The cholesterol myth

In every major long term study, lowering cholesterol through drugs or the standard recommended low fat diet has done more harm than good.

When was the last time you ate an egg without feeling guilty about a soaring cholesterol count? Since the Sixties, when doctors first hypothesized that lowering blood cholesterol levels would prevent heart attacks and strokes, everyone in the Western world has become obsessed with the cholesterol content of every last morsel they put in their mouths. Even McDonald's has recently got in on the act, boasting the low fat content of their hamburgers in advertisements placed in medical journals like the Lancet. But recent evidence now shows that the neither cholesterol lowering drugs nor the recommended diet does anything to prevent heart disease and may even do harm.Epidemiologist Dr James McCormick, Dean of the School of Physics in Trinity College, and President of the Irish College of General Practitioners, along with fellow Trinity professor Petr Skrabanek have co-authored a book entitled Follies and Fallacies in Medicine (Tarragon Press, Glasgow). Although we do not agree with everything in the book, notably an utter dismissal of all forms of alternative medicine, they raise many interesting points about how the "science" of modern medicine is built upon faulty premises. In the following, which is adapted from their book, they lay the cholesterol myth to rest.

Since the Second World War coronary heart disease has become a much more important part of the lives of both doctors and lay people. Epidemiologists have expressed their interest by examining factors which are associated with an increased probability of developing coronary heart disease, so called risk factors. Risk factors, better called risk markers to emphasize that they are associated with an altered probability of developing disease rather than necessarily being causally related, have been described in numerous prospective and case control studies. At the moment some three hundred risk factors for coronary heart disease have been described and the list continues to grow. At the moment it includes: cigarette smoking, high cholesterol, high blood pressure, obesity, diabetes, low levels of high density lipoproteins, selenium, thiazide diuretics, not drinking, not exercising, not having siestas, not eating fish (especially mackerel), living in Scotland, speaking English as a mother tongue, having a high level of phobic anxiety, being scrupulous about keeping appointments, not taking cod liver oil, and snoring. The important associations include being male, a family history of the disease and perhaps most important of all, because it is alterable being poor in the rich world.

Because risk factors are associated with an altered probability of developing disease, it was assumed that an alteration in risk factors would reduce death and morbidity. This led to the believe that identifying risk factors in healthy populations would be a good thing to do. This has turned out to be a dangerous delusion. Dangerous because altering risk factors does little good and may do harm.

The best evidence about the effects of altering risk markers comes from controlled trials. In these trials, the risk status of half the population under study is altered by some intervention, while the other half continues as before; both populations are followed through time to see whether or not they develop disease. Almost all the studies to date have been concerned with middle aged men in whom the risk of coronary disease is reasonably high. Such studies are still difficult and expensive because large numbers of subjects need to be recruited and followed for many years.

There have been five major multiple risk factor intervention trials, all of which were in middle aged men. The duration of follow up was between five and twelve years. The risk factors which were altered by intervention were diet, smoking and blood pressure; in two studies, attempts were also made to reduce weight and increase exercise.

Some studies have examined the effects of drugs or diet alone (that is, the standard World Health Organization recommendation of limiting fat intake to 30 per cent of total dietary intake, with no more than 10 per cent each of saturated, polyunsaturated and monounsaturated fats.

After 828,000 man years of study the results were as follows: 1015 coronary heart disease deaths in the intervention groups, 1049 in the control groups; 2909 total deaths in the intervention groups, 2947 deaths in the control groups, a difference of 36: that is, four less deaths in 10,000 men per year. Such a small difference is well within the limits of chance.

Particularly in America, the main interest now centres on cholesterol and everybody is being encouraged to know not only their blood pressure but also their cholesterol count. There have been three major trials of reducing cholesterol by drugs in middle aged men whose cholesterol was in the upper ranges of normal. The results after 115,176 man years of observation were as follows: 92 coronary heart deaths in the intervention groups, 100 deaths in the control groups, 275 total deaths in the intervention groups but only 240 deaths in the control groups. In other words, lowering cholesterol with drugs did no good and may have done harm.

The other main risk marker is raised blood pressure. Hypertension, as it is known in medicalese, is perhaps the most widespread and damaging of present day non diseases. The only workable definition of hypertension is "blood pressure that I (the doctor) treat". A study of the prevalence of hypertension in those aged 50-64 in Australia and the US concluded that almost 70 per cent of Australians and almost 50 per cent of Americans were hypertensive. (Hypertension was defined as a systolic pressure of greater than 140 mm Hg and/or a diastolic pressure of 90 mm Hg or more. Systolic is the arterial pressure when the heart pumps blood, diastolic when the heart is resting. We still measure blood pressure in millimeters of mercury since doctors used to measure it with a glass column filled with mercury. )

There is good evidence that treating the small minority of people who have sustained blood pressure of more than 105 mm Hg diastolic reduces the subsequent incidence of stroke.

For anyone else, the results of the Medical Research Council Trial (British Medical Journal, 13 July1985) of treating patients with mild to moderate hypertension with either propranolol or bendrofluazide (a drug which increases the excretion of water and salt) against placebo showed no benefit in reduction of all causes of mortality in those who had received active treatment. In the study 117,534 cases underwent 85,572 patient years of observation. Deaths in the treated group were 248, and in the placebo group 253. Side effects among those receiving bendrofluazide and propranolol included gout, diabetes and impotence.

Those who took placebo tablets did not escape side effects. By 12 weeks, 16 per cent of those men taking the diuretic, 14 per cent of those taking propranolol and 9 per cent of those taking the placebo had become impotent; at two years these figures had risen to 23 per cent in the diuretic group and to 10 per cent in those taking the placebo! This suggests that attaching the label hypertensive has, of itself, deleterious effects.

The harm to benefit ratio of treating mild or moderate hypertension is, in the present state of our ignorance, adverse that is, the probability of harm outweighs the possibility of good. The authors of the Medical Council ResearchTrial, which is the best, if imperfect, evidence available, concluded that if 850 mildly hypertensive patients are given antihypertensive drugs for a year, about one stroke will be prevented. It is, however, impossible to predict which people are most likely to have a stroke if untreated, "so this benefit can be achieved only at the expense of involving a substantial percentage of people in adverse reactions to the drugs, mostly but not all minor".

The likelihood of diagnosing non hypertension increases with age. Dr F. H. Messerli and others from the Ochsner Clinic in Louisiana (New England Journal of Medicine, 13 June 1985) found that half of 24 hypertensive patients over the age of 65 had "pseudo-hypertension". Pseudo-hypertension is a false alarm caused by increased resistance to compression of the artery by the blood pressure cuff because of hardening of the arterial wall. The difference between cuff pressure and true pressure (as measured by direct intra arterial measurement), ranged from 10 to 54 mm Hg with a mean of 16 mm Hg for both systolic and diastolic pressures. This would suggest that about half the patients over 65 whose mean cuff blood pressure was 180/100 had a true blood pressure of less than 165/85, which is normal for their age and would not justify treatment. If these patients' blood pressure were to be lowered by drugs it would not only be inappropriate and wasteful of resources but would also place these people at the risk of side effects and even of death.

Those who advocate the identification of risk markers and who believe that coronary heart disease is preventable often cite the North Karelia experiment. North Karelia is a county in Finland, which had the highest known mortality rates from coronary disease. It was decided that there should be a major county wide campaign in North Karelia to reduce risk markers and that the results would be compared with a neighbouring control county, Kuopio. While morality fell in North Karelia, it also fell in Kuopio and in all the other counties of Finland, where there had been no equivalent attempt to alter risk status.

The other argument which is often used in favour of early intervention is the fall in mortality rates which has occurred in the United States and Australia and some other English speaking countries, and which is ascribed to healthier lifestyle, dietary changes, less smoking, treatment of blood pressure and, maybe, better management of heart attacks. These data are based upon what is written in death certificates. However, a recent careful study in Minnesota showed no change in the rates of heart attacks between 1970 and 1980, a time when national rates were said to be falling. In Sweden, despite a reduction in known "risk factors", coronary mortality in men aged between 40 and 74 is increasing. Another discordant finding is that in a number of countries mortality rates in men and women are moving in opposite directions.

A more probable reason for the failure to show that this disease is preventable is insufficient knowledge about its cause. The long list of risk markers is a tribute to our ignorance, rather than a proof of our knowledge. Coronary disease may not be a single disease; relatively young men's heart attacks may not have the same cause as heart attacks in eighty year old women and both may be different from those people who get heart pain on exertion (angina pectoris) but do not suffer heart attacks.

In this circumstance, routine heart checks which measure blood pressure, weight and cholesterol are more likely to do harm than good. It may be that as a result of future research, probably carried out in the laboratory rather than by epidemiologists, we may be in a position to identify high risk groups, for which there are safe and effective ways of reducing risk and preventing disease. In the meantime, activity should be restricted to advising only those individuals who seek help. It is reasonable to advise stopping smoking, particularly heavy cigarette smoking, and if there is a strong family history of coronary disease early in life, to measure cholesterol and other blood lipids. There is no place for mass screening for risk markers or population interventions to alter the eating habits of the nation (at least in terms of standard cholesterol lowering diets. But see our views, p. 8, about the real culprit editor).