Article Text

Download PDFPDF

What explains declining coronary mortality? Lessons and warnings
  1. Simon Capewell,
  2. Martin O’Flaherty
  1. 1
    Division of Public Health, University of Liverpool, Liverpool, UK
  1. Professor S Capewell, Division of Public Health, University of Liverpool, Whelan Building, Quadrangle, Liverpool L69 3GB, UK; capewell{at}

Statistics from

One of the most important health achievements during the 20th century was the unravelling of the causes of vascular diseases (CVDs). Developing the risk factor paradigm helped us to understand the unparalleled rise and fall in CVD mortality rates seen in many Western countries. Despite the substantial decrease in the past 40 years, CVD is still an important burden for societies across the world. Annually in the UK, there are 200 000 deaths from CVD (60 000 premature with health and social costs exceeding 30 billion.1

Dealing with this burden will be challenging. The US healthcare system is already facing economic crisis. Today’s planners require renewed efforts to control this disease. We are very fortunate, because this group of diseases has been extensively studied. We have detailed knowledge of its causation and what effects can come from evidence-based therapeutic interventions.

Since the 1980s, coronary heart disease (CHD) mortality rates have halved in many developed countries and have fallen by 75% in countries like Finland.2 Comparable falls have also been seen in those Mediterranean countries with lower rates, such as France and Italy (fig 1).3 But are these achievements going to persist?

Figure 1 Changes in death rates from coronary heart disease, in men and women aged 35–74, between 1990 and 2000, selected countries. Source: Heartstats.3 Reproduced with permission of the copyright holder.

Modern cardiovascular treatments have played an important role. The explosion in evidence-based treatments since the 1980s has provided us with tools to modestly increase life expectancy in those affected by disease. But it is important to recognise that about half of all CVD related deaths occur outside hospitals, usually before any emergency care can be offered. Furthermore, this is often the first manifestation of CVD.4

Treatments will play an important role in any future CVD strategy. However, our increasingly precise understanding of CHD causation suggests that most of the CHD disease burden can be attributed to modifiable lifestyle and dietary risk factors.57

Furthermore, in the MONICA study of 15 populations experiencing a 27% decline in CHD mortality in a single decade, three-quarters of the decline was due to reduced incidence (reflecting risk factor improvements) and only one-quarter to improved case fatality (reflecting medical treatments).2 8

Moreover, lag times in populations are surprisingly short. Changes in risk factors are closely linked to mortality rates, as shown in a recent analysis of trends in Poland and in the seven-countries study.9 10 In the case of Finland and Sweden, widespread declines in major risk factors have been clearly linked to rapid, large declines in CHD mortality.11

But the MONICA analyses, despite high-quality data, referred to specific cities, which may not represent adequately regions or countries, and focused on 1985–95. Furthermore, many new treatments have been introduced since 1995, along with further, major reductions in smoking and dietary saturated fat intake and increases in obesity and diabetes. What is more important now: modern treatments, population risk factor reductions or both? Trying to understand reducing CVD mortality rates in more detail is a difficult task, since we cannot run a randomised controlled trial—insurmountable ethical and practical issues make randomisation of entire countries impossible.

But the question is still crucial, and the increasing future burden of CVD calls for renewed efforts.

Recent meta-analyses of high-quality cohorts help to validate the risk factor paradigm, and to inform increasingly precise measurement of effect sizes—a 50% CHD mortality decrease for 1 mmol/l reduction in cholesterol, or a 2% decrease for every 1 mm Hg reduction in systolic blood pressure.12 13

A more advanced, analytical approach is then possible: use computer-based policy models to try to summarise, synthesise and then integrate the vast amount of epidemiological and clinical data now available for vascular disease. Several policy models quantifying CHD have been developed.14 In 1982, the US CHD policy model suggested that for the USA changes in risk factors may explain about 50% of the observed decline in the USA between 1976 and 1982.15 More recently, IMPACT has been developed and used in a variety of populations (Scotland, England and Wales, Ireland, Finland, New Zealand and the USA).11 1619 These analyses suggested that we can attribute between 45% and 75% of the CHD mortality decline to changes in risk factors, and the remaining 25–55% to treatments. The most interesting outlier was Finland, where a more substantial decline in cholesterol plus other major risk factors accounted for almost 75% of the fall in mortality. This reflected a comprehensive national strategy which focused on smoking and diet.11

The Finnish experience has demonstrated that a concerted and comprehensive national effort pays huge dividends.20 Poland provides another interesting case study. This country experienced high CVD mortality with increasing rates from 1960 to 1991, reaching a peak of 777 per 106 in men and 257 per 106 in women. Poland then experienced a sharp mortality decline, 38% in men and 42% in women by 2002.9 These changes closely followed decreases in the per capita consumption of saturated fats and, equally important, increases in the consumption of polyunsaturated fats, fruits and vegetables. Smoking hardly changed.9 Similar trends were seen in Poland’s neighbours, Hungary and the Czech Republic.

The evolution of CHD mortality in other parts of the world suggests that the path of continuing decline is not assured. The Eastern European countries experienced an increase in overall and CHD mortality for much of the same period when the Western, more developed countries, experienced continuous decline. Between 1990 and 2000, the Russian federation experienced a 40% increase in CHD mortality rates in men and 30% in women. The explanation for this change is not straightforward, but increases in alcohol consumption, smoking and high-fat/low-vegetable diet probably all played a big role.21 China also recently experienced a surge in CHD mortality. In Beijing between 1984 and 1999, CHD mortality rates increased by 50% in men (and by 27% in women) aged 35–74 years. Most of the increase could be explained by a 1 mmol/l rise in mean cholesterol, which reflected progressive “westernisation” of the previously healthy traditional Chinese diet.22 23

Figure 2 Coronary heart disease deaths prevented or postponed by treatments and risk factor changes in the United States population, 1980–2000. AMI, acute myocardial infarction; BP, blood pressure; CABG, coronary artery bypass grafting; PCI, percutaneous coronary intervention.

A worrying levelling off of the previously falling CHD mortality rates in young adults in England and Wales, the USA, New Zealand and Australia has recently been reported even though the overall trends are still decreasing.17 2427 What does this mean? Are younger cohorts experiencing different levels of exposure to CVD determinants, perhaps associated with the known increasing unfavourable trends in risk factors in these age groups?26 28 If so, we are seeing particularly rapid reversals in the most susceptible young people.

Furthermore, population ageing represents a major challenge. Thus, even if age-specific mortality rates continue to decline, the absolute number of CVD deaths will increase. Predictions up to 2030 suggest that even with an annual decline in mortality rates of about 1%, the absolute number of deaths will increase, attributable solely to population ageing.29 Moreover ischaemic heart disease and cerebrovascular disease will remain the first and second cause of deaths ranking globally.29 A similar projection was recently reported in the Finnish population.30 The combined effect of ageing populations and the reported reversal of the trends in younger adults may therefore increase the absolute burden across the entire age range.


The various CHD policy models consistently suggest that in order to decrease the mass occurrence of the disease, moderate population changes in risk factors are crucial, particularly in the primary prevention setting.31 32

Moreover, if the population increases its proportion of people with a low-risk profile (no smoking, normal cholesterol, blood pressure and no diabetes), the rates of CHD events may fall to non-epidemic levels.6 It is not unreasonable to think that given the current state of our knowledge, the mass occurrence of CHD could be decreased to a point where it would be classified as a “rare disease”.6

Conversely, the “do nothing” option is not attractive. It risks an increase in the numbers of patients with CVD and progressive bankrupting of the National Health Service CVD services, as is currently threatening the US population. The potential contribution of cheap treatments is not minor, like aspirin, β blockers, ACE inhibitors and heart failure treatments. Furthermore, there is great space for improvement because there is still suboptimal uptake of effective evidence-based treatments. A recent IMPACT model analysis showed that in the UK in 2000, for example, barely 50% of patients received appropriate treatments. Increasing the uptake of current treatments to 80%, we could expect to prevent about 21 000 deaths.33 This is already happening in part, thanks to unprecedented funding for the CHD National Service Framework. However, the biggest potential gains will come from low-cost drugs for heart failure and secondary prevention. The most costly therapeutic interventions contribute comparatively little to this objective. In the UK, as in the USA, revascularisation procedures explain only about 5% of the observed fall in CHD mortality (fig 2). Even doubling the numbers of coronary artery bypass grafting procedures would make a disappointingly small difference.16 34 Prioritisation and political courage is therefore needed to support a transfer of finite National Health Service resources from these high-profile and expensive revascularisation strategies to more cost-effective ones.

The nature of CHD pathology and epidemiology suggests that population-wide changes in lifestyle and diet could reduce mortality with surprising rapidity, increasing its appeal to policy makers who need to solve problems within short time frames. If by say, 2010 or 2015, England and Wales achieved the slightly lower levels of risk factors already seen in many other countries like Sweden, Finland, the USA or New Zealand, we could expect about 50 000 fewer CHD deaths a year.31

Prioritising our strategies is thus paramount to help achieve continuing success in controlling CVD. But can we make a decision now? Yes. Do we need to know more before acting? No, because we already know more than enough about the causes of CVD disease. Research into new determinants may be interesting, but it is not essential for us to make sensible policy decisions now.35

Is any research still needed? It seems that in CVD, we are nearer to refining one of the field critical research questions from “What works” to “How can we make this happen”.

We know what the key targets are—reducing risk factors in the primary and secondary prevention settings and selected interventions for patients. However, we need to develop explicit methods to help us to prioritise our strategies in a rigorous way with the aim of changing the disease burden at the entire population level, not just in people at high risk or patients with disease. We also need better forecasts of the future disease burden and risk factors, to help us prepare the workforce and research agenda to tackle this problem well before it impacts. And we need to improve our modelling toolbox in order to take full advantage of the wealth of existing knowledge. We must find the optimal ways of summarising this knowledge in the most useful and accessible way for those making decisions at national, regional and local levels. We also need to know more about how and why people make diet and lifestyle decisions, and why many patients and doctors do not take full advantages of existing therapeutic tools.

Working to explain the recent decline in CVD mortality has taught us many important lessons and has provided unique opportunities to control a serious disease. However, the complex population dynamics show also that we can easily lose control. CVD mortality rates have flattened out in young men and women and may soon rise unless we get a better grip on the major risk factors throughout the whole population. Paradoxically, our successes also create new challenges; thus even with decreasing age-specific rates, we might expect a substantial impact from CVD in the coming decades attributable to population ageing. More important, adverse trends in risk factors and mortality rates among young adults in countries that experienced past success suggest that the healthier environment achieved at the end of the 20th century is changing. However, we are now better prepared to respond. We have the knowledge and skills, but do we have the political courage to advocate policy changes in diet, not just tobacco control?


View Abstract


  • None.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.