![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
Medical College of Georgia, Augusta, Georgia
INTRODUCTION BACKGROUND DIET MODIFICATIONS AND HEART DISEASE RISK AND OUTCOMES INTERVENTION STUDIES WITH WALNUTS OBSERVATIONAL STUDIES ON CONSUMPTION OF WALNUTS AND OTHER NUTS AND CORONARY HEART DISEASE OTHER RELEVANT OBSERVATIONAL STUDIES MECHANISMS WALNUTS AS A FOOD COMPOSITION OF WALNUTS CONTROLLED INTERVENTION STUDIES WITH NUTS OTHER THAN WALNUTS PROTEINS IN NUTS AND AS A FACTOR IN HEART DISEASE RISK OVERVIEW AND SUMMARY OF CLINICAL TRIALS WEIGHT OF EVIDENCE CONCLUSIONS
 |
ABSTRACT |
|---|
 TOP Next |
|---|
KEY WORDS: • walnuts • coronary heart disease • nuts • legumes • polyunsaturated fatty acids
|
INTRODUCTION |
|---|
|
TOPPreviousNext |
|---|
Investigating the relation of diet to disease prevention can encompass evaluating foods for specific nutrients or nonnutrient chemicals, alone or in combination; however, whole foods, or their constituents, may have disparate effects. An example is the relation of fruits and vegetables and their content of antioxidants to the prevention of heart disease or cancer. Because the detailed composition of individual meals consumed by humans is not well documented, attention to individual foods is necessary. Dietary studies rarely enroll sufficient subjects to evaluate the effect of a single change of diet on cardiovascular morbidity or mortality. More often investigators use surrogate endpoints or biomarkers, eg, blood lipids and lipoproteins or blood pressure (BP) that are risk factors for vascular disease. Other risk factors such as cigarette smoking, obesity, family history, glucose intolerance, level of physical activity, age, and gender may be confounders. Atherosclerosis is a progressive disease of the arteries; its pathogenesis continues to undergo active investigation.
|
BACKGROUND |
|---|
|
TOPPreviousNext |
|---|
Research for more than 45 y has defined the relationship of CHD to total cholesterol (TC) and the major lipoprotein fractions. There is almost complete agreement that hypercholesterolemia [elevated TC and low-density lipoprotein-cholesterol (LDL-C)] is a risk factor for CHD. The ranges of concentrations that define the risk are agreed upon, and changes in risk of CHD in relation to changes in cholesterol concentrations have been demonstrated in prospective trials. Changes in the cholesterol concentrations of populations have been related to the incidence of CHD in numerous observational studies. The positive relationship of high-density lipoprotein-cholesterol (HDL-C) to the reduction of heart disease risk also has been defined. The contribution of genetic makeup to the lipid and lipoprotein risk factors is recognized conceptually. Any modulation of lifestyle may be limited by the inherent genetic makeup of the subject (see section on Fats and fatty acids in relation to blood lipids and lipoproteins and artherosclerosis).
CHD is the leading cause of death in the United States: 32% of females and 50% of males will develop CHD, and it is the cause of death in 24% of females and 31% of males. Increasing the serum TC level raises the risk of CHD and decreasing it will reduce risk. The risk of developing CHD is continuous over the range of serum TC levels, with moderate risk associated with levels exceeding 5.2 mmol/L (200 mg/dL), and high risk with >6.2 mmol/L (240 mg/dL). LDL-C levels can be classified similarly into low, moderate, and high risk. Coronary events are reduced by 2-3% for every 1% decrease in LDL-C, as reported in the Helsinki Heart Study (1
). Small, dense LDL particles are associated with a tripling of the risk of myocardial infarction (MI) compared to the larger, more buoyant LDL particles. An elevated level of lipoprotein (a) [Lp (a)] also increases the risk of atherosclerosis. The smaller Lp (a) particles are more atherogenic, and may act by promoting LDL oxidation and decreasing endothelial-dependent vasodilatation and by enhancing retention of LDL in the arterial wall. Lp (a) may also have antifibrinolytic activity, an independent risk factor for atherothrombosis (2
).
In contrast, HDL-C levels are inversely related to the risk of CVD. Risk is appreciably higher in subjects with HDL levels < 0.9 mmol/L (35 mg/dL), and risk decreases as HDL levels increase to > 1.6 mmol/L (60 mg/dL). The highest levels of HDL-C are considered a negative risk factor, ie, they reduce the risk of CVD. Coronary events are reduced by 3% for every 1% increase in HDL-C. The ratios of TC:HDL-C and LDL:HDL-C indicate varying degrees of the risk of CVD.
The results of a more recent multicenter randomized clinical study—the Veterans Affairs Cooperative Studies Program High-Density Lipoprotein Cholesterol Intervention Trial, as reported by Rubins and Robins (3
) and Rubins et al. (4
)—showed the benefit of raising HDL and lowering triacylglycerol levels, independent of effects on the LDL-C level. As a result of dietary intervention in this study, subjects showed no change in LDL-C, a 31% reduction in triacylglycerol, a 6% increase in HDL, a 22% reduction in CHD death and nonfatal MI, and a 29% decrease in stroke. The triacylglycerol level in blood is elevated by increases in energy, fat, carbohydrates, and alcohol. Recent data support the concept that higher triacylglycerol levels increase the risk of CHD, independent of HDL levels or other confounding factors of the dyslipidemic syndrome (ie, glucose intolerance, hyperinsulinism, obesity, and hypertension). This relationship applies especially to diabetic persons and females. Triacylglycerol levels > 2.6 mmol/L (100 mg/dL) increase the risk of CVD, independent of the usual accompanying low HDL. Fasting hypertriglyceridemia may be a stronger predictive risk factor than TC (5
), as reported in a meta-analysis of 17 population-based studies on triacylglycerol levels and CVD. The Physicians’ Health Study (6
) also supports triacylglycerol as an independent risk factor of MI.
Other predictive risk factors for CVD may include levels of fasting plasma insulin and apolipoprotein B (apo B), LDL particle size, and levels of circulating apoproteins. Absolute levels of apoproteins, or changes in lipoprotein particle size, or single amino acid mutations (as in apo-E isoforms) may be better predictors of CHD than lipid levels. It is not known whether these putative risk factors are responsive to dietary change or relevant beyond the specific subpopulations in which they have been evaluated, thus their value for planning dietary interventions is not clear. The relationship of apoprotein polymorphism to diet responsiveness is intriguing and under scrutiny. Elevated triacylglycerol levels may help identify individuals at high risk because of the associated predominance of small, dense LDL particles in these individuals. Elevated levels of insulin as a result of fasting are associated with impaired fibrinolysis and hypercoagulability in individuals with normal or abnormal glucose tolerance, thus enhancing the potential for acute thrombosis.
Although the relationship of the concentration and composition of blood lipids and lipoproteins to cardiovascular risk and atherosclerosis has been suspected for a century, the usefulness of various parameters in predicting disease or CHD events, or their responses to manipulations of lifestyle remain to be elucidated.
Fats and fatty acids in relation to blood lipids and lipoproteins and atherosclerosis
Current scientific consensus supports the value of using lifestyle modifications as a primary prevention mechanism to lower blood cholesterol in order to decrease heart disease risk. Not all of the population is responsive to such interventions, and CHD occurs in the presence of cholesterol levels within the "normal" range.
Feldman (7
) has surveyed the association of diet and CVD. Dietary fats, especially saturated fat (SF) and cholesterol, raise the levels of serum/plasma TC, LDL-C, and triacylglycerol that increase CHD risk. Components of the diet may affect HDL-C, the lipoprotein that lessens CHD risk. Antioxidants (eg, vitamins E and C) may lessen the risk of CVD by decreasing oxidized LDL, which is more atherogenic. High blood levels of the amino acid homocysteine (Hcy) are associated with increased atherosclerosis and are decreased by the intake of folate and vitamins B6 and B12 (see section on Factors and mechanisms other than fats or blood lipids that affect CVD risk).
From the decline in cardiovascular mortality that was observed due to dietary restrictions during the Great Depression and World War II, we can infer the relation of the fat content and the fatty acid composition of the diet to the pathogenesis of atherosclerosis. The dietary fat-heart hypothesis was proposed by Keys (8
) and Keys et al. (9
) and related CHD rates to the intake of dietary fat, especially SF. A recent 25-year follow-up report (10
) of the earlier study by Keys confirmed that the relative increase in CHD mortality due to a given cholesterol increase was similar in all cultures except Japan. A 0.52 mmol (20 mg) increase in serum TC corresponded to an increase in CHD mortality of 12%, increasing to 17% when adjusted for regression dilution bias.
Similarly, diets that are high in total fat (TF), SF, and cholesterol are atherogenic for many animal species. Cholesterol is found in the diet only in animal products and is not present in any plant sources. Long-chain saturated fatty acids (SFAs) in animal or vegetable fats raise plasma cholesterol levels and decrease LDL receptor activity. Paradoxically, these fats also raise HDL-C levels. Liquid vegetable oils with high concentrations of monounsaturated fatty acids (MUFAs) may have a cholesterol-lowering effect and decrease LDL-C, but not HDL-C; some monounsaturated oils may lower triacylglycerol. Polyunsaturated fats (PUFAs) of the n-6 series in liquid vegetable oils decrease LDL-C; increased amounts of PUFA may lower HDL-C. The n-3 series of very long-chain, more unsaturated PUFAs found in fish and fish oils have variable effects on TC, LDL, and HDL-C, and lower the triacylglycerol levels. Trans-fatty acids are produced with some processes of partial hydrogenation of unsaturated liquid vegetable oils (in the United States, predominantly soybean oil). These fatty acids raise LDL-C somewhat less than the long- or medium-chain SFAs of butter, but, in contrast, lower HDL-C.
Theories of atherogenesis propose that LDL-C is the offending agent, delivering cholesterol to the arterial wall and initiating the pathogenic sequelae including endothelial injury. Endothelial injury initiates proliferation of vascular smooth muscle cells (SMCs) and conversion of monocytes to macrophages (cholesterol ester-laden foam cells) with proliferation of fibroblasts under the influence of growth factors and cytokines. Endothelial erosion is associated with atherosclerosis. Oxidized LDL accelerates the formation of foam cells, atheroma, and fibrous plaque. Plaque rupture initiates the events of MI. CHD progression is related directly to levels of TC and LDL-C, and inversely to HDL-C, especially HDL2-C (a fraction of HDL-C) or the ratio of HDL2-C:LDL-C. Levels of triacylglycerol > 6.5 mmol/L (250 mg/dL) increase the risk for MI and warrant intervention; more recent data on risk prevention suggest lowering this risk level, perhaps to 2.6 mmol/L (100 mg/dL). When LDL-C levels are reduced, cholesterol is removed from plaque thereby leading to regression of atherosclerosis. The National Cholesterol Education Program (11
) has set a goal of LDL 
2.6 mmol/L (100 mg/dL) in secondary prevention of CHD (ie, in those who have already suffered an MI). This level of LDL usually parallels a TC value < 4.7 mmol/L (180 mg/dL).
Current theories of atherogenesis implicate plaque rupture with the release of a necrotic lipid core as the precipitating factor for thrombosis at the site and MI. Lipid-lowering, especially of LDL-C, stabilizes the plaque and reduces the risk of MI. Lower lipid levels may also decrease local concentrations of modified lipoproteins that have proinflammatory effects. With intervention, CVD mortality may be decreased by 25%, and incidence of MI by 50%. A significant drop in LDL-C has been shown to halt the progression of coronary atherosclerosis (12
). The percent drop in LDL correlates 1:1 with the decrease in coronary events. Maximal dietary therapy typically reduced LDL-C levels by 0.39-0.65 mmol/L (15-25 mg/dL), or about 5–10%.
The 2001 National Cholesterol Education Program (13
) proposed a multifaceted approach, designated "therapeutic lifestyle changes", to reduce the risk for CHD.
) recommends a total dietary fiber intake of 25–30 g/day from food, about double the current intake in the United States.
) is based on a soluble fiber intake of 3 g/day (3 servings of oatmeal, 28 g each, each with 1 g of ß-glucan soluble fiber). This soluble fiber can decrease TC and LDL-C by approximately 0.13 mmol/L (16
).
,18
).
The American Heart Association recently issued a revised set of dietary guidelines (19
) that are more individualized than before and that are food-based. There are four population-wide goals:
Major guidelines are set for each goal. In the detailed American Heart Association scientific statement, nuts are specified in several guidelines. Nuts are referred to under Goal 1 in a paragraph about soluble fibers, viz., "Grains, vegetables, fruits, legumes, and nuts are good sources of fiber."
For Goal 3, nuts are listed under a guideline that limits foods with a high content of SF and cholesterol. The dietary recommendation is to "substitute grains and unsaturated fatty acids from fish, vegetables, legumes, and nuts." In a more detailed discussion, the report states that meta-analysis showed that 1 g of soluble fiber in the substituted foods (including nuts) would be expected to decrease LDL-C by an average of 0.057 mmol/L (2.2 mg/dL) (16
). This section of the revised American Heart Association guidelines also discusses the evidence that foods rich in long-chain n-3 PUFAs confer cardioprotective effects beyond their effect on the improvement of the lipoprotein profile. These include reduction in sudden death (20
,21
), decreased risk of arrhythmia (22
), lower plasma triacylglycerol levels (23
), and a reduced blood-clotting tendency with eicosapentaenoic acid (20:5n-3, EPA) (24
) and docosahexaenoic acid (22:6n-3, DHA) (25
). In females, 
-linolenic acid (18:3n-3) reduces risk of MI and fatal ischemic heart disease (IHD) (26
,27
). The report also cited the randomized controlled Lyon heart diet trial (28
) that demonstrated the beneficial effects of -linolenic acid on both coronary morbidity and mortality in patients with CHD. Nuts are specified as foods high in n-3 fatty acids, as are other plant sources (eg, flaxseed and flaxseed oil, canola oil, and soybean oil) and fatty fish. Under the heading "issues that merit further research," the section on n-3 fatty acid supplements states that intakes of EPA and DHA of approximately 900 mg/day could beneficially affect CHD mortality rates in patients with coronary disease.
Goal 4 recommends consuming a dietary pattern that emphasizes fruits, vegetables, LF dairy products, and is reduced in fat and cholesterol. The revised American Heart Association guidelines also discuss the study known as the Dietary Approaches to the Study of Hypertension (DASH) (29
). The DASH clinical trial includes nuts in the dietary pattern and results in reduced systolic and diastolic BP, especially in African American hypertensive patients. Recent study results on DASH have been published by Conlin et al. (30
), Sacks et al. (31
), and Svetkey et al. (32
).
Factors and mechanisms other than fats or blood lipids that affect CVD risk
Other proposals have emphasized increasing the intake of n-3 fatty acids in fish and n-3 oils like flaxseed for prevention of CHD, in part because of their favorable action on prostaglandins. Cholesterol-lowering plant sterols, or their derivatives (ie, esters of sitosterol and sitostanol) have been added to fats and salad dressings. The FDA has authorized a health benefit claim for plant sterol/stanol esters and CHD (33
). Other dietary components that lower cholesterol or reduce oxidized cholesterol include 25 g/day of soy protein (34
), and perhaps 37 mg/day of the soy isoflavones (35
), and the antioxidant vitamins E and C (explained below). The amount and type of dietary protein can affect levels of TC and LDL-C. Animal proteins are hypercholesterolemic, and plant proteins are hypocholesterolemic. This may be attributed to the content of the amino acids lysine and methionine in animal proteins and arginine in plant proteins (see section on Proteins in nuts as a factor in heart disease risk).
Moderate intake of alcoholic beverages (1–2/day, 10/week) raises HDL-C levels, and intake of rice bran or olive and canola oils may as well. The tocotrienols, plant sterols, or flavonoids in these oils may be partially responsible for an antiatherogenic effect. As vascular biologists derive more scientific data, they may find that foods influence atherogenesis by mechanisms that are less dependent on lipid:lipoprotein levels and more dependent on vascular reactivity and thrombus formation, in part mediated by nitric oxide (NO). Thus, the diet-heart hypothesis has evolved from the direct relationship of fat-to-heart-disease via blood lipids to more complex associations of foods, nutrients, and phytochemicals, along with genetic and molecular mechanisms of CVD, its pathology and clinical expression.
An elevated blood Hcy level is a risk factor for CVD. Hcy increases in relation to deficient intake or metabolism of folate, and vitamins B6 and B12. The following conclusions have been drawn from several recent studies on Hcy. First, Hcy is an independent risk factor for CHD equivalent in importance to hyperlipidemia and smoking, is a strong predictor of CVD mortality, and accounts for 10% of the attributed risk of CHD (36
). Second, Hcy promotes prothrombotic changes in the vascular environment, arterial narrowing and endothelial cell toxicity, affects platelets and clotting control mechanisms, and stimulates smooth muscle cell proliferation. Third, investigators (37
) have linked hyperhomocysteinemia with premature vascular occlusive diseases, viz., carotid occlusive disease, cerebrovascular disease, CHD, peripheral arterial occlusive disease, and veno-occlusive disease. The issue of Hcy and cardiovascular risk is addressed in a recent editorial (38
) in the American Journal of Clinical Nutrition and in two articles written from opposite standpoints (39
,40
).
The relative risk (RR) of CVD is increased significantly when Hcy levels exceed 
15.8 µmol/L (41
). Apparently, there is a graded effect of the Hcy level on the risk of CVD (42
). Elevated blood levels of Hcy can be normalized with vitamin supplements (0.2–1 mg folic acid, with or without 0.4 mg cyanocobalamin, 10 mg pyridoxal), potentially decreasing cardiovascular risk (1
). Emphasis should be placed on meeting daily requirements for folate and vitamins B6 and B12 in the diet.
The possible role of antioxidant vitamins (ie, vitamins E and C) and dietary supplements in reducing CVD risk is under investigation. The dose of vitamin E that may be effective and safe, and the minimum duration of treatment for protection, are unknown (43
–45
). The GISSI-Prevenzione trial in Italy (46
), a multicenter, open-label design, with random allocation of 11,324 male and female patients to daily n-3 PUFA and/or vitamin E (300 mg) doses, reported that vitamin E administered alone had no statistically significant benefit on survival after MI. This conclusion was based on the combined endpoints and their individual components when analyzed according to the factorial design. However, a statistically significant beneficial effect of vitamin E was determined from the secondary analyses of cardiovascular deaths and the three component subsets, and was comparable to the beneficial effect of n-3 PUFA. The report noted that the vitamin E dose was in excess of any dose achievable through dietary intake, corresponding to 200 tablespoons (3.0 L) of olive oil daily, and is much higher than the recommended dietary allowances for optimum health in adults. Several prospective intervention studies have found no reduction in coronary disease with vitamin C supplementation (43
,47
–50
).
Diets for primary and secondary prevention of coronary heart disease
Diets for primary or secondary prevention of CHD should:
These nutrients can affect not only recognized risk factors such as lipids and lipoproteins, but also vascular mechanisms, clinical symptoms, and heart disease outcomes (such as morbidity and mortality).
|
DIET MODIFICATIONS AND HEART DISEASE RISK AND OUTCOMES |
|---|
|
TOPPreviousNext |
|---|
It is well-established (51
) that SFA increases the risk of heart disease by increasing the levels of TC, LDL-C, apoprotein B (apo B), and increasing the ratios of TC:HDL-C and LDL-C:HDL-C. The main sources of SFA in the diet are animal fats, particularly in dairy and meat products. Decreasing SFA to 10% of energy will reduce TC and LDL-C.
MUFA sources shown to be effective in lowering risk of CVD include tree and ground (legume) nuts as well as popular liquid vegetable oils such as olive oil ("Mediterranean" diet) or peanut oil ("Asian" diet). Neither of these diets specifies a particular diet per se, but references the dietary habits of the inhabitants in the geographic regions where olive or peanut oils predominate and are preferentially used. MUFA-rich foods like avocado also should be acceptable. These additions provide more variety to the standard or "prudent" diet. These MUFA-enriched diets, however, are higher in fat and may be high in energy, thus presenting a potential problem to the overweight. In contrast, substituting high-carbohydrate foods for fat also may add energy (and weight), especially if sugars are increased rather than whole grain complex carbohydrates.
Beneficial effects on the lipid profile of the intake of n-6 PUFAs have been reported in many studies published since the late 1950s (52
–64
). This report is limited to recent reviews and meta-analyses of data from controlled studies. Seminal articles by Keys et al. (65
) and Hegsted et al. (66
) established predictive equations that related levels of fat intake to serum cholesterol levels. This established the initial prediction that PUFAs lowered cholesterol whereas SFAs increased cholesterol.
Salonen et al. (67
), from Finland, studied the relationship of dietary intake of -linolenic acid to BP. Intake obtained from 4-day food records (-linolenic acid ranged from 0.6 to 4.1 g/day, mean 1.7 g/day) inversely related to mean resting BP. The authors did not provide data on the food sources of the fatty acid. The intake of -linolenic acid ranged from 0.6 to 4.1 g/day, mean 1.7 g/day.
Another study, by Berry et al. (68
), investigated the relationship of the fatty acid concentration in adipose tissue and BP in 399 free-living males in New York City. The investigators concluded that -linolenic acid had a disproportionate association with BP, ie, a 1% increase in -linolenic acid reduced BP 5 mm Hg. The authors cited the principal food sources of linolenic acid in the diet as flaxseed (linseed) oil, legumes, and tree nuts (including walnuts and chestnuts).
Kang and Leaf (69
) have reviewed the possible antiarrhythmic effect of n-3 and n-6 PUFAs, with a significant decrease in sudden death associated with their intake. The authors postulated that the fatty acids are incorporated into the sarcolemmal phospholipids of myocytes, perhaps binding to the protein of the sodium channel.
Mutanen and Freese (70
) and Mutanen (71
) reviewed the effects of PUFA on platelet aggregation, citing contradictory results. The evidence concerning linoleic acid (18:2 n-6) is not consistent, but intervention studies show increased platelet aggregation to agonists after high linoleate diets. Intake of -linolenic acid either has no effect or leads to decreased platelet aggregation.
Reviews of the studies relating to the consumption of n-3 fatty acids from fish oils have examined their possible role in preventing restenosis in patients after stent angioplasty. Results were variable, but none of these studies utilized -linolenic acid. Instead, they used long-chain, more polyunsaturated marine oils, compared at times with olive oil. Effects on BP, and production of cytokines and growth factors, were investigated in addition to stenosis. Future studies with plant sources of -linolenic acid would be interesting.
Effects of fats on lipid profile – meta-analyses
Results of a quantitative meta-analysis of 395 metabolic ward controlled studies of fat intake and blood cholesterol indicated that in the British diet, for example, replacing 60% of SFA by other fats and avoiding 60% of dietary cholesterol would reduce blood TC by 10–15% (about 0.8 mmol/L), with 80% of the reduction occurring in LDL-C (72
). Median duration of studies was 1 mo. This dietary change could be accomplished by the isocaloric replacement of 10% of SFA with complex carbohydrates, or the isocaloric replacement of 5% of complex carbohydrates by PUFA, and decreasing total daily dietary cholesterol by 200 mg. A 5% increase in MUFA had no significant effect on TC or LDL-C. A decrease in SFA provided the greatest cholesterol- and LDL-C-lowering effect; an increase in PUFA and decrease in cholesterol had about equal effect, about one quarter that of SFA reduction.
Mensink et al. (73
), in a meta-analysis of 27 controlled diet trials, found that the most favorable lipoprotein risk profile for CHD was achieved if SFA were replaced by unsaturated fatty acid with no decrease in TF intake, assuming no increase in BW. Replacing 10% of total energy (TE) as SFA by carbohydrates would lower LDL-C by 0.34 mmol/L (13 mg/dL) and HDL-C by 0.12 mmol/L (4.67 mg/dL). Replacement by MUFA would reduce LDL-C by 0.39 mmol/L (15 mg/dL) and HDL-C by 0.31 mmol/L (1.2 mg/dL); replacement with PUFA would reduce LDL-C by 0.47 mmol/L (18 mg/dL) and HDL-C by 0.57 mmol/L (2.2 mg/dL).
Gardner and Kraemer (74
) reported a meta-analysis to examine whether oils high in MUFA or PUFA had differential effects on serum lipid levels. The authors analyzed data from 14 small-to-medium diet trials and found no difference in the effect of increasing MUFA or PUFA on lowering TC or LDL-C, or in the effect on HDL-C. There may have been a slight TC-lowering effect of PUFA. They noted that other components in the oils and foods besides fatty acids may influence the outcome. MUFA studies included olive, canola, peanut, sunflower, and high-oleic safflower oils; PUFA studies included grapeseed, high-linoleic safflower oil, and corn oil. The authors considered the effects of decreasing or increasing each type of fatty acid by 10%. The SFA diet was 19% SFA, 13% MUFA, and 6% PUFA; the MUFA diet was 9% SFA, 19% MUFA, and 6% PUFA; and the PUFA diet was 9% SFA, 10% MUFA, and 16% PUFA. The authors compared n-6 with n-3 PUFA and n-3 from plant sources with n-3 from fish, suggesting that triacylglycerol-lowering may be attributed to a higher n-3 content of diets. They indicated that associated micronutrients and factors other than fat in these oils or in foods could modify their effects on blood lipids. The authors reviewed the mixed findings that unsaturated fats may or may not have a differential influence on apoproteins, lipoprotein subclasses, blood glucose, and indexes of thrombosis and cancer. They suggested that outcomes other than serum cholesterol levels should be investigated as the basis for future dietary recommendations for the use of foods high in one unsaturated fat in preference to another.
|
INTERVENTION STUDIES WITH WALNUTS |
|---|
|
TOPPreviousNext |
|---|
Each study is summarized and reviewed from the perspective of the effects on biomarkers and indicators of heart disease (viz., lipids, lipoproteins, BP, pathology, and clinical events). The study is then examined for strengths and/or weaknesses in design, subject demography, dietary intervention, exposure dose, clinical and laboratory measurements, and data analysis. Next the study is evaluated in terms of quality, consistency (ie, agreement) and magnitude of the effect, and the strength and relevance of the association of walnuts with the outcome. Finally, all of the studies are presented in chronological order of publication.
Study 1.
Sabaté et al. (75
) reported the effects of walnuts on serum lipid levels and BP. The investigators studied 18 healthy, normal weight, young (mean 30 y) adult males with "normal" lipid levels (20th–80th percentile) and BP, who did not eat nuts frequently. The subjects were placed on a National Cholesterol Education Program/American Heart Association Step 1 diet (30% fat), either with or without 20% of energy from walnuts, in a single-blind, randomized, crossover design with two 4-week feeding periods. Walnuts reduced the TC, LDL-C, and HDL-C values by 12% (22 mg), 16% (18 mg), and 5% (2 mg), respectively; all changes were statistically significant. There was no effect on BP. Triacylglycerol levels tended to decrease, but not significantly; the ratios of TC:HDL-C and LDL-C:HDL-C decreased significantly. The Walnut diet caused a slight increase in the intake of TF. SFA intake decreased by one-third, PUFA increased 80%, and cholesterol declined 50% in comparison to the Step 1 reference diet (RD), which included no walnuts. Comparing the Walnut and RD periods, the investigators noticed that the fatty acid composition of serum cholesterol esters reflected significant decreases in oleic (18:1 n-9) and arachidonic (20:4 n-6) acids and increases in linoleic and -linolenic acids in the Walnut diet.
Although there was a 5-day run-in period, there was no washout between the 4-week duration of each diet (with or without walnuts). The study personnel were blinded to the patients’ study sequence, ie, one group was consuming the Walnut diet first and then crossed over; the other group was eating the diet without walnuts first and then crossed over. The nutrition research kitchen provided all meals; common, normal food items comprised the menus. The Walnut diet substituted three servings (28 g each) of walnuts daily (for a total of 84 g daily per 2,500 kcal) in place of other foods. Walnuts were used for snacks, mixed in salads and breakfast cereals, or cooked with dinner entrees. The composition of duplicate samples of the two study diets was determined by analyses of samples collected randomly. There were no side effects from the Walnut diet.
Lipid and BP measurements were standardized and validated. The two-tailed t test was used for statistical analyses.
The major difference between the two diets was the increase in PUFA in the Walnut diet, with the percent of -linolenic acid tripling and total PUFA increasing 2.5-times. The response of the subjects was consistent regardless of the sequence of the diets, baseline lipid values, or body size. During the Walnut diet, walnuts contributed 55%, 14%, and 10% of TF, protein, and fiber, respectively.
The investigators concluded that replacing a portion of the fat in a cholesterol-lowering diet with walnuts further lowers serum cholesterol levels and produces a more favorable lipoprotein profile. They suggested that an increase in walnut consumption from a current average of 4 g/week to 28 g/day would lower cholesterol and LDL-C by 4 and 6%, respectively. They also suggested that incorporating walnuts into the diet as snacks, or components of desserts, breads, or entrees would be acceptable to most people as part of a cholesterol-lowering diet.
The investigators noted that their study did not include females or subjects with hypercholesterolemia, and included only younger subjects, and that 4 weeks was a short interval. The results, therefore, could not be extrapolated directly to the population at higher risk of CHD. The issue of free-living subjects was not addressed.
This study was well-designed and controlled, insofar as a diet study with a whole food cannot be blinded to the subject. The small number of subjects, consisted only of young adult males; nevertheless, the dietary intervention was practical for the "real world" and a healthy population. A 4-week study is enough time in which to detect changes in blood lipids, but not long enough to address maintenance. The energy intake was controlled, but the proportion of fat energy from walnuts was too high to be practical. The extrapolation to one-third of that intake (ie, 28 g) and prediction of appreciable lipid-lowering need to be validated by direct study at that level. The magnitude of the lipid effect (cholesterol and LDL-C reduction) is significant and could provide risk reduction for heart disease for the population that is comparable to oatmeal and oat bran (15
,76
) or soy protein (77
,78
). The 5% fall in HDL is noteworthy, but is counterbalanced by the significant decrease in the ratios of cholesterol and LDL:HDL. The trend to decrease TGs by 10% might achieve statistical significance with a small increase in the number of subjects (perhaps to 25 individuals).
Study 2.
Abbey et al. (79
) studied 16 normolipidemic males, mean age of 41 y, with a 36% fat diet for periods of 3 weeks. The investigators compared a diet enriched with peanuts and coconut that was similar to the usual Australian diet in fat composition (SFA:MUFA:PUFA ratios) with an almond-enriched diet and one enriched in walnuts (68 g/day, 46 g fat). The nuts provided about half of the fat energy. All subjects were fed the diets in the same order. The Walnut diet was continued for an additional 3 weeks. Both almonds and walnuts similarly lowered TC and LDL-C significantly, 6 and 10%, respectively, after 3 weeks compared with the RD. The HDL-C and triacylglycerol values did not differ among the three diets. The plasma fatty acid composition with the Walnut diet showed significant decreases in oleic and arachidonic acids and significant increases in linoleic and -linolenic acids. Plasma and dietary fatty acid composition were measured as were plasma lipids and lipoproteins using validated methods. Data were analyzed by paired t tests.
The dietary intake was monitored by food records, with special emphasis on fat intake, and reviewed by the dietitian. The diets were well matched. The investigators were concerned about possible adverse effects of the high PUFA content of the Walnut diet on lipid peroxidation and the formation of oxidized LDL, although this was not studied. The relatively high content of tocopherols in walnuts might be protective (Table 1
). They also discussed the possible influence on the results of elevated myristic acid (14:0) in the RD. They suggest that PUFA- and MUFA-rich nuts should be included in the diet as a replacement for some of the SFA.
|
) prediction of appreciable lipid-lowering, and needs to be validated by a clinical study at that dose. The magnitude of the lipid effect (cholesterol and LDL-C reduction) is significant and could provide risk reduction for heart disease for the population that is comparable to oatmeal and oat bran (15
) or soy protein (78
). The lack of effect on HDL is encouraging and may reflect either the lower level of intake of walnuts or the longer duration of feeding. It is interesting that the results with almonds did not differ from those achieved with walnuts, despite the difference in fatty acid composition, ie, almonds are high in MUFAs and walnuts in PUFAs (Table 1
).
Study 3.
Chisholm et al. (80
) reported a study of 21 males (16 with full dietary records), mean age 45 y, with moderate hypercholesterolemia (polygenic hyperlipidemia), without CHD. The randomized, crossover design compared the effects on lipid metabolism for 4 weeks of two "LF" diets (ie, 30 and 38% TE). A diet with 38% energy from fat is not LF, as described by the investigators. Walnuts (78 g/day) were included only in the 38% LF diet. With the Walnut diet, the significant decreases in TC (4%) and LDL-C (8%) compared to the baseline diet, were about twice that observed with the 30% LF diet. HDL-C increased significantly (42%) with the Walnut diet compared to baseline, and apo B levels decreased significantly (10%) with the Walnut diet compared to baseline, with no significant change from baseline with the 30% LF diet. Neither triacylglycerol nor apo-A levels differed from baseline with either the Walnut or 30% LF diet. Comparing the Walnut diet and RD, the investigators found that the fatty acid composition of plasma triacylglycerol, cholesterol esters, and phospholipids showed significant decreases in 18:1 and 20:4 and increases in 18:2 in the three lipid classes, with an increase in 18:3 in triacylglycerol. Investigators ascribed the absence of more extensive differences between the two diets as possibly being caused by the relatively low levels of myristic acid in both. They noted that the Walnut diet had significantly less SFA, more PUFA, and less cholesterol than the 30% LF diet and recommended that an increase in TF be avoided when nuts are included in the diet.
The experimental design included a run-in period on a LF diet for 1 week before randomization into two groups, with one group continuing on the 30% LF diet, and the other switched to the Walnut diet. There was no washout between crossover periods. Dietary intake was recorded on 2 days of each week of the study. Meals were eaten at home, with individualized menus and recipes; walnuts were provided. Walnuts contributed 20% of the TE of the diet. The lipid measurements were validated for precision and accuracy, and the variances were provided. Data were analyzed using ANOVA with repeated measurements.
Despite the unintended 8% increase in TE from fat on the Walnut diet, the two diets were isocaloric. The subjects did not differ in BW or weight gain with either one of the diets. The investigators concluded that walnuts influence the plasma fatty acid profile in a way expected to reduce the risk of CHD, and is thereby cardioprotective. They were unable to demonstrate convincingly that the Walnut diet was superior to the 30% LF diet in improving the lipoprotein profile. They cautioned that an increase in TF should be avoided in order to improve study design in future interventions with nuts or in advising an increase in nut consumption.
This study did not achieve the desired fat content for the Walnut diet that would provide a better comparison with a LF diet. There was a modest lowering of cholesterol and LDL-C with both diets, especially considering the high intake of walnuts. Again, the number of subjects was small and consisted only of males. These subjects, however, were hyperlipidemic, rather than representative of the healthy population. This 4-week study was adequate to detect changes in blood lipids, but inadequate in terms of addressing sustainability over time. The energy intake was controlled, but the proportion of fat energy from walnuts was too high to be practical. The extrapolation to one-third of that intake (ie, 26 g) would not produce significant enough reduction of cholesterol and LDL-C risk factors for heart disease (ie, 1–2%) to indicate significant protection. The notable increase in HDL-C levels could provide a reduction in risk for heart disease for this population.
Study 4.
Zambón et al. (81
) reported on the Barcelona Walnut trial. The randomized, crossover design feeding study was carried out on 23 females and 26 males with polygenic hypercholesterolemia, mean age 56 y. Patients were recruited from a lipid clinic where they had been placed on a cholesterol-lowering "Mediterranean" diet. They were free of disease other than CHD (7 subjects), took no lipid-lowering medications or dietary supplements, and did not consume nuts frequently. A high-MUFA diet ("Mediterranean" diet consisting of olive oil and natural food stuffs) was compared to a high-PUFA diet (walnuts, 56 g/day) during 6-week feeding periods with random assignment to one of two sequences. There was a 1-week pretrial period for indoctrination and counseling, and no washout period between diets. The MUFA diet was 30% fat, 5% SFA, 21% MUFA, and 4% PUFA, and the PUFA diet was 33% fat, 5% SFA, 16% MUFA, and 12% PUFA. Raw, shelled walnuts were provided daily in packages varying between 41-56 g (8–11 walnuts), affording 18% of TE. Walnuts were consumed in desserts or salads, or as snacks. Diets were monitored by unannounced weekly telephone 24-h recalls. Serum lipids, Lp (a), apoproteins A-1 and B, and LDL resistance in vitro to oxidative stress were measured, as well as the fatty acid composition of LDL lipids in blood samples obtained at 5 and 6 weeks of each dietary period.
TC decreased by 9% on the Walnut diet (by 5% on the control diet); LDL-C declined by 11% (by 6% on the control diet), with parallel changes in apo B levels. Lp (a) levels were decreased (9 vs. 3%) with the control diet. The change in Lp (a) levels was statistically significant only in males and in subjects with levels of 
0.78 mmol/L (30 mg/dL). There were no significant differences in the effects of the two diets on levels of HDL-C, triacylglycerol, or apo A-1 (both diets decreased these variables). The LDL:HDL ratio declined 8% with the Walnut diet and was unchanged with the control diet. Resistance of LDL to oxidative stress was preserved during the Walnut diet. There was no difference between the Walnut and control diets in the -tocopherol content or in the lag time of conjugated dienes (CDs) formation during copper-induced oxidation. Oleic acid was decreased, whereas linoleic and -linolenic acids were increased in LDL-cholesterol ester fatty acids, LDL phospholipid, and triacylglycerol with the Walnut. There was no evidence of a carryover effect.
BW was stable and most subjects tolerated the Walnut diet well, although a few reported mild symptoms of postprandial bloating or heaviness. The energy value and nutrient composition of the two diets (other than the fat) had minor differences besides lower cholesterol content in the Walnut diet, and were insufficient to explain the differences in blood lipids and lipoproteins that were observed with the diets.
Investigators concluded that substituting walnuts for part of the MUFA in a cholesterol-lowering "Mediterranean" diet further reduces TC and LDL-C in males and females with hypercholesterolemia. The Walnut diet effect was observed without changing dietary SFA. They suggested that nonfat components of the walnut matrix may influence blood lipids, and that the lipid effects of whole walnuts should be compared with that of walnut oil. They also noted that the content of -linolenic acid in walnuts may reduce the risk for CHD deaths through its antiarrhythmic properties or other antiatherogenic effects (see below). They proposed that even greater benefits than they encountered studying the "Mediterranean" diet would be obtained by replacing the traditional "Western" dietary fat with walnuts.
Although the results show no change in triacylglycerol levels with either diet, an asterisk in Figure 2 in Zambon (81
) indicates a significant lowering of triacylglycerol with the Walnut diet. This would sustain evidence that walnuts have another favorable lipid effect in addition to those on cholesterol and lipoproteins. This study was well designed and controlled. The number of subjects was greater, included females and males, with a broader age range, and with elevated lipid levels. The 6-week study is not considered long-term. The caloric intake was controlled, but the TE from fat was higher with the Walnut diet, achieving a moderate fat rather than a LF intake. The proportion of fat calories from walnuts was double that considered practical. The extrapolation to half of that intake (ie, 25 g) and prediction of 4-5% cholesterol or LDL cholesterol-lowering needs to be validated by direct clinical study at that intake level. The magnitude of the lipid effect could provide risk reduction for heart disease for the population. The 3% fall in the HDL level is counterbalanced by the significant (8%) decrease in the ratio of LDL:HDL. A significant decrease in triacylglycerol would be an added benefit. The absence of an increase in oxidative stress with the Walnut diet also is favorable. The strength of this intervention diet is that the SFA content was the same for the two diets, so that the increase in PUFA was the significant change and was not accompanied by a decrease in an "unfavorable" nutrient.
Study 5.
Almario et al. (82
) reported the effects of a Walnut diet on plasma lipids and lipoproteins in patients with combined hyperlipidemia. Four diets were studied: the habitual diet (HD) (31% fat) and a LF (20% fat) diet; and the same diets supplemented with walnuts (48 g/8460 kJ). Thirteen postmenopausal females and 5 males, age 60 ± 8 y, completed the four periods. Comparing the Walnut diet with the HD and LF diets, patients lost weight on the LF diet, but did not gain weight with either Walnut diet, despite an increased TE intake of 20 and 23%, respectively.
During the LF + walnuts period, decreases in TC (8%) and LDL-C (12%) were significant, compared with the LF diet alone. Triacylglycerol did not differ among the four diets, and cholesterol was decreased significantly in intermediate density lipoprotein (IDL) and redistributed from the more atherogenic small, dense LDL to larger LDL particles. Plasma apo B was not effected by walnut supplementation but decreased with the LF compared to the HD. The HDL-C decreased (10%) significantly when comparing either the walnut supplemented HD or LF to HD. Apo A-1 increased with the addition of walnuts to the HD. Apo A-1 decreased during the LF diet and remained low after the addition of walnuts. Adding walnuts to the HD or the LF diet alone caused significant shifts in cholesterol that redistributed it from larger into smaller HDL particles. Lipid particle changes occurred in the absence of lipid lowering, suggesting a favorable antiatherogenic effect of walnuts that is independent of any changes in circulating lipid levels.
The patients’ health was stable and their lifestyle was monitored. All patients followed the diet according to the following sequence: 4 weeks on the HD diet then 6 weeks each on the HD + walnuts, LF diet, LF + walnuts, respectively. There were no washout periods. Patients were counseled individually before the LF diet intervention and with group sessions during those periods.
The quantity of walnuts was selected to provide an amount of -linolenic acid similar to the amount that lowered triacylglycerol in fish oil studies. Walnuts were consumed mainly from prepacked rations that were provided. Although the patients were seen and blood samples were drawn at mid- and endpoints of each period, only the terminal samples were used for data analysis. Lipoprotein particle sizes were determined by nuclear magnetic resonance techniques. Diets were monitored from 7-day food records for each period. Data analysis used general linear modeling procedures and correlations.
The walnuts added to the HD increased the fat content from 31 to 37%. Linoleic acid intake increased from 11 to 30 g and -linolenic acid from 1.3 to 5.4 g. The SFA in HD and HD + Walnuts were 11 and 10%, respectively. MUFA were 12 and 13% and PUFA were 6 and 16% in HD and HD + walnuts, respectively. During the LF diet, the TE decreased by about 12% compared to the HD; this was reversed during the LF + walnuts diet. The LF diet contained 20% fat, 8% SFA, 8% MUFA, and 5% PUFA. LF + walnuts contained 34% TF, 8% SFA, 12% MUFA, and 16% PUFA. Addition of walnuts to either the HD or LF diet increased plasma concentrations of linoleic and -linolenic acids and decreased palmitic (16:0), oleic, and arachidonic acids without changing concentrations of EPA and DHA.
The investigators suggest that the absence of effect of the LF diet is because these hyperlipidemic subjects were already following a diet restricted in fat and SF. They noted that the cholesterol-lowering caused by adding walnuts to the LF diet could not be explained in terms of decreasing SFA, but rather was a specific effect of the PUFA and MUFA content. They suggest further examination of possible effects of PUFA on lipoprotein metabolism via effects on lipoprotein lipase, hepatic lipase, lecithin-cholesterol acyltransferase, or cholesterol ester transfer protein, which are all involved in lipoprotein metabolism. Walnuts have a unique elevation of both n-6 (linoleic) and n-3 (-linolenic) acids compared to other nuts. Their metabolic effects should be specifically elucidated and not extrapolated from other unsaturated fats and oils, even of similar fatty acid composition (eg, soybean and wheat germ oils).
This study was well designed and controlled. A small number of hyperlipidemic females and males were studied and fed diets for 6 weeks. Adding walnuts to either diet caused variation in the caloric intake and increased TE by 20%. This might be expected when adding walnuts to a LF diet but not, perhaps, to the HD. Also, the LF diet provided only 82% of the TE of the HD. The walnut intake is about twice that which is practical to include in the diet. The extrapolation to half that intake (24g) and prediction of appreciable lipid-lowering needs to be validated by direct study at that level. The magnitude of the lipid effect (cholesterol and LDL-C reduction) is significant and could reduce the population’s risk for heart disease to an extent that is comparable to oatmeal and oat bran (15
) or soy protein (78
). The fall in HDL is noteworthy, although the LF diet produced a similar decrease. The ratios of cholesterol and LDL:HDL were not calculated. The intriguing findings of changes in lipoprotein particle sizes and composition need confirmation. The results obtained by the investigators from the emerging nuclear magnetic resonance methodology need to be related to the traditional procedures.
Study 6.
In the Jerusalem Nutrition Study (83
), walnuts were one component of a high-PUFA diet study, although there were no data on the amount of walnuts subjects consumed. The intervention study compared 32% fat, high-MUFA and -PUFA diets in 26 healthy, normolipidemic, young males who were randomized to either diet in a 12-week crossover design. The MUFA diet was 8% SFA, 16% MUFA, and 7% PUFA with fat sources from olive oil, avocado, and almonds. The PUFA diet was 7% SFA, 6% MUFA, and 16% PUFA with fat from safflower and soy oils and walnuts. There was a 4-week washout before cross over to the other diet for an additional 12 weeks.
The PUFA diet reduced TC and LDL-C by 16%, whereas the MUFA diet reduction was 10%. There was no change in HDL-C or triacylglycerol. There was an increase in response to oxidative stress as determined by measurements of thiobarbituric acid reactive substances (TBARS) with the PUFA vs. MUFA diets. LDL receptor activity and lipoprotein profiles and composition (studied in about half the subjects chosen at random) were not significantly different between the two diets.
All food, consisting of natural and common food items, was served in a central kitchen. There was a 4-week run-in period on their usual diet (41% fat, 10% SFA, 12% MUFA, and 16% PUFA). Menus were prepared in 12-day rotations. Samples of the diet were analyzed for fatty acid composition by gas-liquid chromatography. Standard procedures were used to measure the lipids and lipoproteins. Statistical analysis was by means of the t test, comparing the means of two determinations at 10 and 12 weeks of the diets and two baseline samples. Dietary compliance by the subjects was monitored by determining the fatty acid composition of the lipids in the RBC membrane. The investigators demonstrated a seasonal effect on fatty acid levels, especially of triacylglycerol, but also of LDL-C, that might be related to dietary changes in Israel vis-à-vis religious observances.
This study focused on the demonstration that MUFAs were equivalent to PUFAs in cholesterol-lowering. In fact, they showed a better effect by PUFAs, which included walnuts as a source. The amount of walnuts was not specified, however, which makes this study interesting but probably not suitable for inclusion in this report as a walnut intervention study. Rather, it supports the direct intervention data. The study resembles, in part, the comparison in Abbey’s study (79
) of an Almond diet with a Walnut one. The study shows that PUFAs do not adversely affect HDL, but also supports an adverse effect of PUFAs on oxidative stress (see section on Possible adverse effects of walnut components).
Summary of clinical human intervention trials with walnuts
Summaries of the six clinical human intervention studies published or in press are presented in
Table 2
. They are consistent in showing decreases in TC and LDL-C that should lower risk of CHD. These results are achieved with intakes of walnuts that would amount to two or three servings daily. There appears to be a null effect on triacylglycerol. Effects on HDL-C are inconsistent, with some studies showing a decrease, others no change, and one an increase. Not all the studies have included the ratios of TC:HDL-C or LDL-C:HDL-C. Where evaluated, these ratios decreased, indicating lessening of risk, even when HDL-C is decreased. In one trial, although HDL-C decreased, apo A-1 increased, providing another favorable outcome. A few studies have included apo B and have shown decreases as well, again favoring risk reduction. These results have been achieved whether the walnuts were included in a usual diet higher in fat, a Step 1 diet, a LF diet, or in a diet already increased in PUFAs.
|
). Among the deficiencies in these trials is that no study included a large number of subjects, encompassing males and females, of a broad age range, and studied over a long period of time. Each involved a level of walnut consumption that is not considered practical in the "real world." No significant adverse effects have been reported in any of the studies. Although the walnut consumption may add calories, net BW gains did not occur. The concern about increased oxidative stress in these studies mirrors the concern with diets that increase PUFAs. The data are inconsistent; the one Walnut trial that examined this possibility showed no adverse effect. The effect of walnut ingestion at a practical level, ie, one serving daily, has not been evaluated. The walnut effect at higher intakes (two to three servings daily) resembles that of oat fiber and soy protein, where the amounts tested in trials exceed usual intakes.
The mechanism of the walnut effect is unclear. The diets used generally reflect a ratio of SFA:PUFA of 1:2, which was a dietary recommendation made early in efforts to reduce the risk of heart disease. Investigators have claimed that the risk reduction with walnuts may exceed that predicted from their fatty acid concentrations and composition, suggesting other factors (see below). Walnuts are unique among nuts in that both -linolenic acid and linoleic acid levels are increased. The dietary intake of walnuts significantly increased levels of both PUFAs in the body, thus demonstrating bioavailability. Risk factors of interest that were addressed only in a single trial include Lp (a), which has not been shown to respond readily to dietary intervention, and LDL and HDL particle sizes, which are emerging as more specific risk factors than the levels of LDL or HDL-C. These particle sizes may have different effects on risk (ie, not all LDLs are "bad" and not all HDLs are " good").
A possible effect on Hcy also needs to be examined in a strong study. Similarly, the significance of the lysine:arginine ratio remains to be defined.
|
OBSERVATIONAL STUDIES ON CONSUMPTION OF WALNUTS AND OTHER NUTS AND CORONARY HEART DISEASE |
|---|
|
TOPPreviousNext |
|---|
.
|
The AHS is a large, prospective study of Seventh-day Adventists (SDAs) in California (84
–87
), in which 31,208 non-Hispanic white subjects were followed for 6 y (1977–1982). Excluding people with heart disease or diabetes, at baseline 26,473 males and females over age 25 were evaluated for first events of MI or CHD. Nut consumption ranged from never to daily. Nut consumption was related inversely to nonfatal MI or death from IHD, or to all-cause mortality. RR of MI and IHD in persons who ate nuts more than 4 times/week were half the RR of subjects who consumed nuts less than once weekly (RR 1). People who ate nuts 1–4 times/week had a 22% reduced risk of acute MI, compared with those eating nuts less than once a week.
The population was predominantly nonsmoking, did not use alcohol, and generally followed a lacto-ovo-vegetarian diet. Diet was assessed by a semiquantitative food frequency questionnaire (FFQ). Participants’ health status was queried annually, and medical records were reviewed for the diagnosis of CVD, nonfatal or fatal, according to the International Classification of Diseases codes. The association of nuts and CVD was undiminished when analyzed in the elderly subgroup, and did not differ by sex, BP, relative weight, nor was there any significant confounding by other foods. Although the investigators did not distinguish between different kinds of nuts, a substudy showed that 32% of nuts consumed were peanuts (legumes or ground nuts), 29% were almonds, 16% were walnuts, and 23% were other kinds of nuts. The authors concluded that frequent consumption of nuts might be protective for both fatal and nonfatal CHD events.
Fraser (85
) has proposed that high nut consumption postpones the development of IHD for several years and confers an 18% lifetime risk reduction of CHD compared with 30% in consumers of low quantities of nuts.
There are several publications by investigators involved in the AHS prospective observational studies on SDA cohort populations. Fraser and colleagues (88
) estimated the effects of particular practices or risk factors, such as consumption of nuts, on the lifetime risk of CHD and the time of first expression in 26,321 non-Hispanic, white SDAs. Their analyses:
A population of African American SDAs in California was analyzed for health habits, risk factors, and all-cause mortality (89
). The same factors were found to operate also in this population. Frequent consumption of nuts appears protective in that population (1,668 subjects), reducing all-cause mortality in both sexes to 0.6, adjusted for age, smoking, and exercise. In the "old-old" cohort (603 female and male subjects, age 84 y (90
), all-cause mortality was decreased to 0.75 (adjusted multivariate RR 0.82) comparing high to low nut consumption. Mortality from CHD was reduced to 0.55 for high vs. low nut consumption (multivariate adjusted RR 0.61).
A report in 1999 (91
) showed that there was a significant protective association between nut consumption and fatal and nonfatal CHD in SDA males and females, RR 0.5 comparing high to low nut consumption, and reduced lifetime risk of CHD by 31% in frequent nut consumers. Results were adjusted for age, sex, smoking, exercise, BMI, hypertension, and consumption of bread, beef, fish, cheese, coffee, legumes, and fruit. Subjects with diabetes were excluded.
The Iowa Women’s Health Study (92
,93
) is a prospective cohort study of postmenopausal females. Prineas reported the effects of nut consumption on CHD mortality in the 41,837 females enrolled. Predominantly white females, 55–69 y of age, completed a FFQ that asked about the frequency of nut consumption as 1-oz (28 g) portions. Over 5-y, 154 of the 34,484 females free of CHD at baseline died of CHD. Coronary mortality was inversely associated with nut intake, with an adjusted RR of 0.6 for eating nuts 1–3-times a mo, 0.75 for eating nuts once a week, and 0.43 for eating nuts 2–4-times a week, p = 0.06.
Kushi et al. (92
) reported the results of follow up for 7 y, with 242 females dying of CHD. The primary finding was a strong inverse association of vitamin E consumption from food with the risk of CHD death. Multivariate adjusted RR in the highest quintile, ingesting 9.64 International Units (IUs) of vitamin E daily, was 0.38. In a subgroup, multivariate analysis of food sources of vitamin E showed the strongest inverse association was with nuts and seeds, exceeding margarine and mayonnaise and creamy salad dressings. RR for the highest quartile of nuts and seeds, >4 x/mo, was 0.60, but adjustment for vitamin E reduced this to 0.72, p = 0.11. The varieties of nuts consumed were not reported.
Combining the AHS and Iowa Women’s Health Study studies (84
,85
), there seems to be a threshold effect of nut consumption at a frequency of once weekly. An inverse, graded relation was observed between nut consumption and CHD events.
In the NHS (27
), of 86,016 female registered nurses (RNs), those consuming at least 5 oz (140 g) of nuts/week had a 35% lowering in nonfatal MI compared with those eating less than 1 oz (28 g) of nuts/mo (rare), RR = 0.65. The RNs ranged in age from 34 to 59 y at baseline. Coronary disease measures included nonfatal MI and fatal CHD. During 14 y of follow-up, 861 cases of MI and 394 deaths occurred. Dietary information was obtained 4 times during the study, and frequency of nut consumption was evaluated for total nuts [later separating peanuts (ground nuts) from tree nuts]. The investigators commented that nut consumption declined over the length of the follow-up period. When peanuts were separated from total nuts, there were few cases of females consuming either peanuts or tree nuts at the highest two quartiles. Multivariate RR for either group was statistically significant (0.79 for other nuts, p = 0.62; 0.66 for peanuts, p = 0.06). Medical records were reviewed for cases, and the World Health Organization criteria were used for diagnosis.
Hu et al. (94
) reported data relating dietary intake of -linolenic acid and risk of IHD among the females enrolled in the NHS cohort study. A higher intake of -linolenic acid was associated with an RR of 0.55 in the lowest quintile, p for trend 0.01. The multivariate risk calculation was adjusted for age, standard coronary risk factors (ie, smoking, BMI, hypertension, diabetes, menopausal status, and parental history of premature MI) and dietary intake of -linoleic acid, alcohol, SF, vitamin C, vitamin E, total energy, and the use of vitamin supplements. The -linolenic acid intake ranged from 0.71 g/day in the lowest quintile to 1.36 g/day in the highest quintile.
In the Physician’s Health Study (95
), of 22,000 males followed prospectively for 12 y, as nut consumption increased the risk for cardiac and sudden cardiac deaths decreased significantly. There were 133 sudden cardiac deaths and 449 cardiac deaths. The investigators suggest that after adjustments their data indicate that nut consumption r
