Flawed research had the world on the run

When John McLachlan's work was published in Science on 7 June 1996, it was considered so important that the editors emphasised it with an additional news story in the magazine. News reports of the study made headlines nationwide and in Europe. Some experts called it a "red flag", and feared dire effects. The US Environmental Protection Agency's grandly titled Office of Prevention, Pesticides and Toxic Substances called the findings "astonishing", and ordered new studies.

The study suggested that, in combination, some common pesticides could "turn on" the oestrogen receptor in human cells - meaning that they would hugely multiply the apparent levels of the hormone in the environment. Given that chemicals that mimic the effect of oestrogen have been implicated in birth defects in male animals, fish turning hermaphrodite, falling human sperm counts and rising levels of breast cancer, it is not surprising that people took note.

The Tulane team tested four pesticides on genetically engineered yeast cells which contained the gene for the human oestrogen receptor, plus a "reporter" gene that made the yeast culture turn blue if a chemical bound to that receptor. On their own, the pesticides - dieldrin, endosulphan, toxaphene and chlordane - showed only a weak response. When they were tried in pairs, the activity shot up by between 160 and 1,600 times. Significant? You bet.

But in its 25 July issue this year, Science had another contribution from McLachlan, who works at Tulane University's Bioenvironmental Research centre, near New Orleans, Louisiana. It was a short announcement, which appeared at the end of the letters section: "Whatever merit [the original paper] contained, and despite the enthusiasm it generated, it is clear that any conclusions drawn from this paper must be suspended until such time, if ever, the data can be substantiated."

In other words, the work was wrong. No other organisation had been able to repeat it. And McLachlan's team had subsequently been unable to cross- confirm it with other work they had done. The university announced "an internal inquiry into the circumstances" of the whole episode. McLachlan declined to talk to the press.

And that was it: no news story (at least, not in Science), and no comments from environmental scientists or the EPA. The work is quietly consigned to the dustbin of "flawed experiments". In his letter, McLachlan added that "it seems evident that there must have been a fundamental flaw in the design of our original experiment".

Defending Science's low-profile treatment of the retraction, Diane Dondershine, the magazine's spokeswoman, said: "It's all part of the scientific process. Other labs tried to duplicate the work and couldn't. We hope other people's work wasn't affected, that they didn't use this to build their own work from." McLachlan approached the magazine in June to retract the paper, she said; the letter's publication had been "a pretty fast turnaround".

True, few scientific careers will have been hurt by this (and McLachlan's prompt action will probably ensure that his reputation remains intact). And as Dondershine pointed out, it is a rare event: in the past two years there have been two or three retractions, out of 1,070 papers published in the magazine.

But what about the people who leapt to action on seeing the results?

In the world of science, it is a year's wasted effort; but in the sense that it confirms scientific method, perhaps that is good. The chemical industry and universities mobilised themselves and began testing pesticide mixtures, and nothing turned up.

But what about the effects in the scientific world? In halls of legislature in Washington and elsewhere, the research shaped environmental debate, creating what one lobbyist called "a real sense of urgency". Eventually, fears amplified by the study caused Congress to put oestrogen research requirements into the Safe Water Act and the Food Safety Protection Act, which were passed last year. The issue also affected legislation in New York state. The EPA is forming a committee to formulate regulations.

Normally, when the feedback loop between research and political effect is this tight, it seems like cause for celebration. But in this case, it was built on a piece of science that turns out to have been faulty. The tight feedback loop seems more like a cause for worry.

We have had our own version of this same scenario in the UK, with the scare last year over "phyto-oestrogens" and phthalates in powdered baby milk. The feedback loop between scientific result and political movement is less tight here; the disaster of BSE and the rising asthma rates linked to car exhausts indicate that our legislators are not prone to jumping to action at the first sign of a scientific paper.

Of course, the issue of environmental oestrogens remains an important one. And other studies have not been overturned, such as the one that indicated that a pesticide spill in a Florida lake shrank the penises of male alligators and increased their oestrogen levels near to that of female alligators; or that from a Columbia River Basin, which found that young otters exposed to environmental pesticides had penises and testicles half the normal size.

But it isn't all one-way traffic. Some American and European scientists doubt that the small amounts of pesticides in the environment can really cause such problems. A new breast cancer study contradicted an earlier finding that linked the disease to higher levels of environmental chemicals. There is even doubt about those studies that suggest sperm levels are falling: some are based on samples from Swedish prisoners - not quite a standard control group.

The trouble is also that legislators hate to repeal legislation once they have it, and lobbyists aren't keen to encourage them. There is no sign that the EPA will try to have the new legislation repealed in the US. It is also a simple fact of modern media life that retractions do not garner much space (unless made by someone whom the media detests, in which case they are hailed as a triumph for the media). So, many people will still think that combinations of pesticides have the effect of huge oestrogen doses; and now there is even some law built on that assumption.

The moral? If our society is built on science, we need to learn how to dismantle it when the foundations are faulty. Perhaps by making more noise when things turn out wrong.