Urinary Polycyclic Aromatic Hydrocarbons Concentrations and Hepatitis B Antibody Serology in the United States (NHANES, 2003–2014)

Abstract

Background:

Polycyclic aromatic hydrocarbons (PAHs) are environmental contaminants that are hepatotoxic and immunotoxic. PAH exposure may modulate hepatitis B immunology.

Objective:

We used data from 6 cycles of the National Health and Nutrition Examination Survey (2003–2014) to evaluate the associations between urinary PAH metabolites and hepatitis B serology.

Methods:

This analysis included individuals who self-reported receiving ≥ 3 doses of hepatitis B vaccine and urinary PAH metabolites (i.e. 1-napthol, 2-napthol, 3-fluorene, 2-fluorene, 1-phenanthrene, 1-pyrene, and total PAH [sum of all metabolites]). Separate logistic regression models assessed the association between hepatitis B vaccination status (i.e. individuals who were immune due to vaccination or susceptible) and tertiles of urinary PAH. Models were adjusted for age, gender, race/ethnicity, survey cycle, family income to poverty ratio, BMI, country of birth, serum cotinine, and urinary creatinine.

Results:

Among participants who reported receiving ≥ 3 doses of vaccine and had no antibodies indicating a history of hepatitis B infection and/or current hepatitis B infection, dose-response relationships were observed where individuals with the lowest odds of serology indicating a response to the hepatitis B vaccine (i.e., anti-HBs⁺, anti-HBc^–, and HBsAg^–) were in the highest tertile of 2-Napthol (adjusted Odds Ratio [aOR]: 0.70, 95% confidence interval [CI]: 0.54, 0.91), 3-Napthol (aOR: 0.68, 95% CI: 0.53, 0.87), 2-Fluorene (aOR: 0.61, 95% CI: 0.54, 0.86), 1-Phenanthrene (aOR: 0.79, 95% CI: 0.65, 0.97), 1-Pyrene (aOR): 0.68, 95% CI: 0.55, 0.83), and total PAH (aOR: 0.73, 95% CI: 0.56, 0.95) had the compared to the lowest tertile.

Conclusion:

This cross-sectional study supports a hypothesis that PAH exposures experienced by the general US population may modulate hepatitis B vaccine induced immunity. Given the ubiquity of PAH exposures in the US, additional research is warranted to explore the effects of chronic PAH exposures on hepatitis B related humoral immunity.

1. Introduction

Hepatitis B is a highly infectious disease of the liver caused by the hepatitis B virus (HBV) and affects approximately 260 million people worldwide (3). Recently, incident HBV infections have increased in regions of the United States (U.S.) that have experienced higher injection drug use during the opioid epidemic (4). Individuals infected with HBV can develop acute or chronic infections. People who experience acute HBV infection may be asymptomatic, experience flu-like symptoms, or jaundice. The infection typically resolves itself and 70–90% of these individuals will develop natural immunity to HBV, but the remainder will develop chronic HBV infection which can substantially increases their risk of cirrhosis and hepatocarcinoma (5). An effective vaccine against HBV was developed in 1982 and a 3-dose vaccination series was recommended for children in the United States starting in 1991 (3,6). The Centers for Disease Control (CDC) reports that the efficacy of the hepatitis B vaccination varies by age, ranging from 90% - 95% in individuals with average immunity for 20 years or more (7).

Several factors have been associated with nonresponse and/or waning of hepatitis B vaccine induced immunity including increasing age, higher body mass index (BMI), current smoking, and vaccine type (7,8). Cigarette smoke contains polycyclic aromatic hydrocarbons (PAHs) that are known to be both hepatotoxic and immunotoxic, raising the question that PAHs may modulate the biological functions that contribute to hepatitis B related immunity (9). PAHs are produced during combustion, and people are exposed to PAHs from the ingestion of grilled or smoked food, inhalation of air pollution, and occupational exposures (10). PAHs are metabolized in the liver through cytochrome P450 (CYP) enzymes and are excreted in urine in a few hours to days (11). Thus, urinary PAH metabolites are considered good biomarkers of recent PAH exposures. Biomonitoring studies conducted in the U.S. show that nearly 100% of the general population have detectable levels of PAHs in urine (9,12–15).

PAHs are toxic to immune cells (Figure 2). Data from in vivo and in vitro studies indicate that PAHs cause immunosuppression by inhibiting differentiation and maturation of innate immune cells and B-cells, which can inhibit T-cell function (11,16–18). For example, benzo[a]pyrene, dimethylben[a]anthracene, and 3-methylcholathrene inhibit he differentiation of blood monocytes into macrophages and dendritic cells (18,19). Studies that exposed cell cultures to PAHs report inhibition and suppression of dendritic B cells which influences how these immune cells differentiate and mature (11,16–18). Epidemiological studies in coke oven workers report that high PAH exposure altered immunological outcomes including T-cell suppression, decreases in IgG, IgA, and IgM but increases in IgE antibodies, and natural killer cell ratios (20–22). Birth cohort studies also report that in utero PAHs exposure lowered T-cell counts and elevated levels of B-cells and IgE antibodies in cord blood followed by altered immune function in children (23–27).

Figure 2:

PAHs can modulate and suppress immunological cell types and processes involved in hepatitis B immunity through two possible pathways. Pathway A depicts immune modulation from PAH exposure through exposure of susceptible individual to natural HBV. Pathway B depicts immune modulation from vaccination to hepatitis B and subsequent PAH exposure.

Given the growing evidence that environmental contaminants can influence immune function (28–33), we hypothesized that PAH exposure was associated with modulated hepatitis B immunity using data from the National Health and Nutrition Examination Survey (NHANES). This database contained information on 6 urinary PAH metabolites, serology that distinguishes between people who are susceptible to HBV, have had prior HBV infection, and who have been vaccinated against hepatitis B. Our a priori hypothesis was that there would be a negative association between urinary PAH concentrations and serology indicating hepatitis B vaccination among people who reported receiving ≥3 doses of HBV vaccine after adjusting for confounders. As a supplementary analysis, we also examined the association between urinary PAHs and natural hepatitis B infection.

2. Methods

2.1. Study Population

Conducted by the National Center for Health Statistics, NHANES assesses population health and nutritional status in the U.S. It is a stratified, random sample survey designed to be representative of the civilian, non-institutionalized, U.S. population (50 states and District of Columbia). Our study combined six cycles of NHANES (2003–2004, 2005–2006, 2007–2008, 2009–2010, 2011–2012, and 2013–2014). While all participants were administered serological testing for hepatitis B, a one-third sub-sample of participants ≥6 years of age provided spot urine samples that were analyzed for PAH metabolites. Participants were identified with hepatitis B serological markers to classify participants as either susceptible to infection, immune due to prior natural HBV infection, or immune due to hepatitis B vaccination (see Table S2). We excluded participants who did not have urinary PAH metabolites (n=35,162), participants positive for HIV (n=93) due to potential immunosuppression, participants with acute HBV infection (n= 51), and participants with unclear hepatitis B antibody results (n=148). This resulted in a total sample size of 15,318 participants who had both HBV serology and urinary PAH data. We created a subsample to test the hypothesis that PAH exposure was associated with decreased vaccine antibody response (See Figure 1). This analytical sample included 6,888 participants who self-reported receiving ≥ 3 doses of hepatitis B vaccination (original question responses included received ≥ 3 known doses, < 3 known doses, no doses, don’t know, and missing) and did not have a history of a previous or current hepatitis B infection. Serological markers were used to classify individuals into those who had mounted an immunological response to HBV vaccination (n=3,045) and those who remained susceptible to hepatitis B (n= 3,082). Additionally, we tested the hypothesis that PAH exposure was positively associated with history of natural HBV infection in a subsample that excluded those who received hepatitis B vaccination. This analysis is summarized in Supplement Material.

Figure 1:

Study population description for analysis.

2.2. Urinary Polycyclic Aromatic Hydrocarbon (PAH) Metabolites

The methods used to quantify urinary PAH metabolites have been described previously (34,35). Briefly, spot urine samples were analyzed for 7 metabolites by gas-chromatography/isotope dilution high-resolution mass spectrometry. Quality control procedures included isotope dilution with internal standards labeled with carbon-13. A total of 6 PAH metabolites were detected in >30% of the population and subsequently included in this analysis (36), specifically: 1- Napthol (CAS 90–15-3), 2-Napthol (CAS 135–19-3), 3-Fluorene (CAS 6344–67-8), 2-Fluorene (CAS 2443–58-5), 1-Phenanthrene (CAS ‎85–01-8), and 1-Pyrene (CAS 5315–79-7). Additionally, we generated a composite measure of total PAH, which was a summation of all 6 PAH metabolites. The limit of detection (LOD) for each metabolite varied between NHANES cycle. Subsequently, the maximal LOD of each metabolite across all 6 NHANES cycles was assigned to standardize LODs (37)(See Appendix A - Table S1). Concentrations below the LOD were replaced with the LOD divided by the square root of 2 for each metabolite. Only 0.05–2% of samples were below the LOD for each metabolite, except for 1-pyrene which had 31% of samples were below the LOD. Intra-individual variations in urine dilution were accounted for by adjusting for urinary creatinine in regression models (38).

2.3. Hepatitis B Serology Assessment

Blood samples from NHANES participants > 2 years of age were tested for the presence of hepatitis B antibodies by the CDC’s Division of Viral Hepatitis. NHANES reports the presence or absence of the following hepatitis B serological markers: hepatitis B surface antibody (anti-HBs), hepatitis B core antibody (anti-HBc), and hepatitis B surface antigen (HBsAG). Anti-HBs was measured using a competitive, solid-phase immunoassay (Ausab, Abbott Laboratories). Anti-HBc was measured using a quantitative enzyme-linked immunosorbent assay (Vitros, anti-HBc ELISA). Lastly, hepatitis B surface antigen (HBsAG) was measured only if an individual received a positive test for anti-HBc using additional immunoassays (Auszyme, Abbott Laboratories). Appendix A - Table S2 shows how these antigens were used to classify participants into three groups: serology indicative of no protective antibodies against HBV (i.e. “susceptible”), serology indicative prior natural HBV infection (i.e. “past HBV infection”), and serology indicative of hepatitis B vaccine (i.e. “immune due to vaccine”).

2.4. Covariates

A literature review identified several covariates that were associated with hepatitis B vaccine response (7,8) that were available in the NHANES database including: age (continuous), sex (male or female), and race/ethnicity (non-Hispanic white, non-Hispanic black, Mexican American, other Hispanic, and other race/multi-racial). Socioeconomic status was measured by family income to poverty ratio (FIPR), which is the self-reported family income divided by region-specific poverty level. Country of birth was determined based on self-reported domestic or foreign birthplace (domestic/outside U.S.). Height and weight measurements were used to calculate body mass index (BMI, kg/m²), which was then used to categorize participants as underweight, normal weight, overweight, and obese based on CDC cutoffs for adults or CDC pediatric growth charts (39,40). Serum cotinine was used to categorize participants as active smokers or exposed to environmental smoke (≥ 10 ng/mL) or as non-smokers (<10 ng/mL) (41).

2.5. Statistical Analyses

To account for the complex survey design of NHANES, all analyses incorporated primary sampling units, strata, and weights. Survey weights for the pooled six cycles included in this analysis were based on lab sample weights which were divided by 6 for the six combined cycles (42). Descriptive statistics including means and frequencies were adjusted for survey design. Urinary PAH concentrations were categorized into tertiles based on the distribution of each individual PAH metabolites and the total PAH in the analytic sample. Multivariate logistic regression models were used to evaluate the odds between urinary PAH tertile and serological markers that would indicate immune due to vaccination (i.e., anti-HBs+, anti-HBc-, and HbSAg-) or not immune (i.e., anti-HBs-, anti-HBC-, and HBsAg-). Note, participants who had a past HBV infection (i.e., anti-Hbs+, anti-HbC+, and HBsAg- or a chronic/acute hepatitis B infection (i.e., anti-HBs-, anti-HBc+, and HBsAg+) were excluded from this analysis. Models were run separately for each PAH metabolite (lowest tertile as referent) and were adjusted for age, sex, race/ethnicity, family income to poverty ratio, country of birth, BMI, serum cotinine, and urinary creatinine. We tested linear trends for dose-response relationships between tertiles of exposure by using orthogonal polynomial coefficients. Effect modification by age, smoking status, and country of origin were examined using interaction terms and stratified models. Statistical significance was evaluated with 95% confidence intervals. All analyses were performed in SAS 9.4 (Cary, NC).

3. Results

The characteristics of individuals who self-reported ≥ 3 doses of hepatitis B vaccination and whose serology indicated that they were either susceptible to HBV or had vaccine protective hepatitis B serology are described in Table 1. There was a higher proportion of individuals over > 45 years, higher BMI (obese), and who were smokers among participants who were susceptible to HBV compared to those with vaccine protective hepatitis B serology. We observed significant exposure-response relationships between hepatitis B vaccination serology and 2-napthol, 1-phenathrene, 1-pyrene, and total PAH (Table 2). Specifically, participants with elevated 2-napthol had 0.82 (highest tertile, 95% CI: 0.69–0.98) and 0.70 (middle tertile, 95% CI: 0.54–0.91) times the odds vaccine protective hepatitis B serology than participants in the lowest tertile of exposure. Those participants with the highest total PAH also had lower odds of vaccine protective hepatitis B serology (aOR: 0.73, 95% CI: 0.56–0.95) compared to participants in the lowest tertile.

Table 1:

Characteristics for NHANES participants (ages 6 years and above, n=6,888) from 2003–2014 who self-reported ≥3 doses of hepatitis B vaccination with serology indicating prior hepatitis B vaccination or susceptibility to HBV with weighted proportions and unweighted sample sizes.

Characteristics	HBV Susceptiblea (n=3,082)	Immunity from Hepatitis B Vaccination*b (n=3,045)
	n (%)	n (%)	P-value
Age (years)			<0.01
6 – 19	1775 (41)	1993 (44)
20 – 44	810 (37)	773 (40)
45 +	497 (22)	279 (16)
Sex			0.05
Male	1521 (48)	1385 (44)
Female	1561 (52)	1660 (56)
Race/Ethnicity			<0.01
Mexican American	678 (11)	679 (9)
Other Hispanic	304 (7)	250 (6)
Non-Hispanic White	1052 (62)	1023 (64)
Non-Hispanic Black	777 (13)	804 (13)
Other race/multi-racial	271 (7)	289 (8)
Family income/poverty Ratio			0.06
Below poverty level ≤ 1	904 (20)	813 (18)
Above poverty level > 1	1981 (80)	2077 (82)
Missing	197	155
Body Mass Index (kg/m 2 ) c			<0.01
Normal Weight	1363 (42)	1631 (51)
Overweight	711(25)	694 (26)
Obese	983 (33)	704 (23)
Missing	25	16
Country of Birth			0.87
United States	2657 (89)	2588 (89)
Elsewhere	425 (11)	457 (11)
Serum Cotinine (ng/mL)			<0.01
Non-smoker (<10)	2589 (79)	2640 (83)
Smoker (≥10)	483 (21)	388 (17)
Missing	10	17

Abbreviations - NHANES: National Health and Nutrition Examination Survey, HBV: Hepatitis B Virus

^*P-value run from Chi-square Test Goodness of Fit, significant at α = 0.05, bolded

^*There are 761 missing participants from the stratification of individuals responding to self-report of greater than 3 doses of hepatitis B vaccination

^aParticipants who are susceptible to HBV infection indicated by no relevant serology

^bParticipants with serology indicating hepatitis B vaccination

^cBMI categories: Normal weight (≤ 24.9 kg/m²), Overweight (25–29.9 kg/m² ), and obese (≥30 kg/m²). Those under 20 years of age were given BMI values based on Centers for Disease Control and Prevention (CDC) growth charts (39)

Table 2:

Unadjusted and adjusted odds ratios^* (OR) and 95% confidence intervals (CI) from multivariate logistic regression of the association between urinary PAH metabolite tertiles and serology indicating hepatitis B vaccination for NHANES participants aged 6 years and older who self-reported ≥ 3 doses of hepatitis B vaccination (n=6,888).

PAH Metabolite Tertiles (Range ng/L)	Serology of Hepatitis B Vaccination
	Crude	p-trend	Adjusted	p-trend
1-Napthol
Low (42 – 901)	1		1
Middle (902 – 2,866)	0.88 (0.74–1.05)		0.81 (0.66–0.98)
High (2,867 – 35,920,000)	0.82 (0.69–0.98)	0.20	0.81 (0.65–1.02)	0.05
2-Napthol
Low (64 – 2,456)	1		1
Middle (2,457 – 6,856)	0.81 (0.70–0.94)		0.82 (0.69–0.98)
High (6,457 – 590,676)	0.68 (0.58–0.79)	0.007	0.70 (0.54–0.91)	0.05
3-Fluorene
Low (7 – 56)	1		1
Middle (57 – 159)	0.98 (0.85–1.14)		0.88 (0.74–1.05)
High (160 – 23,855)	0.77 (0.66–0.89)	0.8	0.68 (0.53–0.87)	0.20
2-Fluorene
Low (7 – 148)	1		1
Middle (149 – 395)	0.95 (0.81–1.11)		0.86 (0.70–1.06)
High (396 – 16,312)	0.77 (0.67–0.89)	0.5	0.62 (0.54–0.86)	0.20
1-Phenanthrene
Low (7 – 88)	1		1
Middle (89 – 191)	0.91 (0.77–1.08)		0.84 (0.69–1.02)
High (193 – 10,427)	0.88 (0.76–1.02)	0.3	0.79 (0.65–0.97)	0.10
1-Pyrene
Low (49 – 83)	1		1
Middle (84 – 198)	0.95 (0.80–1.12)		0.93 (0.77–1.11)
High (199 – 12,518)	0.74 (0.63–0.86)	0.52	0.68 (0.55–0.83)	0.50
Total PAH
Low (293 – 4,648)	1		1
Middle (4,649 – 11,746)	0.81 (0.68–0.96)		0.83 (0.68–1.01)
High (11,747 – 35,928,030)	0.70 (0.59–0.83)	0.02	0.73 (0.56–0.95)	0.10

Abbreviations - NHANES: National Health and Nutrition Examination Survey

Referent for hepatitis B categories is being susceptible

^*P-trend for linear dose-response relationships between tertiles of exposure was run from orthogonal polynomial coefficients, significant at α = 0.05, bolded* All models adjusted for participant age (continuous), gender (male or female), race/ethnicity (Mexican American, other Hispanic, non-Hispanic white, non-Hispanic black, and other race/multi-racial), Family income/poverty ratio (below vs above poverty level), body mass index (kg/m2 for normal weight, overweight, and obese), country of birth (United states or elsewhere), smoking status (serum cotinine level either non-smoker (<10 ng/mL) or smoker (>10 ng/mL), and natural log creatinine as a continuous variable. All covariates were significant in models except serum cotinine.

We observed effect modification between race/ethnicity, BMI, age, and serum cotinine levels for hepatitis B vaccination and select urinary PAH metabolites. For non-Hispanic white individuals, only those within the middle tertile of 1-napthol had a lower odds of vaccine protective hepatitis B serology (aOR: 0.70 95% CI: 0.55 – 0.89). Whereas, only individuals with normal BMI and elevated 2-fluorene had a lower odds of vaccine protective hepatitis B serology (middle tertile aOR: 0.65, 95% CI: 0.48–0.87, highest tertile, aOR: 0.55, 95% CI: 0.38–0.78). Individuals between the ages of 6–19 with the highest total PAH level had 0.70 lower odds of vaccine protective hepatitis B serology (95% CI: 0.50–0.98). Similar effect modification by age were seen for 3-fluorene (aOR: 0.63, 95% CI: 0.45–0.88), 2-flurorene (aOR:0.60, 95% CI: 0.42–0.86), 1-phenanthrene (aOR: 0.73, 95% CI: 0.54–0.99), and 1-pyrene (aOR: 0.63, 95% CI: 0.47–0.84) Lastly, people with serum cotinine levels that indicated a non-smoker and elevated 3-fluorene had 0.73 times the odds of vaccine protective hepatitis B serology (highest tertile 95% CI: 0.57 – 0.95). Individuals with serum cotinine levels that indicated active smoking and elevated 3-fluorene had 0.28 and 0.35 times the odds of vaccine protective hepatitis B serology (middle tertile 95% CI: 0.10–0.80, highest tertile 95% CI: 0.13 – 0.96). Sensitivity analyses that restricted the sample to U.S. born yielded similar results.

4. Discussion

This cross-sectional study within the general U.S. population provides preliminary evidence that PAH exposures may modulate hepatitis B vaccine response. This conclusion is based on the consistent observation that participants with elevated urinary PAH concentrations had lower odds of vaccine protective hepatitis B serology, despite self-reporting that they had received at least 3 doses of the vaccine which is usually sufficient to stimulate humoral immunity (7).

While the mechanism by which PAH exposure may modify hepatitis B immunity is currently unknown, previous studies indicate it is biologically plausible (Figure 2). PAHs could modulate the innate immune system’s response to engulf viral hepatitis B antigens to present to T cells and B cells antibody production (3,5,43,44). Experimentally, in vitro studies report that other environmental exposures such as aflatoxin decreases HBV replication (45). Epidemiological studies demonstrate that dietary aflatoxin exposure interacts with HBV to dramatically increase the risk of chronic liver disease (46).

Furthermore, other chemical exposures have also been associated with reduced response to vaccination in epidemiological studies (28,32,47,48). Prospective studies in a European birth cohort report that childhood vaccination to tetanus and diphtheria toxoid were impaired by polychlorinated biphenyls (PCBs) and perfluorinated compounds (PFCs) (28,47,49). Due to the constraints of the NHANES data, we were unable to evaluate the effects of these persistent organic pollutants in conjunction with PAHs on hepatitis B vaccine response. Other cross-sectional studies indicate associations between arsenic exposure and hepatitis A and B vaccination presentation (50,51). Cross-sectional and cohort observational studies reported exposures to PAHs lead to T-cell suppression and modulated antibody production (20–22,26,27). Additionally, there is a potential for an interaction between arsenic and PAH exposures, although other researchers have shown that PAH might be the driving toxicant in immune dysfunction (52). It is possible that PAH metabolism could hinder the ability of the liver to clear HBV as they similarly target the same organ, or possibly diminish the liver’s ability to repair damage caused by hepatitis B. However, the biological mechanism that underpins these observed associations is currently unknown.

The health effects of chronic low-level exposure to PAHs is an active area of research. This is a large class of chemicals that are commonly found in our outdoor and indoor air, food, secondhand smoke exposure, and some consumer products. For example, parent compounds pyrene, fluorene, and phenanthrene which are metabolized into 1-pyrene, 2-fluorene, 3-fluorene, and 1-phenanthrene, respectively, are found in vehicle exhaust, cigarette smoke, coal and industrial emissions, and smoke from wood burning stoves (9,53). 1 and 2-napthol are metabolites of the parent compound of naphthalene where individuals are exposed to from mothballs, vehicle emissions, cigarette smoke, and open burning (54). Thus, urinary PAH metabolites capture recent PAH exposures from a variety of sources and are a good biomarker of exposure. They are also routinely measured in NHANES and by combining 12 years of surveillance data, we were able to create a large population-based sample that allowed us to examine the association between PAHs and hepatitis B vaccine that was generalizable to the US population. The data collected by NHANES also allowed us to adjust for many confounders and provided good statistical power for this analysis. The NHANES laboratory procedures included several quality controls that produce high quality data. Additionally, we were able to look at effect modification by smoking status using serum cotinine which is more reliable than self-reported smoking status.

There are several limitations to our study. Vaccination status was self-reported and could be subject to recall bias. We were not able to account for the physiological phenomenon by which 5% of the general population do not mount a response to hepatitis B vaccination even when administered multiple booster immunizations (55–57). This phenomenon could lead to false negatives for participants who claimed getting the vaccine but could not produce a serology response, however this would likely yield non-differential bias (58). Additionally, this cross-sectional study design prevents us from determining the temporality between PAH exposure and hepatitis B status. Spot urine samples cannot capture chronic exposure to PAHs and exposure misclassification is likely. We examined each PAH metabolite separately, as well as, a summed total to assess the potential toxicity of each compound. However, this approach produces multiple comparisons that can increase the probability of false positives. Yet, the consistency of the direction and magnitude of the effect size across each metabolite is noteworthy. Lastly, we had an insufficient sample size to assess association between PAHs and current infection (n=51), thus limiting our ability to assess the influence of PAHs on chronic HBV infection. The time that passes between vaccination and serological testing, as well as, duration of PAH exposure likely play a role in the biological response. Unfortunately, this information is not available in the NHANES cross-sectional study design. We saw a suggestion of effect modification by age where younger adults with the highest levels of urinary PAHs had the lowest odds of serology indicating a vaccine response. It is challenging to untangle the effect of age in this study because age may be linked to the duration of time between receiving the vaccine and enrollment in NHANES which could create a structural bias in this data. Additionally, age is linked to PAH exposures through multiple pathways. For instance, older participants tended to have higher urinary PAH levels and report smoking. Even though our analysis controlled for these covariates, cross-sectional analyses are not ideal for examining the effects of chronic exposures. Thus, future studies would benefit from being able to examine the effect of age on hepatitis B vaccine response in a longitudinal study design that includes information about the duration of time between the vaccine delivery, PAH exposures, and serological assessment.

5. Conclusions

In a cross-sectional analysis of the U.S. population, we observed a dose-response association between higher urinary levels of PAHs and lower odds of serology indicating hepatitis B vaccination among the general US population who reported that they had been vaccinated. Additional studies with prospective study designs are needed to further evaluate the differential effects of specific PAH compounds on the immune system, and whether PAH exposure influences susceptibility to hepatitis B.