Perry Sheffield, MD

In the last issue, the reviewed environmental health articles touched on mercury in dental amalgams and organophosphates and DDT from pesticides and their potential role as neurodevelopmental toxins. This review looks at two articles that consider the effect of environmental chemicals on lung function and immune status.

Heilmann C, Grandjean P, Weihe P, Nielsen F, Budtz-Jørgensen E. 2006. Reduced antibody responses to vaccinations in children exposed to polychlorinated biphenyls. PLoS Med 3(8):e311. http://medicine.plosjournals.org/perlserv/?request=get-document&doi=10.1371%2Fjournal.pmed.0030311

Background
While variation of antibody response to vaccinations is well known, the reasons for this variation in otherwise healthy children are not well understood. Some proposed mechanisms include immunotoxic exposure to certain pollutants like polychlorinated biphenyls (PCB). PCBs are organochlorine compounds that, despite a ban several decades ago, persist in the environment and exposures are possible through dietary sources, such as fish, and manufactured sources, such as electrical appliances made before 1977. In limited studies, PCBs have been associated with decreased total immunoglobulins and increased childhood infections.

To explore this potential toxicity further, Heilmann et al designed this prospective, cohort study in the Faroe Islands, an autonomous region of Denmark, located in the North Atlantic, and home to the recent International Conference on Fetal Programming and Developmental Toxicity in May 2007. Some inhabitants of these islands have a 10-fold increase above other Northern European persons in PCB levels due to consumption of pilot whale blubber. Importantly, this population has average Northern European levels of another potential confounding chemical called dioxin. Dioxins are by-products of burned electrical cables and insulation, plastics and household waste. Once formed they, like PCB, persist in the environment for a long period and can enter the food chain. Contaminated fish, meat, dairy products are the most common source of human exposure. This population was chosen to help distinguish PCB and dioxin effects on the immune system. This study examined vaccination response to tetanus and diphtheria vaccines, thymus-dependent neoantigens, as markers of overall immune system efficacy in a population of children exposed in utero and postnatally to elevated PCB levels.

Methods
This prospective, cohort study consisted of two groups of healthy maternal/child pairs. The first cohort of 182 pairs was recruited in 1994-95 with complete data available for 124 pairs. The second cohort included 116 pairs recruited in 1999 – 2001. Maternal serum was collected during the last prenatal visit at 34 weeks gestational age and breast milk was collected after delivery. Both specimen types were analyzed for PCB concentration. Both groups received 3 doses of tetanus and diphtheria vaccinations. None of the vaccines contained mercury-based preservatives. Children in the two cohorts had tetanus and diphtheria toxoid antibody as well as serum PCB levels measured. The first cohort was followed until 7.5 years of age and the second cohort was followed until 18 months of age.

The outcomes measured were the serum specific antibody concentrations measured using enzyme-linked immunosorbent assay.

Data Analysis
As there are numerous different congeners (chemical derivatives) of PCB, a simplified total PCB concentration was calculated. In addition, several of the PCB congeners that have toxic properties similar to dioxins were weighted using toxicity equivalency factors, a standardized concept created to facilitate risk assessment and regulatory control. Standard regression techniques and log transformations of antibody concentrations were used. Models included sex, age, birth weight, maternal smoking during pregnancy, and time from last vaccination. Prenatal and postnatal exposures were analyzed separately and then models that allowed for both exposure variables were used. PCB exposure parameters were log transformed before entry into the model. Log transformation allowed expression of change in percent of the antibody levels per doubling of PCB exposure levels. In addition to the measured PCB exposure, analysis included an exposure variable based on report of maternal whale blubber intake. This inclusion allowed for measurement error and pooling of information from different exposure markers. Prenatal and postnatal exposures were examined separately and then together in one model. The benchmark dose of PCB exposure was determined based on the lower 95% confidence limit of the level that increased the risk of an abnormal antibody response, which in unexposed populations is expected to be 5% – 10%.

Results
The 18-month-old cohort showed a decrease in diphtheria toxoid antibody concentration of 24.4% (95% CI, 1.63% - 41.9%) with a doubling of PCB prenatal exposure. Antibody concentrations were significantly affected by both prenatal and postnatal exposure. The 7.5-year-old group showed a negative correlation between PCB exposure and tetanus antibody concentration. Specifically, there was a decrease of 16.5% of the tetanus antibody concentration for each doubling of the prenatal exposure (95% CI 1.51%-29.3%). While most children maintained sufficient antibody levels to confer protected status, two years following the booster vaccine for the older cohort, 21% (95% CI, 14%-28%) of the children had diphtheria toxoid antibody concentration below the limit for long-term protection. The benchmark dose levels, calculated from maternal serum PCB concentration, for effect on diphtheria toxoid antibody concentration in the younger cohort was 1.14 micrograms/gram (ug/g) lipid (similar to the one based on PCB-related neurodevelopmental deficits). The level for the effect on tetanus toxoid antibody concentration in the older cohort was 2.18 ug/g lipid.

Strengths
Children were from population-based birth cohorts and were in good health, improving the generalizability of the results. This study was a prospective study that used models able to account for measurement error. The population of the Faroe Islands has a higher than average PCB exposure which potentially allows for unmasking of otherwise subclinical toxicity. In addition, serum analyses were intercalibrated between the study laboratories and in sessions organized by the German Society of Occupational Medicine. Examination of the younger and older cohorts together pointed to PCB burden in the prenatal and early postnatal periods as the major determinant of immunotoxic effects. In other words, prenatal and early life PCB exposure seemed to have a larger negative impact on antibody levels than child PCB levels later in life. Such increased vulnerability in early life is not uncommon and points to need of further regulation to protect the youngest children during this period of vulnerability.

Limitations
The width of the confidence intervals for the antibody effects suggests that the results are not very precise. Specifically, differences in PCB exposure effects on tetanus and diphtheria should be interpreted with caution. In addition, the PCB measurements were widely spaced in time between maternal serum/breast milk levels and then follow-up levels in the children. This limits the postnatal exposure assessment and therefore does not allow for delineation of a more specific critical vulnerability window of development.

The study could not identify specific causative PCB congeners. Though the more persistent congeners constitute the majority of the PCB in the samples analyzed, immunotoxic effects could actually have been mediated by other congeners that were no longer present due to their shorter half-life. The persistent congeners could simply have been markers of these other congeners which caused the effects.

Other chemicals such as pesticide metabolite, p,p’-dichlorodiphenyldichloroethylene (p,p’-DDE), and mercury were measured. However, close correlation of PCB and p,p’-DDE levels and potential presence of other chemicals that were not measured did not permit the study to control for other potentially immunotoxic agents. There is also potential for effects from mixed exposures which could not be clearly examined in this study design.

Lastly, the unique diet (pilot whale) of the Faroese that causes their elevated PCB levels makes them an interesting study population but limits the generalizability of the results of this study.

Conclusion
This study provides epidemiological evidence of an association of prenatal and postnatal PCB exposure with decreased antibody response. The authors propose two potential mechanisms of PCB burden in early postnatal period as a major determinant of immunotoxic effects: first, thymus vulnerability both prenatally and in early postnatal life and, second, poor priming of first vaccine before 6 months of age could affect the magnitude of antibody production with subsequent booster vaccines. Further studies with more closely spaced PCB measurements might be able to delineate a more specific window of developmental vulnerability and better differentiate effects of PCB from other contaminants.

Using the benchmark dose levels and a default 10 fold uncertainty factor, the recommended exposure limit would be as low as 0.1 ug/g. In addition, vulnerable groups such as preterm infants or those with chronic infections or other comorbid conditions may be at even increased risk. Although PCB exposure levels have decreased in general, these results suggest that even more efforts may be needed to protect against immune effects.

Irrespective of the lack of strong causation links, this study strengthens the impetus for clinicians to help guide parents (especially breast-feeding mothers) about dietary choices. Clinicians should advise persons, particularly pregnant women about the potential immune effects of PCB. Counseling should include avoidance of dietary sources of PCB such as sport fish and other fish based on regional risk (information available from state health department advisories).

Gauderman, W., Vora, H., McConnell, R., Berhane, K., Gilliland, F., Thomas, D., et al. Effect of exposure to traffic on lung development from 10 to 18 years of age: A cohort study. The Lancet, 369(9561), 571-577. http://www.sciencedirect.com/science?_ob=MImg&_imagekey=B6T1B-4MX4VW6-2-7&_cdi=4886&_user=30742&_orig=search&_coverDate=02%2F23%2F2007&_sk=996300438&view=c&wchp=dGLbVzz-zSkzS&md5=9a9a54db1a723c59ed1dd2453fbe5672&ie=/sdarticle.pdf

Background
Prior air pollution studies show negative effects of urban and regional air pollution on lung function. There is also evidence that local traffic is related to an increased incidence of asthma and other lung diseases in children. This study by Gauderman et al aimed to address the lack of specific evidence regarding the relationship of lung function development and traffic exposure in childhood.

Methods
This prospective, cohort study used data from The Children’s Health Study conducted in 12 southern California communities. Two cohorts of 4th grade students were studied and followed for 8 years. Baseline and follow-up questionnaires to the parents included information about race, ethnicity, income, parental education, doctor diagnosed asthma, and exposure to indoor air pollutants from gas stoves, pets and smoking. Lung function assessments of the children (measuring forced vital capacity (FVC), forced expiratory volume in 1 second (FEV1), and maximum midexpiratory flow rate (MMEF)) were performed yearly by trained technicians at study schools. The exposure data was measured two ways: 1) proximity to a major road and 2) a dispersion model-based estimate of traffic-related air pollution at the child’s residence. The cohorts were categorized into quartiles based on the exposure data. Regional air pollution was continuously monitored within each study community at a central location.

Data Analysis
The statistical techniques used included modeling with three categories of socioeconomic status. The model allowed for separate lung growth curves for each sex, race, ethnic group, cohort and baseline asthma subgroup. Analysis adjusted for height, body-mass index, present asthma status, exercise or respiratory illness on the day of the test, any tobacco smoking by the child in the previous year, and indicator variables for the field technicians.

Four categories were used to describe distance to the freeway (less than 500 meters, 500-1000 meters, 1000-1500 meters and greater than 1500 meters). Similarly distances to nonfreeway roads were categorized based on distances of 75 meters, 150 meters, and 300 meters. Model based estimates of pollution from freeways and nonfreeways were categorized into quartiles based on their respective distributions. Interaction terms in the model allowed for joint estimation of local traffic effects and community long-term average pollutant concentrations. In all cases, negative estimates signified reduced lung function growth with increased exposure (compared to the least exposed category).

The outcomes measured were the annual pulmonary function tests (FVC, FEV1, and MMEF). Predicted FEV1, FVC and MMEF were calculated from models including the observed values and other predictors. Percent predicted values (PPVs) were calculated as observed divided by predicted. Regression models were then used to calculate a mean percent-predicted value for each category of distance to the freeway with adjustment for community. Percent predicted values were scaled so that children living furthest from the freeway had a mean of 100% predicted and others were given means relative to this benchmark.

Results
82% of available students agreed to participate. There were mostly white, non-Hispanic and Hispanic children and equal proportions of male and female participants. 12% of the children lived within 500 meters of a freeway.

Overall, from the two cohorts, 1445 children were observed over the full 8 years. Closer residential distances to the freeway were associated with decreased growth in lung functions. FEV1 of the group living within 500 meters of freeway was 81 ml less (95% CI 18 to 143 ml less; p=0.012) than the group that lived greater than 1500 meters away. In this group living within 500 meters of the freeway, FVC was 63 ml less (95% CI 5 ml greater to 131 ml less) and MMEF was 127 ml/sec less (95% CI 11 to 243 ml/sec less) than the greater than 1500 meters group. Model-based pollution exposure showed deficits in lung function growth but no statistical significance. Non-freeway roads were not associated with deficits. Although low socioeconomic status was associated with increased traffic exposure, adjustment for status induced only a modest change in results (as did adjustment for other indoor air pollutants). Boys were more affected than girls but the “test of effect modification by sex was non-significant (p=0.10).” Only 6 of the 12 communities had substantial numbers of children living within 500 meters of a freeway. The estimated effects of freeway distance on lung development were more pronounced in these six higher traffic communities. Notably, significant lung effects were also seen in children without asthma or history of tobacco use.

Reduced lung function growth was independently associated with both freeway distance and with regional air pollution. Percent-predicted value (PPV) of lung function at 18 years of age showed pronounced deficits. For the group living less than 500 meters from a freeway, PPV FEV1 was 97% (95% CI 94.6 – 99.4; p=0.013 compared to greater than 1500 meters from a freeway). PPV MMEF was 93.4% (89.1-97.7; p=0.006 compared to greater than 1500 meters).

Strengths
This study was a long-term prospective follow-up of two large cohorts with exposure and outcome data consistently obtained. It built on a 2004 study by the same group published in New England Journal of Medicine that assessed the relationship between air pollution measured at central locations in each of 12 communities to lung development. The same equipment and testing protocols were used throughout the study period.

Limitations
The study had an 11% per year attrition rate. Participant attrition is a potential source of bias in cohort studies (although consistent results in the groups followed for the full eight years is reassuring that the results are valid). While the study controlled for socio-economic status and some indicators of indoor air pollution including exposure to environmental tobacco smoke, other confounders are possible for traffic, home and school contributors, and lung function growth. No assessment of the distance of children’s schools from freeways was made. This study was also not able to identify which specific traffic pollutants were responsible for the lung effects, whether there was a mixed pollutant exposure effect, or if some characteristics of traffic beyond just pollutant exposure, such as noise, was associated with the physiologic impact. There was also not a significant association between model-based pollution from a freeway and lung function growth despite large estimated deficits in the highest exposure quartiles. Further study is needed to clarify this question.

Conclusion
Reduced lung function growth was found to be independently associated with freeway distance and regional air pollution. This study strengthens existing evidence that polluted air can have long-term negative effects on children with or without concomitant morbidities.

No evidence was found that traffic effects varied depending on background air quality, suggesting that even in an area with low regional pollution, children living near a major freeway are at increased risk of lung effects. In addition, children who live close to a freeway in a high pollution area experience a combination of adverse developmental effects due to local and regional pollution.

The relevance of this study is how it emphasizes that local (such as neighborhood scale) air pollution (not just regional background air quality) affects lung development in otherwise healthy children and will likely increase adult morbidity and mortality. As clinicians, we can advise our patients, both those with and without pre-existing lung disease, to live as far from traffic as possible when they have a choice. We can advocate on behalf of our patients when new road proposals threaten to affect those who are unable or unwilling to move away. We can also continue to encourage clean air legislation and stricter emissions standards to make the effect of existing roads less detrimental to our pediatric population.

Updated 01/04/08