Environmental exposure and childhood leukemia in Brazil: an exploratory analysis of their association

This study aims to explore the unapparent relations that several factors related to environmental exposure and individual characteristics existing in our environment may have with the process of developing childhood leukemia. From a database obtained from a clinical and epidemiological hospital-based, case-control study on risk factors for childhood leukemia, an exploratory multivariate analysis was performed using principal component analysis and factor analysis. This research is part of a national multicenter study that included 292 cases of leukemia in children aged 0 to 12 years and 541 controls of the same age, hospitalized for non-neoplastic causes in general hospitals near the centers the cases originated in. Information on selected environmental exposure was obtained in interviews with the mothers of both cases and controls by means of a standardized questionnaire. The model with the greatest explanatory power for the variance observed in the data analyzed was of approximately 52%. Three factors were considered most appropriate for predicting leukemogenesis in childhood, each including variables with factor loadings greater than 0.6: factor "conditions related to chemical exposures during pregnancy", which explained 20% of the final variance and included the variables pesticide exposure, exposure to solvents and paint exposure in pregnancy; factor "lifestyle habits during pregnancy" explained 17% of the variance and included exposure to hair dyes and cosmetics for hair straightening; and factor "use of health services during pregnancy", which explained 15% of the variance and included the variables type of delivery (vaginal or caesarean) and use of radiography in pregnancy. Logistic regression analysis revealed a statistically significant association between the development of leukemia in childhood and maternal history of chemical exposure during pregnancy (OR=1.36, 95% CI=1.16-1.59) and use of health services during pregnancy (OR=1.27, 95% CI=1.08-1.49). The results indicate the joint contribution of not just individual but environmental exposure in the development of leukemia in childhood, and are supported by evidence in the literature that the process of carcinogenesis in general and of leukemogenesis in particular, result from effects of multiple mutations related to joint environmental exposure.

Leukemia; Childhood; Environmental exposure; Factor analysis; Logistic regression