Occupational exposure to pesticides and risk of leukemia among offspring in Costa Rica

Abstract

Background: Leukemias are the most common childhood cancers, accounting for 25-35% of the incidence of all childhood cancer in most populations. In early reports on cancer, Costa Rica ranks among the highest incidence of childhood leukemia in the world. Agriculture is an important economic activity in Costa Rica and is characterized by intensive use of pesticides. In this thesis we present a descriptive epidemiologic study of childhood leukemias, two methodological studies on parental pesticide exposure assessment and an etiological analysis based on a population-based control study.

Objective: The main objective of this study was to increase the understanding of potential risk factors of childhood leukemia in Costa Rica, propose improved methods for exposure assessment in cancer research in developing country conditions, and provide epidemiological evidence of association between childhood leukemia and parental occupational pesticide exposure during the prenatal and early postnatal periods of children.

Methods: Childhood leukemia incidence during 1981-1996 was analyzed by histology, gender, birth year, time period of diagnosis, age at diagnosis and region. Case data were extracted from the National Cancer Registry of Costa Rica. Person years at risk were calculated from census data and annual sample survey data. International comparison of rates was performed (Paper I). We describe an icon-calendar interview form (ICF) for assessing retrospective parental exposure to pesticides, collecting data on agricultural pesticide use, job tasks, frequency of exposure and major determinants of pesticide exposure (Paper II). ICF data were combined with external data of pesticide application rates for the country, in two retrospective exposure assessment models (Paper III). Associations between parental exposure to pesticides and the risk of leukemia in the offspring were examined in a population-based case-control study in Costa Rica. All cases of childhood leukemia (N=334), ages 0-14 at diagnosis, during 1995-2000, were identified at the Cancer Registry and the Children's Hospital of Costa Rica (ICD-0-1). Population controls (N=579), were drawn from the National Birth Registry (Paper IV).

Results: The reported cases of childhood leukemia between 1981-1996 represented an overall agestandardized incidence rate of 55 per million person-years. Acute lymphocytic leukemia (ALL) accounted for 79% of all leukemias. A Poisson regression model identified significant excesses of ALL for boys, groups of 1-4 and 5-9 years of age and for three out of seven geographical regions. Based on the first 100 ICF, recall was acceptable in 90% of the interviews, successful identification of special events in 84%, good recovery of pesticide names, frequency of use and application techniques in 90%, report of changes in pesticide use patterns in 63%. Numbers of exposed mothers and fathers for each prioritized pesticide and its exposure intensities for five time windows, were estimated with the exposure assessment models. A sensitivity analysis suggested satisfactory robustness of the model weights. In the case-control study, 876 mothers and 762 fathers were included in the analysis. Mothers' exposures to any pesticides anytime and during year before conception and 1 st and 2nd trimesters, and fathers' exposures to any pesticides during 2nd trimester were associated with excess risks. Excesses of total leukemia were found for mothers and fathers exposed to organophosphates and for fathers exposed to benzimidazoles. There was a suggestion of exposure-response gradients for fathers for picloram, benomyl and paraquat. Age at diagnosis was positively associated with fathers' exposures and inversely with mothers'. Mothers' exposures seemed more important than fathers'.

Conclusions: Based on worldwide epidemiologic data, including ours, we believe that parental and early childhood exposure to pesticides may be associated with risk of leukemia in the offspring. The most likely candidates for etiological pesticidal compounds appear to be organophosphates as group and - still with considerable uncertainty - picloram, benomyl, paraquat, foxim, mancozeb, malathion, dichlorvos and metam sodium. Epidemiological research on childhood cancer needs to be further conducted in developing countries, including etiological research on pesticides. There is a need to study biomarkers for assessing biological processes in leukaemogenesis, susceptibility and exposure. We recommend prevention of pesticide hazards, especially in the population stratum reproductively active and among children.