OBJECTIVE To assess unexpected associations between childhood cancer and pethidine given in labour and the neonatal administration of vitamin K that had emerged in a study performed in the 1970 national birth cohort. METHODS 195 children with cancer diagnosed in 1971-March 1991 and born in the two major Bristol maternity hospitals in 1965-87 were compared with 558 controls identified from the delivery books for the use of pethidine during labour and administration of vitamin K. METHODS Odds ratios for cancer in the presence of administration of pethidine or of intramuscular vitamin K. Both logistic regression and Mantel-Haenszel techniques were used for statistical analyses. RESULTS Children of mothers given pethidine in labour were not at increased risk of cancer (odds ratio 1.05, 95% confidence interval 0.7 to 1.5) after allowing for year and hospital of delivery, but there was a significant association (p = 0.002) with intramuscular vitamin K (odds ratio 1.97, 95% confidence interval 1.3 to 3.0) when compared with oral vitamin K or no vitamin K. There was no significantly increased risk for children who had been given oral vitamin K when compared with no vitamin K (odds ratio 1.15, 95% confidence interval 0.5 to 2.7). These results could not be accounted for by other factors associated with administration of intramuscular vitamin K, such as type of delivery or admission to a special care baby unit. CONCLUSIONS The only two studies so far to have examined the relation between childhood cancer and intramuscular vitamin K have shown similar results, and the relation is biologically plausible. The prophylactic benefits against haemorrhagic disease are unlikely to exceed the potential adverse effects from intramuscular vitamin K. Since oral vitamin K has major benefits but no obvious adverse effects this could be the prophylaxis of choice.