Parental smoking and maternal alcohol and caffeinated beverage consumption are prevalent exposures which may play a role, either directly or through their influence on metabolism, in the aetiology of childhood malignant central nervous system (CNS) tumours. The hypothesis was investigated in the Epidemiological Study on childhood Cancer and Leukemia ESCALE study, a national population-based case–control study carried out in France in 2003–2004. The study included 209 incident cases of CNS tumours and 1681 population-based controls, frequency matched with the cases by age and sex. The data were collected through a standardized telephone interview of the biological mothers. No association between maternal smoking during pregnancy and CNS tumours [odds ratio (OR): 1.1 (0.8–1.6)] was observed. Paternal smoking during the year before birth was associated with CNS tumours (P for trend=0.04), particularly astrocytomas [OR: 3.1 (1.3–7.6)]. Maternal alcohol consumption during pregnancy was not associated with CNS tumours. Associations between ependymomas and the highest consumption of coffee [OR: 2.7 (0.9–8.1)] and tea [OR: 2.5 (1.1–5.9)] were observed. A strong association between CNS tumours and the highest maternal consumption of both coffee and tea during pregnancy was observed [OR: 4.4 (1.5–13)]. The results constitute additional evidence for a role of paternal smoking and suggest that maternal coffee and tea consumption during pregnancy may also increase the risk of CNS tumours. The study does not suggest an increased risk of CNS tumours related to alcohol consumption during pregnancy.