Flower, KB, JA Hoppin, CF Lynch, A Blair, C Knott, DL Shore, DP Sandler. 2004. Cancer Risk and Parental Pesticide Application in Children of Agricultural Health Study Participants. Environmental Health Perspectives 112: 631-635.

Flower et al. report that children of farmers licensed to apply pesticides are greater risk of childhood cancer, particularly lymphoma.

While previous studies have implicated parental pesticide exposure as one possible cause of childhood cancers, this study avoids some of the weaknesses of earlier work by using detailed data on pesticide application practices. It used data from the Agricultural Health Study, a large survey of licensed pesticide applicators and their families in Iowa and North Carolina, to further investigate the relationship between parental pesticide exposure and childhood cancers.

 

What did they do?

Flower et al. analyzed data from the Agricultural Health Study (AHS) on cancer in children of Iowa farmers, and compared those data to information about cancer cases in Iowa's children as a whole.

For the comparison they used an index, the Standardized Incidence Ratio [SIR]. The SIR is the ratio of the observed new cancer cases to the number that would be expected based on the rate in the general population. It is corrected to take into account differences in age, sex, race and time period.

An SIR of one means the observed rate is the same as expected, and thus there is no increased risk. Rates greater than one indicate an increased risk of cancer. For example, a SIR of 1.60 indicates a 60% increase in cancer risk.

 

In the Agricultural Health Study, pesticide applicators (including farmers) and their spouses were asked via questionnaires about pesticide application practices, health, and health-related behaviors. Information on occupational exposures to 50 pesticides and on use of protective equipment, such as chemically resistant gloves, during pesticide application, was included. Female participants, who were either farmers licensed to apply pesticides or spouses of farmers who were licensed, were asked to provide information on children born during or after 1975. A total of 17,357 children from Iowa born during or after 1975 were identified, out of 21,375 total ( Iowa and North Carolina).

The AHS data cover individuals who applied for pesticide application licenses in Iowa and North Carolina between 1993 and 1997. This study focused only on data from Iowa, and only on farmers applying pesticides.

What did they find?

The study identified 50 children with cancer in the Iowa AHS study population.

Children of farmers applying pesticides had elevated rates of cancer compared to children in the general population of Iowa. For all cancer types combined, the SIR was 1.36.

Twice as many children of Iowa farmers licensed to apply pesticides had childhood lymphoma, and 2.5 times as many had Hodgkin’s lymphoma, compared to the number expected. For lymphoma and Hodgkin's lymphoma, SIR values were 2.18 and 2.56, respectively. Statistical analysis concluded both these SIR values, as well as the overall SIR of 1.36, were unlikely to be due to chance.

SIR values greater than one were observed in several other cancer, including brain, bone and germ cell cancer, but there were small numbers of cases of these cancer types and the calculated SIRs fell short of statistical significance.

Children whose fathers did not use chemically-resistant gloves while mixing and/or applying pesticides were twice as likely to have cancer compared to the children of those who did [odds ratio of 1.96, with 95% confidence limits from 1.05–3.76].

The research team examined the data for patterns that might reflect increased risk associated with increased use. No patterns emerged, for example, between the frequency of pesticide application and the risk of childhood cancer.

Flower et al. also looked for associations between cancer risk and specific pesticides that parent farmers reported to have used. The numbers of cancer cases associated with any given one of the 16 specific pesticides applied by parents was small, making statistical analysis difficult. Only one—aldrin—was significantly associated with an increased risk of childhood cancer (OR = 2.66, with the 95% confidence interval extending from 1.08 to 6.59). This result should be viewed cautiously, because the study was not powerful enough to be able to confidently link specific pesticides with cancer (only 6 cases of cancer were linked to paternal use of aldrin) and since other human studies of aldrin have not found a cancer link.

What does it mean?

Including more detailed information on specific pesticide exposures than many previous studies, this research adds to a significant body of literature linking parental pesticide exposure to an increased risk of cancer in children. The results showing heightened risk to Hodgkin's lymphoma reinforce hints from prior work, none of which had yielded statistically significant findings.

The study did not find an association with leukemia or brain/central nervous system tumors, one of the most consistent patterns in earlier studies involving childhood cancers and parental pesticide use. This may have resulted from the relatively small number of cases, or from the fact that most of the parents in the study were men, reducing the study's ability to detect the effect of maternal exposures.

Of practical import, the findings did reveal heightened risk for the children of farmers not using chemically-resistant gloves. Flower et al. note that this increased risk could reflect greater direct exposure of unprotected farmers to pesticides, who subsequently carry the chemicals home and expose their children, or simply less meticulous chemical practices generally, which could increase the opportunity for exposure to children on the farm.