QUICKLINKS AND VIEW OPITONS
Cholesterol Myths by Uffe Ravnskov MD: Myth 6: Lowering Cholesterol Will Lengthen Your Life (Part 1)
Wednesday, November 12, 2014 10:18 am Email this article
“But besides real diseases we are subject to many that are only imaginary, for which the physicians have invented imaginary cures; these have then several names, and so have the drugs that are proper for them.”
—Jonathan Swift (1667-1745)
Time for truth
As one scientific study after another has shown, people can gorge on animal fat for many years and still keep their blood cholesterol low. What we have learned also is that atherosclerosis and heart attacks may occur whether one’s food is meager or fat, and most surprisingly, whether cholesterol is high or low. Given these facts, is there any reason to think that lowering blood cholesterol with diet or medicine can prevent heart attacks?
Based on what I have presented so far, the answer is no. In fairness, however, it still may be possible that high-fat food contains something other than cholesterol and saturated fatty acids that might be dangerous to the heart, or that high blood cholesterol slows the coronary circulation in some way other than by stimulating atherosclerosis. It might just be possible to reach the correct conclusion from the wrong premises.
The diet-heart idea itself is invalid, as I have already demonstrated in several ways. But the best way to know for sure if fat food and a high cholesterol level are dangerous is to use human beings as guinea pigs, to see if coronary heart disease can be induced by feeding these people animal fat or by elevating their blood cholesterol, or to see if heart attacks can be prevented by feeding the experimental subjects a low-fat diet or by lowering their blood cholesterol.
The idea to raise blood cholesterol during several years by dietary means is stillborn no matter how interesting it seems. The ethical committees that must approve all experiments on living creatures should certainly condemn the idea. Fortunately the Masais and other populations already have performed the experiment for us with well-known result.
It is much easier to get a permission to lower blood cholesterol. Many researchers have received permission and have tried although lowering blood cholesterol is possibly more dangerous than increasing it, as I shall soon explain.
To evaluate the effect of lowering blood cholesterol, all other risk factors must remain unchanged. If the test individuals also stop smoking, reduce their body weight or start exercising, or receive treatment for their elevated blood pressure, change their work or get fired, fall in love or get divorced, move to another place with a different climate and culture, or do something else that may influence the condition of their heart or blood vessels, then we do not know what we should attribute the test result to. Is it the cholesterol lowering or is it something else? And this is not the only problem.
The diet-heart proponents say that the prevention of atherosclerosis cannot start too early in life. They add that the best results may be seen if prevention start before the rougher, more rocky deposits develop in our arteries. Here comes a problem, however, because coronary heart disease is uncommon before the age of fifty. To prove that cholesterol lowering prevents heart attacks in young and middle-aged people it is therefore necessary to study many thousands of individuals, preferably those at unusually high risk for heart attacks.
The question we ask is, if fewer heart attacks are seen among people whose cholesterol is lowered by treatment than among untreated people. A cholesterol lowering experiment must therefore include also untreated control subjects. By control subjects, we mean people who have identical risk factors for coronary heart disease, people with, on average, the same blood cholesterol, smoking habits, body weight, and so forth as the individuals who will be manipulated with treatment.
In sufficiently large studies, risk factors usually become evenly distributed by chance, provided that the test subjects and the control subjects are assigned to their two groups on the basis of some random feature such as their day of birth, or by leaving their assignment to a computer. Studies that include randomly selected control subjects are called controlled randomized studies.
As you can see, it is extremely difficult to design even the initial steps of a scientifically acceptable trial. The standards of science are high, however. In fact, they are so high that, even if we manage to select a test group and a control group with almost identical risks for heart disease, we must remember that almost identical and absolutely identical are not the same thing and that we will never know all the factors that may, or may not, stimulate the development of the disease in these people.
To these inevitable problems the trial directors themselves have added one more. If the test individuals are asked not only to lower their blood cholesterol but also to quit smoking, or to lose excess body weight, to get more exercise, or to do something else that we think may be beneficial, we do not know if a possible reduction of heart disease is caused by cholesterol lowering or by something else. Unfortunately, this method, called multiple risk factor intervention, has been used in many trials.
Sighted or blind?
Many researchers have tried to prevent coronary heart disease with diet or drugs. Some of the first trials had so many technical errors that even the diet-heart proponents ignore them when they argue for their idea in their reviews.
One of the more serious errors was that the trials were not blinded. For a trial to be blinded, the patients must not know if they belong to the treatment group or to the control group. In the best experiments, called double-blind trials, not even the doctors know which group any given patient belongs to. Blindness prevents the treated subjects from feeling better merely because they know they are being treated and the control subjects from feeling worse merely because they know they are receiving no treatment; double blindness prevents the doctors who want the treated subjects to benefit, from leaping to conclusions based more on their own hopes than on scientific facts.
Up to 1968 the results of eleven dietary trials were known. Professor Jerome Cornfield at the University of Pittsburg and Dr. Sheila Mitchell at The National Heart, Lung and Blood Institute analyzed these trials. They found that the best results were seen if the doctors knew which group the participants belonged to. In six trials, the doctors knew, and in four of these six the number of heart attacks was reduced. In five trials, the doctors did not know, and in three of these five there was no difference between the number of new heart attacks in the control and treatment groups; in one of these five trials, even more heart attacks and more deaths occurred in the treatment group than in the control group.
Unfortunately, many of the newer trials were neither single nor double-blind, as you will learn soon.
Soybeans against heart attacks
In the 1960s, in London, England, Professor Jeremy Morris led a team of physicians and scientists in an investigation to see if the use of soybean oil instead of animal fat could have some preventive effect on coronary heart disease. This oil is rich in polyunsaturated fatty acids, those that are considered protective against atherosclerosis and coronary heart disease. Enrolled in the trial were about four hundred middle-aged men who had previously been admitted to four London hospitals because of a heart attack; half of these were given a diet containing large amounts of soybean oil. (This is one of the few trials sponsored solely by a government, and not by a drug company or any other vested interest.)
The researchers could see no effect of the oil when the result was analyzed four years later. Although blood cholesterol had decreased considerably in the treatment group, fifteen had died of a heart attack. In the control group, fourteen had died; and the number of nonfatal heart attacks was identical for both groups.
The authors of the report compared their result with a similar but unblinded experiment performed by Dr. Paul Leren, a Norwegian researcher from Oslo. They concluded that, even if Dr. Leren had been more successful, the results of the two trials taken together showed that it was not possible to prevent heart attacks by eating more polyunsaturated fat.
At about the same time, Dr. Seymor Dayton and his team from the University of California in Los Angeles conducted a similar trial. At a nursing home for war veterans they gave a treatment group of four hundred men a diet rich in soybean oil; the four hundred control subjects ate the institution’s usual diet. Great efforts were made to prevent both patients and doctors from knowing who was treated and who was not.
Seven years later, a slightly smaller number of those who had eaten the soybean-oil diet had died from a heart attack. But the lower number of heart attack deaths was balanced by a higher number of cancer deaths.
Moreover, when the researchers analyzed the degree of atherosclerosis and the amount of cholesterol and fat in the arteries of the dead subjects, they discovered something peculiar. Although blood cholesterol had been lowered in the treatment group, there was no difference between the degree of atherosclerosis in the two groups. In fact, those who had eaten the diet laced with soybean-oil had even more cholesterol in the aorta, the chief artery of the arterial system, than those who had eaten the nursing home’s standard fare.
The report from this well-performed trial did not explain why mortality from sclerotic vascular disease had decreased but atherosclerosis itself had not. The authors of the report concluded that the effect of the trial was impressive but that the trial alone was not enough to prove the diet-heart idea; it could not be used as an argument for recommending the diet for the entire population, since only old men had been studied and total mortality hat not been lowered. The authors could also have said that the number of heavy smokers was much higher, indeed significantly higher according to the statistical tests, in the control group. Because smoking is considered a cause of coronary heart disease, the greater number of heart attacks in the control group could, logically, have resulted from smoking and not from diet.
The directors of the trials I have just referred to produced prudent reports about their efforts. It is difficult to find such balanced views from the directors of the trials that were to come.
The Coronary Drug Project
Blood cholesterol can be lowered in many ways. But which way is the most effective, and does it really help? These were the main questions when the American, government-supported National Heart, Lung and Blood Institute started the first mammoth trial to lower blood cholesterol. The year was 1967.
The trial, headed by professor Jeremiah Stamler from Chicago, was called The Coronary Drug Project. The drugs used were nicotinic acid, clofibrate (Atromidin®), thyroid hormone and estrogen (the female sex hormone), the latter given in two different dosages. Because these drugs and hormones lower blood cholesterol, they were considered appropriate for efforts to prevent coronary heart disease.
The subjects in The Coronary Drug Project were more than 8000 middle-aged men who had already had at least one heart attack. About 5500 of these men were randomly assigned to five treatment groups, with the rest assigned to a control group of roughly 2800. Altogether, 53 hospitals from all across the whole US contributed patients to this massive study.
The trial was well prepared. Anything of interest of coronary heart disease was studied by a large number of researchers. In the paper describing the project the list of researchers filled six pages.
Within 18 months after the start of the trial, treatment for those who had received the high dosage of estrogen discontinued because the researchers found that the hormone was causing heart attacks instead of preventing them. And the patients were reluctant to take the estrogen because most of them became impotent and developed feminine-looking breasts. The investigators concluded, “the potential value of this level of estrogen medication is probably limited.”
Those who were treated with half dose estrogen continued the treatment, but a few months later even the smaller dosage was found to be unfavorable. In addition to the side effects cited above, there were also more new cases of cancer.
Treatment with thyroid hormone was discontinued, as well. Although blood cholesterol was lowered, the treatment seemed to induce heart attacks instead of decreasing them, just as was the case with estrogen. The remaining groups continued to the end of the trial.
The result after seven years was depressing. Those who were treated with clofibrate had died just as often as those in the control group and many of them had had serious side effects of the treatment.
Even more side effects were seen after treatment with nicotinic acid. Almost all complained of flushing or skin rashes, half itched, and one in five complained of stomach pains, nausea or other symptoms pertaining to the stomach and bowels. Other common side effects of nicotinic acid were gout (a painful inflammatory disease of the joints), burning pains while urinating, excessive sweating, serious disturbances of the heart rhythm, and various skin diseases. As the directors of the experiment wrote in their report: “Great care and caution must be exercised if this drug (nicotinic acid) is to be used for treatment of persons with coronary heart disease.”
What they left unsaid, though, was how to exercise care and caution. And how can we? How are we doctors to know, before treatment begins, who will experience side effects? The drug was ineffective, anyway, to prevent fatal heart attacks. And why should we use an ineffective drug?
Nicotinic acid is still used today for prevention of coronary heart disease because of a peculiarity that appeared in a study years after The Coronary Drug Project ended. Eight to nine years after all the treatment subjects stopped taking nicotinic acid, a follow-up study showed that fewer in the previously treated group had died from heart attacks and that fewer had died of any cause.
This result—no benefits from the drug during the trial, but fewer deaths years later—stimulated many speculations. One was that perhaps the nasty side effects had concealed its positive effects during the trial. The suggestion here was that perhaps it took many years before a lowering of blood cholesterol would show positive results. The books should be reopened after other trials also.
But clofibrate had lowered blood cholesterol just as much as nicotinic acid had during the trial. Yet clofibrate had prevented no deaths after many years. In fact, as the years passed, the number of fatal heart attacks in the group that had once received this drug was a little greater than in the control group. And nobody mentioned the follow-up findings of another large experiment, the WHO trial (see later). There, more people had died from heart attacks 4-5 years after their treatment with clofibrate ended.
It seems strange also, that a drug could help years after being discontinued, as strange as if the aspirin unsuccessfully taken to relieve a headache on Monday could prevent a headache on Friday.
It is hard to follow the logic in the conclusions from the cholesterol-lowering trials. Sometimes cholesterol lowering results in fewer deaths from heart attacks; sometimes the same degree of lowering results in more deaths. Sometimes the benefit is seen after a short time, sometimes not until years after trial’s end. Or, if there is no benefit—the most common result—the trial directors declare that if the trial had continued longer, there might have been some benefit. However, when trials are sometimes beneficial and sometimes not, the more likely conclusion is that they aren’t effective at all which means that their outcome depends on chance.
Primary versus secondary
The trials I have discussed so far are examples of secondary prevention. “Secondary” means that a disease—coronary heart disease in this instance—has already occurred, and that treatment is aimed at halting further spread of the disease. In contrast, a treatment aimed at preventing the disease in apparently healthy people is called primary.
There is a fundamental difference between primary and secondary prevention. Very often, people who have already had a heart attack are badly frightened and ask themselves, “Will I survive another coronary?” To prevent another heart attack, many are willing to submit to rather unpleasant kinds of treatment. Healthy people whose only defect is high cholesterol are less inclined to exercise, to renounce cigarettes and good food, and, on top of this, to take expensive drugs with unpleasant side effects. Healthy people thus make less compliant treatment subjects in a trial.
In addition, to achieve a significant result, a primary trial requires many more subjects, because the risk of a heart attack is considerably smaller for people who have never had a heart attack than for those who already have suffered one. While the subjects needed for a secondary preventive trial number in the hundreds, those needed for a primary preventive trial are many thousand. And if it is not possible to prevent a heart attack in those who have already had one, it is obviously more difficult to do so in healthy individuals. The results of the primary preventive trials have not been any more successful than those of the secondary preventive trials, even if the diet-heart proponents say otherwise. But now it is up to you, the reader, to judge for yourself.
The Upjohn trial
In the early 1970s, Dr. Albert Dorr and his coworkers at the Upjohn Company, a large pharmaceutical manufacturer in Kalamazoo, Michigan, started a trial to test Upjohn’s new cholesterol-lowering drug colestipol (Lestid®).
At a large number of hospitals in the USA more than 2000 men and women with high cholesterol were selected. After the selection the local doctors consulted the directors of the Upjohn trial. In the offices at Upjohn, after being informed about certain laboratory values of the participants, the directors of the new trial decided which patients would have the drug and which would receive an ineffective sugar pill, the placebo.
To assign participants of a trial in this way obviously may introduce a bias, especially when those who assign them have vested interests. As you will see from the following the distribution of risk factors in the treatment and control groups became far from even.
Two years later the result was analyzed. No effect was seen for the women in the trial. But the effect for the men was amazing. After only two years, the number of heart attacks had been halved for the men in the treatment group. Such remarkable results have never been achieved in any trial before or since.
There was a snag, however. In the control group, the number of individuals with familial hyper?cholesterolemia was greater than in the treatment group. As the prognosis for these people is worse than for others; some of them die young from heart attacks and lack of balance on this matter may well have biased Upjohn’s result.
The WHO Trial
At the same time a similar trial was being performed under the auspices of the WHO. This trial, led by Professor Michael Oliver at the University of Edinburgh, Scotland, used clofibrate, the same drug that was used in the Coronary Drug Project.
For the WHO trial, blood cholesterol was analyzed in 30,000 healthy, middle-aged men in Edinburgh, Prague, and Budapest. The men with the highest blood cholesterol were selected for the treatment, a total of ten thousand individuals. Half of them were treated with clofibrate, the other half with an ineffective placebo.
After about five years treatment, 174 of those who had been taking the placebo had suffered a non-fatal heart attack, but only 131 of those treated with clofibrate. Apparently the drug was a success.
But the number who had died from heart attacks was equal. Worse yet, considerably more of those who had been treated with the drug had died from other diseases. In all three cities more men had died in the treatment group. Taken together, 128 died in the clofibrate group, 87 in the placebo group. And 4-5 years after the trial, even the number who had died from heart attacks was larger in the treatment group.
Clofibrate is still recommended as a useful drug in many countries, however.
Fat food and fit Finns
One of the nations with the highest mortality from coronary heart disease is Finland. The mortality is especially high in the province of North Karelia, for reasons that no one knows. The coronary mortality rate increased year by year up to the 1960s. Of course the Finnish health authorities were concerned. To them, it was self-evident that the cause was high cholesterol, because in Finland it is higher than anywhere else, and some of the highest values are found in North Karelia.
A team of doctors and scientists at the university in Kuopio headed by Professor Pekka Puska decided to do something about the problem. They chose to start in North Karelia. To see if their efforts were beneficial, they used the district of Kuopio as their control, because people in Kuopio died just as often from heart attacks as the people of North Karelia did and their cholesterol was equally high.
In 1972, a public health campaign began throughout North Karelia. Its aim was to prevent heart disease by focusing on smoking, fat food, and high blood pressure. In the mass media, on posters, at public meetings, and through campaigns in schools and work places, the message was proclaimed. In Kuopio nothing was done; here, people were allowed to live as they had traditionally lived.
Five years later the number of heart attacks among North Karelian men had decreased from 0.77 to 0.63 percent each year. The total mortality had also decreased. Possibly this trend could improve in the future, the leaders of the campaign wrote in their report.
There was a problem, however. In Kuopio, where the citizens served as control subjects, the number of heart attacks had decreased even more, among women as well as men, although the townspeople ate and smoked as they had before. In fact, heart mortality had decreased in all the provinces of Finland (figure 6a).
The disappointment of the campaign leaders is easy to imagine. All this enthusiasm and all this work were of no use. On the other hand, negative results can be and are interesting for those whose curiosity is intact and who are more interested in knowledge than in the defense of old positions.
Two conclusions may be drawn from the results of the campaign. First, it could not have been fat food or smoking or high blood pressure that caused the many heart attacks in North Karelia. If it were, the number of heart attacks should have decreased more in North Karelia than in the untreated Kuopio. Second, something had happened in the whole country to cause coronary heart disease to decrease, and it was not an improved diet, it was not reduced smoking and it was not a greater attention to blood pressure.
Unfortunately, the Finnish campaign team did not understand that they had found a track worth examining more closer. Instead, they published further papers with more analyses of their results. More had died in North Karelia during the campaign than in Kuopio, they agreed, but in North Karelia there had been a greater reduction in the number of heart attacks than in all other areas of Finland. They also thought that the small irregularities of the mortality curve for North Karelia proved that their campaign had been of benefit.
Figure 6a. Number of coronary deaths in North Karelia and in the rest of Finland. The circle indicates the starting point of the North Karelia project. Observe that more died from heart attacks during the first three years of the project than during the year before the start, most probably a result of chance. Seen over a longer period of time, heart mortality had declined in North Karelia as well as in all of Finland and in many other countries. This decline had already started before the start of the North Karelia project. After Puska.
But what they did not mention was that the decrease in heart mortality had started several years before the start of the campaign, and that an examination of the curve over a longer period of time clearly revealed that the campaign had made no impact. In fact, if the small irregularities in the curve had any significance at all, heart mortality had increased during the first two years of the campaign. However, the investigators were so convinced of their success that they started similar campaigns in other parts of Finland.
One of the campaign leaders, Dr. Jukka Salonen, had a differing opinion. In a letter to the famous medical journal The Lancet he explained that, although he was a co-author of the North Karelia report, he had not been able to read the optimistic paper before its publication. He did not think it possible to draw the conclusions presented in the paper and admitted that the steeper slope of the mortality curve for North Karelia could be explained in a variety of ways. For instance, he wrote, the increased heart disease seen during the previous decades had come later to North Karelia; the decrease, therefore, had come later, too. The critical factor, however, was whether the campaign had changed the trend in North Karelia (which it had not). Dr. Salonen wrote that the North Karelia project could not be used as evidence to say that the risk factors either caused or did not cause heart disease; the project had merely shown that intervention is possible and can lead to a change in risk factors.
But, as Professor Michael Oliver of Edinburgh, answered: “What was the aim of trying to change risk factors unless they were thought to have some causative role and unless positive results were expected?”
On June 23, 1988, there appeared a full-page advertisement in a great many Finnish newspapers. The ad had been paid for by Valio, a large farm cooperative that markets about 90 percent of all milk products in Finland. It presented “five facts about dietary fat you have wished to hear about but nobody has told you.”
The five facts were as follows.
1. In Finland people eat less fat than in many Western European nations.
2. There is no direct connection between a nation’s intake of animal fat and its mortality from coronary heart disease.
3. Fat intake and mortality from coronary heart disease have changed in opposite directions in many countries.
4. Mortality from coronary heart disease has decreased in Finland, despite the fact that the Finnish people has increased its consumption of animal fat.
5. Finally, a short summary of the North Karelia campaign was given.
It would be an understatement to say that the director of Valio’s research department, Kari Salminen, met stormy weather; a hurricane is more like it. He was attacked in all Finnish newspapers and journals; almost every day during that summer and autumn, critical editorials, articles, and letters appeared in the Finnish press. No one could point to anything incorrect in the advertisement. Rather, it was the morality of the corporation that was rotten. The advertisement was condemned as partial, misleading, and unethical; the claim was that Valio had selected convenient statistics for a deliberate manipulation of scientific data.
As director of the North Karelia project Pekka Puska was particularly offended. He thought that the various efforts of the campaign had produced marked effects at the start (correct, but effects in the wrong direction). The connection between dietary animal fat and heart disease has been better proved than most subjects in medicine, he wrote. The facts in the advertisement, he declared, did not tell the whole truth.
Kari Salminen, Valio’s research director, answered that his company had merely done what the nutrition scientists had been doing for decades. Further, he replied, the advertisement had been designed as an invitation to debate and did not pretend to represent the whole truth.
The invitation was ignored. Debate was replaced by execution, and for many years Pekka Puska appeared in Finnish advertisements for margarine.
The Oslo trial
In Oslo, Norway, Dr. Ingvar Hjermann and his team thought that smoking and a high cholesterol level were the two most important causes of coronary heart disease and they wondered what would happen if smoking was stopped and blood cholesterol was lowered with an appropriate diet. To this end they studied about 1200 middle-aged men, mostly smokers, with high cholesterol. Half of these men received dietary advice and were encouraged to quit smoking. The other half received no treatment.
The result after five years appeared promising. In the group given dietary and smoking advise, 19 died from a heart attack. In the control group, the number was 35; if to the latter group was added a further control participant who had died suddenly of an unknown cause, the difference between the treatment and control groups became statistically significant.
A promising result for the diet-heart supporters. But does the experiment really prove that a faulty diet causes coronary heart disease?
In their paper, the Norwegian researchers pointed to two types of intervention, diet and cessation of smoking. They admitted that if dietary advice had been the only treatment their result would not have been sufficient as evidence.
In fact they had used three types of intervention, because subjects in the treatment group were also advised to reduce their weight. Evidently, this advice was followed: at the end of the trial, the mean weight difference between the two groups was almost seven kilograms (about 15 pounds). Opinions vary as to the importance of a 6-7 kilogram weight difference. It is evident that the risk of diabetes and high blood pressure is greater for overweight people, and diabetes and high blood pressure predispose for coronary heart disease. Most proponents of the diet-heart idea also recognize overweight as a problem in heart disease. Wrote Dr. William Kannel, director of the Framingham project: “Avoidance and correction of obesity deserve a high priority among measures taken to avoid coronary heart disease, since the combined effect of the risk factors it promotes on coronary heart disease incidence is formidable.”
Now to the crucial question. Which of the measures had had the decisive effect in the Oslo trial? Was it lowering of cholesterol by diet, was it the reduction of smoking, or was it the weight loss? Nobody knows.
Possibly you’re asking, “Why didn’t the Oslo trial leaders concentrate on the diet alone?” The answer to that question lies in a previous paper.
About ten years earlier, the researcher who had performed the Norwegian soybean trial, Dr. Paul Leren, had published the latest results of that trial. Although the number of heart attacks was reduced a little, Dr. Morris and his colleagues in England, using a similar treatment, had failed.
But Dr. Leren analysed his own result and found that it might have been a good idea to change more than one risk factor, an approach called multiple risk factor intervention. Diet alone was unsuccessful.
Few of the diet-heart supporters rely on diet only in their trials; diet is often combined with other measures. When the LRC trial was being planned, the scientists stated frankly that diet alone was not enough, that to lower cholesterol drugs were necessary.
It is laudable to try to prevent disease and premature death as effectively as possible. If all the measures are proved to be beneficial the frontier should of course be broadened. But in that case it is not possible to judge the influence of each measure individually. As no one has proved that diet alone is efficient it had perhaps been wiser to exclude the dietary advices, or to study them alone.
MR.FIT—Much ado about nothing
For many years scientists at the National Heart, Lung and Blood Institute had discussed how to prevent heart attacks. But before telling the American public what to do they needed solid proof that their advice would work.
They had rejected the idea of using diet alone. To be successful, they said, it was necessary to attack at least three of the major risk factors: high cholesterol, smoking, and elevated blood pressure. To this end the institute started a gigantic trial called the Multiple Risk Factor Intervention Trial—MR.FIT for short. At its head was once again professor Stamler from Chicago.
The first step was to recruit more than 360,000 middle aged men from eighteen American cities. After a routine investigation the researchers selected about 12,000 men, namely those who were considered especially prone to get a heart attack.
The trial had every chance of succeeding. The test subjects had entered the trial voluntarily and they knew that their condition was considered dangerous. Although they felt hale and hearty (and, by normal standards they were healthy), they were overweight, their blood pressure was too high, and according to the experimenters their cholesterol scores hinted at premature death from heart disease. After the initial analyses, the men took part “with remarkably enthusiastic response.”
However, one of those initial analyses should have stopped the whole MR.FIT trial.
A decade before, a smaller, careful test study had been done. A comparison of the food eaten by the men in that study with the food consumed by the men selected for MR.FIT revealed that the MR.FIT participants had eaten more “healthfully” in all respects, more in accord with the diet-heart idea.
Yet the blood cholesterol of the MR.FIT participants was higher!
Furthermore, initial surveys indicated that those MR.FIT participants who ate less saturated fat and cholesterol tended to have higher blood cholesterol! It was not exactly an encouraging finding for researchers who hoped to lower blood cholesterol by lowering just those components in the diet. But the directors of the trial responded only by declaring that, first, the odd evidence showed that cholesterol and saturated fat should be reduced more than originally planned for the MR.FIT treatment group. Second, they speculated that the fact that blood cholesterol was highest among the MR.FIT subjects who ate the most prudent diet showed that these men must have changed their diet at the last minute, right before the trial began.
Perhaps the directors were correct. The participants could have made eleventh-hour dietary changes in the days just before they were questioned about what they ate. But presented with experimental results contrary to what they have expected, scientists usually want to know “What’s going on here? And why?” In accord with the long tradition of scientific inquiry, most scientist would have asked the MR.FIT participants if they really had shifted to a new diet right before the trial began. More than one hundred million dollars of taxpayers’ money could have been saved if some scientists had asked. If MR.FIT participants were eating as they had always eaten, even just before the trial, they would have demonstrated that diet is unimportant for the blood cholesterol level, and this enormously costly trial could have been cancelled then and there.
But no one asked. Or did they? Perhaps it would have been too heroic for the directors to cancel a trial with all these doctors, nurses, dieticians and, not least, the trial directors themselves, lined up and assured that they would have lucrative and prestigious jobs for the next several years? If anyone asked, they didn’t do it in public, and the trial continued.
The subjects were randomly assigned to two groups of equal size. Those placed in the treatment group and their families met in small groups to learn about the rationale behind the trial, and then to learn how to read food labels, to cook with minimal fat, and to change old recipes to meet new guidelines. In special sessions, the treatment subjects met for an intensive anti-smoking campaign; in selected cases, even hypnosis was used to help participants quit smoking. When necessary, individual counseling was provided by doctors, nutritionists, psychologists, nurses and other health professionals. Every four months, the treatment subjects were called in for blood sampling and to hear if they had fully understood all the new guidelines.
The dietary advice was, of course, aimed at reducing the men’s intake of cholesterol and saturated fat and increasing their intake of polyunsaturated fat. High blood pressure was treated energetically, and subjects with weight problems were taught how to reduce calories and get more exercise. Dietician checked yearly to make sure that the men were really eating as prescribed.
The men in the control group received no advice, but they visited the center once a year for blood sampling and a questionnaire about their eating habits, and the results of these investigations were sent to their own doctors.
After seven years of treatment the effect was analyzed. The trial directors were satisfied that there had been major risk-factor changes. Blood pressure had been lowered considerably and many of the men had quit smoking.
But blood cholesterol had decreased by only seven percent. It had decreased in the control group, too, although the control subjects had scarcely changed their diet at all, so the difference between the two groups was only two percent.
Other risk factors had changed in the control group, as well. The one difference between the groups worth mentioning was that more of the control subjects continued to smoke.
The difference in number of deaths was small, too. In the treatment group 115 had died of coronary heart disease, in the control group 124. According to statistical precepts, such a difference could well have been due to chance. There was no statistical difference, either, in the number of deaths from all causes: 265 in the treatment group, 260 in the control group.
Customarily, when a scientific experiment does not produce results supporting a hypothesis the scientists admit it straight out. But this was not an ordinary experiment. More than a decade of hard work and several hundred millions of dollars had been invested in this most ambitious medical study to date. Hundreds of doctors, professors, statisticians, dieticians, psychologists and others had been engaged. More than fifty scientific reports, most of them mammoth, had been published. And thousands of apparently healthy men, and their families, had been persuaded to take part in time-consuming investigations and to change their diet and way of life for many years. This huge effort could not possibly have been in vain.
And what had been preached for years to the American public about risk factors and heart attacks could not possibly have been wrong.
With a little statistical manipulation, the trial directors improved their results.
The participants were divided into smaller groups. Excluding from the treatment group a sub-group who did especially bad made the overall result appear better. Almost all the other sub-groups had had fewer fatal heart attacks. No great differences, and not all sub-groups, but almost all.
The trial directors concluded that the MR.FIT intervention program might have had a favorable effect for most of the participants. If some of the men had more heart attacks, it was because of the drugs used to lower blood pressure (although in another subgroup treated with such drugs, the outcome was better). It was also obvious that the outcome was favorable for those who had quit smoking, the directors wrote. In fact, the smoking habits explained the whole difference.
Within four years after the end of Mr.FIT, a total of 202 men in the treatment group and 226 in the control group had died from heart disease; again a difference that could be explained by chance (or by the smoking habits). But the investigators claimed that the figures proved the benefit of lowering blood cholesterol.
More prudent diet-heart supporters admit that MR.FIT was a failure, but they usually add that the failure occurred because a two percentage lowering of blood cholesterol is too small to have any effect.
This is a reasonable objection, but with this objection diet is declared worthless as a preventive measure, because the diet had been changed as was aimed. The subjects in the treatment group had almost halved their intake of cholesterol, they had lowered their intake of saturated fat by more than 25 percent, and they had eaten 33 percent more polyunsaturated fat. In the control group the diet was practically unchanged.
If a scientific trial with almost unlimited economic and personal resources cannot lower cholesterol more than two percent over seven years, how is the over-worked general practitioner to succeed with a crammed waiting room and with no dieticians, or experts in behavior modification to hold his hand? And how is the patient to be motivated if he is not rewarded for all his trouble?
MR.FIT demonstrated that it is a good idea to quit smoking. But we already knew that, and most people can manage to quit without such costly help from society.
The final proof
Diet-heart proponents think that if we had a drug that could lower blood cholesterol sufficiently without any serious side effects, we could prevent or at least delay all diseases caused by atherosclerosis.
Here is a dream for all doctors. All that’s necessary is a prescription pad and a gadget for measuring cholesterol. No time-consuming fuss with diet-counseling.
It is a dream, also, for the drug producers. A life-long lowering of cholesterol with expensive drugs in a substantial part of the population is far more profitable than for instance a brief treatment with cheap penicillin. In the offices of the drug manufacturers, the dream calculations are in the billions of dollars.
Large trials using clofibrate had not been especially encouraging, you may say. But other drugs seemed more promising. One of them was cholestyramine (Questran®).
The MR.FIT trial had excluded men with extreme cholesterol values (above 350 mg/dl), as it is considered unethical to place such “patients” in a control group without treatment. In the new jumbo trial called The Lipid Research Clinic Coronary Primary Prevention Trial (LRC), which would use the drug cholestyramine, all individuals with high cholesterol were included. The higher cholesterol, the better.
To solve the ethical dilemma, dietary advice was given in both the LRC’s treatment and control groups. Although this advice would diminish the difference in outcome between the two groups, the degree of cholesterol lowering from diet was expected to be insignificant! The great difference would be created by cholestyramine.
To find about 4000 test individuals, blood cholesterol was determined in almost half a million middle-aged men. Never in history had so many people with such high blood cholesterol levels been involved in a medical experiment. In MR.FIT the upper three percent of the cholesterol range were selected, but men with the highest values were excluded. Here in LRC, only the upper 0.8 percent participated, without exception. Consequently, the mean blood cholesterol before the dietary treatment started was about 40 mg/dl higher than in MR.FIT meaning that most of the participants must have had familial hypercholesterolemia.
All LRC participants were investigated thoroughly, as in MR.FIT. After a few weeks dietary treatment, half of the men were started on cholestyramine medication, the other half on a supposedly inactive placebo powder.
Seven to eight years later, the results were analyzed. Although blood cholesterol in the treatment group had decreased by more than eight percent, the differences in the numbers of heart attacks were so small that only chance could explain them. Of those who had taken cholestyramine, ten percent, or 190 men, had experienced a non-fatal heart attack, as against 11.1 percent, or 212, of the controls, a difference of 1.1 percentage points. As for fatal heart attacks, the figures were 1.7 and 2.3 percent, a difference of 0.6 percentage points, or twelve individuals.
But in the summary of the paper the result was given in another way. The lowering of non-fatal coronary heart attacks was said to be 19 percent and of fatal heart attacks 30 percent. These figures were arrived at by relating the percentage in the treatment group to the percentage in the control group, without any reference to the total number of men involved.
Even the exaggerated figures of the LRC report were a little too optimistic. To reach their 30 percent figure, the LRC directors included the uncertain cases, those who may or may not have died from a heart attack. But to reach their 19 percent figure, they excluded the uncertain cases. If it had been the other way around, the results would have been 24 percent rather than 30, and 15 rather than 19. In other words, they selected the most convenient figures.
Even worse, the LRC directors had lowered their own statistical demands. In a preliminary report written several years before the trial ended, they had stated that, to be convincing, they would accept nothing less than the strongest statistical proof of their findings. In this case, it was a statistical level of 0.01, meaning that the trial results would be 99 percent accurate; and to ensure statistical accuracy, the researchers would use the very demanding two-tailed t-test.
Thus, the directors of the trial had begun by embracing the highest standards. Then, after the fact, when it was clear that the result of the trial did not measure up to their hopes, they shifted their demand for accuracy from 0.01 to the less stringent 0.05, and to the easy one-tailed t-test.
After their results were published, the LRC directors were severely criticized for their lowering of standards. But in response to critical letters to The Journal of the American Medical Association, they simply denied that they had ever declared in writing the high standards that they had originally aimed at. “The term ‘significant’ was not defined in terms of a particular statistical probability level.”
Diet-heart supporters look offended if you tell them that, of half a million men, twelve were rescued from death. In fact, the number rescued was even smaller. Fewer in the treatment group died from heart attacks (32 against 44), but more died by violence or suicide (11 against 4). If we calculate in the ingenious way used by the LRC leaders and other diet-heart proponents, using relative risk and not absolute rate, the excess of violent deaths was huge; after all, eleven is 175 percent greater than four.
Hyping the benefit, minimizing the risks
If all men in the USA with blood cholesterol as high as in the LRC study received the same treatment and got the same result, two hundred lives would be saved per year, provided that the LRC result was not merely due to chance. However, in a 1990 letter to the editors of The Atlantic magazine, Dr. Daniel Steinberg, chairman of the conference that started the publicly funded National Cholesterol Education Campaign against cholesterol in the USA, declared that 100,000 lives could be saved each year. He further claimed that this non-fact had been demonstrated with statistical significance “in a large number of studies.”
Just a few months later, Dr. Basil Rifkind, who had been the director of the LRC study, admitted in a medical journal that the scientific trials had not reduced the number of deaths from coronary heart disease and that “further gains in life expectancy are unlikely in developed countries.”
The LRC results were so feeble that they may well have been caused by mere chance. And both the drug used in the study, cholestyramine, and the supposedly innocent placebo taken by the control group produced some extremely unpleasant side effects. Sixty-eight percent of the men taking the cholestyramine experienced gastrointestinal side effects during their first year of treatment: they had gas, heartburn, belching, bloating, abdominal pain, nausea, and vomiting, and almost fifty percent had constipation or diarrhea. In the control group during the first year, forty-three percent experienced similar side effects, a far higher rate than what occurs if the placebo is truly ineffective.
The readers of the report are reassured that the side effects were not serious and they could be neutralized by standard clinical means, and after seven years the number of these side effects had decreased to only twenty-nine percent. This was not more than after the placebo treatment. Their words suggested that the symptoms from the stomach and the guts had nothing to do with the cholestyramine treatment but was pure imagination or symptoms that the test individuals should have had also without treatment.
In controlled drug trials the control group is usually given an ineffective placebo. The reason is that symptoms considered as side effects may in fact be accidental symptoms unrelated to the treatment; symptoms which by chance appear during the treatment. Accidental symptoms may of course occur in the controls also. Therefore, the percentage of side effects in the placebo group are subtracted from the percentage of side effects in the treatment group to give the true percentage of side effects from the drug.
But here they had given a placebo which certainly not was without side effects; gastrointestinal symptoms in forty-three percent is much more than is usually seen after an innocent placebo. Therefore it is not reassuring to hear that the side effects from the drug equaled the side effects from the placebo.
Neither is it reassuring to learn that, “the side effects were treated by standard clinical means.” These words mean that more than half of these previously healthy individuals in addition to cholestyramine or placebo also took laxatives, antacids or drugs to stop diarrhea or to prevent nausea and vomiting.
A greater number in the treatment group were also admitted to hospital for operations or procedures involving the nervous system. No diagnoses or more specific information was given and as it was impossible for the experimenters to find a reliable explanation to the effect of cholestyramine on the nervous system these side effects were classified as coincidental. The authors did not consider that it should have been the lowering of blood cholesterol and not cholestyramine itself which had given rise to the side effects.
Few people know about the many side effects, which they may get by taking cholestyramine. One reason is of course that they are rarely mentioned to the public. For instance, read another sentence from the letter in The Atlantic by Daniel Steinberg: “The drugs in current use for lowering cholesterol levels have remarkably few side effects and, to my knowledge, no fatal side effects.”
Now we have definitely proved that it is worthwhile to lower blood cholesterol; no more trials are necessary. Now it is time for treatment. In short, this was the message from the experimenters of the LRC. And they considered treatment necessary for most people.
It is a prudent rule in clinical science to be careful with conclusions about other patient groups than those who have been studied, especially concerning a disease with large age and sex variations. If it had been shown (which indeed is questioned) that a treatment is beneficial for middle-aged men with an extremely high blood cholesterol, only this category should be treated until it has been proven that it is also beneficial for other categories of human beings.
In the LRC it was not even middle-aged men with high blood cholesterol who had been studied but to a great part men with inborn errors of cholesterol metabolism. You may probably recall that it was those with the upper 0.8 percent of the blood cholesterol values who had been selected for the trial. As almost one percent of mankind has an inborn abnormality of cholesterol metabolism most of the participants must have belonged to that category. Even if we presume that the treatment was useful it is not evident that a treatment that is useful for such individuals is useful for normal individuals as well.
But nothing was said about that in the paper, neither had the directors of the LRC trial any reservations. Not only middle-aged men should be treated but also other age groups; and not only men with a high blood cholesterol, but also those whose cholesterol was close to normal; and not only men, but also women although women were not studied and although almost all previous studies had shown that a high cholesterol is not a risk factor in women, neither is treatment of any use. The only group that was not mentioned in the report was the children, but this was repaired later.
LRC was not designed to assess directly whether cholesterol lowering by diet prevents heart attacks, they wrote, but the results from the LRC trial taken together with the large volume of evidence relating diet, plasma cholesterol levels, and coronary heart disease, its findings support the view that cholesterol lowering by diet also would be beneficial.
This is a typical argument from diet-heart supporters. Taken together one by one no study has proven that animal fat and high cholesterol is dangerous to the heart, but if you put all the studies together, they do. In the Alice-in-Wonderland atmosphere of the Lipid Research Clinics, nothing plus nothing conveniently equals something.
Science by citation
The high rate of coronary heart disease in Finland has prompted several experimenters to conduct preventive trials; Dr. Tatu Miettinen and his coworkers from Helsinki are among them.
One-half of about 1200 middle-aged, more or less overweight and hypertensive male business executives with high blood cholesterol were advised about smoking, exercise, weight reduction and diet; the other half was used as a control group. If the blood pressure and the blood cholesterol in the treatment group did not become normal they were also treated with various blood pressure and cholesterol lowering drugs.
The experimenters were quite satisfied with the effects of their efforts on risk factors. Blood cholesterol fell by 6.3 percent, the blood pressure by about 5 percent and the tobacco consumption with about 13 percent.
But improved risk factors did not lead to better end results. In the group who exercised, reduced their weight, ate less animal and more vegetable fat, and had quitted smoking, twice as many heart attacks were seen as in the control group.
The investigators believed that the greater number of heart attacks probably was due to clofibrate which some of them had taken, or perhaps to the drugs against high blood pressure. (What a frightening thought that drugs which are used on millions of people to lower the blood pressure or to prevent coronary heart disease could actually cause it instead.)
Their explanation does not jibe with the results of other experiments. In two previous British trials it was said that due to clofibrate the number of heart attacks had decreased, and other studies have shown that the drugs used against high blood pressure protect against coronary heart disease.
The unfavorable result may simply have been due to the fact that the therapy is ineffective. Therefore, the outcome is determined by chance; in one trial the number of heart attacks is a little smaller, in another a little greater. But the diet-heart proponents prefer to look at the supportive studies only and ignore those that are not.
Science Citation Index was an interesting aid for scientists at that time. (Today the PubMed program on the web is even better.) Here you could see who had cited any scientific paper, how often, and where. Editors of medical journals make a point of the papers published in their journal being cited frequently. Frequent citation indicates influence and is prestigious, not only for journals but also for individual scientists. The number of those who have cited papers by Nobel Prize laureates took up many columns each year in the small-typed Science Citation Index.
It was interesting to open the Index and see how often the 1985 paper by Miettinen and colleagues had been cited. Let us compare it with the main report from the LRC trial also, published one year previously. Both papers dealt with the same subject and were published in the same journal and no one has questioned the honesty of the experimenters or the quality of the studies; at least not the Finnish one. Reasonably, they should have been cited almost equally often. That the LRC trial, at least according to its directors, was supportive, and the Miettinen trial was not, is unimportant because the aim of research is to find the truth, whether it supports the current theories or not.
Table 6a. Here you can see how often the two papers have been cited during the first four years after their publication.
|Miettinen and coworkers
The table shows that scientists are like the rest of us; they forget what is awkward and recall only the pleasant memories. A useful quality in private life but nothing to further knowledge.
What do you want: a gastric ulcer or a coronary?
New cholesterol-lowering drugs require new trials. That the previous results were less successful than expected was due to the side effects of the drugs. The fewer deaths from heart disease were balanced by more deaths from other reasons, it was said. Others thought that the cholesterol was not lowered sufficiently.
The new drug gemfibrozil (Lopid®) is chemically close to clofibrate, but was considered more favorable because it lowers the total amount of cholesterol in the blood and at the same time increases the “good” cholesterol. This drug was selected for a new trial in Helsinki, Finland in a project led by Professor Heikki Frick.
Once again investigators chose healthy, middle-aged men with high blood cholesterol. All participants were advised to quit smoking, to exercise and to loose weight; and half of them were given gemfibrozil, the others a placebo drug.
Also in this trial the number of deaths was equal in the two groups, but for the first time a statistically significant reduction of non-fatal heart attacks was seen after cholesterol lowering only. In the Oslo-trial the participants had quitted smoking and lost weight also; in the LRC trial the effect was not significant according to the usual statistical methods; and in the WHO-trial from 1978 the smaller number of non-fatal heart attacks were outnumbered by a greater number of fatal ones.
Has science proved that high cholesterol is the killer? May we use the trial of Professor Frick and his colleagues as an argument to lower cholesterol in a large part of mankind?
According to the diet-heart idea cholesterol is dangerous because it generates atherosclerosis. If this were true then a lowering of blood cholesterol should also influence other vascular diseases caused by atherosclerosis. However, in all the trials the end points used had been fatal and non-fatal coronary heart disease only.
In all the tables of the trials the reader will find a small group of patients classified as “other cardiovascular diseases.” In most trials the number in this category is a little greater in the treatment group. No great differences, but when the effect of the trial is close to the border of statistical significance, as it usually is in the cholesterol trials (when the effect is not directly negative) the small differences between the numbers of “other cardiovascular diseases” take on great importance. If all cardiovascular diseases including coronary heart disease are put together then the result is no longer statistical significant and the result is as before: no difference which could not have been caused by chance.
In addition the treatment gave unpleasant side effects. During the first year 232 or 11.3 percent of the treated individuals had gastrointestinal symptoms. Gradually the side effects abated. The report did not tell whether the test individuals became accustomed to the drug or whether they were treated with other drugs to combat the side effects as in the LRC trial. What we know is that in the treatment group 81 were operated upon because of some gastrointestinal ailment, in the control group 53 only. Thus, if the difference in the number of heart attacks was real and not caused by chance the question is if you prefer an operation of your stomach or gall bladder, or a non-fatal heart attack, because the sum of heart attacks and operations was almost identical in the two groups.
It is a fact that even this trial failed to lower mortality from coronary heart disease and there was no difference between the total number of deaths either; if anything, more had died in the treatment group. But this is not the end of the Helsinki story.
Here are links to the other chapters in the book.
- Cholesterol Myths by Uffe Ravnskov, MD, PhD: Forward to Book by Michael Gurr, PhD
- Cholesterol Myths by Uffe Ravnskov, MD, PhD: Author’s Foreword
- Cholesterol Myths by Uffe Ravnskov MD, PhD: Introduction: The Diet-Heart Idea: A Die-Hard Hypothesis
- Cholesterol Myths by Uffe Ravnskov MD, PhD: Myth 1: High-Fat Foods Cause Heart Disease
- Cholesterol Myths by Uffe Ravnskov MD, PhD: Triglycerides
- Cholesterol Myths by Uffe Ravnskov MD, PhD: Myth 2: High Cholesterol Causes Heart Disease
- Cholesterol Myths by Uffe Ravnskov MD: Familial hypercholesterolemia—not as risky as you may think
- Cholesterol Myths by Uffe Ravnskov MD, PhD: Myth 3: High-Fat Foods Raise Blood Cholesterol
- Cholesterol Myths by Uffe Ravnskov MD, PhD: Myth 4: High Cholesterol Blocks Arteries
- Cholesterol Myths by Uffe Ravnskov MD, PhD: Myth 5: Animal Studies Prove the Diet-Heart Idea
- Cholesterol Myths by Uffe Ravnskov MD, PhD: Cholesterol lowering in children
- Cholesterol Myths by Uffe Ravnskov MD: Myth 6: Lowering Cholesterol Will Lengthen Your Life (Part 1)
- Cholesterol Myths by Uffe Ravnskov MD: Myth 6: Lowering Cholesterol Will Lengthen Your Life (Part 2)
- Cholesterol Myths by Uffe Ravnskov MD PhD: Myth 7: The Statins — Gift to Mankind (Part 1)
- Cholesterol Myths by Uffe Ravnskov MD PhD: Myth 7: The Statins — Gift to Mankind (Part 2)
- Cholesterol Myths by Uffe Ravnskov MD PhD: “The most exact data base”—the screenee
- Cholesterol Myths by Uffe Ravnskov MD PhD: Myth 8: Polyunsaturated Oils are Good for You
- Cholesterol Myths by Uffe Ravnskov MD PhD: Dr. Ornish and The Lifestyle Heart trial
- Cholesterol Myths by Uffe Ravnskov MD PhD: Myth 9: The Cholesterol Campaign is Based on Good Science
- Cholesterol Myths by Uffe Ravnskov MD PhD: Insider Insight
- Cholesterol Myths by Uffe Ravnskov MD PhD: Myth 10: All Scientists Support the Diet-Heart Idea
- Cholesterol Myths by Uffe Ravnskov MD PhD: Epilogue
- Cholesterol Myths by Uffe Ravnskov MD PhD: References
Articles on the same subject can be found here:
Please feel free to share your comments about this article.
© Copyright 2003-2017 - Larry Hobbs - All Rights Reserved.