Dietary Fat and Coronary Heart Disease. Is There a Connection?

The problem of the relationship between dietary fat and health is exceedingly complex. Since the principal health question regarding fat relates to its association with the risk of coronary heart disease (CHD), this becomes the central issue that must be examined. Not only are there several distinct types of fat, but there are a number of different consumers, i.e., old, young and in between, men and women, pre- and postmenopausal women, individuals with the so-called metabolic syndrome or Syndrome X, the obese and the overweight, those with normal and those with abnormal blood lipids (the several cholesterol fractions and triglycerides), diabetics, individuals who have had a heart attack and individuals who are thought to be free of all coronary heart disease, smokers and non-smokers, etc. Most studies look at only a few sub-sets of this general population. Since fat consumption alters cholesterol levels, the fat-heart disease question must also involve a consideration of the relationship between fat consumption and cholesterol on the one hand, and cholesterol and heart disease on the other. Thus the subject is far from simple.

The history of the anti-fat movement will be reviewed first, followed by an examination of what modern nutritional epidemiology and nutritional science has revealed about the merits and risks of dietary fat, especially in the context of coronary heart disease. Then we will discuss low-fat and high-fat diets with these results in mind.

THE ANTI-FAT MOVEMENT

In the last twenty-five to thirty years, fat has gone from being taken for granted as a normal constituent of Western diets to being branded public enemy number one. This change came about through the efforts of the US Government, mainly the USDA, and various groups such as the American Heart Association, the National Academy of Science, and the American Dietetics Association. The anti-fat movement was also greatly assisted by the medical profession and the media, both of whom were active in spreading the word. The battle cry--fat causes heart disease, strokes and cancer--was, as might be expected, very effective. The food industry picked up on this movement very early and an ever-increasing stream of low-fat, lite, low- cholesterol, and zero-fat products have appeared on the shelves of the supermarkets. Fat substitutes have become popular--they taste like fat but are not digestible. Books on low-fat diets have appeared by the dozens. An almost zero-fat diet was claimed to reverse heart disease, which the general public has been lead to believe is caused by dietary fat and cholesterol particles plugging up the coronary arteries (unfortunately, it is not that simple). The USDA came out with its famous food pyramid, which basically favors a high-carbohydrate and low-fat diet. Today the notion that fat is bad is securely established in the minds of the general public. In fact, the movement has many of the attributes of an organized religion. It is therefore quite interesting that the lumping of all fat into one category and labeling it as bad has virtually no basis in modern scientific studies, and the elimination of some fats from the diet can have very serious health consequences. Gary Taubs, in his recent feature article in the journal Science (1) called it "The Soft Science of Dietary Fat." His article contains an excellent review of the anti-fat movement.

A problem with the early studies that led to the condemnation of fat was that the various fats were lumped together as simply fat, without regard for the fact that there were thought to be good fats and bad fats. The need for and actions of the good fats were, to some extent, recognized and understood even at the start of the anti-fat campaign, and this was the basis of the opposition from a number of scientists. It turned out that their concerns were ignored. Today it is absolutely clear that any discussion or study concerning fats and health must deal with the individual fats, that is, the saturated fats, the monounsaturated fats, the polyunsaturated fats (with special attention to the omega-3 and omega-6 fats), and the trans-fats, which are mostly man-made (2). This obviously complicates matters quite considerably. In fact, when the decision was made to condemn all fats, part of the motivation was the view that the general public was not able to deal with these complexities, which was probably correct. A decrease in consumption of the good fats was considered the price that had to be paid to decrease the consumption of what were viewed as the bad fats. The molders of public opinion considered that on balance there would be benefit, and they would not listen to those who disagreed and pointed out the risks. Also, how bad the trans-fats were was not fully appreciated in the early days of the low-fat movement, when part of the gospel was to replace butter with margarine, which was high in trans-fat. More on this later.

Another problem with the early studies that led to the condemnation of all fat was that, by modern standards, they were rather crude. The science of nutritional epidemiology has changed significantly during the last twenty years, with the result that study design has vastly improved, statistical power increased, and the ability to correct for confounding factors has reached a very high level (3). An example of the confounding factor problem is as follows. Let's say a study finds that heavy drinking of alcohol is positively associated with lung cancer. But since heavy drinkers also tend to be smokers, frequently heavy smokers, the data must be corrected for the known risk of lung cancer associated with smoking. Otherwise the observed association of drinking with lung cancer is meaningless. Studies of dietary fat and health are confounded by numerous factors associated with both lifestyle and other nutritional factors. Early studies by and large failed to deal effectively with this problem.

Finally, in the period when the early studies were being performed, little was known about the relationship between the various blood lipid fractions such as HDL, LDL, and the triglycerides, and the risk of cardiovascular disease. Only total cholesterol was measured. At present, there is considerable information on the relationship of the various lipid fractions to CHD risk, and this continues to be an active area of research.

Today, the early studies of the relationship between fat and health are regarded by many nutritional epidemiologists as mainly of historical interest, although these studies are still quoted by proponents of the "all fat is bad" religion as they struggle to maintain their viewpoint. In what follows, the subject of dietary fat will be explored in some detail, with special emphasis on potential health problems associated with low-fat diets. The questions that need to be addressed are as follows:

• Which fats, if any, are bad, which fats are good? What is the actual evidence?
• Do fats really cause heart disease and cancer?
• Are there dangers associated with low and very low fat diets?
• On the basis of modern nutritional epidemiology, what are the recommendations as regards optimum balance of the various types of fat, carbohydrate and protein in the diet if the goal is to remain healthy?

FAT CONSUMPTION AND HEART DISEASE

There are three types of study that are of importance in this context. One involves following large groups of individuals, the so-called cohorts, using food frequency questionnaires and in some cases interviews. Subjects are followed for a number of years, and the participants in any given study may number in the thousands to tens of thousands. Associations are sought between diet and the incidence of one or more health problems that occur at a later date (the prospective, cohort study). The second type of study involves an intervention. In the present context, the two common ones are (a) reducing the total intake of fat in general, and (b) replacing saturated fat with unsaturated fat, keeping the total fat calories approximately constant. In connection with the diet-heart hypothesis, the investigators then look for changes in the incidence of events related to heart disease, such as a survived or fatal heart attack. From such studies, one can determine if there is an association between fat intake and the incidence of adverse coronary events. A third type of study, called the case-control study, involves recruiting both a group of individuals with a specific problem or problems and a control group presumed free of the problem or problems. Factors are then sought that are present in the group with problems and absent in the controls.

In connection with the questions of whether or not scientific studies support the hypothesis that there is an association between adverse heart events and the consumption of either saturated or polyunsaturated fat, Willett (3) has reviewed 21 prospective cohort studies involving over a quarter-million participants. Four of the nineteen studies supported the hypotheses that saturated fat was bad, 10/19 showed no significant association, and the balance provided insufficient information. In connection with the hypotheses that polyunsaturated fat was beneficial, 4/19 supported this hypothesis, 9/19 showed no association, and the balance provided insufficient information. There are two other similar reviews in the recent literature (4,5). One essentially agrees with the Willett analysis, the other finds a slight suggestion that saturated fat may have an adverse association, but the association is weak. There is a small but significant variation in the studies included in these two last mentioned reviews, but the general picture is clear. If saturated fat were tried before a judge and jury on the basis of this evidence, it would no doubt be acquitted! The evidence is simply inconclusive.

Nine dietary intervention studies have also been reviewed by Hu and coworkers at Harvard (6). In the studies reviewed, groups of individuals were randomly assigned to an intervention or control group, and the incidence of events in the intervention group was compared with the number found in the control group. Two of the studies involved presumably healthy individuals, and seven involved patients who had experienced a heart attack. Two of the studies involved reducing total fat. One found a benefit (reduced adverse coronary events), the other found none. In both cases the changes were very small. Of the seven other studies included in the review, in six the intervention involved decreasing saturated fat and increasing unsaturated fat. In one study, the intervention involved adopting a Mediterranean type diet augmented with alpha- linolenic acid, one of the omega-3 essential fatty acids. Six of the seven studies found a significant benefit whereas one found none.

The study that found the largest decrease in adverse coronary events was the one that involved switching to a Mediterranean diet. In this study the changes in total cholesterol were also tabulated. The change in the incidence of adverse coronary events failed completely to correlate with the change in the serum cholesterol. If the data are plotted, one obtains a scatter diagram with no visible trend or correlation. In the study where there was the largest decrease in adverse coronary events (73%), there was no on-average change in the serum cholesterol. Note that seven of the nine studies involved heart attack patients--so- called secondary prevention. Also note that in the studies where saturated fat was replaced with unsaturated fat, two variables are being simultaneously changed. Also, only two of the studies where saturated fat was replaced by unsaturated fat involved healthy individuals (primary prevention), and one of these found no effect. These studies nevertheless are universally viewed as providing convincing evidence that the action of replacing saturated fat with unsaturated fat is a beneficial move in the context of heart disease, especially preexisting heart disease. In fact, it is hard to argue with the overall benefit observed, which was quite large. However, such studies fail to prove that saturated fat is bad, since that question was not studied in isolation. The prospective cohort studies mentioned earlier attempted to pull out saturated fat as an independent variable but failed to provide any conclusive or consistent evidence that it was bad. Conclusive evidence that polyunsaturated fat was good was also not obtained in these cohort studies. The advice to decrease saturated fat and increase polyunsaturated fat is now incorporated in most modern recommendations as a presumably beneficial dietary modification.

It seems clear that on the basis of the above studies, one cannot conclude that the consumption of saturated fat has been proven to be bad in the context of heart disease. In fact, it would appear close to neutral. If it is indeed dangerous to the heart, the risk is too small to stand out even in a large number of studies. After all, the prospective cohort studies discussed above involved over a quarter-million subjects! The fact that this point of view is at variance with the older studies has been explained by the failure to take into account confounding factors. Also, selection of only data favorable to the fat-heart disease hypothesis appears to have taken place. In the most notorious case, a comparison of the incidence of heart disease, as a function of fat consumption for a group of different nations, led to a remarkable, straight-line correlation where, in fact, when all the data available at the time are used, a scatter plot with no correlation is obtained (7). Yet this study was for years the basis of the belief that fat consumption caused heart disease and is still quoted by those who adhere faithfully to this dogma.

Polyunsaturated fats, as a class, include the so-called trans-fats. Most of the trans-fats encountered in the modern diet are man-made by the partial hydrogenation of polyunsaturated fats using high temperature and a metal catalyst. While the question of whether or not saturated fat is bad will no doubt continue to be debated, there appears no question whatsoever that trans-fats are bad, in fact really bad. Yet they permeate the modern diet. Part of the motivation for their creation was vastly extended shelf life, since natural unsaturated fats oxidize easily and thus become rancid. Also, this was a way to make liquid fat into solid fat, such as Crisco and stick margarine. The body treats trans-fats like real fats, and, among other things, incorporates them into the cellular lipid bilayer. They are found in fried foods such as French fries, baked goods, and in many prepared foods. There has been a lot of research on the problems caused by the consumption and metabolism of trans-fats. For example (2):

• Trans-fatty acids lower HDL in a dose response fashion, i.e. the more you eat, the lower it gets (the wrong direction for the so-called good cholesterol).
• Trans-fatty acids increase LDL (the so-called bad blood lipid).
• Trans-fatty acids raise lipoprotein(a), which is regarded as undesirable
• Trans-fatty acids raise total cholesterol in the blood by as much as 20 to 30 mg/dL.
• Trans-fatty acids decrease the levels of testosterone in males and increase the number of abnormal sperm.
• Trans-fatty acids inhibit the function of a number of essential enzymes in the body associated with the chemistry of omega-3 and omega-6 fatty acids.
• Trans-fatty acids interfere with the action of enzymes that metabolize toxic chemicals.
• Trans-fatty acids decrease the response of cells to insulin. They hamper the proper function of the insulin receptors by changing the fluidity of the lipid bilayer and other cellular membranes. Thus they may contribute to insulin resistance.

This is a horrible report card for something that today is a major constituent of the typical North American diet. It is estimated that for a typical teenage diet with 3000 calories and 35% from fat, that 40% of the fat calories are from trans-fats. It may be even higher. Trans-fats totally permeate the processed food production. Until the trans-fat content is included as a part of the label on all processed foods, all one can do is look for the words partially hydrogenated and avoid all products that contain such ingredients. Individuals wishing to avoid trans-fats should avoid all commercial baked goods, reject margarine totally (unless it is the new trans-fat free product) and avoid all deep fried foods as well unless they are deep fried at home, using fresh unsaturated oils at a minimum temperature to accomplish the cooking. Some scientists who have studied the trans-fat phenomenon in detail consider it to be potentially one of the major health problems of the 21st century.

FATS AND CHOLESTEROL

Aside from the indication that it is beneficial to replace saturated fat with unsaturated fat, the cohort and intervention studies fail to provide much guidance. It is instructive to turn to the relationship between fat consumption and serum (blood) cholesterol levels. It appears to be an uncontested fact that the consumption of saturated fat increases the levels of low-density lipoprotein (LDL) cholesterol, which is regarded as the bad lipoprotein. However, life is never simple, since saturated fat also increases the high- density lipoprotein (HDL) cholesterol fraction as well, and HDL is considered very beneficial. The extent to which this results in a draw, so to speak, is not clear, but the conventional wisdom ignores the increase in HDL, focuses on the increase in LDL, and thus concludes that saturated fat as bad. However, if an individual had a normal level of LDL and a low level of HDL, raising the HDL might on balance be beneficial even if the LDL level also went up. The percentage decrease in risk is 3% for every 1% increase in HDL, whereas there is only a 1% increase in risk for each 1% increase in LDL (8). The changes in both LDL and HDL with the ingestion of saturated fat will depend on the nature of the fat, that is, the actual saturated fats involved, and as well, will depend on the individual, since there is considerable variation. Perhaps it is because of the combination of good and bad effects of saturated fats, coupled with individual variation, that the studies discussed above provided no definitive guidance as to the question of saturated fat increasing the risk of CHD, but only inconsistent findings.

Both monounsaturated fats (olive oil is high in monounsaturates) and polyunsaturated fats turn out to be beneficial in the context of cholesterol levels. They raise HDL and lower LDL. The interplay between diet and HDL, LDL, total cholesterol, and the other important blood lipids, the triglycerides, merits further discussion, not only since it relates the "bad" label attached to saturated fat and the "good" label associated with the mono- and polyunsaturates, but also because it relates to the question of high vs. low fat diets which can modify these blood lipid levels.

Studies far too numerous to list have established, probably beyond reasonable doubt, that high levels of total cholesterol and LDL cholesterol present a risk for the development of CHD. Since LDL cholesterol represents typically more than 50% of the total cholesterol, this total is considered a surrogate for LDL. Thus the studies that measured only total cholesterol remain valid today and are part of the case for high serum cholesterol being a risk factor. Four quite different and more or less independent approaches to lowering both total cholesterol and LDL produce reduced risk of CHD (surgery, sequesterant drugs, statin drugs and diet). This is one of the cornerstones of the argument that LDL is bad and that lowering it represents a beneficial intervention. There is also considerable literature indicating that low HDL carries added risk for CHD and that raising HDL decreases the risk (9-12).

However, there are some disturbing aspects to this picture. In fact, Ravnskov (4,7) and a few others have been very vocal critics of the hypothesis connecting cholesterol, fat and CHD, but their views are perhaps a bit extreme. Nevertheless, as any doctor involved with heart patients will tell you, something like 50% of all heart attacks occur in individuals who have a normal blood lipid profile, i.e. LDL is not elevated and HDL is not low. In addition, there is frequently no correlation between the extent of LDL lowering and the decrease in risk, and the decrease in risk in some studies is independent of the level of risk in the subjects studied (13,14). This latter phenomenon is also seen in the diet intervention studies, as discussed above. In addition, a significant decrease in CHD risk is found by lowering LDL in individuals that are already at low risk as judged by their blood lipid profile (15,16).

It does not help that the details of the mechanism of the adverse action of LDL is poorly understood, as is the overall mechanism of the formation of atherosclerotic plaques, from their beginnings as fatty streaks in arteries to large, unstable plaques, the rupture of which is thought to be related to heart attacks. In fact, the conventional wisdom that atherosclerosis is caused by high cholesterol has recently been challenged by Ravnskov in the Quarterly Journal of Medicine. He documents the following points (17):

• Serum cholesterol does not predict the degree of atherosclerosis at autopsy.
• Serum cholesterol does not correlate with the degree of coronary atherosclerosis on angiography.
• Serum cholesterol does not correlate with the degree of coronary calcification.
• Serum cholesterol does not correlate with the degree of peripheral atherosclerosis.
• Changes in serum cholesterol concentrations are not followed by parallel changes in atherosclerosis growth--i.e. there is no exposure response.

Thus an important question: why does high cholesterol predict cardiovascular disease if LDL levels and changes in LDL do not correlate with the degree of atherosclerosis or with atherosclerosis growth? As Ravnskov points out, there can be many reasons for these observations. One is that high LDL or total cholesterol may be secondary to other factors that promote cardiovascular disease (17). On-going research will no doubt eventually produce a more satisfactory picture.

It seems clear that LDL is bad cholesterol and HDL is good cholesterol. Saturated fat consumption raises both and might be viewed as neutral. Eating mono- or polyunsaturated fat decreases LDL and raises HDL. From this one might conclude that it is pointless to worry about fat. However, to not worry about fat at all would be to ignore the apparently beneficial effects of the unsaturated fats and the studies that demonstrated a significant decrease in the risk of CHD when saturated fat is replaced with unsaturated fat-- a benefit that does not appear to prove that the reduction of saturated fat was indeed beneficial, but that merely the combined action has merit. In this context, it is important to understand that if one merely increases the consumption of unsaturated fat and does not decrease the saturated component, there may be a risk of adversely disturbing the energy balance with an associated weight gain, unless other macronutrients such as protein or carbohydrate are decreased.

LOW-FAT DIETS

Once the "fat is bad" religion was firmly established and supported by the US government, the medical profession, the nutritionist community, the media, and, with great enthusiasm, the food industry, the so- called low-fat or very low-fat diet was a natural result. Since most people tend to maintain a certain level of caloric intake, low fat in fact meant a dietary change that involved replacing fat with carbohydrate. Protein in general was not viewed as a candidate since most common protein sources contained lots of fat. Carbohydrate has roughly half the calories per gram as compared to fat (4 vs. 9 kcal/gram), so the low fat diet was also a high carbohydrate diet, and this was embraced by millions as the path to health and longevity. Lack of knowledge among the general public as to the various types of carbohydrates and their relative effects on blood glucose, insulin, and the development of insulin resistance led to the indiscriminate consumption of increased amounts of carbohydrates, in many cases carbohydrates from sources such as potatoes, white rice, ordinary pasta, white bread, bagels, baked goods, sugar-rich deserts, non-diet soft drinks and juices, etc. Much of the carbohydrate that replaced fat was of the type that was rapidly digested to yield large increases in post-meal (called postprandial) blood glucose and insulin, which frequently leads to insulin resistance and eventually the so-called metabolic syndrome or Syndrome-X (18). Carbohydrate not needed for energy was turned into fat and stored--something the general public was probably not aware of, since the "fat is bad" religion did not include information on the metabolism of carbohydrates. The notion that eating fat made one fat, and avoiding fat therefore would result in weight stability or weight loss, was part of the conventional wisdom of the masses.

The replacement of fat with carbohydrate has been studied a great deal since the "fat is bad" religion became commonplace. In many individuals, such a change in diet results in a significant decrease in serum HDL, an equally significant increase in serum triglycerides, and eventually the development of insulin resistance, where the body requires larger and larger amounts of insulin to manage the load of serum glucose that arises from the heavy carbohydrate component of the diet. Insulin resistance frequently leads to type 2 diabetes, which in turn results in a whole host of serious health problems and a well documented huge increase in the risk of heart disease. Most diabetics in fact die of heart disease. The above-described changes in HDL and triglycerides also are in the direction of increasing the risk of CHD. It actually turns out to be difficult to separate the effects on CHD risk of high triglycerides and low HDL, since these two conditions frequently occur together, but there are studies that indicate that high triglycerides can be viewed as an independent risk factor (19). Suffice to say that there are a number of studies in the recent literature that indicate a very high level of increased risk when HDL is low and triglycerides are high (9-12). Interestingly enough, men with conventional risk factors for CHD actually have low risk if they have low triglycerides and high HDL (11). Some investigators think that in fact low HDL and high triglycerides can present a greater risk than either high total cholesterol or high LDL (8). In some individuals, low HDL and high triglycerides are accompanied by a change to smaller LDL particles. There is considerable evidence that these are more dangerous in the context of heart disease than large, less dense LDL particles (20). In fact, this is one of the frequently encountered arguments against the low-fat high-carbohydrate diet. However, not all studies find that CHD risk higher when the serum concentration of small, dense LDL particles predominates (21).

ESSENTIAL FATTY ACIDS

By restricting fat consumption, there also arises the risk of a deficiency in the essential fatty acids, which are critical to health. The omega-3 and omega-6 (n-3 and n-6) polyunsaturated fatty acids require a more detailed discussion. They are critically important because they are involved in the production of what are called eicosanoids, which are transient hormones that play key roles in human biochemistry. There are a large number of different eicosanoids, and a detailed discussion (22) is beyond the scope of this review. Suffice to say that the subject can be simplified by thinking in terms of two classes of eicosanoids, which are frequently called "good" and "bad" although this simple pair of labels conceals the fact that good and bad eicosanoids are both necessary. It is thus not a matter of minimizing the bad and maximizing the good, but of keeping the two classes in balance. How they complement one another can be seen from the following table:

These lists are far from complete but nevertheless illustrate the complimentary nature of the two classes as well as the central role they play in essential biochemical processes. The connection with cardiovascular disease is also evident. The relationship between the n-3 and n-6 polyunsaturated fatty acids and the eicosanoids is that the n-3 fats are involved in the production and regulation of the good eicosanoids whereas the n-6 fats come into play with the bad eicosanoids. Since balance is the name of the game, it is important not only to get enough of each type of fat, but also to get them in roughly the correct proportion for optimum functioning of all the systems that depend on eicosanoids and the hormones derived from them. One school of thought on this subject maintains that the n-3 to n-6 ratio should by around 1:2 to 1:4. There is some evidence that our genetic makeup evolved on a diet that had a ratio in this range (23). While the optimum ratio is debatable (24), what is clear is that the typical Western diet is very heavy in the n-6 fats and very light in the n-3s. This is due to the high consumption of vegetable oils that are rich in n-6 fatty acids and the low consumption of fish, nuts, and the oils that are rich in n-3s. There is a great deal of literature on the beneficial effects of increasing the intake of n-3 fats, and most of this is in the context of cardiovascular disease. The typical Western diet has been estimated to have an n-3 to n-6 ratio of 1 to 20. Thus there seems to be little risk associated with dietary modifications that reduce n-6 consumption, and there appears to be considerable room to increase the n-3 content of diets, a move that should prove beneficial.

An interesting twist to the n-3 story is that the parent acid, alpha-linolenic acid, is converted with variable efficiency in humans into two critically important n-3 fatty acids, EPA and DHA. Fish, however, are rich in these two n-3 fats, and they can be found in concentrated form in fish oil. There are dozens of studies that have shown that n-3 fatty acids from fish and other sources are active in preventing heart attacks and sudden cardiac deaths by preventing erratic heart rhythms, reducing the tendency of blood to clot inside arteries (a cause of heart attacks), improving the blood lipid profile (total cholesterol, HDL, LDL, and triglycerides), and favorably influencing inflammation which is thought to play an important role in the development of atherosclerosis (plaque formation). By far the strongest connection between the n-3 fatty acids and heart disease is the reduction of sudden deaths, presumably because of the strong anti- arrhythmic effects of these fats (25). This is very important since for a significant number of victims, sudden death is the first indication of heart disease. It is clear from the above that by reducing the dietary fat content to a very low value, there is a significant danger of insufficient n-3 essential fatty acids and of having a very high ratio of n-6 to n-3, because what fat there is in the low-fat diet will generally be rich in the omega-6 fats.

In the above discussion of the connection between dietary fat and CHD, one study was mentioned that produced an astounding 73% reduction in risk of a second adverse coronary event as compared to the control group in patients who had already experienced one heart attack. The dietary approach in this study was to switch the treated group to a Mediterranean style diet and to add supplementary alpha-linolenic acid in the form of an enriched margarine (free of trans-fats) thus increasing the n-3 content of the diet. This study, known as the "Lyon Diet Heart Study" underscores the importance of the n-3 essential fatty acids in connection with heart disease (26,27).

In summary, many experts in nutrition, who base their views and recommendations on the recent literature, are suggesting that the popular movement to replace fat with carbohydrate was in fact a serious mistake in the context of public health, that the recommendation in fact has no basis in valid scientific studies, and that the general public would be well advised to instead replace some saturated fat with unsaturated fat and limit the consumption of rapidly digested (high glycemic index) foods (24). The consumption of the right mix of fat, protein and slowly digested carbohydrates such as are found in many vegetables, as well as low- glycemic index fruits, is recommended as a prudent approach to a diet that is more in tune with current knowledge regarding the prevention of CHD than the low-fat high-carbohydrate diet.

In the past 25 years, eggs have been held up as a prime example of fat, cholesterol and caloric excesses in the American diet. The widespread bad publicity resulted in almost a 50% drop in egg consumption. Eggs typically contain about 200 mg of cholesterol per egg, and are a major source of dietary cholesterol in Western diets. In addition, the egg contains many other nutrients including unsaturated fats, essential amino acids, folic acid and other B vitamins. Eggs are also low in saturated fat. The so-called omega-3 eggs from chickens fed on a diet that includes flax seed in addition have a high content n-3 polyunsaturates and in fact a very favorable ratio of n-3 to n-6 essential fatty acids. How deserving of blanket condemnation eggs really are has been addressed in a number of studies. They fall into two classes. First there are the studies that examine the relationship between dietary cholesterol and serum cholesterol. The other type of study actually examines the question of an association between egg consumption and heart disease and stroke. As regards cholesterol intake and serum levels, there have been a number of studies that find dietary cholesterol raises levels of total and LDL cholesterol, but the effects are very small, especially compared to the effects of saturated and trans-fatty acids (28). An interesting aspect of the cholesterol feeding studies is the observation that dietary cholesterol raises both LDL and HDL with little change in the LDL:HDL ratio. This provides another argument that dietary cholesterol is unimportant, except for the small minority who adversely respond to its ingestion (29). In fact there is a large variation in individual response of serum levels to dietary cholesterol, with many showing no change at all.

There have been 10 important prospective cohort studies of the effect of dietary cholesterol on the risk of heart disease. Only two provided a statistically significant indication that there was an increased risk, and these were not controlled for confounding from total energy or fiber (30). The two studies with corrections for confounding by total energy and fat or by total energy, fat and fiber both failed to provide a statistically significant indication of increased risk associated with an increment of 200 mg cholesterol per 1000 kcal caloric intake (about 400 to 500 mg/day). These two studies involved 37,851 men and 80,082 women. The study involving these two cohorts also looked at egg consumption in excess of one egg per day, and found no significant risk (28). Also, no risk associated with egg consumption was found for ischemic or hemorrhagic stroke. It is worth remarking that the eggs in these studies were not the new omega-3 eggs. These eggs are of recent origin and the studies in question involved a number of years of follow-up. In spite of these studies, the notion that eggs are associated with heart disease persists (31).

FAT AND CANCER

The above discussion makes it clear that it is hard to document a convincing case, in connection with heart disease, against any kind of fat other than trans-fat. But how about the notion that fat causes cancer? As mentioned at the beginning of this review, the same types of studies done twenty or thirty years ago that suggested a link between fat and heart disease also implicated fat with the incidence of cancer. These studies suffered from the same weaknesses as those that attempted to associate fat and heart disease. The conclusions from these early studies have not been confirmed by modern epidemiology. Unfortunately, the only extensive data relates to breast and colon cancer, although there is significant data concerning prostate cancer. Other sites have not been studied with the statistical power found in studies of cancer epidemiology of these three sites, so any discussion of diet and cancer is of necessity incomplete. As regards breast cancer, the clearest and most consistent finding is that high calorie intake, regardless of the food source, is far more important than dietary fat. In the Nurses' Health Study, which involved a large cohort studied over a long period, there was no hint of an increase in breast cancer with higher dietary fat. In fact those with a very low intake of fat (less than 15% of total calories) had a significant increased risk of breast cancer (3,32,33).

The early studies connecting colon cancer and dietary fat have also not held up against modern studies. The only connection appears to be with red meat, but no one knows if it is something in the red meat or has to do with chemicals produced on cooking. There is on-going research that is looking at this connection, especially the degree of "doneness" associated with cooking or broiling red meat. Again, the strongest link is with too many calories in relation to the exercise level--the problem of a positive energy balance (3,24,32,33).

The situation with prostate cancer is complex. The incidence and mortality of prostate cancer shows remarkable variation across geographical and ethnic groups, and changes in risk seen among migrants has inspired a search for dietary factors that might influence the development of this common disease. Fat is one of the most extensively studied dietary factors in this context. The current status has been reviewed by Moyad (34), Schulman et al (35) and earlier by Kolonel et al (36). Moyad reviews eleven prospective cohort studies involving over 200,000 subjects. While three studies showed positive association with red meat, the association failed to reach statistical significance. The one study that found a statistically significant connection between high-fat foods and prostate cancer has been criticized for an inadequate questionnaire and for the absence of high intakes of milk and beef that prevented the comparison with low intakes.

Case-control studies, on the other hand, have shown a connection between fat or fat-type food as well as alpha-linolenic acid and the incidence of prostate cancer. However, these studies are viewed as suspect because of recall bias and confounding. Also, prospective studies fail to confirm the positive risk connection with alpha-linolenic acid. Thus it seems clear that the association between dietary fat and the incidence of prostate cancer is very weak if not totally absent. However there is a positive association with red meat and dairy products and prostate cancer in individuals who have metastatic disease (37-40). In what appears to be the most recent large prospective study in this context, intakes of red meat and dairy products were not associated with total or advanced cancer (stage A2, B, C), but consumers of red meat had a significantly elevated risk of metastatic prostate cancer (stage D and fatal). A high intake of dairy products was also associated with an increased risk of metastatic disease, but the investigators were able to explain this association by known confounders (38). This study involved over 51,000 men followed for over 10 years. However, it is not clear if the connection with red meat is due to the meat per se or something in the meat or something associated with the cooking process. The association of prostate cancer with dairy products is thought to be confounded by the positive association with calcium intake observed is some studies of this disease (39,41). There does not appear to be any link with prostate cancer and vegetable fats (33).

With regard to other cancer sites, there is no significant evidence that there is a link between fat and cancer, but the data is very sparse (33,42).

Thus, if there is a relationship between dietary fat and cancer, it is very weak, and appears to be limited to animal fat and red meat. Limiting red meat and animal fat consumption (but not eliminating it) is part of many prudent diets, since replacing saturated fat with unsaturated fat in order to reduce the risk of CHD would normally involve decreasing the consumption of red meat and animal fat. Because red meat does in fact turn up in studies as a possible risk factor for both colon and prostate cancer, it may indeed be wise to limit consumption of this type of food until more definitive information is available. Willett holds to the view that the important factors in the relationship between diet and cancer appear to be a positive energy balance, reflected in part in early height and weight gain, early onset of menstruation for women and weight gain as an adult (24).

FAT IS BAD BECAUSE FAT MAKES YOU FAT

This is a popular concept that has a strong logical appeal to the layman. However, it appears to be an oversimplification of a complex subject. Viewed in its most general form, there is of course an element of truth in the statement that eating fat makes one fat. Total energy intake is related to the consumption of fat, protein and carbohydrate, and fat being more energy dense, can contribute to an energy imbalance where excess macronutrients are stored as fat. But this is not an argument for low-fat diets. One of the highest profile opponents of the fat makes you fat dogma is Harvard's Walter Willett. He points out that there has been a significant decline in the percentage of energy obtained from fat in the US population in the past two decades, as would be expected from the success of the anti-fat movement. However, at the same time there has been a significant increase in obesity. In short-term dietary intervention trials, individuals who are assigned to a diet that had a lower percentage of energy from fat generally had a modest reduction in weight. However, compensatory mechanisms appear to come into play, since in trials lasting over a year with fat consumption in the range of 18-40% (a very wide range indeed) of total energy, there appears to be little if any effect on the ultimate degree of fatness. Thus the available evidence points to the fact that diets high in fat do not appear to be the primary cause of the high degree of excess body fat in the US population, and reductions in fat will not be the solution (24,43).

If we turn to anecdotal evidence, the most successful long-term weight loss programs in North America, if not worldwide, appear to involve diets that are not low in fat but low or very low in carbohydrates, especially carbohydrates that are rapidly digested with concomitant adverse swings in both serum glucose and insulin. The diet programs with the highest profile include those of Atkins (44), the Eades (45), the Hellers (46), and Sears (22). Atkins pioneered the low-carbohydrate diet. While his anecdotal evidence is extensive with over 60,000 patients treated in his Manhattan clinic, nutritional scientists simply do not like anecdotal evidence, even when it is supported by thousands of case histories from other advocates of the same protocol. But progress is being made on this front. There are now two papers presented at national meetings that provide scientific evidence for the effectiveness and safety of the Atkin's type diet (47,48). There is also one study in the context of adolescent obesity (49). The Atkins' and other low-carbohydrate diets provide a counter argument to the fat makes you fat view since these are diets that have a higher percentage of fat than is common in North American diets, and patients are losing weight, not gaining it, without, for most individuals, adverse effects on blood lipid profiles.

Obviously, weight gain and obesity are complex subjects. However, there seems little doubt that there is a strong relationship with carbohydrate metabolism, blood levels of glucose and insulin, and the development of insulin resistance, which frequently accompanies obesity or the state of being overweight (18). To focus on fat alone would appear to be a serious mistake.

WHAT DO THE EXPERTS RECOMMEND?

Expert Panel on the Identification, Evaluation and Treatment of Overweight and Obesity in Adults (50)

In the context of strategies for weight loss and weight maintenance, they comment:

"Reducing the percentage of dietary fat alone will not produce weight loss unless total energy intake is also reduced. Isoenergetic replacement of fat with carbohydrates will reduce the energy fraction from fat but will not cause weight loss. Reducing intake of dietary fat, along with reducing dietary carbohydrate, usually will be needed to produce the energy deficit needed for an acceptable weight loss. When fat intake is reduced, priority should be given to reducing saturated fat to enhance lowering of LDL-C levels." Note: In this note, LDL-C is LDL.

Executive Summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on the Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) (51)

"ATP III's TLC (Therapeutic Lifestyle Changes) diet generally contains the recommendations embodied in the Dietary Guidelines for Americans 2000. One exception is that total fat is allowed to range from 25-35% of total calories provided saturated fats and trans fatty acids are kept low. A higher intake of total fat, mostly in the form of unsaturated fat, can help reduce triglycerides and raise HDL cholesterol in persons with the metabolic syndrome."

Note: The suggestion to increase fat intake is no doubt directed at the large population of individuals on low or very low fat diets that have high triglycerides and low HDL. Metabolic syndrome is also known as Syndrome-X.

Sense for Women. Your Plan for Natural Prevention and Treatment (52)

"Like many cardiologists, I used to recommend low-fat, high-carbohydrate foods to my cardiac patients. I was caught up in the low-fat, high-carbohydrate craze that swept the country ten years ago. Boy, was I off the mark! Many of my patients did initially lose weight on the no-fat, low-fat diets, but over time their HDL "good" cholesterol decreased and their triglycerides shot up, and they often regained weight."

Note: Stephen Sinatra, M.D., is Director of Medical Education at Manchester Memorial Hospital and Assistant clinical Professor of Medicine at the University of Connecticut School of Medicine. He is the author of several popular books.

Eat, Drink and Be Healthy. The Harvard Medical School Guide to Healthy Eating (24)

"PUTTING IT INTO PRACTICE:

• Remember that not all fats are bad--unsaturated fats protect against heart disease and other chronic conditions.
• Make decisions about dietary fats based on their proven impact on heart disease, not by their weak- -if any--connection with cancer.
• Limit the amount of saturated fat in your diet, as the American Heart Association, National Cholesterol Education Program and others recommend. But there is no good evidence that replacing saturated fat with carbohydrates will lower rates of heart disease, while there is solid proof that replacing saturated fat with unsaturated fat will.
• Reduce saturated fats by limiting the amount of full-fat diary products you eat and replace red meat with nuts, legumes, poultry and fish whenever possible.
• Use liquid vegetables oils in cooking and at the table.
• Eat one or more good sources of n-3 fatty acids every day--fish, walnuts, canola or soybean oil, ground flaxseeds or flaxseed oil."

Note: Walter C. Willet, M.D., Dr. PH. is one of the world's leading nutritional epidemiologists. He is a professor of medicine at Harvard University and chairman of the Department of Nutrition, Harvard School of Public Health. He has co-directed or directed a number of large and highly significant studies on the relationship of diet and health. He is the author of the definitive textbook on this subject (3).

FINAL REMARKS

This discussion should have made it clear that conducting studies that yield clear-cut answers is not the norm in this area. In fact the attempts to examine the question of the link between fat and heart disease have been remarkably inconsistent, when viewed as a whole. There have been a very large number of studies, especially if one counts those dating back to the 50s, and yet if one searched for one word to characterize the whole lot, it would probably be "inconclusive." Also a recurrent observation is that prospective cohort studies are at odds with case-control studies. While it is suspected that this is due, in case-control studies, to recall bias and failure to correct for confounding factors, this explanation can hardly be considered well established. Even the dietary intervention studies that replaced saturated fat with unsaturated fat were in a sense inconclusive, since two variables were simultaneously changed, but in this case, the overall result was clear enough--in fact convincing enough to become a popular recommendation from many authorities. It should also be apparent that by selecting studies, it is possible to support practically any point of view. Unfortunately, this is not a totally unknown practice(53).

As was mentioned above, there are many puzzling features of the cholesterol--fat--CHD picture, especially in the context of the initiation and progression of atherosclerosis. The most important "new vista" that has opened up in the past decade associates inflammation with atherosclerosis and the risk of heart attacks (54). This has become an exceedingly active area of research, and inflammation may well push cholesterol off its long held position at center stage. Research now suggests some association of inflammation with the adverse aspects of hypertension, diabetes, obesity and abnormal blood lipid levels as they relate to heart disease (54). In fact, the leading researchers in the field of inflammation and CHD predict that quite soon the measurement of serum C-reactive protein will become as routine as cholesterol level measurements. There are also other markers of inflammation that are currently under investigation in connection with CHD, and in addition, a lot of attention is being given to the role of chronic infections (e.g. gingivitis, prostatitis, bronchitis, etc.) in the formation and acceleration of atherosclerotic lesions.