Methodological problems in assessing health-related, neuropsychological effects of lead absorption in a very low-level exposed area.

Methodological problems in the low level lead studies are reviewed using the Aarhus lead study as an example. It is shown that a lead effect can be found in an area where the background blood lead level is as low as 37 micrograms/l (geometric mean). Even in such an area it is worthwhile looking for populations at risk. It is shown that attrition causes confounding with a directional bias towards the null-hypothesis. Longitudinal studies are associated with this type of bias. Misclassification as for past exposures will also have a bias of this type. Studies depending on blood-lead measures are liable to have this type of bias. This is the case to a lesser degree in studies using cumulated indices. Subsamples of tooth dentin (circumpulpal dentin) satisfies the requirement for an index for cumulated lead absorption. Since lead absorption requires motor abilities, medical factors that are risk factors for motor development as well as for the other developmental outcome are also confined with a directional bias towards the null-hypothesis. It is suggested to exclude for such factors. Lead seems to be a ubiquitous noxious substance. Any community should have a strategy to minimize its effects.