<<

European Journal of Clinical Nutrition (1998) 52, 12±16 ß 1998 Stockton Press. All rights reserved 0954±3007/98 $12.00

A diet rich in walnuts favourably in¯uences plasma fatty pro®le in moderately hyperlipidaemic subjects

A Chisholm1, J Mann1, M Skeaff1, C Frampton3, W Sutherland2, A Duncan1 and S Tiszavari1

Departments of 1Human Nutrition and 2Medicine, University of Otago; and 3Centre for Computing and Biometrics, Lincoln University

Objective: To compare two low diets one rich in walnuts on parameters of in a group of hyperlipidaemic subjects. Design: A randomised cross over study. Setting: Department of , University of Otago, Dunedin, New Zealand Subjects: Twenty one men with mean (s.d) levels of total and LDL of 6.58 (0.60) and 4.63 (0.58) respectively. Interventions: For two periods of four weeks subjects were asked to consume two low fat diets (fat 30% total energy), one containing, on average, 78 g=d walnuts. Walnuts obtained through Lincoln University and the Walnut Growers Group (South Canterbury). Results: Participants reported a higher total fat intake on the walnut diet (38% compared with 30% on the low fat diet P < 0.01) The most consistent change in fatty acid pro®le of triacylglycerol, and cholesterol on the walnut diet was a signi®cant (P < 0.01) increase in linoleic acid. Triacylglycerol linolenate also increased signi®cantly (P < 0.01). Total and LDL cholesterol were lower on both experimental diets than at baseline, 0.25 mmol=l and 0.36 mmol=l respectively on the walnut diet and 0.13 mmol=l and 0.20 mmol=l respectively on the low fat diet. High density cholesterol was higher on both the walnut and low fat diets when compared to baseline (0.15 mmol=l and 0.12 mmol=l, respectively). When comparing the walnut and low fat diets only apo B was signi®cantly lower (P < 0.05) on the walnut diet. Conclusions: Despite an unintended increase in the total fat intake on the walnut diet, pro®le of the major lipid fractions showed changes which might be expected to reduce risk of cardiovascular disease. The reduction of apolipoprotein B suggests a reduction in lipoprotein mediated risk, the relatively low content of both diets perhaps explaining the absence of more extensive differences in lipoprotein levels on the two diets. Sponsorship: Nutrition Department University of Otago, New Zealand. Descriptors: walnut diet; low fat diet; plasma fatty ; and

Introduction teins in normolipidaemic male volunteers in an attempt to clarify the mechanism by which nuts might exert their Diets low in total and have long been recom- protective effect. We have persued this issue further by mended for reduction of total and low density lipoprotein examining the effects of walnuts on lipids, lipoproteins and cholesterol and to reduce coronary heart disease (CHD) triacylglycerol fatty acids in men with moderately raised risk (National Cholesterol Education Programme, 1991, cholesterol levels. Recommendations of the European Atherosclerosis Society, 1992). More recently it has become apparent that certain long chain unsaturated fatty acids may protect Methods against CHD and that this protective effect may be as strong as the deleterious effect of a high saturated fat diet. Subjects Several epidemiological studies have shown reduced rates Twenty one men were recruited following advertisement in of heart disease among vegetarians (Burr & Butland, 1988; the local news media in Dunedin, New Zealand or from Thorogood et al, 1994) but there is no clear evidence that among participants in previous studies conducted in the elimination of , ®sh, and products derived from Department of Human Nutrition, University of Otago. them, accounts for this bene®cial effect. Indeed, prospec- Subjects were required to be below 65 y of age and to tive studies of Seventh Day Adventists suggest that regular have levels of total cholesterol between 5.5±7.5 mmol=l. At consumption of nuts and wholemeal may be impor- the time of recruitment all were free of symptoms of tant protective factors (Fraser et al, 1992). Two research disease and none had a strong family history of premature groups (Sabate et al, 1993; Abbey et al, 1994) have coronary heart disease. Those with familial hypercholester- examined the effects of various nuts on lipids and lipopro- olaemia, familial combined hyperlipidaemia or secondary hyperlipidaemia were also excluded. Thus all were con- sidered to have polygenic hyperlipidaemia. Ethical Correspondence: Dr J Mann, Department of Human Nutrition, University approval for the study was granted by the Ethics Committee of Otago, Dunedin, New Zealand. Received 16 February 1997; revised 19 August 1997; accepted 28 August (Otago) of the Southern Regional Health Authority and all 1997 subjects gave informed consent on the understanding that Walnuts and plasma fatty acids A Chisholm et al 13 they could withdraw from the study at any time. The Table 1 Mean fatty acid composition, of the New Zealand grown Walnut subjects initial mean (s.d.) age was 45 (6.8) y and their 2 mean body mass index 28.4 (4.3) kg=m . At the time of Fatty acid % total fatty acids Amount in g=100 g walnuts recruitment mean (s.d.) levels for total cholesterol (TC), low density lipoprotein (LDL) cholesterol and high density 16:0 7.7 5.2 lipoprotein (HDL) cholesterol were 6.58 (0.6), 4.63 (0.58), 18:0 2.2 1.5 18:1 20.8 14.2 and 1.05 (0.31) mmol=l, respectively. 18:2n-6 58.0 39.7 18:3n-3 11.1 7.6 Experimental design A randomised cross over design was employed. All parti- cipants consumed both the walnut and reference diets. The of the methodology and its practical application, including study involved a run in period of one week during which a recipes. intake was calculated using the computer standard lipid lowering diet was advised. The purpose of program `Diet Entry and Storage=Diet Cruncher' (Mar- the run in period was to allow participants time to practise shall,1993) and data from the New Zealand Food Composi- the dietary modi®cations, to achieve a diet in which total tion database (Burlingam et al, 1993). New Zealand fat provided 30% total energy, but of short enough duration walnuts obtained through Lincoln University and the not to cause a marked change in lipid levels. Thereafter Walnut Growers Group (South Canterbury) were supplied, participants were randomised into two groups, Group 1 already shelled and weighed out to the participants. The continued on the `Low Fat Diet' ®rst and Group 2 the fatty acid composition of the walnuts used in this study is `Walnut Diet'. After four weeks on the ®rst diet the shown in Table 1. participants crossed over to the alternate diet. sam- ples were taken after an overnight fast once at the end of Laboratory methods the run in period (baseline) and twice during the ®nal week Cholesterol concentrations in plasma and lipoprotein frac- of each phase of the study. The values for the various tions were measured by enzymatic methods using kits and parameters measured on each diet are the mean of the calibrators from Boehringer Mannheim, and triacylglycerol measurements made on the two occasions. After the initial were measured enzymatically with Roche Diagnostic kits four day diet record during the run in period, eight day diet on a Cobas Fara analyser (Roche Diagnostics). Calibration records were collected during each experimental period by and quality control was maintained by participation in the recording food eaten on two different days each week. Royal Australasian College of Pathologists Quality Assur- ance Programme. Coef®cient of variation for measurement Diets of cholesterol was 1.6% and for triacylglycerol 3.4%. High As participants in this study ate all meals at home and density lipoprotein cholesterol was measured in the super- followed their usual activities, detailed dietary counselling natant after precipitation of apo BÐcontaining lipoproteins was provided to individuals, and where appropriate their with phosphotungstate=magnesium chloride (Assmann et partners. The experimental diets were individually pre- al, 1983). LDL-C was calculated using the Friedewald scribed and based on estimated energy requirement and formula (Friedewald et al, 1972). Apolipoprotein AI and reported intake during the baseline. Diets were calculated apolipoprotein B concentrations were determined by im- (in increments of 100 kcal=419 kJ, to cover the range from munoturbidimetry using commercial kits from Boehringer 1800 kcal=7533 kJ to 3000 kcal=12556 kJ) on the basis of (Coef®cient of variation 2.6 and 6.0% respectively). Plasma information obtained from the four day diet record. Wal- lipids were extracted according to the procedure of Bligh & nuts contributed 20% of the total energy and 55% of energy Dyer (1959) and the various lipid fractions separated by from fat in the walnut diet with the remaining fat coming thin layer chromatography. Fatty acid composition of each from chosen by the participants according to pre- fraction was analysed by gas liquid chromatography (Holub ference and their daily eating patterns. During the Low Fat & Skeaff, 1987). The precision of the fatty acid measure- Diet all fat came from a variety of foods other than nuts. ment was determined by repeated analysis of a pooled The dietary methodology used a food exchange system, plasma sample. The coef®cient of variation for all fatty developed in this department (Chisholm et al, 1992), in acids present at greater than 1 mol % in the pooled sample order to facilitate the achievement of fat targets appropriate ranged from 1±4%. to a particular experimental diet. The system permitted the subjects to utilize a wide range of fat containing foods from meat or poultry, dairy, ®sh, made up dishes or baked Statistics product sources in order to achieve their fat from food For all variables, the two measurements made during the targets. A selection of recipes, with a fat per serving last week of each period were not signi®cantly different analysis, were supplied together with sample menu plans from one another so the means of the two values were used. for both the walnut and low fat diet phases. Some of the Preliminary analyses of the data included a repeated mea- recipes were for use in both phases of the study while sures ANOVA to determine whether the order of the others incorporating walnuts were designed speci®cally for interventions in¯uenced the relative treatment response. this intervention. It was assumed that these recipes would The means shown to compare the two diets are pooled be utilized not only by the study subjects but also for their over the two sequences. Differences between the two diets families, so information on how to adjust the fat per serving were compared using ANOVA with repeated measures. according to the number of servings a recipe may yield was also supplied. For example a 2500 kcal low fat diet Results permitted 83 gm fat from food. The walnut diet at the same energy levels was calculated to contain 46 gm fat All 21 subjects attended all appointments but only 16 from walnuts and 37 gm fat from food. Participants provided a full set of dietary records. Data are presented received an individual instruction folder containing details for these subjects. Body weight on the two experimental Walnuts and plasma fatty acids A Chisholm et al 14 Table 2 Mean (s.d.) plasma lipid and lipoprotein values on the baseline, Table 3 Fatty acid composition (mol%) of plasma walnut and low fat diets triacylglycerol

Baseline diet Walnut diet Low fat diet Fatty acid Low-fat Walnut

TC 6.24 5.99 6.11 C14:0 3.0 (1.3) 2.8 (1.3) mmol=l (0.61) (0.45)*b (0.75) C16:0 28.8 (2.4) 27.9 (4.3) C16:1 5.4 (1.2) 4.6 (1.0)* LDL 4.30 3.94 4.10 b C18:0 3.6 (1.4) 2.8 (1.0) mmol=l (0.62) (0.46)** (0.63) C18:1 37.7 (2.9) 33.3 (4.0)* HDL 1.06 1.21 1.18 C18:2n-6 13.8 (4.5) 19.8 (6.3)* mmol=l (0.33) (0.10)**b (0.11)**c C18:3n-3 1.1 (0.7) 2.0 (1.0)* C20:3n-6 0.3 (0.1) 0.4 (0.2) TG 1.95 2.00 1.86 C20:4n-6 0.9 (0.4) 1.0 (0.3) mmol=l (0.75) (0.80) (0.48) C20:5n-3 0.1 (0.1) 0.2 (0.1) C22:6n-3 0.4 (0.2) 0.4 (0.2) APO A 129.5 131.5 130.0 mg=dl (26.88) (22.28) (23.14) Values are reported as means (s.d.). Unless other- APO B 100.5 87.22 92.16 wise reported, fatty acids contributing less than mg=dl (30.93) (24.51)*a (21.37) 0.5 mol% are omitted. *P < 0.01. *P < 0.05; **P < 0.01. aWalnut vs low fat, baseline vs walnut. Table 4 Fatty acid composition (mol%) of plasma bBaseline vs walnut and low fat. cholesterol ester cBaseline vs low fat. Key Fatty acid Low-fat Walnut TC ˆ total cholesterol LDL ˆ low density lipoprotein cholesterol C14:0 1.1 (0.4) 1.1 (0.4) HDL ˆ high density lipoprotein cholesterol C16:0 13.8 (2.1) 12.2 (1.2)* TG ˆ total C16:1 4.1 (1.4) 3.4 (1.1) APO A ˆ apolipoprotein A1 C18:0 1.2 (1.0) 0.8 (0.2) APO B ˆ apolipoprotein B100. C18:1 19.1 (2.8) 15.7 (2.9)** C18:2n-6 49.4 (5.6) 55.3 (4.7)** C18:3n-3 0.8 (0.2) 0.8 (0.2) diets was the same, weight [mean (s.d.)] on the walnut diet C20:3n-6 0.6 (0.1) 0.5 (0.1) and low fat diet was 87.3 (15.6) and 87.3 (15.0), respec- C20:4n-6 4.8 (1.2) 4.8 (1.2) C20:5n-3 0.8 (0.4) 0.9 (0.3) tively. Plasma lipids, and apolipoproteins are presented in C22:6n-3 0.4 (0.1) 0.4 (0.1) Table 2. On both experimental diets total and LDL choles- terol and apo B were lower than at baseline, but the Values are reported as means (s.d.). Unless other- differences were signi®cant only for the walnut diet. Apo wise reported, fatty acids contributing less than 0.5 mol% are omitted. B was signi®cantly lower at the end of the walnut diet in *P < 0.05; **P < 0.01. comparison to the low fat diet (P < 0.05). Total and LDL cholesterol were slightly lower at the end of the walnut diet in comparison to the low fat diet but these differences were Table 5 Fatty acid composition (mol%) of plasma phospholipid not statistically signi®cant. The HDL values were similar on both experimental diets and signi®cantly higher than at Fatty acid Low-fat Walnut baseline (P < 0.01). Those who received the walnut diet C14:0 0.5 (0.1) 0.5 (0.2) ®rst (group 2) showed a reduction in both TC and LDL-C C16:0 31.4 (2.4) 31.5 (2.2) from baseline to the end of the walnut diet (70.17 and C16:1 0.9 (0.2) 0.8 (0.2)* 70.27 mmol=l, respectively) followed by a reduction from C18:0 14.2 (0.9) 9.0 (1.2)** the end of the walnut diet to the end of the low fat diet of C18:1 10.2 (1.2) 15.8 (2.9)** 70.27 and 70.24 mmol=l respectively. Group 1 receiving C18:2n-6 16.9 (2.1) 18.9 (2.1)** C18:3n-3 0.2 (0.1) 0.1 (0.1) the low fat diet ®rst showed a slight rise in TC and LDL-C C20:3n-6 2.4 (0.4) 2.3 (0.5) values from baseline ( ‡ 0.16 and ‡ 0.10 mmol=l re- C20:4n-6 1.4 (0.3) 1.4 (0.3) spectively) to the end of this diet phase but then a C20:5n-3 0.7 (0.2) 0.8 (0.2) marked drop from the end of the low fat diet of 70.50 C22:6n-3 2.4 (0.7) 2.0 (0.5) and 70.63 mmol=l to the end of the walnut diet. Fatty Values are reported as means (s.d.). Unless other- acids were measured in plasma triacylglycerol, cholesterol wise reported, fatty acids contributing less than and . The analysis of the plasma 0.5 mol% are omitted. triacylglycerol fatty acid values showed signi®cant changes *P < 0.05; **P < 0.01 in the levels of C16:1, C18:1, C18:2n-6 and C18:3n-3; the plasma cholesterol ester in levels of C16:0, C18:1, and low fat diet. The mean daily intake of walnuts was 78 g per C18:2n-6; and the plasma phospholipid in levels of C16:1, day on the walnut diet. This supplied 53 g of fat, 30 g of C18:0, C18:1, and C18:2n-6, between the walnut and low which came from and 5.9 g from linolenic fat diets. (Tables 3, 4 and 5). The mean nutrient intake acid. values on the two diets are shown in Table 6. Intakes of total, saturated and polyunsaturated fatty acids, carbohy- drate and (expressed as a percentage of total Discussion energy) were signi®cantly different on the two diets. The low fat diet was higher in saturated fat and dietary choles- Our aim in this study was to compare two diets of similar terol, the walnut diet higher in total and polyunsaturated energy content both meeting current recommendations for fat. and protein intakes were higher on the macronutrient intakes (National Cholesterol Education Pro- Walnuts and plasma fatty acids A Chisholm et al 15 Table 6 Mean (s.d.) daily nutrient intake on the walnut and low fat diets 0.4 after adjustment for total fat intake, this ®nding was the (n ˆ 16) most striking dietary factor associated with reduced risk Walnut Diet Low Fat Diet (Aschiro et al, 1996). In a secondary prevention trial de Lorgeril et al, 1994 randomised post myocardial infarction Energy (MJ=kcals) 9.4 (2.4)= 9.0 (2.4)= patients to a conventional `prudent' diet or a `Mediterra- 2255 (586) 2171 (596) nean' diet rich in . There was an Protein (% total energy) 17 (3) 19 (3)* Carbohydrate (% total energy) 40 (5) 46 (5)** approximately two thirds reduction in non fatal myocardial Cholesterol (mg) 230 (93) 320 (134)** infarction and cardiac death in the latter group. Although Fibre (g) 30 (8) 30 (10) con®dence intervals were relatively wide and it is dif®cult Total fat (% total energy) 38 (4) 30 (4)** to be certain that the results were entirely due to the effects SFA (% total energy) 10 (2) 12 (3)** MUFA (% total energy) 10 (2) 10 (2) of alpha linolenate rather than the other characteristics of PUFA (% total energy) 16 (3) 5 (1)** the Mediterranean diet, the data certainly support the Alcohol (% total energy) 4 (4) 5 (4) suggestion that this fatty acid has the potential to reduce cardiovascular risk (de Lorgeril et al, 1994). Alpha lino- *P < 0.05; **P < 0.01. lenic acid may be protective directly, in which case, the Key increased alpha linolenic content of plasma triacylglycerol SFA ˆ saturated fatty acids MUFA ˆ monounsaturated fatty acids during the walnut diet would be bene®cial. Alternately, the PUFA ˆ polyunsaturated fatty acids desaturation and elongation of alpha linolenic to eicosa- pentaenoic acid, leading to increased tissue levels of may be protective (Schmidt & Dyer- gramme, 1994). In the walnut diet the intention was for berg, 1994). In this regard the walnut diet did not produce participants to replace various other fat sources with the increased eicosapentaenoic acid levels in plasma triacyl- nuts provided. Compliance with the consumption of wal- , phospholipid or cholesterol ester. This ®nding is nuts was very good, as evidenced by the large relative consistent with that of Abbey et al (1994) who found that increases of linoleic acid in plasma triacylglycerol, by walnuts did not increase plasma eicosapentaenoic acid 6.0 mol% and alpha linolenic acid, by 0.8 mol%. The when compared to a reference diet with a P=S ratio of marked decrease in content of plasma triacylgly- 0.4-similar to our low fat diet-but is in contrast to that of cerol is another characteristic marker of increased linoleic Sabate et al (1993). The n-6=n-3 polyunsaturated fatty acid acid intake (Abbey et al, 1994; Kelley et al, 1993; Vasta et ratio in walnuts is roughly 5:1 (Quigley et al, 1995), al, 1996) and further con®rmation that compliance with the whereas in canola and ¯axseed , also advocated for walnut regimen was high. Regrettably, despite detailed their alpha linolenic acid content, the ratios are roughly 2:1 dietary instructions and regular reinforcement throughout (Quigley et al, 1995) and 1:4 (Layne et al, 1996), respec- the experimental period total energy from fat was higher on tively. Chan et al (1993) and others (Abbey et al, 1994; the walnut diet. The data given in Table 6 and examination Layne et al, 1996; Lichtenstein et al, 1993, Valsta et al, of the food records suggest that instead of replacing other 1996) have reported that when using alpha linolenic acid high fat foods with walnuts the participants were consum- rich oils if the n-6=n-3 ratio in the overall diet is high ing the raw nuts in addition to their usual food. Never- (namely > 5) an increase in plasma alpha linolenic acid can theless, the diets were essentially isocaloric and some food occur, but there is little, if any, increase in eicosapentaenoic substitution was evident. On the low fat diet in comparison acid content of plasma lipids above that of a standard to the walnut diet mean amounts consumed were increased reference diet. Increasing eicosapentaenoic acid levels by for meat, (164 vs 102 g=d), dairy products (228 vs consuming alpha linolenic acid-rich oils appears to require 217 g=d), bakery products (179 vs 164 g=d), and a low n-6=n-3 ratio (namely < 3) in the overall diet as well (167 vs 132 g=d). With hindsight it would have been as a high intake of alpha linolenic acid (Chan et al, 1993; preferable to have adopted an approach which did not Mantzioris et al, 1994, 1995), not easily achieved by rely on the fat exchange concept. However, our success supplementing a low fat diet with walnuts, as in the present with this approach in previous studies (Chisholm et al, experiment. 1992; Cox et al, 1995), encouraged its continued use and The ®ndings relating to lipids and lipoproteins are more we believe that despite a higher fat intake in the walnut dif®cult to interpret. It should be noted that on both group interesting ®ndings have nevertheless emerged. experimental diets total and LDL cholesterol and apolipo- The most relevant ®ndings concern the plasma fatty protein B were lower than at baseline, however this was acids. These ®ndings are broadly comparable with those of only signi®cant for the walnut diet. The only statistically Sabate et al (1993) and Abbey et al (1994) who studied signi®cant difference between the walnut and low fat diets healthy young men at low risk of coronary heart disease was a lower level of apo B though there was also a trend (CHD). The most striking change on the walnut diet was towards lower LDL and a higher HDL on the walnut diet. the increase in linoleic acid in all plasma fatty acid An explanation may be found in the lack of difference, fractions and the increase in alpha linolenic acid in between the two treatment groups in the myristic acid plasma triacylglycerol. There is reasonably convincing composition of plasma triacylglycerol. In a previous evidence that linoleic acid may be protective against study (Cox et al, 1995), in which the effects of , CHD (Riemersma et al, 1986; Hetzel et al, 1989; Roberts , and saf¯ower oil on plasma lipids were et al, 1993) and there is emerging evidence of the impor- compared, we reported that a saf¯ower oil diet reduced tance of dietary alpha linolenic acid. plasma total cholesterol by 10% in comparison to a butter The cardioprotective effects of linolenic acid have diet. Myristic acid content of plasma triacylglycerol on the recently been highlighted by results from the Health Pro- butter diet was 4.4 mol%, whereas on the saf¯ower oil diet fessionals study. Aschiro et al, (1996) found that a one it was 2.6 mol% (P < 0.01). In the present study, the percent increase in total energy from linolenic acid was myristic acid content of plasma triacylglycerol on the associated with a relative risk of myocardial infarction of walnut and low fat diets were 2.8 and 3.0 mol%, respec- Walnuts and plasma fatty acids A Chisholm et al 16 tively. Myristic acid has the strongest cholesterol raising Fraser GE, Sabate J, Beeson WL & Strahen M (1992): A possible properties of the saturated fatty acids (Katan et al, 1994) protective effect of consumption on risk of coronary heart disease. The Adventist Health Study. Arch. Intern. Med. 152, 1416±1424. and it is tempting to speculate that the equally low level of Friedewald WT, Levy RI & Fredrickson DS (1972): Estimation of the myristic acid in plasma triacylglycerol of the walnut and concentration of low density lipoprotein cholesterol in plasma, with- low fat groups explains why, despite an increase in linoleic out the use of the preparative ultracentrifuge. Clin. Chem. 18, 499± acid content of plasma triacylglycerol (by 6.0 mol%) simi- 502. lar to that found between the butter and saf¯ower oil diets Hetzel BS, Charnock JS, Dwyer T & McLennan PL (1989): Fall in coronary heart disease mortality in U.S.A. and Australia due to (by 6.8 mol%) in our earlier study, the walnut diet did not sudden death: evidence for the role of polyunsaturated fat. J. Clin. reduce plasma cholesterol concentrations. Epidemiol. 42, 885±93. Holub BJ & Skeaff CM (1987): Nutritional regulation of cellular phos- phatidylinositol. Methods in Enzymol. 141, 234±244. Katan MB, Zock PL & Mensink RP (1994): Effects of and fatty acids Conclusions on in humans: an overview. Am. J. Clin. Nutr. 60 (suppl), 1017S±1022S. Thus our data con®rm the potential for walnuts to in¯uence Kahn RH, Phillips RL, Snowdon DA & Choi W (1984): Association the fatty acid pro®le of plasma lipids in a way which might between reported diet and all cause mortality: twenty-one year follow be expected to reduce the risk of coronary heart disease. up on 27,530 adult Seventh-Day Adventists. Am. J. Epidemiol. 119, This may explain the cardioprotective effect of nuts 775±787. Kelley DS, Nelson GJ, Love JE Branch LB, Taylor PC, Schmidt PC, observed in studies of vegetarians. Other studies have Mackey BE & Iacono JM (1993): Dietary-linolenic acid alters tissue suggested an improvement in the overall lipoprotein pro®le fatty acid composition but not blood lipids, lipoproteins or coagulation in those consuming substantial quantities of walnuts. Fail- status in humans. Lipids 28, 533±537. ure to show more convincing differences between the Layne KS, Goh YK, Jumpsen JA, Ryan EA, Chow P & Clandinin MT walnut and low fat diets in this study may have been due (1996): Normal subjects consuming physiological levels of 18:3 (n-3) and 20:5 (n-3) from ¯axseed or ®sh oils have characteristic differences to the relatively similar intakes of myristic acid on both in plasma lipid and lipoprotein fatty acid levels. J. Nutr. 126, 2130± diets. Clearly any advice to increase nut consumption 2140. which might follow further con®rmation of these ®ndings Lichtenstein AH, Ausman LM, Carrasco W, Jenner JL, Gaultieri LJ, should indicate a note of caution to avoid an increase in Goldin BR, Ordovas JM, Schaefer EJ (1993): Effects of canola, corn and olive oils on fasting and postprandial plasma lipoproteins in humans total fat intake. as part of a National Cholesterol Education Program Step 2 diet. Arteriosclerosis and Thrombosis 13, 1533±1542. AcknowledgementsÐWe thank the participants for their commitment and Mantzioris E, James MJ, Gibson RA & Cleland LG (1994): Dietary enthusiasm. Mrs Margaret Waldron for expert venipuncture and support of substitution with an alpha-linolenic acid-rich increases participants. Mrs Jenny Lawrence, Lincoln University & the Walnut eicosapentaenoic acid concentrations in tissues. Am. J. Clin. Nutr. 59, Growers Group (South Canterbury) for the supply of the walnuts, and 1304±1309. Dr Geoffery Savage and Ms Lisa Schwartz of Lincoln University for the Mantzioris E, James MJ, Gibson RA & Cleland LG (1995): Differences exist in the relationships between dietary linoleic and ± linolenic acids analysis of the walnut fatty acids. and their respective metabolities. Am. J. Clin. Nutr. 61, 320±324. Marshall R (1993): Diet entry and storage = Diet Cruncher. Dunedin, New Zealand. References National Cholesterol Education Programme (1994): Second report of the expert panel on detection, evaluation, and treatment of high blood Abbey M, Noakes M, Belling B & Nestel P (1994): Partial replacement of cholesterol in adults (Adult treatment panel 11). Circulation 89, 1336± saturated fatty acids with or walnuts lowers total plasma 1432. cholesterol and low-density-lipoprotein cholesterol. Am. J. Clin. Nutr. Quigley RJ, Burlingham BA, Milligan GC & Gibson JJ (1995): Fats and 59, 995±999. Fatty Acids in New Zealand Foods. (1st ed.) Palmerston North, NZ: Ascherio A, Rimm EB, Giovannucci EL, Spiegelman D, New Zealand Institute for Crop & Food Research. Stampfer M & Willet WC (1996): Dietary fat and risk of coronary Recommendations of the European Atherosclerosis Society (1992): Pre- heart disease in men:cohort follow up study in the . BMJ vention of coronary heart disease: Scienti®c background and new 313, 84±90. clinical guidelines. Nutrition Metabolism and Cardiovascular Diseases Assmann G, Schriewer H, Schmidt G & Hegele E (1993): Quanti®cation 2, 113±156. of high density lipoprotein cholesterol by precipitation with phospho- Riemersma RA, DA, Butler S, Elton RA, Oliver M, Salo M, Nikkari tungstig acid=MgCl2. Clin. Chem. 29, 2026±2030. T, Vartiainen E, Puska P, Gey F, Rubba P, Mancini M & Fidanza Bligh DG & Dyer WJ (1959): A rapid method of total lipid extraction and F(1986): Linoleic acid content in and coronary heart puri®cation. Can. J. Biochem. Physiol. 37, 911±917 disease. Br. Med. J. Clin. Res. Ed. 292 (6533), 1423±1427. Burlingham BA, Milligan GC, Apimerika DE & Arthur JM (1993): Roberts TL, Wood DA, Riemersma RA, Gallagher PJ & Lampe FC FOOD®les Palmerston North, NZ: New Zealand Institute for Crop & (1993): Linoleic acid and risk of sudden cardiac death. Br. Heart J. Food Research. 70 (6), 524±9. Burr ML & Butland BK (1988): Heart disease in British vegetarians. Am. Sabate J, Fraser GE, Burke K, Knutsen SF, Bennett H & Lindsted KD J. Clin. Nutr. 48, 830±832. (1993): Effects of walnuts on serum lipid levels and blood pressure in Chan JK, McDonald BE, Gerrard JM, Bruce VM, Weaver BJ & Holub BJ normal men. N. Engl. J. Med. 328, 603±607. (1993): Effect of dietary alpha-linolenic acid and its ratio to linoleic Sabate J (1993): Does nut consumption protect against ischaemic heart acid on platelet and plasma fatty acids and thrombosis. Lipids 28, 811± disease? Eur. J. Clin. Nutr. 47 (suppl 1), S71±S75. 817. Schmidt EB & Dyerberg J (1994): -3 fatty acids. Current status in Chisholm A, Mann JI, Sutherland W, Williams S & Ball M (1992): Dietary cardiovascular medicine. Drugs 47, 405±424. management of patients with Familial Hypercholesterolaemia treated Thorogood M, Carter R, Ben®eld L, McPherson K & Mann JI (1987): with Simvistatin. Q. J. Med. 308±309, 824±831. Plasma lipids and lipoprotein cholesterol concentrations in people with Cox CM, Mann JI, Sutherland WH, Chisholm A & Skeaff CM (1995): different diets in Britain. BMJ 295 (6594), 351±353. Effects of , butter, and saf¯ower oil on lipids and lipopro- Thorogood M, Mann J, Appleby P & McPherson K (1994): Risk of death teins in persons with moderately elevated cholesterol levels. J. Lipid from cancer and ischaemic heart disease in meat and non-meat eaters. Res. 36, 1787±1795. BMJ 308 (6945), 1667±1670. de Lorgeril M, Renaud S., Mamelle N, Salen P, Martin J-L, Monjaud I, Vasta LM, Salminen I, Aro A & Mutanen M (1996): Alpha-linolenic acid Guidollet J, Touboul P & Delaye J (1994): Mediterranean alpha- in rapeseed oil partly compensates for the effect of ®sh restriction linolenic acid-rich diet in secondary prevention of coronary heart on plasma long chain n-3 fatty acids. Eur. J. Clin. Nutr. 50, 229± disease. Lancet 343, 1454±1459. 235.