Lead in the average Joe: no increase in cancer risk. (Science Selections).

Although classified as a possible carcinogen in humans, lead is an established carcinogen in experimental animals; the administration of inorganic lead in rats and mice has resulted in kidney, nervous system, and lung tumors. So far, however, a definitive blood lead-cancer relationship has not been established in humans. In this issue, Ahmedin Jemal of the National Cancer Institute and his colleagues investigated the association between blood lead concentrations and cancer mortality among Caucasians (whites) in the general U.S. population [EHP 110:325-329]. Although a mild dose-response relationship was noted among women with very high blood lead concentrations, none of the site-specific cancers examined showed a statistically significant increase in risk with chronic lead exposure.

Many of the epidemiologic studies that relate lead with cancer have been conducted in occupational settings; no previous study is known to have examined the association between lead and cancer in the general population. Jemal and colleagues used data from the National Health and Nutrition Examination Survey (NHANES) II Mortality Study of 1992, which was designed to examine the association between factors measured at baseline with overall or cause-specific mortality among a large general population. They restricted their study to whites because both blood lead concentrations and cancer deaths vary by race, and the number of deaths in other races in the Mortality Study was too small to provide reliable estimates.

The researchers investigated a total of 203 cancer deaths (117 men, 86 women) among 3,592 white people, with an average followup of 13.3 years until death from cancer. They used a statistical method known as Cox proportional hazard regression modeling to estimate dose-response relationships between blood lead and all cancer mortality. Relative risks were estimated for site-specific cancers (such as kidney tumors) by categorizing blood lead concentrations as high, medium, and low (for classification into groups only). The study also took into account nutrition, consumption of alcoholic beverages, and tobacco use.

A statistical test of data from the followup period indicated that cancer risks did not increase with increasing blood lead concentrations among men and women combined and among men alone, but there was a marginally significant increase among women alone. Women seemed to show a possible threshold effect at about the 94th percentile of lead, which corresponds to a blood concentration of 24 [micro]g/dL among all study subjects. Further evaluation among women, however, showed virtually similar relative risks.

With respect to the association of blood lead concentration with selected site-specific cancer mortality, none of the site-specific cancers showed a statistically significant excess risk with elevated concentrations. Among the combined sexes and men, risks did increase with increases in blood lead concentration, but not significantly for prostate cancers among men and brain cancers among both sexes. Among women, there was no clear pattern.

Jemal lists several limitations in the database and analyses used. For instance, in the Mortality Study, people not known to be deceased were assumed to be alive; thus, a misclassification of vital status was possible. Also, measurement of blood lead occurred once and may not reflect cumulative exposure, but provides only an approximation. This becomes important because blood lead concentrations have been decreasing in the U.S. population, due, for instance, to the discontinued use of leaded gasoline, paint, and solder.

Whether lead causes cancer in humans is thus still not well established. The apparent dose-response relationship found in women with the highest blood lead concentrations could be either a chance finding or attributable to certain confounding factors. The authors write that this finding "has no clear biologic explanation; further studies of populations with sufficiently high lead exposure are needed to replicate the finding among women before it is believable."