Objectives The authors performed a meta-analysis ofcaseecontrol and cohort studies to clarify the possiblerelationship between exposure to pesticides andchildhood cancers.Methods Two cohort and 38 caseecontrol studies wereselected for the first meta-analysis. After evaluatinghomogeneity among studies using the Cochran Q test,the authors calculated a pooled meta-OR stratified oneach cancer site. The authors then constructed a list ofvariables believed to play an important role in explainingthe relation between parental exposure to pesticide andchildhood cancer, and performed a series ofmeta-analyses. The authors also performed a distinctmeta-analysis for three cohort studies with RR data.Results Meta-analysis of the three cohort studies didnot show any positive links between parental pesticideexposure and childhood cancer incidence. However, themeta-analysis of the 40 studies with OR values showedthat the risk of lymphoma and leukaemia increasedsignificantly in exposed children when their mother wasexposed during the prenatal period (OR¼1.53; 95% CI1.22 to 1.91 and OR¼1.48; 95% CI 1.26 to 1.75). Therisk of brain cancer was correlated with paternalexposure either before or after birth (OR¼1.49; 95% CI1.23 to 1.79 and OR¼1.66; 95% CI 1.11 to 2.49). TheOR of leukaemia and lymphoma was higher when themother was exposed to pesticides (through householduse or professional exposure). Conversely, the incidenceof brain cancer was influenced by the father’s exposure(occupational activity or use of household or gardenpesticides).Conclusion Despite some limitations in this study, theincidence of childhood cancer does appear to beassociated with parental exposure during the prenatalperiod.