Contents - Previous - Next

Cholesterolaemic effects of the saturated fatty acids of palm oil

Pramod Khosla and K. C. Hayes

Abstract

Dietary saturated fats are implicated as a major risk factor in hypercholesterolaemia and cardiovascular disease. Palm oil is a major source of the world's supply of oils and fats, but because of its relatively high content of saturated fatty acids (principally palmitic acid), its consumption has come under intense scrutiny over the last decade owing to potential health implications. Based on studies carried out more than thirty years ago, the hypothesis was developed that lauric, myristic, and palmitic acid were the three principal cholesterol-raising saturated fatty acids. Since palmitic acid is the most abundant fatty acid in the diet, the cholesterol-raising effect of all saturated fatty acids was accordingly assigned to it. However, recent studies from both humans and experimental animals suggest that not all saturated fatty acids are cholesterol-raising. When all dietary fatty acids are equalized, with the exception of the two being tested, palmitic acid appears to have no impact on the plasma cholesterol in normocholesterolaemic subjects when dietary cholesterol intake is below a certain critical level (400 mg per day). Only when cholesterol consumption exceeds this level, or when hypercholesterolaemic subjects are studied, does palmitic acid appear to increase the plasma cholesterol. These differential effects of palmitic acid on plasma cholesterol are thought to reflect differences in LDL-receptor status. Collectively these data imply that, for most of the world's population, palm oil would be an inexpensive and readily metabolized source of dietary energy with minimal impact on cholesterol metabolism.

Introduction

A substantial body of data implicates dietary saturated fat as one of the risk factors in hypercholesterolaemia and cardiovascular disease [1, 2]. However, the debate over what constitutes the "ideal" fat and, more specifically, its fatty acid profile has generated much controversy and confusion among both scientists and laymen. The subject is complicated further by economics. Since dietary fat is derived invariably from the consumption of various oils and meat and dairy products, advice from the scientific community affects production and distribution trends, and, in certain instances, specific national interests.

Palm oil is a major contributor to the world's supply of fats and oils and is arguably the most cost-effective source of edible fat. Because it has a relatively high content of saturated fatty acids compared with most other oils, the principal one being palmitic acid, the growing presence of palm oil in the world market-place has been the focus of much discussion over the last decade. Although palm oil (with other so-called tropical oils) has typically represented less than 3% of the total fat consumed in the United States, it is a major source of dietary fat in Latin American, South-East Asia, China, India and Pakistan, parts of West Asia, and Africa. Thus, emerging evidence on its metabolic impact based on carefully constructed scientific studies both in animals and in clinical settings will have far-reaching consequences affecting two-thirds of the world population. Due in part to palm oil's potential as a cost-effective source of fat in human nutrition, the scientific community must guard against intentional or unintentional bias and maintain a responsible perspective when reporting its findings or making recommendations concerning consumption of the oil.

It is apparent from recent data on lipoprotein metabolism in humans and animals that focusing on specific fat classes (saturated, mono-unsaturated, polyunsaturated) is a gross over simplification of the effect of dietary fat on cholesterol metabolism, including plasma lipoproteins. Even a superficial analysis of some of the so-called saturated fats (e.g., palm oil, lard, tallow, butter, coconut oil) reveals that they have distinct profiles (table 1) and empirically exert different metabolic effects. Accordingly, research in recent years has shifted toward elucidating the effects of specific dietary fatty acids in triglycerides, as opposed to specific classes of fats, on plasma lipids and lipoprotein metabolism. Detailed reviews on palm oil per se have been published recently [3-5]. Our purpose is to summarize the current knowledge concerning its impact on lipid metabolism from the perspective of its fatty acid profile.

TABLE 1. Fatty acid composition of so-called saturated fats (percentages)

Both RDB and red palm oils have fatty acid compositions similar to that of crude palm oil.

Dietary fats and serum cholesterol

Cardiovascular disease accounts for almost half a million deaths annually in the United States. One of the most easily measured indicators of risk is the serum or plasma cholesterol concentration, specifically the level of low-density lipoprotein (LDL) cholesterol, called the "bad" cholesterol. An elevated level of LDL cholesterol is a major risk factor. Conversely, an elevated level of high-density lipoprotein (HDL) cholesterol (the "good" cholesterol) is believed to confer protection. Hence, in any individual with elevated cholesterol, the primary goal is to lower the LDL cholesterol level to reduce the risk of cardiovascular disease. Although this objective may be achieved by drug therapy, one of the first interventions is dietary modification. Since 1908 we have known that diet affects serum cholesterol levels and, in the case of laboratory animals, the ability to develop atherosclerosis [6]. Although numerous dietary factors have been implicated on the basis of epidemiological studies, the single most important variable that has come under the most scrutiny is fat.

Saturated, mono-unsaturated, and polyunsaturated fats

Classification of fats has typically been based on their constituent fatty acids. Hence, fats in which the fatty acids with no double bonds (those most frequently encountered are 1218°C, lauric, myristic, palmitic, and stearic respectively) represent more than 50% of the total fatty acids are referred to as saturated; those in which the majority of total fatty acids have one double bond (usually oleic acid) are designated mono-unsaturated; and those in which fatty acids with two or more double bonds are the majority (usually linoleic acid) are referred to as polyunsaturated. Therefore, although two different fats may both be referred to as saturated, they may have distinctly different fatty acid profiles-for example, coconut oil, rich in lauric and myristic acid; palm oil, rich in palmitic acid; and cocoa butter, rich in stearic acid. The two most abundant fatty acids in nature are oleic and palmitic, which raises serious doubts that either would be considered detrimental to normal metabolic processes.

Since the 1950s numerous studies both in humans and in animals have investigated the effects of dietary fat saturation on cholesterolaemia [1, 2, 7-11]. The human studies were complicated by numerous variables, including the age and sex of the subjects, whether they were carried out in a metabolic ward or had free-living subjects, whether the subjects consumed liquid-formula diets or solid diets, and so on. The general consensus emerged, however, that saturated fats were twice as effective in elevating serum cholesterol as polyunsaturated fats were in lowering it. Mono-unsaturated fats were considered neutral (i.e., as having no effect on serum cholesterol). These observations led to a massive introduction of polyunsaturated fats in the marketplace from the 1950s, which doubled the typical polyunsaturated consumption between 1940 and 1985 from 2.5% to 5.5% of energy (en%) [12]. This rise in intake was associated with a peak in serum cholesterol and a decline in coronary heart disease [1].

Regression equations

Two independent research groups [8, 10] translated these early results into mathematical regression equations that have been used to predict the average change in serum cholesterol that might be expected for a given change in the percentage of energy consumed from a specific class of fatty acids. In addition, the equations included a cholesterol-elevating contribution from dietary cholesterol itself. These early studies also assigned essentially equal cholesterol-raising power to three saturated fatty acids, 12:0, 14:0, and 16:0, whereas the saturated fatty acids 10:0 and 18:0 were considered neutral. Even though an initial study and regression analysis showed myristic acid (14:0) to be four times as potent as palmitic acid (16:0) in raising serum cholesterol [5], a subsequent study with modified (transesterified) fat led to a revised opinion and to the labelling of 12:0, 14:0, and 16:0 saturated fatty acids as equivalent [13]. The fact that palmitic acid is the most abundant fatty acid in the food supply has meant that the cholesterol-raising property of all saturated fats has generally been attributed to their palmitic acid content. By the same argument, because myristic acid (and lauric acid) typically represent less than 2% of the energy in the American diet, their cholesterol-raising potential has been overlooked or dismissed as having any impact of consequence.

These studies focused on types of fats and oils, from which inferences were made about fatty acids and their ability to raise and lower serum cholesterol. We now have substantially more information about lipoprotein metabolism. This is important because the LDL:HDL ratio appears to be critical to the atherogenic potential of the lipoproteins. In theory it is conceivable that a proper balance in the fats (i.e., fatty acids) consumed will greatly enhance the circulating lipoprotein profile. In addition, the discovery of the LDL receptor [14] revealed a complex metabolic pathway that must be appreciated to understand fully the impact that dietary fatty acids have on lipoprotein metabolism. Although regression equations [9,10] have proved useful in attempts to sort out the saturated and unsaturated fatty acid effects on serum cholesterol among populations, they provide minimal information on how dietary fat effects lipoprotein metabolism, especially as it pertains to individuals. Also, it is now apparent that over the full range of potential 18:2 intakes (1-30 en %) the resulting decrease in plasma cholesterol may be non-linear.

Is mono-unsaturated fat as good as polyunsaturated fat?

In recent years a series of reports have suggested that the cholesterol-lowering potential of mono-unsaturated fat compares favourably with that of polyunsaturates, lowering LDL without lowering HDL [15,16]. This claim is surprising in light of data indicating that monounsaturated fatty acids are relatively neutral. We believe that total substitution with a mono-unsaturated fat in the experimental settings (which seldom occurs in Western diets because other fatty acids are present) does two important things: it potentially removes all the myristic acid from the diet, and it supplies more than enough 18:2 (about 4 en%) to maximize the LDL-receptor efficiency in the absence of 14:0 [16]. Our data suggest that 18:1 is not as effective as 18:2 when either 14:0 or cholesterol has down-regulated the receptors at a fatty acid intake below this critical 18:2 threshold.

In addition, on the basis of the above premise and as discussed previously [17, 18], it follows that any data obtained with a ratio of dietary polyunsaturated to saturated fat outside the normal range in the human diet (0.2-1.0) is likely to generate spurious results for the reasons stated. That is why feeding all the fat as safflower oil or coconut oil is not a legitimate, practical, or clinically meaningful evaluation of a saturated or polyunsaturated fat effect. Clearly, to derive valid information about the physiological impact of dietary fat, and specifically fatty acids, it should be fed at the levels that the body normally encounters.

Is palmitic acid cholesterol-elevating?

In an initial study with three different species of monkeys [19], tallow and lard (as saturated fats) were not much more cholesterolaemic than corn oil (a polyunsaturated fat) and were less so than other saturated fats, coconut oil or butter, even though both lard and tallow contain appreciable amounts of saturated fatty acids. Analysis revealed distinctly different profiles of the saturated fatty acids. This prompted us to question the generally held belief that the 12-16°C fatty acids were equivalent in terms of their cholesterol-raising ability. On further investigation with diets using blends of oils in which total saturated, mono-unsaturated, and polyunsaturated fatty acids were held constant, the exchange of dietary 16:0 for 12:0 + 14:0 [20] caused a decrease in the plasma cholesterol (table 2). This result clearly suggested that palmitic acid was not cholesterolaemic but neutral under those conditions, and that the widely held belief that all saturated fatty acids are the same was invalid. In a collaborative study, the same result was obtained in normocholesterolaemic humans, even with 300 mg of cholesterol in the diet [21]. Essentially similar results were obtained for hamsters fed blends of fats to control for specific fatty acids. Furthermore, the HDL cholesterol and the mRNA abundance for the LDL receptor were increased by 16:0 [22].

TABLE 2. Effect of exchanging 16:0 for 12:0 + 14:0 on lipid values in 21 monkeys of three species fed cholesterol-free purified

Adapted from ret 20.

Diets were formulated to give identical levels of total saturated, mono-unsaturated, and polyunsaturated fatty acids, with 16:0 increased at the expense of 12:0 + 14:0 in going from diet A to C.

Values are mean ± SEM.

* Means in the same column sharing an asterisk are significantly different.

In a subsequent study [23], monkeys were fed diets rich in either 12:0 + 14:0 or 16:0 + 18:1 and simultaneously injected with homologous ¹²⁵I-VLDL and ¹³¹I-LDL to assess apo B (and therefore very low-density lipoprotein [VLDL] and LDL) metabolism. Analysis of apo B specific activity data showed that monkeys fed the 16:0 + 18:1-rich diet had increases in the pool size of VLDL apo B and its transport rate and decreases in the pool size of LDL apo B and its total transport rate. The irreversible fractionated catabolic rate (FCR) for VLDL apo B and LDL apo B was similar between dietary groups (table 3). Although the total apo B and VLDL apo B transport rates were increased, LDL apo B concentration was reduced because of a decrease in the mass and proportion of LDL apo B derived independently of VLDL catabolism. This study further suggested that 16:0 is unlike 12:0 or 14:0, and clearly indicated that saturation of dietary fat has distinct effects on the transport of LDL apo B from VLDL-dependent and -independent pathways.

These studies [20, 22, 23] used cholesterol-free diets and normocholesterolaemic animals, and the results suggested that 12:0 + 14:0 is more cholesterolaemic than 16:0, and that 16:0 and 18:1 are neutral in terms of their effects on plasma cholesterol, as originally suggested [9]. A reappraisal of the literature, especially reports that developed the notion that palmitic acid was a cholesterol-elevating fatty acid, revealed two telling points. First, most studies used patients with mild (>220 mg/dl) to severe (>250 mg/dl) hypercholesterolaemia [9, 15], and many employed a design in which fat was fed in a background of dietary cholesterol. Thus, studies in both hypercholesterolaemic human subjects [15] and non-human primates fed cholesterol-containing diets [24] suggested 16:0 was hypercholesterolaemic compared with 18:1.

In both these situations some degree of down-regulation of the LDL receptor would be expected [14]. Since the LDL receptor, in addition to clearing circulating LDL, is responsible for clearing VLDL remnants [2, 25], when its activity is compromised it will fail to clear VLDL remnants. Consequently, the latter would be further metabolized to lead to an expanded LDL pool. However, when LDL receptor activity is not compromised by cholesterol feeding or other environmental-genetic interactions, as in our rhesus study [23], no expansion of the LDL pool occurs, since VLDL remnants are effectively removed to preclude their conversion to LDL. As a consequence, no elevation in plasma cholesterol was apparent when 16:0 was fed [19, 20, 22, 23].

As a working hypothesis, we suspected that in cases of normal LDL-receptor activity (i.e., in the absence of dietary cholesterol), 16:0 and 18:1 would exert similar effects on receptor-mediated LDL clearance. To test this, normocholesterolaemic cebus monkeys fed diets rich in 16:0,18:1, or 18:2 without cholesterol, were coinjected with radiolabelled native and methylated LDL (table 4). Receptor-mediated LDL clearance was similar for all three diets [26]. The total cholesterol was lower in cebus fed the 18:2-rich diet, but this was totally attributable to decreased HDL. The LDL concentrations and clearance were similar for all three diets. In fact, the 16:0-rich diet produced the lowest LDL:HDL ratio, significantly better than 18:2-rich diet, supporting our previous finding in hamsters [22] that 16:0 may be the saturated fatty acid responsible for the rise in HDL associated with saturated fat consumption. On the basis of our cebus data [26], we exchanged 16:0 for 18:1 in normocholesterolaemic humans (7 en%) and again observed no differences in LDL or HDL or total cholesterol, whereas increasing 12:0 + 14:0 caused a significant rise in LDL and total cholesterol [27], just as in our earlier monkey study [20].

TABLE 3. VLDL and LDL kinetic values in rhesus monkeys fed cholesterol-free diets rich in 12:0 + 14:0 or 16:0 + 18:1

Adapted from ref. 23.

Values are the mean ± SD of four monkeys per dietary group.

a. Values in the VLDL rows are transport to LDL; those in the LDL rows, from VLDL.
* Significantly different from the 12:0 + 14:0-rich diet.

TABLE 4. LDL kinetic values in cebus monkeys fed cholesterol-free diets rich in 18:2,18:1, or 16:0

Adapted from ref. 26.

Values are the mean ± SD of nine monkeys per dietary group. Figures in parentheses represent the percentage of the total FCR that is attributable to non-receptor- or receptor-mediated pathways.

Using accumulated data from the feeding of 16 different cholesterol-free fat blends, we generated regression equations (of the type originally developed by Hegsted) for the fatty acid impact on the plasma cholesterol response in our cebus monkeys [28]. The dietary 14:0 and 18:2 intakes alone were able to explain almost 92% of the observed variation in plasma cholesterol, with 16:0 and 18:1 appearing to be neutral. In view of our working hypothesis concerning the importance of the LDL receptor, we decided to re-examine the report in which the entire dietary fatty acid profile (not just saturated versus polyunsaturated) was published together with the cholesterol response in a large number of dietary manipulations (36 diets) using the same people [9]. Just as originally reported for all 36 diets, we found that 14:0 was four times as cholesterolaemic as 16:0, with 18:2 the only fatty acid that lowered cholesterol. However, based on our hypothesis that 16:0 was neutral when LDL-receptor activity was not compromised (e.g., by dietary cholesterol), we analysed the data at low (<= 300 mg) or high (>400 mg) cholesterol intakes. In the 17 human diets in which cholesterol intake was 300 mg or less, 85% of the observed variation in serum cholesterol could be explained solely on the basis of 14:0 and 18:2. However, in the 19 human diets containing more than 400 mg cholesterol, 16:0 appeared slightly cholesterolaemic. We now also have accumulated data from normocholesterolaemic gerbils fed a total of 33 cholesterol-free diets [29]. Again, 14 0 and 18:2 explain almost 90% of the observed variation in plasma cholesterol. Including dietary 16:0 and 18:1 in the regression failed to improve the predictability of the regression on the cholesterol response.

As an additional test of our hypothesis, we recently fed normocholesterolaemic cebus monkeys cholesterol-free diets rich in 16:0, 18:1, or equivalent amounts of 16:0 + 18:1. Again, no differences were noted in plasma lipid or LDL and HDL kinetic values among the groups [30]. Only when LDL receptors were down-regulated with dietary cholesterol (0.3% w/w) was a hypercholesterolaemic effect of 16:0 apparent (authors' personal observation), similar to the observations of others in monkeys fed similar diets [24].

Summary

Palmitic acid is best considered a transitional fatty acid; no apparent abnormality in cholesterol metabolism develops when energy flow is normal and fat is transported and cleared under normal physiological circumstances. Circumstances may develop, however, as in hypercholesterolaemic persons, wherein lipoprotein production or clearance becomes impaired, such as obesity and hyperinsulinaemia where compromised LDL-receptor activity is a factor. In such individuals palmitic acid may add to the cholesterolaemia because it represents the primary stimulus for fat transport as triglycerides among the fatty acids, thereby contributing to the pool of lipoproteins that must subsequently be subjected to an impaired clearance process. Only then does 16:0 appear to have a negative impact on cholesterol metabolism [31, 32]. The inference is that for most of the world's population, in whom adequate energy consumption and not energy storage (adiposity) is the problem, palm oil represents an ideal, inexpensive, highly palatable source of energy in the food supply.

References

1. Committee on Diet and Health, Food and Nutrition Board, Commission on Life Sciences, National Research Council. Fats and other lipids. In: Diet and health: implications for reducing chronic disease risk. Washington, DC: National Academy Press, 1989:159258.

2. Grundy SM, Denke MA. Dietary influences on serum lipids and lipoproteins. J Lipid Res 1990;31:1149-72.

3. Cottrell RC. Introduction: nutritional aspects of palm oil. Am J Clin Nutr 1991;53(suppl):989S-1009S.

4. Elson CJ. Tropical oils: nutritional and scientific issues. Crit Rev Food Sci Nutr 1992;31 :79-102.

5. Wood R. Biological effects of palm oil in humans. In: Chow CK, ed. Fatty acids in food and their health implications. New York: Marcel Dekker, 1992:647-61.

6. Ignatovski AI. Influence of animal food on the organism of rabbits. Izv Imp Voyenno-Med Akad Peter 1908;16:154-76.

7. Ahrens EH Jr, Insull W Jr, Blomstrand R et al. The influence of dietary fats on serum lipid levels in man. Lancet 1957;1:943-53.

8. Keys A, Anderson JT, Grande F. Prediction of serum cholesterol responses of man to changes in fats in the diet. Lancet 1957;2:959-66.

9. Keys A, Anderson JT, Grande F. Effect on serum cholesterol in man of mono-ene fatty acids (oleic acid) in the diet. Proc Soc Exp Biol Med 1958;98:387-91.

10. Hegsted DM, McGandy RB, Myers ML, Stare FJ. Quantitative effects of dietary fat on serum cholesterol in man. Am J Clin Nutr 1965;17:281-95.

11. Keys A, Anderson JT, Grande F. Serum cholesterol response to changes in the diet: IV. Particular saturated fatty acids in the diet. Metabolism 1965;14:776-87.

12. Stephen AM, Wald NJ. Trends in individual consumption of dietary fat in the United States, 1920-1984. Am J Clin Nutr 1990;52:457-69.

13. McGandy RB, Hegsted DM, Meyers ML. Use of semi-synthetic fats in determining the effects of specific dietary fatty acids on serum lipids in man. Am J Clin Nutr 1970;23:1288-98.

14. Goldstein JL, Brown MS. The low density lipoprotein pathway and its relation to atherosclerosis. Ann Rev Biochem 1977;46:897-930.

15. Mattson FH, Grundy SM. Comparison of the effects of dietary saturated, monounsaturated, and polyunsaturated fatty acids on plasma lipids and lipoproteins in man. J Lipid Res 1985;26:194-202.

16. Mensink RP, Katan MB. Effect of a diet enriched with monounsaturated or polyunsaturated fatty acids on levels of low-density and high-density lipoprotein cholesterol in healthy women and men. N Eng J Med 1989;321 :436-41.

17. Hayes KC. Dietary saturated fatty acids and low density or high density lipoprotein cholesterol. (Letter to the editor) N Eng J Med 1989;322:402-04.

18. Keys A. Lowering plasma cholesterol by diet. (Letter) N Eng J Med 1986;315:585.

19. Pronczuk A, Patton GM, Stephan ZF, Hayes KC. Species variation in the atherogenic profile of monkeys: relationship between dietary fats, lipoproteins and platelet aggregation. Lipids 1991;26:213-22.

20. Hayes KC, Pronczuk A, Lindsey S, Diersen-Schade D. Dietary saturated fatty acids (12:0, 14:0, 16:0) differ in their impact on plasma cholesterol and lipoproteins in nonhuman primates. Am J Clin Nutr 1991;53:491-98.

21. Sundram K, Hassan AH, Siru OH, Hayes KC. Dietary palmitate lowers cholesterol relative to laurate and myristate in humans. Arterioscler Thromb 1991;11: 1614a.

22. Lindsey S. Benattar J. Pronczuk A, Hayes KC. Dietary palmitic acid (16:0) enhances high density lipoprotein cholesterol and low density lipoprotein receptor mRNA abundance in hamsters. Proc Soc Exp Biol Med 1990;195:261-69.

23. Khosla P. Hayes KC. Dietary fat saturation in rhesus monkeys affects LDL concentrations by modulating the independent production of LDL apolipoprotein B. Biochem Biophys Acta 1991;1083:46-54.

24. Rudel LL, Haines JL, Sawyer JK. Effects on plasma lipoproteins of monounsaturated, saturated, and polyunsaturated fatty acids in the diet of African green monkeys. J Lipid Res 1990;31:1873-82.

25. Havel RJ. The formation of LDL: mechanisms and regulation. J Lipid Res 1984;25:157076.

26. Khosla P. Hayes KC. Comparison between the effects of dietary saturated (16:0), monounsaturated (18:1) and polyunsaturated (18:2) fatty acids on plasma lipoprotein metabolism in cebus and rhesus monkeys fed cholesterol-free diets. Am J Clin Nutr 1992;55:51-62.

27. Ng TK, Hayes KC, deWitt GF et al. Dietary palmitic and oleic acids exert similar effects on serum cholesterol and lipoprotein profiles in normocholesterolemic men and women. J Am Coll Nutr 1992;11(4):383-90.

28. Hayes KC, Khosla P. Dietary fatty acid thresholds and cholesterolemia. FASEB J 1992;6:2600-07.

29. Pronczuk A, Liang JS, Hayes KC. Gerbils are more sensitive and less variable than hamsters in their plasma cholesterol response to dietary fatty acids and cholesterol. FASEB J 1992;6:A1104 (#972).

30. Khosla P. Hayes KC. Exchanging 10% dietary energy between palmitic (16:0) and oleic (18:0) acids fails to impact lipoprotein metabolism in cebus monkeys fed cholesterol-free purified diets. FASEB J 1992;6:A1361 (#2462).

31. Hayes KC, Khosla P. Dietary fats and blood lipids: misconceptions, emerging issues surrounding saturated fats. In: Anderson GH, Kakis G. eds. Proceedings of the symposium Optimizing Heart Health: The Diet Connection. Toronto: University of Toronto, 1992:2325.

32. Hayes KC, Khosla P. Pronczuk A, Lindsey S. Reexamination of the dietary fatty acid-plasma cholesterol issue: Is palmitic acid (16:0) neutral? In: Gold P. Grover SA, Roncari DAK, eds. Cholesterol and coronary heart disease: the great debate. Lancaster, UK: Parthenon, 1992:189-206.

Contents - Previous - Next