Vom Saal is an endocrinologist at the University of Missouri at Columbia.

The endocrine system is a complex network through which chemicals send signals to many organs in the body. While we tend to think that hormones kick in with a vengeance first at puberty, in fact the developing fetus is exquisitely sensitive to these hormonal messages.

The major concern regarding endocrine disruptors is with exposure during critical times in fetal development when organs are forming. While exposure in adults to synthetic hormones such as the birth control pill has reversible effects, even very small doses of endocrine disruptors during fetal development may result in diminished or absent reproductive capability that can result in the loss of entire populations.

Decades ago, studies in the Great Lakes in North America confirmed suspicions that pollutants were threatening the survival of several species of fish and birds. Needless to say, these findings led scientists to wonder whether the human population, which is exposed to low doses of contaminants, might also be at risk. These concerns spawned testing of laboratory animals in an attempt to understand the potential risk of exposure to exogenous hormones.

Until recently, testing of endocrine-mimicking chemicals commonly found in fabrics, pesticides, detergents, and plastics has been conducted on animals by giving them high doses of these compounds. The traditional assumption underlying such testing measures has been that there is a simple dose-response relationship, that is, the higher the dose, the larger the potential for adverse outcomes.

But new research indicates the testing model has been based on erroneous assumptions. There are now numerous findings that conflict with prior toxicological studies that concluded that the concentrations of endocrine disrupting chemicals in the environment are safe.

Consider, for example, recent experiments by my colleagues and I on fetal male mice. During fetal development, a very small amount of the natural hormone estradiol is present in mouse serum. Exposure of male fetuses to an extremely small supplement of estrogen resulted in an abnormally enlarged fetal prostate. Furthermore, the effect was permanent since the adult mice developed enlarged prostates relative to untreated mice.

How small a dose does it take to produce adverse effects? While the normal levels of estradiol are two tenths of a picogram per milliliter of serum, the addition of one tenth of a picogram, a 50 percent increase but still only a tenth of a trillionth of a gram, caused abnormal prostate development.

I have also investigated the effects of the estrogen-mimicking chemical bisphenol-A, which is found in the lining of food and beverage cans, dental sealants, and plastic bottles, including baby bottles. Previous studies conducted by the plastics industry tested for doses much higher than those encountered in the environment by most people, and that research revealed that high doses were not in fact harmful, thus leading the U.S. government to believe the ambient levels of bisphenol-A were safe.

However, recent studies involving the use of much lower doses and examination of organ function, which was not previously examined, produced numerous effects in male adult offspring of pregnant mice, including enlarged prostates and lower testicular sperm production due to exposures to doses similar to those encountered by people using products made from bisphenol-A, such as canned foods, dental sealants, and polycarbonate food and beverage containers.

Other studies also produced early onset of puberty in females and changes in the breast and pituitary gland at the same low doses described above. In fact, doses 25,000 times lower than doses used in studies by the plastics industry produced these effects.

In short, research conducted in the past decade has turned the concept of "safe doses" of endocrine-mimicking hormones on its head. The current risk- assessment model that provides the basis for the design of toxicological tests assumes that there is a simple relationship between dose and response; that is, response increases or stays the same with each increase in dose. This model cannot accommodate data sets found in experiments with hormones where response first increases then decreases as the dose increases.

Using information from high-dose studies to predict the probability of effects at lower levels is to use circular reasoning. Our experimental results show that in a traditional toxicology study design, the endocrine disrupting effects at low doses, in many cases similar to amounts being consumed by people, could be missed.

Furthermore, since we know that breast cancer can occur from natural levels of estrogen, women are already over the threshold of the hormone required to induce cancer. Therefore, we must assume that, in the case of endocrine-mimicking hormones, the lack of a threshold dose means that every dose carries risk; there is no safe dose, only levels of risk that society deems acceptable.

Vom Saal's article "A Challenge to Risk Assessment: Low-dose Effects of Endocrine Disruptors" will appear in the Fall 1998 issue of FORUM for Applied Research and Public Policy.