METHUSELAH MICE: Do the data from laboratory animals reflect results in humans?

Across the nation, insomniacs, the sick and the simply curious are gobbling up melatonin. The hormone is being touted in the popular press as a natural way to get a better night's sleep, to improve one's sex life, to live longer and to fight the ravages of AIDS, Alzheimer's disease and cancer, among other afflictions.

Boosted by cover stories in Newsweek and by a rash of uncritical books, melatonin seems to have taken over from beta carotene as pop medicine's Next Big Thing. (Beta carotene, another natural substance, recently disappointed enthusiasts when a clinical trial found that it not only fails to prevents cancer in smokers, but may actually promote it.)

A backlash against melatonin hype is now under way. Critics warn that the doses of melatonin commonly sold in health food stores--which raise levels of the substance in the blood 30-fold higher than their normal peak--could be dangerous for some people. And many researchers and fad-watchers point out that the more startling claims being made for the substance are unsupported by studies on patients. They are, rather, starry-eyed extrapolations from experiments conducted on rats and mice. These animals differ from people in many ways, especially as regards sleep, the one aspect of human physiology that melatonin clearly affects.

Richard Wurtman, a researcher at the Massachusetts Institute of Technology, has spent years studying melatonin and its biological effects. He has described some of that work in a Scientific American article ("Carbohydrates and Depression" by Richard J. Wurtman and Judith J. Wurtman in SA, January 1989). Wurtman says there is "no controversy" that the substance, in doses of a fraction of a milligram, can induce sleep and shift the sleep cycle. Summaries of his findings appear in an M.I.T. press release.

But Wurtman, the named inventor on an M.I.T. patent covering the use of melatonin for controlling sleep, says there is "no evidence" that it has any effect on human life expectancy, and "only marginal" evidence that it promotes longevity in mice. Wurtman further maintains that health-food store doses of melatonin might diminish sex drive. He also denounces as "wicked" the suggestion that people with cancer or AIDS should take the hormone: Wurtman thinks melatonin is as likely to worsen those conditions as to ease them.

Wurtman is not the only researcher criticizing melatonin mania. In an article published in Cell, Steven M. Reppert and David R. Weaver of Harvard Medical School recently described as "seriously flawed" an experiment by Walter Pierpaoli, William Regelson and others that kindled hopes melatonin might extend life.

The investigators prolonged the life of elderly mice by giving them transplants of tissue from the pineal gland (an organ that responds to melatonin in the bloodstream) taken from younger mice. Pierpaoli and Regelson hypothesized that the transplanted tissue was more responsive to melatonin and somehow revitalized the old mice. But according to Reppert and Weaver, the mice used in the experiments have a genetic defect that means they cannot make melatonin; attributing the effect to the hormone is therefore "absurd.

Regelson and Pierpaoli's book The Melatonin Miracle is one of the most unrestrained of the current crop. Regelson asserts that he and Pierpauli wrote their book because pharmaceutical companies are likely to drag their heels on melatonin research, owing to the difficulty of patenting a substance found in nature, whereas "I am 70 years old and I have a time constraint."

But Fred W. Turek of Northwestern University, writing in Nature, points out that "not one study" in humans supports Pierpaoli and Regelson's claim that, for example, melatonin helps to prevent heart attacks. Victor Herbert and Ruth Cava of the American Council on Science and Health caution that children, women who are nursing or who may become pregnant, and people who have immune-system disorders should avoid taking the hormone because of uncertainties about its effects.