Considering Environmental Exposures to Per- and Polyfluoroalkyl Substances (PFAS) as Risk Factors for Hypertensive Disorders of Pregnancy

Abstract

Hypertensive disorders of pregnancy (HDP), including preeclampsia and gestational hypertension, lead to significant maternal morbidity and in some cases, maternal mortality. Environmental toxicants, especially those that disrupt normal placental and endothelial function, are emerging as potential risk factors for HDP. Per- and polyfluoroalkyl substances (PFAS) are a large group of ubiquitous chemicals found in consumer products, the environment, and increasingly in drinking water. PFAS have been associated with a multitude of adverse health effects, including dyslipidemia, hypertension, and more recently, HDP. In this review, we present epidemiological and mechanistic evidence for the link between PFAS and HDP and recommend next steps for research and prevention efforts. To date, epidemiological studies have assessed associations between only ten of the thousands of PFAS and HDP. Positive associations between six PFAS (PFOA, perfluorooctanoic acid; PFOS, perfluorooctane sulfonic acid; PFHxS, perfluorohexane sulfonic acid; PFHpA, perfluoroheptanoic acid; PFBS, perfluorobutanesulfonic acid; and PFNA, perfluoronanoic acid) and risk for HDP have been reported in some, but not all, studies. PFAS disrupt placental and immune function, cause oxidative stress, and disrupt lipid metabolism. These physiological disruptions may be mechanisms through which PFAS can lead to HDP. Overall, limited epidemiological evidence and plausible mechanisms support PFAS as risk factors for HDP. More research is needed in diverse, well-powered cohorts that assess exposures to as many PFAS as possible. Such research should consider not only individual PFAS but also the totality of exposures to PFAS and other environmental chemicals. Pregnant women may be a group that is vulnerable to PFAS exposure, and as such HDP risk should be considered by policymakers setting PFAS exposure limits. In the interim, medical and public health professionals in regions with PFAS contamination could provide short-term solutions in the form of patient-level prevention, increased monitoring, and early intervention for HDP.

Introduction

Hypertensive Disorders of Pregnancy (HDP)

Cardiovascular disease (CVD) is the leading cause of death for women during and after pregnancy (1–4). Hypertensive disorders of pregnancy (HDP), including gestational hypertension; preeclampsia with and without severe features; eclampsia; and hemolysis, elevated liver enzymes, low platelet (HELLP) syndrome, are responsible for 8.3% (5) to 16% (6) of maternal deaths in developed countries and are associated with severe maternal morbidity (7). In the U.S. from 1993 to 2014, the number of deliveries complicated by HDP steadily increased from 5.3% to 9.1% (8). For preeclampsia alone, the cost to the U.S. health care system for complications in the first postpartum year is estimated at $1.03 billion for women, in addition to $1.15 billion to care for the infants born to these mothers (9).

Beyond the immediate health and financial cost of HDP, there are significant immediate and long-term health implications for women (10–12). A retrospective cohort study of all hospital admissions after singleton births among over 1.4 million women in Florida (U.S.) found that women who suffered an HDP had twice the risk of an admission for acute myocardial infarction, stroke, or heart failure within three years post-delivery, with the highest risk among African American patients (12). In a cohort study in the Netherlands following women for 2.5 years postpartum, those with HDP (n=306) had higher cardiovascular risk profile, including increased blood glucose levels, glycosylated hemoglobin, insulin, total cholesterol, high-density lipoprotein cholesterol, triglycerides, and high-sensitivity C-reactive protein compared to women who did not have an HDP (n=99) (13). Thus, finding ways to prevent HDP would not only improve maternal outcomes, but would also serve as an important step toward primary prevention of CVD in young women.

Environmental Chemical Exposures and Pathophysiology of HDP

As the number of women suffering from HDP increases, it is imperative to identify its causes—including those that are potentially modifiable—and to develop broad scale prevention efforts. Medical risk factors for HDP include prior diagnosis of preeclampsia, nulliparity, prior hypertension, diabetes type 1 and 2, chronic kidney disease, systemic lupus erythematosus, antiphospholipid syndrome, advanced maternal age, BMI >35 kg/m², polycystic ovarian syndrome, and low calcium intake (14–16). However, many women with HDP do not have any of these risk factors. Cases with unknown etiology are hypothesized to develop through interactions between environmental and genetic factors that adversely affect the placenta or other maternal physiology (17). Chemical or ‘toxicant’ exposures are emerging as significant environmental factors contributing to HDP risk. Importantly, chemical exposures are modifiable risk factors—both in the short term through behavioral and lifestyle changes and in the long term through environmental policy and regulations.

While research on environmental toxicants and HDP risk is a nascent field, some exposures of concern have already been identified, with many chemicals yet untested. For example, third trimester exposure to ambient fine particulate matter (PM 2.5) was associated with elevated risk for preeclampsia in a meta-analysis including nine cohort studies (18). There is also evidence for metals (lead, cadmium), organochlorine pesticides, and polycyclic biphenyls and preeclampsia risk, as summarized in two recent reviews (17, 19). Phthalates and bisphenols are additional classes of ubiquitous endocrine-disrupting chemicals with mixed evidence as HDP risk factors (20–22).

The pathophysiology of preeclampsia includes endothelial dysfunction, altered immune response, exaggerated inflammatory response, abnormal coagulation, and increased oxidative stress (9, 16, 23). Toxicants can disrupt normal placental function through many of these mechanisms. Exposure to a multitude of toxicants such as tricholorethylene, polybrominated diphenyl ethers, metals, bisphenols, and phthalates have been shown to increase oxidative stress and inflammatory markers, and to negatively impact trophoblast function (17, 24–26). Deficient trophoblast function leads to abnormal remodeling of the maternal uterine spiral arteries; the subsequent poor perfusion leads to increased placental oxidative stress and endothelial dysfunction (17, 27), which can then result in HDP (23).

Objectives of This Review

The aforementioned studies provide epidemiological and mechanistic evidence for toxicants as preventable contributors to HDP. Even so, the types of chemicals studied thus far represent only a fraction of those that women encounter. There is a need to identify hazardous exposures that are widespread among populations. Per- and polyfluoroalkyl substances (PFAS) are a large group of chemicals that are highly ubiquitous in the environment and consumer products and are detected in human populations around the world (28). PFAS are associated with a multitude of adverse health effects including immunotoxicity, cancers, dyslipidemia, and liver and kidney diseases (29). Several of these toxic effects were first discovered in the C8 Health Study, a study comprised of nearly 70,000 individuals with high levels of environmental exposure to one PFAS - perfluorooctanoic acid (PFOA) - in the Mid-Ohio Valley (30). In 2011, the C8 Science Panel concluded that there was a probable link between PFOA exposure and HDP (31), yet other PFAS were not evaluated. In their 2020 review, Blake and Fenton posited the placenta is a target of toxicity for many PFAS that may lead to adverse health outcomes for both the child and mother, including HDP (32). The objectives of our review are to highlight the current epidemiological and mechanistic evidence for multiple PFAS as risk factors for HDP with an emphasis on the state-of-the-science since the C8 Science Panel’s statement, and recommend next steps for research, environmental health prevention efforts, and clinical practice to reduce maternal risk from these exposures.

PFAS as Global Exposures of Concern

PFAS are a group of man-made chemicals with useful industrial properties including being flame retardant and resistant to stains, water, and/or grease. Due to these properties, they are widely found in products including non-stick cookware, fast food packaging, waterproof fabrics, aqueous film-forming foams used to fight fires at airports and military bases, cleaning products, and more (33, 34). An estimated 4,000 different PFAS have existed since the 1940s from a variety of classes and subgroups (Table 1) (35). Currently, environmental scientists only have methods to identify and quantify concentrations of a small fraction of PFAS (<200) in environmental or human samples. While the structures, properties, and, likely, the toxicity of these PFAS vary widely, they all share strong Carbon-Fluorine (C-F) bonds. Due to their resistance to degradation from these strong bonds and their high water solubility, PFAS contaminate groundwater, drinking water, and wildlife in an increasing number of communities throughout the U.S. and the world. For example, in 2016, an estimated 6 million Americans were drinking water that exceeded the U.S. Environmental Protection Agency’s current lifetime health advisory limit for two legacy PFAS: perfluorooctanoic acid (PFOA) and perfluorooctane sulfonic acid (PFOS) (36). Due to improving analytical methods and consideration of additional PFAS, it is now estimated that >100 million Americans may be drinking PFAS-contaminated water (37).

Table 1.

Major classes of per- and polyfluoroalkyl substances (PFAS).

PFAS Groups and Sub-Groupsa	Included in Epidemiological Studies of HDP?b
Perfluoroalkyl Acids and Perfluoroalkylether Acids (PFAA)
Perfluoroalkyl Carboxylic Acids (PFCA)	yes
Perfluoroalkylether Carboxylic Acids (PFECA)	no
Perfluoroalkylether Sulfonic Acids (PFESA)	no
Perfluoroalkyl Phosphinic Acids (PFPiA)	no
Perfluoroalkyl Phosphonic Acids (PFPA)	no
Perfluoroalkyl Sulfonic Acids (PFSA)	yes
Precursors to PFAA
n:2 Fluorotelemer-based substances	no
Hydrofluoroethers	no
Non-polymers	no
Perfluoroalkyl alcohols	no
Side-chain fluorinated polymers	no
Emerging/Replacement PFAS
GenX, ADONA, etc.	noc
Some Additional PFAS
Fluoropolymers	no
Perfluoroalkanes	no
Perfluoropolyethers	no

^aIt is estimated that >4000 PFAS are or have ever been synthesized, and this list is not extensive. Each sub-group contains multiple individual PFAS.

^bWhether specific PFAS analytes within these classes and sub-groups have been measured in samples from human studies focused on hypertensive disorders of pregnancy (HDP) is noted.

^cWhile not included in epidemiological studies of HDP, GenX has been included in in vitro and animal studies focused on mechanisms that may underlie pregnancy complications.

The toxicity of PFAS was not widely known until the 2000s, when PFOA and PFOS were identified in human and environmental samples and linked to adverse health outcomes (28). PFOA and PFOS are associated with a myriad of health effects including elevated cholesterol (38, 39), altered immune response (40), reproductive toxicity (41), and urogenital cancers (42, 43). While PFOA and PFOS are no longer produced in most countries, they are still detected in humans (28, 44). Called ‘forever chemicals,’ the structure of many PFAS makes them resistant to degradation from strong acids, alkalis, oxidizing agents, and photolysis. Some PFAS have biological half-lives as long as 3–8 years and persist in soil and water. Thus, even when phased out of production due to health risks, human populations remain exposed to PFAS for decades (45). According to the U.S. National Health and Nutrition Examination Survey (NHANES), which measures environmental exposures every few years in a representative sample of the U.S. population, PFOA and PFOS were detected in >95% of tested serum samples between 1999 and 2008—even after their production was discontinued (46). While the average concentrations of PFOS and PFOA are decreasing over time in the U.S. population, environmental scientists are now able to measure additional PFAS in human samples—several of which are increasing over time in the U.S. (46, 47) and elsewhere. Importantly, the fraction of PFAS that are typically measured in human populations and assessed for health effects is tiny compared to the number of PFAS in existence and come primarily from only two groups: perfluoroalkyl carboxylic acids (PFCA) and perfluoroalkyl sulfonic acids (PFSA). See Table 1 for major classes of PFAS.

PFAS Exposure and CVD

In non-pregnant adults, there is inconsistent evidence for associations between PFAS and hypertension or other CVD. To give a few examples of the studies on this topic, Min and colleagues examined the cross-sectional association of PFOA and hypertension using data from over 2,000 adults participating in the 2003–2006 U.S. NHANE S and reported that serum PFOA concentrations were positively associated with hypertension in a dose-dependent fashion (48). Pitter et al. examined the relationship of four highly detected PFAS (PFOA, PFOS, PFHxS [perfluorohexane sulfonic acid], and PFNA [perfluoronanoic acid]) and hypertension in over 16,000 Italian adults aged 20–39 years in a cross-sectional survey. PFOA concentrations were associated with higher systolic and diastolic blood pressure in all participants, and all four PFAS were significantly associated with hypertension among males (49).

Conversely, there was no association of cumulative PFOA exposure with hypertension in a cohort of over 30,000 adults from the C8 Health Project, a study comprised of individuals with high levels of environmental exposure to PFOA in the Mid-Ohio Valley. However, exposure levels were estimated via modeling and not direct measures (50). Lin and colleagues also found no statistically significant relationship after correcting for multiple comparisons between plasma concentrations of six PFAS (PFOA, PFOS, PFHxS, PFNA, EtFOSAA, and MeFOSAA) and hypertension among 957 pre-diabetic and overweight adults from the Diabetic Prevention Program cohort (51). In a study of male fisherman in the U.S. state of Wisconsin, it was noted that increasing concentrations of PFNA, a long-chain PFAS, were associated with a lower risk of hypertension, while seven other PFAS measured were not significantly associated with the outcome (52). Additionally, when Mattson et al. examined a cohort of rural male farmers in a longitudinal study, perfluoroheptanoic acid (PFHpA) was statistically significantly higher among individuals who developed coronary heart disease, but levels of seven other PFAS measured did not differ by disease status (53).

There is also preliminary mechanistic evidence supporting a relationship between certain PFAS and cardiovascular health effects. As one example, an in vitro study demonstrated that PFOS treatment of human microvascular endothelial cells induced production of reactive oxygen species (ROS), actin filament remodeling, and changes in endothelial permeability (54). These processes are involved in inflammatory processes and CVD development and provide a window into the potential cellular mechanism of PFAS inducing CVD (54). Overall, epidemiological and mechanistic evidence to date suggests a modest link between PFAS and CVD among non-pregnant adults that may vary based on the exact PFAS, the population, and other confounding or moderating factors yet to be identified. Given the importance of the placenta—a potential target organ of PFAS toxicity—in pregnancy-specific hypertensive disorders, it is imperative we understand the risk of PFAS exposure to the cardiovascular health of pregnant women.

PFAS Exposure and HDP

At least nine epidemiological studies have examined the association between PFAS exposure and the development of HDP. Table 2 presents details on each of these studies while Supplemental Table 1 provides a high-level overview of their findings. In 2009, Stein and colleagues (55) evaluated approximately 2,000 women who were part of the C8 Health Project. They identified modest associations between the highest levels of PFOA and PFOS in women’s serum measured approximately five years after pregnancy with their reported histories of preeclampsia. While this study was unable to demonstrate causation, as the toxicant levels were measured after women had experienced an HDP, this study inspired others to investigate the link. In 2012, Savitz et al. used a larger sample of women from the C8 Health Project to reexamine this association (56). Using modeling-based estimates of exposure to PFOA, they identified a weak association between PFOA and self-reported preeclampsia among women who had a singleton, live birth. This study added to the concern over the relationship between PFAS and HDP, and contributed to the C8 Science Panel’s decision to declare a ‘probable link’ between PFOA and pregnancy-induced hypertension (31). Even so, both studies were disadvantaged by relying on women’s recollection of having a diagnosis of preeclampsia and retrospective exposure assessment.

Table 2.

Details from nine epidemiological studies assessing associations between gestational PFAS exposure and hypertensive disorders of pregnancy.

Reference	Cohort Region*	Cohort Name	Total Sample Size	Number of HDP Cases	HDP included	PFAS Assessed	Exposure Assessment Timing	Statistically Significant Associations between PFAS and HDP
Birukov et al., 2021 Environ Int	Denmark	Odenese Child Cohort	1436	153 (49 gHTN, 104 PreE)	Repeat BP measures, gHTN, and PreE (from medical chart)	PFOS, PFOA, PFHxS, PFNA, PFDA	First trimester serum	In adjusted models, doubling of PFOS or PFOA associated with higher repeated measures of SBP and DBP. No statistically significant associations between PFAS and PE or gHTN in adjusted models. Doubling of PFOS associated with increased DBP (0.58 [0.13, 1.04] mmHg). Doubling of PFHxS associated with decreased SBP (−0.51 [−0.97, [0.04]).
Borghese et al., 2020 Environ Int	Canada	MIREC	1739	176 (127 gHTN, 49 PreE)	gHTN and PreE (real-time evaluation by study team and medical chart review)	PFOS, PFOA, PFHxS	First trimester plasma	In adjusted models, PFHxS associated with PreE (OR 1.32 [1.03–1.70]) and near-statistical significance for association with gHTN (OR 1.15 [0.98–1.35] per log-2 unit increase in PFHxS). Offspring sex-dependent associations were also reported for PFOS and PFHxS with gHTN (males).
Huang et al., 2019 Environ Health	China	N/A	686	42	gHTN, PreE, and overall HDP (from medical chart)	PFOA, PFOS, PFNA, PFUA, PFDA, PFHxS, PFDoA, PFBS	Umbilical cord blood plasma	In single exposure models, the third tertile of PFBS was associated with odds for HDP (OR 2.21 [1.00–4.88]). In multiexposure models, PFBS was also significantly associated with HDP or PreE alone.
Huo et al., 2020 Environ Int	China	Shanghai Birth Cohort	3220	135	gHTN, PreE, and overall HDP (from medical chart)	PFOA, PFOS, PFNA, PFDA, PFUnDA, PFHxS, PFHpA, PFBS, PFDoA	Early pregnancy serum	In adjusted logistic regression, there was elevated risk for gHTN (OR 1.38 [1.01–1.87]) with each in-unit increase in PFHpA. No PFAS were identified as predictors of HDP in the exposure mixture model (using elastic net regression).
Rylander et al., 2020 Toxics	Sweden	Southern Sweden Maternal Biobank	876	296	PreE (from medical chart)	PFOA, PFOS, PFNA, PFHxS	Early pregnancy serum	PreE cases had significantly higher concentrations of PFNA among multiparous women compared to controls, and the association between PFNA and PreE was near statistically significant in adjusted models. Women in the third quartile of PFHxS exposure had significantly higher odds for PreE in the adjusted model (OR 1.67 [1.02–2.74]).
Savitz et al., 2012 Epidemiology	U.S.	C8 Health Project	10,186	730	PreE (self-reported)	PFOA	Historical estimates based on modeling	OR 1.13 (1.00–1.28) for an interquartile shift in log-transformed PFOA. Association strengthened in nulliparous women alone.
Starling et al., 2014a, Am J Epid	Norway	Norwegian Mother and Child Cohort Study	976	466	PreE (from medical chart)	PFOA, PFOS, PFDA, PFNA, PFUnDA, PFHxS, PFHpS	Mid-pregnancy plasma	Mostly null associations. There were positive associations between PFHpS and PFOS in adjusted models that were near statistical significance. There was an inverse association between PFUnDA and PreE that was statistically significant when modeling quartiles of exposure. All women were nulliparous
Stein et al., 2009 Am J Epid	U.S.	C8 Health Project	1589 (with PFOA); 4566 (with PFOS)	156 (with PFOA); 407 (with PFOS)	PreE (self-reported)	PFOS, PFOA	Maternal serum levels within 5 years after the index pregnancy	In adjusted models, OR for PFOS and PFOA with PreE was positive, but the only statistically significant association was observed between PFOS concentrations and PreE among women above the 50th percentile or 90th percentile.
Wikström et al., 2019 Sci Rep	Sweden	SELMA	1773	64	PreE (from medical chart)	PFOS, PFOA, First trimester PFHxS, PFNA, serum PFDA, PFUnDA, PFHpA, PFDoDA	First trimester serum	One log-transformed unit of PFOS was associated with an adjusted OR of 1.53 (95% CI 1.07–2.20) for PreE. PFNA was also associated with PreE (OR 1.38 [1.01–1.89]). Both associations were stronger among nulliparous women only.

^*Indicates which country the cohorts were recruited from.

PFAS=Per- and Polyfluoroalkyl Substances; HDP = hypertensive disorders of pregnancy; gHTN = gestational hypertension; MIREC = Maternal-Infant Research on Environmental Chemicals; N/A = not available; study population does not have a specific name; OR = odds ratio; PreE = preeclampsia; SELMA=Swedish Environmental Longitudinal, Mother and child, Asthma and allergy; PFHpA = Perfluoroheptanoic acid; PFOA= Perfluorooctanoic acid; PFNA= Perfluoronanoic acid; PFDA=Perfluorodecanoic acid; PFUnDA = Perfluoroundecanoic acid; PFDoA = Perfluorododecanoic acid; PFBS= Perfluorobutanesulfonic acid, PFHxS = Perfluorohexane sulfonic acid, PFHpS = Perfluoroheptanesulfonic acid; PFOS= Perfluorooctane sulfonic acid.

In 2014, Starling and colleagues used data from the Norwegian Mother and Child Cohort Study conducted by the Norwegian Institute of Public Health to examine the association between seven PFAS chemicals measured in the second half of pregnancy and preeclampsia. Though their definition of preeclampsia differed from accepted guidelines in the U.S., they were able to validate the diagnosis with chart review among the cases, but not the controls. This study found mixed results, with some chemicals showing a positive and one showing a negative association with HDP, though only an inverse association between perfluoroundecanoic acid (PFUnDA) and preeclampsia was statistically significant in confounder-adjusted models (57). Ultimately, these authors concluded that concentrations of PFAS in lowly-exposed populations were not linked to the development of HDP.

In 2019, Huang and colleagues examined 674 women from China and measured concentrations of six PFAS in umbilical cord blood (58). One of the PFAS measured—perfluorobutanesulfonic acid (PFBS)—was significantly associated with HDP in logistic regression of both the single exposure and a multiple exposure model. This study validated patient diagnoses from medical records, but exposure was measured at the time of delivery, making it difficult to attribute causation. In 2020, Huo and colleagues built upon this work with a prospective study design. They measured concentrations of 10 PFAS in early gestational serum from 3,220 women, including 135 who went on to develop an HDP (59). Results were mostly null, though PFHpA was associated with an elevated risk for gestational hypertension (OR 1.38 [95% CI 1.01–1.87]).

In 2019, Wikström and colleagues (60) reported on the association between first trimester serum levels of eight PFAS and patient-reported preeclampsia among women in the Swedish Environmental Longitudinal, Mother and child, Asthma and allergy (SELMA) cohort. They reported an adjusted odds ratio of 1.53 (95% C I 1.07–2.20) for preeclampsia, with an increase of PFOS concentration from the first to fourth quartile of exposure. Associations were stronger among nulliparous women. Patient-reported outcomes were again used for the diagnosis of preeclampsia, though less time had passed between the pregnancies of interest and the conducted study.

In another Swedish study, Rylander and colleagues utilized a case-control design to assess associations between early pregnancy PFOA, PFOS, PFHxS, and PFNA and HDP among 296 cases and 580 controls (61). While most associations were not statistically significant, among multiparous women, PFNA concentrations were higher among cases compared with controls. In adjusted models, quartiles of PFNA and PFHxS were associated with increased odds for preeclampsia; this only obtained statistical significance for the third quartile of PFHxS exposure (OR 1.67 [95% CI 1.02–2.74]).

In 2020, Borghese and colleagues evaluated the association between first trimester levels of PFOA, PFOS, and PFHxS and preeclampsia and gestational hypertension in the Canadian Maternal-Infant Research on Environmental Chemicals (MIREC) cohort (62). Medical records were reviewed to confirm pregnancy outcomes. This study identified a statistically significant association between PFHxS and preeclampsia when modeling the exposure as continuous or categorized by tertiles. The association remained near statistically significant when stratified by offspring sex in both male (p=0.12) and female (p=0.11) groups. Among women carrying a male fetus, PFOS and PFHxS were also significantly associated with gestational hypertension. This study was the first to evaluate the relationship between PFAS and HDP stratified by offspring sex.

In 2021, Birukov and colleagues examined the association between 5 PFAS measured in first trimester serum (PFOA, PFOS, PFHxS, PFNA, and PFDA) with gestational hypertension (49 cases out of 1436 women) and preeclampsia (104 cases) (63). Participants were part of the Odenese Child Cohort, recruited in Denmark, and diagnosis was from their medical charts. No statistically significant associations were reported between any PFAS and either HDP when adjusting for maternal age, smoking, parity, and pre-pregnancy BMI. However, this team also evaluated associations between repeat clinical measures of blood pressure taken throughout pregnancy among all the women. A doubling of PFOS was associated with 0.58 mmHg increase in diastolic blood pressure (95% CI [0.13, 1.04]). A doubling of PFHxS was associated with lower systolic blood pressure (−0.51 mmHg [−0.97, 0.04]). While small, the authors concluded that this type of change could be significant on a population level especially among populations with higher levels of exposure to PFAS.

Collectively, these nine studies assessed associations between 10 different PFAS and preeclampsia and/or additional HDP (Table 2). While not consistent across all studies, there is some evidence pointing to associations between PFHpA, PFOA, PFNA, PFBS, PFHxS, and PFOS with HDP. The legacy toxicant, PFOS, which remains detectable in nearly all populations worldwide, was significantly associated with risk for HDP in three out of eight studies that evaluated it and was also associated with increased diastolic blood pressure during pregnancy. PFHxS is another legacy PFAS that was associated with HDP in two out of seven studies. Importantly, the health impacts of PFHpA, PFNA, PFBS, and PFHpS are only recently being evaluated; each of these PFAS had evidence for association with HDP. PFNA was significantly associated with HDP in two out of six studies. PFHpA and PFBS were associated in one out of two studies each. PFHpS was only included in one study, and the association between this chemical and preeclampsia was nearly statistically significant (57). There was no evidence for a positive association between long-chain PFCAs (PFDA, PFUnDA, PFDoA) and HDP; each was included in at least three studies.

Inconsistences between studies including the same PFAS may reflect inadequate statistical power or differences in outcome assessment, analysis methods, and/or study population. While most measured PFAS in early gestation, there were differences in the timing of exposure assessment across studies. PFAS plasma/serum concentrations are known to decrease across pregnancy due to plasma volume expansion and distribution of PFAS into the fetal compartment (64). As far as the outcomes are concerned, five studies focused on only preeclampsia and four included gestational hypertension or other HDP. Only one study assessed the association with repeated blood pressure measures during pregnancy. Recently, women from Project Viva were followed up to 3 years postpartum, and associations between gestational PFAS and postpartum outcomes were assessed. A doubling of PFOS was associated with 1.2 mmHg higher postpartum systolic blood pressure (0.3, 2.2), signifying the need to evaluate women’s health beyond pregnancy when considering the potential impacts of PFAS (65).

Depending on the study, cases were identified via medical chart review or self-report; the latter of which can be unreliable (66). The set of confounders included in analyses across the nine studies were similar; all included maternal age, maternal weight or pre-pregnancy BMI, parity and smoking status or exposure. A few included fetal sex, time of exposure measurement, and education. Several studies demonstrated effect modification by parity. Only two studies stratified by fetal sex, and this needs to be considered as an effect modifier in the future. Dietary factors and physical activity could be confounders or effect modifiers in relationships between PFAS and HDP, yet these data were not available in any of the reported studies. Study populations were from China or five predominantly Caucasian nations (Canada, Denmark, Norway, Sweden, and U.S.), and race or ethnicity was not considered in any analysis. The burden of exposure to PFAS within these study populations was not equivalent, and susceptibility to PFAS toxicity may vary.

Even with these limitations, there is evidence for exposure to PFSAs and PFCAs with carbon chain lengths up to nine as risk factors for HDP (Table 2 and Supplemental Table 1). In the case of the widely studied 8-carbon PFOA and PFOS, PFOS has more consistent evidence for HDP risk compared with PFOA. PFSAs and PFCAs represent only two sub-groups of PFAS (Table 1); it is completely unknown whether PFAS from other groups influence HDP. In vitro studies suggest that newer PFAS, such as GenX, exhibit toxicity (67). Mechanistic studies can shine light on PFAS of concern that are yet to be evaluated in epidemiological studies.

Potential Mechanisms of Action for PFAS Leading to HDP

PFAS may lead to HDP through multiple biological mechanisms of action including disruptions to placental function or dyslipidemia, in part through peroxisome proliferator-activated receptor (PPAR) activation. Immunotoxicity and oxidative stress are also relevant mechanisms of action to HDP. In a recent review, the placenta was posited as a target organ for PFAS that mediates the effects of PFAS on HDP and other adverse reproductive and developmental outcomes (32). In mice, exposure to PFOA or its emerging replacement PFAS, GenX, had dose-dependent effects on placenta weights and histopathology (67). Normal placental invasion is dependent on active regulation by the maternal immune system and alterations to invasion and vascularization are known features of HDP (68) that could be disrupted by PFAS. Decreased inflammatory signaling in trophoblasts is one of the ways in which PFAS may affect placental development. In vitro, PFOS, PFOA, and GenX treatment each decreased trophoblast migration and expression of chemokines and inflammatory enzymes necessary for migration (69). This mechanistic evidence is concerning since abnormal trophoblast invasion and dysfunction are characteristic of preeclampsia (70). In another in vitro study, PFOA (but not PFOS) treatment reduced vascularization in a vascular network 3D model using the HTR8/SVneo trophoblast cell line (71). There is evidence that other placental cell types are sensitive to dysregulation by PFAS. Using the BeWo cell line of multinucleated syncytial cells, PFOS and PFBS alter regulation of key functional genes (72).

PFAS are known to influence lipid metabolism and utilization, with most studies to date focusing on cholesterol. Large studies including the C8 Health Project and the U.S. NHANES have established associations between elevated cholesterol (total and LDL)—a risk factor for hypertensive disorders—and PFOS, PFOA, PFNA, PFDA, and PFUnDA concentrations in children, adolescents, and/or adults (38, 39, 50, 73, 74). Two cross-sectional studies have examined this association in pregnant women and found that PFOS or PFOA were associated with increased total cholesterol, PFOS and PFNA were inversely associated with triglycerides, and five PFAS were associated with elevated HDL levels during pregnancy (75, 76). Elevated first and third trimester triglycerides are positively associated with pregnancy-induced hypertension and preeclampsia (77, 78). A lipidomics study profiling multiple classes of lipids identified first trimester lipids that predict preeclampsia; how these lipids contribute to pathogenesis of HDP is not yet known (79). Preliminary evidence suggests statins may prevent the development of preeclampsia; a randomized clinical trial of pravastatin for the prevention of preeclampsia in pregnant women is currently ongoing (ClinicalTrials.gov: NCT03944512).

One mechanism of PFAS toxicity that may mediate the effects on both placental function and lipid homeostasis is PPAR activation. This mechanism was recently reviewed by Szilagyi et al. (80). Briefly, PPAR are nuclear receptors with critical roles in regulating fatty acids and metabolism. PPAR-alpha and PPAR–gamma are important for placental function. In vitro and in vivo studies have provided evidence that PFAS activate various PPAR, including in the placenta, and this may one of the molecular mechanisms underlying its effects.

Healthy pregnancy requires a delicate immune system balance for successful fetal and placental tolerance, protection against infection, and, ultimately, labor (81). PFAS, however, may interfere with normal immune function during pregnancy. In rodents and humans, chronic inflammation, characterized by increased oxidative stress and pro-inflammatory cytokines, precedes HDP (82–84). While there have been few studies of PFAS and immune response in pregnant women, one study of obese and overweight pregnant women from California reported an association between early gestation PFAS and increased interleukin-6—an inflammatory biomarker measured in early gestation and at three and six months postpartum (85). Another related mechanism by which PFAS may be linked to HDP is through an increase in reactive oxygen species (ROS). In endothelial cells, PFAS induce the production of ROS and increase endothelial permeability in vitro (54). Cultured trophoblasts have also demonstrated an increase in ROS generation after exposure to PFOS (86). Other studies have linked elevated levels of various PFAS chemicals to markers of increased oxidative stress in fish, birds, and occupationally-exposed adults (87–89). Whether oxidative stress is increased by PFAS in pregnant women remains unexplored.

Summary and Future Needs

Summary of Evidence to Date for PFAS and HDP

HDP are morbid processes that have been increasing at alarming frequency with both immediate and long-term detrimental health effects (8, 12). In addition to medical risk factors, exposures to environmental toxicants are emerging as preventable risk factors for HDP. In this review, we present evidence for PFAS exposures as risk factors for HDP. We discussed the results of nine epidemiological studies that collectively assessed associations between 10 PFAS and preeclampsia or other HDP (Table 2). These studies provide some evidence for legacy PFAS (PFOA, PFOS, PFHxS) and PFAS with ongoing uses (PFHpA, PFBS, and PFNA) as risk factors for HDP. However, evidence is inconsistent for PFAS that are included in multiple studies. Inconsistences may arise due to differences in timing and method of exposure assessment (e.g., direct measurement versus approximation for PFOA studies), statistical power due to sample size, population differences, and method for determining diagnosis (e.g., some studies relied on self-reported diagnoses, while others had access to medical records). While patient-provided history is important, prior work has demonstrated that the positive predictive value of a positive response to a question about history of preeclampsia or gestational hypertension may be as low as 56% and 64% respectively (66).

Six of the PFAS included in HDP-focused research were only included in one to four studies each. Among these, PFHpS and long-chain PFCAs (PFDA, PFUnDA, and PFDoA) were not associated with HDP. However, PFHpA and PFBS were associated with HDP in one out of two studies each. Since these were not included in most studies, and there is also in vitro evidence for a mechanistic basis by which PFBS could contribute to HDP (72), it is imperative that the risk of these PFAS are assessed in future studies.

While the link between PFAS exposure and HDP is compelling, the association with general CVD is less clear. There is mechanistic evidence that the placenta is a unique target of PFAS, which may lead to the increased risk of HDP (32), and this may occur in part through disruptions to PPAR signaling (80). Other mechanistic links between PFAS exposure and HDP include a disrupted inflammatory response, increased ROS formation, and dyslipidemia. How or whether each of these mechanisms leads to PFAS-induced HDP remains to be determined. Along with developing children, pregnant women may be particularly vulnerable to PFAS toxicity, and prevention efforts and environmental regulations should be set to ensure their safety. Importantly, PFAS from all but two sub-groups (Table 1) have never been evaluated for their association with HDP; this assessment is urgently needed.

Recommendations for Future Research

While some epidemiological evidence and plausible mechanisms support PFAS as risk factors for HDP, much research needs to be done to solidify this connection and inform appropriate medical recommendations and environmental policymaking. Several challenges to this type of research include logistics with sample collection, timing, and appropriately selecting specific PFAS chemicals to study. Studies obtaining appropriately timed samples from women and validating their diagnosis of an HDP are challenging and expensive to conduct. The studies conducted to date demonstrate modification or interaction by parity and fetal sex (56, 60, 62); these and other factors need to be considered to understand the totality of risk from HDP. We recommend at a minimum that future studies assess effect modification by fetal sex, parity, race/ethnicity and that studies also include timing of PFAS measurement during pregnancy (64) as a covariate in analyses. Maternal age, smoking status, and pre-pregnancy BMI are necessary confounders to consider, and modification by diet and physical activity is an important consideration for future research in this area. Studies based in the U.S. have reported differences in exposure levels to PFAS by race and ethnicity (90–92). Inequitable exposure burden could be one of many factors contributing to racial disparities in maternal morbidity, including HDP. The association between PFAS and HDP needs to be tested in racially and ethnically diverse populations in order to understand risk and design prevention efforts for the most vulnerable.

To fully appreciate the impact of PFAS on any health outcome, studies are hindered by the fact that >4,000 types of PFAS exist and much of the existing literature focuses on a small number of PFAS chemicals (Table 1). As noted by Sunderland and colleagues, future research needs to include a broader number of PFAS analytes and their cumulative burden to fully understand the impact of these chemicals (28). The cost of analyzing multiple PFAS, while decreasing over time, can be prohibitively expensive. Exposure assessment methods and strategies for grouping PFAS when performing risk assessment are in development (93, 94). These and computational toxicology approaches could be applied to assess whether PFAS exposures impact HDP risk, including inferences about whether PFAS with limited human data would be expected to have similar effects based on chemical structure and in vitro data. In epidemiological studies with multiple PFAS quantified, methods for modeling exposure mixtures will be of increasing importance, and development of this methods is a growing area. Some of the available statistical methods include but are not limited to: environmental risk score (95), toxicant score (96), supervised principal component analysis proceeded by classification and regression tree (97), joint weighed quantile sum (WQS) regression (98, 99), Bayesian kernel machine regression (100), and machine learning methods such as lasso or adaptive elastic net (101). The correlation structure between PFAS is key when determining which method to use. For instance, joint WQS has good sensitivity and specificity to identify predictors from among a correlated mixture while some other methods are better for non- or weakly-correlated exposures (102–104).

Future Implications for Environmental Policy

Understanding the impact of PFAS on HDP is critical for risk assessment managers and environmental policymaking as governing bodies of communities, states, and countries navigate, revise, and develop standards for PFAS in groundwater and drinking water supplies. Pregnant women may be a group that is particularly vulnerable to the effects of PFAS, and as such, HDP risk should be considered in policymaking. Fortunately, since the C8 Science Panel reported a probable link between PFOA and HDP (31), risk assessment for some PFAS and policies by some governing bodies have considered HDP risk in their recommendations (i.e. for PFAS limits in drinking water and food). However, this assessment has been limited to very few PFAS – typically only PFOA and PFOS (105). Presented in this review, there is now some evidence, for several other PFAS and HDP risk (Table 2). Even so, the majority of the thousands of PFAS in production and use throughout the world have never been measured in human populations or assessed for human health risks (Table 1). Methods to infer, based on chemical structures, which PFAS may impact the same health outcomes can be employed, and regulations will ultimately need to consider exposure mixtures to capture the real-world exposure scenario for pregnant women (94).

Environmental risk assessment continues to evolve, with the goal of considering environmental exposures from multiple dimensions and their collective impact when making public health and policy decisions to protect populations. New methods for analyzing the ‘exposome’—or the totality of external exposures—will be instrumental in identifying the most hazardous exposures. For example, in a large study using >800,000 birth records for the U.S. state of Florida, ambient pollutants along with neighborhood-level and sociodemographic status variables were selected as the most associated with risk for HDP (106). However, these largescale efforts have not yet included chemicals such as PFAS. In order to decrease maternal morbidity related to HDP from environmental exposures, risk assessment and environmental policymakers need to identify hazardous exposures—which may include PFAS—and reduce their sources of exposure to protect populations.

Future Implications for Medical Care

Before environmental policy and industrial practices change, medical and public health practitioners may be uniquely positioned to provide short-term solutions to environmental-related HDP in the form of patient-level prevention and early intervention. With increased environmental monitoring of PFAS in the U.S. and some other countries, many ‘hotspots,’ or communities with high exposure burden, are known. Medical professionals working with women in these regions could educate their patients about PFAS chemicals and direct them to resources about how to avoid PFAS in consumer products and clean up drinking water. Furthermore, patient populations in these high-risk communities could be monitored closely throughout the peripartum period for development of hypertension or preeclampsia in order to provide intervention as soon as necessary. In the future, new insights into the mechanism of disease can provide a window into therapeutic targets to mitigate the risk of HDP, such as the use of aspirin for HDP risk reduction in women with known high exposure levels (16). Through short- and long-term efforts aimed at both prevention and intervention, the prospect of reducing HDP by targeting hazardous environmental chemicals such as PFAS could lead to immediate and long-term maternal health benefits.

Abbreviations:

CVD

Cardiovascular disease

HDP

Hypertensive disorders of pregnancy

HELLP

hemolysis, elevated liver enzymes, low platelet

PFAS

Per- and polyfluoroalkyl substances

C-F

Carbon-Fluorine

PFOA

perfluorooctanoic acid

PFOS

perfluorooctane sulfonic acid

NHANES

National Health and Nutrition Examination Survey

PFCA

perfluoroalkyl carboxylic acids

PFSA

perfluoroalkyl sulfonic acids

PFHpA

perfluoroheptanoic acid

ROS

reactive oxygen species