Response: Coronary heart disease prevention: act now, research at leisure

Robert Beaglehole and Paul Magnus

We welcome the response elicited by our article. Our most important point is that if we could dramatically reduce population levels of the established risk factors we would drastically cut coronary heart disease (CHD) rates in developed countries and probably curb the emerging epidemic in developing countries. We based this conclusion on a persuasive body of epidemiological evidence that in turn builds upon the evidence from non-epidemiological research.

We used the population attributable fraction approach to conclude that at least 75% of CHD could be explained by the major risk factors and that the contribution of these factors has often been grossly downplayed in the past. This restricted our focus mainly to studies of middle-aged and older people. But it did allow us to quantify the contribution of the established risk factors in a way that other approaches do not. We are reassured that the commentators either accepted this central argument or at least failed to mount any plausible case against it.

We then asked whether the decades-long search for new or emerging risk factors offered much extra scope for prevention. We found little encouraging evidence and therefore stressed the need to apply more fully what is already known. We argued that much has been and could be achieved in practice with present knowledge, and that the priority for further research should be given to increasing the application of existing knowledge and for more understanding of the upstream determinants of the major risk factors. We also speculated in passing on the priorities of many epidemiologists working in this area.

Among those invited to comment on our article, Greenland et al. and Law et al. present further evidence that complements our main arguments. They focus on atherosclerosis and its development throughout life and advocate a related life course approach to prevention. While agreeing with the logic of the life course approach to prevention, we reiterate that in the short to medium term the greatest prevention gains will come from concentrating on reducing the risk of CHD in the adult population. The life course approach will eventually bring great benefits but only after decades of effective prevention programmes.

We are pleased that Marmot, in his thoughtful comments, reinforces the need for epidemiologists to become more actively involved with the development, implementation, and evaluation of population-wide approaches to the prevention of CHD. Without strong epidemiological input and leadership, cardiologists will continue to dominate the prevention debate and individual approaches will remain the priority.

However, Marmot risks obscuring what we mean when we say that the established risk factors explain 75% of the population occurrence of CHD. It simply means that in their absence there would be 75% less CHD—or even less. In fact, Stamler et al.’s finding from the massive US cohorts1 is that low-risk subcohorts defined in terms of the major risk factors had CHD mortality rates below epidemic levels. It is reasonable that the established risk factors should largely explain the difference between epidemic and non-epidemic levels and not be able to equally explain much smaller variations that occur within an epidemic, such as those between social groups in the Whitehall studies.

Marmot applies our point about a ‘myth of the 50%’ to his interesting work on social gradients, in a way that may appear to suggest that the myth does not exist. But this is an inappropriate test of the myth. We did not aim to explain the social gradient in the first place—it is a different issue to the main one we are addressing. Elsewhere Marmot agrees that the underlying causes of the CHD epidemic are social, economic, and cultural. But he then disagrees that these causes operate ‘principally’ through the established risk factors. This view seems much at odds with Marmot’s own finding that CHD mortality in the first Whitehall cohort would have been two-thirds lower if everyone were at low risk as defined in terms of the three main risk factors (Marmot M personal communication, 24 October 2000).

Since social differences can be very important, however, we do not negate the value of continuing the search for the reasons for individual and group differences in CHD rates within populations. We do, however, suggest that this is a much lower priority than acting on what we already know to reduce the overall epidemic. This also applies to directing more effort to reducing the known explanations of social group differences—as identified by Whitehall II.

Nieto’s comments sometimes border on the sarcastic and he seriously misrepresents us in several instances. Among other things, he found us ‘presumptuous’ and ‘blatantly unscientific’. He attributes to us things that we patently did not write and views we do not have. For example, he says we asserted that there is ‘nothing new to be learned’ about CHD risk factors. We did not say that, and of course we do not think it. We are also said to see research on new factors as ‘irrelevant’ or ‘worthless’ or in some cases ‘utterly superfluous’. Again, we used no such words and this is a cavalier if not wilful misrepresentation of our case. The issue is not a matter of dismissing all such research, but of priorities in the face of a great epidemic. Just how relevant and how worthy and how promising is much of this work when compared with research and action to reduce the known risk factors?

Nieto further makes the remarkable implication that we are ‘ignoring everything’ outside epidemiological evidence. In fact we spoke of satisfying ‘evidence of causality with the totality of evidence being strong and consistent’. Elsewhere we talked of biological plausibility and a full range of laboratory and clinical studies, along with epidemiology.

Responding to our claim that at least 75% of CHD cases are explained by the established risk factors, Nieto shifts the goal posts. He notes that the studies we cited can only inform us of the predictors and correlates of the ‘late phases’ of the disease natural history; there is ‘absolutely no proof’ that they could equally explain the actual initiation of atherosclerosis, which he says is the true target of primary prevention. This may raise the bar of proof beyond reach, given the methodological challenges. Yet even here he would do well to read the evidence from Greenland et al. And in any case, wouldn’t the capacity to greatly reduce CHD among adults amount to a good start?

In summary, we restate our position that there is an urgent need to apply what we already know to the prevention and control of the CHD epidemics. Selected further research is important but should not be an excuse for a limited public health response. A co-ordinated and comprehensive response has the potential to significantly improve the health of all populations in only a few years.

References

1 Stamler J, Stamler R, Neaton JD et al. Low risk factor profile and long-term cardiovascular and noncardiovascular mortality and life expectancy: findings for 5 large cohorts of young adult and middle-aged men and women. JAMA 1999;282:2012–18.[Abstract/Free Full Text]