Framingham is the paradigmatic study describing the relationship of the major risk factors to CHD incidence, and the Seven Countries Study showed that they are important in a range of populations, finding recently confirmed by the Asia Pacific Cohorts, the Prospective Studies Collaboration and the INTERHEART study.
However not many studies have addressed the question of what drives these trend as raised by the Bethesda conference. Because this question cannot be practically addressed with randomized controlled trials, it has been mainly studied using observational designs and more recently, modelling studies.
I will describe some of the major studies that provided evidence to explain what drives CHD mortality trends. One of them is the MONICA, a large international observational study and two modelling studies, the US CHD Policy model and the IMPACT model. Finally, I will discuss Finland, a country where both observational and modelling evidence is available.
The MONICA Study20:
Together with the Seven Countries Study and the Framingham Study, the MONICA Project (Multinational MONItoring of trends and determinants in CArdiovascular disease) helped to establish our current understanding of the epidemiology and control of cardiovascular disease. MONICA objectives were first to observe the trends in CVD mortality and morbidity and second to evaluate the extent to which these trends were related to changes in known risk factors, including daily living habits, health care, and major socioeconomic features. The study was conducted in 38 populations in 21 different countries starting in 1980 and finalizing data collection by the late 1990s. They assessed the contribution of risk factors to CHD incidence rates by looking at the association between 10 year trends in major risk factors (smoking, blood pressure and blood cholesterol) and 10 years trends in incidence (fatal and non fatal events). Careful biochemical measurements and strict event ascertainment and death certification procedures were key features of this exemplar study. They monitored almost 13 million people over 10 years, and more than 300,000 men and women were sampled and examined for risk factors. During the 10 year period they registered 166,000 myocardial infarctions.
In the MONICA populations, CHD mortality rates fell 4% per year. About two thirds of the observed fall could be attributed to the fall in event rates, while one third could be attributed to a fall in case fatality20. However, they were unable to precisely quantify how much of these falls could be attributed to changes in risk factors18 or to evidence based treatments.19 The effect of individual risk factors on risk was lower than the observed in the cohort studies. In part, this can be attributed that to some extent, the analysis in the MONICA study are at the ecological level, and that they were not corrected for dilution regression bias. When the association in cohort studies is explored without taking into account this bias, the results are similar.148,197
US CHD Policy Model198
The US CHD Policy model is a state-transition, cell based model developed in the1980s.198 It was initially used to examine trends in CHD mortality199,200 and expected gains in life expectancy from risk factor modifications.201 This model was also used to evaluate the cost-effectiveness of specific medical interventions for primary and secondary prevention of CHD202-204, salt reduction policies205 and health promotion activities.206. The model showed that in the US population and for the period 1980-1990 risk factor changes contributed 50% to the mortality decline while treatments contributed 43%.
IMPACT
IMPACT is a cell-based model originally developed by Capewell and colleagues in 1996.5 Using a MS EXCEL spreadsheet, this model combines data from many sources on patient numbers, treatment uptake, treatment effectiveness, risk factor trends and consequent mortality effects. The deaths prevented or postponed (DPPs) over a specified period are then estimated. The model can be used to estimate the proportion of change in mortality attributable to specific treatments or risk factor changes. It can also estimate the future consequences of altering treatment strategies and changing population risk. The model also estimates life years gained and cost-effectiveness for specific interventions.
IMPACT, an ongoing project, has been used to explore the contributions of risk factors and treatments in over 10 countries. In most of the studied countries (New Zealand, Scotland, England & Wales, Sweden, Italy, Spain, Iceland, USA and Canada) CHD mortality rates has been declining. The IMPACT model consistently found that about 40 to 72% of the fall in deaths could be attributed to risk factors changes and 23 to 55% to treatments.144 An interesting observation from IMPACT modelling related to the City of Beijing, were CHD trends were increasing, essentially driven by a
huge increase in cholesterol levels. This might be related to the rapid adoption of a “Westernized” diet, rich in saturated fats.186
Particularly interesting are the recent findings in central European populations, were rapid declines in mortality has been observed after decades of increasing rates and linked to profound socio-economic changes resulting in substantial modification of the exposure to CHD risk factors. I will describe in chapter 6 an analysis using the IMPACT model of the trend determinants for one of the most interesting countries in the region, Poland, which experienced a dramatic decline in CHD mortality since the 90s. I will also use the IMPACT model to study recent English trends by socioeconomic status (chapter 7) Methodological details of the model will be presented in the relevant sections and in appendices 2 and 4.
Finland: Observational and modelling studies
Finland experienced a marked decline in CHD mortality during the 20th century, associated with the implementation of nationwide, population level policies. Trends in mortality and in the major risk factors were then closely monitored. Thus detailed analysis of the period of sustained declined between 1972-2006 is available.207
Serum cholesterol declined significantly in both men and women over that period. Blood pressure declined up to 2002, but levelled afterward. Smoking followed a more complex pattern, declining in men but increasing in women until 2002, and levelling off since then. BMI increase in men throughout the period and in women started to increase in 1982.
Collectively, the changes in risk factors explained about 60% of the 80% observed decline in CHD mortality.
The Finnish IMPACT model148 found similar contributions of risk factors to that observed by Vartianen et al. In this modelling exercise, risk factors explained 48 to72% of the observed fall in coronary heart disease mortality between 1982-1997, while medical treatments explained 23%, consistent with the observational data.148
4.5
CONCLUSIONS
Although still there is not a definitive answer to the question posed at the Bethesda conference, insights coming from observational and modelling studies suggest that both risk factors and medical treatments contributed substantially to the observed trends in CHD mortality.
The fact that the current epidemics across the globe are at different stages and that rates are increasing in many countries makes current study of these trends pertinent, especially given the alarming increases in obesity and diabetes.
However, the current emphasis on monitoring trends using age-adjusted rates might present an incomplete picture of the state of the epidemic in individual countries, because this might conceal important differences in age-specific patterns that might mark the beginning of a new phase in the CHD epidemic.
The decades long decline in age adjusted trends in the Western world convey the idea that trends are invariably longstanding, almost set in stone. In the next chapter, I will challenge this cosy concept, by studying recent changes in the pattern of the CHD epidemic in a variety of Western countries.