refers to PFAS with a carbon chain length of 2–3 carbon atoms. They possess an extremely strong ability to migrate across different media [87]. This characteristic makes it more likely for ultra-short-chain PFAS to be present in high concentrations in groundwater, as PFAS with longer carbon chains are more likely to be intercepted and adsorbed by soil or other geological structures, preventing them from infiltrating into groundwater. Due to the fact that PFAS alternatives often use short-chain PFAS, this may accelerate the presence level of ultra-short-chain PFAS in the water environment. Short-chain PFAS are designed to have lower persistence in the body than long-chain PFAS, and theoretically have lower toxicity. However, studies have shown that even short-chain PFAS can be rapidly absorbed by the human body after exposure and strongly bind to serum albumin, thereby enhancing their intracellular distribution [88]. Furthermore, short-chain alternatives may have greater liver toxicity than long-chain PFAS, which further emphasizes the necessity of further exploring their risks to human health. Therefore, as the use of short-chain PFAS as alternatives increases, the presence level of ultra-short-chain PFAS in the water environment may also rise, requiring us to closely monitor and study it.

3.3. Global Risk Assessment of Potential Drinking Water Sources

The global risk distribution for PFAS was calculated by averaging and maximizing the median PFAS concentrations obtained from multiple sampling sites in each country, excluding heavily contaminated water bodies (see the statistical range shown in Figure 2b). Subsequently, these two concentrations were used to assess the RQ_RSC in the water body. Currently, standards for PFAS in drinking water are still being proposed and refined. Between 2020 and 2024, China, the United States, the European Union, Australia, and Canada have proposed or updated PFAS drinking water standards and recommended health standards (Table S5), with the most stringent and comprehensive being the U.S. EPA National Primary Drinking Water. The limit values are 10 times smaller than the corresponding PFOS recommendation in China (80 ng/L), 20 times smaller than the PFOA recommendation (40 ng/L), and 140 times smaller than the PFOA standard in Australia (560 ng/L), whereas the EU and Canadian recommendations are more stringent. In contrast, the EU and Canadian recommendations prefer to control the total PFOS content. In addition to the continuous updating of drinking water standards, with the in-depth study of PFAS biotoxicity, several countries or organizations have further proposed or revised daily reference concentrations that are more stringent than the drinking water standards based on toxicity data (Tables S3 and S4). Therefore, to truly reflect the risk of exposure due to PFAS toxicity, the U.S. EPA Reference Dose for Recent Oral Exposure (RfD), which is based on animal toxicity analyses, was chosen as the Allowable Daily Intake (ADI) for calculating the RQ_RSC in the present study.

The PFAS standard proposed by the EPA is nearly the most stringent in the world. Based on this standard, the risk assessment was performed using the PFAS median mean concentration values, as shown in Figure 4. The number of samples at greater risk (RQ_RSC > 1) from direct consumption of these potentially contaminated water sources accounted for 45.6% of the total. RQ_RSC < 0.2, which is generally considered to be a low risk, accounted for 48.4% of the total number of samples. The remaining 6% of the samples were subject to further risk exposure studies based on specific use scenarios. PFOA levels showed significant risk in almost all regions, especially in South Africa, Australia, India, and Norway, where RQ_RSC values exceeded 20. For PFOS, half of the regions were at higher risk (RQ_RSC > 1) and the most contaminated areas overlapped with the PFOA distribution. In contrast, the risk value of PFDA is much higher than that of PFOA and PFOS, with the RQ_RSC exceeding 1 in all regions, and for infants aged 0–1 year, the RQ_RSC can be as high as 1556.2. This phenomenon stems from the fact that, on the one hand, the high biotoxicity of PFDA has led to the setting of a low RfD value, which raises the standard, and, on the other hand, since the younger the population, the more tolerant it is to the toxicity of PFAS (and therefore faces a greater risk of contamination), the more severe the contamination, the greater the risk. Capacity is lower; they are therefore exposed to greater health risks. With regard to GenX, according to the available data, the risk of exposure in the United States is relatively high, but manageable overall. This situation is due to the fact that most countries have not yet mastered GenX monitoring technology, which may result in large quantities of the substance being present in the environment but not being detected, thus biasing the data. For example, in a recent study, GenX levels of 447,000 ng/L were detected at the outfall of a wastewater treatment plant near Lake Taihu in China, with an increasing trend over the years [89].

Among the potential drinking water bodies, PFHxA (RQ_RSC < 0.014), PFBS (RQ_RSC < 0.25), and PFBA (RQ_RSC < 0.05) were basically at a low-risk level, as shown in Figure S3. PFPrA, similar to GenX, has a small monitoring sample size and therefore there is uncertainty about future risk trends. However, based on the data monitored to date, the overall risk can be considered negligible (RQ_RSC < 0.03). Risk was assessed using the maximum median PFAS concentration, as shown in Figure S4. For these eight evaluated PFAS, the overall risk increased with increasing PFAS concentration. The RQ_RSC for PFHxA, PFBA, and PFPrA were all less than 0.2, whereas the risk for PFBS was slightly higher only in the Indian region (0.2 < RQ_RSC < 1). For PFOS, PFOA, PFDA, and GenX, which originally had higher mean risks, their calculated risks at the maximum median concentration were even more significantly elevated. China became the region with the highest risk for PFOA, followed by India and South Africa; for PFOS, with 86.7% of regions having an RQ_RSC > 1, China also ranked among the top three high-risk areas. In all regions, PFDA was at high risk (RQ_RSC > 10) and the number of very high risk regions was large enough to warrant special attention. In this data set of the largest median PFAS values, in addition to the United States, China also clearly reflects a higher risk in terms of GenX exposure. Considering the large industrial scale of China, it is expected that as the amount of GenX monitoring data increases, the high risk level will become more apparent.

The RfD is the indicator that best reflects the toxicity of PFAS, and national criteria, based on a variety of factors, are gradually converging towards it. However, according to this criterion, nearly half of the world’s water sources are potentially at risk. This not only indicates that the standard is too high for many regions and is difficult to use as a basis for uniform restrictions, but also highlights the importance of strengthening the protection and monitoring of potential drinking water sources. In addition, the difficulty of removing PFAS contamination from water with conventional drinking water treatment technologies limits the ability to reduce the risk of PFAS exposure.

3.4. The Health Impacts of PFAS

PFAS are highly persistent synthetic organic pollutants that resist environmental degradation, leading to long-term accumulation in soil, water bodies, and living organisms. Human exposure primarily occurs through contaminated drinking water, posing significant threats to public health. Epidemiological studies have consistently demonstrated that chronic exposure to PFAS-contaminated drinking water is associated with a range of adverse health outcomes [90]. For instance, the EPA and numerous cohort studies have reported significant associations between elevated serum PFAS concentrations and hepatotoxicity, immunosuppression, thyroid hormone disruption, hypercholesterolemia, pregnancy-induced hypertension, and increased risks of certain cancers—particularly renal and testicular carcinomas [90,91,92]. Of particular concern is the heightened vulnerability of children and pregnant women; prenatal exposure to PFAS has been linked to impaired neurodevelopment, reduced birth weight, and long-term metabolic dysregulation in offspring.

At the molecular and cellular levels, PFAS exert toxicity largely by disrupting the endogenous antioxidant defense system, thereby inducing oxidative stress. Under physiological conditions, cells maintain redox homeostasis through enzymatic antioxidants such as superoxide dismutase (SOD), glutathione peroxidase (GPx), and catalase (CAT), which effectively scavenge reactive oxygen species (ROS) [93]. However, experimental evidence confirms that common PFAS compounds—including PFOA and PFOS—significantly suppress the activity of these critical antioxidant enzymes, resulting in excessive ROS accumulation [93,94]. Elevated ROS levels cause oxidative damage to lipids, proteins, and DNA, and further activate pro-inflammatory signaling pathways such as nuclear factor-kappa B (NF-κB) and mitogen-activated protein kinase (MAPK), amplifying inflammatory responses and tissue injury. Moreover, PFAS impair mitochondrial function, reducing ATP production efficiency and further compromising the cell’s capacity to cope with oxidative stress [95].

More critically, PFAS-induced oxidative stress can trigger multiple forms of regulated cell death. Under conditions of low-dose chronic exposure, cells predominantly undergo apoptosis—a programmed cell death pathway characterized by caspase cascade activation, DNA fragmentation, and cellular shrinkage. In contrast, acute high-dose exposure may induce alternative death modalities such as necroptosis or ferroptosis [96]. Ferroptosis, an iron-dependent form of cell death driven by lipid peroxidation, has recently emerged as a key mechanism in PFAS toxicity. These non-physiological cell death processes not only compromise organ function (notably in the liver, kidneys, and lungs), but may also serve as pathological underpinnings for the development of chronic diseases [97].

Upon entering the human body via drinking water, PFAS not only directly interfere with endocrine and immune functions, but also profoundly disrupt cellular redox balance by inhibiting endogenous antioxidant enzymes and activating diverse cell death pathways, ultimately leading to multi-organ toxicity. Given their environmental persistence, bioaccumulative potential, and widespread human exposure, there is an urgent need to enhance the monitoring and remediation of PFAS in drinking water supplies and to deepen mechanistic investigations into their toxicological profiles. Such efforts are essential to inform the development of more stringent regulatory standards and evidence-based public health protection strategies.

3.5. The Challenge of Reducing the Risk of PFAS

The main challenges to reducing the risk of PFAS in potential drinking water sources include detection, identification, and treatment technologies, as well as appropriate regulatory systems.

The pre-treatment and detection process of PFAS in drinking water is relatively complex. Since the concentration of PFAS in water bodies is usually at the level of ng/L or μg/L, the samples generally need to be enriched as necessary before detection. Currently, the commonly used pre-treatment for PFAS is solid phase extraction (SPE) to achieve the enrichment of PFAS. The detection technique for PFAS is usually based on high-performance liquid chromatography (HPLC) instruments, with different types of instruments being selected according to the analytical focus. For example, for PFAS-targeted analysis, LC-MS/MS is usually used for detection and quantification of PFAS by known standards. For non-targeted analysis, on the other hand, full-spectrum scanning using high-resolution mass spectrometry (HRMS) is required, and novel PFAS are identified by comparing mass spectral signals. Since there is often a lack of relevant standards for non-targeting studies, techniques such as Kendrick Mass Defect analysis, homologous series detection, and fragmentation analysis are often used to identify PFSA structures, and the reliability of the identification results is assessed according to the confidence level (CL) [8,98]. The limited number of standards in targeted studies and the complexity of the procedures for identifying novel PFAS in non-targeted studies pose challenges for monitoring. In addition, the long detection and analysis cycle time, coupled with the inability of current technology to identify all types of PFAS comprehensively, leads to a lag in contaminant identification and control, which increases the risk of PFAS toxicity.

Drinking water treatment technologies include traditional and advanced processes. Traditional treatment processes usually use biological treatment, coagulation and sedimentation, and filtration and disinfection units, but these technologies are ineffective in removing PFAS. Meanwhile, the biological treatment and disinfection process may break the C-C bond, while it is difficult to effectively break the C-F bond, increasing the short-chain PFAS content in the water, and thus an increase in the total concentration of PFAS in the effluent [8]. This is one of the reasons why high concentrations of PFAS have been detected in drinking water in many locations. For example, the highest concentration of PFAS in drinking water in North Carolina was 458 ng/L [72], and a similar situation exists in the Nanjing [71], Changshu [67], and Taihu Lake [74] areas in China, where the highest concentrations were 271.0 ng/L, 155.4 ng/L, and 188.35 ng/L, respectively. Advanced treatment processes mainly include activated carbon adsorption (AC), advanced oxidation technology (AOP), and reverse osmosis (RO)-based membrane filtration technology, etc. At present, most of the research on PFAS by AOP focuses on laboratory simulation of water bodies, and although it can effectively decompose part of PFAS, it has not yet been practically applied in drinking water plants due to idealized research conditions, a single research object, and higher treatment costs and operational difficulties [99,100,101,102]. Activated carbon adsorption has been applied in general water plants, but, due to its lack of selectivity and poor adsorption effect on short-chain PFAS, short-chain PFAS are still left in the effluent water [103]. RO is the best choice for reducing the risk of PFAS in drinking water because of its excellent retention effect on low-molecular-weight pollutants, which can almost completely remove PFAS [104]. However, when treating drinking water on a large scale, the cost of the RO process is a major limiting factor.

Globally, policies for the management of PFAS are constantly being proposed and revised. Policy restrictions on the use of PFAS in industrial production are based on PFAS toxicity studies, emission standards are set based on PFAS contamination, and drinking water standards and health recommendations are set based on PFAS risk assessment. However, the development of policies has always lagged behind the occurrence of pollution events. Anticipating contamination is a fundamental strategy for controlling it at the source. Currently, many fields have begun to develop relevant Artificial Intelligence (AI) models, a trend that may provide new opportunities for predicting pollution events and regulating policy innovation. Despite the predictability benefits of AI, there are a number of challenges. First, there is a lack of comprehensive cross-sectoral statistics on the types and scale of raw materials used by industries, and second, there is insufficient data on the true and comprehensive levels of contamination, not only of PFAS but of all other pollutants. The most critical and complex part of this process is the migration and transformation of many substances before they are produced and exposed to the human body, as well as the risk of harm that may be posed once they enter the human body. This part involves multiple fields and is unimaginably complex. However, machine learning [105,106], model training [107,108], and large-scale AI modeling are already planting the seeds in various fields [109,110], so pollution foresight prediction may not be that far away.

4. Conclusions

PFAS have significantly increased concentrations in potential drinking water sources due to their widespread industrial applications and environmental release, placing an increasing number of people at risk for PFAS exposure. This study provides a comprehensive review of the literature on PFAS levels in potential drinking water sources over the past decade and a risk assessment of PFAS based on the RfD proposed by the U.S. EPA. The purpose of this study was to track the latest developments in global PFAS contamination and to identify the PFAS species with high detection frequencies, concentrations, or exposure risks to guide research on PFAS treatment technologies. While PFAS treatment technologies primarily target PFOA and PFOS, a variety of PFAS have been detected at high concentrations in potential drinking water sources globally, including, but not limited to, PFHxA, PFBS, PFHpA, PFHxS, PFPeA, PFNA, and GenX, which also require targeted treatment technologies. For example, high levels of PFBS in India, PFOS and PFHxS in Israel, and PFOS, PFHxS, and 6:2FTS contamination in Sweden require special attention. In addition, China, Brazil, Germany, South Africa, and the Danube River Basin are areas with high levels of total PFAS contamination, particularly dominated by PFOA, PFOS, PFHxS, and PFBS, with China having particularly high levels of contamination. The risk assessment of potential drinking water sources showed that 45.6% of the global region is at high risk (RQ_RSC > 1), while 48.4% of the region is at low risk (RQ_RSC < 0.2). Of these, PFOA, PFOS, PFDA, and GenX were associated with a higher risk of toxicity exposure. Young children are at the highest risk of PFAS exposure due to their physiologic characteristics. In order to reduce the health risks associated with high PFAS exposure, it is important to research targeted treatment technologies based on the current presence of PFAS and to take action in simplifying detection techniques, promoting large-scale safe treatment technologies, and reducing the lag in regulatory measures. In the future, we need more advanced technologies and interdisciplinary collaboration, such as the use of Artificial Intelligence models, to predict the level of PFAS presence and its migratory transformation and to provide timely feedback.