Serum Polyfluoroalkyl Concentrations, Asthma Outcomes, and Immunological Markers in a Case–Control Study of Taiwanese Children

1Perfluorinated compounds (PFCs) include a class of human-made organic chemicals composed of a fluorinated carbon backbone of varying length, terminated by a carboxylate or sulfonate functional group. Such PFCs are extremely stable, thermally, biologically, and chemically, and additionally possess hydrophobic and lipophobic characteristics that enable products coated in them to repel both oil and water and resist staining. Accordingly, PFCs are widely used, for example in surfactants, emulsifiers, food packaging, nonstick pan coatings, fire-fighting foams, paper and textile coatings, and personal care products.

2This combination of extreme resistance to degradation and environmental ubiquity has raised concerns in recent years. Furthermore, studies have shown that PFCs accumulate among the higher trophic level of the food chain, such as predators and human beings. Although data from the National Health and Nutrition Examination Survey have indicated a decrease in serum PFC concentrations in the general U.S. population since the production of some PFCs has been phased out [for example, the average concentration of perfluorooctane sulfonate (PFOS) decreased from 30.4 ng/mL in 1999 to 13.2 ng/mL in 2008], PFCs are still manufactured abroad. PFCs bioaccumulate by binding to proteins in the liver and serum, in contrast with many other persistent organic pollutants that persist primarily in adipose tissue, and they are slowly eliminated without biotransformation. Serum half-life estimates in an occupationally exposed cohort ranged from 5.4 years for PFOS to 8.5 years for perfluorohexane sulfonic acid (PFHxS).

3Several attempts have been made to understand the toxicological hazards that may be associated with exposure. Early animal studies focused almost exclusively on exposure to PFOS and perfluorooctanic acid (PFOA), two of the most common PFCs. Evidence of several potential effects has been reported based on experimental studies, including hepatotoxicity, immunotoxicity, developmental toxicity, reproductive toxicity, neurotoxicity, endocrine toxicity, and tumors of the liver, thyroid, and mammary glands.

4Preliminary data suggest that PFCs have the potential to exacerbate atopic diseases such as asthma. In a murine model of asthma, found that PFOA increased serum levels of immunoglobulin E (IgE) and enhanced the hypersensitivity response to ovalbumin, suggesting that PFOA exposure may augment the IgE response to environmental allergens. Another recent study reported that PFOS exposure decreased baseline airway resistance but significantly increased airway responsiveness in an allergic murine model. Furthermore, in our experimental studies, in vivo exposure to PFOS was associated with decreased secretion of T-helper (TH)1-type cytokines [interleukin (IL)-2 and interferon (IFN)-γ] and increased secretion of TH2-type cytokines (IL-4 and IL-10) and IgE, which suggested that PFOS exposure might shift the host’s immune state toward a more TH2-like state. TH1/TH2 polarization (toward TH2 response) is a hallmark of atopy diseases, and IgE is known to play a role in mediated type 1 hypersensitivity reactions, including asthma. Accordingly, we hypothesized that exposure to PFCs may have a role in asthma development in humans.

5Asthma is the most common respiratory disease in young children, and although recent studies indicated that asthma prevalence has plateaued or may be declining, it is still a major public health problem among young people. The case–control design of the Genetic and Biomarkers study for Childhood Asthma (GBCA) provided an opportunity to explore the association between PFCs exposure and asthma in children. In addition, two validated questionnaires [asthma severity score questionnaire and Asthma Control Test (ACT)] were used to examine the association between PFCs and asthma severity and control in asthmatic children.

6Study population. The GBCA was conducted between 2009 and 2010. A total of 231 10- to 15-year-old children with asthma diagnosed by a physician in the previous year were recruited from two hospitals in northern Taiwan. Controls were selected from our previous cohort study population in seven public schools of Northern Taiwan. These schools had diverse geographical and socioeconomic settings, being located in city, rural, and high-altitude communities, respectively. In each targeted school, children of the same age range and without a personal or family history of asthma were invited to participate, and 225 nonasthmatic controls enrolled in the study (response rate was 72% among those contacted by phone). Information pertaining to demographic variables, environmental exposures, and asthma outcomes was collected from questionnaires. We also collected urine and serum samples for each child after 8 hr of fasting. A trained fieldworker measured each child’s height, weight, waist circumference, and blood pressure. All participants and their parents provided written informed consent. The study protocol was approved by the Institutional Review Board (National Taiwan University Hospital Research Ethics Committee) and complied with the principles outlined in the Helsinki Declaration.

7Serum IgE, AEC, ECP level detection. Venous blood was placed into EDTA tubes, a peroxydase coloration was performed, and the absolute eosinophil counts (AEC) were calculated using automatic analyzer (> 106 cells/L; XE2100; Sysmex, Kobe, Japan). Serum samples were stored at –80°C until total IgE and eosinophilic cationic protein (ECP) levels were analyzed. Serum total IgE levels were determined using a Pharmacia UniCap assay test system (Pharmacia Diagnostics, Uppsala, Sweden). Total IgE concentrations < 0.35 IU/mL were defined as absent or undetectable. ELISA kits were used to measure ECP levels in serum samples according to the manufacturer’s instructions (R&D Systems Europe, Abingdon, UK). The limit of quantitation for ECP concentrations was 0.125 µg/mL.

8Asthma control test and asthma severity evaluation. The ACT, a five-item questionnaire used to assess asthma control in the previous 4 weeks, was administered to the asthmatic children. Questions pertaining to asthmatic symptoms, use of rescue medication, and limitation of daily activities are used to ascertain asthma management. The reliability, empirical validity, and discriminative power in assessing the control of asthma by Chinese children are good. The sum of the scores of the five questions gives the total ACT score (range, 5–25); the higher the score, the better controlled the disease. We also used a 13-item asthma severity score questionnaire to evaluate four overall components of asthma severity in the asthmatic children, including frequency of current asthma symptoms, use of systemic corticosteroids, use of other medications (besides systemic corticosteroids), and history of hospitalizations and intubations. Possible total scores range from 0 to 28, with higher scores reflecting more severe asthma.

9PFC concentrations. PFCs were measured from 0.5 mL of serum using Agilent high-performance liquid chromatography (HPLC)–in tandem with an Agilent 6410 Triple Quadruple (QQQ) mass spectrometer (MS/MS) (Agilent, Palo Alto, CA, USA). Detailed information about standards and reagents, sample preparation and extraction, instrumental analysis, quality assurance and quality control, and recovery experiments in the present study is provided in Supplemental Material, pp. 2–4 (http://dx.doi.org/10.1289/ehp.1205351) and is described elsewhere. Ten PFCs were analyzed in serum samples: PFOS, PFOA, perfluorobutane sulfonate (PFBS), perfluorodecanoic acid (PFDA), perfluorododecanoic acid (PFDoA), perfluoroheptanoic acid (PFHpA), perfluorohexane acid (PFHxA), PFHxS, perfluorononanoic acid (PFNA), and perfluorotetradecanoic acid (PFTA). The limit of quantification (LOQ) for PFOS, PFOA, and PFNA was 0.03 ng/mL; for PFBS and PFHxS 0.07 ng/mL; for PFDA and PFDoA 0.1 ng/mL; for PFHpA and PFHxA 0.05 ng/mL; and for PFTA 0.02 ng/mL. All tests were duplicated and average of the two measures was calculated as the concentrations of PFC.

10Statistical analysis. Statistical analysis was performed using SAS software (version 9.2; SAS Institute Inc., Cary, NC, USA). Concentrations of PFCs and biomarkers below the LOQ were assigned a value equal to the LOQ divided by the square root of 2 for statistical analyses. We calculated univariate statistics, including median, interquartile range (IQR), and range for each PFC. Because PFC concentrations were highly skewed, we utilized the Wilcoxon rank-sum test to compare PFC concentrations between children with and without asthma.

11We used logistic regression to estimate odds ratios (ORs) and 95% CIs for each PFC quartile relative to the lowest quartile, with a priori adjustment for child age and sex. To determine the magnitude of other potential confounding, we examined the following variables using a backward deletion strategy: parental education, body mass index (BMI), environmental tobacco smoke (ETS) exposure, and month of survey. If the estimated PFC effect changed by at least 10% when a covariate was included in the base model, the covariate was included in the final model. Multiple general linear models were used to estimate associations with continuous outcomes (IgE, AEC, and ECP) in PFC quartiles, with the lowest PFC quartile as reference group and adjusted for identified covariates. These models were applied separately for cases and controls. We modeled an ordinal variable assigned the median value for each corresponding quartile to estimate p-values for trend. A p-value of < 0.05 was considered statistically significant.

12Compared with children who did not have asthma, asthmatic children tended to be younger and less likely to report ETS exposure (Table 1). In addition, asthmatic children had significantly higher median plasma concentrations of IgE, AEC, and ECP.

13Nearly all study participants had detectable serum concentrations of all PFCs (> 94% of PFCs) except for PFDoA (84.4% in children with and without asthma) and PFHpA (53.3% in unasthmatic children, and 70.6% in asthmatic children) (Table 2). Because of the large numbers of samples below the LOQ, we did not conduct further analyses of PFHpA. Serum concentrations of PFCs were significantly higher in asthmatic children than in nonasthmatic children (p < 0.05), except for PFHxA concentrations, which were similar in both groups, and PFTA concentrations, which were significantly higher in nonasthmatic children.

14Crude and adjusted ORs for asthma in association with the highest versus lowest quartile of exposure were significantly elevated for all PFCs except for PFHxA and PFTA (Table 3). In general, the data suggest increasing odds of asthma with increasing PFCs, with the strongest associations for exposures in the fourth quartile. Specifically, adjusted ORs for the highest versus lowest quartile were 2.63 (95% CI: 1.48, 4.69) for PFOS, 4.05 (95% CI: 2.21, 7.42) for PFOA, 1.90 (95% CI: 1.08, 3.37) for PFBS, 3.22 (95% CI: 1.75, 5.94) for PFDA, 1.81 (95% CI: 1.02, 3.23) for PFDoA, 3.83 (95% CI: 2.11, 6.93) for PFHxS, and 2.56 (95% CI: 1.41, 4.65) for PFNA.

15None of the PFCs were significantly associated with serum levels of IgE or AEC among children without asthma, but serum ECP concentration was positively associated with PFDA and PFDoA (Table 4). In contrast, among children with asthma, all three biomarkers were positively associated with PFOS and PFOA, with significant monotonic trends with increasing exposure (Table 5). For example, asthmatic children in the highest of PFOS quartile had mean IgE levels of 877.3 IU/mL (95% CI: 695.2, 1059.5), compared with 517.9 IU/mL (95% CI: 336.7, 699.2) for in the lowest quartile (Figure 1). In addition, with the exception of PFHxA, which was not associated with any of the biomarkers, all the remaining PFCs were associated with two of the three biomarkers evaluated.

16Among asthmatic children, positive trends for associations with asthma severity scores were significant for PFOS, PFDA, PFDoA, and PFTA, but none of the PFCs was associated with ACT scores (Table 6).

17Serum concentrations of PFCs were significantly higher in asthmatic children compared with controls, and among children with asthma, all but one of the PFCs evaluated were positively associated with at least two of the three immunological biomarkers (IgE, AEC, and ECP). Although temporality cannot be determined due to the cross-sectional nature of the data, and noncausal associations due to uncontrolled confounding or other sources of bias cannot be ruled out, the robust associations of PFCs with asthma and asthma related biomarkers in children suggest that a causal relationship may be present. However, concentrations of individual PFCs were positively correlated [see Supplemental Material, Table S1 (http://dx.doi.org/10.1289/ehp.1205351)], and therefore it is not possible to determine whether associations apply to multiple PFCs or to only a subset of individual PFCs.

18There is little information in the literature about associations between environmental PFCs and asthma or asthma-related biomarkers in children. In a systematic Medline search (http://www.nlm.nih.gov/bsd/pmresources.html), we identified only two studies of PFCs and atopic disease in humans. The first was a cross-sectional study of 566 residents with prolonged exposure to PFOA in their drinking water. In that study, respiratory illness was evaluated by questionnaire, and standardized prevalence ratios (SPR) using National Health and Nutrition Examination Survey (NHANES) data for comparison rates suggested an increased prevalence of asthma among the exposed participants than in the general U.S. population (SPR = 1.82, 95% CI: 1.47, 2.25). The second was a cohort study of prenatal exposure to PFCs in association with IgE levels and atopic dermatitis in 244 newborns. In that study, prenatal PFOA and PFOS exposures were positively correlated with cord blood IgE levels, but were not significantly associated with atopic dermatitis.

19In a murine model of asthma, evaluated the effects of PFOA dermal exposure on the hypersensitivity response to ovalbumin (OVA). The authors reported that IgE increased to a greater extent in animals exposed to PFOA and OVA, and that the severity of the OVA-specific airway hyperreactivity response, and a pleiotropic cell response characterized by eosinophilia and mucin production, increased with increasing concentrations of PFOA. evaluated the effect of prenatal PFOS exposure on maturation of the terminal airway epithelium in rats based on histological and morphometric examination, and reported that there were significant histologic and morphometric differences between control and PFOS-treated lungs in newborns, suggesting that PFOS may inhibit or delay perinatal lung development. Also, a recent study using an allergic murine model to evaluate effects of PFOS exposure on pulmonary function and airway responsiveness reported that PFOS exposure decreased baseline airway resistance, but significantly increased airway responsiveness in allergic mice. In our experimental studies, in vivo exposure to PFOS was linked decreased secretion of TH1-type cytokines (IL-2 and IFN-γ), and increased secretion of TH2-type cytokines (IL-4 and IL-10) and serum IgE, which suggested that PFOS exposure might shift immune responses toward a more TH2-like state. TH1/TH2 polarization toward TH2 responses is a hallmark of atopy diseases, and IgE plays a role in mediating type 1 hypersensitivity reactions, including asthma. Therefore, we hypothesize that exposure to PFCs may augment the TH2 response, resulting in airway hyper-reactivity to environmental allergens.

20Potential mechanisms for effects of PFCs on immune response and asthma development in humans are uncertain. One possibility is an effect of PFCs on regulatory T cells that influence the development of immune-related diseases including asthma and allergy. Another possible mechanism for effects of PFCs on immune responses pertains to the peroxisome proliferator–activated receptors (PPARs) signaling pathway. PFCs are known as agonists for PPARs (Vanden huevel et al. 2006). Both PPAR-α and PPAR-γ are expressed on cells of the monocyte/macrophage lineage, suggesting a possible role in immune function. Although PPAR-γ activation has been implicated as an important contributor to the pathogenesis of the toluene diisocyanate–induced asthma phenotype in a female BALB/c mice model, PFOS and PFOA have been shown to significantly increase activation of mouse or human PPAR-α and PPAR-β, but not of PPAR-γ, in vitro. reported that PPAR-α agonists, including gemfibrozil, ciprofibrate, and fenofibrate, can increase the production of the TH2 cytokine IL-4, and suppress MBP (myelin basic protein) Ac1-11–induced proliferation by T-cell receptor transgenic T cells. In addition, gemfibrozil shifted cytokine secretion by inhibiting IFN-γ and promoting IL-4 secretion in human T-cell lines. These results suggest that PFCs may potentially augment the TH2 response and subsequent airway hyperreactivity to environmental allergens through a PPAR-mediated mechanism.

21In the present study, nearly all study participants had detectable serum concentrations of all PFCs, including both asthmatic children and controls. PFOS was the most abundant PFC in the serum measured in 2009–2010 in these Taiwanese children, with median concentrations (28.9 ng/mL in controls and 33.9 ng/mL in asthmatics) that were similar to levels reported for children 12–19 years of age in the general U.S. population in 1999–2000 (29.4 ng/mL), but higher than the median concentration reported for U.S. children in 2007–2008 (11.3 ng/mL) and concentrations reported for other populations of children sampled during the mid- to late- 2000s [see Supplemental Material, Table S2 (http://dx.doi.org/10.1289/ehp.1205351)]. In contrast, median concentrations of PFOA in the present study population (0.2 ng/mL in controls and 1.2 ng/mL in asthmatics) were lower than reported for other populations of children during the late 2000s (e.g., 4.0 ng/mL based on NHANES data for 2007–2008), and substantially lower than median concentrations reported for children living in a U.S. community near a manufacturing facility (26.3–32.6 ng/mL). Differences in PFC concentrations among populations may reflect changes in exposures over time, as well as differences in diet and other sources of exposure and individual differences in rates or patterns of metabolism or excretion.

22The limitations of these analyses should be noted. We based the PFCs measures on a single serum sample, and although PFCs have a half-life of 5.4–8.5 years, samples taken at several time points might be more accurate than a single sample for classifying exposure. As previously noted, this is a cross-sectional study, and temporal relationships between exposures and outcomes cannot be established. In addition, associations with individual PFCs may be biased due to correlations with other PFCs. Finally, cases were recruited from two hospitals in northern Taiwan, whereas controls were recruited from schools in the same region. Therefore, estimates also may have been influenced by selection bias or uncontrolled confounding.

23In conclusion, we observed positive associations between serum PFCs and asthma, and positive associations between PFCs and IgE, AEC, and ECP levels, and (to a lesser extent) asthma severity scores, in asthmatic children. These findings suggest that exposure to PFCs may not be related only to asthma outcomes but also to asthma severity. Although the production of some PFCs has been phased out in the United States and Europe, PFCs are still manufactured in Asia and elsewhere, and exposure may also result from the breakdown of similar compounds to PFCs in the environment. Therefore, continued exposure is of public health concern, and additional research on potential immunotoxic effects of PFCs is warranted.

24In the manuscript originally published online, the unit for IgE was given as IU/dL instead of IU/mL, and the unit for absolute eosinophil count (AEC) was given as 109/L instead of 106/L. They have been corrected here.