Summary of the Scientific Conference on the Efficacy of Hypocholesterolemic Dietary Interventions
Diet is the primary therapeutic approach for persons who are at increased risk of premature heart disease as a result of elevated cholesterol levels (>200 mg/dL). This conference was held to reassess the efficacy of hypocholesterolemic diets based primarily on reduced intake of saturated fat and cholesterol and to evaluate the effects on lipoproteins and cholesterol of dietary components other than cholesterol and saturated fat.
Cholesterol Lowering With Step I and Step II Diets
The Step I and Step II diets are designed to reduce dietary intake of saturated fat and cholesterol and thereby on average lower total and LDL cholesterol levels. The Step I diet lowers LDL cholesterol approximately 7% to 9%. The Step II diet can lower LDL cholesterol 10% to 20%. Available evidence suggests that further reductions in the range of ≥20% to 25% can be achieved with a very low fat diet (also very low in saturated fat and cholesterol). Individual responses can vary markedly. Many Americans may be unable to achieve the maximum cholesterol-lowering potential expected of these dietary recommendations and thus will require pharmacological intervention. This variation is the result of the interplay of many physiological variables that are not yet fully understood. Low-fat diets can result in reductions in HDL cholesterol and sometimes apolipoprotein (apo) A-1 as well as in plasma triglycerides. Scientists disagree as to whether or not this is harmful.
Factors Affecting Cholesterol and Triglyceride Responses
Several mechanisms have been proposed for fatty acid influences on lipoproteins: (1) saturated fat interferes with formation of cholesterol esters in the liver, which leads to an increase in unesterified cholesterol that suppresses LDL receptor activity; (2) a large change in polyunsaturated fat intake relative to saturated fat results in decreased apoB production; and (3) polyunsaturated fats reduce production of cholesterol-carrying lipoproteins. The mechanism of cholesterol response may vary, depending on the fatty acid (eg, lauric, myristic, and palmitic acids, each of which is hypercholesterolemic).
The issue of whether or not monounsaturated fatty acids in the diet are beneficial relative to polyunsaturated fatty acids remains unresolved. For lowering LDL cholesterol, the differences are small. Based on in vitro studies, monounsaturated fatty acids may have an advantage over omega-3 or omega-6 polyunsaturated fatty acids; for coagulation factors, omega-3 polyunsaturates might be favored over omega-6 or omega-9 polyunsaturated fatty acids. The AHA diet recommends that intake of polyunsaturated fatty acids equal up to 10% of calories. Some have questioned whether essential fatty acid deficiency can be prevented at this level. The case has not been made for a high degree of essential fatty acid deficiency in the US population. The new World Health Organization guidelines set a range of 4% to 10% for polyunsaturated fatty acid intake; the higher levels are for populations that eat higher amounts of saturated fat. The important point is the relation of one fat to another, not just the addition of monounsaturated or polyunsaturated fatty acids.
Randomization is a process by which the fatty acids of a triglyceride are rearranged so that every fatty acid is present in each position of the triglyceride at one third of its total concentration. In terms of absorption, the randomized fats are absorbed and diluted with the other fats. A randomized fat seems to produce little effect on cholesterol levels, but the manner in which the cholesterol is carried is different.
Dietary cholesterol raises plasma cholesterol, on average, a small but measurable amount in humans. Some persons are totally resistant, whereas others are highly sensitive. Responsiveness is inversely proportional to the baseline intake. No simple marker identifies the genetically hypersensitive individual. The large variability in responsiveness complicates the issue of numerical targets. Adaptation may lead to less lowering of plasma cholesterol levels by modified diets than predicted from short-term experiments.
The plant sterol family includes a number of compounds structurally related to cholesterol. Several (ie, sitosterol, desmosterol, campesterol, stigmasterol) possess the cholesterol nucleus but have double bonds and/or additional alkyl groups in the side chain. Other compounds (eg, ergosterol) differ from cholesterol in both nucleus and side chain. A number of feeding studies using plant sterols have been conducted in animals and humans in varying states of health. In most of these studies cholesterol levels were reduced.
Several mechanisms of hypocholesterolemic action have been proposed. These include formation of mixed cholesterol-sitosterol crystals, competition for absorption sites, influence on cholesterol-rich micelles, competitive esterification, enhanced cholesterol excretion, and interference with cholesterol biosynthesis. Little work has been done in this area.
Low-fat diets are unavoidably high in carbohydrates. The DELTA study (Dietary Effects on Lipids and Thrombogenic Activity) produced a variable, not statistically significant increase in triglyceride levels. This study, unlike other diet studies, indicated that low-fat diets lower some hemostatic factors. It is unclear what these hemostatic factors mean in relation to disease. Low-fat, high-carbohydrate diets raise VLDL and triglycerides and change particle size. However, if the diet results in weight reduction, hypertriglyceridemia may not be a problem. In addition, if the high-carbohydrate diet is high in fiber and complex carbohydrates and simple sugars are kept to a minimum, then triglyceride elevations in most persons are very small and may be clinically significant.
Hypocholesterolemic Effects of Other Dietary Variables
Dietary fiber includes the nonstarch polysaccharides (NSPs) found in plant foods that are predominantly associated with plant cell walls. NSPs differ in chemical-physical properties, based on their dispensability and water-holding capacity, viscosity, binding capacity, and fermentability by micro-organisms in the large bowel. These differences modify the physiological response to consumption of dietary fiber.
The results of investigations to determine the contribution of dietary fiber to lowering risk of cardiovascular disease have demonstrated that viscous polysaccharides can lower plasma cholesterol and LDL-cholesterol concentrations, independent of other dietary modifications. Most studies suggest a reduction of 3% to 5%. Various polysaccharides also slow absorption of lipid from the small intestine. One of the most important contributions of fiber to risk reduction is the lowering of total fat, saturated fatty acid, and cholesterol content of the diet.
Replacing some carbohydrates in the diet with protein may have a beneficial effect with respect to lipid levels. HDL levels can increase and triglycerides can be reduced.
There is clear evidence that the amount and type of dietary protein can influence the levels of total cholesterol and LDL cholesterol in animals. The differing cholesterolemic effects of animal and plant proteins seem to be due largely to their amino acid composition. Some essential amino acids, such as lysine and methionine, which are generally higher in animal proteins, produced a hypercholesterolemic response in rabbits, whereas arginine, which is generally higher in plant proteins, appeared to be hypocholesterolemic. The hypercholesterolemic effect of dietary amino acids is associated with downregulation of hepatic LDL receptors.
In these animal studies the dietary effect of protein was not entirely independent of fat and cholesterol intake because a relatively low fat intake was used. With a high saturated fat intake, the experiment could demonstrate the difference between casein and soy protein; with a very low-fat diet the effect can be demonstrated quite readily in rabbits; in other species, eg, the pig, cholesterol is needed in the diet to demonstrate the difference.
In studies with soy protein and animal protein used at a ratio of 1:1, the lipid effect is the same as with all soy protein. With half and half soy/animal, the animal protein boosts the level of some B vitamins and trace minerals. Protein interacts with fiber in the diet. If the fiber is cellulose, soy protein has a greater cholesterol-lowering effect than does casein. If the fiber is alfalfa, the two become identical.
Soy protein also contains trace components, which may have some significant effects on risk for cardiovascular disease. The isoflavonoids are tyrosine kinase inhibitors and thus may influence cardiovascular risk, even without substantial effects on cholesterol levels. There is evidence that isoflavonoids are related to phytoestrogens, and they also may tend to reduce the clumping of platelets and perhaps reduce thrombus formation. They may also have some effect on what happens in the cell wall, because tyrosine kinases are related to cell proliferation. Soy flavonoids have also been implicated as being potent antioxidants.
The role of dietary protein in the control of blood cholesterol levels in humans is more controversial. Numerous studies have shown that substitution of soybean protein for animal protein in the diet can reduce the level of plasma or serum cholesterol, but the results have varied from >20% to little or no effect. The hypocholesterolemic effect of substituting soy protein for animal protein is more marked in hypercholesterolemic subjects and with diets containing high levels of protein. Some of the observed variations may also be due to interactions between protein and other dietary constituents.
Fat substitutes are classified into categories: emulsifiers, hydrocolloids, starch derivatives, hemicelluloses, β-glucans, soluble bulking agents, microparticulates, and synthetic fats.
The synthetic fat substitutes are the most similar in organoleptic properties and functionality of fats. The FDA recently approved one—olestra—for use as an additive in snack foods only.
The obvious advantage of fat substitutes is to reduce the total amount of fat in the diet while providing pleasant fatlike organoleptic properties. The major disadvantage of synthetic fat substitutes is their long-term safety and tolerability in humans, especially in binge situations. Some synthetic fat substitutes are not absorbed in the intestines and are excreted in the feces. This can result in abdominal discomfort, steatorrhea, diarrhea, and anal leakage. In addition, the potential for fat-soluble vitamin deficiencies must be carefully evaluated. Long-term studies are needed to confirm product safety and tolerability of high doses to ascertain their safety in the American diet.
Nontriglyceride Food Components
Many nontriglyceride components are present in vegetables, even fiber-rich foods. For example, tocotrienols, which are found in rice bran oil, have been shown to lower cholesterol and to inhibit tumors in vivo and in vitro. The form of rice oil is important. That found in the domestic market has been alkali-refined, a process that removes 80% to 90% of the unsaponifiables, thus reducing the ability of the oil to lower cholesterol.
Dietary calcium has been shown to modestly reduce total cholesterol and LDL cholesterol levels. Supplementation of a diet rich in saturated fatty acids with 2 g of elemental calcium has produced an average 18 mg/dL decrease in LDL cholesterol levels, and at a more moderate intake of saturated acids, an 8 mg/dL decrease in LDL. The exact mechanism of action is unclear; there is evidence that excess dietary calcium can minimally interfere with fat absorption, particularly saturated fatty acids, which require a longer time for absorption.
Physiological Conditions Affecting Cholesterol Response
Obesity and Dietary Response
The prevalence of obesity is rapidly increasing in the United States. It is estimated that approximately 33% of adults are obese. The problem is greatest in black women. There appears to be a consistent and graded relation between obesity, as determined by body mass index (BMI), total cholesterol, and LDL cholesterol.
High blood pressure, hyperlipidemia, and diabetes are more prevalent in the obese. Body fat distribution is important in assessing risk, especially the waist/hip ratio.
Studies suggest that diet responsiveness in women varies according to changes that occur in various life cycle stages. Factors such as the amount and location of body fat and the use of oral contraceptives and hormone replacement therapy also influence diet responsiveness. Low concentrations of HDL cholesterol seem to be a better predictor of risk in women compared with men, and high LDL cholesterol may be a less important risk factor for women before menopause and during hormone replacement therapy. Elevated triglycerides appear to be an independent risk factor in postmenopausal women, although not all scientists agree. Gender differences in diet response may vary with the diet strategy used. In addition, genetic phenotypes associated with coronary heart disease (CHD) risk may modify responsiveness to diet in men and women. Distinctions need to be made in interpretation of results from different types of studies.
Genetic variation can influence plasma lipid and lipoprotein responses to hypocholesterolemic dietary interventions and may contribute substantially to variations in dietary responsiveness within and among population groups. As yet there is limited information as to specific genes that modulate diet-induced changes in LDL cholesterol. As new human genetic variants affecting lipoprotein metabolism are identified, studies in large population groups, or perhaps in suitable animal models, will be necessary to determine their impact on dietary response and their interactions with other genetic, environmental, and hormonal factors.
Essentially all individuals with diabetes mellitus should follow an AHA/National Cholesterol Education Program Step I diet, and many should follow the Step II diet. The primary goal is to reduce saturated fat. In persons with insulin-dependent diabetes, the source of calories, after reduction of saturated fat, can be based on patient preference. In persons with non–insulin-dependent diabetes (NIDDM), the issues are not so clear. The recent American Diabetes Association guidelines could lead to confusion, suggesting lower carbohydrate intake for elevated triglycerides and lower fat for obesity when these occur concomitantly in the same patient. At this time the primary focus should be on decreased saturated fat and cholesterol in the diet for persons with NIDDM.
Modification of the AHA Diet?
Among conference participants it was generally agreed that any changes in AHA dietary recommendations should be carefully considered. Safety and prevention are two aspects of the recommendations that must be kept in mind. Although useful as a population goal, for those at high risk it is unlikely that cholesterol lowering achieved with a diet containing 8% to 10% of calories as saturated fat is sufficient. For these patients, significantly greater reductions in saturated fat (<7% of calories) will be necessary. Some studies show that a 15% reduction in total or LDL cholesterol can be expected only by reducing saturated fat to <7% of calories. Although very low-fat diets have been used in high-risk populations, it is questionable whether the data support low-fat intake, eg, ≤20% of total fat in the general population. Data are needed on the effect of such a diet on healthy women, children, and the elderly compared with high-risk men. Individual needs must be considered. Achievement and maintenance of a healthy weight should receive more emphasis, because the major nutritional problem in the United States is obesity.
The prudence of developing a lower fat or AHA Step III diet was considered. This has not gained favor as it places the emphasis on total rather than saturated fat. Rather than changing the total fat recommendation, the modification should be to reduce saturated fat intake to, perhaps, 5% of total calories. If this is done, a major task will be to identify individuals who are at risk. One possibility is to look at population-specific guidelines and then consider some sort of graded or specific target groups. The fat restriction may be synonymous with caloric density. For example, such choices in older people may be a significant determinant of caloric intake. More data are needed because whether and under what conditions a low-fat intake affects weight status is still very controversial.
For persons with low HDL cholesterol and high triglycerides, the population recommendation for diet is appropriate with some modification, depending on the type of triglyceride abnormality. For example, when dealing with hyperinsulinemia of “Syndrome X” phenomenon, restriction of refined sugar and initiation and/or maintenance of an exercise program are very important.
Optimal AHA Goal for Dietary Advice
If saturated fat and dietary cholesterol intakes are low, intake of fiber-rich foods is increased, and body weight is controlled, there will be an improvement in the cholesterol levels of most people, and risk of CHD will be reduced. Refinements in advice, such as the need for more physical activity (although exercise in terms of the LDL-cholesterol profile is effective only with a reduction in adiposity), making appropriate food choices such as eating more vegetables and fruits and becoming familiar with food labels for information on food composition, including caloric content of food choices, are desirable.
Cost-Effectiveness of Diet Versus Drugs
The issue of cost-effectiveness is complex: can an intervention be effective without having every individual see a dietitian at 3-month intervals, which is neither possible nor feasible. The real goal remains the lowering of the population's mean LDL cholesterol or total cholesterol. If the mean total cholesterol could be reduced from 205 mg/dL to 180 mg/dL, the number of persons with hypercholesterolemia would drop dramatically and the amount of atherosclerotic CHD would decline tremendously.
Impressive progress has been made in establishing the role of diet in lowering LDL-cholesterol levels. The papers presented at this conference suggest that greater responses can be achieved as a better understanding is gained of the other factors that are involved. The presentations at this conference together with other emerging evidence should provide an impetus to investigators to continue to pursue the question of how diet can modify CHD risk.
A single reprint is available by calling 800-242-8721 (US only) or writing the American Heart Association, Public Information, 7272 Greenville Avenue, Dallas, TX 75231-4596. Ask for reprint No. 71-0100. To purchase additional reprints: up to 999 copies, call 800-611-6083 (US only) or fax 413-665-2671; 1000 or more copies, call 214-706-1466, fax 214-691-6342, or E-mail firstname.lastname@example.org. To make photocopies for personal or educational use, call the Copyright Clearance Center, 508-750-8400.
- Copyright © 1996 by American Heart Association