BLOG

2. Blood cholesterol has nothing to do with atherosclerosis

One of the most surprising facts about cholesterol is that there is no relationship between the blood cholesterol level and the degree of atherosclerosis in the vessels. If a high cholesterol really did promote atherosclerosis, then people with a high cholesterol should evidently be more atherosclerotic than people with a low. But it isn´t so.
The pathologist Dr. Kurt Landé and the biochemist Dr. Warren Sperry at the Department of Forensic Medicine of New York University were the first to study that question (25). The year was 1936. To their surprise, they found absolutely no correlation between the amount of cholesterol in the blood and the degree of atherosclerosis in the arteries of a large number of individuals who had died violently. In age group after age group their diagrams looked like the starry sky.

Drs. Landé and Sperry are never mentioned by the proponents of the diet-heart idea, or they misquote them and claim that they found a connection (26), or they ignore their results by arguing that cholesterol values in the dead are not identical with those in living people.

That problem was solved by Dr. J. C. Paterson from London, Canada and his team (27). For many years they followed about 800 war veterans. Over the years, Dr. Paterson and his coworkers regularly analyzed blood samples from these veterans. Because they restricted their study to veterans who had died between the ages of sixty and seventy, the scientists were informed about the cholesterol level over a large part of the time when atherosclerosis normally develops.

Dr. Paterson and his colleagues did not find any connection either between the degree of atherosclerosis and the blood cholesterol level; those who had had a low cholesterol were just as atherosclerotic when they died as those who had had a high cholesterol.

Similar studies have been performed in India (28), Poland (29), Guatemala (30), and in the USA (31), all with the same result: no correlation between the level of cholesterol in the blood stream and the amount of atherosclerosis in the vessels.

But a correlation has been found in a few studies. One of these was the famous study from Framingham, Massachusetts (32). The correlation found by the Framingham investigators was minimal, however. In statistical terms, the correlation coefficient there was only 0.36. Such a low coefficient indicates a desperately weak relationship between variables, in this case, of course, between cholesterol and atherosclerosis. Usually, scientists demand a much higher correlation coefficient before they conclude that there is a biologically important relationship between two variables.

The very low correlation coefficient was arrived at after much study. First, many of the townspeople of Framingham had their cholesterol tested several times over a period of several years. Then, Dr. Manning Feinleib of the National Heart, Lung, and Blood Institute, led a team of coworkers in studying the coronary vessels of those who had died. The researchers were eager to learn which of the many factors they had studied was most important in the development of atherosclerosis in these dead people from Framingham. Was it blood cholesterol or the number of cigarettes smoked, or something else?

After carefully describing the atherosclerosis in the coronary arteries of the dead people, Dr. Feinleib and his associates concluded that the cholesterol level of the blood best predicted the degree of atherosclerosis. Neither age nor weight nor blood pressure nor any other factor was as good as blood cholesterol. But again, the correlation coefficient between cholesterol and atherosclerosis was a mere 0.36.

The written report of the study offered no diagrams and no information about the cholesterol and atherosclerosis of each of the individuals whose bodies had been examined. And the report did not discuss the very low correlation coefficient; it didn’t even comment upon that matter.

When scientists reach a result contrary to all previous studies, it is routine–not merely usual but routine–to provide a detailed report about the result and also to discuss any possible ways in which the study may have been biased away from accuracy and truth. In the Framingham case, there was an especially great need for this routine scientific procedure to be followed. Not only was the correlation coefficient so trivial, but this study, funded with millions of taxpayers’ dollars by The National Institute of Health, could have a major impact on national health care and the American economy. If there was no connection between cholesterol and atherosclerosis, as the previous studies had shown, then there was no reason to bother about cholesterol or the diet. And billions of taxpayers’ dollars could have been spent more wisely than in lowering the cholesterol of healthy people.

But the scientists conducting the Framingham study had no reservations. They were eager to stress their own excellence and to highlight the weaknesses of Dr. Paterson’s study of Canadian war veterans. In their report, they did not mention the studies of Drs. Landé and Sperry at all, nor the studes from India, Poland, Guatemala or the USA. When the Framingham study authors mentioned their opponents, it was only to criticize without putting their own cards on the table. Some of those hidden cards are fascinating to wonder about.

How were the dead of Framingham chosen for postmortem examination, for example? From 914 dead individuals, the researchers examined only 281. And from the 281, they selected 127 (14 per cent of all dead) who became the subjects of an autopsy program especially designed to investigate the heart and its vessels.

Thus, those chosen for autopsy in the Framingham study were not a random sampling of the population, as they had been in the previous studies. The report from Framingham said nothing about the selection criteria, although scientific studies routinely do. Usually the determining factor is age. A postmortem is seldom performed on people who have died peacefully in old age, as most of us will. Primarily, a postmortem is restricted to young and middle-aged people, who have died before their time, and so it was in the Framingham study. Almost half of those autopsied were younger than 65 years. For this reason, the autopsied subjects had to have included a relatively large number with familial hypercholesterolemia, the unusual genetic disease of cholesterol metabolism. Furthermore, people with this disease are of special interest to scientists studying the cholesterol problem and were probably chosen for autopsy in a program tailored to investigate coronary disease.

With only 14% of the Framingham dead chosen for autopsy, the risk of bias must have been great because there is one exception from the above rule: patients with the rare disease familial hypercholesterolemia have much atherosclerosis, and very high cholesterol levels in their blood. If many such patients are included in a study of cholesterol and atherosclerosis, a correlation will be found.

The question about blood cholesterol and atherosclerosis has been studied by coronary angiography also. It seems as if every specialist in coronary angiography in America has performed his own study, funded with federal tax money awarded by the National Heart, Lung and Blood Institute. In paper after paper published in various medical journals, using almost identical words, these medical specialists emphasize the importance of the blood cholesterol level for the development of atherosclerosis (33).

But the reports offer no individual figures, only correlation coefficients, and these are never above a minimal 0.36, usually even smaller. And they never mention any of the previous studies that found no association between degree of atherosclerosis and level of blood cholesterol.

Studies based on coronary angiography are fundamentally flawed if their findings are meant to be applied to the general population. Coronary angiographies are performed, mainly, on young and middle-aged patients with symptoms of heart disease, which means that a relatively large number of patients with familial hypercholesterolemia must have been included. Again, there is an obvious risk for the kind of bias that I described above. The fact that this objection is justified was demonstrated in a Swedish study performed by Dr. Kim Cramér and his group in Gothenburg, Sweden (34). As in most other angiographic studies the patients with the highest cholesterol values had on average the most arteriosclerotic coronary vessels.

But if those who were treated with cholesterol-lowering drugs were excluded, and almost certainly this group must have included all patients with familial hypercholesterolemia, the correlation between blood cholesterol and degree of atherosclerosis disappeared.

In Japan the food is meager, blood cholesterol is low and the risk of getting a heart attack is much smaller than in any other country. Given these facts you will most probably say that in Japan atherosclerosis must be rare.

The condition of the arteries of American and Japanese people was studied in the fifties by Professors Ira Gore and A. E. Hirst at Harvard Medical School (35) and Professor Yahei Koseki from Sapporo, Japan. At that time US people on average had a blood cholesterol of 220 whereas Japanese had about 170.

The aorta, the main artery of the body, from 659 American and 260 Japanese people were studied after death. Meticulously all signs of atherosclerosis were recorded and graded. As expected, atherosclerosis increased from age 40 and upwards, both in Americans and in Japanese. Now to the surprising fact.

When degree of atherosclerosis was compared in each age group there was hardly any difference between American and Japanese people. Between age forty and sixty Americans were a little more arteriosclerotic than Japanese; between sixty and eighty there was practically no difference, and above eighty Japanese were a little more arteriosclerotic than Americans.

A similar study was conducted by Dr J.A. Resch from Minneapolis and Dr.s N. Okabe and K. Kimoto from Kyushu, Japan (36). They studied the arteries of the brain in 1408 Japanese and in more than 5000 American people and found that in all age groups Japanese people were more arteriosclerotic than were Americans.

The conclusion from these studies is of course that the level of cholesterol in the blood has little importance for the development of atherosclerosis, if any at all.

The Cholesterol Myths #3

References