980546120941Toxicol SciToxicol. Sci.Toxicological sciences : an official journal of the Society of Toxicology1096-60801096-092924284791472420610.1093/toxsci/kft269HHSPA722624ArticleInfluenza Vaccine Response in Adults Exposed to Perfluorooctanoate and PerfluorooctanesulfonateLookerClaire*LusterMichael I.†1CalafatAntonia M.‡JohnsonVictor J.§BurlesonGary R.§BurlesonFlorence G.§FletcherTony*Department of Social and Environmental Health Research, London School of Hygiene and Tropical Medicine, London, WCIH 9SH, UKSchool of Public Health, West Virginia University, Morgantown, West Virginia 26505Division of Laboratory Sciences, National Center for Environmental Health, Centers for Disease Control and Prevention, Atlanta, Georgia 30333Burleson Research Technologies, Inc., Morrisville, North Carolina 27560To whom correspondence should be addressed at School of Public Health, West Virginia University, 39 Quail Road, Morgantown, WV 26508. Fax: 304-212-4701. miklus22@comcast.net13120162711201332014231201613817688For permissions, please journals.permissions@oup.com.

Supported by several epidemiological studies and a large number of animal studies, certain polyfluorinated alkyl acids are believed to be immunotoxic, affecting particularly humoral immunity. Our aim was to investigate the relationship between the antibody response following vaccination with an inactivated trivalent influenza vaccine and circulating levels of Perfluorooctanoate (PFOA) and perfluorooctanesulfonate (PFOS). The study population consisted of 411 adults living in the mid-Ohio region of Ohio and West Virginia where public drinking water had been inadvertently contaminated with PFOA. They participated in a larger cross-sectional study in 2005/2006 and were followed up in 2010, by which time serum levels of PFOA had been substantially reduced but were still well above those found in the general population. Hemagglutination inhibition tests were conducted on serum samples collected preinfluenza vaccination and 21 ± 3 days postvaccination in 2010. Serum samples were also analyzed for PFOA and PFOS concentrations (median: 31.5 and 9.2 ng/ml, respectively). Questionnaires were conducted regarding the occurrence and frequency of recent (during the last 12 months) respiratory infections. Our findings indicated that elevated PFOA serum concentrations are associated with reduced antibody titer rise, particularly to A/H3N2 influenza virus, and an increased risk of not attaining the antibody threshold considered to offer long-term protection. Although the direct relationship between weakened antibody response and clinical risk of influenza is not clear, we did not find evidence for an association between self-reported colds or influenza and PFOA levels nor between PFOS serum concentrations and any of the endpoints examined.

immunotoxicologyinfluenza vaccinepolyfluorinated alkyl acidsPerfluorooctanoateperfluorooctanesulfonate

Polyfluorinated alkyl acids (PFAAs) are organofluorine compounds used in various manufacturing processes that are readily found in the environment. Perfluorooctanoate (PFOA or C8) and perfluorooctanesulfonate (PFOS) are two of the most common PFAAs. They have long half-lives, initially reported to be 3.5 years (Olsen et al., 2007), and their persistence in the blood of animals and humans has been linked to a number of adverse health outcomes (Steenland et al., 2010). There is evidence of widespread exposure to PFOA and PFOS in the general population although levels vary between countries and over time (Calafat et al., 2007; Fromme et al., 2007; Kannan et al., 2004; Kato et al., 2011b; Olsen et al., 2012). In 2007–2008, the mean serum concentrations in the general population of the United States were 13.2 ng/ml (PFOS) and 4.1 ng/ml (PFOA) (Kato et al., 2011b). Although mean serum concentrations of PFOA and PFOS have decreased in recent years in developed countries (Glynn et al., 2012; Kato et al., 2011b; Schroter-Kermani et al., 2012), likely due to decreased production and use of PFOS and PFOA, the stability of the compounds and recent increases in serum PFOA concentrations observed in Korea and PFAAs in China (Harada et al., 2010; Jin et al., 2007), along with their potential human health effects, are of concern.

The Dupont Company in West Virginia had used PFOA and related substances in manufacturing processes since 1951. PFOA contamination of water supplies downwind and downstream of the Dupont plant, located near the Ohio River, has been observed since the 1980s. Although production has been drastically reduced and manufacturing processes have been changed, investigations of possible health outcomes related to this contamination continue. In 2001, a class action lawsuit was filed by residents from affected communities alleging health effects from the contaminated water supplies. As part of a class action settlement, the C8 Health Project was established to investigate possible health effects in exposed populations, and follow-up studies in a subgroup of this population were conducted in 2009–2010.

Animal and human studies have suggested PFAAs are immunotoxic, affecting particularly humoral immunity. For example, in laboratory rodent studies, exposure to either PFOS or PFOA, even at blood concentrations similar to those detected in highly exposed humans, suppressed antibody responses (DeWitt et al., 2008, 2012; Fair et al., 2011; Yang et al., 2002). Using human leukocytes, in vitro exposure to PFOA inhibited interleukin (IL)-4 and IL-10, cytokines which help regulate immune responses (Corsini et al., 2011; Fletcher et al., 2011). Grandjean et al. (2012) found that increased PFOA or PFOS serum concentrations were associated with reduced antibody responses to childhood diphtheria and tetanus vaccines. Similarly, a small Norwegian study recently reported a negative association between maternal PFOA and PFOS serum concentrations and rubella vaccine response, as well as an increase in certain infections in children (Granum et al., 2013).

One of the most commonly used vaccines in the developed world is the seasonal influenza vaccine. Influenza is a viral illness with the potential for serious morbidity and mortality, particularly in children and the elderly, resulting in more than 30 000 fatalities per year in the United States alone (Blanciforti, 2006). The CDC currently recommends universal influenza vaccination for all individuals older than 6 months who do not have specific contraindications (eg, severe egg allergy, history of Guillain-Barré) (NCIRD, 2011). In the United Kingdom, vaccination is recommended for those in specific high-risk groups (Department of Health, 2011). Factors associated with a reduced antibody response to influenza vaccination include older age (Goodwin et al., 2006; Goronzy et al., 2001; Murasko et al., 2002; Weksler and Szabo, 2000; Wick and Grubeck-Loebenstein, 1997), smoking status (Cruijf, 1999), use of immunosuppressive therapies, and chronic inflammatory illnesses (Admon et al., 1997; Blumberg et al., 1998; Cullen et al., 2012; Dengler et al., 1998; Holvast et al., 2009; Robertson et al., 2000; Salles et al., 2010; Tiu et al., 2011; Vilchez et al., 2002). Chronic stress has also been found to be negatively associated with vaccine antibody titers in both young and old (Pedersen et al., 2009; Vedhara et al., 1999).

In contrast to childhood vaccines, for which pre-existing immunity is typically low, prevaccination antibody levels to influenza can be highly variable in adults due to previous infections or vaccinations with cross-reacting epitopes. For this reason, measurement of both pre- and postvaccination antibody levels is important to establish an individual's response to a vaccine. This study offered a unique opportunity to assess the immune response to influenza vaccine in a group of residents who lived in the 6 contaminated water districts surrounding the Dupont chemical plant.

Materials and MethodsStudy Design

This study was conducted in 2010/2011 as part of the C8 Science Panel Studies and was approved by the appropriate Institutional Review Board (IRB) and comply with all relevant national, state, and local regulations. The C8 Health Project initially collected data on 69 030 eligible subjects between August 2005 and August 2006. Eligible participants were those who had consumed water for at least 1 year between 1950 and December 3, 2004, while living, working, or going to school in one of the following 6 water districts: Little Hocking Water Association of Ohio; City of Belpre, Ohio; Tupper Plains–Chester District of Ohio; Village of Pomeroy, Ohio; Lubeck Public Service District of West Virginia; Mason County Public Service District of West Virginia; or private water sources within aforementioned districts and areas of documented PFOA contamination. Details of the study enrollment process, including consenting procedures, have been described elsewhere (Frisbee et al., 2009). In 2010, 755 participants were recalled and invited to participate in a second interview and provide a second blood sample for PFAA analysis (Fitz-Simon et al., 2013). This interview included questions regarding the occurrence and frequency of a number of recent (during the last 12 months) common infections, including coughs, colds, flu, and other upper respiratory infections. Frequency information was only gathered for cold episodes but not influenza specifically. A subset of these participants who had not yet received the annual flu vaccine were offered the vaccine and asked to consent to pre- and postvaccination blood sampling to determine virus-specific antibody titers.

The study was conducted with every participant receiving an influenza vaccine and short health questionnaire. The prevaccination serum sample was collected at the time of vaccination, and the postvaccination sample was collected 21 ± 3 days later (FDA, 2007). Subjects (411) were healthy adults aged over 18 years who did not have a history of influenza vaccination in the past 3 months or a relatively high risk of influenza complications (cancer, recent history of influenza, and vaccine allergy).

Vaccine

A single lot of FLUVIRIN (Novartis, Cambridge, Massachusetts), an inactivated trivalent vaccine containing influenza serotypes A/H3N2 and A/H1N1 (swine flu) and influenza B, was used in the study.

Hemagglutination Inhibition Assays

Sera were tested for influenza-specific antibody by a hemagglutination inhibition (HI) assay using turkey red blood cells (tRBCs) for A/H3N2 and A/H1N1 and influenza B. Nonspecific HI activity was removed by treating the serum samples with receptor destroying enzyme (RDE, ATCC) overnight at 4°C. The RDE-treated sera were tested for residual HI activity by adding 50 μl of the treated serum to 50 μl of the 0.5% tRBCs. This mixture was incubated for 30–60 min and then read to determine whether HI activity was present. If HI activity was present, the nonspecific agglutination inhibitors were removed by adsorption of the serum with packed tRBCs for 1 h at 4°C. Patient serum was prepared by using a 2-fold dilution curve starting at 1:10 and ending at 1:10 240. The serum dilution series (25 μl/dilution) was mixed with 25 μl of optimized dilutions of the 3 influenza antigens in a V-bottom 96-well plate and incubated for 45–60 min at 25°C. Fifty microliters of standardized 0.5% tRBCs were then added to the appropriate wells and incubated at room temperature for 30–60 min to allow the RBCs to settle to the bottom of the well. Positive (known titer of influenza-specific antibodies) and negative (no influenza-specific antibodies) control sera were run in parallel to the test subject sera. Quality control plates were prepared to confirm the optimized antigen. If there were no hemagglutinins present in the sample, as in the control wells that did not contain antigen, the tRBC pellet would stream across the bottom of the well forming a distinct “teardrop” shape. The plates were read when the “teardrop” in the control wells (no antigen) extended from the middle of the well to the outer edge of the well. If influenza-specific antibody was present in the serum sample, the influenza antigen would be bound by the antibody before it could interact with the tRBCs to cause agglutination. The dilution of serum at which a complete “teardrop” formed was designated at the influenza-specific antibody titer for the antigen being tested.

PFOA and PFOS Concentrations

Laboratory measurement of PFOA and PFOS used online solid phase extraction coupled with reversed-phase high-performance liquid chromatography separation and detection by isotope-dilution tandem mass spectrometry (Kato et al., 2011a). The detection limits were 0.5 ng/ml (PFOA) and 0.2 ng/ml (PFOS). At the laboratory, interday precision, calculated as the relative standard deviation of 60 repeated measurements in a 6-month period, was below 8% (PFOA: 7.6% [3.1 ng/ml] and 5.8% [11.7 ng/ml]; PFOS: 7.3% [8.2 ng/ml] and 7.6% [27.5 ng/ml]). Intraday precision, calculated as the relative standard deviation of 5 repeated measurements within 1 day, ranged from approximately 2% to 6% (PFOA: 2.8% [3.1 ng/ml] and 1.7% [11.7 ng/ml]; PFOS: 5.8% [8.2 ng/ml] and 4.9% [27.5 ng/ml]).

AnalysisVaccine response measures

The geometric mean antibody titer (GMT) rises for different participant characteristics are described. Statistical significance between participant groups for GMTs was determined by examining the bounds of the 2-sided 95% confidence intervals to determine whether they overlapped. To explore the associations between antibody titers and PFOA and PFOS levels, several measures of vaccine response were used. Firstly, the antibody titer rise following vaccination was calculated by subtracting the prevaccination titer from the postvaccination titer. Secondly, the ratio of pre- and postvaccination titers was examined by dividing the postvaccine titer by the prevaccine titer. The antibody titer rise and ratio for antibodies specific for each strain showed skewed distributions and were log10 transformed for the analysis.

Additional analyses used dichotomous surrogate outcomes to measure vaccine response. The associations between PFOA/PFOS and seroconversion, defined as a 4-fold or greater increase in antibody titer following vaccination, and seroprotection, defined as a postvaccination HI titer > 1:40 (irrespective of prevaccination levels) were examined. The seroprotection threshold is usually considered indicative of protection from a clinical testing perspective (Beyer et al., 2004).

Exposure measures

Concentrations of PFOA and PFOS showed skewed distribution and were log10 transformed. PFOA and PFOS were examined as linear variables in both the transformed and untransformed form and grouped as categorical exposure groups of quartiles of PFOA and PFOS levels.

Regression analysis

The associations between the antibody titer rise or ratio and untransformed, log10 transformed and concentration quartiles of PFOA and PFOS were examined using linear regression. Given the significant skew in the PFOA measurements, only log10-transformed and quartile concentrations of PFOA and PFOS are presented in the main text. Initially, each association was examined considering only PFOA or PFOS as an exposure (unadjusted results for linear models only are shown). Age and gender were then included as obligatory covariates. The effect of age on antibody titer rise and ratio was found to be nonlinear; therefore, age was fitted using a restricted cubic spline function with 7 knots in the linear regression model. The possible effect of additional a priori confounders including smoking status, any previous influenza vaccination (participants were excluded if this had occurred in the last 3 months), specific H1N1 vaccination in the previous year, day of serum sample collection, coexisting medical conditions and common anti-inflammatory, and pain relief medications was also considered. Potential confounders associated with the log10-transformed antibody titer rise or antibody titer ratio of that vaccine strain (p < .20) were tested in the model and retained if they remained independently associated with the outcome (p < .05). The confounders included in the final linear regression models were age (fitted as cubic spline), gender, mobility (as measured by the number of addresses since 1970 or birth), and a history of previous influenza vaccination.

Multivariable logistic regression models were used to calculate odds ratios to assess the effect of PFOA and PFOS on the likelihood of achieving seroconversion or seroprotection following vaccination. The odds ratio represents a close approximation of relative risk. A priori confounders associated with seroconversion or seroprotection (p < .20) in univariate analysis were included in a multivariate regression model and retained if they remained independently associated with the outcome (p < .05). In the logistic regression model, age was modeled as a categorical variable in 10-year age bands. Only adjusted models are presented in the text.

Multivariable logistic regression models were also used to calculate odds ratios to assess the effect of serum PFAA levels on the likelihood of influenza and colds in the past year. Because the symptoms of influenza are not readily distinguished from other respiratory infections (Call et al., 2005), both self-reported colds and influenza were considered in this analysis. Ordered multi-variable logistic regression was used to test whether there was a risk between increasing frequency of cold and influenza episodes during the year and increased PFAA levels. Models for the occurrence of adult influenza were first fitted adjusted for age and sex only. The possible effect of other a priori con-founders including smoking status, alcohol intake, body mass index, diagnosis of diabetes, and educational level was also considered but was not found to significantly affect the final model. All statistical analysis was conducted using Stata 12.1 (StataCorp, Texas).

ResultsBaseline

A total of 411 adults had a prevaccination antibody titer taken and received the influenza vaccination as part of the study (Table 1). The population median log10 serum PFOA concentration was 1.50 (IQR: 1.14, 1.95) and PFOS, 0.96 (IQR: 0.76, 1.16), reflecting a geometric mean concentration of 33.74 ng/ml (95% confidence interval [CI]: 29.78, 38.22) and 8.32 ng/ml (95% CI: 7.65, 9.05), respectively. There was a modest trend of higher PFOA and PFOS concentrations among older participants (Table 1). Postvaccination serology was available for 403 (98.1%) participants on whom most analysis is based. Participants had higher prevaccination antibody GMTs for A/H3N2 than for A/H1N1 and B (20.69 [95% CI: 17.87, 23.95], 16.11 [95% CI: 13.90, 18.68], and 8.72 [95% CI: 8.02, 9.49]), respectively (not shown). Prevaccination GMTs were higher among those who reported previous influenza vaccination from last year than those who did not (10.17, 19.81, and 31.42 compared with 6.04, 9.83, and 7.60 for B, A/H1N1, and A/H3N2 antigens, respectively), and there was a tendency toward greater pre-existing immunity (prevaccination titers) among those who reported greater residential mobility (a higher number of previous addresses) (Table 1). Following vaccination, there was strong evidence of a larger response to A/H1N1 (overall GMT rise 343.08 [95% CI: 296.07, 397.56]) compared with A/H3N2 (151.49 [95% CI: 126.19, 181.86]) or B (44.21 [95% CI: 38.58, 50.66]). Furthermore, the antibody response to A/H1N1 was strongly associated with age, being higher among younger participants (Supplementary Table S1). For Flu B, a higher antibody response was seen among men than women (54.47 [95% CI: 44.92, 66.05] and 35.32 [95% CI: 29.26, 42.64], respectively), and for both A/H1N1 and B, there was strong evidence that those who reported previous influenza vaccination had lower antibody responses than those who did not (34.91 and 296.68 compared with 74.41 and 467.28 for B and H1N1, respectively) (Supplementary Table S1).

Details of prevaccination antibody GMTs by medical comorbidity and medication groups can be found in Supplementary Table S2. Participants with a history of asthma had a higher prevaccination GMT for H1N1 than those without, reflecting the higher rate of previous A/H1N1 vaccination reported among those with asthma (65% vs 35.2%, p = .03 [not shown on table]). Similarly, those with diabetes reported both higher rates of previous influenza vaccination and had a higher prevaccination GMT for A/H3N2 than those without diabetes (41.38 [95% CI: 25.37, 67.48] vs 19.16 [95% CI: 16.45, 22.30]).

Vaccine ResponseGMT rise

Vaccine response as measured by the GMT rise against each of the 3 vaccine components is shown in Table 2 by quartile of PFAA. Individuals in the fourth quartile of PFOA exposure had a lower GMT rise in Flu B; however, this trend was less evident for A/H1N1 and PFOA. For other A/H3N2 and PFOA and for any type with PFOS, there was no evident pattern with rise in antibody titers.

Log<sub>10</sub>-transformed antibody titer rise and ratio

There was a negative association between log10-transformed A/H3N2 antibody titer ratio and PFOA when log10-transformed PFOA was fitted as a continuous variable in the regression model (adjusted coefficient −0.12 [95% CI: −0.25, 0.02, p = .09]) (Table 3). By PFOA quartile, the A/H3N2 antibody titer ratio showed a largest reduction in response for the highest quartile of PFOA concentrations (adjusted coefficient −0.22 [95% CI: −0.43, −0.01]) (Table 3).

For titer rise, there was no overall trend with exposure, but the second and third quartiles of PFOA concentration had lower rises in log10-transformed A/H3N2 antibody titers than those in the first quartile (adjusted coefficients −0.28 [95% CI: −0.51, −0.06] for second quartile and −0.37 [95% CI: −0.60, 0.13], respectively) (Table 3).

Although unadjusted results suggested a negative association between log10-transformed A/H1N1 antibody titer rise with quartiles of PFOA concentrations, both PFOA blood concentrations and A/H1N1 antibody response are related to age, and this association did not persist in the adjusted model (Table 3). The effects on A/H3N2 antibody titers were not confounded by age. There was no suggestion from the linear regression analysis of an association between influenza B antibody titer rise or ratio and PFOA or PFOS concentrations (Table 3).

Seroconversion/seroprotection

Seroconversion for influenza vaccines was defined as having a 4-fold increase in antibody titer. Seroprotection was defined as an HI antibody titer ≥ 1:40 postvaccination independent of the prevaccine HI titer. Both are used as evidence of efficacy to help support the licensing of seasonal inactivated influenza vaccines (FDA, 2007). The proportions of participants who seroconverted following vaccination and associations with PFAA are shown in Table 4. Overall, a higher proportion of participants seroconverted to H1N1 than H3N2 and B (84.12% [95% CI: 80.54, 87.70] compared with 64.76% [60.08, 69.45] and 62.03% [57.28, 66.79] for H3N2 and B antigens, respectively). As expected, previous influenza vaccination was associated with a decreased likelihood of seroconverting to all 3 components of the vaccine (ORs 0.25, 0.13, and 0.31 for B, A/H1N1, and A/H3N2, respectively; p < .01 for all) (Supplementary Table S4). Increased age was also associated with decreased odds of seroconverting to influenza B (p = .04) and a higher proportion of men seroconverted to B than women (68.3% vs 55.6) (Supplementary Table S4). There was weak evidence of odds of seroconversion to A/H1N1 increasing with PFOA concentration with the odds ratio rising to 2.23 (0.90–5.53) in the fourth quartile, but all confidence intervals were wide, and we observed no other associations of note between PFOA or PFOS and likelihood of seroconverting to the vaccine components.

The proportions of participants who were seroprotected (HI titer ≥ 40) at their postvaccination blood test were 95.5% for A/H1N1, 83.9% for A/H3N2, and 66% for B. (Table 5). Unlike B and A/H3N2, the response to A/H1N1 was not strongly associated with whether the participant had received a previous influenza vaccination (Supplementary Table S5). Participants who reported previous influenza vaccination had a decreased likelihood of seroprotection for B (odds ratio [OR]: 0.50 [95% CI: 0.29, 0.84], p = .01), but an increased likelihood of seroprotection to A/H3N2 (OR: 2.09 [95% CI: 1.18, 3.69], p = .01) (Supplementary Table S5). Increasing age was also strongly associated with decreased seroprotection for B and H1N1 (Supplementary Table S5).

After adjustment for age, gender, and past vaccination, there was some evidence of seroprotection for decreased likelihood of seroprotection from A/H3N2 in relation to PFOA (by PFOA quartile the odds ratios were 0.34 [0.14, 0.83], p = .02, 0.28 [0.11, 0.70], p = .01, and 0.39 [0.15, 0.99], p = .05, respectively). Conversely, there was a suggestion of an increasing trend of seroprotection for A/H1N1, with an overall OR of 2.34 (0.91–6.07) per log unit of PFOA. No other strong associations between seroprotection and PFOA or PFOS levels were found.

Self-reported cold and influenza episodes

Data were available on the frequency of respiratory infections in the population of the entire follow-up study (n = 755), which included those who participated in the vaccination study. One hundred and sixty-three (21.6%) adult participants reported an episode of influenza in the 12 months preceding the questionnaire, whereas 538 (71.3%) reported a cold in the same time period (Tables 6 and 7). Age was a major confounder with younger participants who more likely to report colds and flu and more likely to have had lower PFAA levels (Table 1). Adjusted models assessed either colds or flu, comparing those with and without a reported infection in relation to either log PFAA as a continuous variable or quartiles of PFAA concentration. For colds, the frequency of reported episodes was also assessed. There was no evidence for any association between PFAA concentrations and either self-reported colds or flu (Tables 6 and 7).

Discussion

This study offered a unique opportunity to help determine whether PFAA concentrations affected the humoral antibody response in humans as determined by HI titers following adult seasonal influenza vaccine. The opportunity to determine pre-and postvaccination HI titers allowed 4 measures of immune response—the rise, the ratio, seroconversion, and seroprotection—to be evaluated in relation to concurrent measured serum concentrations of PFOA and PFOS. This design is particularly valuable in the assessment of influenza vaccines, which are designed to protect against commonly circulating viruses, and in adults, for whom immune function is a product of the recent study vaccine response, prior vaccinations, and pre-existing viral exposures. This design allowed assessment of the association of circulating levels of PFAA and the humoral immune response following immunization to 3 influenza subtypes relative to previous influenza-specific antibody levels and also with the attainment of a relative protection antibody threshold. It is important to note that HI titers in vaccine studies serve only as a surrogate for protection from influenza illness or disease, and the absolute antibody titer that confers protection has not been defined.

Age was found to be highly correlated with both pre-existing immunity and vaccine response to some but not all of the subtypes examined. The finding of decreasing antibody response to influenza vaccination with increasing age and history of influenza vaccination in previous years is consistent with a number of studies of influenza vaccine response. Age was carefully adjusted for by using a cubic spline term for age in the adjusted linear regression models. However, the higher vaccine response to influenza B found among men has not been reported elsewhere. This gender difference in the study was isolated to this virus serotype and is not consistent with other published studies (Cook, 2008; Engler et al., 2008; Klein et al., 2010), suggesting that this may be a chance finding.

The most consistent finding in relation to the 2 exposures considered and the 3 flu strains was evidence of a reduced antibody response to A/H3N2 influenza vaccine by higher PFOA concentration, refected in the results for titer rise, titer ratio, and seroprotection, though not seroconversion. confidence intervals were relatively wide, especially when separated into quartile groups, but some results, most notably the reduction in seroprotection, were significantly lower in higher serum PFOA concentration quartiles. These results suggest that individuals with raised PFOA concentrations have an increased risk of not attaining the antibody threshold considered to offer long-term protection from this virus strain. There were no other consistent associations, the only other associations coming close to statistical significance being increases in immune response in for A/H1N1 in relation to PFOS (titer rise) and PFOA (seroconversion and seroprotection, top quartile only).

Previous animal and human studies have suggested PFAAs are associated with a general suppression of humoral immune response; however, we only observed strong evidence of suppression to the A/H3N2 vaccine component. The differences in the ability of the various influenza strains to show a significant effect in relationship to PFOA concentrations may be due to differences in their antigenic determinants. Those strains with more conserved antigenic determinants (epitopes) would likely have a more significant memory cell response. It is generally thought that memory cell responses are less sensitive to chemical immunosuppression than the primary, and thus, response to strains containing more conserved antigenic determinants than other strains might be less likely to show an effect or at least behave kinetically different in their immune response due to a more significant memory cell response. Nonetheless, the lack of consistency between the various endpoints increases the plausibility that the results are due to chance.

In addition to age, a number of medical comorbidities and medications were also considered as potential contributors to, or confounders of, vaccine response; however, the regression models failed to find any evidence of these associations, likely due in part to the low prevalence of these factors in this population. A potential limitation in the study was the reporting of previous vaccination. Analysis revealed the strong effect of previous influenza vaccination on immune response. This information was self-reported by participants using a questionnaire. Without validation, this factor has the potential for recall bias, and the variation in seasonal influenza vaccine year to year would alter the degree to which previous vaccination infers pre-existing immunity against the study vaccine. However, we consider it unlikely that any problems with recall of past vaccinations would be associated with PFAA category.

The reduced influenza vaccine response observed in this study in adults with elevated PFOA concentrations is of note given the findings reported previously in children by Grandjean et al. (2012) and Granum et al. (2013) showing similar effects. In addition to the differences in the ages of the participants, the most significant difference between our study and theirs was in the experimental design. The latter studies collected serum for analysis only prior and not following booster immunizations. Boosters for childhood vaccinations are normally conducted at least 3 years or more following the initial immunization to stimulate memory cell responses. Hence, significant decay of the primary antibody response would have occurred as peak IgG levels appear 21 days postimmunization. Swartz et al. (2003) in a large study monitoring vaccine titers in children following infant immunization showed that the antibody titer decays by > 90% from the primary immunization within 3 years prior to booster. Thus, the titers reported from the previous PFAA studies in children represent differences detected from residual antibody and not the peak response that we monitored. On the other hand, participants in our study had prevaccination titers for all vaccine viruses, especially A/H1N1 which would not have been expected following infant immunization. This pre-existing immunity may have limited the power of our study to capture differences between exposure groups in antibody titer rises.

Our study found no association between PFAA concentration and recent self-reported cold or influenza episodes. It is possible that the extent of suppression in the vaccine response associated with PFOA exposure is insufficient to change in an individual's risk of infections, particularly in a small population. Given the high background rate of respiratory infections, with over 70% of adults reported having experienced an episode in the preceding 12 months, there is not much scope to detect an increase in infections. Furthermore, self-report respiratory infection, particularly influenza, has poor specificity and sensitivity, particularly during times of pandemic virus circulation and subsequent heightened public awareness (Jutel et al., 2011). During the 2009/2010 influenza season in the United States, the main circulating virus was the 2009 pandemic influenza A (H1N1), and very few other seasonal influenza viruses were detected (CDC, 2010). Therefore, if the impact of the A/H3N2 strain of influenza specifically was affected by PFOA serum concentrations, it is plausible that there was too little of this virus type circulating to detect this association in self-reported influenza cases.

In any case, our findings provide evidence that PFOA at serum concentrations between 13.7 and 90 ng/ml, about 4- to 5-fold above the current levels found in the general U.S. population (Kato et al., 2011b), are associated with a reduced HI response to A/H3N2 influenza virus, a commonly occurring flu virus. We found no evidence of an association between self-reported colds or influenza and PFAA concentrations. Furthermore, we saw no evidence that PFOS serum concentrations are associated with reduced vaccine responses.

Supplementary Material

The authors wish to thank the following CDC professionals for their help, advice, and for reagents and viruses: Carolyn B. Bridges, MD, Xiyan Xu, MD, Mary Hoelscher, Kathy Hancock, Jacqueline M. Katz, PhD, Nancy Cox, PhD, and Rubin O. Donis, PhD, Influenza Division, NCIRD, CCID, Centers for Disease Control and Prevention 1600 Clifton Road, Atlanta, GA 30333.

Funding: This research was supported by the C8 Class Action Settlement Agreement (Circuit Court of Wood County, WV) between DuPont and Plaintiffs, which resulted from releases into drinking water of the chemical perfluorooctanoic acid (PFOA, or C8). Funds are administered by an agency which reports to the court. Our work and conclusions are independent of either party to the lawsuit.

Disclaimer The findings and conclusions in this paper are those of the authors and do not necessarily represent the official position of the CDC. The authors declare they have no actual or potential competing financial interests.

Supplementary Data: Supplementary data are available online at http://toxsci.oxfordjournals.org/.

AdmonDEngelhardDStraussNGoldmanNZakay-RonesZ1997Antibody response to influenza immunization in patients after heart transplantationVaccine15151815229330462BeyerWEPalacheAMLuchtersGNautaJOsterhausAD2004Seroprotection rate, mean fold increase, seroconversion rate: Which parameter adequately expresses seroresponse to influenza vaccination?Virus Res10312513215163500BlancifortiLLusterMI2006Considerations in estimating social and economic impacts of immunotoxic agentsHuman Ecol Risk Assess12888903BlumbergEAFitzpatrickJStutmanPCHaydenFGBrozenaSC1998Safety of influenza vaccine in heart transplant recipientsJ Heart Lung Transplant17107510809855446CalafatAMKuklenyikZReidyJACaudillSPTullyJSNeedhamLL2007Serum concentrations of 11 polyfluoroalkyl compounds in the U.S. population: Data from the national health and nutrition examination survey (NHANES)Environ Sci Technol412237224217438769CallSAVollenweiderMAHornungCASimelDLMcKinneyWP2005Does this patient have influenza?JAMA29398799715728170CDC2010Update: Influenza activity—United States, 2009–10 seasonMMWR5990190820671661Centers for Disease Control and Prevention (CDC) and National Centre for Immunization and Respiratory Diseases (NCIRD)2011Seasonal Influenza (Flu)—Key Facts About Seasonal Flu VaccineAvailable at: http://www.cdc.gov/FLU/protect/keyfacts.htmAccessed April 5, 2012CookIF2008Sexual dimorphism of humoral immunity with human vaccinesVaccine262930CorsiniEAvogadroAGalbiatiVdell'AgliMMarinovichMGalliCLGermolecDR2011In vitro evaluation of the immunotoxic potential of perfluorinated compounds (PFCs)Toxicol Appl Pharmacol25010811621075133CruijffMThijsCGovaertTAretzKDinantGJKnottnerusA1999The effect of smoking on influenza, influenza vaccination efficacy and on the antibody response to influenza vaccinationVaccine1742643210073719CullenGBaderCKorzenikJRSandsBE2012Serological response to the 2009 H1N1 influenza vaccination in patients with inflammatory bowel diseaseGut6138539121757451DenglerTJStrnadNBuhringIZimmermannRGirgsdiesOKublerWEZielenS1998Differential immune response to influenza and pneumococcal vaccination in immunosuppressed patients after heart transplantationTransplantation66134013479846520Department of Health2011Seasonal flu plan Winter 2011/12Available at: http://www.dh.gov.uk/en/Publicationsandstatistics/Publications/Publications-PolicyAndGuidance/DH_127051Accessed July 12, 2011DewittJCCopelandCBStrynarMJLuebkeRW2008Perfluorooctanoic acid-induced immunomodulation in adult C57BL/6J or C57BL/6N female miceEnviron Health Perspect11664465018470313DeWittJCPeden-AdamsMMKellerJMGermolecDR2012Immunotoxicity of perfluorinated compounds: Recent developmentsToxicol Pathol4030031122109712EnglerRJNelsonMRKloteMMVanRadenMJHuangCYCoxNJKlimovAKeitelWANicholKLCarrWW2008Half- vs full-dose trivalent inactivated influenza vaccine (2004–2005): Age, dose, and sex effects on immune responsesArch Intern Med1682405241419064822FairPADriscollEMollenhauerMABradshawSGYunSHKannanKBossartGDKeilDEPeden-AdamsMM2011Effects of environmentally-relevant levels of perfluorooctane sulfonate on clinical parameters and immunological functions in B6C3F1 miceJ Immunotoxicol8172921261439FDA, Office of Communications2007Guidance for Industry, Clinical Data Needed to Support the Licensure of Seasonal Inactivated Influenza VaccinesAvailable at: http://www.fda.gov/cber/vaccine/vacpubs.htmAccessed November 14, 2009Fitz-SimonNFletcherTLusterMISteenlandKCalafatAMKatoKArmstrongB2013Reductions in serum lipids with a 4-year decline in serum perfluorooctanoic acid and perfluorooctanesulfonic acidEpidemiology2456957623685825FletcherTSteenlandKSavitzD2011C8 Science Panel. Status Report: PFOA and Immune Biomarkers in Adults Exposed to PFOA in Drinking Water in the Mid Ohio ValleyAvailable at: http://www.c8science-panel.org/pdfs/Status_Report_C8_and_Immune_markers_March2009.pdfAccessed April 5, 2012FrisbeeSJBrooksAPJrMaherAFlensborgPArnoldSFletcherTSteenlandKShankarAKnoxSSPollardC2009The C8 health project: Design, methods, and participantsEnviron Health Perspect1171873188220049206FrommeHMidaschOTwardellaDAngererJBoehmerSLieblB2007Occurrence of perfluorinated substances in an adult German population in southern BavariaInt Arch Occup Environ Health8031331916915390GlynnABergerUBignertAUllahSAuneMLignellSDarnerudPO2012Perfluorinated alkyl acids in blood serum from primiparous women in Sweden: Serial sampling during pregnancy and nursing, and temporal trends 1996–2010Environ Sci Technol469071907922770559GoodwinKViboudCSimonsenL2006Antibody response to inflenza vaccination in the elderly: A quantitative reviewVaccine241159116916213065GoronzyJJFulbrightJWCrowsonCSPolandGAO'FallonWMWeyandCM2001Value of immunological markers in predicting responsiveness to influenza vaccination in elderly individualsJ Virol75121821218711711609GrandjeanPAndersenEWBudtz-JorgensenENielsenFMolbakKWeihePHeilmannC2012Serum vaccine antibody concentrations in children exposed to perfluorinated compoundsJAMA30739139722274686GranumBSolvorLSStølevikBThomsenCAabergeISvan LoverenHNygaardUC2013Pre-natal exposure to perfluoroalkyl substances may be associated with altered vaccine antibody levels and immune-related health outcomes in early childhoodJ Immunotoxicol1037337923350954HaradaKHYangHRMoonCSHungNNHitomiTInoueKNiisoeTWatanabeTKamiyamaSTakenakaK2010Levels of perfluorooctane sulfonate and perfluorooctanoic acid in female serum samples from Japan in 2008, Korea in 1994–2008 and Vietnam in 2007–2008Chemosphere7931431920149408HolvastAvan AssenSde HaanAHuckriedeABenneCAWestraJPalacheAWilschutJKallenbergCGBijlM2009Studies of cell-mediated immune responses to influenza vaccination in systemic lupus erythematosusArthritis Rheum602438244719644961JinYSaitoNHaradaKHInoueKKoizumiA2007Historical trends in human serum levels of perfluorooctanoate and perfluorooctane sulfonate in Shenyang, ChinaTohoku J Exp Med212637017464105JutelABakerMGStanleyJHuangQSBandaranayakeD2011Self-diagnosis of influenza during a pandemic: A cross-sectional surveyBMJ Open1e00023410.1136/bmjopen-2011-000234KannanKCorsoliniSFalandyszJFillmannGKumarKSLoganathanBGMohdMAOliveroJVan WouweNYangJH2004Perfluorooctanesulfonate and related fluorochemicals in human blood from several countriesEnviron Sci Technol384489449515461154KatoKBasdenBJNeedhamLLCalafatAM2011aImproved selectivity for the analysis of maternal serum and cord serum for polyfluoroalkyl chemicalsJ Chromat A12182133213710.1016/j.chroma.2010.10.051KatoKWongLYJiaLTKuklenyikZCalafatAM2011bTrends in exposure to polyfluoroalkyl chemicals in the U.S. Population: 1999–2008Environ Sci Technol458037804521469664KleinSLPassarettiCAnkerMOlukoyaPPekoszA2010The impact of sex, gender and pregnancy on 2009 H1N1 diseaseBiol Sex Diff1511MuraskoDMBernsteinEDGardnerEMGrossPMunkGDranSAbrutynE2002Role of humoral and cell-mediated immunity in protection from influenza disease after immunization of healthy elderlyExp Gerontol3742743911772530OlsenGWLangeCCEllefsonMEMairDCChurchTRGoldbergCLHerronRMMedhdizadekashiZNobilettiJBRiosJA2012Temporal trends of perfluoroalkyl concentrations in American red cross adult blood donors, 2000–2010Environ Sci Technol466330633822554481OlsenGWMairDCReagenWKEllefsonMEEhresmanDJButenhoffJLZobelLR2007Preliminary evidence of a decline in perfluorooctanesulfonate (PFOS) and perfluorooctanoate (PFOA) concentrations in American Red Cross blood donorsChemosphere6810511117267015PedersenAFZachariaeRBovbjergDH2009Psychological stress and antibody response to influenza vaccination: A meta-analysisBrain Behav Immun2342743319486657RobertsonJDNageshKJowittSNDougalMAndersonHMuttonKZambonMScarffeJH2000Immunogenicity of vaccination against influenza, Streptococcus pneumoniae and Haemophilus influenzae type B in patients with multiple myelomaBr J Cancer821261126510755398SallesMJSensYABoasLSMachadoCM2010Influenza virus vaccination in kidney transplant recipients: Serum antibody response to different immunosuppressive drugsClin Transplant24E17E2319758368Schroter-KermaniCMullerJJurlingHConradASchulteC2012Retrospective monitoring of perfluorocarboxylates and perfluorosulfonates in human plasma archived by the German Environmental Specimen BankInt J Hyg Environ Health1010161022SteenlandKFletcherTSavitzDA2010Epidemiologic evidence on the health effects of perfluorooctanoic acid (PFOA)Environ Health Perspect1181100110820423814SwartzTASaliouPCatznelsonEBlondeauCGilIPeledTHavkinOFletcherM2003Immune response to a diphtheria and tetanus toxoid administration in a three-dose diphtheria tetanus whole-cell pertussis/enhanced inactivated poliovirus vaccination schedule: A 7-year follow upEur J Epidemiol1882783312974560TiuCTLinYSPagalaMGhitanMTreanorJJFitzgeraldTXiaoYYUshumirskiySWongSSJavierEM2011Antibody response to inactivated influenza A (H1N1) 2009 monovalent vaccine in patients with and without HIVJ Acquir Immune Defc Syndr58e99e102VedharaKCoxNKWilcockGKPerksPHuntMAndersonSLightmanSLShanksNM1999Chronic stress in elderly carers of dementia patients and antibody response to influenza vaccinationLancet35362763110030328VilchezRAFungJKusneS2002The pathogenesis and management of influenza virus infection in organ transplant recipientsTranspl Infect Dis417718212535259WeihePKatoKCalafatAMNielsenFWanigatungaAANeedhamLLGrandjeanP2008Serum concentrations of polyfluoroalkyl compounds in Faroese whale meat consumersEnviron Sci Technol426291629518767701WekslerMESzaboP2000The effect of age on the B-cell repertoireJ Clin Immunol2024024910939711WickGGrubeck-LoebensteinB1997Primary and secondary alterations of immune reactivity in the elderly: Impact of dietary factors and diseaseImmunol Rev1601711849476675YangQAbedi-ValugerdiMXieYZhaoXYMollerGNelsonBDDePierreJW2002Potent suppression of the adaptive immune response in mice upon dietary exposure to the potent peroxisome proliferator, perfluorooctanoic acidInt Immunopharmacol238939711811941Baseline Participant Characteristics, PFOA/PFOS Serum Concentrations (ng/ml), and Prevaccination GMTs (<italic>n</italic> = 411)
Number (%)Median Log10 PFOA Serum Concentration at Vaccination (IQR)Median Log10 PFOS Serum Concentration at Vaccination (IQR)Influenza Type B GMT (95% CI)Influenza A H1N1 GMT (95% CI)Influenza A H3N2 GMT (95% CI)
Gender
 Female200 (48.7)1.37 (1.03, 1.77)0.84 (0.64, 1.01)9.36 (8.28, 10.58)16.64 (13.56, 20.43)26.21 (21.35, 32.17)
 Male211 (51.3)1.64 (1.28, 2.03)1.09 (0.91, 1.23)8.16 (7.26, 9.17)15.63 (12.62, 19.36)16.53 (13.46, 20.31)
Age (years)
 < 3049 (11.9)1.43 (1.05, 1.66)0.93 (0.82, 1.05)8.80 (6.79, 11.41)18.37 (12.37, 27.28)16.88 (10.76, 26.47)
 31–40107 (26.0)1.23 (1.03, 1.79)0.89 (0.64, 1.09)9.13 (7.63, 10.93)19.62 (14.09, 27.30)18.87 (14.30, 24.88)
 41–50109 (26.5)1.54 (1.15, 1.81)0.95 (0.74, 1.14)8.81 (7.56, 10.26)18.18 (13.17, 25.09)16.74 (12.57, 22.30)
 51–60107 (26.0)1.59 (1.26, 2.05)1.04 (0.84, 1.21)8.50 (7.24, 9.99)12.71 (10.01, 16.13)25.75 (19.43, 34.13)
 > 6039 (9.5)1.83 (1.65, 2.26)1.09 (0.92, 1.35)7.94 (5.87, 10.74)10.93 (7.45, 16.03)34.09 (20.28, 57.29)
Current smoker
 No331 (80.5)1.52 (1.14, 1.96)0.96 (0.76, 1.17)8.49 (7.73, 9.34)15.52 (13.17, 18.30)21.39 (18.17, 25.17)
 Yes80 (19.5)1.45 (1.11, 1.73)0.96 (0.76, 1.12)9.74 (8.08, 11.75)18.82 (13.36, 26.52)18.03 (12.84, 25.31)
Any previous influenza vaccine (not within last 3 months)
 No121 (29.4)1.46 (1.15, 1.90)1.05 (0.79, 1.22)6.04 (5.49, 6.64)9.83 (7.99, 12.10)7.60 (6.51, 8.87)
 Yes290 (70.6)1.52 (1.13, 1.98)0.94 (0.74, 1.13)10.17(9.12, 11.33)19.81 (16.44, 23.87)31.42 (26.34, 37.48)
H1N1 vaccine in last influenza season
 No181 (62.4)1.56 (1.16, 2.03)0.96 (0.77, 1.14)9.85 (8.63, 11.24)11.61 (9.74, 13.84)26.25 (21.03, 32.76)
 Yes108 (37.2)1.44 (1.11, 1.84)0.91 (0.72, 1.12)10.59(8.76, 12.81)47.57 (33.61, 67.32)42.65 (31.93, 56.97)
 Don't know1 (0.3)0.82 (0.82, 0.82)0.04 (0.04, 0.04)40.00 (—, —)160.00 (—, —)20.00 (—, —)
Mobility (number of addresses since birth or 1970 if born prior to 1970)
 1–356 (13.6)1.79 (1.40, 2.11)1.06 (0.86, 1.26)7.62 (6.11, 9.49)12.50 (8.51, 18.36)21.28 (14.65, 30.91)
 4–6205 (49.9)1.49 (1.17, 1.86)0.94 (0.76, 1.16)8.65 (7.73, 9.68)15.42 (12.52, 18.98)20.76 (16.86, 25.55)
 7–987 (21.2)1.43 (1.03, 1.86)0.96 (0.79, 1.17)8.33 (6.77, 10.24)17.33 (12.41, 24.19)17.47 (12.72, 23.98)
 ≥ 1063 (15.3)1.35 (1.03, 1.82)0.94 (0.67, 1.11)10.80 (8.58, 13.60)21.13 (14.21, 31.43)25.20 (16.64, 38.15)
PFOA and PFOS serum concentrations at time of vaccination
 PFOAa
  First quartile103 (25.1)0.93 (0.67, 1.04)0.75 (0.45, 0.93)9.60 (8.04, 11.47)18.95 (14.01, 25.64)23.99 (18.09, 31.80)
  Second quartile103 (25.1)1.32 (1.21, 1.42)0.99 (0.76, 1.20)7.95 (6.85, 9.24)18.08 (13.27, 24.64)17.13 (12.73, 23.05)
 Third quartile103 (25.1)1.69 (1.60, 1.81)1.04 (0.81, 1.20)8.34 (7.07, 9.84)16.57 (12.14, 22.60)15.49 (11.98, 20.02)
  Fourth quartile102 (24.8)2.17 (2.07, 2.34)1.09 (0.91, 1.25)9.09 (7.54, 10.96)11.85 (9.10, 15.43)28.87 (20.75, 40.15)
 PFOSb
  First quartile106 (25.8)1.11 (0.75, 1.44)0.58 (0.34, 0.67)9.87 (8.34, 11.69)17.21 (12.75, 23.23)21.92 (16.17, 29.70)
  Second quartile101 (24.6)1.57 (1.11, 2.00)0.88 (0.83, 0.92)8.96 (7.55, 10.64)15.95 (11.77, 21.60)23.26 (17.58, 30.78)
  Third quartile102 (24.8)1.62 (1.25, 2.05)1.06 (1.02, 1.10)8.16 (6.86, 9.70)17.94 (13.09, 24.59)22.15 (16.32, 30.04)
  Fourth quartile102 (24.8)1.72 (1.43, 2.12)1.27 (1.21, 1.38)7.99 (6.76, 9.45)13.67 (10.39, 17.99)16.20 (12.13, 21.65)

Note. Range of PFAA concentrations within quartiles:

PFOA: First quartile: 0.25–13.7 ng/ml; second quartile: 13.8–31.5 ng/ml; third quartile: 31.6–90 ng/ml; fourth quartile: 90.4–2140 ng/ml.

PFOS: First quartile: 0.1–5.8 ng/ml; second quartile: 5.9–9.2 ng/ml; third quartile: 9.3–14.5 ng/ml; fourth quartile: 14.7–42.3 ng/ml.

GMT Rise Following Vaccination (95% CI) by Quartile of PFOA/PFOS Serum Concentration (<italic>n</italic> = 403)
Influenza Type BInfluenza Type A H1N1Influenza A H3N2
GMT (95% CI)GMT (95% CI)GMT (95% CI)
PFOA and PFOS serum concentrations at time of vaccination
PFOAa
 First quartile49.46 (38.14, 64.12)476.23 (360.77, 628.65)228.86 (161.53, 324.27)
 Second quartile46.00 (35.28, 59.97)352.22 (255.33, 485.88)125.36 (86.01, 182.73)
 Third quartile43.55 (33.08, 57.33)306.32 (232.58, 403.44)104.13 (72.47, 149.64)
 Fourth quartile20.90 (16.58, 28.24)274.79 (202.85, 372.23)183.73 (127.28, 265.23)
PFOSb
 First quartile42.25 (33.41, 53.42)342.31 (256.04, 457.65)137.69 (98.70, 192.09)
 Second quartile41.48 (30.73, 55.99)280.44 (197.64, 397.92)147.27 (99.27, 218.47)
 Third quartile41.13 (31.65, 53.44)417.74 (318.99, 547.07)211.02 (141.24, 315.27)
 Fourth quartile52.83 (38.92, 71.71)341.82 (258.03, 452.81)126.69 (88.85, 180.65)

Note. Bold values represent significant associations. Range of PFAA concentrations within quartiles:

PFOA: First quartile: 0.25–13.7 ng/ml; second quartile: 13.8–31.5 ng/ml; third quartile: 31.6–90 ng/ml; fourth quartile: 90.4–2140 ng/ml.

PFOS: First quartile: 0.1–5.8 ng/ml; second quartile: 5.9–9.2 ng/ml; third quartile: 9.3–14.5 ng/ml; fourth quartile: 14.7–42.3 ng/ml.

Linear regression Coefficients of Log<sub>10</sub>-Transformed influenza Antibody Titer rise and Log<sub>10</sub>-Transformed influenza Antibody Titer ratio With unit increase in Log<sub>10</sub> Transformed and Quartiles of PFOA and PFOS Serum Concentration (<italic>n</italic> = 403)
Log PFOA (ng/ml)Log PFOS (ng/ml)
Regression Coefficient95% CIp ValueRegression Coefficient95% CIp Value
a. Log10-transformed PFOA/PFOS (as continuous variable)
Influenza type B
 Log10-transformed antibody titer rise (n = 359)a
  Unadjusted−.05(−0.16, 0.05).33.09(−0.07, 0.24).29
  Adjustedb−.02(−0.13, 0.09).73.05(−0.11, 0.21).56
 Log10-transformed antibody titer ratio (postvaccine: prevaccine)
  Unadjusted−.05(−0.15, 0.05).30.10(−0.04, 0.25).16
  Adjustedb−.02(−0.11, 0.08).73.05(−0.09, 0.18).52
Influenza A/H1N1
 Log10-transformed antibody titer rise (n = 322)a
  Unadjusted−.11(−0.22, 0.01).07.09(−0.08, 0.25).32
  Adjustedb−.03(−0.14, 0.09).63.15(−0.02, 0.32).08
 Log10-transformed antibody titer ratio (postvaccine: prevaccine)
  Unadjusted.02(−0.12, 0.15).79.12(−0.07, 0.32).22
  Adjustedb.07(−0.06, 0.21).30.10(−0.11, 0.30).36
Influenza A/H3N2
 Log10-transformed antibody titer rise(n = 372)a
  Unadjusted−.02(−0.16, 0.13).81.08(−0.13, 0.28).47
  Adjustedb−.01(−0.17, 0.14).86.09(−0.13, 0.32).42
 Log10-transformed antibody titer ratio (postvaccine: prevaccine)
  Unadjusted−.15(−0.28, −0.02).03.04(−0.15, 0.24).68
  Adjustedb−.12(−0.25, 0.02).09−.005(−0.20, 0.19).96
PFOA Quartiles (ng/ml)cPFOS Quartiles (ng/ml)d
Regression Coefficient95% CIp ValueRegression Coefficient95% CIp Value
b. Quartiles of PFOA/PFOS concentration
Influenza type B
 Log10-transformed antibody titer rise (n = 359)e
  Adjustedb
   First quartileReferentReferent
   Second quartile−.03(−0.19, 0.13).69.02(−0.13, 0.18).76
   Third quartile−.02(−0.19, 0.15).82−.03(−0.19, 0.14).73
   Fourth quartile−.07(−0.24, 0.10).42.04(−0.14, 0.21).68
 Log10-transformed antibody titer ratio
  Adjustedb
   First quartileReferentReferent
   Second quartile.05(−0.09, 0.19).53.004(−0.14, 0.14).96
   Third quartile.07(−0.07, 0.22).32−.02(−0.16, 0.12).78
   Fourth quartile−.03(−0.17, 0.12).71.03(−0.12, 0.18).71
Influenza A/H1N1
 Log10-transformed antibody titer rise (n = 322)e
  Adjustedb
   First quartileReferentReferent
   Second quartile−.09(−0.27, 0.08).31−.04(−0.21, 0.14).68
   Third quartile−.10(−0.28, 0.09).30.13(−0.04, 0.31).14
   Fourth quartile−.12(−0.30, 0.06).19.10(−0.09, 0.29).30
 Log10-transformed antibody titer ratio
  Adjustedb
   First quartileReferentReferent
   Second quartile−.08(−0.29, 0.12).43−.07(−0.28, 0.13).47
   Third quartile−.04(−0.25, 0.18).72.03(−0.18, 0.24).78
   Fourth quartile.07(−0.14, 0.29).51.03(−0.19, 0.26).77
Influenza A/H3N2
 Log10-transformed antibody titer rise (n = 372)e
  Adjustedb
   First quartileReferentReferent
   Second quartile−.28(−0.51, -0.06).02.03(−0.19, 0.26).78
   Third quartile−.37(−0.60, −0.13).002.18(−0.06, 0.41).14
   Fourth quartile−.12(−0.36, 0.13).35−.04(−0.28, 0.21).77
 Log10-transformed antibody titer ratio
  Adjustedb
   First quartileReferentReferent
   Second quartile−.10(−0.30, 0.10).31−.06(−0.26, 0.14).56
   Third quartile−.07(−0.28, 0.14).49.02(−0.18, 0.23).84
   Fourth quartile−.22(−0.43, -0.01).04−.03(−0.24, 0.19).82

Note. Log10-transformed antibody titer rise missing for participants with untransformed antibody titer rise of 0.

Adjusted for age (cubic spline), gender, mobility, and history of previous influenza vaccination.

Range of PFAA concentrations within quartiles:

PFOA: First quartile: 0.25–13.7 ng/ml; second quartile: 13.8–31.5 ng/ml; third quartile: 31.6–90 ng/ml; fourth quartile: 90.4–2140 ng/ml.

PFOS: First quartile: 0.1–5.8 ng/ml; second quartile: 5.9–9.2 ng/ml; third quartile: 9.3–14.5 ng/ml; fourth quartile: 14.7–42.3 ng/ml.

Log10-transformed antibody titer rise missing for participants with untransformed antibody titer rise of 0.

Logistic Regression OR (95% CI) of Seroconversion (4-Fold increase in Antibody Titer Following Vaccination) for Log<sub>10</sub> Transformed and Quartiles of PFOA and PFOS Concentration (ng/ml) (<italic>n</italic> = 403)
Number Who Seroconvert Following Vaccination (Row%)Adjusted OR (95% CI)ap Valuep Value for Heterogeneity Across Quartilesp Value for Trend Over Quartiles
Influenza type B—odds ratio for seroconversion
 Log10 transformed (as continuous variable, odds ratio per unit rise in PFOC measure)
  Log10 PFOA250/403 (62.0%)0.80 (0.53, 1.21).30
  Log10 PFOS250/403 (62.0%)1.17 (0.63, 2.17).63
 Quartiles (as categorical variable)
  PFOAa
   First quartile61/100 (61.0%)1.09.28
   Second quartile71/102 (69.6%)1.43 (0.76, 2.70).26
   Third quartile66/100 (66.0%)1.39 (0.73, 2.66).32
   Fourth quartile52/101 (51.5%)0.71 (0.38, 1.36).30
  PFOSb
   First quartile67/104 (64.4%)1.76.75
   Second quartile54/97 (55.7%)0.72 (0.39, 1.33).29
   Third quartile61/100 (61.0%)0.81 (0.42, 1.53).51
   Fourth quartile68/102 (66.7%)0.87 (0.43, 1.74).69
Influenza A H1N1—odds ratio for seroconversion
 Log10 transformed (as continuous variable, odds ratio per unit rise in PFAA measure)
  Log10 PFOA339/403 (84.1%)1.51 (0.89, 2.56).12
  Log10 PFOS339/403 (84.1%)1.10 (0.51, 2.37).80
 Quartiles (as categorical variable)
  PFOA
   First quartile83/100 (83.0%)1
   Second quartile82/102 (80.4%)0.74 (0.34, 1.59).44
   Third quartile83/100 (83.0%)1.11 (0.49, 2.50).80
   Fourth quartile91/101 (90.1%)2.23 (0.90, 5.53).08.07.05
  PFOS
   First quartile88/104 (84.6%)1
   Second quartile81/97 (83.5%)0.97 (0.44, 2.14).94
   Third quartile82/100 (82.0%)0.78 (0.35, 1.75).55
   Fourth quartile88/102 (86.3%)0.94 (0.38, 2.31).90.93.76
Influenza A H3N2—odds ratio for seroconversion
 Log10 transformed (as continuous variable, odds ratio per unit rise in PFAA measure)
  Log10 PFOA261/403 (64.8%)0.76 (0.51, 1.15).19
   Log10 PFOS261/403 (64.8%)1.17 (0.63, 2.15).62
 Quartiles (as categorical variable)
  PFOA
   First quartile68/100 (68.0%)1
   Second quartile68/102 (66.7%)0.90 (0.48, 1.68).75
   Third quartile69/100 (69.0%)1.13 (0.59, 2.17).71
   Fourth quartile56/101 (55.4%)0.62 (0.33, 1.16).14.22.20
  PFOS
   First quartile65/104 (62.5%)1
   Second quartile60/97 (61.9%)1.08 (0.59, 1.97).81
   Third quartile64/100 (64.0%)1.10 (0.59, 2.06).76
   Fourth quartile72/102 (70.6%)1.41 (0.72, 2.78).32.78.35

Note. Adjusted for age (cubic spline), gender, mobility, and history of previous influenza vaccination. Range of PFAA concentrations within quartiles: PFOA: First quartile: 0.25–13.7 ng/ml; second quartile: 13.8–31.5 ng/ml; third quartile: 31.6–90 ng/ml; fourth quartile: 90.4–2140 ng/ml.

PFOS: First quartile: 0.1–5.8 ng/ml; second quartile: 5.9–9.2 ng/ml; third quartile: 9.3–14.5 ng/ml; fourth quartile: 14.7–42.3 ng/ml.

Logistic Regression OR (95% CI) of Seroprotection (HI Antibody Titer ≥ 1:40 Following Vaccination) for Log<sub>10</sub> Transformed and Quartiles of PFOA and PFOS Concentration (ng/ml) (<italic>n</italic> = 403)
Patient DemographicNumber Are Seroprotected Following Vaccination (Row%)Adjusted OR (95% CI)app Value for Heterogeneity Across Quartilesp Value of for TrendOver Quartiles
Influenza type B—odds ratio for seroprotection
 Log10 transformed (as continuous variable, odds ratio per unit rise in PFAA measure)
  Log10 PFOAa266/403 (66.0%)1.04 (0.68, 1.60).85
  Log10 PFOSb266/403 (66.0%)0.85 (0.44, 1.64).63
 Quartiles (as categorical variable)
  PFOA
   First quartile72/100 (72.0%)1.52.68
   Second quartile66/102 (64.7%)0.76 (0.40, 1.45).41
   Third quartile68/100 (68.0%)1.13 (0.57, 2.23).73
   Fourth quartile60/101 (59.4%)0.77 (0.39, 1.50).44
  PFOS
   First quartile75/104 (72.1%)1.63.50
   Second quartile59/97 (60.8%)0.67 (0.35, 1.25).21
   Third quartile66/100 (66.0%)0.82 (0.42, 1.59).55
   Fourth quartile66/102 (64.7%)0.73 (0.36, 1.47).38
Influenza A H1N1—odds ratio for seroprotection
 Log10 transformed (as continuous variable, odds ratio per unit rise in PFAA measure)
  Log10 PFOA385/403 (95.5%)2.34 (0.91, 6.07).08
   Log10 PFOS385/403 (95.5%)0.93 (0.23, 3.71).91
 Quartiles (as categorical variable)
  PFOA
   First quartile97/100 (97.0%)1.04.02
   Second quartile95/102 (93.1%)0.74 (0.17, 3.28).69
   Third quartile94/100 (94.0%)1.59 (0.33, 7.70).57
   Fourth quartile99/101 (98.0%)6.47 (0.91, 45.85).06
  PFOS
   First quartile101/104 (97.1%)1.40.41
   Second quartile90/97 (92.8%)0.55 (0.13, 2.37).42
   Third quartile97/100 (97.0%)1.81 (0.32, 10.22).50
   Fourth quartile97/102 (95.1%)1.26 (0.24, 6.61).79
Influenza A H3N2—odds ratio for seroprotection
 Log10 transformed (as continuous variable, odds ratio per unit rise in PFAA measure)
  Log10 PFOA338/403 (83.9%)0.66 (0.39, 1.12).12
  Log10 PFOS338/403 (83.9%)0.63 (0.26, 1.49).29
 Quartiles (as categorical variable)
  PFOA
   First quartile91/100 (91.0%)1.03.07
   Second quartile82/102 (80.4%)0.34 (0.14, 0.83).02
   Third quartile80/100 (80.0%)0.28 (0.11, 0.70).01
   Fourth quartile85/101 (84.2%)0.39 (0.15, 0.99).05
  PFOS
   First quartile89/104 (85.6%)1.36.24
   Second quartile82/97 (84.5%)0.85 (0.38, 1.88).68
   Third quartile87/100 (87.0%)1.09 (0.47, 2.56).84
   Fourth quartile80/102 (78.4%)0.56 (0.24, 1.28).17

Note. aAdjusted for age (cubic spline), gender, mobility, and history of previous influenza vaccination.

Range of PFAA concentrations within quartiles:

PFOA: First quartile: 0.25–13.7 ng/ml; second quartile: 13.8–31.5 ng/ml; third quartile: 31.6–90 ng/ml; fourth quartile: 90.4–2140 ng/ml.

PFOS: First quartile: 0.1–5.8 ng/ml; second quartile: 5.9–9.2 ng/ml; third quartile: 9.3–14.5 ng/ml; fourth quartile: 14.7–42.3 ng/ml.

Self-reported Cold and Flu infections in 12 Months Preceding Questionnaire (<italic>n</italic> = 755)
PFOAPFOS
Number Reporting Episode in Last 12 Months (%)Adjusted OR (95%CI)ap ValueNumber Reporting Episode in Last 12 Months (%)Adjusted OR (95%CI)ap Value
OR for any “flu” infection in last 12 months
 Log10-transformed PFOA/PFOS (continuous)163/755 (21.6%)0.98 (0.70, 1.38).92163/755 (21.6%)0.97 (0.58, 1.63).91
 PFOA/PFOS exposure group (categorical)
  First quartile45/191 (23.6%)151/193 (26.4%)1
  Second quartile37/189 (19.6%)0.91 (0.55, 1.51)39/187 (20.9%)0.83 (0.51, 1.35)
  Third quartile44/188 (23.4%)1.30 (0.79, 2.13)33/190 (17.4%)0.74 (0.44, 1.24)
  Fourth quartile37/187 (19.8%)1.09 (0.65, 1.83).5740/185 (21.6%)1.20 (0.70, 2.04).29
OR for any cold in last 12 months
 Log10-transformed PFOA (continuous)538/755 (71.3%)0.83 (0.61, 1.13).23538/755 (71.3%)0.83 (0.51, 1.34).44
 PFOA/PFOS exposure group (categorical)
   First quartile145/191 (75.9%)1140/193 (72.5%)1
   Second quartile141/189 (74.6%)1.13 (0.69, 1.83)146/187 (78.1%)1.60 (0.99, 2.60)
   Third quartile131/188 (69.7%)0.97 (0.60, 1.57)134/190 (70.5%)1.20 (0.75, 1.92)
   Fourth quartile121/187 (64.7%)0.80 (0.50, 1.29).53118/185 (63.8%)1.09 (0.67, 1.78).24
OR for either a cold or “flu” in last 12 months
 Log10-transformed PFOA (continuous)554/755 (73.4%)0.85 (0.62, 1.16).31554/755 (73.4%)0.90 (0.55, 1.48).68
 PFOA/PFOS exposure group (categorical)
   First quartile148/191 (77.5%)1145/193 (75.1%)1
   Second quartile145/189 (76.7%)1.21 (0.73, 2.00)150/187 (80.2%)1.66 (1.00, 2.75)
   Third quartile137/188 (72.9%)1.10 (0.67, 1.81)137/190 (72.1%)1.19 (0.74, 1.94)
   Fourth quartile124/187 (66.3%)0.84 (0.52, 1.36).46122/185 (66.0%)1.15 (0.69, 1.91).24

Note. Adjusted for age (cubic spline) and gender.

Frequency of Cold Infections in 12 Months Preceding Questionnaire (<italic>n</italic> = 755)
Number of Colds Reported in Last 12 MonthsParticipants Reporting This Frequency of Colds (%)
0217 (28.7%)
1404 (53.5%)
2122 (16.2%)
39 (1.2%)
43 (0.4%)
PFOAPFOS
Adjusted OR (95%CI)aLRTb for Null vs General Association p ValueAdjusted OR (95%CI)aLRT for Null vs General Association p Value
OR for frequency of colds in last 12 months
 Log10-transformed PFOA/PFOS (continuous)0.91 (0.70, 1.19).510.89 (0.60, 1.33).58
 PFOA/PFOS exposure group (categorical)
  First quartile11
  Second quartile0.93 (0.63, 1.38)1.56 (1.05, 2.31)
  Third quartile0.95 (0.63, 1.41)1.05 (0.70, 1.56)
  Fourth quartile0.90 (0.60, 1.36).971.10 (0.71, 1.69).10

Note. Adjusted for age (cubic spline) and gender.

Likelihood-ratio test.