Members of the Childhood Cancer Research Core have conducted, or collaborated on numerous studies that have addressed questions relating to home or parental occupational exposures. The exposures of interest are wide-ranging, but have included pesticides, polynuclear aromatic hydrocarbons (PAHs), metals, radon, electromagnetic fields (EMF), environmental tobacco smoke (ETS), infectious agents, N-nitroso compounds, and solvents. Often, the study has been designed to explore, in a semi-directed fashion hypotheses related to these agents. For example, Dr. Buckley obtained house-dust samples from homes of children with acute lymphoblastic leukemia (ALL) and control homes with a primary aim of determining whether the level of exposure to specific pesticides in this dust was higher for cases than controls. Thus the study was directed specifically to address the question of pesticide exposure in ALL etiology, but it was also hypothesis generating in that there was little prior knowledge that could be used to focus on any specific pesticide or class of compounds.

In some instances, the studies were more focused, and were indeed hypothesis testing. Examples of this would include studies of radon, EMF in leukemia, and N-nitroso exposure in brain tumors.

A significant consideration, for some studies, was the community and public health importance of the putative association. Thus EMF and radon would probably not top the list of agents deserving of highest priority for study, from a purely scientific perspective, but were extremely important from a public health perspective. The possible hazards from EMF exposures, in particular, had the potential to cost the electric power industry and others billions of dollars in legal and ameliorative costs.

The mainstay of these studies continues to be data collection through self-administered questionnaires or (more commonly) interviews of the parents of children with specific cancers and a matched control group. This approach is cost-effective and can obtain, from the one source, information about a wide range of factors, including parental occupational exposure, home exposures of parent and child, medication taken by child, parental smoking habits, maternal exposures during pregnancy and the medical history of the child and his/her family. However, the quality of these data are variable, in part because many exposures (such as pesticides) are inherently difficult to specify and quantify, and in part because of the difficulty in recalling events that occurred years in the past. Some of these inaccuracies can be reduced with careful questionnaire design, but even with the best design we can expect considerable exposure misclassification.

Reliance on parental recall also opens up the possibility of differential recall of case and control parents, leading to a bias in the risk estimates. Fortunately, for the most part, recall bias does not seem to be a widespread problem in our studies, although it is always a concern.

For the above reasons, we have endeavored to use, where practical, alternative approaches that rely less on parental recall and that can provide more objective and quantifiable measures of exposure. A major limitation on this effort has been broad geographic spread of the cases and controls in most of our studies. Because of the rarity of childhood cancer, sizable case series can only be achieved by cooperation on a multi-state or national scale. One way to get objective data is to access written records, where possible, such as medical records that document medical exposure.

Another is to make direct, in-home measurements. This is possible for local studies, but is not generally feasible on a national scale. An exception has been a study that was completed recently of pesticides, PAHs, metals and cotinine in house dust of children with ALL (and controls) in a nine-state region of the U.S. This study was economically feasible only because the NCI was investing substantial resources sending teams into the homes of these subjects to measure EMF levels, and the incremental cost of dust sampling was modest.

Lastly, we recognize the potential of the new molecular techniques to provide insights into the processes of tumorogenesis and/or to quantitate in vivo exposure to environmental insult, and to do so on a cost-effective basis in large epidemiological studies. We are just starting a case-control study of non-Hodgkin's lymphoma in children that will assess pesticide exposure and will measure the frequency of two specific chromosomal rearrangements mediated by the VDJ recombinase enzyme system.

Project A: Risk of ALL Associated with Exposures to Substances in Household Dust: Pesticides, Polynuclear Aromatic Hydrocarbons and Metals

Dust samples have been obtained from the homes of 250 cases of acute lymphoblastic leukemia and 300 controls to determine whether increased risk can be seen for a range of pesticide residues, PAHs and metals in the dust. Young children are at particular risk of exposure to substances in housedust, since they spend much time on the floor, crawling or playing, and commonly place objects in their mouth. Two manuscripts are in development.

Project B: Correlates of Pesticide, PAH and Metal Residues in House Dust

Related to, but separate from the above project is one that asks what sociodemographic and other factors correlate with high levels of pesticides, PAHs and specific metals in the home, irrespective of case-control status. In particular we are interested in correlating the responses of parents to extensive questionnaires on the use of pesticides in the home and at work with the measurements of these substances in samples taken from the home. An extensive analysis of the PAH component of this project has been completed, under contract from the EPA.

Project C: Risk of ALL Associated with Passive Smoke Exposure

Cotinine was also measured in the house dust. This has allowed for an objective estimate of ALL associated with environmental tobacco smoke exposure and for a comparison with self-reported smoking habit. A manuscript on this analysis is in preparation.

Project A: Role of Atypical Antigenic Stimulation or Responsiveness in NHL Etiology

The first NHL case-control study focused primarily on immune responsiveness, looking for evidence of abnormal levels of stimulation or altered patterns of response. These were assessed via surrogates measures, such as evidence for increased allergy or autoimmune disease in the family, or evidence for differences in patterns of infection, particularly in early childhood. There was some indication of an increased rate of infection in the cases in first two years of life, but the evidence was inconsistent and did not appear to support the primary study hypotheses. One association that did come through strongly, however, was with pesticide exposure of the child and parents, at home. Two manuscripts have been submitted from this study.

Project B: Pesticides and VDJ Recombinase-Mediated Mutation Rates in Children with NHL

As a follow-on to Project A we were funded to conduct a larger case-control study, focusing on the possible role of pesticide exposure and using a molecular marker (aberrant VDJ-recombinase activity) that may correlate both with the exposures of interest and with the risk of chromosomal recombinations that lead to NHL. This study is in progress.

Project A: Case-Control Studies of Leukemia - Risks Related to Pesticide Exposure

Dr. Buckley is responsible for analysis of data from two large case-control studies of acute leukemia (Dr. Leslie Robison, P.I.) to determine whether self-reports of use of pesticides in the home and in the parent's workplace are associated with an increased risk.

Project B: Risks of Childhood Cancer Associated with Radon Exposures

These same studies have been used to place radon detectors in the homes of cases and controls to establish whether this exposure is a risk factor.

Project C: Infant Leukemia Case/Control Study

This is a collaborative study, conducted out of the University of Minnesota (Dr. Julie Ross, P.I.) to study the epidemiology of infant leukemia, with emphasis on the role of exposure to topoisomerase inhibitors. In particular topo II inhibitors in the diet will be closely examined.

Project D: Downs Syndrome and Downs-leukemia

This is another study conducted by Dr. Julie Ross to study the risk factors for leukemia in children with Downs syndrome. It also will obtain interview data that will provide information on risk factors for Downs syndrome itself.

Project A: West Coast Childhood Brain Tumor Study

Children under age 20 living in 19 counties in the Los Angeles, San Francisco and Seattle areas with a primary brain tumor diagnosed from 1984 -1991 (540 cases and 840 matched controls) were studied. Papers published include: maternal consumption of cured meats and vitamin supplements during pregnancy; exposure to tobacco smoke; use of electric blankets and water beds; and head injury (see publications). Additional analyses include: epilepsy and anticonvulsant use; and residential water source and nitrite and nitrate content of household water as assessed from dipstick measurements (manuscripts in review); medication use by mothers during the pregnancy and by the children; birth characteristics (including birthweight, parental age, birth order and history of birth defects in the index child or his siblings); use of rubber baby bottle nipples and pacifiers; child's diet and other early exposures.

Project B: Pesticide Exposure and Childhood Brain Tumors

This study involved telephone interviews with mothers of 450 cases and their controls from the Los Angeles portion of the West Coast study of children to get details of pesticide exposure during gestation and childhood. The manuscript reporting findings is in preparation.

Project C: Maternal Diet and PNET: CCG - based study

Data collection has been completed for this nationwide telephone interview study of primitive neuroectodermal tumors (PNET), being conducted by Dr. Bunin (Children's Hospital of Philadelphia). As part of our collaboration, we arranged for the translation of study questionnaires and other materials into Spanish.

Project D: International Collaborative Case-Control Study of Childhood Brain Tumors and N-nitroso Exposures

Data collection for this 9-center international study is now complete and a pooled dataset is under preparation. Interview data are available from mothers and fathers of 1218 children with primary brain tumors and 2223 control children. Dr. Preston-Martin has received a grant from NINDS to support analysis of these data.

Project A: Childhood Brain Tumors and Residential Exposure to Magnetic Fields in Los Angeles

This study of 300 case and 300 control children involved visits to measure magnetic fields and map the wiring configurations at 1131 homes in Los Angeles County where these 600 children had lived. Findings were reported in 1996 along with a response to an invited commentary on the evolution of epidemiologic evidence on magnetic fields and childhood cancer. This study utilizes the expertise of the Study Design and Statistical Methodology Research Core of the Center.

Project B: Control Selection Methods for EMF Studies

Random digit dialing (RDD) may not be an unbiased method of control selection for case-control studies. In a study of childhood leukemia and EMF using RDD controls we found an association between wiring configurations and leukemia risk. To investigate whether differential telephone answering patterns influenced these findings we are investigating a new control selection procedure, called the case-specular approach using the "mirror-image" of the house of the cases and controls. This project used the expertise of the Study Design and Statistical Methodology Research Core.