Water Health Assessment<BR>(Dover Township) Toms River, New Jersey

Case-Control Study of Childhood Cancers in Dover Township, December 2001
(Dover Township) Toms River, New Jersey

Volume 1: Summary of the Final Technical Report

Draft for Public Comment
Case-control Study of Childhood Cancers
in Dover Township (Ocean County), New Jersey
Volume 1: Summary of the Final Technical Report
Public Comment Period
December 19, 2001 through February 19, 2002
Division of Epidemiology, Environmental and Occupational Health
New Jersey Department of Health and Senior Services
in cooperation with
Agency for Toxic Substances and Disease Registry
U.S. Department of Health and Human Services

Introduction

In 1995, at the request of the federal Agency for Toxic Substances and Disease Registry (ATSDR), the New Jersey Department of Health and Senior Services (NJDHSS) evaluated the occurrence of childhood cancer in Dover Township and found that the incidence was significantly higher than expected for the period 1979 through 1991, Consequently, the NJDHSS and ATSDR, in cooperation with the local Citizen Action Committee for Childhood Cancer Cluster, developed a Public Health Response Plan (PHRP) detailing a systematic process to Investigate the elevation of childhood cancer in Dover Township. The PHRP's purpose was twofold: 1) to update and re-evaluate information on childhood cancer incidence in Dover Township; and 2) to evaluate possible community exposures to toxic chemicals in the environment, called exposure pathways, in order to generate hypotheses which could be assessed in an epidemiologic study.

The results of the cancer update, Childhood Cancer Incidence Health Con5ultation: A Review and Analysis of Cancer Registry Data, 1979-1995 for Dover Township (Ocean County), New Jersey (1997), confirmed that the overall childhood cancer incidence rate in Dover Township was statistically significantly elevated for the period 1979 through 1995, primarily due to excesses of leukemia (over 9 times higher than expected) and brain and central nervous system cancer (11.5 times higher than expected) in females residing in the Toms River section of the Township.

Findings of the PHRP evaluation of potential exposure pathways in the community indicated that past releases of toxic chemicals into the environment had resulted in exposure to residents in Dover Township. Consequently, NJDHSS and ATSDR concluded that the Reich Farm Superfund site was a public health hazard due to past human exposure to toxic chemicals due to groundwater contamination which affected both private wells and wells in the public water distribution system. Early in the course of evaluating these exposure pathways, testing of the community water supply revealed a previously undiscovered contaminant, styrene-acrylonitrile (SAN) trimer, attributable to the Reich Farm Site, in groundwater from two of the wells in the United Water Toms River (UWTR) Parkway well field. NJDHSS and ATSDR also concluded that the Ciba-Geigy Corporation Superfund site represented a public health hazard due to past human exposures to toxic chemicals from operations at the facility, through past groundwater contamination affecting public water wells and private wells used for irrigation. Past air pollution emissions from the facility were also of public health concern.

In 1997, based on these findings, NJDHSS and ATSDR decided to conduct an epidemiologic study to evaluate the relationship between the completed environmental exposure pathways and the elevated childhood cancer incidence in this community. The primary hypotheses in this study are that childhood cancers were associated with environmental exposure pathways which were identified in earlier reports:

*contamination of potable water from UWTR Parkway well field by the Reich Farm Superfund site;
*contamination of potable water from UWTR Holly Street well field, due to wastewater discharge into the Toms River from the Ciba-Geigy site;
*contamination of potable water affecting private wells near the Reich Farm and Ciba-Geigy sites and other areas in Dover Township from other or unknown sources; and
*air pollution emissions front the Ciba-Geigy facility.

Potential exposure to other environmental factors are also considered in order to address additional community concerns. These environmental factors included:

*potential exposure to ambient air releases from the Oyster Creek Nuclear (Generating Station, located approximately ten miles South of Dover Township; and
*residential proximity to sites in the area which concerned the community:
*the Ciba-Geigy Superfund site;
*a section of the Toms River where treated wastewater was discharged (1952 to 1966) from the Ciba-Geigy facility;
*the Ciba-Geigy pipeline which transported treated wastewater across Dover Township (1966 to 1991) and experienced three documented breaks (1984, 1988, and 1989) during its usage;
*the Reich Farm Superfund site;
*the Dover Township Municipal Landfill;
*the Ocean County Landfill; and
*the Toms River Coal Gas site.

The epidemiologic study uses a case-control design to evaluate possible risk factors and the magnitude of their association with childhood cancers in Dover Township. The risk factors evaluated include the environmental exposures comprising the primary hypotheses arid other factors which have been evaluated in other studies of childhood cancer. A case-control study design was selected because it is the best epidemiological method for studying rare diseases.

The study has two separate components: the Interview Study and the Birth Records Study. In the Interview Study, residential histories were collected by telephone interview using a structured questionnaire. In the Birth Records Study, residence of the mother at the time of the child's birth was obtained from the birth certificate. The primary data used to address the environmental hypotheses were drawn from residential histories (Interview Study) or from birth certificates (Birth Records Study), combined with modeled water distribution or air Pollution exposure estimates. Information on other factors was collected through the questionnaire (Interview Study) or the birth certificate (Birth Records Study) because these potential risk factors have been previously studied in other investigations of childhood cancer or are potential confounders of the environmental factors.

The relative risk (odds ratio) of childhood cancers was computed using conditional logistic regression to evaluate the degree to which exposure factors were associated with disease. The odds ratio (OR) is the ratio of the risk of disease among the exposed compared to the risk among the unexposed. An OR greater than one means that the exposure factor was more common in cases than controls. Conversely, an OR less than one means that the exposure factor was less common in cases than controls. The 95% confidence interval was calculated to assess statistical association between exposure and disease. The width of the confidence interval is a measure of the certainty of the estimated OR. Because of the small number of subjects in this study, a relatively large odds ratio is needed before it reaches statistical significance; this is especially true for certain small strata analyses.

Odds ratios were computed for two age groups (children diagnosed prior to age 20, and children diagnosed prior to age five). For the Interview Study, ORs were computed for each of four diagnostic cancer groupings: leukemia and nervous system cancers combined; leukemia: all nervous system Cancers; and brain and central nervous system cancers. For the Birth Records Study, Ors were computed for each of five cancer groupings: all cancers combined; leukemia; all nervous system cancers; brain and central nervous system cancers; and all other cancers.

The potential association of a risk factor with an outcome was evaluated using a combination of criteria, including: strength of the association, statistical significance, consistency of findings of multiple measures for an exposure, apparent dose response effect, and evidence of a completed exposure pathway.

Interview Study

The Interview Study focused solely on the types of cancers that were found to be elevated in Dover Township. A case was defined as a child who was diagnosed between 1979 and 1996 with leukemia or nervous system cancer before 20 years of age while residing in Dover Township. Cases were identified from New Jersey State Cancer Registry records. Four controls were matched to each case by age, sex, and residence in Dover Township at the time the case was diagnosed. Controls were identified from Toms River School District student rosters. A structured questionnaire was developed and administered to the parent or guardian over the telephone by experienced NJDHSS interviewers. Information was collected for the time period comprising one year prior to birth to the month and year of the case's diagnosis. Consequently, a few parents were required to recall information from as early as 1962.

Interviews were conducted for a total of 40 children who met the case definition: 22 with leukemia and 18 with nervous system cancers (5 of which were sympathetic nervous system cancers and 13 of which were brain and central nervous system cancers). There were no interviews for 39 potentially eligible controls, primarily due to parent refusals or lack of response after repeated contact attempts. Interviews were conducted for a total of 159 controls. The overall study participation rate was 84% (100% of cases and 80% of controls).

Using information from birth certificates, the participating and nonparticipating controls were found to be generally similar for average weight at birth, race, and maternal age at the time of child's birth. However, nonparticipating eligible controls were found to have been born earlier in the study time period, were less likely to have a mother residing in Dover Township at the time of their births, and were less likely to be the mother's first born child than the participating controls.

The excellent response rates and cooperation of study families were strengths of the Interview Study. Limitations of the Interview Study included the study's low statistical power to detect significant results and potential recall bias.

Birth Records Study

In order to address community concerns that children may have moved away from Dover Township prior to being diagnosed with cancer, the Birth Records Study was conducted and evaluated all types of childhood cancer. A case was defined as a child who was diagnosed with any type of cancer before 20 years of age and whose mother was a resident of Dover Township at the time of the child's birth. All births from 1966 through 1996 among residents of Dover Township, identified from New Vital Statistics records, were compared with the New Jersey State Cancer Registry records, 1979 through 1996, and nine other state cancer registries. Although not all states have cancer registries, these nine states were included because they have cancer registries and were the destination of 70% of Ocean County out-migrants between 1985 and 1990, according to U.S. Census data. Ten controls, who matched each case by year of birth and sex, were randomly selected from all births to Dover residents. The cases were then each compared to their ten matched controls with respect to variables contained on the birth certificate.

A total of 528 children were enrolled in the Birth Records Study (48 cases and 480 controls). There were a total of 16 children with leukemia's, 13 children with nervous system cancers, and 19 children with some other type of cancer. Of the 48 cases, 41 were residents of Dover Township when diagnosed with cancer, five of the cases resided in another Ocean County Community, one resided in another New Jersey county, and one resided in another state at the time of their diagnosis. Of the 41 cases who were residents of Dover Township both at birth and diagnosis, 24 were diagnosed with leukemia or nervous system cancer, and thus were also included in the Interview Study.

Since birth certificates were used as the source of the information collected for study children, a strength of the Birth Records Study is that this type of data is free of exposure misclassification due to recall error. Another strength is that case and control selection was likely unbiased. Limitations of the Birth Records Study include; lack of available information on the child's postnatal exposures, limited amount of prenatal information, the birth address may not be the mother's residence throughout the pregnancy, and low statistical power.

Environmental Factors

Household Drinking Water Source Exposure

Three major study hypotheses were evaluated regarding association with childhood cancer incidence: exposure to the United Water Toms River (UWTR) public water distribution system's Parkway well field water; exposure to the UWTR's Holly Street well field water; and exposure Lo private well water in areas of Dover Township with known groundwater contamination. Since historical contamination of the Holly Street and Parkway well fields has been documented, human exposure to the contaminated water became one of the critical risk factors to be examined. However, because contamination was not continuously present at either the Parkway or Holly Street well fields throughout the entire study time period, the study also evaluated potential contaminant time windows regarding the possible period of contamination of each well field. Exposure indices developed without the use of contaminant time windows will be referred to as unadjusted while those indices developed with contaminant time windows will be referred to time-specific in later discussions.

For the Holly Street well field, pollution was documented for the mid-1960's, but the duration of contamination was unknown. The source of the Holly Street well field contamination was believed to be Ciba-Geigy wastewater discharged into the Toms River from 1952 to 1966, By the mid-1 970s very little water from contaminated wells of the Holly Street well field was being pumped into the water distribution system. Consequently, for the purpose of the study the period of 1962 through 1975 was assumed to comprise the time frame of potential exposure to contaminated Holly Street well field water.

The Parkway well field became contaminated at some point after the illegal disposal of thousands of drums of chemical waste at the Reich Farm site, which occurred in 1971. Although contamination of the Parkway well field was first documented in 1986, it is unknown when the contamination first affected the well field. It has been estimated that groundwater contamination from the Reich Farm site may have taken approximately ten to fifteen years to reach the Parkway well field. Therefore, for the purposes of this study, the period of 1982 through 1996(the end of the Study period) was assumed to comprise the time frame of potential exposure to contaminated Parkway well field water. Other alternate time frames, assuming a contaminant transit time of as little as six years (1978) to as long as 14 years (1986), were also evaluated.

The chemical composition of the groundwater contamination which resulted from the dumping of chemical wastes at Reich Farm had not been well characterized previous to the current investigation. A variety of chemicals including two carcinogens, trichoroethylene and tetrachloroethylene, and a previously unknown material, styene-acrylonitrile (SAN) trimer, have been found in the plume. However, since the concentration of the contaminants impacting the Parkway well field over the years is unknown, the risk potential to the community is not well understood. Consequently, the true carcinogenic and non-carcinogcnic potential of the contamination mixture in this completed exposure pathway remains unknown.

Exposure to public water from the eight additional UVVTR well fields (Anchorage, Berkeley, Brookside, Indian Head, Route 70,Silver Bay, South Toms River, and Windsor) was also evaluated. To evaluate potential exposure to contamination from private well use, I 1 areas in Dover Township with a history of groundwater contamination were delineated (for more detail see Volume IV, Appendix D).

Computer-based reconstruction modeling conducted by ATSDR was used to derive monthly estimates of the percent of water delivered to each study residence from each well field in the public water distribution system from 1962 through 1996(for more detail see Volume IV, Appendix D). From these modeled data, the average percent of public water delivered (called the exposure index) from each of the well fields was calculated for each study subject using residential information. Three exposure categories were formed: low (receiving less than 10% of their water from the specified public well field); medium (receiving 10% to 49.9% of their water from the specified public well field); and high (receiving 50% or more of their water from the specified public well field).

In the Interview Study, information was also collected on the average daily amount of tap water consumed by the mother during pregnancy and by the child from birth to the date of diagnosis. Additional exposure indices were created using the tap water consumption information combined with the average public water source percentages, called the water source/consumption indices.

Although computer modeling of the public water distribution system was the most objective method available to estimate past exposure to well field water (the vehicle for the actual distribution of environmental contaminants), there are limitations in the exposure assessment. In developing the historic model reconstructions, a number of assumptions were made on how the system operated overtime. Altering system operations, such as when wells were pumping, might affect the estimated percentage of water delivered to a particular location from any well field. In order to evaluate the uncertainty and variability in system operations, ATSDR used multiple system assumptions to recalculate water percentage estimates. This sensitivity analysis found that estimates were very similar regardless of which system operations were employed. Consequently, the use of computer modeling for the water exposure assessment was a strength in the Interview and Birth Records Studies.

Another strength of the Interview Study was its ability to evaluate potential exposure during specific critical time periods: the total study time period (one year prior to birth to the date of diagnosis); the prenatal time period; and the postnatal time period. For the Birth Records Study, only the residence at birth was available to assess exposure during the prenatal time period. Additionally, in the Birth Records Study the source of water used in the home (private well or public water) was unknown and was assumed to be public water if the water distribution system was on the street of the residence during the year of the child's birth.

......Please click on next to read the rest of this report......

BACK || CONTENTS || NEXT