Authors
Xiaomei Ma, Patricia A. Buffler, Robert B. Gunier, Gary Dahl, Martyn T. Smith, Kyndaron Reinier, and Peggy Reynolds.

Title
Critical windows of exposure to household pesticides and the risk of childhood leukemia

Journal
Environmental Health Perspective. 110: 955-960. 2002.

Summary
The goal of Northern California Childhood Leukemia Study (NCCLS) is to explore the etiological role of household pesticide exposure and risk of childhood leukemia. Although the relationship between household pesticide exposure and childhood leukemia has been examined in numerous studies the results have been inconclusive or inconsistent, making it difficult to clarify the role of pesticides in the etiology of this cancer. The inconsistencies of these previous studies have been due to limitations including small numbers of cases, study design, problem in recall and reporting of household pesticide exposure, matching criteria, method and timing of data collecting, definition of exposure, enlistment control, and the use of alternate respondents. The NCCLS was designed to collect data collection on household pesticide exposure using in-home personal interviews.

This study used a total of 162 newly diagnosed leukemia patients during 1995-1999 and 162 matched controls which were randomly selected from the birth registry. In the NCCLS each pesticide was classified according to the name and the purpose of use, frequency of use, and the time of use (3 months before pregnancy, pregnancy, years 1, 2, and 3 after birth). Pesticides and insecticides were categorized according to professional services or individual uses, indoor, and outdoor use. The insecticide exposure include professional pest control services, use of different products to control ants, flies, cockroaches, spider, termites, and plant/tree insects. Herbicide exposures consisted of professional lawn and weed control services, flea collars, indoor fogger, and flea control products.

The results of this study suggest that the use of professional pest control services is associated with an increased risk of leukemia, when exposures occur at any time from one year before birth to three years after (OR= 2.8 and 95% confidence interval 1.4-5.7). The indoor pesticide exposure was associated with an elevated risk of childhood leukemia while no significant association was observed with outdoor pesticide exposure. There was no association between use of flea control products and the risk of leukemia.

The authors of the study demonstrated that the exposure to insecticide and risk of leukemia varied with the highest risk observed for exposure during pregnancy (OR= 2.1; 95% CI, 1.1-4.3) and lowest during year 3 (OR= 1.2; 95% CI, 0.7-2.2). Exposure to insecticide significantly increased the risk of leukemia during the 4 year period (from 1yr before birth to age 3) this study (OR= 2.2; 95% CI, 1.1-4.3) examined. Also more frequent exposures to insecticides were associated with a higher risk of leukemia. However, leukemia risk following exposure to insecticide during pregnancy was higher than exposure after birth, suggesting that the fetus is potentially more sensitive to carcinogens in the insecticides used. It has been hypothesized that some chromosome translocation events related to childhood leukemia have a prenatal or even a preconceptual origin.

Overall, the design of this case-control study was to reduce the general biases in epidemiological studies by exposure classification error and distinguishing between the risks associated with different types of pest control, and identifies the importance of the timing and location of exposure. The NCCLS is one of the first studies that addresses critical window of exposure, that is, the timing of household pesticide exposures. However, a limitation in this study can be related to use of broad subgroups of pesticides whereas no specific chemical was measured. Also, the small sample size used in this study reduces the confidence in the study results. Regardless, this study does bring to light an import point in the endocrine disruption and cancer field. Specifically, in vitro and animal tests for carcinogenicity of the active ingredients in these pesticides tend to yield negative results and suggest that these agents will not be carcinogenic in humans. If confirmed in future research, the results of the present study would suggest several potential problems such as: (1) current testing strategies are not sensitive enough to detect compounds that will be carcinogenic in humans; (2) there are ingredients present in the commercial mixture that are carcinogenic and have not been evaluated; or (3) that chemicals present in the commercial pesticide mixture interact to induce cancers. It may therefore be necessary to study commercial pesticide preparations for their potential to cause cancer in animals as a first step in unraveling this paradox.