Childhood leukemia and residential magnetic fields: are pooled analyses more valid than the original studies?

The association seen in epidemiological studies between childhood leukemia and magnetic field strength in the child's home has been very important in influencing reviews of international groups and standard setting organizations. This association is usually based on the results of two published pooled analyses, which use definitions of exposure that differ from those of some the original studies. However, the results and conclusions of the pooled analyses differ from those of the three largest recent studies, which have the most sophisticated methodology and together account for the majority of the exposed cases at high exposure levels in the pooled analyses. These recent studies, using the exposure methods and the cut-off levels set a priori, each concluded that there was little evidence of any association. The pooled analyses, using different exposure measures and different cut-offs, conclude that an association exists at high exposure levels. It is not clear if the results of the pooled analysis are more valid than those of the recent major studies, although this has been often assumed in influential reviews.