Search

This thesis was designed to evaluate the effects of birth weight, maternal age and parental social class on the risk of childhood leukemia using a population based case control study of 798 participants in five Canadian provinces. Children between the ages of 0-14, who were diagnosed between 1990 and 1994 in Alberta, Saskatchewan and Manitoba, or 1990 to June 1995 in British Columbia and Quebec were eligible for inclusion in the study. Controls were matched on age, gender and province of residence. Response rate was 90% for cases and 76% for controls and data was collected via personal interview with parents. The primary area of investigation was the assessment of exposure to electromagnetic fields as a risk factor for childhood leukemia. Interview data on a large number of variables were available allowing for consideration of other potential risk factors. Univariate analysis suggested a case group with lower maternal age at birth, lower levels of household income, maternal and paternal education and occupational status. There was no evidence of a significant effect for birthweight on disease risk. In a multivariate analysis, four models were constructed encompassing variables related to social class, behaviour, demographics and environmental tobacco smoke. These were individually assessed using logistic regression techniques. The persisting variables were then combined and assessed in a final logistic model. Variables shown to affect disease risk in the final model were Asian ethnic origin, (OR = 3.65, CI = 1.27-10.52, p value = .02); maternal smoking of 10-20 cigarettes per day during pregnancy, (OR = 1.80, CI = 1.12-2.90, p value = .02); mother's usual occupation in high SES category, OR = 0.48, (CI = 0.26 - 0.88, p value = .02); maternal alcohol use of 1-2 drinks per week in the month before pregnancy, OR = (1.77, CI = 1.17 - 2.70, p value = .007).