Background Parental ages, parity, and social class have been found in some studies to be associated with particular childhood cancers. Further investigation is warranted because of conflicting findings, biases, and the need to test specific hypotheses.

Methods A case-control study was conducted (England and Wales, ages 0–14 years). Cases were ascertained from the National Registry of Childhood Tumours, and were born and diagnosed during 1968–1986. Birth record controls were matched 1:1 to cases on date of birth, sex and area. Information on variables of interest for both groups came from birth records. In all, 10 162 pairs could contribute to matched analyses.

Results The odds ratio (OR) for retinoblastoma resulting from assumed new germ cell mutations among children of fathers aged ≥45 years was 3.0 (95% CI : 0.2–41.7). The risk of childhood acute lymphoblastic leukaemia (ALL) was significantly higher among children of older mothers and fathers, and significant trends with increasing mothers' (P < 0.001) and fathers' (P = 0.002) ages were found. There was a strong and significant protective effect of increasing parity on risk of childhood ALL. The adjusted OR for parity of ≥5 (versus 0) was 0.5 (95% CI : 0.3–0.8). Children in more deprived communities had a lower risk of ALL; but this was not significant after confounders were allowed for. There was no significant effect of social class based on parental occupation on ALL risk, but the numbers were small in those analyses.

Conclusions The associations between ALL and parental ages did not disappear when children with Down syndrome were excluded, suggesting an additional explanation beyond known links. The strong ALL association with parity may be because of an unknown environmental risk factor.