Every vitamin devotee everywhere will be outraged. Anyone who has ever pondered the life-enhancing claims on the seductive tubs of supplements lining the shelves of the pharmacists must be doing a double take. According to a group of scientists, some should come with a health warning: These supplements could shorten your life.

The five scientists in the Cochrane review team, based in Denmark, say that certain antioxidants, which many have speculated could hold the answer to longer life, actually endanger it. Beta-carotene, vitamin A and vitamin E, "given singly or combined with other antioxidant supplements significantly increase mortality," they conclude.

Vitamin C comes out of it better, although the team says there is no evidence it increases lifespan. "We lack evidence to refute a potential negative effect of vitamin C on survival," it says. Selenium, the fifth antioxidant in the analysis, appeared to reduce mortality, but the researchers cast doubt on the quality of the evidence.

The findings have been met with howls of protest from the vitamin and health-food lobby who make a living by promoting supplements. Their denials are likely to give comfort to the large numbers of people who want to think vitamins make them healthier. We are all urged to eat more fruit and vegetables, we know that vitamins are essential to our well-being, so how can this bunch of scientists possibly be right?

Make no mistake, the Cochrane team is a heavyweight bunch, enjoying great respect in scientific circles. It searched for every published paper it could find on the five antioxidants - and wrote to the manufacturers in case they were sitting on unpublished data (none replied). This turned up references to 815 trials. The vast majority were excluded from the review - only 67 made it.

"The paper's conclusions are drawn on less than 9 percent of available evidence," complained the Health Food Manufacturers' Association. "In no way can this review be considered comprehensive." But these charges miss the point: Only studies that had been adequately run were included.

The 67 were selected because they were randomized controlled trials - the so-called gold standard, in which half the participants are randomly given a drug (or in this case vitamin) and the other half are given a placebo. If the trial is good, it is double blind, so that nobody knows who has taken what. At the end of the trial, the blind is broken and statisticians can work out how well the drug or vitamin performed.