Doctors working on the world's largest study of heart disease - involving 150,000 people in 21 countries - told a cardiology conference in Vienna that the fall in heart disease among the populations studied did not exactly match the decline in the classic risk factors, namely smoking, raised blood pressure, high cholesterol and obesity. In some places where the risk factors declined steeply, the incidence of disease fell only slightly; in others, where the change in risky behaviour was small, the fall was sharp.
So does this mean that we can all ignore the health educators' exhortations to give up smoking and eat less fatty foods? Absolutely not, says Professor Hugh Tunstall-Pedoe, director of the cardiovascular epidemiology unit at Ninewells Hospital, Dundee, and one of the study's chief organisers.
"None of our results suggest that the campaigns against fatty food and smoking were wrong, just that there is more to the story than that. If we had shown a perfect correlation between the classic risk factors and heart disease, that would have made the subject of heart disease rather boring. Instead, our results leave room for the matter to be more complicated," he says.
The study, called the WHO Monica Project (from MONItoring CArdiovascular disease), which covers countries as diverse as China, Russia, Canada and Australia, showed that blood pressure and smoking were coming down in most of the populations studied, cholesterol levels were not changing much and people were getting fatter. But the reductions in blood pressure and smoking did not seem to match the fall in heart attacks.
"There were large differences in the rate of decline in populations with similar trends in risk factors," adds Prof Tunstall-Pedoe. "For example, the reduction in risk factors was similar in Glasgow and north Karelia in Finland, yet the fall in deaths from heart disease was much higher in Karelia."
The professor says that scientists know from hundreds of other studies that the classic risk factors are important. He and his team in Dundee recently published, in the British Medical Journal, a study comparing the importance of the classic risk factors with 20 other factors that had been put forward in recent years.
"The classic ones came out on top," he says, "with the exception of one or two minor differences. A diet rich in potassium seemed to have a protective effect against cardiovascular disease and, for women, having a `type A' - a driven, ambitious - personality, seemed to confer some benefit.
So what is distorting the Monica results? Why isn't the match better? There are four reasons, according to Prof Tunstall-Pedoe. "The first problem is one of measurement. In a study with 38 centres in 21 countries there is a huge problem in standardising measurements. Also, personnel change, so as soon as one team has learnt what to do, its members move on.
All results were sent to a data centre in Helsinki, and procedures and results were scrutinised by quality control centres. Although serious failure led to exclusion from the study, if we had excluded every centre with any problems at all, we would have ended up with too few centres.
"Secondly, because most trends - in smoking, blood pressure and heart disease - were going down, we did not have as great a heterogeneity of trends as we would have liked. The changes we were measuring were not that large compared to the possible errors in measurement. For research purposes - though not for other reasons - we would have liked a better spread of trends, with some going up."
The third problem was one of time lag. If people reduce their risk of heart disease by changing their lifestyle, there is a time lag before the effects are seen.
This seems to vary according to different populations and their characteristics (some populations have naturally low levels of cholesterol, for example), which makes comparisons difficult. The reduction in risk factors among the Scottish population is now paying off in reduced rates of disease, but it has taken longer in that country than in some others.
Finally, Prof Tunstall-Pedoe admits that there may be other determinants of heart disease, apart from the classic risk factors. Some of these, such as diet, were known about when the study was set up 20 years ago, but were too difficult to measure. Others have emerged more recently.
"The importance of eating a diet rich in fruit and vegetables is very strong from the epidemiological standpoint," he says. Such a diet, full of anti-oxidant vitamins, seems to be useful in preventing disease, but no one knows exactly what dose of which vitamin is playing the crucial part.
Another possible cause of heart disease is the presence of low-level chronic infection, according to Professor Brian Pentecost, the medical director of the British Heart Foundation, who welcomed the study. But this factor has emerged in recent years, and was not suspected when the Monica study was set up.
"There have been a number of pilot studies of antibiotic treatment being given to people who have had heart attacks and who have evidence of infection, which have shown some benefit in preventing further attacks," says Prof Pentecost said. Various infections have been implicated, including chronic periodontal disease and chlamydia, but it is not yet known which are important.
If chronic low-grade infection were found to be a culprit in heart disease, it would help to explain why the disease is more prevalent among socially deprived populations than among the wealthy, but the fact that smoking is also more common among the poor confuses the picture.
The Monica study also considered the efficacy of the different treatments, but was unable to say which of them was best. It could only say that "those populations which showed the most rapid increase in new treatments tended to be those in which heart attack survival and mortality were improving most".
Another factor may have also been in play. "Those countries that adopted the most modern treatments were also the wealthiest, so it may have been the wealth and material quality of life that played as great a part in reducing death as the treatments themselves," says Prof Pentecost.
The study showed that heart disease is still not completely understood. But that does not mean that doctors are merely clinging to the wreckage when they reiterate the health messages that they have sent out for the last 20 years. It means that there may be additional factors that need to be taken into account," he adds.
Or, as Professor Tunstall-Pedoe puts it: "If you get eaten by a crocodile when you are expecting lions and tigers, it does not mean that big cats have rubber teeth."
People love the idea of a study that overturns all the health rules of the last 20 years, partly because they want permission to indulge their habits but also because they like the idea that all those clever doctors got it wrong. But they cannot seize on this one as an answer to their prayers.