Coincidence or a straw in the wind? In March, two of the world's leading research journals carried reports that make gloomy reading for anyone with an interest in public health. One was a systematic review of published research, the other a journalistic essay. But both pointed to quietly accumulating weaknesses in one of the central planks of health advice: the amount of fat we eat.

The systematic review, from an authoritative group of nutritionists and statisticians, can be found in the British Medical Journal (322: 757). It looks at the outcome of 27 attempts to reduce cardiovascular disease and death by persuading people to lower or modify their fat or cholesterol intake. It concludes that only those trials lasting at least two years have shown any significant reduction in disease of the heart and blood vessels. But that modest encouragement is dampened by its other findings: that the trials have had 'little effect on total mortality'; and, more generally, that despite decades of effort there is still only 'limited and inconclusive' evidence of the benefits of modifying fat consumption.

Systematic reviews are not, of course, infallible. But this possible route of escape becomes less sustainable in the light of a detailed and revealing article in Science (291, 2536). Criticism of the 30-year-old diet-heart hypothesis is nothing new; what the Science piece underlines is the full extent of its long history of ignoring onslaughts, and surviving. Even the now falling rate of death from heart disease does not provide unequivocal evidence of the benefits of dietary change because these have been accompanied by improvements in treatment.

The virtue of the eat-less-fat message is its simplicity. To a public growing ever more sceptical of 'official' advice on health matters, any revision now of what has been preached for the best part of 30 years might be too much to swallow. Bloody, but thus far unbowed, the hypothesis soldiers on.