Chemosphere 367 (2024) 143606

Contents lists available at ScienceDirect

Chemosphere
journal homepage: www.elsevier.com/locate/chemosphere

Low-salt diets and salt-free cooking help reduce exposure to Per- and
polyfluoroalkyl substances (PFAS)
Shuai Zhang a,b,* , Hanhan Tang c
a
Department of Male Reproductive Health, Lianyungang Maternal and Child Health Hospital, Qindongmen Avenue, Haizhou District, Lianyungang, 222000, China
b
Clinical Center of Reproductive Medicine, Lianyungang Maternal and Child Health Hospital, Qindongmen Avenue, Haizhou District, Lianyungang city, 222000, China
c
Plastic Surgery Department, Xuzhou Central Hospital, No. 209, Tongshan Road, Xuzhou city, 221004, China

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• A low-salt or low-sodium diet is associ­

ated with a lower risk of PFAS exposure.
• Salt-free cooking is associated with a
lower risk of PFAS exposure.
• Improved dietary and cooking habits
may help reduce the risk of PFAS
exposure.
• Multiple sensitivity analyses were con­
ducted, including PSM.

A R T I C L E I N F O A B S T R A C T

Handling editor: Prof. A. Gies Background: The ubiquity of Per- and polyfluoroalkyl substances (PFAS) in various consumer and industrial
products poses a significant public health challenge, but effective strategies to reduce human exposure to PFAS
Keywords: are limited.
Per- and polyfluoroalkyl substances Objectives: This study aims to evaluate the association between dietary patterns, specifically low-salt diets and
Low-salt or low-sodium diet
salt-free cooking, and serum PFAS levels in the general population.
Salt-free cooking
Methods: The study analyzed data from 11,137 participants from the National Health and Nutrition Examination
Exposure risk
Dietary patterns Survey (NHANES) using weighted linear regression. We assessed associations between low-salt or low-sodium
NHANES dietary patterns and the way salt was used during cooking or food preparation and serum levels of five highly
detectable PFAS compounds: perfluorooctanoic acid (PFOA), perfluorooctanesulfonic acid (PFOS), per­
fluorohexanesulfonic acid (PFHxS), perfluorodecanoic acid (PFDA), and perfluorononanoic acid (PFNA). Since
consuming fish and shellfish is a major source of PFAS exposure in humans, the intake status of these foods was

* Corresponding author.
E-mail address: senomega@hotmail.com (S. Zhang).

https://doi.org/10.1016/j.chemosphere.2024.143606
Received 27 May 2024; Received in revised form 20 October 2024; Accepted 21 October 2024
Available online 22 October 2024
0045-6535/© 2024 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
S. Zhang and H. Tang Chemosphere 367 (2024) 143606

adjusted for in the sensitivity analysis. Additionally, other sensitivity analyses, including propensity score
matching, were conducted.
Results: The analyses showed a significant negative association between low-salt or low-sodium diet and serum
levels of the five PFAS compounds. In contrast, regular use of salt in cooking or food preparation was signifi­
cantly and positively associated with higher serum levels of PFAS. These findings were consistent across all
models. Also consistent were the results of sensitivity analyses based on participants’ consumption of fish and
shellfish and propensity score matching.
Conclusions: Low-salt or low-sodium dietary patterns, and salt-free cooking may be are associated with a reduced
risk of PFAS exposure in the general population. While this study offers new insights into mitigating PFAS
exposure, further validation in additional datasets is necessary, along with confirmation through intervention
studies designed based on this hypothesis.

1. Introduction These reports indicate that different cooking methods and the use of
seasonings such as salt during cooking may influence the release of PFAS
Per- and polyfluoroalkyl substances (PFAS) are a new class of pol­ substances from food, thereby affecting human exposure risk to PFAS.
lutants known as "forever chemicals” because of their strong carbon- According to existing data, the relationship between specific dietary
fluorine bonds, which make PFAS extremely stable to thermal, chemi­ habits, such as a low-salt or low-sodium diet, and PFAS exposure risk has
cal, and biological degradation (Buck et al., 2011). PFAS are widely used not been systematically studied in the general population. Additionally,
in a variety of living and production activities, such as daily chemical the impact of salt usage habits, such as salt-free cooking, on PFAS
products, mouthwash, food packaging materials, nonstick cookware, exposure is also rarely reported in the literature. Therefore, this study
surfactants, firefighting foams, carpets, and furniture (Pelch et al., analyzes the correlation between a low-salt or low-sodium diet and
2019). Nowadays, PFAS are almost ubiquitous in the environment, and serum levels of five common PFAS compounds—PFOA, PFOS, per­
diet is the main route of human exposure to PFAS, including ingestion of fluorohexane sulfonic acid (PFHxS), perfluorodecanoic acid (PFDA), and
contaminated food and water (Sunderland et al., 2019). PFNA—in the U.S. adult population, to explore the potential impact of
PFAS are a class of endocrine disruptors that pose widespread health such diets on PFAS exposure. Moreover, this study further investigates
risks, raising concerns about the potential health hazards associated the association between salt usage methods during cooking and food
with PFAS exposure. Extensive epidemiological evidence suggests that preparation and PFAS exposure risk.
exposure to PFAS is closely linked to an increased risk of various dis­
eases, including but not limited to infertility (Green et al., 2021), dia­ 2. Methods
betes (Roth and Petriello, 2022), hypertension (Pan et al., 2023; Averina
et al., 2021), autoimmune diseases (Qiao et al., 2024), and cancer (van 2.1. Study population
Gerwen et al., 2023; Steenland and Winquist, 2021). However, due to
the current lack of effective methods to degrade PFAS, how to effectively The National Health and Nutrition Examination Survey (NHANES) is
reduce the exposure of PFAS remains a significant public health chal­ a research program that employs a complex sampling design to assess
lenge, aside from limiting its use. the health and nutritional status of American adults and children.
Due to the challenges in degrading PFAS, their accumulation effect in Conducted biennially, this study is organized and implemented by the
the biosphere is particularly pronounced (Miranda et al., 2021), espe­ National Center for Health Statistics (NCHS). All participants in the
cially among mammals, fish, and marine life (Onteeru et al., 2022). study provide written informed consent.
Notably, variations in individual dietary habits not only influence the Since the 2003–2004 cycle, NHANES has begun to collect and pub­
risk of some diseases but also affect the frequency and dosage of expo­ lish dietary interview questionnaire data concerning participants’ spe­
sure to hazardous pollutants. Previous studies have linked frequent cial dietary patterns. Consequently, this study includes NHANES data
consumption of restaurant foods and popcorn to higher serum concen­ from eight consecutive cycles spanning from 2003 to 2018, involving a
trations of PFAS (Susmann et al., 2019). Similarly, regular consumption total of 80,312 participants. During the screening process, we excluded
of fish and shellfish is associated with significant increases in serum participants under the age of 20 (N = 35,522), those lacking data on
PFAS levels (Christensen et al., 2017). Additionally, an increase in di­ PFAS compounds (N = 31,822), and those without information on low-
etary fiber can promote the excretion of perfluorooctanoic acid (PFOA), salt or low-sodium dietary patterns (N = 808). Additionally, we
perfluorooctane sulfonic acid (PFOS), and perfluorononanoic acid excluded participants with incomplete covariate information (N =
(PFNA) (Mw et al., 2021). A study comparing vegetarians and omni­ 1023). After these exclusions, a total of 11,137 participants were
vores found significantly lower levels of PFOS and PFNA in vegetarians included in the final analysis (see Fig. 1). Among them, 196 individuals
(Menzel et al., 2021). These findings suggest that dietary improvements (1.8%) reported following a low-salt or low-sodium diet. This is a cross-
could potentially reduce PFAS exposure and, consequently, the risk of sectional study, and all participants provided only one set of survey
certain diseases, as PFAS exposure may play a mediating role between information for inclusion in this research.
the two.
The processing and cooking of food can influence the release of PFAS 2.2. Low-salt or low-sodium diet information
substances within the food (Vendl et al., 2022). Different cooking
methods may cause changes in PFAS concentrations in food. Some Since the 2003–2004 cycle, NHANES has been collecting information
studies have found that the PFAS concentration in baked foods increases from participants during dietary interviews on whether they are
while frying results in a decrease in PFAS levels (Chen et al., 2023). currently following any diets for weight loss or other health-related
Steaming bullfrogs can reduce PFAS content by over 66% (Sun et al., reasons. Participants are asked, "Are you currently on any kind of diet,
2023). Conversely, the concentrations of PFOA, PFHxS, and PFOS in either to lose weight or for some other health-related reason?” If they
shrimp boiled in brine actually doubled after boiling (Taylor et al., respond affirmatively, they are further asked about the type of diet they
2019). Furthermore, a recent survey study of middle-aged women in the are on. If a participant chooses "DRQSDT3 - Low-salt or low-sodium diet
United States found that frequent use of salty foods was significantly (including diet to lower blood pressure or hypertension)" in response to
associated with higher serum concentrations of PFOA, PFOS, and the question "What kind of diet are you on?” it indicates that they are
2-(N-ethyl-perfluorooctane sulfonamido)acetic acid (Park et al., 2019). following a low-salt or low-sodium diet (a).

2
S. Zhang and H. Tang Chemosphere 367 (2024) 143606

In addition to collecting data on low-salt or low-sodium diets, guidelines, total PFOA is calculated as n-PFOA + Sb-PFOA, and total
NHANES also investigated the usage of salt during the cooking or food PFOS as n-PFOS + Sm-PFOS. The aggregate results are consistent with
preparation process (a) This primarily inquired about the frequency total PFOA and PFOS levels from previous cycles. Detailed information
with which participants use salt when cooking or preparing food at on PFAS detection is available in Supplementary Table 1.
home, excluding food prepared outdoors. Specifically, the question,
labeled as DRQSPREP, asked how often regular or seasoned salt is added
2.4. Covariates
during home cooking or food preparation, with response options
including never, rarely, occasionally, and very often.
Referencing prior studies (Calafat et al., 2007; Zhu et al., 2021), the
statistical analysis adjusted for potential confounders that could affect
outcomes, including age (continuous), gender (male and female),
2.3. PFAS levels
race/ethnicity (categorized as Mexican American, Non-Hispanic White,
Non-Hispanic Black, Other Hispanic, and Other groups), education
Since 1999, NHANES has been conducting ongoing tracking of PFAS
attainment (classified as Less than 9th Grade, 9th-11th Grade, High
exposure within the general U.S. population. This study selected five
School Graduate/GED or Equivalent, Some College or AA Degree, and
PFAS compounds—PFOA, PFOS, PFHxS, PFDA, and PFNA—which have
College Graduate or Above), smoking status (categorized as never,
a serum detection positivity rate exceeding 70%, to assess levels of PFAS
former, current), poverty income ratio (PIR) (<1, 1–2, and ≥2), and
exposure. Among these, PFDA had the lowest detection positivity rate at
NHANES survey cycles.
73.22%, while PFOS was the highest at 99.42%. Details on the sample
collection, processing, and analysis can be found on the NHANES official
website (b). If the serum concentration is below the lower limit of 2.5. Statistical analysis
detection (LLOD), NHANES substitutes the value with LLOD/sqrt(2).
Starting from the 2013–2014 cycle, NHANES measured concentrations Given that NHANES employs a complex multistage probability
of linear PFOA (n-PFOA), and sum of branched isomers of PFOA (Sb- sampling design, we accounted for the weights of the two-year PFAS
PFOA), as well as linear PFOS (n-PFOS), and the sum of per­ subsample in subsequent statistical analyses to ensure that the results
fluoromethylheptane sulfonate isomers (Sm-PFOS). According to are representative of the general U.S. population. Due to the

Fig. 1. Participant screening flowchart.

3
S. Zhang and H. Tang Chemosphere 367 (2024) 143606

significantly skewed distribution of the five PFAS compounds, they were Table 1
subjected to natural logarithmic transformation to enhance their Baseline characteristics of participants, (N = 11,137), NHANES, USA.
normality. Overall low-salt or low-sodium diet
After adjusting for age, gender, race/ethnicity, education attain­
Characteristic Overall, N = No, N = Yes, N = p-
ment, smoking status, PIR, and NHANES cycle, we used weighted linear 11,137 10,941 196 (1.8%)a valueb
regression models to calculate the covariate-adjusted geometric means (100%)a (98.2%)a
of the five PFAS concentrations and their 95% confidence intervals to Age (years) 46.0 (33.0, 46.0 (33.0, 58.0 (47.0, <0.001
compare differences between serum PFAS levels among participants 59.0) 59.0) 70.0)
categorized by low-salt or low-sodium diet and salt use habits during Sex ​ ​ ​ 0.2
cooking or food preparation (Kang et al., 2024). In analyzing the dif­ female 5765 (52%) 5665 (52%) 100 (46%) ​
male 5372 (48%) 5276 (48%) 96 (54%)
ferences in serum PFAS levels among participants with varying salt ​
Race/ethnicity ​ ​ ​ 0.12
usage habits during cooking or food preparation, participants who never Non-Hispanic 5119 (70%) 5031 (70%) 88 (67%) ​
added salt to their food were used as the reference group to compare the White
differences between the other three groups relative to this group. Non-Hispanic 2306 (10%) 2260 (10%) 46 (13%) ​
A weighted multivariable linear regression model was used to assess Black
Mexican 1752 (8.0%) 1729 (8.1%) 23 (4.5%)
the association between serum PFAS concentrations and a low-salt or
​
American
low-sodium diet. We constructed three models. Model I included only Other/multiracial 1023 (6.5%) 1002 (6.5%) 21 (10%) ​
age and gender. Model II added race and ethnicity, education level, Other Hispanic 937 (5.1%) 919 (5.1%) 18 (5.2%) ​
smoking status, and poverty income ratio. Model III was further adjusted Smoking status ​ ​ ​ 0.005
Never smoker 6115 (55%) 6027 (55%) 88 (45%)
for survey cycles based on Model II. To assess whether the method of salt ​
Former smoker 2758 (24%) 2679 (24%) 79 (38%) ​
usage during cooking or food preparation influences PFAS serum con­ Current smoker 2264 (21%) 2235 (21%) 29 (16%) ​
centrations, we employed a weighted multivariable linear regression Education ​ ​ ​ 0.4
model for analysis, using only the fully adjusted model. attainment
Some College or 3315 (31%) 3259 (31%) 56 (32%) ​
AA degree
2.6. Sensitivity analysis High School 2551 (24%) 2507 (24%) 44 (20%) ​
Grad/GED
Firstly, as the consumption of fish and shellfish is a primary source of College Graduate 2526 (29%) 2486 (29%) 40 (28%) ​
human PFAS exposure, we further adjusted for participants’ consump­ or above
9-11th Grade 1590 (10%) 1562 (10%) 28 (12%)
tion of shellfish and fish in the past 30 days and then repeated the
​
Less Than 9th 1155 (5.2%) 1127 (5.1%) 28 (8.8%) ​
aforementioned analysis. Grade
Secondly, to balance the baseline differences between participants PIR ​ ​ ​ >0.9
on low-salt or low-sodium diets and those not on such a diet, propensity ≥2 5878 (66%) 5783 (66%) 95 (66%) ​
1–2 3041 (21%) 2981 (21%) 60 (21%)
score matching (PSM) was employed for a 1:4 matching. We then further ​
<1 2218 (13%) 2177 (13%) 41 (13%)
adjusted for age, gender, race/ethnicity, education attainment, smoking
​
NHANES cycle ​ ​ ​ 0.057
status, PIR, consumption of shellfish and fish in the past 30 days, and 2003–2004 1309 (12%) 1292 (12%) 17 (7.0%) ​
survey cycle in the matched population, and conducted a multivariable 2005–2006 1374 (13%) 1352 (13%) 22 (13%) ​
linear regression analysis to test the robustness of the results (Xu et al., 2007–2008 1521 (12%) 1497 (12%) 24 (8.9%) ​
2009–2010 1620 (12%) 1593 (12%) 27 (13%)
2020).
​
2011–2012 1290 (12%) 1264 (12%) 26 (10%) ​
Third, we also examined the sodium intake of participants across the 2013–2014 1356 (13%) 1335 (13%) 21 (14%) ​
eight consecutive NHANES cycles from 2003 to 2018 (with sodium 2015–2016 1371 (13%) 1335 (13%) 36 (23%) ​
intake calculated as the average from two days of dietary recall data 2017–2018 1296 (13%) 1273 (13%) 23 (10%) ​
(Jackson et al., 2016)) and performed a Spearman correlation analysis a
Median (IQR); n (unweighted) (%).
between sodium intake and the serum concentrations of the five PFAS. b
Wilcoxon rank-sum test for complex survey samples; chi-squared test with
Additionally, for each NHANES survey cycle, we plotted line graphs Rao & Scott’s second-order correction.
using the median concentrations of the five serum PFAS and sodium
intake to observe trends in these indicators over time. than that of the other groups, while PFDA concentrations are relatively
All analyses were performed using R version 4.3.2, considering a low. There is a strong correlation among the five PFAS compounds, with
two-sided p-value <0.05 as statistically significant. The weighted anal­ Spearman correlation coefficients greater than 0.69 between PFOA and
ysis employed the "survey” package, and the "MatchIt” package was used PFOS, PFOA and PFNA, as well as PFDA and PFNA.
for PSM analysis.
3.2. Serum PFAS levels categorized by low-salt or low-sodium diet and
3. Results salt usage habits during cooking or food preparation

3.1. General participant characteristics and serum PFAS profiles The differences in the levels of five serum PFAS, categorized by low-
salt or low-sodium diet as well as salt use habits during cooking or food
Table 1 presents the baseline characteristics of the participants. This preparation, are presented in Table 2. Participants on a low-salt or low-
study included 11,137 U.S. adults aged 20 and older. Of these, 1.8% (N sodium diet demonstrated significantly lower geometric means of the
= 196) reported being on a low-salt or low-sodium diet. The gender and five serum PFAS compared to those not on such a diet. Furthermore,
race/ethnicity composition were similar between the two groups. Par­ participants who occasionally or very often used salt in cooking or food
ticipants on a low-salt or low-sodium diet were significantly older (58.0 preparation had significantly higher serum levels of PFOS, PFDA, and
vs 46.0 years) than those not on such a diet. The proportion of current PFNA than those who never used salt.
smokers was significantly higher among participants not on a low-salt or
low-sodium diet. The differences between the survey cycles of the two
groups were also quite apparent.
Fig. 2 displays the distribution and correlations of five serum PFAS
compounds. The serum concentration of PFOS is significantly higher

4
S. Zhang and H. Tang Chemosphere 367 (2024) 143606

Fig. 2. The distribution and correlations of five serum PFAS compounds. Note: A: violin plot of the distribution of serum concentrations of the five PFAS compounds;
B: correlation between serum concentrations of the five PFAS compounds.‘***’p < 0.001.

3.3. Association between serum PFAS levels and low-salt or low-sodium low-sodium diet and the serum levels of four ln-PFAS compounds,
diet except for PFOA, which approached significance (p = 0.069)
(Supplementary Table 5). Additionally, occasionally or very often add­
Table 3 shows the association between the serum levels of five ln- ing salt during cooking was significantly positively correlated with
PFAS compounds and a low-salt or low-sodium diet within a weighted increased serum levels of PFOA, PFOS, PFDA, and PFNA (Supplementary
multivariable linear regression model. After adjusting for covariates, a Table 6).
significant negative correlation between the low-salt or low-sodium diet As shown in Supplementary Fig. 1, the Spearman correlation analysis
and the serum levels of all five ln-PFAS compounds was observed across indicated that sodium intake was positively correlated with all five
all three models. serum PFAS, although the correlation coefficients were relatively low.
Except for PFOS, the concentrations of the other four serum PFAS were
3.4. Association between serum PFAS levels and salt usage methods significantly associated with sodium intake. As illustrated in Supple­
during cooking mentary Fig. 2, over the eight survey cycles, all five serum PFAS
exhibited a decreasing trend, while sodium intake remained generally
In the fully adjusted model, we explored the association between the stable.
method of salt use during cooking and ln-PFAS levels. The results indi­
cate a significant positive correlation between adding salt occasionally 4. Discussion
or very often during cooking or food preparation and increased serum
levels of PFOA, PFOS, PFDA, and PFNA, compared to participants who To our knowledge, this is the first study to investigate the association
never add salt to their food (Table 4). between a low-salt or low-sodium diet and the risk of PFAS exposure.
Additionally, the study further analyzed the association between the
3.5. Sensitivity analysis methods of salt use during cooking and PFAS exposure risk. The results
indicate that a low-salt or low-sodium diet is significantly associated
In the sensitivity analysis, we further adjusted for the consumption of with reduced exposure risks for five serum PFAS compounds, including
fish and shellfish in the past 30 days, yielding results similar to previous PFOA, PFOS, PFHxS, PFDA, and PFNA. Cooking without salt is also
findings. We still observed a significant negative correlation between a associated with a reduced risk of PFAS exposure in the general popu­
low-salt or low-sodium diet and serum levels of five ln-PFAS compounds lation. The possible explanations for these results might include 1) This
(Supplementary Table 2). Similarly, occasionally or very often adding dietary pattern results in the consumption of foods that contain rela­
salt during cooking was significantly positively correlated with tively lower amounts of PFAS compounds. 2) Adding salt during cooking
increased serum levels of PFOA, PFOS, PFDA, and PFNA (Supplementary may facilitate the release of some PFAS compounds from food or uten­
Table 3). sils. 3) Influenced by other healthy dietary habits associated with low-
As shown in Supplementary Table 4, the baseline differences be­ salt and low-sodium diets, such as vegetarian diets and high-fiber diets
tween the two groups after propensity score matching were not statis­ (Bowman, 2020).
tically significant. The unadjusted differences in serum PFAS levels Dietary patterns have been established to be closely associated with
between the two groups were also not significant. Multivariable disease risk (Hu, 2002). Human intake of PFAS primarily originates from
regression analysis revealed that after adjusting for relevant covariates, food, drinking water, cooking utensils, and food packaging materials
a significant negative correlation was observed between a low-salt or (Sunderland et al., 2019). Variations in dietary patterns may lead to

5
S. Zhang and H. Tang Chemosphere 367 (2024) 143606

differences in PFAS exposure risk among individuals, thereby mediating

Per- and poly-fluoroalkyl substances (PFAS), Perfluorooctanoic acid (PFOA), Perfluorooctane sulfonic acid (PFOS), Perfluorohexane sulfonic acid (PFHxS), Perfluorodecanoic acid (PFDA), Perfluorononanoic acid (PFNA).
Reference
the relationship with the risk of disease onset. A cross-sectional study of

p-value
the elderly population in Sweden found that the WHO-recommended

0.023

0.004

0.003
0.55
dietary pattern is positively correlated with PFHxS levels, while a
low-carbohydrate, high-protein diet is positively correlated with PFOS,
PFNA, PFDA, and perfluoroundecanoic acid (PFUnDA) levels. Addi­
tionally, the Mediterranean diet is positively associated with PFOA,
(0.817,0.886)

(0.649,0.836)

(0.737,0.831)

(0.769,0.851)

(0.821,0.897)

(0.824,0.914)
perfluorooctane sulfonamide (PFOSA), PFHxS, PFNA, PFDA, and
PFNA1

PFUnDA levels (Sjogren et al., 2016). Another study focusing on chil­

0.851

0.737

0.783

0.809

0.858

0.868
dren’s dietary patterns and PFAS exposure risk found that diets high in
packaged foods and fish can lead to increased plasma concentrations of
Reference

all PFAS, with notable increases in MeFOSAA and PFOS (Seshasayee

<0.001
p-value

0.334
et al., 2021). Diets rich in animal products, such as fish, poultry, and red
0.03

0.01 meat, are associated with increased prenatal PFAS levels (Eick et al.,
2023). Previous reports have indicated that vegetarians have signifi­
0.22(0.207,0.233)

cantly lower serum PFAS levels compared to omnivores (Menzel et al.,

(0.234,0.252)

(0.191,0.239)

(0.218,0.241)

(0.229,0.248)

(0.243,0.266)
2021). The intake of high-fiber foods in the diet is associated with lower
serum PFAS concentrations (Mw et al., 2021). Additionally, it has been
PFDA1

reported that individuals consuming foods like pizza and salty snacks
0.243

0.214

0.229

0.238

0.254

have significantly higher levels of PFOA and PFOS in their serum sam­
Comparison of differences in serum levels of five PFAS according to low-salt or low-sodium diets and salt use habits during cooking or preparing foods.

ples compared to those who do not consume these foods (Park et al.,
Reference

2019). However, in the sensitivity analysis, we observed that sodium

p-value

0.021

0.962

intake remained relatively stable across the eight consecutive cycles,

0.75

which is consistent with previous studies (Clarke et al., 2021). No

1

similar temporal trends were noted between sodium intake and the
1.38(1.317,1.446)

concentrations of the five serum PFAS. This suggests that, on a broader

(1.349,1.456)

(1.021,1.374)

(1.303,1.542)

(1.341,1.496)

(1.338,1.463)

timescale, the decline in serum PFAS levels among the general popula­
tion in the United States is more likely attributable to government re­
PFHxS1

1.401

1.184

1.417

1.416

1.399

strictions on the production and use of PFAS substances rather than

changes in salt intake.
1 Models were adjusted for age, gender, race/ethnicity, education attainment, smoking status, PIR, and NHANES cycle.

A low-salt or low-sodium diet may indirectly result in reduced water

Reference

intake. PFAS are water-soluble, and contaminated water sources are

p-value

0.037

0.219

0.028

0.048

another significant source of human PFAS exposure (Domingo and

Nadal, 2019). Research indicates that the intake of PFAS from drinking
water ranks second only to that from food (Kim et al., 2019), with serum
PFAS levels positively correlated with public water supply systems and
(8.445,8.991)

(6.152,8.703)

(7.701,8.575)

(8.146,8.984)

(8.423,9.139)

(8.389,9.079)

tap water consumption (Nair et al., 2021). The primary sources of PFAS
water contamination include industrial production, commercial air­
PFOS1

8.714

7.317

8.126

8.555

8.774

8.727

ports, military bases, and the use of PFAS-containing aqueous

film-forming foam (AFFF) during firefighting activities (Rosen et al.,
2022; Anderson et al., 2016; McGuire et al., 2014). In the United States,
Reference

residents primarily obtain drinking water from surface water and

p-value

0.046

0.378

0.098

0.254

groundwater. It is estimated that nearly 80 million people nationwide

have drinking water containing PFOA and PFOS at concentrations
exceeding 10 ng/L, with over 200 million potentially using water
2.48(2.405,2.558)

contaminated with these substances (Andrews and Naidenko, 2020). A

PFAS (ng/ml)

(2.398,2.525)

(1.861,2.466)

(2.229,2.479)

(2.348,2.546)

(2.367,2.538)

survey investigating PFAS in shallow groundwater consumed by resi­

dents in Wisconsin revealed that nearly 71% of samples from 450
PFOA1

2.461

2.142

2.351

2.445

2.451

households contained at least one PFAS compound (Silver et al., 2023).

From 1999 to 2017, the number of sites contaminated with PFAS across
Note: Geometric means and 95% confidence intervals.

the United States has shown an increasing trend, further exacerbating

Very Often, (N = 4650)

water pollution issues (Sunderland et al., 2019). In addition to direct

Rarely, (N = 1969)

Occasionally, (N =

contamination of water sources, the treatment of PFAS in wastewater

Never, (N = 899)
No, (N = 10941)

Yes, (N = 196)

also warrants attention. Studies have shown that PFAS concentrations

are notably high in contaminated wastewater and sludge in the U.S.
Once PFAS, including PFOA and PFOS, enter water systems, they are
3461)

typically not removed by conventional wastewater treatment methods

such as biodegradation or flocculation. In fact, their concentrations in
treated wastewater often become higher, potentially due to the con­
Salt used in preparation
Low salt or low sodium

version of some precursor substances into PFAS during biological

treatment processes (Arvaniti and Stasinakis, 2015).
The majority of individuals, including both adults and children, have
detectable PFAS exposure (Kotlarz et al., 2020). Once absorbed by the
human body, PFAS demonstrates significant bioaccumulative proper­
Table 2

diet

ties. PFOA has a reported average half-life of 2.7 years, while the
half-life of PFHxS is twice that length (Li et al., 2018). Aside from fish,

6
S. Zhang and H. Tang Chemosphere 367 (2024) 143606

Table 3
Association between serum levels of five PFAS and low-salt or low-sodium diets.
PFAS (ng/ml) Low-salt or low-sodium diet Model I Model II Model III

β (95%CI) p-value β (95%CI) p-value β (95%CI) p-value

PFOA No Reference ​ Reference ​ Reference ​

Yes − 0.215(-0.364,-0.066) 0.006 − 0.196(-0.343,-0.049) 0.01 − 0.139(-0.275,-0.002) 0.048
PFOS No Reference ​ Reference ​ Reference ​
Yes − 0.263(-0.455,-0.072) 0.008 − 0.259(-0.448,-0.07) 0.009 − 0.175(-0.339,-0.01) 0.04
PFHxS No Reference ​ Reference ​ Reference ​
Yes − 0.215(-0.358,-0.071) 0.004 − 0.194(-0.341,-0.048) 0.01 − 0.168(-0.311,-0.025) 0.023
PFDA No Reference ​ Reference ​ Reference ​
Yes − 0.156(-0.289,-0.023) 0.023 − 0.175(-0.305,-0.044) 0.01 − 0.128(-0.244,-0.012) 0.032
PFNA No Reference ​ Reference ​ Reference ​
Yes − 0.169(-0.308,-0.031) 0.018 − 0.176(-0.315,-0.037) 0.014 − 0.144(-0.268,-0.02) 0.025

Note：Model I adjusts for age, gender; Model II adjusts for age, gender, race/ethnicity, education attainment, smoking status, PIR; Model III adjusts for age, gender,
race/ethnicity, education attainment, smoking status, PIR, and NHANES cycle. Per- and poly-fluoroalkyl substances (PFAS), Perfluorooctanoic acid (PFOA), Per­
fluorooctane sulfonic acid (PFOS), Perfluorohexane sulfonic acid (PFHxS), Perfluorodecanoic acid (PFDA), Perfluorononanoic acid (PFNA).

sodium chloride levels, low pH, or contain emulsifiers, the migration of

Table 4
PFAS from FCMs to the food increases (Ramírez Carnero et al., 2021). An
Association between salt usage during cooking or preparing foods and serum
increase in salinity reduces the repulsive forces between PFAS at the
levels of five PFAS.
air-liquid interface, making it easier for PFAS to migrate or transfer from
PFAS (ng/ml) Salt used in preparation β (95%CI) p-value contaminated utensils or liquids to food (Morrison et al., 2023; Eze et al.,
PFOA Never Reference ​ 2024). A low-salt or low-sodium diet and salt-free cooking can result in
Rarely 0.039(-0.021,0.1) 0.207 reduced salinity in food, which may help decrease the migration of
Occasionally 0.054(0.001,0.106) 0.048
PFAS.
Very Often 0.042(-0.012,0.096) 0.133
PFOS Never Reference ​ The results of this study found a negative correlation between low-
Rarely 0.051(-0.011,0.114) 0.11 salt or low-sodium dietary patterns and serum PFAS concentrations,
Occasionally 0.077(0.017,0.136) 0.014 with a higher proportion of males and older participants compared to
Very Often 0.071(0.011,0.132) 0.023 those not following a low-salt or low-sodium diet. The adoption of a low-
PFHxS Never Reference
salt or low-sodium diet may be primarily driven by health concerns,
​
Rarely − 0.001(-0.09,0.088) 0.988
Occasionally − 0.013(-0.101,0.075) 0.774 such as hypertension, overweight, or obesity. However, some studies
Very Often − 0.027(-0.108,0.054) 0.519 have reported that serum PFAS concentrations may be higher in elderly
PFDA Never Reference ​ and male populations compared to other groups (Schultz et al., 2023).
Rarely 0.043(-0.019,0.104) 0.175
We suggest that the reasons for this disparity may be related not only to
Occasionally 0.081(0.026,0.137) 0.005
Very Often 0.146(0.085,0.208) <0.001 the low-salt or low-sodium dietary patterns adopted by these partici­
PFNA Never Reference ​ pants but also to their engagement in healthier lifestyle choices. This
Rarely 0.033(-0.031,0.098) 0.311 group may consume more healthful foods and use healthier consumer
Occasionally 0.092(0.035,0.15) 0.002 products, thereby reducing their exposure to PFAS.
Very Often 0.103(0.041,0.165) 0.002
This study has several unique advantages. Firstly, NHANES is a na­
Note: Models were adjusted for age, gender, race/ethnicity, education attain­ tionally representative survey. Second, beyond examining the associa­
ment, smoking status, PIR, and NHANES cycle. tion between a low-salt or low-sodium diet and PFAS exposure risk, it
Per- and poly-fluoroalkyl substances (PFAS), Perfluorooctanoic acid (PFOA), further explores the association between salt usage habits in the general
Perfluorooctane sulfonic acid (PFOS), Perfluorohexane sulfonic acid (PFHxS),
U.S. population and PFAS exposure risk. Third, the study conducted
Perfluorodecanoic acid (PFDA), Perfluorononanoic acid (PFNA).
multiple sensitivity analyses, employed propensity score matching, and
adjusted for the potential impacts of consuming fish and shellfish.
seafood, and shellfish (EFSA Panel on Contaminants in the Food Chain Fourth, although NHANES is a cross-sectional design, which is less
(CONTAM) et al., 2018), common foods such as meat, milk, vegetables, conducive to causal inference, it is unlikely that a reverse causal asso­
and eggs are also impacted by PFAS contamination (Zhang et al., 2010; ciation will hold for this study. It is unlikely that reductions in PFAS
Macheka et al., 2021; Brown et al., 2020; Zafeiraki et al., 2016). The would influence changes in dietary habits. Furthermore, participants’
effects of different cooking methods on PFAS levels in food vary during dietary habits and salt usage habits are long-term observational in­
processing. Steaming can increase PFOS content in fish, but reduce dicators, and since PFAS are known to accumulate in biological organ­
concentrations of other shorter-chain PFAS (Hu et al., 2020). For sea­ isms, their levels more accurately reflect long-term exposure. It is
food, baking raises the total concentration of PFAS, whereas steaming important to note that while a significant association was identified, the
and frying lower it (Chen et al., 2023). Additionally, evidence suggests causality of this association remains to be determined. Because the po­
that cooked vegetables have lower PFAS levels than raw vegetables, tential influence of unobserved confounders or biases on the results
likely due to volatilization losses during cooking (Liu et al., 2023). PFAS cannot be excluded. However, this study has some limitations. First,
can become trapped in protein aggregates due to heat treatment, reports of special diets are based on participants’ self-reports, which
reducing their bioavailability; steaming is more effective than frying in may introduce potential bias. Second, although adjustments were made
achieving this effect (Zhao et al., 2023). Human salt intake primarily for potential confounding factors, not all factors influencing PFAS
comes from diet. PFAS are effective ionic surfactants, heavily concen­ exposure could be excluded. Third, the data originated from the general
trated at material surfaces and interfaces, with changes in salinity U.S. population, and further validation is needed to determine its
significantly affecting their adsorption (Le et al., 2022). Additionally, representativeness for other populations. Fourth, the test positivity rate
another significant source of PFAS in food is food contact materials of PFDA was low, at 73.22%. When the distribution of values below the
(FCMs), which are widely used for their fat resistance, moisture barrier, LLOD is not centered around the replacement value (LLOD/√2), this
and non-stick properties. When foods have a high-fat content, elevated may result in biased statistical results and/or reduced statistical efficacy

7
S. Zhang and H. Tang Chemosphere 367 (2024) 143606

(Richardson and Ciampi, 2003; Nie et al., 2010). Finally, due to Andrews, D.Q., Naidenko, O.V., 2020. Population-wide exposure to per- and
polyfluoroalkyl substances from drinking water in the United States. Environ. Sci.
restricted access to geographic coding data for sampling locations, we
Technol. Lett. 7, 931–936.
are unable to obtain this information about the participants. However, Arvaniti, O.S., Stasinakis, A.S., 2015. Review on the occurrence, fate and removal of
the potential impact of dietary patterns and PFAS exposure levels perfluorinated compounds during wastewater treatment. Sci. Total Environ.
resulting from geographic differences should not be overlooked (Nair 524–525, 81–92.
Averina, M., Brox, J., Huber, S., Furberg, A.-S., 2021. Exposure to perfluoroalkyl
et al., 2021; Schultz et al., 2023; McCullough et al., 2022; Liddie et al., substances (PFAS) and dyslipidemia, hypertension and obesity in adolescents. The
2023). Fit Futures study. Environ. Res. 195, 110740.
Bowman, S.A., 2020. A vegetarian-style dietary pattern is associated with lower energy,
saturated fat, and sodium intakes; and higher whole grains, legumes, nuts, and soy
5. Conclusion intakes by adults: national health and nutrition examination surveys 2013-2016.
Nutrients 12, 2668.
In summary, after an in-depth analysis of large-sample data from Brown, J.B., Conder, J.M., Arblaster, J.A., Higgins, C.P., 2020. Assessing human health
risks from per- and polyfluoroalkyl substance (PFAS)-Impacted vegetable
NHANES, this study found that a low-salt or low-sodium diet is closely consumption: a tiered modeling approach. Environ. Sci. Technol. 54, 15202–15214.
associated with reduced exposure risks to five major serum PFAS com­ Buck, R.C., et al., 2011. Perfluoroalkyl and polyfluoroalkyl substances in the
pounds, including PFOA, PFOS, PFHxS, PFDA, and PFNA. Additionally, environment: terminology, classification, and origins. Integr Environ Assess Manag
7, 513–541.
the results suggest that salt-free cooking may help reduce PFAS expo­ Calafat, A.M., Wong, L.-Y., Kuklenyik, Z., Reidy, J.A., Needham, L.L., 2007.
sure. Although this study provides new insights into reducing PFAS Polyfluoroalkyl chemicals in the U.S. Population: data from the national health and
exposure among populations, further evidence is needed to confirm nutrition examination survey (NHANES) 2003-2004 and comparisons with NHANES
1999-2000. Environ. Health Perspect. 115, 1596–1602.
these findings and to inform the development of policies and guidelines Chen, Z., et al., 2023. Bioaccumulation and risk mitigation of legacy and novel
for public health agencies. perfluoroalkyl substances in seafood: insights from trophic transfer and cooking
method. Environ. Int. 177, 108023.
Christensen, K.Y., et al., 2017. Perfluoroalkyl substances and fish consumption. Environ.
CRediT authorship contribution statement Res. 154, 145–151.
Clarke, L.S., et al., 2021. Temporal trends in dietary sodium intake among adults aged
Shuai Zhang: Writing – review & editing, Writing – original draft, ≥19 Years - United States, 2003-2016. MMWR Morb. Mortal. Wkly. Rep. 70,
1478–1482.
Visualization, Resources, Project administration, Methodology, Formal Domingo, J.L., Nadal, M., 2019. Human exposure to per- and polyfluoroalkyl substances
analysis, Conceptualization. Hanhan Tang: Writing – review & editing, (PFAS) through drinking water: a review of the recent scientific literature. Environ.
Writing – original draft, Visualization, Methodology, Formal analysis, Res. 177, 108648.
Efsa Panel on Contaminants in the Food Chain (Contam) et al., 2018. Risk to human
Conceptualization.
health related to the presence of perfluorooctane sulfonic acid and perfluorooctanoic
acid in food. EFSA J. 16, e05194.
Ethics statement Eick, S.M., et al., 2023. Dietary predictors of prenatal per- and poly-fluoroalkyl
substances exposure. J. Expo. Sci. Environ. Epidemiol. 33, 32–39.
Elevated levels of serum per- and poly-fluoroalkyl substances (PFAS) in contact lens users
NHANES study was conducted under the authorization of the Na­ of U.S. young adults. Chemosphere 350, 2024, 141134.
tional Center for Health Statistics (NCHS) Ethics Review Committee, and Eze, C.G., et al., 2024. Emerging contaminants in food matrices: an overview of the
occurrence, pathways, impacts and detection techniques of per- and polyfluoroalkyl
all participants provided informed consent. The patients/participants
substances. Toxicol Rep 12, 436–447.
provided their written informed consent to participate in this study. Green, M.P., Harvey, A.J., Finger, B.J., Tarulli, G.A., 2021. Endocrine disrupting
chemicals: impacts on human fertility and fecundity during the peri-conception
period. Environ. Res. 194, 110694.
Funding Hu, F.B., 2002. Dietary pattern analysis: a new direction in nutritional epidemiology.
Curr. Opin. Lipidol. 13, 3–9.
This research received no external funding. Hu, Y., et al., 2020. Cooking methods affect the intake of per- and polyfluoroalkyl
substances (PFASs) from grass carp. Ecotoxicol. Environ. Saf. 203, 111003.
Jackson, S.L., King, S.M.C., Zhao, L., Cogswell, M.E., 2016. Prevalence of excess sodium
Declaration of competing interest intake in the United States - NHANES, 2009-2012. MMWR Morb. Mortal. Wkly. Rep.
64, 1393–1397.
Kim, D.-H., Lee, J.-H., Oh, J.-E., 2019. Assessment of individual-based perfluoroalkly
The authors declare that they have no known competing financial substances exposure by multiple human exposure sources. J. Hazard Mater. 365,
interests or personal relationships that could have appeared to influence 26–33.
Kotlarz, N., et al., 2020. Measurement of novel, drinking water-associated PFAS in blood
the work reported in this paper. from adults and children in wilmington, North Carolina. Environ. Health Perspect.
128, 77005.
Acknowledgements Le, S.-T., Gao, Y., Kibbey, T.C.G., O’Carroll, D.M., 2022. Calculating PFAS interfacial
adsorption as a function of salt concentration using model parameters determined
from chemical structure. Sci. Total Environ. 848, 157663.
Thanks to the editors and reviewers for their precious time and Li, Y., et al., 2018. Half-lives of PFOS, PFHxS and PFOA after end of exposure to
constructive comments. The authors thank the participants and staff of contaminated drinking water. Occup. Environ. Med. 75, 46–51.
Liddie, J.M., Schaider, L.A., Sunderland, E.M., 2023. Sociodemographic factors are
NHANES. associated with the abundance of PFAS sources and detection in U.S. Community
water systems. Environ. Sci. Technol. 57, 7902–7912.
Appendix A. Supplementary data Liu, S., et al., 2023. Transport and transformation of perfluoroalkyl acids, isomer profiles,
novel alternatives and unknown precursors from factories to dinner plates in China:
new insights into crop bioaccumulation prediction and risk assessment. Environ. Int.
Supplementary data to this article can be found online at https://doi. 172, 107795.
org/10.1016/j.chemosphere.2024.143606. Macheka, L.R., Olowoyo, J.O., Mugivhisa, L.L., Abafe, O.A., 2021. Determination and
assessment of human dietary intake of per and polyfluoroalkyl substances in retail
dairy milk and infant formula from South Africa. Sci. Total Environ. 755, 142697.
Data availability McCullough, M.L., et al., 2022. Association of socioeconomic and geographic factors with
diet quality in US adults. JAMA Netw. Open 5, e2216406.
McGuire, M.E., et al., 2014. Evidence of remediation-induced alteration of subsurface
Data will be made available on request.
poly- and perfluoroalkyl substance distribution at a former firefighter training area.
Environ. Sci. Technol. 48, 6644–6652.
References Menzel, J., et al., 2021. Internal exposure to perfluoroalkyl substances (PFAS) in vegans
and omnivores. Int. J. Hyg Environ. Health 237, 113808.
Miranda, D.A., et al., 2021. Bioaccumulation of Per- and polyfluoroalkyl substances
Anderson, R.H., Long, G.C., Porter, R.C., Anderson, J.K., 2016. Occurrence of select
(PFASs) in a tropical estuarine food web. Sci. Total Environ. 754, 142146.
perfluoroalkyl substances at U.S. Air Force aqueous film-forming foam release sites
Morrison, A.L., et al., 2023. Impact of salinity and temperature on removal of PFAS
other than fire-training areas: field-validation of critical fate and transport
species from water by aeration in the absence of additional surfactants: a novel
properties. Chemosphere 150, 678–685.

8
S. Zhang and H. Tang Chemosphere 367 (2024) 143606

application of green chemistry using adsorptive bubble fractionation. Ind. Eng. Seshasayee, S.M., et al., 2021. Dietary patterns and PFAS plasma concentrations in
Chem. Res. 62, 5635–5645. childhood: project Viva, USA. Environ. Int. 151, 106415.
Mw, D., Dr, K., Mp, L., 2021. The concentration of several perfluoroalkyl acids in serum Silver, M., et al., 2023. Prevalence and source tracing of PFAS in shallow groundwater
appears to be reduced by dietary fiber. Environ. Int. 146. used for drinking water in Wisconsin, USA. Environ. Sci. Technol. 57, 17415–17426.
Nair, A.S., Ma, Z.-Q., Watkins, S.M., Wood, S.S., 2021. Demographic and exposure Sjogren, P., et al., 2016. Circulating levels of perfluoroalkyl substances are associated
characteristics as predictors of serum per- and polyfluoroalkyl substances (PFASs) with dietary patterns - a cross sectional study in elderly Swedish men and women.
levels - a community-level biomonitoring project in Pennsylvania. Int. J. Hyg Environ. Res. 150, 59–65.
Environ. Health 231, 113631. Steenland, K., Winquist, A., 2021. PFAS and cancer, a scoping review of the
Nie, L., et al., 2010. Linear regression with an independent variable subject to a detection epidemiologic evidence. Environ. Res. 194, 110690.
limit. Epidemiology 21, S17–S24. Sun, Q., et al., 2023. Cooking methods effectively alter perfluoroalkyl substances and
Onteeru, M., et al., 2022. Association between fish oil supplements use and serum per- nutrients in cultured and wild bullfrogs. J. Hazard Mater. 445, 130555.
and polyfluoroalkyl substances (PFAS): results from the National Health and Sunderland, E.M., et al., 2019. A review of the pathways of human exposure to poly- and
Nutrition Examination Survey. Environ. Res. 215, 114205. perfluoroalkyl substances (PFASs) and present understanding of health effects.
Pan, K., et al., 2023. The relationship between perfluoroalkyl substances and J. Expo. Sci. Environ. Epidemiol. 29, 131–147.
hypertension: a systematic review and meta-analysis. Environ. Res. 232, 116362. Susmann, H.P., Schaider, L.A., Rodgers, K.M., Rudel, R.A., 2019. Dietary habits related to
Park, S.K., Peng, Q., Ding, N., Mukherjee, B., Harlow, S.D., 2019. Determinants of per- food packaging and population exposure to PFASs. Environ. Health Perspect. 127,
and polyfluoroalkyl substances (PFAS) in midlife women: evidence of racial/ethnic 107003.
and geographic differences in PFAS exposure. Environ. Res. 175, 186–199. Taylor, M.D., et al., 2019. Do conventional cooking methods alter concentrations of per-
Pelch, K.E., Reade, A., Wolffe, T.A.M., Kwiatkowski, C.F., 2019. PFAS health effects and polyfluoroalkyl substances (PFASs) in seafood? Food Chem. Toxicol. 127,
database: protocol for a systematic evidence map. Environ. Int. 130, 104851. 280–287.
Qiao, J.-C., et al., 2024. Associations of per- and polyfluoroalkyl substances (PFAS) and van Gerwen, M., et al., 2023. Per- and polyfluoroalkyl substances (PFAS) exposure and
their mixture with risk of rheumatoid arthritis in the U.S. adult population. Environ. thyroid cancer risk. EBioMedicine 97, 104831.
Health 23, 38. Vendl, C., et al., 2022. Thermal processing reduces PFAS concentrations in blue food - a
Ramírez Carnero, A., et al., 2021. Presence of perfluoroalkyl and polyfluoroalkyl systematic review and meta-analysis. Environ Pollut 304, 119081.
substances (PFAS) in food contact materials (FCM) and its migration to food. Foods Xu, K.Y., Hartz, S.M., Borodovsky, J.T., Bierut, L.J., Grucza, R.A., 2020. Association
10, 1443. between benzodiazepine use with or without opioid use and all-cause mortality in
Richardson, D.B., Ciampi, A., 2003. Effects of exposure measurement error when an the United States, 1999-2015. JAMA Netw. Open 3, e2028557.
exposure variable is constrained by a lower limit. Am. J. Epidemiol. 157, 355–363. Zafeiraki, E., et al., 2016. Perfluoroalkylated substances (PFASs) in home and
Rosen, E.M., et al., 2022. Drinking water-associated PFAS and fluoroethers and lipid commercially produced chicken eggs from The Netherlands and Greece.
outcomes in the GenX exposure study. Environ. Health Perspect. 130, 97002. Chemosphere 144, 2106–2112.
Roth, K., Petriello, M.C., 2022. Exposure to per- and polyfluoroalkyl substances (PFAS) Zhang, T., et al., 2010. Perfluorochemicals in meat, eggs and indoor dust in China:
and type 2 diabetes risk. Front. Endocrinol. 13, 965384. assessment of sources and pathways of human exposure to perfluorochemicals.
Schultz, A.A., et al., 2023. Biomonitoring of perfluoroalkyl and polyfluoroalkyl Environ. Sci. Technol. 44, 3572–3579.
substances (PFAS) from the survey of the health of Wisconsin (SHOW) 2014-2016 Zhao, A., et al., 2023. Tracing the bioaccessibility of per- and polyfluoroalkyl substances
and comparison with the national health and nutrition examination survey in fish during cooking treatment. J. Agric. Food Chem. 71, 19066–19077.
(NHANES). J. Exposure Sci. Environ. Epidemiol. 33, 766–777. Zhu, Y., Ro, A., Bartell, S.M., 2021. Household low pile carpet usage was associated with
increased serum PFAS concentrations in 2005-2006. Environ. Res. 195, 110758.