The Association of Diet and Thrombotic Risk Factors in Healthy Male Vegetarians and Meat-Eaters

The association of diet and thrombotic risk factors in healthy male vegetarians and meat-eaters

DLi1, A Sinclair1*, N Mann1, A Turner2, M Ball3, F Kelly1, L Abedin1 and A Wilson1

1Departments of Food Science, RMIT University, Melbourne 3001, Australia; 2Medical Laboratory Science, RMIT University, Melbourne 3001, Australia; and 3School of Nutrition and Public Health, Deakin University, Malvern, VIC 3144, Australia

Objective: The aim of this study was to assess thrombosis tendency in subjects who were habitual meat-eaters compared with those who were habitual vegetarians. Design: Cross-sectional comparision of habitual meat-eaters and habitual vegetarians. Setting: Free living subjects. Subjects: One hundred and thirty-nine healthy male subjects (vegans n 18, ovolacto vegetarians n 43, moderate-meat-eaters n 60 and high-meat-eaters n 18) aged 20 ± 55 y who were recruited in Melbourne. Outcome measures: Dietary intake was assessed using a semi-quantitative Food Frequency Questionnaire. The parameters of thrombosis were measured by standard methods. Results: Saturated fat and cholesterol intakes were signi®cantly higher and polyunsaturated fat (PUFA) was signi®cantly lower in the meat-eaters compared with vegetarians. In the meat-eaters, the platelet phospholipids AA levels were signi®cantly higher than in the vegetarians, but there was no increase in ex vivo platelet aggregation and plasma 11-dehydro thromboxane B2 levels. Vegetarians, especially the vegans, had a signi®cantly increased mean collagen and ADP stimulated ex vivo whole blood platelet aggregation compared with meat-eaters. The vegan group had a signi®cantly higher mean platelet volume than the other three dietary groups. However, meat-eaters had a signi®cantly higher cluster of cardiovascular risk factors compared with vegetarians, including increased body mass index, waist to hip ratio, plasma total cholesterol (TC), triacylglycerol and LDL-C levels, ratio of TC=HDL-C and LDL-C=HDL-C and plasma factor VII activity. Conclusions: Consumption of meat is not associated with an increased platelet aggregation compared with vegetarian subjects. Sponsorship: Meat Research Corporation, Australia. Descriptors: vegetarians; meat; hemostatic factors; platelet aggregation; lipoprotein lipids; fatty acids

Introduction Arterial thrombosis is generally recognised to play a major role in the transition from stable to acute ischaemic The major cause of premature mortality in Western Society heart disease, manifested by unstable angina, acute myo- is cardiovascular disease. High fat, particularly saturated fat cardial infarction and sudden death. Beside local stimuli (SFA) intake has been implicated as one of the major such as disruption of plaque, systemic thrombogenic factors causes primarily because it raises blood low density lipo- (platelet hyperreactivity, increased ®brinogen (FIB) con- protein cholesterol (LDL-C) levels. High meat consump- centration and factor VII activity (VII), defective ®brino- tion has been linked with high plasma cholesterol levels, lysis and abnormalities of blood ¯ow) may contribute to the since most meats are sold with attached depot or visible fat occurrence, extent, and persistence of coronary thrombosis which is rich in SFA (Sinclair & O'Dea, 1990). Acute and its clinical sequelae. Therefore, it may be bene®cial to clinical disorders of the cardiovascular system are often detect thrombosis tendency in both patients and healthy caused by formation of thrombus and subsequent occlusion people by using appropriate laboratory tests. of a blood vessel. The traditional cardiovascular disease Platelet aggregation is an early event in the development risk factors of a positive family genetic history, cigarette of thrombosis. It is initiated by the production of throm- smoking, elevated blood LDL-C and blood pressure do not boxane A (TXA ) produced from arachidonic acid (AA, account for the majority of cases at risk for future throm- 2 2 20:4 n-6), a long chain n-6 polyunsaturated fatty acid botic events. (PUFA) in the platelet membrane. Since the Western diet contains an excess of n-6 PUFA, it is argued that this situation may lead to an overproduction of TXA2 and *Correspondence: Dr A Sinclair, Department of Food Science, RMIT therefore an increase in thrombosis tendency (Berner, University, GPO Box 2476V, Melbourne, VIC 3001, Australia. 1993). Meats are the major source of dietary arachidonic Guarantor: A Sinclair Contributors: D Li, A Sinclair, N Mann, A Turner and M Ball initiated the acid and this arachidonic acid may contribute to the pool of study. D Li prepared the drafts of paper, did laboratory assays, collected tissue arachidonic acid and thereby to an increased risk of data, and did the statistical analysis. A Sinclair selected the study site, platelet aggregation and thrombosis (Mann et al, 1995). On supervised the project, and secured the funding. N Mann, F Kelly, L the other hand, lean meats and ®sh are also the major Abedin and A Wilson helped data collection and provided technical sources of eicosapentaenoic acid (EPA, 20:5 n-3) and assistance. All investigators contributed to the drafts of paper. Received 7 November 1998; revised 15 February 1999; accepted docosahexaenoic acid (DHA, 22:6 n-3) which may provide 23 February 1999 a protective effect against thrombin formation. Habitual diet and thrombotic risk factors in man DLiet al 613 In general, vegetarians have lower blood pressure, lower Blood specimen collections fat intake, lower body mass index, and higher antioxidant Subjects attended the RMIT Medical Centre in the morning intake than omnivores. These factors are known to reduce following an overnight fast. Subjects were allowed to sit the risk of cardiovascular disease (CVD). However, vege- relaxed for 10 min, and then venous blood was taken in tarians also have a relatively high intake of linoleic acid plain, EDTA, citrate and CTAD (citrate, theophylline, (18:2 n-6, a precursor of AA) and low EPA and DHA adenosine and dipyridamole) vacuum tubes with 21-gauge intakes (Sanders & Roshanai, 1992). needles. After blood collection, the subject's weight, Previous short term intervention studies from this group height, waist=hip ratio and blood pressure were measured. have indicated that low fat diets enriched in AA signi®- Whole blood platelet aggregation, full blood examination cantly increased plasma AA levels and TXA2 production, and platelet isolation were performed during the 3 h follow- whereas diets containing AA and long chain (LC) n-3 ing blood sampling. Plasma and serum samples were PUFA or diets rich in LC n-3 PUFA do not raise TXA2 prepared during the 2 h after blood was drawn, aliquoted levels (Sinclair & Mann, 1996; Mann et al, 1997). One of into separate tubes and stored at 720C until analysis. the dif®culties with previous studies is that they have been only conducted for short periods (2 ± 3 weeks), due to the Full blood examination dif®culty subjects had consuming higher amounts of meat A full blood cell count was performed on a Coulter STKR than normal. Therefore, it was decided to conduct a cross- analyser (Coulter Electronics Inc, Hialeah, USA). sectional study of thrombosis risk indicators in healthy male subjects who habitually consumed high quantities of Plasma lipid and lipoprotein lipids meat. In this study, we compared blood cell counts, Triacylglycerol (TAG) and total cholesterol (TC) concen- lipoprotein lipids, coagulation factors, plasma and platelet tration of fasting EDTA plasma were determined by stan- phospholipid fatty acid pro®les, plasma 11-dehydro throm- dard enzymatic methods on a centrifugal autoanalyser boxane B2 (11-dehydro TXB2) and ex vivo agonist induced (Hitachi Autoanalyser System 705, Japan), using commer- whole blood platelet aggregation in meat consumers and cially available kits (Boehringer Mannheim, Sydney, Aus- vegetarians. The latter group were chosen because they tralia) as published previously (Sravropoulous & Crouch, have a low intake of AA. We hypothesised that meat 1974). High density lipoprotein cholesterol (HDL-C) was consumption does increase the levels of blood indicators determined after precipitating all plasma lipoproteins of increased thrombosis potential risk. except HDL with polyethylene glycol 6000 (PEG 6000). LDL-C was calculated using the Friedwald equation as developed by DeLong et al (1986). Methods Platelet and plasma fatty acids Two CTAD tubes of whole blood were spun at 900 rpm for Subjects 10 min. Platelet rich plasma (PRP) was removed and This project was approved by the Human Research Ethics platelets were isolated from PRP using the method pub- Committee of RMIT University, and all subjects gave lished by Castaldi & Smith (1980). Remaining platelet poor written informed consent. One hundred and forty seven blood was further spun at 3000 rpm for 15 min, and platelet healthy male non-smokers aged between 20 ± 50 y were poor plasma (PPP) was removed and stored at 7 20C for recruited through advertisements in University newsletters later analysis of 11-dehydro TXB2. Total lipid of platelets and local newspapers. The exclusion criteria for this study and EDTA plasma was extracted with chloroform:methanol were: evidence of CVD, hypertension, renal disease, hyper- 1:1 (C:M, v=v) containing 10 mg=L of butylated hydroxy- lipaemia, hematological disorders, diabetes, family history toluene (Labco, VIC, Australia), and 10 mg=L of C17:0 of CVD, excess alcohol intake and drug therapy. According phospholipid (diheptadecanoyl) (Nu-Chek-Prep, INC., MN, to their habitual dietary intake (based on the semi-quanti- USA) as internal standard as reported previously (Sinclair tative Food Frequency Questionnaire (FFQ)), the subjects et al, 1987). The platelet and plasma phospholipid (PL) were divided into four groups vegan (n 18), ovolacto fractions were separated by thin-layer chromatography. The vegetarian (n 46), moderate-meat-eater (n 65) and methyl esters of the fatty acids of the platelet and plasma high-meat-eater (n 18). A vegan was de®ned as someone PL fractions were prepared by saponi®cation using KOH who ate no meat and eggs and dairy products less than six (0.68 mol=L in methanol) followed by transesteri®cation times per year. An ovolacto vegetarian was de®ned as with 20% boron tri¯uoride (BF3) in methanol (Sinclair et someone who ate meat not more than six times per year al, 1987). The fatty acid compositions of platelet and but consumed eggs and dairy products freely. A high-meat- plasma PL were determined by gas-liquid chromatography eater was de®ned as those who consumed 285 g meat=d as described previously (Sinclair et al, 1987). and moderate-meat-eater was de®ned as consuming < 285 g meat=d (raw weight). For the subjects to be classi- Plasma 11-dehydro thromboxane B2 concentration ®ed into these categories, they had to have been practising The concentration of 11-dehydro TXB2 in platelet poor their diets for at least six months prior to the study. plasma (PPP) was determined using an enzyme immuno- assay (EIA) method with commercially available EIA kits (Cayman Chemical Company, MI, USA) as described Dietary intake elsewhere (Pradelles et al, 1985). The dietary intake data for each subject was assessed using the semi-quantitative FFQ (Fidanza et al, 1994) and calcu- Plasma hemostatic factors lated using the Diet=1 Version 4 software (Xyris Software, Coagulation and ®brinolysis parameters (prothrombin time Pty Ltd, Highgate Hill, QLD, Australia) with NUTTAB 95 (PT), ®brinogen (FIB), activated partial thromboplastin database based on Composition of Foods, Australia. time (APTT), factor VII activity (VII), plasminogen and Habitual diet and thrombotic risk factors in man DLiet al 614 antithrombin III (ATIII)) were determined in citrated Table 2 reports dietary intake data of the four dietary plasma using the ACL 200 system with commercially groups. There was no meat consumption in the vegans. available kits (Coulter=IL Ltd. Italy) as described by Three ovolacto vegetarians had an intake of a small amount Messmore (1983). of white meat (less than three times a year). Both ovolacto vegetarian and vegan groups had signi®cantly lower ®sh Agonist induced ex vivo whole blood platelet aggregation intakes than both high- and moderate-meat-eater groups. The ex vivo agonist stimulated citrated whole blood platelet Only ten ovolacto vegetarians and three vegan subjects aggregation using collagen (2 mg=L), arachidonic acid consumed ®sh. High-meat-eaters had a signi®cantly higher (1 mmol=L) and adenosine diphosphate (ADP, 8 mmol=L energy intake than the other three dietary groups. However, and 17 mmol=L) as agonists was measured as maximal the energy sources were quite different between the four aggregation (O at 5 min) and rate of aggregation (slope, dietary groups. Protein, total fat, SFA and monounsaturated O=min) using a two channel Whole Blood Aggregometer fat (MUFA) as % of energy, showed a signi®cant decreas- (Model 540VS, Chrono-log Corporation, USA) as reported ing trend from high-meat-eaters to moderate-meat-eaters to by Ingerman-Wojenski & Silver (1984). ovolacto vegetarians and vegans. However, PUFA as % of energy was signi®cantly increased from high-meat-eaters to moderate-meat-eaters to ovolacto vegetarians and vegans. Statistical analyses Vegans do not eat dairy products and meats, and their fat The data analyses were performed using a StatView soft- intake is derived mostly from seeds, nuts, avocado and ware program. Descriptive statistics were initially per- vegetable oils. formed. ANOVA was used to establish if differences Cholesterol intake showed a signi®cant decreasing trend existed between the dietary groups for each parameter. If from high-meat-eaters to moderate-meat-eaters to ovolacto a signi®cant difference was found, a further multiple vegetarians to vegans. The small amount of cholesterol in comparison test was performed using Fisher's post hoc the vegan diet was derived from egg-containing pastry. tests to determine differences between each pair of dietary Dietary non-meat SFA densities showed a signi®cant groups. The values are reported as meanÆ s.d. P values decrease from high-meat-eaters to moderate-meat-eaters were two tailed and P < 0.05 was considered as signi®cant. and ovolacto vegetarians to vegans. Non-meat PUFA densities showed an opposite trend to SFA. Both high- and moderate-meat eater groups had a sig- Results ni®cantly higher red blood cell count (RBC) than both One hundred and forty seven healthy male subjects com- ovolacto vegetarian and vegan groups (Table 3). The vegan pleted the study. However, eight subjects were excluded group had a signi®cantly lower platelet count (PLT) than from the ®nal results due to hypertension, consumption of the ovolacto vegetarian group. Mean platelet volume dietary supplements, unreliable reporting of dietary intakes (MPV) was signi®cantly greater in the vegan group than (based on energy calculation of basal metabolic require- in the other three dietary groups. Vegans had a signi®cantly ment, if below 20% subjects were excluded) and evidence lower plasma factor VII activity (VII) than moderate-meat- of family history of CVD. Therefore, a of total one hundred eaters. There were no signi®cant differences between the and thirty nine subjects (18 high-meat-eaters, 60 moderate- four dietary groups in white blood cell count (WBC), meat-eaters, 43 ovolacto vegetarians and 18 vegans) were plasma prothrombin time (PT), activated partial thrombo- included in the ®nal results. plastin time (APTT), ®brinogen (FIB), plasminogen and Table 1 shows the physiological characteristics of the antithrombin III (ATIII) levels. Both ovolacto vegetarian four dietary groups. The median age (range) were 34.5 y and vegan groups had signi®cantly lower plasma TC and (21 ± 50) for the high-meat-eater group, 38 (21 ± 55) for the LDL-C levels, TC=HDL-C and LDL-C=HDL-C ratios moderate-meat-eater group, 31 (22 ± 54) for the ovolacto compared with both high- and moderate-meat-eater vegetarian group and 31 (22 ± 50) for the vegan group. The groups. Vegans had a signi®cantly lower HDL-C than mean body mass index (BMI) was signi®cantly greater in moderate-meat-eaters and TAG than high-meat-eaters. both the high- and moderate-meat-eater groups than in both There was no signi®cant difference between the four diet- the ovolacto and vegan groups. The mean ratio of waist to ary groups for plasma 11-dehydro TXB2 concentrations; hip measurement showed a decreasing trend from both however, both the ovolacto vegetarian and the vegan high- and moderate-meat-eater groups to ovolacto vegetar- groups tended to have a higher mean level than meat ians and to vegans. Mean systolic and diastolic blood eating groups. pressure showed a decreasing trend from high-meat-eaters Fatty acid compositions of plasma PL are shown in the to moderate-meat-eaters to ovolacto vegetarians to vegans. Table 4. Plasma PL SFA were signi®cantly lower in the

Table 1 Characteristics of subjects in the dietary groups (meanÆ s.d.)

High meat Moderate meat Ovolacto Vegan (n 18) (n 60) (n 43) (n 18)

Age (y) 34.2Æ 9.4 38.3Æ 7.3 34.9Æ 9.0moa 33.0Æ 7.7moa BMI 27.0Æ 3.4 26.4Æ 3.4 23.6Æ 2.8hoc, moc 23.3Æ 3.5hv, mvc Waist=Hip 0.88Æ 0.06 0.88Æ 0.05 0.86Æ 0.04moa 0.85Æ 0.05mva Systolic BP 129.8Æ 14.6 125.7Æ 10.5 124.4Æ 13.4 122.1Æ 9.7 Diastolic BP 83.4Æ 8.0 81.7Æ 7.9 77.7Æ 10.4hoa, moa 77.7Æ 7.8

High meat Moderate meat Ovolacto Vegan (n 18) (n 60) (n 43) (n 18)

White meat (g) 109.6Æ 51.6 53.7Æ 27.1hmc 0.2Æ 0.7hoc, moc 0.0Æ 0.0hvc, mvc Red meat (g) 270.4Æ 118.0 108.9Æ 41.1hmc 0.0Æ 0.0hoc, moc 0.0Æ 0.0hvc, mvc W R (g)d 387.2Æ 122.0 163.2Æ 48.4hmc 0.2Æ 0.7hoc, moc 0.0Æ 0.0hvc, mvc Liver (g) 2.8Æ 6.6 1.7Æ 4.8 0.0Æ 0.0hoc, moc 0.0Æ 0.0hvc, mvc Kidney (g) 0.7Æ 1.9 0.2Æ 1.0 0.0Æ 0.0hoc, moc 0.0Æ 0.0hvc, mvc Fish (g) 27.7Æ 23.9 26.1Æ 18.3 1.4Æ 3.4hoc, moc 0.7Æ 1.8hvc, mvc Energy (MJ) 1.64Æ 3.31 1.13Æ 2.48hmc 1.15Æ 2.46hoc 1.18Æ 2.44hvc Protein (% of E) 19.4Æ 1.3 17.9Æ 2.2hmb 14.7Æ 2.2hoc, moc 14.1Æ 2.4hvc, mvc Carbo. (% of E)e 40.3Æ 4.4 45.7Æ 6.8hmb 50.9Æ 6.4hoc, moc 57.4Æ 5.4hvc, mvc, ovc Total fat (% of E) 37.8Æ 4.2 32.8Æ 6.1hmb 32.7Æ 6.2hob 28.2Æ 4.3hvc, mvb, ovb SFA (% of E) 17.4Æ 2.7 14.3Æ 3.2hmc 11.9Æ 4.0hoc, moc 6.6Æ 1.6hvc, mvc, ovc MUFA (% of E) 14.9Æ 1.4 13.0Æ 2.8hma 12.9Æ 3.2hca 11.7Æ 2.7hvc PUFA (% of E) 5.6Æ 1.4 5.6Æ 2.3 7.9Æ 2.5hoc, moc 9.8Æ 3.1hvc, mvc, ovb Alcohol (% of E) 2.6Æ 1.9 3.2Æ 3.4 1.7Æ 3.1mob 0.4Æ 0.7hva, mvc Cholesterol (mg) 611Æ 150 332Æ 107hmc 197Æ 117hoc, moc 22Æ 22hvc, mvc, ovc non-meat densityf Total fat (g=MJ) 8.6Æ 1.5 7.7Æ 1.7 8.6Æ 2.0moa 7.4Æ 1.2hva, ova SFA (g=MJ) 3.8Æ 0.9 3.1Æ 0.9hmb 3.1Æ 1.0hob 1.7Æ 0.4hvc, mvc, ovc MUFA (g=MJ) 2.8Æ 0.5 2.7Æ 0.7 3.4Æ 1.0hoa, moc 3.1Æ 0.7 PUFA (g=MJ) 1.4Æ 0.5 1.3Æ 0.5 2.1Æ 0.7hoc, moc 2.6Æ 0.8hvc, mvc, ovb Cholesterol (mg=MJ) 24.1Æ 7.6 20.7Æ 10.9 17.3Æ 10.1hoa 1.9Æ 2.0hvc, mvc, ovc

aP < 0.05; bP < 0.01; cP < 0.001. dW R white meat red meat, raw weights. eCarbo. carbohydrate. fThe nutrient densities of the non-meat component were calculated by removing the meat proportion from the original dietary data. hmhigh-meat vs moderate-meat; hohigh-meat vs ovolacto; momoderate-meat vs ovolacto; hvhigh-meat vs vegan; mvmoderate- meat vs vegan; ovovolacto vs vegan.

Table 3 Hematological and biochemical parameters of four dietary groups (meanÆ s.d)

High meat Moderate meat Ovolacto Vegan Parameters (n 18) (n 60) (n 43) (n 18)

WBC (6109}=L) 6.66Æ 1.03 6.07Æ 1.22 6.36Æ 1.29 5.96Æ 1.69 RBC (61012=L) 5.17Æ 0.19 5.02Æ 0.30 4.83Æ 0.37hoc, mob 4.76Æ 0.24hvc, mvb PLT (6109=L) 200.5Æ 38.2 214.3Æ 46.5 223.4Æ 55.4 192.5Æ 51.7ova MPV (fL) 8.33Æ 0.87 8.53Æ 0.92 8.60Æ 0.96 9.53Æ 1.30hvc, mvc, ovb Fibrinogen (g=L) 3.31Æ 0.82 3.46Æ 0.80 3.45Æ 0.77 3.26Æ 0.67 PT (INR)d 0.98Æ 0.09 0.97Æ 0.09 1.01Æ 0.16 0.99Æ 0.06 APTT (s) 35.9Æ 3.7 35.8Æ 3.4 35.9Æ 4.6 35.0Æ 2.4 AT III (%) 94.6Æ 12.4 90.8Æ 11.8 91.3Æ 11.4 90.4Æ 10.4 Plasminogen (%) 102.2Æ 10.9 102.1Æ 14.4 100.1Æ 15.2 94.7Æ 12.4 Factor VII (%) 98.8Æ 22.0 102.8Æ 19.3 96.2Æ 20.4 89.0Æ 14.3mvb TC (mmol=L) 4.59Æ 0.75 4.50Æ 0.73 3.60Æ 0.80hoc, moc 3.52Æ 0.83hvc, mvc LDL-C (mmol=L) 2.94Æ 0.68 2.93Æ 0.69 2.11Æ 0.64hoc, moc 2.14Æ 0.70hvc, mvc HDL-C (mmol=L) 1.08Æ 0.28 1.09Æ 0.25 1.03Æ 0.24 0.96Æ 0.14mva TC=HDL-C 4.46Æ 1.24 4.34Æ 1.16 3.66Æ 1.19hoc, mob 3.69Æ 0.90hva, mva LDL-C=HDL-C 2.89Æ 0.91 2.86Æ 0.96 2.19Æ 0.93hoc, mcc 2.25Æ 0.71hva, mva TAG (mmol=L) 1.49Æ 0.86 1.27Æ 0.50 1.20Æ 0.66 1.08Æ 0.52hva TAG=HDL-C 1.54Æ 1.19 1.29Æ 0.72 1.27Æ 0.82 1.18Æ 0.68 11-dehydro TXB2 (ng=L) 89Æ 34 84Æ 32 97Æ 40 101Æ 54 aP < 0.05; bP < cP < 0.001; dINR International Normalised Ratio hmhigh-meat vs moderate-meat; hohigh-meat vs ovolacto; momoderate-meat vs ovolacto; hvhigh-meat vs vegan; mvmoderate-meat vs vegan; ovovolacto vs vegan. vegan group than in both the high- and moderate-meat- high- and the moderate-meat-eater groups. The proportion eater groups. The moderate-meat-eater group had a sig- of 22:4 n-6 was signi®cantly lower in the moderate-meat- ni®cantly higher plasma palmitic acid (16:0) in plasma than eater group than in both the ovolacto vegetarian and the the vegan group. There was a signi®cant trend for arachidic vegan groups. The proportion of total n-3 PUFA, 20:5 n-3, acid (20:0) to decrease from the high-meat-eaters to the 22:5 n-3, and 22:6 n-3, and n-3=n-6 ratio were signi®cantly moderate-meat-eaters to the ovolacto vegetarians to higher and the AA=EPA ratio were signi®cantly lower in vegans. The PUFA level in the plasma PL was signi®cantly both the high- and the moderate-meat-eater groups than in higher in both the ovolacto vegetarian and the vegan groups both the ovolacto vegetarian and the vegan groups. than in both the high- and the moderate-meat-eater groups. Table 5 reports the results of fatty acid compositions of Total n-6 PUFA and 18:2 n-6 proportions showed a sig- platelet PL. The moderate-meat eater group had a signi®- ni®cant trend to increase from the high- and the moderate- cantly lower total MUFA proportion than the both the meat-eater to the ovolacto vegetarian to the vegan group. ovolacto vegetarian and the vegan groups. The proportions The ovolacto vegetarian group had a signi®cantly lower of 18:2 n-6 and 22:4 n-6 were signi®cantly higher in both proportion of plasma PL 20:4 n-6 compared with both the the ovolacto vegetarian and the vegan groups than in both Habitual diet and thrombotic risk factors in man DLiet al 616 Table 4 Fatty acid composition of plasma phospholipids of the four dietary groups (% of total fatty acid, meanÆ s.d.)

High meat Moderate meat Ovolacto Vegan Fatty acid (n 18) (n 60) (n 43) (n 18)

14:0 0.3Æ 0.1 0.3Æ 0.1 0.3Æ 0.1moa 0.3Æ 0.1 16:0 27.2Æ 0.9 27.3Æ 1.3 27.0Æ 1.4 26.4Æ 1.6mva 18:0 13.3Æ 1.1 13.1Æ 0.9 13.1Æ 1.0 13.2Æ 1.0 20:0 0.3Æ 0.2 0.3Æ 0.1hma 0.2Æ 0.1hoc, moc 0.1Æ 0.0hoc, mvc, ovb Total SFA 41.2Æ 1.4 40.9Æ 1.2 40.6Æ 1.4 40.0Æ 1.1hvb, mva 16:1 n-9 0.2Æ 0.0 0.2Æ 0.0 0.2Æ 0.0 0.2Æ 0.0 16:1 n-7 0.6Æ 0.2 0.6Æ 0.3 0.5Æ 0.2 0.4Æ 0.1hva, mvb 17:1 n-9 0.2Æ 0.0 0.1Æ 0.0hmb 0.1Æ 0.0hoc, moc 0.1Æ 0.0hvc, mvc 18:1 n-9t 0.2Æ 0.1 0.3Æ 0.1 0.3Æ 0.21hob 0.3Æ 0.2 18:1 n-7t 0.5Æ 0.1 0.4Æ 0.1 0.5Æ 0.2 0.4Æ 0.2hva, ova Total Trans 0.7Æ 0.2 0.7Æ 0.2 0.8Æ 0.3 0.7Æ 0.3ova 18:1 n-9c 9.8Æ 1.1 9.4Æ 1.4 9.6Æ 1.2 9.6Æ 0.9 18:1 n-7c 1.4Æ 0.2 1.5Æ 0.2 1.4Æ 0.2 1.7Æ 0.5hvc, mvb, ovb Total MUFA 12.8Æ 1.0 12.6Æ 1.2 12.6Æ 1.3 12.6Æ 1.2 18:2 n-6 20.6Æ 2.6 22.4Æ 2.8hma 26.0Æ 3.1hoc, moc 26.1Æ 2.5hvc, mvc 20:3 n-6 3.6Æ 0.8 3.5Æ 0.7 3.3Æ 0.9 3.4Æ 0.9 20:4 n-6 10.6Æ 1.6 10.5Æ 1.7 9.5Æ 1.9hoa, moa 10.6Æ 1.5 22:4 n-6 0.4Æ 0.1 0.4Æ 0.1 0.5Æ 0.1moa 0.5Æ 0.1mvb 22:5 n-6 0.2Æ 0.1 0.2Æ 0.1 0.3Æ 0.1 0.2Æ 0.1 Total n-6 35.5Æ 2.1 37.0Æ 2.1hma 39.5Æ 2.6hoc, moc 40.8Æ 1.7hvc, mvc, ova 18:3 n-3 0.2Æ 0.1 0.2Æ 0.1 0.3Æ 0.1moc 0.3Æ 0.1mvb 20:5 n-3 1.1Æ 0.5 1.0Æ 0.3 0.7Æ 0.3hoc, moc 0.6Æ 0.3hvc, mvc 22:5 n-3 1.3Æ 0.2 1.2Æ 0.2 1.1Æ 0.2hob, mob 1.0Æ 0.3hvb, mvb 22:6 n-3 3.4Æ 1.0 3.3Æ 0.8 2.2Æ 0.7hoc, moc 2.0Æ 0.4hvc, mvc Total n-3 6.0Æ 1.3 5.8Æ 1.0 4.2Æ 1.0hoa, moc 4.0Æ 0.7hvc, mvc Total PUFA 41.5Æ 2.0 42.7Æ 1.8hma 43.8Æ 2.2hoc, mob 44.8Æ 1.3hvc, mvc n-3=n-6 0.17Æ 0.04 0.16Æ 0.03 0.11Æ 0.03hoc, moc 0.10Æ 0.02hvc, mvc AA=EPA 9.8Æ 2.9 11.2Æ 3.3 17.5Æ 11.5hoc, moc 21.7Æ 10.4hvc, mvc SFA=PUFA 0.10Æ 0.07 0.96Æ 0.06 0.93Æ 0.08hob, moa 0.89Æ 0.06hvc, mvc

aP < 0.05; bP < 0.01; cP < 0.001. hmhigh-meat vs moderate-meat; hohigh-meat vs ovolacto; momoderate-meat vs ovolacto; hvhigh-meat vs vegan; mvmoderate-meat vs vegan; ovovolacto vs vegan.

the high- and the moderate-meat-eater groups. There was a speculated by Berner (1993) and Okuyama et al (1996) that signi®cant trend for 20:4 n-6 to decrease from the high- and the AA intake from meat and eggs increases the risk of the moderate-meat-eater groups to ovolacto vegetarian to thrombotic events. We hypothesised that subjects who vegan groups. Proportions of 20:5 n-3, 22:5 n-3, 22:6 n-3 consumed a high intake of meat would have high levels and total n-3 PUFA, and ratio of n-3=n-6 were signi®cantly of platelet AA relative to non-meat eaters, and these lower in the vegans group than the ovolacto vegetarian, subjects would have a higher platelet aggregability and high- and moderate-meat-eater groups. thromboxane production compared with non-meat eaters. Agonist induced whole blood platelet aggregation The results showed that high-meat-eaters consumed results are reported in Table 6. Collagen induced whole approximately 191 mg of AA per day compared with blood platelet aggregation (expressed as O at 5 min and as 100 mg for the moderate-meat-eaters, 30 mg for the ovo- slope O=min) was signi®cantly greater in both ovolacto lacto vegetarians and 1 mg per day for the vegans. The high vegetarian and vegan groups than in both high- and AA intake was associated with signi®cantly increased moderate-meat-eater groups. The high-meat-eater group platelet AA levels in both the high- and moderate-meat- experienced the weakest collagen-induced whole blood eaters compared with the vegans but these differences were platelet aggregation. Whole blood platelet aggregation relatively small (24.4% and 24.5% vs 23.0%). Other studies induced by two ADP concentrations (8 mM and 17 mM) of vegetarians and omnivores have also reported small gave similar results. It increased stepwise from high-meat- differences in the plasma or platelet AA levels, with the eaters to moderate-meat-eaters to ovolacto vegetarians to omnivores having about a 5 ± 10% increase in the propor- vegans in O at 5 min. tion of AA (Phinney et al, 1990; Sanders & Roshanai, 1992). These data indicate that dietary AA from high intakes of meat is not associated with major increases in Discussion platelet or plasma AA levels ( < 10% difference) compared Results from experimental, epidemiological and cross-cul- with vegans (zero intake of AA), nor were there increases tural studies suggest that diet plays an important role in the in platelet aggregation. This is in contrast with the assertion etiology of CHD (Keys, 1980; Kromhout et al, 1982; Kushi of Berner (1993). In this study, the high dietary LA (nearly et al, 1985). Acute clinical events of the cardiovascular 10% of energy in the vegetarian subjects) provided suf®- system generally involve thrombus formation, migration cient substrate for nearly maximal accretion of AA in and subsequent vascular occlusion. The aim of this study platelets, as predicted by Lands et al (1992). The data was to assess the thrombosis potential risk factors of showed that there was a relationship between meat intake healthy subjects who were habitual meat-eaters compared and platelet aggregation and TXA2 production but it was with those who were habitual vegetarians, since it has been opposite to that predicted. That is, it was found that there Habitual diet and thrombotic risk factors in man DLiet al 617 Table 5 Fatty acid composition of platelet phospholipids of the four dietary groups (% of total fatty acid, meanÆ s.d.)

High meat Moderate meat Ovolacto Vegan Fatty acid (n 18) (n 60) (n 43) (n 18)

14:0 0.3Æ 0.2 0.2Æ 0.1hma 0.2Æ 0.1 0.2Æ 0.1 16:0 13.6Æ 0.6 14.0Æ 1.0 14.4Æ 1.1hoa 14.4Æ 1.6hva 17:0 0.6Æ 0.1 0.6Æ 0.1 0.5Æ 0.1hoa, moc 0.5Æ 0.1hvb, mvc 18:0 21.4Æ 1.5 20.9Æ 1.5 20.0Æ 1.2hoc, mob 20.5Æ 1.9 20:0 0.9Æ 0.2 0.9Æ 0.1 0.9Æ 0.1 1.0Æ 0.1 Total SFA 36.8Æ 1.8 36.6Æ 1.7 36.0Æ 1.7 36.5Æ 2.9 16:1 n-9 0.3Æ 0.1 0.2Æ 0.0 0.2Æ 0.0 0.2Æ 0.0 16:1 n-7 0.3Æ 0.1 0.3Æ 0.1 0.2Æ 0.1hob 0.2Æ 0.1hvc, mvc, ova 17:1 n-9 1.0Æ 1.2 0.7Æ 0.6 1.3Æ 1.1mob 0.8Æ 0.8 18:1 n-9t 0.3Æ 0.1 0.3Æ 0.1 0.4Æ 0.1 0.3Æ 0.1 18:1 n-7t 0.7Æ 0.2 0.6Æ 0.2 0.6Æ 0.2 0.5Æ 0.2hvc, mvc, ovb Total Trans 1.0Æ 0.2 0.9Æ 0.2 1.0Æ 0.3 0.7Æ 0.4hvb, mvb, ovb 18:1 n-9c 14.1Æ 0.9 14.0Æ 0.6 14.1Æ 0.9 15.1Æ 1.3hvc, mvc, ovb 18:1 n-7c 1.0Æ 0.2 1.0Æ 0.2 1.0Æ 0.2 1.2Æ 0.3hva, mvb, ovb Total MUFA 17.4Æ 1.0 16.9Æ 1.0 17.6Æ 1.1mob 17.9Æ 1.4mvc 18:2 n-6 5.3Æ 0.7 5.2Æ 0.6 6.1Æ 0.9hoc, moc 6.3Æ 0.5hvc, mvc 20:3 n-6 1.5Æ 0.3 1.5Æ 0.3 1.5Æ 0.4 1.5Æ 0.4 20:4 n-6 24.4Æ 1.4 24.5Æ 1.2 23.9Æ 1.3moa 23.0Æ 1.7hvb, mvc, ova 22:4 n-6 2.3Æ 0.3 2.4Æ 0.3 2.6Æ 0.4hob, moa 2.7Æ 0.5hvb, mvb 22:5 n-6 0.2Æ 0.0 0.3Æ 0.2 0.3Æ 0.1 0.2Æ 0.1 Total n-6 33.7Æ 1.9 33.9Æ 1.2 34.4Æ 1.5 33.8Æ 2.2 18:3 n-3 0.0Æ 0.0 0.1Æ 0.1 0.1Æ 0.0 0.1Æ 0.2 20:5 n-3 0.4Æ 0.1 0.4Æ 0.1 0.3Æ 0.1hob, moc 0.2Æ 0.1hvc, mvc, ova 22:5 n-3 1.9Æ 0.3 1.9Æ 0.3 1.9Æ 0.5 1.5Æ 0.4hvb, mvc, ovc 22:6 n-3 1.5Æ 0.4 1.6Æ 0.4 1.2Æ 0.4hoa, moc 0.9Æ 0.3hvc, mvc, ovb Total n-3 3.7Æ 0.5 3.9Æ 0.5 3.3Æ 0.8hoa, moc 2.7Æ 0.5hvc, mvc, ovc Total PUFA 37.4Æ 2.0 37.8Æ 1.1 37.8Æ 1.3 36.6Æ 2.3mvb, ovb n-3=n-6 0.11Æ 0.01 0.12Æ 0.02 0.10Æ 0.03hoa, moc 0.08Æ 0.02hvc, mvc, ovb AA=EPA 67.1Æ 19.0 65.5Æ 15.8 97.2Æ 48.8hob, moc 113.9Æ 38.2hvc, mvc SFA=PUFA 0.99Æ 0.08 0.97Æ 0.05 0.95Æ 0.07 1.00Æ 0.05ovb

aP < 0.05; bP < 0.01; cP < 0.001. hmhigh-meat vs moderate-meat; hohigh-meat vs ovolacto; momoderate-meat vs ovolacto; hvhigh-meat vs vegan; mvmoderate-meat vs vegan; ovovolacto vs vegan.

Table 6 Agonist induced ex vivo whole blood platelet aggregation in the four dietary groups (meanÆ s.d.)

High meat Moderate meat Ovolacto Vegan Agonists (n 18) (n 60) (n 43) (n 18)

By height (O at 5 min) Collagen (2 mg=mL) 14.2Æ 3.7 15.1Æ 3.8 17.6Æ 3.5hob, moc 17.5Æ 3.7hvb, mva Arachidonate (1.0 mM) 11.9Æ 4.7 11.7Æ 2.8 12.0Æ 3.9 12.9Æ 2.2 ADP (8 mM) 4.9Æ 2.9 6.4Æ 3.5 7.5Æ 3.4hoa 8.6Æ 2.7hvb, mva ADP (17 mM) 6.0Æ 2.9 7.6Æ 3.3 8.9Æ 3.9hob 9.2Æ 2.1hva By slope (O=min) Collagen (2 mg=mL) 6.7Æ 2.7 8.0Æ 2.9 9.5Æ 3.1hoc, mob 9.4Æ 1.9hvb Arachidonate (1.0 mM) 7.5Æ 3.8 8.5Æ 3.2 9.3Æ 4.3 9.4Æ 2.4 ADP (8 mM) 2.3Æ 1.7 3.6Æ 2.6 5.2Æ 3.7hob, moa 4.5Æ 2.2hva ADP (17 mM) 3.0Æ 1.8 4.5Æ 3.0 5.9Æ 3.5hob,moa 5.1Æ 1.9

aP < 0.05; bP < 0.01; cP < 0.001. hmhigh-meat vs moderate-meat; hohigh-meat vs ovolacto; momoderate-meat vs ovolacto; hvhigh-meat vs vegan; mvmoderate-meat vs vegan; ovovolacto vs vegan. was a signi®cantly higher platelet aggregation and a ten- preference for AA to occupy the sn-2 position of platelet dency to higher (but not signi®cant) mean plasma 11- membrane PL (Goodnight et al, 1981). There would be dehydro TXB2 levels in the vegetarians, despite lower lowered competition with AA as a substrate for cyclo- platelet AA levels. If the explanation for this ®nding is oxygenase (Fisher & Weber, 1985). Also EPA competi- related to platelet fatty acids, clearly the AA levels are tively binds to cyclo-oxygenase excluding AA from the unlikely to involved. A more likely rationale would be the active site (Corey et al, 1983). Both EPA and DHA lower LC n-3 PUFA levels in platelet of vegetarians. In the interfere with the binding of TXA2 to its platelet receptor vegans, the EPA and DHA levels in the platelets were (Swann et al, 1989). The DHA lipoxygenase product 14- nearly half those in the meat eaters, leading to an elevated hydroxy-DHA binds to TXA2 receptor sites inhibiting AA=EPA ratio. The lowered levels of platelet LC n-3 aggregation and smooth muscle cell contraction (Croset PUFA in the vegans would contribute to increased throm- et al, 1988). The conclusions from this aspect of the study boxane production and platelet aggregation through a are that a high consumption of meat does not markedly number of mechanisms. Firstly there would be increased increase platelet AA levels nor does it increase platelet Habitual diet and thrombotic risk factors in man DLiet al 618 aggregation or thromboxane production compared with to a lower plasma factor VII activity compared with an moderate- or non-meat eaters. However, based on the omnivorous diet in this study population. data produced in this study, the meat-eaters did appear to Most acute clinical cases of cardiovascular disease are have a higher number of other risk factors normally caused by the formation of a thrombus (Cahill & Newland, associated with CVD. 1993) with platelet aggregation being the initial step in The dietary intakes of the subjects were assessed using a these events (Benditt & Schwartz, 1994). An increased semi-quantitative FFQ. While there is no ideal method for platelet aggregability is signi®cantly associated with CHD the assessment of individual nutrient intake as all current mortality (Thaulow et al, 1991). Previously, only one study methods involve some systematic error (Willett, 1990), the (Fisher et al, 1986) has investigated platelet function in FFQ is rapid with a low respondent burden and high vegetarians. They found no signi®cant difference between response rate. The semi-quantitative FFQ used was that the vegetarians and the omnivores in vitro collagen-, AA-, based on the work of Fidanza et al (1994) which contains ADP- and epinephrine-stimulated platelet aggregation. In photographs of different food portion sizes and can be a the present study, there was an unexpected increase in ex useful method for evaluating food intake and ranking vivo platelet aggregability in both ovolacto vegetarian and subjects according to habitual intake of certain nutrients. vegan groups. However, Fisher et al (1986) used different The data showed that the meat eaters relative to the methodology to that used in this present study. They used vegetarians had signi®cantly increased levels of dietary the traditional optical method for platelet aggregation test fat, SFA, cholesterol and a lower linoleic acid intake. The (using plasma) and they also adjusted the platelet count. It PUFA:SFA ratio in the high meat eater group was 0.32 has been reported that dilution of platelet rich plasma can compared with 1.48 in the vegans. These present data show cause change in platelet responsiveness (Thaulow et al, that the dietary fat intake pro®le of both ovolacto vegetar- 1991). The increased MPV in vegans suggests the presence ian and vegan groups approaches the National Health & of larger, activated platelets. When platelets become acti- Medical Research Council (NHMRC) Recommended Daily vated, they change from their normal resting disc-like Intakes (1991). This data is consistent with the ®ndings of structure to assume a spherical shape and their volume previous vegetarian studies in Australia (Beilin et al, 1988), increases substantially. This, in conjunction with the the UK (Sanders & Key, 1987) and USA (Dwyer, 1988). increased platelet aggregability, suggests what should be These vegetarian dietary intake patterns are associated with an increased thrombosis tendency in vegans, and in the case lower blood pressure, BMI, waist=hip ratio and plasma of the platelet aggregation may be related to low dietary TC, TAG, LDL-C, LDL-C=TC, LDL-C=HDL-C and intake of n-3 PUFA. However, the vegan group had lower TAG=HDL-C, all of which have a protective effect on RBC and PLT which may overcome the thrombotic poten- both thrombosis and atherosclerosis (Stampfer et al, 1991). tial by lowering blood viscosity thus improving blood ¯ow Is a high consumption of meat per se responsible for the and by decreased hemostatic capacity due to reduced increased intake of fat, saturated fat and low PUFA or is it platelet numbers. These ®ndings require further exploration likely to be the whole=total diet which has these character- in future studies, however given that the vegetarian groups istics? The non-meat density data in Table 2, indicates that have a low incidence of heart disease due to the absence or the subjects consuming the high meat diets had signi®- low level of many of the usual risk factors, we are cantly higher levels of saturated fat than the three other interpreting the thrombosis risk indicators with caution. groups and that both meat eater groups had a lower PUFA intake in the non-meat portions of their diet. This suggests that the overall diet was rich in saturated fat rather than the Conclusions meat being responsible for these diets characteristics. For In this study population, both the high- and moderate-meat- example, meat consumed in a prepared meal such as a stew eater groups have a cluster of thrombotic and atherosclero- or pasta sauce or in a take-away form could contain a tic risk factors higher than both ovolacto vegetarian and higher proportion of saturated fat compared with steak or vegan groups. These factors include BMI, waist=hip ratio, chops which might be fat-trimmed either in the super- blood pressure, coagulation factor VII activity, plasma TC, market or at home. Several studies have indicated that LDL-C and TAG concentration, ratio of TC=HDL-C, LDL- lean meat itself does not increase plasma cholesterol levels, C=HDL-C and TAG=HDL-C. However, compared with rather the total level of saturated fat in the diet is the major both high- and moderate-meat-eater groups, vegetarians determinant in this connection (Sinclair et al, 1987; Watts have an increased ex vivo collagen- and ADP-induced et al, 1988; Scott et al, 1994). platelet aggregability and lower platelet and plasma PL n- Many epidemiological and clinical studies have pro- 3 PUFA content which may be associated with an increased vided evidence showing that increased coagulation factors thrombotic risk. Based on the present data, we are not and impaired ®brinolysis system are important predictors of suggesting that meat eaters alter their diet to a vegetarian CVD (Miller, 1992; Thompson et al, 1995), especially diet. However, it could be suggested that meat eaters reduce plasma ®brinogen concentration and factor VII activity their SFA intake to improve their lipid pro®le, factor VII (Meade et al, 1986), due to their pivotal role in blood activity and possibly their weight. Vegetarians, and espe- coagulation (Ruddock & Meade, 1994; Miller, 1995). cially vegans, might be advised to increase their dietary Lower plasma factor VII activity in the vegan groups intake of n-3 fatty acids to increase their platelet PL n-3 corresponded to long-term low total fat and saturated fat PUFA and hence perhaps reduce their platelet diet of vegans. Plasma ®brinogen concentration was not aggregability. in¯uenced by habitual dietary intake. These results are consistent with the ®ndings of others which have been References reviewed by Miller (1998). The present results indicate that Beilin LJ, Rouse IL, Armsrong BK, Margetts BM & Vandongen R (1988): both ovolacto vegetarian and vegan diets have a protective Vegetarian diet and blood pressure levels: incidental or causal associa- effect against thrombosis tendency because they contribute tion? Am. J. Clin. Nutr. 48, 806 ± 810. Habitual diet and thrombotic risk factors in man DLiet al 619 Benditt EP & Schwartz SM (1994): Blood vessels. In Pathology, 2nd edn, Messmore HL Jr (1983): Automation in coagulation testing. Clinical eds. E Rubin & JL Farber, 455 ± 501. Philadelphia: J. B. Lippincott applications. Semin. Thromb. Hemost. 9, 341 ± 345. Company. Miller GJ (1992): Haemostasis and cardiovascular risk. The British and Berner LA (1993): Roundtable discussion on milk fat, dairy foods and European experience. Arch. Path. Lab. Med. 116, 1318 ± 1321. coronary heart disease risk. J. Nutr. 123, 1175 ± 1185. Miller GJ (1995): Risk of thrombosis and the prethrombotic state. Cahill MR & Newland AC (1993): Platelet activation in coronary artery Atherosclerosis 10, 801 ± 804. disease. Br. J. Biomed. Sci. 50, 221 ± 214. Miller GJ (1998): Effect of diet composition on coagulation pathways. Am. Castaldi PA & Smith IL (1980): The effect of platelets on the in vitro J. Clin. Nutr. 67, S542 ± S545. response to prothrombin complex concentrates in F. VIII inhibitor National Health and Medical Research Council (1991): Recommended plasma. Pathology 12, 111 ± 118. Dietary Intakes for use in Australia. Canberra: AGPS. Corey EJ, Shih C & Cashman JR (1983): Docosahexaenoic acid is a strong Okuyama H, Kobayashi T & Watanabe S (1996): Dietary fatty acids Ð inhibitor of prostaglandin but not leukotriene biosynthesis. Proc. Natl. the n-6=n-3 balance and chronic elderly disease. Excess linoleic acid Acad. Sci. USA. 80, 3581 ± 3587. and relative n-3 de®ciency syndrome seen in Japan. Prog. Lipid. Res. Croset M, Sala A, Folco & Lagarde M (1988): Inhibition by lipoxygenase 35, 409 ± 457. products of TXA2 like responses of platelets and vascular smooth Phinney SD, Odin RS, Johnson SB & Holman RT (1990): Reduced muscle, 14 hydroxy 22:6n-3 is more potent than 12-HETE. Biochem. arachidonate in serum phospholipids and cholesterol esters associated Pharmacol. 37, 1275 ± 1280. with vegetarian diets in humans. Am. J. Clin. Nutr. 51, 385 ± 392. DeLong DM, Delong ER, Wood DD, Lippel K & Rifkind BM (1986): Pradelles P, Grassi J & Maclouf J (1985): Enzyme immunoassays of A comparison of the methods for the estimation of plasma low and very eicosanoids using acetylcholine esterase as label: an alternative to low lipoprotein cholesterol. The LRC prevalence study. J. Am. Med. radioimmunoassay. Anal. Chem. 57, 1170 ± 1173. Assoc. 256, 2372 ± 2377. Ruddock V & Meade TW (1994): Factor VII activity and ischaemic heart Dwyer JT (1988): Health aspects of vegetarian diets. Am. J. Clin. Nutr. 48, disease Ð fatal and non-fatal events. Q.J. Med. 87, 403 ± 406. 712 ± 738. Sanders TA & Roshanai F (1992): Platelet phospholipid fatty acid Fidanza F, Gentile MG & Porrini M (1994): A self-administered semi- composition and function in vegans compared with age and sex quantitative food-frequency questionnaire with optical reading and its matched omnivore controls. Eur. J. Clin. Nutr. 46, 823 ± 831. concurrent validation. Eur. J. Epidemiol. 11,1±8. Sanders TAB & Key TJA (1987). Blood pressure, plasma renin activity Fisher S & Weber PC (1985): Thromboxane (TX) A3 and prostaglandin and aldosterone concentrations in vegans and omnivore controls. Hum. (PG) I3 are formed in men after dietary eicosapentaenoic acid: Identi- Nutr.: Appl. Nutr. 41A, 204 ± 211. ®cation and quanti®cation by capillary gas chromatography-electron Scott LW, Dunn JK, Pownall HJ, Brauchi DJ, Mcmann MC, Herd JA, impact mass spectrometry. Biomed. Mass. Spectrom. 12, 470 ± 476. Harris KB, Savell JW, Cross HR & Gotto AM (1994): Effect of beef Fisher M, Levine PH, Weiner B, Ockene IS, Johnson B, Johnson MH, and chicken consumption on plasma lipid levels in hypercholesterole- Natale AM, Vaudreuil CH & Hoogasian J (1986): The effect of mic men. Arch. Intern. Med. 154, 1261 ± 1267. vegetarian diets on plasma lipids and platelet levels. Arch. Intern. Sinclair AJ & Mann NJ (1996): Short-term diets rich in arachidonic acid Med. 146, 1193 ± 1197. in¯uence plasma phospholipid polyunsaturated fatty acid levels and Goodnight SH, Harris WS & Connor WE (1981): The effects of dietary prostacyclin and thromboxane production in humans. J. Nutr. 126, omega-3 fatty acids on platelet composition and function in men: A S1110 ± S1114. prospective, controlled study. Blood 58, 880 ± 885. Sinclair AJ & O'Dea K (1990). Fats in human diets through history: is the Ingerman-Wojenski CM & Silver MJ (1984): A quick method for screen- western diet out of step? In Reducing Fat in Meat Animals, eds. JD ing platelet dysfunctions using the whole blood lumi-aggregometer. Wood & AV Fisher, pp 1 ± 47. London and New York: Elsevier Applied Thromb. Haemost. 51, 154 ± 156. Science. Keys A (1980): Seven Countries: A Multivariate Analysis of Death and Sinclair AJ, O'Dea K, Dunstan G, Ireland PD & Niall M (1987): Effect on Coronary Heart Disease. Cambridge, Mass: Harvard University Press. plasma lipids and fatty acid composition of very low fat diets enriched Kromhout D, Bosschieter EB & de Lezenne Coulander C (1982): Dietary with ®sh or kangaroo meat. Lipids 22, 523 ± 529. ®bre and 10-year mortality from coronary heart disease, cancer and all Stampfer MJ, Sacks FM, Salvini S, Willett WC & Hennekens CH (1991): causes: the Zutphen Study. Lancet 2, 518 ± 521. A prospective study of cholesterol, apolipoproteins, and risk of myo- Kushi LH, Lew RA, Stare FJ, Ellison CR, Lozy ME, Bourke G, Daly L, cardial infarction. N. Engl. J. Med. 325, 373 ± 381. Graham I, Hickey N. Mulcahy R & Kevaney J (1985): Diet and 20-year Stavropoulous WS & Crouch RD (1974): A new colorimetric procedure mortality from coronary heart disease. The Ireland-Boston Diet-Heart for the determination of serum triglycerides. Clin. Chem. 20, 857. Study. N. Engl. J. Med. 312, 811 ± 818. Swann PG, Venton DL & Le Breton GC (1989): Eicosapentaenoic Lands WEM, Libelt B, Morris A, Kramer NC, Prewitt TE, Bowen P, acid and docosahexaenoic acid are antagonists at the thromboxane Schmeisser D, Davidson MH & Burns JH (1992): Maintenance of lower A=prostaglandin H2 receptor in human platelets. FEBS 243, 244 ± proportions of (n-6) eicosanoid precursors in phospholipids of human 247. plasma in response to added dietary (n-3) fatty acids. Biochim. Biophys. Thaulow E, Erikssen J, Sandvik L, Stormorken H & Cohn PF Acta. 1180, 147 ± 162. (1991): Blood platelet count and function are related to total and Mann NJ, Johnson LG, Warrick GE & Sinclair AJ (1995): The arachidonic cardiovascular death in apparently healthy men. Circulation 84, 613 ± acid content of the Australian diet is lower than previously estimated. 617. J. Nutr. 125, 2528 ± 2535. Thompson SG, Kienast J, Pyke SDM, Haverkate F & Van De Loo JCW Mann NJ, Sinclair AJ, Pille M, Johnson LG, Warrick GE, Reder E & (1995): Hemostatic factors and the risk of myocardial infarction or Lorenz R (1997): The effect of short-term diets rich in ®sh, red meat, or sudden death in patients with angina pectoris. N. Engl. J. Med. 332, white meat on thromboxane and prostacyclin synthesis in humans. 635 ± 641. Lipids 32, 635 ± 644. Watts GF, Ahmed W, Quiney J, Houlston R, Jackson P, Iles C & Lewis Meade TW, Mellows S, Brozovic M, Miller GJ, Chakrabarti RR, North (1988): Effective lipid lowering diets including lean meat. Br. Med. J. WRS, Haines AP, Stirling Y, Imeson JD & Thompson SG (1986): 296, 235 ± 237. Hemostatic function and ischaemic heart disease: principal results of the Willett WC (1990): Nutritional epidemiology. In Monographs in Epide- Northwick Park Heart Study. Lancet 2, 533 ± 537. miology and Biostatistics, Vol. 15. Oxford: Oxford University Press.