Abstract

OBJECTIVES: To investigate the hypothesis that children of men whose jobs involve contact with many people (particularly children) are at an increased risk of leukaemia. METHODS: A population based dataset obtained from routinely collected death certificates involving 14,168 cancer deaths occurring before the age of 15 years registered in England and Wales between 1959-63 and 1970-90. Associations were assessed with the proportional cancer mortality ratio (PCMR), with all childhood cancer deaths forming the standard for comparison. The PCMRs were adjusted, by stratification, for age and year of death (in 1-year bands) and paternal social class (nine categories). Analyses were performed by estimated level of paternal occupational social contact (high, medium, and low) for all leukaemias, leukaemia subtype, age at death, year of death, and individual occupation. RESULTS: Out of 223 occupations, 36 (16%) were identified as having potentially high levels of social contact, and 27 (12%) as having potentially medium levels of social contact. No associations were found between paternal occupational social contact and death during childhood from leukaemia (high social contact: PCMR 94, 95% confidence interval (95% CI) 87 to 102; medium social contact: 101, 95 to 106). No associations were found when the data were analysed by leukaemia subtype, age at death, year of death, or individual occupation. CONCLUSION: The findings presented here do not support the suggestion that childhood leukaemia is related to the amount of social contact that fathers experience at work.