Melissa H Olken & Joel D Howell. Cambridge World History of Food. Editor: Kenneth F Kiple & Kriemhild Conee Ornelas. Volume 1. Cambridge, UK: Cambridge University Press, 2000.
Over the course of the twentieth century, cardiovascular disease (CVD) has become the leading cause of death in the United States. CVD is also a significant cause of morbidity and mortality in many other industrialized countries and regions, such as Scandinavia, the United Kingdom, Australia, and Canada. Most CVD is manifested as coronary artery disease (CAD), usually based on atherosclerosis. This “epidemic” of CVD has been attributed to the poor lifestyle habits of members of late-twentieth-century industrialized, urban society, who smoke tobacco, exercise rarely, and indulge in fat-laden diets (Kannel 1987).
A striking similarity of these factors leading to disease is that each—in most cases—can be modified by an individual at risk for coronary disease, even without professional guidance and in the absence of public health initiatives. But risk factors are not always easily eliminated. Addiction to tobacco is difficult to overcome. Exercise may be problematic for some people, given constraints on time posed by other obligations. Everyone, however, must eat, and perhaps for this reason, of all the possible causes of heart-related diseases, diet has received the most attention.
This chapter explores the relationship between nutrition and heart-related diseases by describing selected nutrients that have been implicated in the pathogenesis, prevention, or treatment of CAD. Most of the available data come from population studies. It appears unlikely that any single nutrient will soon be identified as the specific dietary agent that causes atherosclerotic diseases. Moreover, any individual nutrient is but a small part of a larger group of chemicals that make up any particular “food.” As of this writing, the overall relationship between diet and heart disease remains obscure. It will require many years of medical, nutritional, and food science research before we can hope to weave into a meaningful tapestry all of the puzzling threads of nutritional biochemistry and CAD pathophysiology.
Pathogenesis of Atherosclerosis
In the middle of the nineteenth century, it was discovered that lipids constituted a major component of atherosclerotic plaque (Vogel 1847). This observation led investigators to wonder how this lipid collection was formed and how it might be related to disease. They noted that when people with CAD died, examination of the vessels that supply blood to the heart (the coronary arterial endothelium) revealed greasy, yellow “plaques” adhering to the endothelial cells. These plaques varied in thickness and composition, reflecting both the age of the subjects and the severity of their disease. A few decades into the twentieth century it was widely recognized that CAD lesions were due to atherosclerosis, with cell death within the lesion (Leary 1935). Additional clues to disease severity—and the longevity of the disease process—were later found in the presence of blood substances, such as lipids and macrophages, and in structural damage to the surrounding endothelium.
Noting precisely which endothelium is diseased may provide clues to the disease process. Although CAD is often conceptualized as a systemic disease, the location of atherosclerotic lesions is not random, nor are they necessarily “systemic.” Some areas of the blood vessels are more “lesion-prone” than others, and these sections differ structurally and functionally from “non-lesion-prone” sites (Schwartz, Valente, and Sprague 1993). Lesion-prone sites are more permeable to substances that lead to the development of CAD. Any adequate explanation of the disease process must explain these morphological and topographical features.
Although autopsy findings may serve to document an underlying disease process, they represent only one moment in the lifetime of the affected individual and cannot alone explain the natural history of the disease. One approach to understanding the disease process is to conduct experiments using an animal that exhibits a similar set of lesions. In 1943, D.V. Dawber and L. N. Katz studied chickens that developed atheromata and suggested that their findings might be relevant to spontaneous disease (Dawber and Katz 1943).
Pigs also provide a useful model for studying human atherosclerotic CAD—a model used to good effect by investigators in the 1960s and early 1970s. When pigs were fed a low-fat, cholesterol-free diet, researchers initially noted a degeneration of smooth muscle cells. As the pigs aged, the degenerated smooth muscle cells accumulated and caused intimal thickening. If there was no injury to the blood vessel during this thickening stage, the pigs experienced no disruption of the endothelium of the artery. However, if the investigators mechanically injured the endothelium, intimal thickening progressed to form a more substantial plaque—one that could impede blood flow within the artery.
This early plaque contained not only degenerated smooth muscle cells but also collagen, a fibrous tissue that made the plaque less compliant. Plaque thickening could be further accelerated (in concert with mechanical injury) with dietary factors, including oxidized sterols, such as vitamin D3 and 25-hydroxy cholesterol (a derivative of pure cholesterol). In advanced stages of atherosclerosis, lipid infiltrated the plaque. Other factors were found to influence development of the lesion. As the vessel lumen narrowed during plaque development, reduced blood flow through the area reduced oxygen delivery to the structures supplied by that blood vessel, and the hypoxemia thus produced sped up lipid accumulation in the plaque. As the plaque continued to thicken, the endothelium was stretched thin, sometimes thin enough to rupture.
The gap left by this process was then vulnerable to infiltration by lipoproteins. In animals that did not have the intimal injury, there was no progression past the intimal thickening. When investigators used electron microscopy to compare the experimentally induced atherosclerotic vessels of the mature pig with those obtained from a human who had undergone open heart surgery for CAD, they could detect no histological differences.
These animal studies were vital to our present understanding of the atherosclerotic process in humans. Recent research supports the concept that the arterial endothelium must sustain a mechanical injury to begin the process of pathologically significant plaque formation. Other blood components now known to be involved in this process are monocytes, platelets, and lipid-protein complexes, such as low-density lipoprotein (LDL) and lipoprotein(a). Also part of the process of plaque formation are chemical mediators, such as adhesive cytokines, chemoattractants, free radicals, and proteolytic enzymes.
Currently, most investigators believe that in early atherogenesis, monocytes (a type of white blood cell) are recruited to a “lesion-prone” area of the artery’s innermost wall, or intima. Before the plaque can be formed, the monocyte has to attach to the endothelium, a process orchestrated by various chemoattractant substances (for example, oxidatively modified LDL cholesterol) and adhesive cytokines (for example, interleukin 1-beta). Once attached, the monocytes migrate through the endothelium and attach to its underside, away from the portion of the endothelium that is in contact with the bloodstream.
In this new space, the monocytes are transformed into a different type of cell, known as macrophages. As long ago as the early 1900s, pathologists observed macrophages associated with mature atherosclerotic plaques, but their presence was believed to be incidental. In the late twentieth century, however, macrophages have been assigned a pivotal role in atherogenesis. The macrophage synthesizes a variety of substances that activate the inflammatory response in the affected area, such as oxygen free radicals, pro-teases, and lipases. These cause the macrophage to take up oxidatively modified LDL and change its appearance so that the cell has a foamy appearance when viewed with a microscope. The cells are thus called “foam cells.”
The next step in the process of atherosclerosis is necrosis of the foam cell, likely due to cytotoxicity from the oxidatively modified LDL. Smooth muscle migration and proliferation are mediated by a platelet-derived growth factor, a potent chemoattractant. Fibroblastic growth factors probably regulate smooth muscle cell proliferation. In the area of the plaque where there is the greatest density of macrophages, rupture may occur due to the high local concentration of macrophage-derived metalloproteases. This process ultimately results in blood clots that can obstruct blood flow—mural or occlusive thrombosis. Occlusion, whether temporary or not, creates a locally hypoxemic environment, which further enhances plaque growth. Oxidized LDL may initiate an autoimmune process that also adds to the inflammatory reaction occurring in the lesion-prone area.
The intricate atherosclerotic process provides investigators with many avenues of exploration for prevention and intervention. From a nutritional standpoint, the most obvious factors are cholesterol, dietary fat, and the biochemical precursors for the substances that may enhance or intervene in plaque formation: omega-3 fatty acids, protein, “antioxidant” nutrients (for example, vitamin E, beta-carotene, vita-min C), and other dietary substances. As an alternative approach to dietary prevention, one might consider dietary changes that would decrease the activities of the substances that enhance arteriosclerosis.
Early Epidemiological Studies
In the mid-1940s, Ancel Keys (1963) and William Kannel (1971) and their colleagues designed two large-scale studies to examine longitudinally what caused people to develop coronary artery disease. These investigators believed that by recording the essential characteristics of a large group of at-risk people and by comparing those who went on to develop CAD with those who did not, they could identify particular risk factors for the disease. These two sets of studies have had a profound influence on the development of the field.
Keys and his associates studied 281 Minnesota business and professional men who were 45 to 55 years of age and clinically healthy at the start of a 15-year study period. Each year, the men were given a detailed physical examination, with particular attention to the cardiovascular system. Examiners noted each man’s weight, relative body fatness, blood pressure, and serum cholesterol concentration.They measured two different pools of cholesterol in addition to total cholesterol: high-density lipoprotein (HDL) and low-density lipoprotein (LDL) cholesterol. During the 15 study years, 32 deaths occurred, 17 of which were directly attributable to CAD. Body weight and fatness were not predictive of disease, but the men in the upper part of the blood pressure distribution had a greater risk of CAD. Most compelling were the findings with regard to cholesterol. They indicated a direct relationship between blood cholesterol and risk of CAD, extending over all levels. The men who remained healthy had significantly higher HDL cholesterol than did subjects who developed CAD (mean 45.29 milligrams [mg] per deciliter as compared with 39.45). Detailed statistical analysis, however, failed to support the hypothesis that HDL cholesterol level was an independent predictor of CAD risk. Rather, high LDL and total cholesterol were more important predictors of CVD.
The Framingham Study (Kannel et al. 1971) followed 2,282 men and 2,845 women in Framingham, Massachusetts, for 14 years, starting in 1948. As in the Keys study, investigators were attempting to identify factors related to the onset of clinical coronary disease. Subjects were free of known coronary disease at the time of enrollment and were subdivided into groups based on their serum lipid content. After 14 years, 14 percent of the men and 6 percent of the women had developed some clinical manifestation of CAD. The incidence of CAD increased with age, and the baseline serum lipids and lipoproteins were higher in the CAD group than in the other subjects. Furthermore, the lipid profiles in the CAD subjects were high when compared to profiles in other parts of the world, such as France and Japan, where low CAD rates have been reported.
Investigators were unable to demonstrate that a particular lipid played a greater role than others. However, they did observe differences between men and women. In men and younger women (<55 years), the differences between those with CAD and those without the disease were more related to the total cholesterol. In older women (>55 years), prebeta lipoprotein (very low density lipoprotein or VLDL) discriminated better than total cholesterol between women with and without CAD. These two studies, coupled with what was known about atherosclerotic plaque composition, further supported the idea that one key to solving the mystery of atherosclerosis was to determine the relationship between blood lipids and plaque formation. Moreover, of all blood lipids, cholesterol appeared to be the most important.
Sources of Cholesterol
Cholesterol occurs naturally in eukaryotic cells (cells with a membrane-bound nucleus). In humans, cholesterol serves many vital functions. It acts as an integral part of the cellular membrane, serves as a chemical backbone for essential substances (for example, steroid hormones, vitamin D), and assists in digestion through its role in the formation of bile salts. Given the central role of cholesterol in biochemical and physiological functions, it is not surprising that the substance became a central focus for biochemical researchers. Although all human cholesterol was once thought to be ingested (Leary 1935), we now know that humans are able to manufacture cholesterol de novo, primarily in the liver and, to a lesser extent, in intestinal and other cells. Thus, even if one consumed a very low-cholesterol diet, the adult liver and intestine would still manufacture approximately 800 milligrams of cholesterol per day, which is enough for normal human functions. But cholesterol also reaches the body through dietary means for most humans, except those who are strict vegetarians and consume neither flesh nor dairy or egg products. Thus, cholesterol in atherosclerotic plaque formation could come from either de novo synthesis or dietary sources. The interaction between de novo cholesterol synthesis and dietary cholesterol, and the subsequent metabolism of cholesterol, is an intricate and fascinating phenomenon.
Serum cholesterol is transported bound to protein, such as apoprotein, along with phospholipids and other circulating fat-soluble compounds. These lipoproteins (or lipid:apoprotein molecules) are classified according to increasing density: From lower to higher density they are called chylomicrons, very low density lipoproteins (VLDL), low-density lipoproteins (LDL), and high-density lipoproteins (HDL). LDL typically contains 60 to 70 percent of the total serum cholesterol and HDL about 15 to 20 percent. The remaining cholesterol is carried in VLDL and chylomicrons.
The apoprotein portions contain biochemical signals that regulate the entry and exit of particular lipids at specific targets. For instance, chylomicrons transport dietary cholesterol, triglycerides, and other lipids from the intestine to the liver and adipose tissue. VLDL transports de novo synthesized cholesterol and triglycerides from the liver to adipose tissue. The residue is transformed into LDL, which is very rich in cholesterol. LDL moves cholesterol to peripheral tissues and regulates de novo cholesterol synthesis. Nonhepatic target cells possess specific LDL receptors that allow the cell to take up LDL cholesterol. The receptor number can be increased and decreased depending on the needs of the target cell. HDL presumably carries cholesterol from peripheral tissues back to the liver.
LDL regulates cholesterol metabolism through feedback inhibition. First, the cholesterol that is released from the LDL after it is inside the cell suppresses the key synthetic enzyme, HMG-CoA reductase, slowing de novo cholesterol formation. Second, when cholesterol concentration within a cell is adequate for its needs, the cell shuts down the manufacture of additional LDL receptors, thus preventing it from taking up more cholesterol.
Relationship of Cholesterol to CAD
Perhaps the clearest link between cholesterol, LDL, and arteriosclerosis comes from people with homozygous or heterozygous familial hypercholesterolemia. People who inherit both genes of this autosomal recessive disorder (homozygotes) have extraordinarily high levels of total and LDL cholesterol in circulation. Depending on the type of inheritance, they suffer from premature CAD either as a child or as a young adult. The defect in most cases is an absence or deficiency of functional LDL receptors, impeding the movement of LDL-cholesterol into target cells. Because LDL cannot satisfactorily get cholesterol inside a cell, there is no mechanism to turn off HMGCoA reductase, and cholesterol biosynthesis continues unchecked. The excess cholesterol, primarily found with LDL, is deposited in various tissues, including arterial endothelium. These cause myocardial infarctions at a very young age. Because dietary cholesterol restriction does not affect de novo synthesis, it is of very limited value alone in the management of these individuals.
Heterozygous familial hypercholesterolemia is relatively common (1:500 births), but is certainly far less common than CAD. Nonetheless, the association between cholesterol and CAD at extreme levels supports the idea that cholesterol is a factor in the development of CAD, particularly as it is consistent with numerous epidemiological studies that demonstrate a relationship between total and LDL cholesterol and CVD. Some of these are within-population studies, such as the Framingham and Minnesota studies discussed in the section “Early Epidemiological Studies.” Others compare different populations or people who move from one population to another. These studies often consider the importance of culture—particularly as manifested in what people eat—and, thus, bring us closer to the overall topic of this chapter and this work, which is diet.
The observation that different communities have different rates of CAD provides an important counterbalance to overly deterministic theories of disease causation, as well as engendering a series of ideas leading toward an understanding of the possible impact of diet on cholesterol and on CAD. C. D. De Langen, a Dutch public health physician, noted in the 1910s and 1920s that the incidence of CAD was extraordinarily low among residents of the island of Java (Snapper 1963). He explained part of this difference as being the result of differences in diet; Javanese stewards who worked on Dutch steamships soon developed a pattern of coronary health similar to that of the native-born Dutchmen who worked on the ships.
In 1941, I. Snapper made a similar series of observations about people native to northern China. He said that the difference could, perhaps, be attributed to the “equanimity of the Chinese,” but suggested that diet was probably a more important explanation. These observations led the American scientist Ancel Keys to wonder about the geographical variation in the incidence of CAD (Keys 1983). He initiated a massive study, published as Seven Countries: A Multivariate Analysis of Death and Coronary Heart Disease. This detailed work was begun in 1947 and eventually resulted in the study of a total of 12,763 men over 10 years in Yugoslavia, Finland, Italy, the Netherlands, Greece, the United States, and Japan. Conclusions are complex (on a scale with the study itself), but they show a clear relationship between the dietary intake of saturated fat and cholesterol and the incidence of coronary heart disease. Other between-population studies have shown that in countries where CVD rates are low, the serum LDL levels also tend to be low (for example, rural Japan and China compared to the United States and Finland).
Other investigations (following the lead of De Langen) have examined changes in CAD incidence with migration. Most have shown that when people move from a region with a low incidence of CAD to a region with a high incidence of CAD, and adopt the lifestyle of their new country, their likelihood of having CAD approaches that of the region to which they have moved. For instance, Japanese people living in their native Japan, in Hawaii, and in San Francisco, have an increasingly higher incidence of CVD: Age-adjusted rates among Japanese people were 1.6 per 1,000 person-years in Japan, 3.0 in Hawaii, and 3.7 in San Francisco (Kato et al. 1973). Saturated fat intake as a percentage of calories in the three populations was 7, 23, and 26 percent, respectively. The dietary trends among these populations emphasize the importance of the type of fat intake: Foods high in saturated fat also tend to be high in cholesterol. What happened to Japanese people moving from Japan to Hawaii to San Francisco seems clear enough. As they adopted a progressively more Western diet with a higher intake of saturated fats, their incidence of heart disease increased.
If the changing of diet results in a changing rate of heart disease, then public education efforts to heighten personal awareness of cholesterol levels may be serving as a broad and effective means of intervention. There are some encouraging signs. From 1980 to 1987, in the Minneapolis-St. Paul area, there was a significant decrease in total serum cholesterol and in the numbers of individuals with significant hypercholesterolemia (greater than 5.04 micromoles [mmol] per liter or 195 mg per deciliter) (Burke et al. 1991). However, despite the downward trend of cholesterol in the study population as a whole, many individuals (67 percent) with cholesterol levels high enough to require dietary and/or medical therapy remained unaware of their condition.
On a national level, the National Health and Nutrition Examination Survey (NHANES), undertaken between 1976 and 1980 (NHANES II) and from 1988 to 1991 (NHANES III), found an increase in the number of people with total cholesterol values less than 5.17 mmol per liter (200 mg per deciliter) and a decrease in those with total cholesterol values greater than 6.21 mmol per liter (240 mg per deciliter).
Although these population-based observations may reflect national public-health efforts to reduce total and saturated fat and cholesterol, such hopeful trends probably do not completely explain the decline in mortality from CVD. During the 1980s in the United States, there was also a decline in the use of tobacco, a major risk factor for CAD. In addition, there were changes in the diagnosis and management of CAD, including more widely available cardiac catheterization, angioplasty, functional cardiac studies (for example, dipyridamole and thallium stress echocardio-grams), and development and utilization of new medications (for instance, thrombolytics, cardiospecific beta-blockers, and angiotensin converting enzyme inhibitors). Thus, at present, the relative importance of population-based cholesterol lowering in bringing about a declining CVD death rate remains unknown.
Nonetheless, although many factors could alter mortality from CAD, the lowering of cholesterol would intuitively seem both important and appropriate. Prevention of CAD among a community of individuals can be viewed from two perspectives: primary prevention and secondary prevention.
Efforts to prevent persons from developing a particular disease fall under the rubric of primary prevention. In the case of CAD, primary prevention implies reducing the disease incidence for people with risk factors but who have not yet developed it. The Lipid Research Clinics Program (1984) and the Helsinki Heart Study (Frick et al. 1987) were large, randomized, controlled clinical trials that indicated that reducing serum cholesterol in persons without known CAD reduces the incidence of CAD onset. However, these trials were short (less than 10 years) and did not (or could not) demonstrate that primary prevention lengthens the life span by delaying or preventing the onset of CAD. Individuals who have hypercholesterolemia as their only risk factor for CAD likely do not add many years to their life by reducing their cholesterol level (Browner, Westen-house, and Tice 1991). Indeed, some studies have even suggested a rise in the incidence of violent deaths among individuals with the lowest cholesterol levels, although the relevance of such observations remains controversial.
On average, however, 3 to 4 years of life can be gained by individuals who have CAD and reduce their serum cholesterol by dietary and/or by medical means. This approach—intervention in the presence of known disease—defines secondary prevention.
The idea that diet has some relationship with heart disease, and that changing the diet can help people with heart disease, has been around for some time. What entered and left the body were long held as being critically important to its general functioning. Thus, it should come as no surprise to see comments about treating heart disease with diet from, essentially, the very first speculations about heart diseases in general (Fothergill 1776, cited in Leibowitz 1970). We should not attempt, however, to read these comments as having the same sort of specificity we now associate with statements about disease causation and diet.
For the early twentieth century, some insight into generally accepted notions of diet and heart disease may be derived by consulting a series of articles published in 1913 by the American Medical Association (AMA), the “cordial reception” of which led to their presentation in book form. The opinions expressed in these articles may be regarded as reflecting a general consensus, at least among allopathic physicians. Some of what was advised for the patient with heart disease seems logical (if not necessarily appropriate) to the late-twentieth-century reader, such as the suggestion that coffee should be avoided or that obese patients should lose weight in order to help the functions of their hearts.
Other approaches, such as administering hydro-chloric acid to aid in digestion, now seem clearly out of place. But what is most striking is the emphasis on the idea that diet should be individualized, at least in part because it is limited by “what the patient will do.” Rather than being dogmatic about what constitutes a “correct” diet, advisers would do better to tailor their recommendations to the specific social and cultural setting of the person under treatment.
Across the Atlantic, the noted British physician Sir Clifford Allbutt, Regius Professor at Cambridge University in England, was also concerned about individual variation in dietary suggestions (Allbutt 1915, particularly 238-54 of Vol. 1). He considered the question of overfeeding to be a relative one: What was a good diet for one person might be gluttony for another. He noted that cholesterol was associated with atherosclerosis, but he saw also the wide variation in the diets of people who developed atherosclerosis.
More recently, many trials have examined the value of cholesterol reduction in the management of CAD. These studies usually randomized patients with known CAD to receive some type of treatment (diet, drug, lifestyle change) versus “usual care” or placebo. The outcomes of interest in these studies tended to be occurrence of angina, myocardial infarction, death, and regression of atherosclerosis. Most subjects had total serum cholesterol concentrations greater than 5.00 mmol per liter. To determine regression or progression of atherosclerosis, baseline coronary angiograms were compared with those at the end of the study period. The angiograms were assessed with quantitative angiography and/or global score assessment. Both methods asked the same question: Are the areas of stenosis the same, less, or greater than at baseline, and is there a relationship between the extent of plaque regression and the amount of cholesterol lowering?
The Cholesterol-Lowering Atherosclerosis Study (CLAS) (Blankenhorn et al. 1987) investigated the effects of drug therapy in a randomized, placebo-controlled trial using men ages 40 to 59 years who had undergone coronary artery bypass graft surgery. After two years of treatment, patients who had been given the cholesterol-lowering drugs colestipol and niacin instead of placebos exhibited a reduction in total cholesterol, a reduction in LDL-cholesterol, and a significant amount of atherosclerotic plaque regression and preservation of native coronary arteries. However, the incidence of coronary events was no different between the two groups.
A fairly recent study by D. Ornish and colleagues (1990) explored the influence of lifestyle changes on reversal of CAD in a randomized controlled trial that followed men and women for one year. The treatment group experienced aggressive lifestyle interventions, including the imposition of a very low cholesterol, strict vegetarian diet. Their overall dietary fat was reduced to 10 percent of calories, far lower than in the usual American diet, which contains 35 to 40 percent of calories from fat. The treatment group demonstrated regression of atherosclerotic lesions and reduction of serum cholesterol and LDL-cholesterol.
The investigation has been criticized for flaws in randomization. Morever, the practicality of implementing such a drastic dietary intervention for the general population with CAD has also been questioned. However, the findings are provocative and suggest that a radical reduction in dietary fat can greatly influence serum cholesterol and plaque regression. The study subjects also received psychological and behavioral interventions and increased their exercise. Whether such changes can alter morbidity and mortality from CAD over one’s lifetime to a greater extent than more conservative, yet aggressive, approaches to cholesterol lowering remains to be elucidated.
The St. Thomas Atherosclerosis Regression Study (STARS) (Watts et al. 1992) focused on 90 men with CAD and mild hypercholesterolemia for approximately three years. It investigated the effects of usual care versus two interventions: a low-cholesterol diet and a low-cholesterol diet with the cholesterol-lowering drug, cholestyramine. Both interventions reduced the frequency of cardiovascular events and the progression of coronary artery narrowing.
The Monitored Atherosclerosis Regression Study (MARS) (Blankenhorn et al. 1993) was a randomized, double-blind, placebo-controlled study of 270 men and women with CAD that evaluated the effect of a low-cholesterol, reduced-fat diet with and without the HMG-CoA reductase inhibitor lovastatin. This trial demonstrated regression of atherosclerotic plaques; however, there was no difference in cardiac events between the treatment and placebo groups. At first glance, these results may seem to contradict the STARS trial, but the diet in the control group of the STARS study was not one designed to lower plasma lipids. In the STARS trial, both the lipid-lowering diet and the diet plus cholestyramine resin reduced coronary events to the same degree. They were significantly different from the usual care group, but not from each other.
The data on the role of cholesterol, particularly LDL-cholesterol, in the pathogenesis of CAD is compelling. Epidemiological population studies strongly suggest a positive relationship between elevated serum cholesterol and mortality rates from CAD. Primary and secondary prevention trials support the contention that reducing LDL-cholesterol, through a lipid-lowering diet with or without adjuvant medication, alters disease progression in middle-aged men. It is less clear if therapy will significantly reduce the incidence of subsequent cardiac events. The trials have not demonstrated an impact on longevity. However, it is certainly possible that future studies will do so. Studies in the Scandinavian Simvastatin Survival Study (known as 4S) seem to show an overall decrease in the death rate for people who already had heart disease when treated with potent cholesterol-lowering drugs (Scandinavian Simvastatin Survival Study Group 1994).
The influence of diet on CAD may reflect the amount of fat a person ingests. But the type of fat consumed may be just as important as the amount. Cholesterol is a type of fat found only in animal cells, not in plant cells. Low-cholesterol diets stress not only a reduction in dietary cholesterol but also a reduction in overall dietary fat, especially saturated fat. This approach is necessary because animal fat is laden with cholesterol. Animal fat is also primarily a saturated fat. Thus, the current recommendations by several national advisory panels are to decrease the consumption of saturated, cholesterol-rich animal fat, and to increase the amount of monounsaturated and polyunsaturated, or plant-derived, fat. Because poultry and fish contain more polyunsaturated fat than beef or pork, the recommendations also suggest consuming more poultry and fish as protein sources.
In dietary recommendations made in 1970 in the “Report of the Inter-Society Commission for Heart Disease Resources” (which were incorporated into “Dietary Goals for the United States by the United States Senate Select Committee on Nutrition and Human Needs”), the consumption of fat was expected to decrease from 42 to 30 percent of calories. Since 1909, Americans had consumed an average of 600 mg of cholesterol per day, but with these recommendations, cholesterol consumption was projected to decrease to 300 mg per capita per day. Such recommendations were considered radical at the time, and those who proposed diets based on them were concerned about the ability of patients to consume such a diet while maintaining normal protein nutrition.
More recently, however, the National Research Council’s 1992 recommendations have emphasized that in order to better meet the national goal of reducing total cholesterol concentrations to less than 5.04 mmol per liter, daily calories from fat should not exceed 30 percent of total calories, and two-thirds of that fat should be monounsaturated and polyunsaturated.
Saturated versus Unsaturated Fats
Can the rise in CVD mortality seen in countries like the United States and Scandinavia be attributed to the type of dietary fat, as well as to dietary cholesterol? The answer to the question seems to be yes. In the first half of this century, although U.S. CVD mortality rates rose dramatically, the increase in consumption of total and saturated fat was much more modest. However, during the same period, consumption of polyunsaturated fat rose two to three times (Friend 1967; Page and Marston 1979). In Europe during World War II, as wartime deprivation curtailed animal fat and, in fact, total fat consumption, the number of deaths due to CAD fell dramatically. In the 1970s, rural Romanians consumed 900 mg of cholesterol per day—50 percent more than Americans at the time—but Romanian CVD mortality rates were approximately 20 percent lower, perhaps because their dietary fat intake was 30 percent less than the average in the United States (WHO 1976).
Such observations implicate dietary fat, not just dietary cholesterol, in atherogenesis. Further, the nature of the fat (that is, saturated or unsaturated) may be important. Whether a fat is saturated or not has to do with how many hydrogen atoms are bound to carbon. Carbon atoms form the “backbone” of a fat molecule and can maximally bind to 4 other atoms. If 4 binding sites are used, the carbon atom is said to be “saturated.” If only 2 binding sites are used, the carbon atom forms a double bond with another carbon atom and is said to be “unsaturated.” Saturated fatty acids have no carbon double bonds, monounsaturated fatty acids have one carbon double bond, and polyunsatu-rated fatty acids have two or more carbon double bonds.
Epidemiological studies have observed that mortality from CVD is lower in southern European countries. There, total fat consumption is not remarkably low compared to other industrialized nations, but there is a greater consumption of monounsaturated fats, such as olive oil.
In 1928, two Chinese biochemists suggested that the low rate of CAD in China might be explained by the polyunsaturated fat linoleic acid (Snapper 1963). Others have later shown that this substance can lower serum cholesterol (Ahrens et al. 1959). In the United States, increased consumption of linoleic acid, primarily through greater consumption of corn oil products, parallels the decline in population serum cholesterol concentrations.
Fish oils appear to lower serum lipids in animals and humans (Bronte-Stewart et al. 1956; Nelson 1972). In the 1970s, epidemiologists attempted to explain a striking difference in the incidence of CAD between Eskimos, who rarely suffer from CVD, and Danes. They found a major dietary difference, not in total fat consumption but in the amount of fish oil consumption. Eskimos who have diets high in omega-3 fatty acids (found in fish oil) have prolonged bleeding times, as well as a decreased number and aggregation of platelets—all features associated with a reduced incidence of coronary thrombosis. Several hypotheses have been presented to explain the antiatherogenic effects of marine oils. They include altered plasma cholesterol and triglyceride concentrations, altered metabolism of prostaglandins and leukotrienes, and a variety of other physiological responses (Zhu and Parmley 1990).
Laboratory experiments have suggested possible mechanisms for the effects of marine oils. Animals fed a diet high in omega-3 long-chain polyunsaturated fatty acids and dietary cholesterol had a reduction in the number and size of atherosclerotic lesions. This may have resulted from chemical changes that affect how blood cells adhere to lesion-prone areas. Associated with these physiological changes were alterations in the prostaglandin synthetic pathway reflected by decreased thromboxane A2, increased prostacyclin, decreased leukotriene B4, and increased leukotriene B5. Most of these studies have also demonstrated significant lowering of plasma total cholesterol, LDL-cholesterol, and triglyceride levels. HDL-cholesterol has generally been unchanged.
Clinical investigations have shown that the consumption of fish oil in normal volunteers and in patients with hyperlipidemia can remarkably decrease plasma triglyceride levels, with inconsistent effects on plasma cholesterol and HDL-cholesterol. Perhaps as a result, fish oil supplements were widely marketed in the 1980s. But although epidemiological studies of Eskimos suggest a chemoprotective effect of fish oil, it is unclear that increasing fish oil in an American or northern European diet will impact CVD mortality rates. In addition, problems with fish oil supplements, including a bad flavor and unpalatable belching, pose compliance problems for patients with CVD risk.
Other Dietary Components
Sources of dietary protein include meat (for example, beef, pork, lamb, and wild game), poultry (such as chicken, turkey, and quail), fish, seafood, dairy products, eggs, grains, legumes (for example, peanuts, soy, and dried beans), nuts, and seeds. Depending on one’s culture, religion, ethnicity, and other socioeconomic variables, the primary daily protein source will vary. Different protein sources contain varying amounts and types of fat. The discussion in the section “Dietary Fat” has pointed out that saturated animal fat is most relevant to hyperlipidemia, unlike plant lipids, which are unsaturated.
Milk protein has been implicated as an important nutrient in atherogenesis for more than 20 years, most notably in S. Seely’s epidemiological analyses (1981, 1988). By reviewing the CVD mortality rates in men from 24 countries along with food consumption data, he found a significant correlation between CVD and the consumption of unfermented milk proteins, the only exception being cheese (1981). In 1988, Seely studied food consumption patterns of men and women in 21 countries that might be associated with CVD mortality. He observed significant positive correlations between CVD mortality and milk, milk products, sugar, and oats. There were negative correlations with fish proteins, vegetable proteins, and fish fat.
Proteins are made of amino acids, and one of those amino acids—arginine—has become the subject of intense scrutiny. The discovery in the 1980s and 1990s that mammals endogenously synthesize nitric oxide, a toxic substance, has led to an enormous body of literature describing the physiology, immunology, and biochemistry of this substance. It is produced by the enzymatic conversion of L-arginine to nitric oxide by nitric oxide synthetase. This reaction occurs in several tissue types, including endothelium, macrophages, and the brain. Nitric oxide produced in endothelium causes relaxation of the vessel and, hence, vasodilation.
L. Castillo and colleagues (1995) demonstrated that up to 16 percent of dietary arginine may be converted to nitric oxide in healthy volunteers. Several investigators have reported abnormal coronary arterial vasodilation, probably related to nitric oxide production in patients with increased cholesterol (Creager et al. 1990), hypertension (Linder et al. 1990), and CAD (Zeiher et al. 1991; Egashira et al. 1993). M. Jeserich and colleagues (1992) found significantly lower plasma L-arginine concentrations in patients with hypercholesterolemia (greater than 270 mg per deciliter), as compared to patients with normal cholesterol levels (less than 220 mg per deciliter). The relationship of these findings to CVD pathogenesis remains to be elucidated. It is unclear whether this difference in arginine levels is because of diet or because of the process of atherogenesis.
Dietary Carbohydrates and Fiber
Carbohydrate foods provide the bulk of daily energy needs. Animal studies have found that if the ratio of energy to protein is low, serum cholesterol tends to be lower. In humans, decreased energy intake may be associated with less obesity, a lower incidence of diabetes, and reduced cholesterol (total and LDL) (Kannel 1987).
Thus, modification of these atherogenic risk factors by modifying energy intake may be beneficial. The other main dietary energy source is fat, which provides about twice the number of calories per gram as carbohydrate does. Thus, to reduce overall energy intake and reduce atherosclerotic risk, it is desirable to obtain energy calories primarily from carbohydrate sources.
The association of fiber with hyperlipidemia has also received attention. Soluble fiber consumption is associated with reduced serum cholesterol. D. J. A. Jenkins and colleagues (1993) studied soluble fiber supplementation in conjunction with a lipid-lowering diet in 43 healthy men and women of normal weight. They observed a significant reduction in total cholesterol in both groups compared to baseline, but the soluble fiber group attained an even greater reduction.
“Antioxidant” nutrients have recently gotten a great deal of media and scientific coverage. These compounds include beta-carotene, vitamin E, and vitamin C. There are several reasons for this interest. LDL-cholesterol is oxidatively modified through chemical mediators released from the macrophage during atherosclerotic pathogenesis. Nonnutrient antioxidant substances (for example, probucol, butylated hydroxytoluene) inhibit the progression of atherogenesis in rabbits, presumably due to an alteration in oxidatively modified LDL-cholesterol (Bjorkhem et al. 1991; Mao et al. 1991). Antioxidants such as vitamin E work by “grabbing” a hydrogen atom. Having done so, they are now the oxidized molecule, and they have thus prevented another moiety, such as LDL, from becoming oxidized. Epidemiological studies suggest that people with high plasma levels of vitamin E have a lower risk of CVD. Similar results have been obtained when examining beta-carotene intake and CVD incidence. In the Nurses’ Health Study, CVD risk was reduced 30 to 40 percent in those individuals with high calculated intakes of vitamin E or beta-carotene. The data on vitamin C were unconvincing (Stampfer et al. 1987; Stampfer et al. 1993).
M. Abbey, P. J. Nestel, and P. A. Baghurst (1993) investigated LDL oxidation in nonsmoking men and women who were randomized to receive either a placebo or an “antioxidant” vitamin supplement (18 mg beta-carotene, 250 mg d-alpha-tocopheryl succinate, and 12 mg zinc) for 6 months. During the 6-month study period, the supplemented group’s LDL was less able to oxidize than the control group. I. Jialal and S. M. Grundy (1993) also demonstrated a decrease in LDL oxidation rate in men supplemented with vitamin E.
There are, however, several problems with recommending supplements of vitamin E or any other of the “antioxidants.” First, many of these vitamins have biochemically important interactions. For instance, vitamin C may reduce vitamin E needs by reducing oxidized vitamin E. The chemically reduced vitamin E can then be “recycled” and used again as an antioxidant. Second, no well-controlled randomized clinical trials have determined what dose of antioxidant may be needed to accomplish atherosclerosis prevention without adverse side effects. We do not even know which nutrient (or mixture of nutrients) should be tested. Further, it is not known if in vitro LDL oxidation is the best marker for what these nutrients may do in atherogenesis.
Thus, while it is probable that some of these nutrients play important roles in lipid metabolism (which may affect CVD), we simply do not know enough about them to make sweeping recommendations. If individuals would reduce their dietary fat to 30 percent of calories and increase their consumption of complex carbohydrates to 55 percent of calories (including fruits and vegetables), they would likely increase their ingestion of these “antioxidants,” with numerous overall health benefits and without risk.
Several studies suggest a possible benefit of alcohol ingestion. LDL-cholesterol levels appear to be lower in individuals who consume moderate amounts of alcohol: 1 to 3 bottles of beer, glasses of wine, or shots of liquor per day. At higher levels of alcohol ingestion, LDL is further reduced, probably due to the replacement of alcohol for energy calories at the expense of ingesting foods containing cholesterol and fat. Alcohol is the only substance humans ingest (nonpharma-cologically) that can raise HDL-cholesterol, a finding that probably explains the beneficial effect of alcohol ingestion in reducing CVD mortality. However, in addition to its beneficial effects on blood lipids, alcohol is a drug with significant potential for abuse. Thus, most public health officials, physicians, and scientists are concerned about advising alcohol as a therapeutic modality for hyperlipidemia and CVD prevention.
Miscellaneous Dietary Components
S. Warshafsky, R. S. Kamer, and S. L. Sivak (1993) examined five placebo-controlled randomized trials on the effects of garlic supplementation on serum cholesterol. About one-half to one clove of garlic per day decreases serum cholesterol by about 9 percent. Proposed mechanisms of garlic’s effects on blood lipids include increased bile acid excretion and reduced HMG-CoA reductase activity in the liver. However, carefully controlled studies of garlic remain to be done.
In concert with the antioxidant hypothesis, certain minerals have been implicated in atherogenesis. Selenium is able to act as a free-radical scavenger through its role as a cofactor in the enzyme glutathione peroxidase. However, the epidemiological studies on selenium consumption and CVD mortality do not bear out a probable relationship.
We have reviewed several dietary components that have been implicated in the pathogenesis of cardiovascular disease. The contribution of most nutrients appears to revolve around their ability to influence the serum lipid profile—total cholesterol, LDL-cholesterol, HDL-cholesterol, triglycerides, and the various apoproteins.
There is a need to look beyond serum lipids, however. Besides influencing a specific set of measurable laboratory parameters, nutrients must impact the pathogenesis of disease through biochemical and molecular alterations at the cellular level. This milieu of biochemical reactions and interactions is ultimately the most consequential to the dietary pathogenesis, prevention, and therapy of cardiovascular disease. We will doubtless continue to receive information about the important benefits of “nutrient A” on “function B,” and physicians, scientists, patients, and the public will try to make sense of these assertions.
Yet one must remember that nutrients are but small components of the whole chemical “package” that makes up a particular food or group of foods. Other properties of additional components in food may work synergistically or antagonistically, or act as innocent bystanders in cardiovascular disease chemo-protection and pathogenesis. It is unlikely that in years to come a single nutrient—will be uncovered as “the cause” of CAD. More likely, further investigation will describe the subtle interactions of various dietary components with an individual’s genome, in concert with other cardiovascular risk factors, such as tobacco use, diabetes mellitus, hypertension, hypoestrogenemia, stress, and obesity. Researchers probably will demonstrate that it is not just one or two nutrients or dietary factors that are critical, but the diet as a whole.
Consider an 88-year-old, generally healthy man who had a verified consumption of 20 to 30 whole eggs per day (Kern 1991). Despite this incredibly high daily cholesterol intake (12,953 micromoles), his total cholesterol was only 5.18 mmol per liter (200 mg per deciliter), and his LDL-cholesterol was 3.68 mmol per liter (142 mg per deciliter). It appears that he had compensated for his dietary excess by excreting excess cholesterol bound in bile acids in his stool. Thus, how an individual utilizes the diet biochemically may be as important as the diet itself.
We are not yet in a position to predict which people will do well with high cholesterol intake and which will not. In the meantime, it is prudent to follow the recent dietary guidelines of the National Research Council, which stress a diet containing 15 to 20 percent of calories from protein, 25 to 30 percent of calories from dietary fat, and the remaining calories from complex carbohydrates, such as whole grains, vegetables, and fruits.
These guidelines were put forward as death rates from CAD were falling, and there is no shortage of advocacy groups ready and willing to take credit for this reduction. But it is most probably the result of many lifestyle changes, among them declining use of tobacco (at least in the United States, when considered for all members of the population combined); better emergency medical services; improved inhospital treatments; and, of course, dietary modifications. As noted, change in diet has a particular appeal as an explanation because it is the only one of these factors that daily affects each and every member of society. And, as the new millennium dawns, it would seem that dietary intervention to prevent CAD has proved to be a successful means of secondary prevention for those people who have already incurred a cardiac event, even if the utility of such intervention as a means of primary prevention, applied to entire populations, remains controversial.
We should also bear in mind that the debate over diet and heart disease is being conducted in a very public arena, and an arena in which diet and heart disease are only one aspect of a more general tension over questions focusing on personal risk and responsibility. In the case of heart disease, these questions have stirred a broad-based popular reaction. No longer are discussions of diet and the heart confined to the pages of medical journals, but they can now be found in practically every issue of widely circulating newspapers and magazines. One good example is an extensive and detailed analysis questioning the overall importance of cholesterol in heart disease—indeed, dismissing “The Cholesterol Myth”—in the Atlantic Magazine (Moore 1989a). In the same year that this essay appeared, it was also published in expanded form as a book (Moore 1989b).
Public skepticism about the importance of diet in heart disease is also apparent in cartoons. One, in the New Yorker magazine (January 16, 1989: 39), shows a man starting to eat a huge oatmeal muffin. The caption reads: “Wellness update: Thirty-year-old man starting on the twenty-five-thousand-pound oat-bran muffin he must consume over forty years in order to reduce significantly his risk of death from high cholesterol.”
This and countless other less-than-reverent cartoons constitute widespread cultural markers that express a number of popular reactions to the diet-heart hypothesis. First, they ventilate frustration at the gospel of eating for one’s heart (or for health in general). And given a culture in which the average person is daily bombarded with images of food, with much of the easiest food to acquire and consume that which is likely to be least beneficial for preventing heart disease, this frustration is understandable. Such cartoons also reflect a general lack of conviction that altered diets will “work” against heart disease—a skepticism that primary prevention trials have thus far failed to address convincingly. At a lay level, most people know (or have heard of) someone who has lived to a ripe old age while engaging in near-constant dietary indiscretions, whereas someone who ate “right” and stayed fit may have suffered an early death. Epidemiologists might discount the relevance of such anecdotes, but they can have a major impact on popular perceptions of risk and disease, especially when the behavior that is being advocated may be neither easy nor (seemingly) pleasurable.
The popular reaction against dietary constraints also raises fundamental questions of personal versus public responsibility. If someone has CAD, whose fault is it? The choices people make about what to eat are limited by the cultural world in which they live, so to what extent is the larger society to be held responsible? And, if one is personally accountable for what foods are consumed, then should a “healthy lifestyle” make a difference in terms of how much one pays for health or life insurance?
Finally, what is the disease here? Is it CAD, or is it high cholesterol? We should not lose sight of the fact that blood lipids are merely surrogates for what is most important—the disability and death that come from CAD and CVD. Keeping track of these differences is important. First, it can help to maintain focus on the ultimate goals of therapy and not allow us to become sidetracked by that which is easier to effect. Second, it will prevent the labeling of a large percentage of the population as “diseased” simply because of a high lipid profile. The answers to the questions raised in this chapter are unlikely to come purely from the accumulation of more and more data, because the prevention controversies are also fueled by “hidden ideological, structural, and professional factors” (Aronowitz 1994).