Studies in the US, Italy, UK and other countries have reported increases over time of the incidence rates of acute lymphatic leukaemia (ALL). We have previously reviewed the findings of studies published up to 2003: patterns have not been fully consistent across countries, and independent analyses of the US Surveillance Epidemiology and End Results programme (ie, the greatest available case series) have not provided unequivocal results.¹ A pooled analysis of European data carried out at the International Agency for Research on Cancer² estimated a statistically significant increase of 1.4% per year starting in 1970. This and other studies suggest an increase over time for all childhood cancers considered together and for most individual childhood cancer types, but with somewhat smaller increases than for ALL.

How should these observations be interpreted? Most studies are based on analyses of routine statistics such as those produced by population-based cancer registries. The data underlying the European findings² had been subjected to standardised procedures to evaluate quality of cancer registration, to uniform …