(notes: references not included; the letter "omega" may be missing due to Internet conversion problems) Visit our key definitions for abbreviations.
It is well established that increased dietary polyunsaturated fatty acids (PUFAs) reduce Total/HDL Cholesterol (TC/HDLC). Dietary intervention studies have provided evidence that eating PUFAs and fish oils reduces the biochemical risk factors associated with cardiovascular disease. Optimizing lipoprotein levels with diet or drugs regresses CAD and reduces the risk of angiographic progression, a risk reduction attributed primarily to a reduction in TC/HDLC (unpublished communications by other authors at recent scientific meetings).
It is not known whether substantial reductions in TC/HDLC, similar to those achieved with drugs, can be achieved with the EFAs linoleic acid (18:2n-6, LA), alpha-linolenic acid (18:3n-3, ALA) or fatty acids found in fish oil (ALA n-3 derivatives). A computer search found no published data on the relationships between plasma fatty acids and TC/HDLC levels, although many equations have been described linking dietary fatty acids to cholesterol. The problem with dietary data alone is that it is impossible or impractical to accurately track dietary fatty acids, conversion of calories to PUFA or other fats, calorie utilization, or metabolism. Thus, dietary intake data is only a rough approximation of PUFA consumption and accumulation, even when subjects comply with experimental requirements.
Since the 1960's, Kingsbury et al, among others, have found fatty acid abnormalities in patients with different types of atherosclerosis and myocardial infarction. Serum fatty acid patterns have been shown to be good predictors of myocardial infarction, and low dietary intake of linoleic acid correlates with the predisposition to myocardial infarction. Moreover, among several variables, linoleic acid in serum total lipids was found to be the single most effective predictor of cardiovascular death in postinfarction middle-aged men. Lower levels of PUFAs were found in phospholipids of subjects with sudden cardiac death.
In addition to preventing or correcting physiological abnormalities that may arise from low plasma PUFA levels, correcting abnormally low levels of PUFAs with dietary PUFAs has additional clinical applications if it can be shown to reduce cardiovascular endpoints (such as myocardial infarct, angina, stroke, and hypertension). Lowering TC/HDLC with diet or drugs can prevent heart disease morbidity and mortality and regress atherosclerotic disease.
Using food to reduce the TC/HDLC ratio, considered one of the best surrogate indices for heart disease risk, would have implications for health care costs. Lowering TC/HDLC with drugs can cost more than $2,000 per year. With more than 50,000,000 Americans having abnormal cholesterol, treating everybody with drugs would cost more than $100 billion per year!
In a cross sectional study of about 100 subjects, criticized as having a small sample size, Siguel and Lerman found an inverse relationship between PUFA and TC/HDLC. To confirm and expand those results, we report here the analyses of 519 samples from the Framingham Heart Study. In a recent study, Mantzioris et al fed 30 subjects n-3 or n-6 fatty acid-rich diets. They found what appeared to be a "neutral effect on cholesterol". We analyzed both sets of data using a different model to characterize fatty acid status and found that increasing low PUFA levels with dietary PUFA drastically decreased TC/HDLC in the manner we predicted(12). These relationships have major implications for nutrition policy and medical care.
These results show that in cross section (different subjects, one time blood measurement), as well as in longitudinal studies (same subjects on different diets for different time periods), TC/HDLC declines when PUFAs increase, in a direct and highly linear inverse manner. Together, these results provide evidence that plasma PUFA levels have a significant effect on TC/HDLC, and that TC/HDLC levels can be modulated by changes in dietary PUFAs. A related study feeding diets low or high in fat and high in PUFAs found similar relationships, with subjects eating the low-fat diet having higher TC/HDLC and lower PUFAs than subjects eating a high PUFA diet. Although not shown, both the percent and concentration of SFAs and MUFAs are positively correlated with TC/HDLC in the Framingham samples. Future reports will describe the relationships between individual lipids, the percents and concentrations of specific fatty acids, age, sex, height and weight.
Mantzioris et al reported a different model of the relationship between cholesterol and fatty acid profiles. Instead of comparing TC/HDLC, the authors compared total and HDL cholesterol. Comparing only changes in cholesterol across treatment groups, as done by the previous authors, is ineffective because individuals rarely follow the treatment diet as well as the investigators expect, and thus the expected group differences due to diet may not materialize when gauged by plasma PUFAs. Differences in exercise, environmental temperature, body metabolism, and eating behavior drastically affect the net balance of PUFAs in the body.
Blood lipid values are difficult to predict from dietary intake, because the same diet will produce different changes in body fatty acid composition in different individuals (it is impossible to have identical life styles, etc.). Studies that evaluate the risk-protecting effects of different mixtures of dietary fatty acids or of low-fat diets ought to analyze subject's plasma fatty acid compositions. This would allow researchers to determine the net effects of carbohydrate and fat intake, and their interaction with exercise, food processing (i.e., hydrogenation), oxidation, and other variables affecting net cis EFA intake.
Dietary intake data may misrepresent what the subject actually eats. For example, n-3 intake is almost impossible to ascertain from dietary data, because the n-3 composition of oils varies greatly in foods-- even within the same brand (due to factors such as hydrogenation and cooking). Also, there is a lack of data on foods rich in n-3 such as peas and eggs. Food manufacturers are not required to maintain EFA composition constant (even on their own brands); the EFA composition of common foods such as eggs or chicken varies according to the feed used, which in turn depends on global market prices for corn (no n-3), soy (has n-3) and other chicken feeds.
Thus, it is hazardous to explain physiological changes only on the basis of presumed dietary intake; it is essential to do plasma fatty acid analysis to obtain a better gauge of changes in body fatty acid composition. Plotting TC/HDLC vs. plasma PUFA values relates two physiological variables measured on the same sample and can be a better indicator of the effects of dietary fatty acids on plasma lipids.
In our experience, whole plasma is the best single indicator of EFA status(18). Whole plasma can identify subjects with EFA abnormalities and can monitor the correction of those abnormalities in response to treatment. The percent of PUFA in whole plasma varies between 41% (25 percentile) and 47% (75 percentile), with an average of about 44% in the Framingham Heart Study (Framingham subjects are neither healthy nor subjects with optimal lipid values). Total PUFA is often below 35% for patients with intestinal malabsorption, and above 50% for healthy subjects (Siguel, unpublished data). Given the ranges of PUFAs found in the US population, a 50% reduction in TC/HDLC to reach optimal ratios may be achieved by dietary manipulation (see Fig 1).
These results do not mean that all PUFAs have equivalent effects on TC/HDLC. The subjects in these feeding studies had substantially elevated TC/HDLC ratios. Their PUFAs were also substantially below desirable levels. When both n-6 and n-3 levels are low, most mixtures of n-6 and n-3 may decrease TC/HDLC. The effects of different n-6 and n-3 fatty acid mixtures are likely to be different for subjects less deficient in PUFAs. These matters require further study. Moreover, the ratio of 20:5n-3/20:4n-6, associated with platelet aggregation and thrombosis, is an additional risk factor for thrombosis or stroke, independent of TC/HDLC(12).
What is noteworthy is that the % of plasma PUFA accounts for most of the variability in TC/HDLC, as can be seen from Figures 1 and 2. These findings support the hypotheses of Siguel(12) and Sinclair: that PUFA levels are the most important factor in atherosclerotic and coronary heart disease.
As can be seen in Table 1, HDL increases and triglycerides decrease when PUFA increases. This was predicted by Siguel and Lerman, who proposed that "plasma cholesterol and triglyceride levels are determined or regulated, in great part, by EFA metabolism, which is in part determined by the total amounts of each type of fatty acid in the body"(12).
Some scientists argue that most Americans eat plenty of EFAs. In an article commenting on Siguel's research, Grundy was quoted by Time magazine as stating that there is no reason to believe that EFAD is widespread. Dr. Grundy told me that he was referring to severe EFAD. Although many Americans may eat enough EFAs, a significant proportion do not, and could benefit from improving their EFA status.
EFA requirements should be established as grams/day rather than as a percent of calories, because the body's need for EFAs is more likely to depend on the number of cells in the body than on caloric intake. People who eat enough w6s often eat so much saturated fat and calories from carbohydrates that their body % of PUFA is low and their cells cannot get enough EFAs(12).
Siguel published fatty acid distributions from approximately 200 samples from the Framingham Heart Study(21). Different criteria can be used to judge EFA status. We may consider the 5%, 25%, 75% and 95% levels of a population sample such as the Framingham Heart Study. We can deem a subject to have EFAI when indicators of EFA status place him outside values of a reference population. Or we can consider as "healthy" the values of a group of subjects with lipid levels within currently established "normal" guidelines. According to Siguel, more than 25%(21) of the Framingham subjects had T/T > 0.02, indicating n-6 EFA deficiency(18).
We hypothesize that low calorie diets reduce TC/HDLC in many overweight subjects by burning SFAs preferentially over PUFAs, thereby increasing the % of total plasma PUFAs. This effect is most pronounced when overweight subjects with significant adipose tissue reserves of PUFAs lose weight.
However, when normal weight individuals maintain their weight on low-calorie, low-fat diets low in EFAs, plasma PUFAs are likely to decline and TC/HDLC may increase(19,27). Replacing fat with calories from low-fat foods may decrease plasma PUFA levels and therefore increase TC/HDLC. Consistent with this hypothesis, a study by Schaefer et al found that low fat diets, without weight loss, lead to a substantial decline in HDL, increase in TG, and increase in TC/HDLC. During their study (about 3 months), 2/29 subjects (about 6%) developed evidence of heart disease; one had an infarct.
We recently reviewed two patients who had been placed on low fat diets by their physicians. One followed his diet for approximately three months, after which his total PUFA = 39.6%. In this period, TC/HDLC changed from 2.5 to 4 and triglycerides from 93 to 138 mg/dL. Another patient followed a low fat diet for more than 5 years. When measured, his TC/HDLC was 7.45, plasma PUFA = 29% (one of the lowest values we have seen for a patient without intestinal disease causing fat malabsorption), T/T = 0.036, and 16:1n-7 = 3.9%. [In a healthy reference population, PUFA > 50%; T/T < 0.02; 16:1n-7 < 2.5%(14,18).]
These results suggest that very low-fat diets are effective to reduce TC/HDLC or triglycerides only if they increase the percent of PUFAs by decreasing body fat (i.e., with weight loss). This conclusion can be confirmed by measuring fatty acid profiles in studies using low-fat or low-calorie diets. Because the EFAs are essential nutrients while the body makes lipoproteins, cause and effect can be determined. We hypothesize that, in most people, low plasma levels of EFAs are caused by insufficient EFA intake, while lipoprotein abnormalities are caused by EFA abnormalities. However, in a small % of individuals with significant genetic abnormalities, or in the presence of diseases affecting lipid metabolism (such as insulin dependent diabetes mellitus or thyroid disease), the lipoproteins could become inefficient carriers of the EFAs. This would lead to EFAD in specific tissues, while EFA levels remained normal in plasma and adipose tissue.
There is overwhelming evidence that diets rich in PUFAs can lower TC/HDLC. PUFAs, particularly n-3 derivatives, have been proposed for the treatment of a variety of cardiovascular diseases. The active conversion of EFA precursors, ALA and LA, to their derivatives(14) suggests that many people obtain adequate n-3 and n-6 derivatives from ALA and LA, and that these fatty acids can produce drastic declines in TC/HDLC. Subjects with low body levels of EFAs or those who need EFA derivatives could be at high risk on a low-fat or strict vegetarian diet deprived of EFA derivatives. Reduced formation of EFA derivatives has been associated with insulin dependent diabetes, alcoholism, and aging. For these subjects, a low-fat or vegetarian diet deprived of EFA-derivatives could lead to higher TC/HDLC, higher TG, and perhaps an increased incidence of heart disease.
It is generally accepted that proper dietary treatment requires a mixture of n-3 and n-6 fatty acids and their derivatives, because ideal health depends on an optimal ratio of n-3/n-6 fatty acids. Siguel and Lerman(12) recommend a treatment diet that will bring the plasma fatty acid profile of a patient closer to the profile of a healthy reference population.
A trial period of an oil high in both 18:2n-6 and 18:3n-3 (i.e., soybean oil), together with appropriate antioxidants, may be the best way to raise EFAs and decrease cholesterol. After soybean oil supplementation for several months, fatty acid analysis helps to determine whether 18:3n-3 is sufficient to increase 20:5n-3 in blood, and lower triglycerides, or whether fish oils are necessary to lower triglycerides. One must also recognize that the dietary balance of n-3 and n-6 can be a modulator of eicosanoid activity, platelet aggregation(3), and incorporation of PUFAs in cell membranes(19).
Linear regression is a customary statistical method and is easy to interpret, but its value is limited to the range analyzed. Averaging TC/HDLC ratios reduces the influence of extreme points and narrows the range of PUFA %. Individuals can have plasma values of PUFA below 39% or above 48%, not shown in Figure 1. Obviously, Figure 1 does not imply that when PUFA = 60%, TC/HDLC = 0. The biological relationship between TC/HDLC and PUFA % most likely follows a curve that would plateau at TC/HDLC slightly below 2. Finding a more "biological" relationship would require further studies.
Individual patients have a great deal of biological "noise" (variability) in their measurements, and they may not respond to maneuvers which seem very plausible on a population basis. While increasing PUFA may lower TC/HDLC on a statistical basis, the results are less predictable on an individual basis.
Physiological and measurement variability in cholesterol and PUFA% may exceed 30%. Thus, the specific response of levels of TC/HDLC in one individual to dietary changes in PUFAs may not follow the relationship shown in this paper, particularly over a short time period. Reasonable PUFA intake over a period of several months is small compared with the large reserves of fatty acids in adipose tissue of most people.
Many physicians believe that drug therapy is needed because dietary changes alone are not sufficient to drastically reduce TC/HDLC (personal communications to the author by cardiologists and physicians at annual meetings of the Am. College of Cardiology and the Am. Heart Assoc). The American Heart Association supports the view that diet can only lower cholesterol by 15%, so drug therapy may be necessary.
This belief is derived from the fact that conventional treatment diets have relatively small amounts of EFAs (often below 10 grams/day), not enough to significantly change PUFA levels. Gronn(19) used 36 grams/day of n-6 and 6 grams/day of n-3 derivatives to achieve minor reductions in TC/HDLC and total TG over a 4 week period. Substantial changes in body fat composition may take years to accomplish by dietary change. Given these factors and the random variability in lipid and fatty acid measurements, individual reductions in abnormal lipids may not reflect "average" statistical results and perseverance is required for dietary changes to affect lipids.
This study shows that an inexpensive diet high in PUFA can drastically reduce TC/HDLC. Furthermore, eating PUFAs to bring plasma levels closer to those of a healthy population may improve life expectancy through mechanisms other than their effect on TC/HDLC. Siscovick et al have shown that reduced levels of n-3 PUFA in RBCs are associated with increased risk of cardiac arrest. However, patients who have low HDL and very high TC/HDLC (usually greater than 7), probably due to a genetic defect in making HDL rather than acquired factors (i.e., obesity or low PUFA), may not respond as expected to dietary changes in PUFA.
How much PUFAs to provide to subjects and the importance of achieving ideal weight in the context of a high PUFA diet needs further investigation. Overweight individuals may not reduce their risk of heart disease by eating more PUFA without losing weight. Increased weight imposes strain on the heart by mechanical forces independent of PUFA composition. Because PUFA-rich foods are obviously high in fat (calories), and most patients with increased risk of heart disease are overweight, increasing plasma PUFA levels often requires a significant reduction in food intake.
Treatment of cardiovascular disease through nutrition is less expensive than drugs or surgery, and it can be equally or more effective. Diagnosing lipid abnormalities through fatty acid analysis may lead to a new classification of dyslipidemia based on biochemical processes. Treating lipid abnormalities by manipulating PUFA levels may change preventive medicine approaches and reduce health care costs. More research is needed to explore the relationship between lipids, cis and trans fatty acids, and the biochemical or genetic abnormalities associated with low HDL.
Background. Dietary and plasma fatty acids have been linked to total cholesterol but not to the ratio of Total/HDL Cholesterol (TC/HDLC).
Methods. To evaluate the relationship between dietary and plasma levels of PUFAs and TC/HDLC, we analyzed cross-sectional and longitudinal data using 519 plasma samples (50% men, 50% women) from subjects participating in the Framingham Heart Study and results from a study feeding diets rich in either n-6 linoleic (LA) or n-3 alpha-linolenic acid (ALA) with or without fish oil supplements (n-3 derivatives).
Results. The values of TC/HDLC are inversely related to the percent of plasma PUFA when both variables are measured at the same time in different subjects, R = 0.82, p < 0.000001. PUFAs in phospholipids increase in response to increased dietary intake of different PUFA, either n-3, or n-6 or fish oils. There was a highly significant inverse relationship between TC/HDLC and the percent of PUFA in phospholipids, R = 0.97, p < 0.001. The relationship was similar regardless of the source and type of dietary fatty acids. A similar relationship existed when only the baseline points were considered.
Conclusions. When plasma PUFA % increases, either in response to a diet high in PUFA or across different subjects, TC/HDLC ratios decline. Evaluation of plasma fatty acid profiles and increased balanced dietary intake of PUFA to bring fatty acid profiles of subjects with low PUFA plasma levels closer to the profile of a healthy reference group is an effective approach to reduce high TC/HDLC. Reductions of more than 50% in TC/HDLC appear feasible with dietary modification alone. Further research into fatty acid metabolic activity may determine the biochemical basis of common dyslipidemias.
|
All information on this website is copyrighted; see use and permission to reproduce. The information in this website is not medical advice, merely a general scientific discussion. See warnings & disclaimers. |
|
|
© 1998 Edward N. Siguel. All rights reserved |
modified 9/12/98 |