Of Mice and Men

The Environmental Protection Agency has compiled a good record in dealing with such large-scale pollutants of air and streams as sulfur dioxide, nitrogen oxides, and biochemical oxygen demand. Public concern now has shifted to the possible effects of the more than 65,000 different chemicals listed as having been in commercial production since 1945. Current statistics on mortality and morbidity do not indicate a cancer epidemic or other major health problems traceable to chemicals. The public harbors fears, nevertheless, and the EPA must deal with them.

Assessing risk

In 1984, the Agency took a giant constructive step by announcing and beginning implementation of a program of risk assessment separate from risk management. Emphasis now is being placed on identifying and reducing major risks. Risk assessment includes taking into consideration such scientific evidence as is available. In many cases, of course, the evidence is scanty or flawed.

In the early days of concern about possible carcinogenic chemicals, a super cautious approach could be justified on the basis of lack of information. However, as the years pass, pleading ignorance is less and less a tenable excuse. The EPA should be more active in supporting research.

In particular, a hard look needs to be taken at the soundness of present treatment of toxicological data. The current procedure of extrapolating from massive doses in highly inbred tumorgenic mice and rats to low doses in heterogeneous human populations seems unlikely to yield reliable results. An additional questionable practice is the assumption of no threshold for all cases of chemical carcinogenicity. This assumption is only an unproven hypothesis. In contrast, there is good evidence of thresholds, for example, in the effects of selenium and formaldehyde. Vitamin A offers another example: small amounts are essential to health; very large doses cause birth defects.

Real-world data

In the midst of studies involving untold numbers of rodents, the federal agencies should not forget that the species of importance is humans. Thus, more emphasis should be placed on human data than is now the case. Granted that epidemiological evidence is hard to interpret because of uncertain data and confounding factors. But has this resource been adequately exploited? For example, for decades a large number of people have used diet drinks containing saccharin. Others have not, and populations might well be available for different studies. When many students took courses in organic chemistry twenty to fifty years ago, they were exposed to substantial quantities of what now are labelled carcinogens. Another group are the veteran operators of chemical plants. Establishing exposure levels would be difficult, but even with the uncertainties involved, the data could prove useful.

In reducing our ignorance of which chemicals do what and to whom, those we entrust with protecting the public's health and environment should pay at least equal attention to the uncollected and unanalyzed data of the real world as they do to the painstaking experiments of the laboratory. Rats and mice may offer clues, but not conclusions.