Per- and Polyfluoroalkyl Substances in Eutopic Endometrium Tissue and Risk of Endometriosis: Findings from the Investigating Mixtures of Pollutants and Endometriosis in Tissue (IMPLANT) Study

Among 434 participants, 181 were diagnosed with incident endometriosis and 253 without endometriosis ( Table 1 ). The mean age for all participants was 33 years. Race and ethnicity were similar across endometriosis diagnosis, with 75% of those with endometriosis identifying as non-Hispanic white, as compared to 76.3% of those without endometriosis. The most common primary reason for surgery was pelvic pain ( n = 197 ). When stratified by endometriosis subtype, 15.1% of participants with endometriosis had deep infiltrating endometriosis (DE), 35.1% had superficial endometriosis (SE), and 12.6% had ovarian endometrioma (OE). Of the participants with endometriosis, 27.6% had stages 3 and 4, and 72.4% had stages 1 and 2 (Table S1). Slightly less than half of participants were parous and breastfed (45.4%), while 13.2% were parous but did not breastfeed. Ectopic tissue samples were only sufficient to obtain PFAS values for four individuals. Participant characteristics in the Investigating Mixtures of Pollutants and Endometriosis in Tissue (IMPLANT) study. 33 ± 7 32 ± 7 33 ± 7 Note: AAPI/ANAI, Asian American Pacific Islander/Alaska Native/American Indian; BMI, body mass index; SD, standard deviation. Serum cotinine ≥ 10  ng / mL indicating current smoking. Descriptive statistics of per- and polyfluoroalkyl substances in the eutopic endometrium of participants from the Investigating Mixtures of Pollutants and Endometriosis in Tissue (IMPLANT) study ( n = 434 ). Note: —, no data; IQR, interquartile range; LOD, limit of detection; PFAS, per- and polyfluoroalkyl substances; PFDA, perfluorodecanoic acid; PFDoDA, perfluorododecanoic acid; PFHpA, perfluoroheptanoic acid; PFHxS, perfluorohexane sulfonate; PFNA, perfluorononanoic acid; PFOA, perfluorooctanoic acid; PFOS, perfluorooctanesulfonic acid; PFOSA, perfluorooctane sulfonamide; PFUnDA, perfluoroundecanoic acid. Not analyzed due to large percent below the limit of detection. Five PFAS were widely detected in eutopic endometrium tissue: PFOS ( 100 % > LOD ), PFOA ( 99.5 % > LOD ), PFHxS ( 98.6 % > LOD ), PFNA ( 94.2 % > LOD ), and PFOSA ( 80.4 % > LOD ) ( Table 2 ). PFAS with > 50 % below the LOD were not included in further analyses: PFDA ( 41.0 % > LOD ), PFUnDA ( 27.4 % > LOD ), PFHpA ( 23.7 % > LOD ), and PFDoDA ( 15.2 % > LOD ). The strongest correlations were found between PFOA and PFNA ( ρ = 0.78 ), followed by PFOA and PFOS ( ρ = 0.73 ) and PFNA and PFOS ( ρ = 0.60 ). When evaluating correlations by endometriosis diagnosis, findings were largely similar. However, PFOSA and PFOS were more strongly correlated among those with endometriosis ( ρ = 0.42 ) compared to those without ( ρ = 0.20 ) (Table S2). PFOS had the highest median concentration [ 6.58  ng / g ; interquartile range (IQR): 6.44  ng / g ] in eutopic endometrium tissue, while PFOSA, with a median of 0.12  ng / g and an IQR of 0.18, had the lowest concentration ( Table 2 ). Median concentrations of PFAS were similar across endometriosis diagnosis ( Table 2 ). Only four participants’ ectopic endometrial tissue could be analyzed for PFAS concentrations. PFAS in ectopic tissue was higher than in eutopic tissue for PFHxS, PFOA, PFNA, PFDA, and PFUnDA (Table S3). When stratifying by ASRM staging, stages 1 and 2 had slightly lower PFAS concentrations compared to stages 3 and 4, except PFNA, which was similar across ASRM stages ( Table 3 ). For example, median PFHxS concentrations were 0.56  ng / g (IQR: 0.73  ng / g ) for stages 1 and 2 and 0.67  ng / g (IQR: 0.87  ng / g ) for stages 3 and 4, representing an approximate 16% increase in PFHxS levels in stages 3 and 4. Few differences were noted by endometriosis typology. Individuals with deep infiltrating endometriosis had slightly lower eutopic endometrium concentrations of PFOA (median: 1.71  ng / g , IQR: 1.61  ng / g ) and PFOS (median: 6.38  ng / g , IQR: 5.81  ng / g ) compared to those with superficial endometriosis (PFOA: 1.91  ng / g , IQR: 1.93  ng / g ; PFOS: 6.86  ng / g , IQR: 7.39  ng / g ), corresponding to an 11% decrease in PFOA and a 7% decrease in PFOS levels. PFNA levels were also lower in the deep group compared to the superficial group, with a decrease of ∼ 7 % . None of the five PFAS were independently associated with a risk of endometriosis for both unadjusted and adjusted models ( Table 4 ). For instance, PFOA had an adjusted risk ratio of 1.01 (95% CI: 0.90, 1.15) and PFOS had a risk ratio of 0.98 (95% CI: 0.85, 1.13). After adjusting for the other PFAS in co-exposure models, results were largely unchanged, although PFOA concentration in eutopic endometrium tissue was associated with a slightly higher risk of incident endometriosis diagnosis, with a wide confidence interval, likely driven by correlations with PFNA and PFOS ( Table 4 ). In contrast, when we evaluated those with endometriosis by ASRM staging, several PFAS were associated ( Table 5 ). There was a greater risk of ASRM stage 3 and 4 endometriosis compared with stage 1 and 2 for PFHxS ( RR = 1.37 ; 95% CI: 1.12, 1.68), PFOS ( RR = 1.36 ; 95% CI: 1.02, 1.81), and PFOSA ( RR = 1.23 ; 95% CI: 1.09, 1.40) after adjusting for age, BMI, race/ethnicity, smoking, and site ( Table 5 ). In contrast, there were no clear associations of PFAS with endometriosis subtypes. For example, PFOS with deep vs. superficial endometriosis risk ratio was 0.91 (95% CI: 0.69, 1.19) ( Table 5 ). Median and interquartile range of PFAS concentrations (ng/g) by American Society for Reproductive Medicine (ASRM) staging and endometriosis subtype among participants in the Investigating Mixtures of Pollutants and Endometriosis in Tissue (IMPLANT) study ( n = 434 ). Note: PFAS, per- and polyfluoroalkyl substances; PFHxS, perfluorohexane sulfonate; PFNA, perfluorononanoic acid; PFOA, perfluorooctanoic acid; PFOS, perfluorooctanesulfonic acid; PFOSA, perfluorooctane sulfonamide. Modified Poisson regression models of PFAS and endometriosis diagnosis among participants in the Investigating Mixtures of Pollutants and Endometriosis in Tissue (IMPLANT) study ( n = 434 ). Note: CI, confidence interval; multiple, co-exposure model adjusted for other PFAS and BMI, race/ethnicity, site, and smoking; PFAS, per- and polyfluoroalkyl substances; PFHxS, perfluorohexane sulfonate; PFNA, perfluorononanoic acid; PFOA, perfluorooctanoic acid; PFOS, perfluorooctanesulfonic acid; PFOSA, perfluorooctane sulfonamide; RR, risk ratio. Adjusted for age, BMI, race/ethnicity, site, and smoking. Adjusted modified Poisson regression models for eutopic PFAS concentrations and American Society for Reproductive Medicine (ASRM) staging of endometriosis and endometriosis subtype among participants in the Investigating Mixtures of Pollutants and Endometriosis in Tissue (IMPLANT) study ( n = 434 ). Note: Adjusted for age, BMI, race/ethnicity, site, and smoking. CI, confidence interval; PFAS, per- and polyfluoroalkyl substances; PFHxS, perfluorohexane sulfonate; PFNA, perfluorononanoic acid; PFOA, perfluorooctanoic acid; PFOS, perfluorooctanesulfonic acid; PFOSA, perfluorooctane sulfonamide; RR, risk ratio. The PFAS mixture analysis using BKMR showed results similar to the main analysis, indicating a less noticeable association with incident endometriosis (Figure S2). We observed a positive association between ASRM stage and the eutopic PFAS mixture (Figure S3), suggesting that higher levels of PFAS exposure are associated with more advanced ASRM stages. Furthermore, in this study, the trend appears to be nonlinear, suggesting that the effect of PFAS on the ASRM stage may vary at different exposure levels, though confidence intervals were wide. We did not find strong evidence of nonlinearity for individual PFAS while holding other PFAS at their median, as nonlinear trends were associated with wide confidence interval (Figures S4 and S5). We also did not find evidence of interactions: the associations of each PFAS were similar by quantiles of other PFAS (Figures S6 and S7). Our results were robust to imputing concentrations < LOD with LOD / √ 2 ; results were comparable to single imputation at the LOD, at zero, and multiply imputing values (Figure S6). PFAS dichotomized at the LOD were not associated with endometriosis (Figure S7). When adjusting for breastfeeding and parity, the estimated RR for PFAS and endometriosis did not change (Table S4). Consistent with our overall findings, there were no clear associations between PFAS concentrations and endometriosis when we restricted by a surgical indication of pelvic pain (Table S5). Overall, the Pearson correlation coefficients for each PFAS measured in serum and eutopic endometrial tissue were moderate and positive (ranging from 0.43 to 0.49). For example, PFHxS in serum correlated with PFHxS in eutopic tissue ( ρ = 0.48 ), while serum PFNA, PFOA, and PFOS also exhibited similarly moderate correlations ( ρ = 0.49 , 0.43, and 0.46, respectively). We were not able to determine correlations for PFOSA because, in serum, 0% of concentrations were above the LOD.

The Investigating Mixtures in Pollutants and Endometriosis in Tissue (IMPLANT) study was nested in the ENDO study. The George Mason University institutional review board approved the IMPLANT study. The ENDO study participants provided consent for future use of stored biospecimens and participants. IMPLANT included participants 18–44 years of age who underwent laparoscopic surgery or laparotomy for any indication in one of the 14 surgical sites located in Salt Lake City, Utah, or San Francisco, California, during 2007–2009. 24 Exclusion criteria included a prior diagnosis of laparoscopy- or laparotomy-confirmed endometriosis, current pregnancy, breastfeeding in the past 6 months, postmenopause, history of hormonal treatment in the past 2 years, history of cancer, male sex assigned at birth, and inability to consent. 24 Eutopic endometrium specimens were obtained for all participants during laparoscopy or laparotomy using a Pipelle. Endometrial tissue was placed in a 2 -mL cryogenic vial. When ectopic, endometrial-like tissue was present, it was removed by excision rather than laser. Biopsy samples were obtained and stored similarly. Participants with ectopic tissue samples also provided eutopic tissue samples. The tissue was preserved with formalin and frozen at − 80 ° C until shipment for analysis. Using the American Society of Reproductive Medicine (ASRM) staging and classification, surgeons reported the severity and type of endometriosis. 25 Participants with endometriosis were classified as subtypes superficial or deep infiltrating endometriosis and ovarian endometrioma based on surgical report. Since participants can have superficial, deep, and ovarian endometrioma at the same time, participants with endometriosis were classified as having subtypes superficial or deep infiltrating endometriosis and ovarian endometrioma. The eutopic tissue ( 0.1 g ) was homogenized in 1 mL of Milli-Q water using a homogenizer (Fisher Homogenizer 850) until the tissue was completely ground. The homogenized solution was transferred into a 15 -mL PP tube, spiked with 13 C labeled internal standards (IS) and sonicated for 60 min. To this mixture, 1 mL of 0.5 M tetrabutyl ammonium hydrogen sulfate solution (adjusted to pH 10), 2 mL of 0.25 M sodium carbonate/sodium bicarbonate buffer, and 5 mL of methyl tert-butyl ether (MTBE) were added. The samples were shaken in a reciprocal shaker (Eberbach Corp.) at 250 oscillations/min for 1 h. Then, the samples were centrifuged at 4,500  rpm for 10 min (Eppendorf Centrifuge 5804), and the organic layer was transferred into a new tube. The extraction was repeated twice with 4 mL of MTBE. The extracts were combined and evaporated to near dryness under a gentle nitrogen stream. The samples were reconstituted in 250 μ L methanol, vortexed, filtered through a 0.22 - μ m nylon filter (Corning) and transferred into a vial for analysis using high-performance liquid chromatography–tandem mass spectrometry (HPLC-MS/MS). Concentrations of target chemicals were determined using an ABSCIEX 5500 electrospray tandem mass spectrometer (ESI-MS/MS) (SCIEX), interfaced with a Nexera X2 LC-30AD series HPLC (Shimadzu). The analytes were separated using a BEH C18 column ( 2.1 × 50 mm , 1.7 μ m ; Waters). MeOH and 0.1% wt/vol ammonium acetate in water were used as mobile phases. Target analytes were detected using multiple reaction monitoring in negative ionization mode. Further details of mass spectrometric conditions and HPLC parameters are described in Honda et al. 26 Nine PFAS, perfluorodecanoic acid (PFDA), perfluorohexane sulfonic acid (PFHxS), perfluorononanoic acid (PFNA), perfluorooctanoic acid (PFOA), perfluorooctane sulfonic acid (PFOS), perfluorododecanoic acid (PFDoDA), perfluoroheptanoic acid (PFHpA), perfluorooctanesulfonamide (PFOSA), and perfluoroundecanoic acid (PFUnDA), were analyzed. The target analytes were quantified by isotopic dilution method and a 12-point calibration (at concentrations ranging from 0.02 to 100  ng / mL ) with a regression coefficient of ≥ 0.999 . A pure solvent (MeOH) and a mid-point calibration standard ( 5  ng / mL ) were injected after every 10 samples to check for carryover of target chemicals and instrumental drift in sensitivity. Several procedural blanks were analyzed to monitor for contamination that can arise from reagents and materials used in sample preparation steps. Two replicates of Standard Reference Material (SRM1957 and SRM1958, NIST; IS spiked) containing certified values for PFHxS, PFOS, PFHpA, PFOA, PFNA, PFDA, and PFUnDA for SRM1957 and PFHxS, PFOS, PFOA, and PFNA for SRM1958 were processed. The measured concentrations in SRMs were within ± 10 % of the certified values. The limits of detection (LOD) in nanograms per gram tissue were 0.008 for PFOS, 0.011 for PFOA, 0.004 for PFHxS, 0.01 for PFNA, 0.004 for PFOSA, 0.012 for PFDA, 0.009 for PFUnDA, 0.011 for PFHpA, and 0.008 for PFDoDA. Incident endometriosis was diagnosed by gold standard surgical visualization. 27 Surgeons completed an operative report immediately after surgery to confirm the presence or absence of endometriosis. Endometriosis was confirmed when feasible by histopathologic analysis. Endometriosis typology was based on the revised American Society for Reproductive Medicine criteria, and endometriosis subtype was defined as superficial, deep infiltrating, and ovarian endometrioma as previously described by Byun et al. 25 We evaluated endometriosis by ASRM staging (stages 3 and 4 vs. stages 1 and 2). To examine associations, we created indicator variables for different subtypes. We classified participants diagnosed with deep infiltrating endometriosis as having deep infiltrating endometriosis, regardless of whether they also had superficial endometriosis or ovarian endometrioma. Participants with ovarian endometrioma were classified as such, even if they also had another co-diagnosis of deep infiltrating or superficial endometriosis. Participants classified as having superficial endometriosis could not have a co-diagnosis of ovarian endometrioma or deep infiltrating endometriosis. Additionally, we compared deep infiltrating endometriosis vs. superficial and ovarian endometrioma vs. superficial endometriosis. Prior history of breastfeeding and parity was combined as follows: breastfed, parous; no breastfed, parous; nulliparous. Self-reported race and ethnicity were collected during study enrollment using predefined categories, including non-Hispanic white, non-Hispanic black or African American, Hispanic, Asian American, Pacific Islander, and other. Due to the small number of participants who identified as non-Hispanic black, Hispanic, Asian American, Pacific Islander, or other, we condensed these groups into a single category for analysis. At the baseline visit, trained study staff measured participant weight and height and collected ∼ 24 mL of blood via venipuncture. Cotinine levels in serum were analyzed using HPLC-MS/MS, applying isotope dilution and calibration against external standards. 28 Multiple imputation was used for missing covariates. Descriptive statistics of age, body mass index (BMI) ( kg / m 2 ), enrollment site, breastfeeding/parity, normal pelvis status based on surgical report, and surgical indication based on the diagnosis of endometriosis were obtained. Smoking status was determined by serum cotinine levels, with ≥ 10  ng / mL indicating current smoking. 29 PFAS with < 20 % of samples below the LOD were included in the main analysis. Missing PFAS values were replaced by the LOD / √ 2 . PFAS were log-transformed to approximate a normal distribution. 30 Pearson correlation coefficients between PFAS were determined. PFAS correlation coefficients were evaluated by diagnosis of endometriosis and between serum and tissue PFAS. Descriptive analysis included medians and interquartile ranges of PFAS by endometriosis status and stage. In addition, we evaluated demographic characteristics by ASRM staging 1 and 2 compared to 3 and 4, as well as by endometriosis subtype. Modified Poisson regression models were fitted to determine the relative risk (RR) of endometriosis diagnosis, with 95% confidence intervals (CIs), for each PFAS individually. Additional co-exposure models were adjusted and included all PFAS with 95% CI reported to assist in the interpretation of subsequent flexible models of exposure mixtures. Confounders were selected based on a priori information and directed acyclic graphs (Figure S1). For parsimony, the final adjusted model did not include breastfeeding/parity. The final models included age (years, continuous), BMI ( kg / m 2 , continuous), race and ethnicity (white, all other race/ethnicities), smoking status (current, nonsmoker), and study site (Utah, California). Race/ethnicity and study site were included as confounders based on evidence of variation in environmental exposures and health disparities across different racial/ethnic groups and geographic locations. 31 , 32 By including these variables, we aim to account for potential unmeasured differences in PFAS exposure and health outcomes linked to these factors. In secondary analysis of PFAS with > 10 % below the LOD, models were fitted to binary exposures of PFAS above compared to below the LOD. For ectopic tissue, the direct lab PFAS value was reported. Further, we used Bayesian kernel machine regression (BKMR), which allowed us to model the association between the overall PFAS mixture and endometriosis, accounting for potential nonlinearity, nonadditivity, and interaction effects. 33 Two BKMR models, adjusted for the same confounders as our modified Poisson models, were created for this analysis. The first model estimated the change in latent endometriosis for 5% increases in the PFAS mixture relative to the 25th percentile of the mixture. The second model estimated the change in latent ASRM staging criteria for those with endometriosis with each 5% increase in the PFAS mixture compared to the PFAS mixture at its 25th percentile. Nonlinearity and interactions were explored for each outcome. The goal of this exploratory analysis was to estimate associations with the overall PFAS mixture, so variable selection within BKMR, which identifies individual PFAS most associated with the outcome, was not used. All analyses were completed using R (version 4.2.2; R Development Core Team). A sensitivity analysis was conducted to compare the influence of treating values below the LOD with other commonly used methods, including when they were imputed using at zero, imputed using the LOD, and multiply imputed. 34 These models were also adjusted. In addition to the individual PFAS risks in ASRM staging and status, a co-exposure model that included all PFAS that had values > 80 % above the LOD was used. Further sensitivity analysis included evaluating breastfeeding and parity as a potential confounding factor, as this has been shown to be a factor of PFAS concentrations. 35 In addition, we evaluated how PFAS affect the risk of endometriosis among participants who had surgery specifically for pelvic pain. Lastly, we evaluated correlations between serum and endometrium PFAS levels.

Endometriosis affects ∼ 11 % of individuals assigned female at birth in the United States and causes pain and infertility. 1 , 2 Endometriosis is estrogen-dependent and defined by uterine glands and stroma growing outside of the uterus. 3 The etiology of endometriosis is still poorly understood but historically has been hypothesized to be caused by retrograde menstruation, which is when menses flow through the fallopian tubes instead of exiting through the cervix into the vagina. 4 Other theories of endometriosis origin include being driven by hormonal levels. 5 In endometriosis, estrogen and progesterone signaling is often disturbed, leading to progesterone resistance and an excess of estrogen. 6 , 7 Therefore, endocrine-disrupting chemicals (EDCs), due to their ability to affect endogenous hormones, may play a role in endometriosis incidence or severity. EDCs can bind to hormone receptors and disrupt hormonal signaling. 8 EDCs may influence endometriosis via their ability to activate the generation of endothelial cells related to endometrial agenesis. 9 Per- and polyfluoroalkyl substances (PFAS) are persistent EDCs widely present in the environment. 10 Due to their water- and oil-repelling properties, PFAS have been extensively used in consumer products like stain-resistant fabrics, firefighting foams, nonstick cookware, and cosmetics since the 1950s. Evidence suggests that PFAS may contribute to endometriosis by affecting reproductive hormones and promoting inflammation. 11 – 14 In mice, perfluorooctanoic acid (PFOA) increased uterine weight, indicating estrogenic activity. 15 In human cells, PFOA disrupted progesterone-dependent gene expression. 16 In fish, several types of PFAS activated estrogen receptor-alpha, with 43 PFAS showing ligand binding activity, albeit weaker than 17 beta-estradiol. 17 , 18 Additionally, PFAS can induce oxidative stress, 19 which has been linked to a potential pathway for developing endometriosis. 20 Several studies suggest that PFAS in serum may play a role in the etiology of endometriosis. 21 – 23 While serum or blood measures are convenient and widely accepted as a surrogate for PFAS exposure, characterizing accumulation in target tissue (i.e., at the endometrium) would strengthen our understanding of the role of PFAS in the etiology of endometriosis. However, such analysis is often hindered by the lack of availability of target tissue. This cohort has reported on PFAS in serum in relation to endometriosis. 24 The Endometriosis, Natural History and Diagnosis (ENDO) study consisted of women undergoing laparoscopy or laparotomy with a surgical sample ( n = 434 ) of endometrium providing the opportunity to measure PFAS in the target tissue. 2 Therefore, our goal is to elucidate whether PFAS exposure at the target endometrium is associated with incident endometriosis and endometriosis staging. We measured PFAS in eutopic endometrium tissue in relation to incident diagnosis of endometriosis among a cohort of participants undergoing laparoscopic surgery for any indication.

PFAS were widely detected in eutopic endometrium tissue among women with and without endometriosis. Median concentrations of PFOS, PFOA, PFHxS, PFNA, and PFOSA ranged from 0.11  to  6.69  ng / g . Eutopic PFAS concentration was not associated with risk of incident endometriosis diagnosis in individual or multiple co-exposure PFAS adjusted models. However, among those with endometriosis, PFAS were associated with a higher risk for more advanced endometriosis stage. Specifically, PFOS, PFHxS, and PFOSA were associated with a higher risk of ASRM stage 3 and 4 endometriosis compared to ASRM stage 1 and 2. Using BKMR, the mixture of PFAS in eutopic tissue was associated with latent ASRM stage. Taken together, these findings underscore that PFAS concentrations in eutopic tissue may play a role in endometriosis severity, with little indication of associations with endometriosis incidence in this study. It is challenging to compare concentrations of PFAS in this cohort with other studies as, to our knowledge, no other studies have measured PFAS in endometrium tissue or staging and subtype. Other studies that evaluated PFAS in relation to endometriosis measured PFAS in blood. PFAS were detected in placenta samples in one study 36 but were much lower than eutopic endometrium tissue in our study (median PFOS, 6.98 vs. 0.2  ng / g ; PFOA, 1.93 vs. 0.06  ng / g ; and PFHxS, 0.65 vs. 0.03  ng / g ). In contrast to our findings in tissue, in the ENDO Study, serum levels of PFOS, PFOA, PFNA, and PFDA were associated with higher odds of endometriosis. 24 This may be because PFAS in tissue reflects different exposure timing compared to PFAS in serum. Campbell et al. 21 found tertiles of PFOA, PFOS, and PFNA in serum were associated with endometriosis in the National Health and Nutrition Examination Survey cross-sectional data, although PFAS were measured several years after the diagnosis of endometriosis. 21 A hospital-based cross-sectional study in China found that PFOA ( median = 8.51  ng / mL ) was associated with increased odds of endometriosis. 14 Although our findings differed for endometriosis diagnosis and PFOA, we observed a higher risk of having ASRM stages 3 and 4 compared to stages 1 and 2, with higher levels of PFOA and other PFAS. In line with our findings, living in an area with high PFAS levels in drinking water was not associated with subsequent endometriosis diagnosis in Ronneby, Sweden. 37 However, our exposure assessment differed from this study, as we measured PFAS in endometrium tissue while residential address records were used to categorize high exposure in the Ronneby cohort. Perfluorobutanesulfonic acid (PFBS) was associated with endometriosis related infertility in a cohort of Chinese women, while PFNA was associated with a reduced risk of endometriosis. 23 We were unable to corroborate this in our findings as we did not measure PFBS, but our finding that PFNA was associated with a slightly reduced risk of incident endometriosis corroborates findings by Wang et al. 23 This study had several limitations. Of note, there was limited participant diversity, with most participants identifying as non-Hispanic white. This is a limitation because environmental justice concerns suggest that certain racial and ethnic groups may experience higher PFAS exposures due to disparities in environmental contamination and occupational settings, which may not be captured in this study. Few participants had deep infiltrating and ovarian endometriomas compared to superficial endometriomas, which limited our power to detect associations of PFAS with endometriosis subtypes. Future studies should include more participants with deep infiltrating endometriosis and ovarian endometrioma subtypes. While we measured PFAS in the endometrium at the time of the biopsy, we could not determine the concentrations before endometriosis developed. However, the median PFAS half-life is 1.5–8 years, 38 , 39 and endometriosis diagnosis takes, on average, 6.7 years from symptom onset to diagnosis, so levels may accurately reflect exposure during that time for select PFAS. 40 In addition, while this study tested for nine specific PFAS, there are more than 4,000 in production, 41 and while we did recruit participants from two different sites, the locations might not be representative of more-contaminated sites. However, the study also had several strengths. One strength is the availability of eutopic tissue, which is closer to the site where endometriosis occurs. 42 Another strength is our large sample size compared to a few other studies that have ranged from 87 to 335. 23 , 37 , 43 We were able to adjust for a range of confounding factors, and we were able to estimate associations of endometriosis with mixtures of PFAS as well. In conclusion, we did not observe that PFAS in eutopic endometrium were associated with incident diagnosis of endometriosis. Among women with endometriosis, we found that PFOS, PFHxS, and PFOSA were associated with more advanced ASRM staging. More research is needed to better understand these findings. Although limited by few samples, our finding of higher PFHxS, PFOA, PFNA, PFDA, and PFUnDA levels in ectopic compared to eutopic tissue are intriguing and aid the hypothesis that PFAS may contribute to endometriosis progression. Endometriosis is a debilitating syndrome, with more than 60% of those affected experiencing chronic pain and many also facing infertility. 44 , 45 Individuals who have gone through menopause can also sometimes experience symptoms of endometriosis, implicating that this type of research is important across the different life stages. 46 In addition, several PFAS compounds can adversely affect the immune system. 47 , 48 Endometriosis has autoimmune-like features and is associated with chronic inflammation. 3 Immunotoxicity may play a role; however, more research is needed on the role of the immune system and the severity and type of endometriosis. Lastly, PFAS have come under scrutiny by the Environmental Protection Agency for their possible effects on many diseases. This research is important, as it highlights the potential link between PFAS exposure and advanced stages of endometriosis, emphasizing the need for further research on the role of environmental endocrine-disrupting chemicals in endometriosis severity.