The causal role of blood lipids in the aetiology of coronary heart disease – an epidemiologist's perspective

Abstract

Background. Cardiovascular (CVD) mortality decreased abruptly in Norway during WW II but increased faster than in other European countries from the 1950's. Mean life expectancy for middle-aged men declined during 1960's. Objectives. To give a short overview of CVD epidemiology, especially changes in coronary heart disease (CHD) mortality, total serum cholesterol and diet during the latter half of the 20th century. Methods. Review of mortality statistics, reports from the previous National Health Screening Service and papers concerning risk factor and dietary changes. Data on CHD morbidity are not available. Discussion and conclusions. CHD mortality reached its peak during 1966–1970. It declined during the next 30 years bringing Norwegian mortality rates to levels comparable to some Mediterranean countries. The main causes for the decline during the first twenty years of this time period are reduced total cholesterol levels from the 1970's to the 1990's, as well as a declined prevalence of smoking in the male population. Improved medical and interventional treatment are likely to explain the changes occurring during the last decade of the 20th century.

• Cardiovascular mortality
• cholesterol
• diet

Demographic evolution and coronary heart disease

“Female life expectancy in the record-holding country has risen for 160 years at a steady pace of almost 3 months per year…life expectancy has also risen for men albeit more slowly.” 1. The authors have accessed mortality data from a number of developed countries and analysed the trends of life expectancy from 1840 to 2000. Consider their expression; “a steady state”. They underscore that the four-decade increase in life expectancy over 16 decades is extraordinarily linear, and this is the main focus of their paper. There are however a couple of deviations from the linearity which have had profound impact on global health. One was the rapid decline in life expectancy in 1919–1920, and the second is the attenuation of the increase in life expectancy which occurred during the 1950–1970's. It is seen from Figure 1 1 that the life expectancy for men dropped like a stone in 1919 whereas the slope flattened out three decades later. This last deviation was the epidemic of coronary heart disease (CHD). It struck first in the US, and thereafter in northern and mid-Europe, Australia and New Zealand. The southern part of Europe was not hit to the same extent at that time. Why did this occur? Why did the frequency of this disorder suddenly explode as it did in developed countries – with a few exceptions – during such a short time period? The increase was not a function of increasing age, as seen from the decline in life expectancy. The coronary mortality rates in Norway came to their highest levels in 1966–1970. The CHD mortality rates had doubled for men since the 1951–1955, and were 50% higher for women. These increases in coronary heart disease mortality had as already mentioned profound impact on the life expectance of middle-aged men. A corresponding decline was not seen for women, mainly due to the lower absolute number of deaths from this cause in middle-aged women. It was not until the break-up of the communist regimes in Eastern Europe that a similar decline in life expectancy was observed during a relatively short time period 2. Rapid changes as those occurring first during WWII and the 1950's could not be ascribed genetic causes even if that causal concept was well ingrained in the minds of the medical establishment, in particular in Norway when it came to coronary heart disease, and for a reason 3, 4.

Figure 1.  Level of male and female life expectancy at birth in selected countries 1840 to 2000. From Oeppen J, Vaupel JW. Demography. Broken limits to life expectancy. Science 2002;296:1029–1031.

From genetics to nutrition and smoking

In 1936 the Norwegian physician Carl Müller (1886–1983) identified 17 families with xantomas, hypercholesterolaemia and cardiovascular disease in Oslo 4. He described this disorder as a monogenic autosomal dominant disease. His first conclusions were that this was a common disorder – mainly because he had been able to identify such a large number of families. The concept of cardiovascular diseases being the effect of mainly inherited traits was strong until well after the Second World War, even if urban life style and stress would also be considered possible causes for the observed increase in cardiovascular mortality during the 1930's. The emphasis on familial hyper-cholesterolaemia is reflected in the teaching of Norwegian medical students, who are drilled to look for xanthomas, high cholesterol levels and premature coronary heart disease, even if the likelihood for ever finding a case is remote, only 1 in 500 carries one of the deleterious alleles.

The World War II Experiment

The causal emphasis shifted from inheritance to external factors a few years after the end of the Second World War. In 1950 Axel Strøm, then professor of social medicine in Oslo reported together with Jensen that a sharp fall in mortality of cardiovascular diseases had occurred in Norway during the war 3. The war had of course led to major changes in lifestyle and living conditions, of which the most remarkable were declines in fat consumption and cigarette smoking as well as increased physical activity. The decline in coronary mortality during the Second World War has been called ”The Norwegian World War II Experiment”, but these changes have also been observed in other war torn countries such as the Netherlands during WWII and Berlin in WWI where especially the incidence of diabetes declined. Parallel to the changes in coronary mortality in Norway was also a decline of post-operative thrombo-embolic postoperative complications, most likely due to a slimmer, more physically active and less smoking population, all pointing towards a more efficient thrombolysis.

The first epidemiological studies

The rapid changes during WWII and the later increasing mortality rates, initiated the systematic studies on the possible effect of total serum cholesterol, as well as blood pressure as risk factors for coronary mortality. The principal investigator was Knut Westlund who started the studies in the 1950's and had the first results published in the 1960's 5, 6. These longitudinal studies demonstrated that total cholesterol was indeed a major determinant for coronary heart disease deaths, but also that differing levels could explain differences in mortality rates within the country.

The Norwegian government took his findings very seriously and 1960 a committee addressing the possible association between dietary fat intake and cardiovascular disorders was organised under the auspices of the government 7. The committee concluded in 1963 that the dramatic increase in disease occurrence with 5 times higher admission rates for coronary disease in men aged 40 – 59 years which took place in Oslo from 1947 to 1960 was real. It was also shown that the increase seen in women aged 60 – 69 years corresponded to that observed in men ten years younger. The committee concluded that the causes for this increase were likely to be age, blood pressure, cigarette smoking, and stress, but their main emphasis was on serum cholesterol and its association to dietary fat intake. The empirical evidence for these conclusions was national statistics and above mentioned Norwegian epidemiological research, but two reports from abroad, one from the “The Central Committee for medical and community program of the American Heart Association” (AHA) of January 1961, and another from World Health Organisation (WHO) from 1962 were also considered by the committee. The WHO report was coloured by negativism and insecurity whereas the AHA report came up with dietary advice aimed at particularly exposed subjects. The Norwegian committee however goes further than both these reports and concludes with a set of recommendations aimed at the total population. One of the recommendations was that the average energy intake from dietary fat should not exceed 30%, which corresponded to the levels in countries with low coronary mortality rates. It was also recommended that polyunsaturated fat should constitute one third of the total fat intake. The committee emphasised the need for monitoring nutrition, cardiovascular morbidity and mortality at a population level. The latter was not taken as seriously as the dietary recommendations. The dietary recommendations were incorporated in later White Papers (1975 and 1981) explicitly formulating a nutritional policy with health aspects as the main objectives.

Mechanisms and causes of cardiovascular diseases

Three different theories on the causes of atherosclerosis and cardiovascular disease dominated by the end of the 19th century; the incrustation theory developed by von Rokitansky putting most weight upon thrombosis and role of platelets, the irritation theory from von Virchow with emphasis on chronic inflammatory processes, and the lipid infiltration theory based upon Anitschkov's and Chalatov's feeding experiments on rabbits at the beginning of the 20th century 8, 9.

The lipid theory was strong in Norway, especially because of the previously quoted epidemiological studies. Westlund showed for instance that the 10 year risk of angina or myocardial infarction for men aged 40 – 49 increased from 2.3 to 37.4% when serum cholesterol increased from 4.4 to 8.9 mmol/l 10. He also showed that cholesterol was less important than blood pressure for cerebro-vascular events, and that increasing body weight was associated with increased diabetes incidence.

The nationwide population based surveys and randomized controlled trials

In 1974 the Cardiovascular Disease Studies, beginning with a survey in Finnmark, the northernmost county in Norway started the series of systematic and standardized studies were under the guidance of the National Health Screening Service 11. The aims of these studies were three-fold: the identification of high risk subjects for further individual treatment, to influence the life-style in the total population, and to monitor and initiate research concerning coronary risk factors at a population level. During 1974 to 2001 more than 400 000 individuals were examined, forming the basis for one of the largest amount of standardized collected information on cardiovascular risk factors in the world 12

A landmark randomized intervention study was initiated during the early 1970's in Oslo:”Oslo-Diet and Stop Smoking Study” in order provide a basis for individual dietary advice and treatment firmly bedded on scientific data 13.

The starting point for the “Oslo-Diet and Stop Smoking Study” was the results of a randomized study on more than 400 myocardial infarction patients undertaken by Paul Leren with Knut Westlund as advisor 14. The results indicated that dietary changes with fat reduction and increased intake of polyunsaturated fatty acids protected against secondary heart attacks. The study sample for the “Oslo-Diet and Stop Smoking Study” was drawn among participants among in the Oslo Study of 1972 – 1973 where all men aged 40 – 49 years and a 7% sample of men aged 20 – 39 years had been invited to screening for coronary risk factors comprising: family history, physical activity at work and in leisure time, stress, smoking habits, chest pain questionnaire, total cholesterol, triglycerides, blood glucose and blood pressure. Thirty years ago, only 15% had what was then considered optimal cholesterol levels (<200 mg/dl or 4.8 mmol/l), 56% were daily smokers. And maybe more important, there were close correlations between total cholesterol and blood pressure, as well blood pressure, body weight and glucose.

The actively treated group was advised to reduce the intake of saturated fat, slightly increase polyunsaturated fatty acids as well as fish and vegetables, and when deemed appropriate, reduce weight. The smokers were in addition advised to stop smoking. No advice was given to the control group. Eighty per cent of the participants were smokers, and the total cholesterol ranged from 6.9 to 9.0 mmol/l. They were all followed for 5 years. Total cholesterol declined by almost 1 mmol (13%), compared to the control group which declined 3%, resulting in a net reduction of 10%, as well as 2.5 kg weight reduction whereas quitting smoking only differed by 6% (23 versus 17%). The 16 year follow-up showed a 50% lower mortality in the intervention group even if the intervention as such only lasted for five years, and the control group was informed of the results at the end of the study 15. The first results convinced the Norwegian health authorities of the benefit of dietary changes, and underscored the need for more systematic dietary changes on national level.

The declining coronary heart disease mortality rates and changes in diet and smoking

From the 1970's the CHD mortality rates started to decline, and have been reduced by more than half for men aged 40 – 69 years during the last 30 years 16. Pedersen et al. 17 calculated that during the period 1996 – 2000 and for all age groups, 30 903 fewer deaths occurred than expected, had the mortality remained the same as during 1971 – 1975. The declining death rates are also reflected in the subsequent increase in life expectancy of middle-aged men. But it was not until the 1990's that the life expectancy exceeded that of 1951 – 1955. Pedersen et al. have estimated the contributions from dietary changes as well as the declining smoking prevalence to explain the declining coronary mortality rates. They observed that the proportion of total fat in the diet went down from 41 to 34% of energy and that the proportion of unsaturated to saturated plus trans fatty acids had increased, mainly due to changes in milk fat and margarine consumption. These changes together with the switch from drinking unfiltered coffee (boiled coffee) to filtered coffee would result in reductions of total cholesterol in the order of 0.8 mmol/l. This corresponds closely to the observed 0.5 to 1 mmol/l decline in the total cholesterol observed in the population surveys. Other factors that are likely to have contributed to the decline, at least until 1990 are reduced smoking (in men), a small reduction in blood pressure, increased consumption of fruit, vegetables, cod liver-and fish oil. The effect of improved treatment is more likely to have contributed substantially after 1990 when statins as well as improved revascularisation had been introduced.

The almost futuristic approach taken by the governmental committee established as far back as 1960 gave the Norwegians an early start when it came to preventive efforts.

Are blood lipids causal factors or only innocent bystanders?

Above is briefly described how a chronic and fatal disorder reached epidemic proportions and returned to lower levels after about two decades. The factors which were associated with its increase and decline in Norway are the same as those observed in other countries and populations where this evolution has taken place; blood lipids, smoking, body weight and diabetes. Are these factors causally important, or only innocent bystanders? They are consistently associated with disease frequency across populations and at a graded level. They can be influenced by external measures, and these measures can if properly applied, affect subsequent risk of disease. This latter fact makes them more than bystanders and has led us to the risk factor concept. A risk factor can be defined as any variable being statistically associated with a disease end-point. The term major risk factor however comprises traits that are common in the population, are amenable to prevention and control, and have a graded strong and independent relationship to disease incidence or other risk factor. Six major risk factors are established to-day; adverse dietary pattern, dietary patterns affecting total cholesterol, obesity, high blood glucose or diabetes, blood pressure and smoking 18. It is seen from this list that these factors are not exclusive entities. They are both overlapping and interrelated whereby any analysis of their impact on disease frequency and effect of intervention may be difficult when it comes to make a distinction of each ones contribution. Two of the factors are obviously affecting the others and stand out as the most important explanatory factors as to why the CHD epidemic reached such dramatic levels; the adverse dietary pattern, and the dietary patterns affecting total cholesterol. Causality is more than only explaining associations; it also provides us with an instrument for change. Our problem with coronary heart disease to-day is to lift the complexity of atherogenesis, inflammation, thrombosis and fibrinolysis up to the individual and population level. This can only be understood by studying populations and the corresponding individuals. Studies at population level do indicate that the explanation as to why suffering and dying from CHD is less probable in developed countries to-day involves a decline in total cholesterol, declining smoking prevalence, declining blood pressure- and in this order 19. Re-vascularisation on the other hand contributes only to a minor degree to the declining mortality of CHD. Opponents to the diet-CHD theory will have to explain the rise of the CHD epidemic as well as the decline. One possible avenue would be to argue that the susceptibility towards particular dietary patterns increased for certain birth cohorts due to dramatic external factors- as has actually been suggested 20, or due to a particular exposure e.g. life events or stress. This however does not exclude the role of diet; it only points at the need for discussing to what extent a single factor is necessary or sufficient to trigger the disease event given the putative combination of factors causing a disease.