Prenatal exposure to perfluorooctane sulfonate and perfluorooctanoate and the risk of preeclampsia: a nested case-control study in Shanghai, China.

Abstract

To the Editor: Preeclampsia presents with new-onset hypertension and maternal multi-organ injury that is associated with substantial maternal and fetal morbidity and mortality.[1] Although the etiology remains a mystery, environmental toxicants, especially those that interact with genetic factors and disrupt normal placental function, are emerging as potential risk factors for preeclampsia. Perfluoroalkyl and polyfluoroalkyl substances (PFAS), a kind of aliphatic chemicals, have been widely used in industries and consumer products for their excellent properties of heat-, water- and oil-resistance, leading to ubiquitous contamination and inevitable human exposure. The long-chain perfluorooctane sulfonate (PFOS) and perfluorooctanoate (PFOA), as the most widely used representatives, have been detected in the blood, human cord blood, and breast milk of the general population.[2] Previously, several epidemiological studies have examined the associations between PFOS/PFOA and preeclampsia, but have found inconsistent results, which may reflect differences in outcome assessments, analysis methods, and/or study population.[3] Since the 3M Company, one of the largest PFAS manufacturers, phased out PFOS and PFOA production from 2002, PFOS and PFOA production volumes have been increasing in China. However, the risk of preeclampsia with PFOS and PFOA exposure is much rarely described in China, especially in the most intensively industrialized regions, such as the eastern coast. Thus, we explore the association between maternal serum concentrations of PFOS and PFOA and the risk of developing preeclampsia in Shanghai, China. A nested case–control study was conducted in a longitudinal cohort of pregnant women enrolled in the year 2016 at the Obstetrics and Gynecology Hospital of Fudan University. Eligible subjects were women with singleton pregnancies who were later delivered in the study hospital. Those with complications including diabetes mellitus, fetal malformation, chronic hypertension, kidney disease, or any additional vital pre-existing chronic disorders were excluded. The diagnosis of preeclampsia was based on the 2013 American College of Obstetricians and Gynecologists (ACOG) guidelines. Maternal serum samples of 5 mL were collected into vacuum polypropylene tubes without additives at 16 to 20 gestational weeks (the time of Down's screening). Immediately, the serum samples were sent blindly on dry ice to the laboratories. Serum PFOS and PFOA concentrations were evaluated by high-performance liquid chromatography–mass spectrometry. The limit of detection (LOD) for PFOS and PFOA was 0.1 ng/mL. We substituted the value by dividing the LOD (0.1 ng/mL) by two if the concentration was less than the LOD. The epidemiological and clinical data were extracted from the electronic medical records. Our protocol was approved by the Ethics Committee of the Obstetrics and Gynecology Hospital of Fudan University, and each subject signed informed consent. Statistical analyses were carried out using IBM SPSS Statistics for Windows, Version 20.0 (IBM Corp, NY, USA). The general characteristics of the participants were calculated and reported as mean ± standard deviation for continuous variables and n (%) for categorical variables. The differences between groups were analyzed by the Student's t test for normally distributed continuous variables and by the Mann–Whitney U test for non-normally distributed continuous variables. The chi-squared test was used for categorical variables. Multiple logistic regression analysis was conducted to explore the preeclampsia risk for PFOS and PFOA levels by calculating odds ratios (ORs) and 95% confidence intervals (95% CIs). Furthermore, we used restricted cubic splines with three knots at the 10th, 50th, and 90th percentiles of the PFOS and PFOA concentrations, with the reference value (OR = 1) set at the 50th percentile. The results were adjusted for potential confounders including maternal age, parity, and body mass index (BMI), which have been reported to be associated with preeclampsia. P value <0.05 represented statistical significance. A total of 4716 pregnant women were enrolled. 1147 women were excluded due to gestational or pregestational diabetes mellitus (N = 849), fetal malformation (N = 76), chronic hypertension (N = 54), kidney disease (N = 27), or any additional vital pre-existing chronic disorders (N = 141). After exclusion, 3569 participants remained, among whom 88 developed preeclampsia. Controls (n = 88) were selected from healthy full-term pregnant women of similar ages and blood collection dates. Two samples of preeclampsia cases were missed and therefore removed for additional analysis [Supplementary Figure 1, https://links.lww.com/CM9/B220]. There were no statistically significant differences between the groups regarding maternal age, educational level, parity, and sampling time. Participants in the preeclampsia group had a higher BMI at the first visit, which was corrected in subsequent regression analysis. In addition, compared with the results of the controls, cases had earlier gestational delivery weeks and a higher proportion of cesarean section. However, these two factors were outcome indicators and had a negligible impact on our analysis. No participants consumed alcohol or smoked cigarettes [Supplementary Table 1, https://links.lww.com/CM9/B220]. Only two samples (1.2%) in the control group had lower concentrations of PFOS than LOD (<0.1 ng/mL). The median (25th, 75th) concentrations in all subjects were 7.1 ng/mL (3.9–23.5 ng/mL) for PFOS and 17.6 ng/mL (13.3–25.7 ng/mL) for PFOA. The preeclampsia cases had higher PFOS and PFOA concentrations than controls. In subgroup analysis, all subsets of preeclampsia had significantly increased serum PFOS concentrations compared to controls (all P values <0.001). Significantly higher PFOA concentrations were observed in severe and late-onset preeclampsia compared with controls (P < 0.05). Among pre-eclampsia subjects, significantly higher PFOS concentrations were observed in mild-pre-eclampsia cases compared with severe-pre-eclampsia cases (P < 0.001) [Table 1]. Table 1 - Maternal serum PFOS and PFOA concentrations in the study population. Population Group N PFOS concentrations (ng/mL) P value∗ P value† PFOA concentrations (ng/mL) P-value∗ P-value† Controls 88 4.6 (3.6, 7.1) Ref / 16.1 (13.2–21.4) Ref / Preeclampsia 86 23.3 (7.7, 34.3) <0.001 / 19.3 (14.3–27.0) 0.012 / Severity Mild 53 18.3 (4.4, 29.9) <0.001 Ref 18.0 (11.3–26.4) 0.127 Ref Severe 33 29.2 (19.8, 48.4) <0.001 <0.001 23.0 (16.4–31.1) 0.002 0.065 Onset time Early 15 23.4 (15.7, 32.2) <0.001 Ref 19.1 (16.2–30.0) 0.095 Ref Late 71 23.3 (7.2, 35.3) <0.001 0.277 20.1 (12.5–26.9) 0.017 0.165 Data are presented as n or median (25th, 75th).∗Preeclampsia and the subsets were separately compared with controls.†Comparisons were conducted among subsets.PFOA: Perfluorooctanoate; PFOS: Perfluorooctane sulfonate; Ref: Reference. In the logistic regression models, the estimated ORs (95% CI) of preeclampsia risk significantly increased 7.7-fold (3.8–15.7) compared with low PFOS concentrations after adjusting for maternal age, BMI, and parity when the median value of 7.1 ng/mL was selected as the point of interception. On the other hand, higher PFOA exposure increased 1.8-fold the risk of preeclampsia after adjusting for potential confounders, although it was not significant [Supplementary Table 2, https://links.lww.com/CM9/B220]. Restricted cubic splines showed significant linear dose-response relationships between PFOS and the risk of preeclampsia. The association between PFOA and preeclampsia was not statistically significant [Supplementary Figure 2, https://links.lww.com/CM9/B220]. In our nested case–control study, we found that the maternal serum concentrations of PFOS were significantly positively associated with higher odds of preeclampsia, while the association of PFOA was mild. These findings provided new targets for the etiological prevention of preeclampsia and put forward a basis for the formulation of public health policy. Studies of highly exposed communities in the US-based C8 Health Project reported PFOS and PFOA had modestly associated with maternally self-reported preeclampsia.[4] However, there were certain inherent limitations in these data, mainly including exposure measurement bias on PFOS and PFOA contaminated drinking water and information error caused by the self-reporting outcomes, so the results may not be generalizable to other populations. In a Norwegian MoBa cohort, increased levels of PFOS had minor associations with preeclampsia risk in nulliparous women at mid-pregnancy (around 17–20 gestational weeks).[5] However, preeclampsia diagnosis was obtained from registers. Likewise, another Chinese study found no association between cord blood PFOS and PFOA concentrations and preeclampsia, while cord blood may not be a good biomarker for early pregnancy exposure.[6] To overcome these shortcomings, our study was performed on the general population and maternal blood samples were used to determine PFOS and PFOA levels, which is believed to be a better biomarker to reflect internal exposure in pregnant women. Besides, the serum collection time was chosen at the first half of pregnancy (16–20 gestational weeks) in order to coincide with the window of spiral arteries remodeling and placental development. In addition, the nested case–control design enabled us to select all preeclampsia cases and matched control subjects to avoid selection bias. These data provide further evidence of a causal relationship between PFOS and PFOA exposure and preeclampsia. Our study showed that almost all pregnant women in Shanghai are widely exposed to PFOS and PFOA. With the use of PFOS and PFOA being restricted by the government in the United States in 2001 and Europe in 2008, more chemical industries have shifted and relocated from Europe and America to Asia. Actually, the annual production of PFOS and PFOA has skyrocketed since 2003 in China due to increasing demand in both domestic and overseas markets. Thus, we will be more likely to be exposed to PFOS and PFOA in the future. The PFOS and PFOA concentrations in our study were almost at the same level as those obtained in the Shanghai Birth Cohort Study conducted at Xinhua Hospital between 2013 and 2015 (PFOS 8.2 ng/mL; PFOA 11.6 ng/mL).[7] Consequently, stronger and harsher monitoring should be adopted to control PFOS and PFOA pollution for the preservation of human health. Several limitations should be mentioned in this study. First, the firm causal relationship cannot be established owing to the observational design. Second, the blood samples were only obtained at a single time point of 16 to 20 gestational weeks, which may not comprehensively reflect PFAS exposure before and during pregnancy. However, considering its long half-life in the body, a single estimation may have minimal influence on the results. Third, we cannot, as with all observational studies, exclude the effect of other unmeasured factors on the results, even though we adjusted for many potential confounders. Finally, since the participants included in our study were of Han ethnicity and enrolled from only Shanghai, China, the results may not be generalizable to other racial groups or other localities. Overall, this study confirmed that pregnant women in Shanghai, China were extensively exposed to PFOS and PFOA, and that serum PFOS concentrations were closely associated with the risk of developing preeclampsia. The result supports the development of a public health policy with a view to protecting maternal and fetal health by restricting the use of PFOS and PFOA. Moreover, our study provides a base for future research on the mechanisms that will provide biological plausibility on the causality between PFAS exposure and preeclampsia. Statement All the authors have read the manuscript and approved for submission. All the authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Funding This work was supported by grants from the National Science Fund of China (81200449) and the National Science Fund of Shanghai, China (12ZR1403700). Conflicts of interest None.