This paper challenges a common basic assumption in risk assessment, that there is a threshold level of contamination below which no effects are seen or caused. Their results and analysis present compelling proof that this assumption is wrong.

What did they find?
Their fundamental finding is that any addition of an estrogenic contaminant with will cause an effect. For the system in which they worked, there is no threshold. This is because in their experimental system, endogenous estrogen is already at a high enough level to exceed the threshold for causing an effect. Endogenous estrogen is already activating the system. A contaminant doesn't have to exceed the threshold because endogenous estrogen already does.

Sheehan et al. show this to be the case experimentally, with a very large sample size, and then show through theoretical analysis why it is so. The combination of experimentation and theoretical analysis is powerful. They conclude that their results demonstrate "that no exogenous estrogen is without risk."

What does this mean?
The "threshold assumption" is central to the entire approach used by risk assessment for establishing the health safety of chemicals. It has been widely accepted but rarely tested. Its use in regulatory science has been a pragmatic step, not something based on theory or on fact. This paper stands out as a strong test of that assumption--strong because of the sample size and the analysis--and the assumption comes up short.

From Sheehan et al.: "Over 70,000 man-made chemicals that have an aggregate value of billions of dollars are found in food, water, air, or soil. Given the central role of the threshold assumption in evalution of health safety, the exposure of all organisms to synthetic chemicals, the importance of these chemicals in modern society, and their huge production volume and economic value, it is surprising that the threshold assumption has been so widely accepted and so rarely tested."

That is surprising. And it is deeply disturbing that when the assumption is tested, as Sheehan et al. have done, it is disproven.

Why is this important?
This assumption is a key part of the way that safety standards are set. In experiments designed to explore toxicity, an important goal is to establish a "no observed adverse effect level" or NOAEL. This is the level of exposure that produces no statistically significant increase in adverse effects. Typically experiments start at high contamination levels and work downward along the exposure curve until no statistical effect is seen. That's the NOAEL. A fudge factor is then added because of differences among species. Typically this means that the NOAEL is divided by some number, often 100. The assumption is that an exposure level calculated in this fashion is safe, and it is used to determine acceptable exposure levels.

What Sheehan et al. show is that this approach is wrong if the contaminants share a common mechanism with endogenous chemicals that are already at a level above a concentration sufficient to cause an impact. (Note: there is another assumption in this approach--that the dose-response relationship is monotonic--that science is showing is also commonly violated by endocrine disrupting compounds). That condition (sharing a common mechanism...) is likely to be quite common for endocrine disrupting chemicals.

So the basic message of this paper is that one of the fundamental assumptions used to guide risk assessment doesn't work with endocrine disruption. It's simply wrong.

What did they do?
Sheehan et al. worked experimentally with sex control in the red-eared slider, a turtle in which sex determination is normally controlled by temperature (via a mechanism in which the hormonal processes involved in sex determination are temperature dependent; more...). They exposed a series of turtle eggs at 28.6°C to a range of doses of 17ß-estradiol. The temperature they chose normally would have resulted in mostly males but some females. They then determined the sex of each egg at hatching. They analyzed the results using a theoretical construct based on the Michaelis-Menten equation, which has been developed in basic chemistry to model enzyme kinetic studies. The data from the large experiment fit the M-M model extraordinarily well. Their analyses showed that any addition of exogenous estrogen caused a change in the sex ratio of pool of eggs.

Talk to us: email