sign in sign up
<<
Home , Fatty acid

European Journal of Clinical Nutrition (2000) 54, 768±773 ß 2000 Macmillan Publishers Ltd All rights reserved 0954±3007/00 $15.00 www.nature.com/ejcn

Serum fatty , pro®le and dietary intake of Hong Kong Chinese omnivores and vegetarians

HY Lee1, J Woo1*, ZY Chen2, SF Leung3 and XH Peng3

1Department of Medicine and Therapeutics, The Chinese University of Hong Kong, Hong Kong; 2Department of , The Chinese University of Hong Kong, Hong Kong; and 3Department of Paediatries, The Chinese University of Hong Kong, Hong Kong

Objective: To examine the serum and lipid pro®les and dietary intake of Hong Kong Chinese omnivores and vegetarians with respect to cardiovascular health. Design: Random population survey strati®ed by age and sex. Subjects: One-hundred and ninety-four omnivore subjects (81 men, 113 women) age 25 ± 70 y, and 60 ovo-lacto- vegetarian adults (15 men, 45 women) age 30 ± 55 y. Measurements: quantitation was by a frequency method. Serum fatty were analysed by , and serum lipid by standard laboratory methods. Results: Compared with omnivores, vegetarians had higher serum concentrations of polyunsaturated (PUFA) and monosaturated fatty acids (MUFA), and lower saturated fatty acids (SFA), long chain -3 and trans fatty acids (TFA). They also had lower serum and higher apoA-I concentrations, but the LDL=HDL ratio was not different. The ratio of polyunsaturated to saturated fatty acids intake was higher in vegetarians. Compared with results from populations with higher incidences of coronary heart disease, while lower myristic and concentrations and higher eicosapentaneoic (EPA) and docosahexanoic acid (DHA) may partly account for the difference in incidence, concentration was higher. Although the Chinese vegetarian diet may be bene®cial for heart health in that antioxidant and ®bre intakes are higher and saturated lower, the low EPA and DHA due to omission from dietary source and suppressed formation by high linoleic acid level, and the presence of TFA in the diet, may exert an opposite effect. Conclusion: There are some favourable features in the serum fatty acid pro®le in the Hong Kong Chinese population with respect to cardiovascular health, but the consumption of TFA is of concern. The Chinese vegetarian diet also contains some adverse features. Sponsorship: This study is partly sponsored by the Hong Kong Health Services Research Grant. Descriptors: serum fatty acids; Hong Kong Chinese; vegetarian; lipid pro®le European Journal of Clinical Nutrition (2000) 54, 768±773

Introduction 1992). To date the serum fatty acid pro®le has not been documented for the Chinese population. As part of a The relationship between dietary factors such as the territory-wide cardiovascular risk factor study in Hong adverse effect of and consumption and Kong, we investigated the serum fatty acid composition, the bene®cial effect of vegetables, fruits and nuts, towards serum and dietary intake of Chinese omnivores and coronary heart disease has been well documented in Cau- vegetarians with respect to cardiovascular health. casian populations (Bonanome & Grundy, 1988; Key et al, 1996; Fraser, 1999). Studies also showed that vegetarians have lower incidence of cardiovascular diseases (Key et al, Methods 1996; Fraser, 1999). The incidence and mortality from coronary heart disease in Hong Kong Chinese is one- Subjects eighth to one-quarter of that in Caucasian populations One hundred and ninety-four (81 men and 113 women) (Woo & Donnan, 1989; Department of Health, 1994). aged 25 ± 70 y healthy community omnivore subjects were This may in part be explained by dietary differences. randomly selected, strati®ed for age and sex, from the Dietary fat and fatty acid composition have been documen- territory-wide cardiovascular risk factor study (CRFS) ted as risk factors for coronary heart disease (Ascherio 1995 (Woo et al, 1997). The number of subjects in each et al, 1994, 1999; Simon et al, 1995). Fatty acid intake age and sex group was according to the percentage dis- estimated by dietary record or duplicated meal measure- tribution in the CRFS (Table 1). For comparison between ment alone is either unreliable or too complicated (Rosha- omnivores and vegetarians, serum fatty acid analysis was nai & Sanders, 1985; Hunter et al, 1992). However, body carried out in 133 (56 men and 77 women) omnivore fatty acid composition re¯ects the quality of dietary fat and subjects, matched for age and body mass index with the the type of fat consumed over a long period (Hunter et al, vegetarian subjects. Data from all omnivore subjects were used for comparison with other countries. Sixty healthy vegetarian adults (15 men and 45 women) *Correspondence: J Woo, Department of Medicine and Therapeutics, aged 25 ± 55 y, were recruited from the local vegetarian Prince of Wales Hospital, Shatin NT, Hong Kong. societies and religious groups by advertising through letters E-mail: [email protected]. and telephone contacts to the respective representatives. Dietary intake of Hong Kong Chinese HY Lee et al 769 Table 1 Age distribution (%) of 194 omnivore subjects and in the 14 ± 22 were measured. Individual fatty acid level was territory-wide cardiovascular risk factor study (CRFS) expressed as the percentage of the total serum fatty acids. Men Women Dietary assessment Age (y) n ˆ 81 CRFS (n ˆ 1399) n ˆ 113 CRFS (n ˆ 1482) A food frequency questionnaire consisting of most fre- quently consumed food items in the following seven 20 ± 29 8.6 8.7 9.7 7.2 30 ± 39 32.1 28.7 31.9 30.6 categories was used: bread=pasta=rice (16 items); vegeta- 40 ± 49 32.1 25.8 26.6 29.7 bles (63 items); fruits (26 items); meat (39 items); ®sh (31 50 ± 59 17.3 15.9 19.5 15.7 items); eggs (5 items); beverages (37 items); dimsum= 60 ± 69 9.9 16.7 12.4 13.2 snacks (39 items); soups (10 items); and ==sauces.  70 0.0 4.1 0.0 3.7 For vegetarians, meat and meat products were removed from the food questionnaire and vegetarian food that are commonly consumed in Hong Kong were added. Wherever The criteria for inclusion as vegetarian subjects were that possible, subjects were told prior to the visit that a survey they consumed no meat or ®sh but may consume dairy on a week's diet would be carried out and were advised to products or eggs, and had maintained a vegetarian diet for record brie¯y their diet to help the interview. On the day of at least one year. the interview, each subject was asked to complete the questionnaire, including the food item, the size of each Serum lipid analysis portion, and the frequency of consumption on a daily or Twenty millilitres of venous were taken after an weekly basis. Data were crossed-checked by examining the overnight fast. Sera were collected for exam- dietary pattern (for example, if meals were skipped) to see ination and the remainders were stored at 70C before if it corresponded to the number of times staple such fatty acid analysis. Cholesterol concentration was measured as rice or noodles were consumed over a one week period. on the Hitachi 717 analyser using the cholesterol oxidation The amount of was estimated according to the method (CV2 2.9% at 3.3 mmol=l and 2.5% at 7.0 mmol=l), method of preparing different foods: 0.2 tablespoon for and by the same instrument using the = steaming ®sh or stir-frying half a portion of vegetables or kinase method (CV 3.3% at 0.63 mmol=l). High- one portion of meat. Quantitation of dietary lipids was density- cholesterol (HDL-C) was measured on carried out using food tables for Hong Kong which were the Hitachi 717 analyser after precipitation of very-low- compiled from McCane and Widdowson (Holland et al, density-lipoprotein cholesterol (VLDL-C), and low density 1992), and two food tables used in China, published by lipoprotein (LDL-C) by PEG 6000 (CV Zhongshan University (Tsang & Fung, 1991) and the 4.2% at 1.64 mmol=l). LDL-C was calculated by the Institute of Health of the Chinese Medical Science Institute Friedewald equation (Friedewald et al, 1972), provided (1992). Only crude ®bre intake instead of soluble ®bre total did not exceed 4.0 mmol=l. The serum could be quanti®ed, since many vegetables are not found in apolipoprotein-A (apoA-I) and apolipoprotein B (apoB) English food tables and only appear in the Chinese tables, concentrations were assayed using an immunoradiometric which provide crude ®bre content. assay (IRMA, Mercadia apo(a) RIA kit, Uppsala, Sweden, previously known as Pharmacia apo(a) (RIA). The method Comparison between local omnivores and Caucasians, has an average interassay variation of 3 ± 5%. Japanese and Eskimos Serum fatty acids of the studied omnivores (n ˆ 194) were compared with other populations with different dietary fat Serum fatty acid analysis consumption patterns. Due to lack of comprehensive data Sera were thawed to room for fatty acids for Caucasians, data from Austin, Minnesota in 1975, analysis. Serum lipids were extracted from blood samples including 103 healthy males and 101 healthy females with 5 ml chloroform-methanol (2 : 1, v=v). Heptadecanoic with 10 subjects per sex for each decade of age from 0 to acid (17:0) was added as an internal standard to quantify 90 y, was used for analysis (Holman et al, 1979). The serum the fatty acids in serum samples. This fatty acid was used fatty acid concentration was presented as regression model because it is not presented in human plasma and its with age, therefore, the mean age of our studied males and recovery rate is greater than 98%. Samples were then females was substituted in the regression model for each centrifuged at 1000 rpm for 15 min. The chloroform phase serum fatty acid and the values obtained were used for containing the extracted lipids was separated by aspiration comparison with our studied data. The Japanese data were and transferred to the methylation tube. Fatty acids in from a study of 110 healthy normal individuals (55 males serum were converted to their corresponding fatty acid and 55 females) recruited in Aomori, Japan, whose eating methyl using an analytical evaporator (N-EVAP habits were believed to be representative of the average Model No. 111 USA), by addition of 2 ml of 14% boron Japanese diet (Nakamura et al, 1995). The serum fatty acid tri¯uoride in methanol and 1 ml of toluene (2 : 1, v=v) under data were shown by age groups (30 s, 40 s, 50 s and 60 s). nitrogen at 90C for 45 min. After cooling down the tube at Data for Eskimos included 20 adults (10 males and 10 room temperature, 5 ml of hexane and 2 ml of saline were females) aged 17 ± 40 y (mean age: 27.0 y), randomly added into the tube. The samples were then vortexed and selected from a coastal village with a population of 370 centrifuged at 1500 rpm for 10 min. The top hexane layer people, in Southwestern Alaska (Parkinson et al, 1994). was taken out and transferred into SP-2560 ¯exible fused silica capillary column. Fatty acids were quanti®ed by Statistical analysis gas ± liquid chromatography (Hewlett Packard 5890 Series Statistical analysis was done using the SPSS program II, USA), using hydrogen as the carrier gas at a head (spss version 6.1). Results are expressed as mean+ s.d. pressure of 20 psi. Fatty acids from chain length Comparisons between groups were made using Student's t-

European Journal of Clinical Nutrition Dietary intake of Hong Kong Chinese HY Lee et al 770 test. Partial correlation was used to assess the relationship was signi®cantly lower in the vegetarian group and the between serum fatty acids and serum lipids, controlling for mean intake was only half that of the omnivore group. age, sex and body mass index (BMI). P values were two- Increase in dietary PUFA (percentage dietary fat) was tailed and P-value < 0.05 was considered signi®cant. re¯ected as an increase in serum PUFA (r ˆ 0.15, Written consent was obtained from each subject and the P ˆ 0.038), after controlling for age, sex and BMI. A study was approved by the Ethnic Committee. trend of higher concentrations of serum SFA with higher consumption of saturated fat (percentage energy) was also observed, although this was not statistically signi®cant Results (r ˆ 0.13, P ˆ 0.071). Subjects For 194 omnivore subjects, the mean age was 43.4+ 11.2 y Serum lipid pro®le for male and 43.8+ 11.6 y for female. The mean body The serum lipid pro®les of the vegetarians and omnivores mass index (BMI) was 24.4+ 3.6 for male and 23.6+ 3.8 are shown in Table 4. The vegetarians had a signi®cantly for female. The mean age and BMI of omnivores and lower concentration of serum total cholesterol, LDL-C and vegetarians at the age between 30 and 55 y are shown in HDL-C. However, the LDL-C=HDL-C ratio was not dif- Table 2. All vegetarian subjects were lacto-ovo-vegetarians ferent. ApoA-I concentration was signi®cantly higher in the and had practised vegetarian diet for more than 1 y. There vegetarians. Such differences were more marked between was no signi®cant difference in age and BMI between the vegetarian and omnivore women (1.48 vs 1.37, omnivores and vegetarians. P ˆ 0.011). Serum lipids showed signi®cant correlation with indivi- dual serum fatty acids after controlling for age, sex and Dietary intake body mass index. Triglyceride correlated positively with Nutrient intakes in vegetarians and omnivores are shown in (14:0; r ˆ 0.16, P ˆ 0.032), palmitic acid Table 3. The vegetarian group had a signi®cantly lower (16:0; r ˆ 0.20, P ˆ 0.007), (16:1n-7; intake of energy and (energy-adjusted) but a r ˆ 0.27, P < 0.001), oleic acid (18:1n-9; r ˆ 0.29, higher carbohydrate intake compared with the omnivore P < 0.001), total trans (r ˆ 0.17, P ˆ 0.022) and total group. Intake of saturated, monounsaturated and polyunsa- MUFA (r ˆ 0.39, P < 0.001), but correlated negatively turated fatty acids were higher in the omnivore group as with linoleic acid (18:2n-6; r ˆ0.16, P ˆ 0.026), arachi- compared with the vegetarian group. However, the percen- donic acid (20:4n-6; r ˆ0.26, P < 0.001) and total PUFA tage of polyunsaturated fatty acids contributing to the total (r ˆ0.21, P ˆ 0.004). The lowering effect of linoleic acid dietary fat intake was signi®cantly higher in the vegetarian and total PUFA on serum triglyceride was more marked in group than in the omnivore group, as shown by the ratio of vegetarians (r ˆ0.49, P < 0.001 and r ˆ0.37, polyunsaturated to saturated fatty acids. Cholesterol intake P ˆ 0.006, respectively). Conversely, HDL-C correlated positively with eicosa- Table 2 Age and BMI (mean+ s.d.) in vegetarians and omnivores pentaenoic acid (EPA, 20:5n-3; r ˆ 0.19, P ˆ 0.008), ara- chidonic acid (r ˆ 0.18, P ˆ 0.014) and (18:0, Male Female r ˆ 0.26, P < 0.001), but correlated negatively with (r ˆ0.35, P < 0.001, alpha-linolenic acid (18:3n-3; Vegetarians Omnivores Vegetarians Omnivores r ˆ0.16, P ˆ 0.033) and total MUFA (r ˆ0.36, n 15 56 45 77 P < 0.001). Age (y) 39.7+ 4.8 40.0+ 6.4 39.2+ 4.7 40.5+ 6.8 Serum total cholesterol and LDL-C showed positive BMI (kg=m2) 23.1+ 3.1 24.3+ 3.6 23.2+ 3.4 23.7+ 3.6 correlation with (22:5n-6; both BMI, body mass index. r ˆ 0.16, P ˆ 0.024). Moreover, serum total cholesterol correlated positively with EPA (r ˆ 0.19, P ˆ 0.011) but negative correlation with oleic acid (r ˆ0.15, P ˆ 0.040). Table 3 Nutrient intakes (mean+ s.d.) in vegetarians and omnivores Vegetarians (n ˆ 60) Omnivores (n ˆ 133) Serum fatty acid pro®le Energy (kcal) 1835.6+ 556.2 2148.3+ 807.8* Sixteen fatty acids in serum were identi®ed (Table 5). The Carbohydrate (g=1000 kcal) 166.2+ 16.5 129.6+ 18.2* most abundant serum fatty acid in the studied population Protein (g=1000 kcal) 33.8+ 7.3 47.0+ 7.5* was linoleic acid (18:2n-6), followed by oleic acid (18:1n- Total fat (g=1000 kcal) 24.2+ 5.5 33.7+ 5.8* 9), palmitic acid (16:0) and stearic acid (18:0). Unidenti®ed SFA (g=1000 kcal) 5.9+ 2.0 9.1+ 2.1* peaks accounted for less than 0.5% of the total fatty acids. MUFA (g=1000 kcal) 7.2+ 1.9 11.7+ 2.6* PUFA (g=1000 kcal) 6.8+ 1.8 7.9+ 1.7* Cholesterol (g=1000 kcal) 87.2+ 66.7 161.8+ 50.1* Table 4 Serum lipid pro®le of vegetarians and omnivores P=S ratio 1.25+ 0.44 0.91+ 0.30* Percentage of energy Vegetarians (n ˆ 60) Omnivores (n ˆ 133) Carbohydrate (%) 66.5+ 6.6 51.8+ 7.3* Protein (%) 13.5+ 2.9 18.8+ 3.0* TC (mmol=l) 4.51+ 0.74 4.84+ 0.90* Fat (%) 21.8+ 5.0 30.3+ 5.2* TG (mmol=l) 1.13+ 0.53 1.05+ 0.58 SFA (%) 5.3+ 1.8 8.2+ 1.8* LDL-C (mmol=l) 2.80+ 0.70 3.00+ 0.80* MUFA (%) 6.5+ 1.7 10.5+ 2.4* HDL-C (mmol=l) 1.20+ 0.25 1.32+ 0.31* PUFA (%) 6.1+ 1.6 7.1+ 1.6* LDL-C=HDL-C 2.44+ 0.83 2.46+ 0.99 VLDL-C (mmol=l) 0.51+ 0.24 0.48+ 0.42 Abbreviations: s.d., standard deviation; SFA, saturated fatty acid; MUFA, ApoA-I (g=l) 1.45+ 0.20 1.34+ 0.21** monounsaturated fatty acid; PUFA, polyunsaturated fatty acid; P=S ratio, ApoB (g=l) 0.92+ 0.23 0.89+ 0.31 polyunsaturated to saturated fatty acid ratio. *P < 0.001. *P < 0.05, **P < 0.01.

European Journal of Clinical Nutrition Dietary intake of Hong Kong Chinese HY Lee et al 771 Table 5 Serum fatty acid composition (mean+ s.d.) of Hong Kong (1.1+ 0.5 vs 1.0+ 0.5; P ˆ 0.037) than women. For vege- Chinese omnivores and vegetarians tarians, men had lower concentration of eicosadienoic acid Vegetarians Omnivores than women (0.1+ 0.2 vs 0.7+ 1.8; P ˆ 0.041). (n ˆ 60) (n ˆ 133)

SFA: Comparison between local omnivores and Caucasians, Myristic (14:0) 0.3+ 0.3 0.5+ 0.4* Palmitic (16:0) 19.0+ 2.4 17.8+ 3.9** Japanese and Eskimo Stearic (18:0) 7.1+ 0.9 13.0+ 5.6** Compared with that in Caucasians, the analysis of serum Total SFA 26.5+ 2.6 31.3+ 6.7*** fatty acids showed a favourable pro®le with respect to MUFA: coronary heart disease (Table 6). Hong Kong Chinese had Palmitoleic (16:1n-7) 0.7+ 0.6 1.0+ 0.5*** signi®cantly lower concentrations of myristic (14:0) and Oleic (18:1n-9) 22.3+ 3.7 18.8+ 4.5*** Total trans 0.03+ 0.12 0.50+ 0.86*** palmitic acid (16:0) but higher concentrations of eicosa- Total MUFA 24.0+ 3.7 21.8+ 4.9** pentaenoic (EPA, 20:5n-3) and PUFA: (DHA, 22:6n-3). Omega-6: Oleic acid, which was found in this study to be posi- Linoleic (18:2n-6) 38.2+ 4.1 29.5+ 7.6*** Gamma-linolenic (18:3n-6) 0.9+ 1.2 0.5+ 1.1 tively associated with serum triglyceride but negatively Eicosadienoic (20:2n-6) 0.5+ 1.6 0.2+ 0.2 associated with HDL-C and total cholesterol, was also Dihomogammalinolenic (20:3n-6) 1.0+ 0.7 1.1+ 0.6 signi®cantly lower. The difference in serum fatty acid Arachidonic (20:4n-6) 5.3+ 1.6 6.3+ 2.2*** pro®le between local and Caucasian populations re¯ects Docosatetraenoic (22:4n-6) 0.0+ 0.2 0.4+ 0.7*** dietary differences. A comparison with the dietary pattern Docosapentaenoic (22:5n-6) 0.1+ 0.4 0.2+ 0.4 Total omega-6 46.0+ 4.7 38.8+ 8.6*** from an American adult population survey in 1989 ± 1991, Omega-3: which included strati®ed samples of the US population in -linolenic (18:3n-3) 1.7+ 2.0 0.8+ 0.6*** de®ned geographic areas, showed that Hong Kong Chinese Eicosapentaenoic (20:5n-3) 0.2+ 0.5 1.3+ 1.3*** consumed more ®sh, fruits and vegetables but less milk and Docosahexaenoic (22:6n-3) 1.7+ 2.5 3.4+ 2.2* Total omega-3 3.6+ 3.3 5.5+ 2.3*** animal fat (Table 7, Popkin et al, 1996). Total PUFA 49.5+ 4.6 44.2+ 9.3*** Matching the mean age of our studied omnivores, the serum fatty acid pro®le at age 40 s (mean age: 43.7+ 2.6 y *P < 0.05, **P < 0.01, ***P < 0.001. for male and 42.7+ 2.5 y for female) from the Japanese

Signi®cant difference was found between omnivores and Table 7 Comparison of the selected nutrient intake and food vegetarians. Serum polyunsaturated fatty acids (PUFA) and consumption of the studied Hong Kong Chinese omnivores and Americans monounsaturated fatty acids (MUFA) in vegetarians were Hong Kong (n ˆ 194) USA (n ˆ 4610) signi®cantly higher than those in the omnivores. Saturated fatty acids (SFA), (20:4n-6) and long Percentage energy from: chain omega-3 fatty acids (EPA, 20:5n-3; DHA, 22:6n-3) Fat 29.7 34.1 ± 34.2 SFA 8.0 11.6 ± 11.9 were lower in the vegetarians than those in omnivores. Food items (g=day): Trans fatty acids were found in all the omnivore subjects Meat 77.3 51 ± 59 but only in 42% of the vegetarian subjects and were Poultry 65.6 24 ± 36 signi®cantly higher in omnivores. 83.8 9 ± 21 Eggs 21.9 17 ± 22 Signi®cant gender differences were also found in both Whole milk Total milk: 46.3 0.2 ± 0.3 omnivore and vegetarian groups. Among the omnivores, Low-fat milk Ð 120 ± 163 men had signi®cantly higher concentration of total MUFA Animal fat Ð 18 ± 22 (23.6+ 4.2 vs 20.4+ 5.1; P < 0.001), oleic acid Fruit 249.5 97 ± 154 (20.3+ 3.8 vs 17.6+ 4.6; P ˆ 0.001), eicosadienoic acid Vegetables 252.8 111 ± 133 (0.2+ 0.2 vs 0.1+ 0.2; P ˆ 0.001), gamma-linoleic acid SFA, saturated fatty acid. (0.8+ 1.5 vs 0.3+ 0.5; P ˆ 0.022) and palmitoleic acid aSource: Popkin et al (1996).

Table 6 Comparison of serum fatty acid compositions of the studied Hong Kong Chinese omnivores with Americans, Japanese and Eskimos

Hong Kong USAa Japanb Eskimoc

M(nˆ 81) F (n ˆ 113) M (n ˆ 103) F (n ˆ 101) M (n ˆ 13) F (n ˆ 13) Both sexes (n ˆ 20)

SFA 14:0 0.5+ 0.5 0.4+ 0.3 1.0+ 0.7* 1.0+ 0.7* 0.8+ 0.4 0.7+ 0.4 Ð 16:0 17.5+ 3.9 17.4+ 3.6 22.1+ 3.0* 21.5+ 3.4* 26.5+ 3.1* 24.4+ 2.6* 16.6+ 2.3 18:0 12.7+ 6.1 13.4+ 5.7 6.6+ 1.3* 6.7+ 1.3* 9.3+ 1.4 9.1+ 1.0 7.5+ 1.2* MUFA 18:1n-9 20.7+ 3.8 18.0+ 4.6 23.8+ 5.2* 23.5+ 3.8* 23.0+ 2.8* 20.5+ 2.6 18.2+ 1.6 Omega-6 18:2n-6 29.5+ 5.7 29.7+ 7.8 25.4+ 5.3* 25.8+ 3.9* 19.5+ 3.1* 22.3+ 3.1* 30.6+ 7.5 18:3n-6 0.7+ 1.1 0.4+ 0.7 0.4+ 0.3 0.5+ 0.6 0.5+ 0.7 1.0+ 0.4 0.2+ 0.1 20:2n-6 0.2+ 0.2 0.2+ 0.2 0.2+ 0.2 0.3+ 0.2 Ð Ð Ð 20:4n-6 5.8+ 1.7 6.0+ 1.8 8.1+ 2.4* 8.3+ 2.1* 5.4+ 1.1 6.2+ 0.9 5.3+ 2.0 Omega-3 18:3n-3 0.9+ 0.4 0.9+ 0.6 0.5+ 0.3* 0.6+ 0.3* Ð Ð 0.4+ 0.1 20:5n-3 1.0+ 0.9 1.3+ 1.3 0.8+ 0.6* 0.7+ 0.5* 3.4+ 1.1* 3.8+ 1.5* 6.4+ 4.1* 22:6n-3 3.2+ 2.1 3.4+ 2.2 1.6+ 1.3* 1.7+ 1.2* 7.2+ 3.2* 7.4+ 2.0* 5.2+ 1.5*

SFA, saturated fatty acid; MUFA, monounsaturated fatty acid; Omega-6, omega-6 polyunsaturated fatty acid; Omega-3, omega-3 polyunsaturated fatty acid. Sources: aHolman et al (1979); bNakamura et al (1995); cParkinson et al (1994).

European Journal of Clinical Nutrition Dietary intake of Hong Kong Chinese HY Lee et al 772 data were used for comparison. Chinese omnivores had to coronary heart disease risk. Oleic acid (18:1) induces a signi®cantly lower concentrations of serum palmitic acid, concentration-dependent reduction of NO synthase activity oleic acid, EPA and DHA, but higher concentrations of in endothelial in animal models (Davda et al, 1995). linoleic acid. Our studied omnivores had signi®cantly higher However, in a human umbilical vein endothelial cell concentrations of palmitic acid but lower concentrations of model, oleic acid was found to inhibit the expression of DHA and EPA than Alaskan Eskimos. gene products involved in leukocyte recruitment and early atherogenesis (Carluccio et al, 1999). Moreover, both alpha- and oleic acid have been shown to Discussion increase arterial compliance, resulting in a functional Features of Chinese diet are less than 30% energy from fat, improvement in the systematic arterial circulation despite less than 10% from saturated fat, and cholesterol intake not a rise in LDL oxidizability (Nestel et al, 1997). These greater than 180 mg=1000 Kcal (Woo et al, 1997), which observation may provide a basis for the cardioprotective may contribute towards lower mortality rate from coronary effect of a Mediterranean-type diet (which is high in the heart diseases (Woo et al, 1989). This study shows that content of these fatty acids, De Lorgeril M et al, 1996). differences in the serum fatty acid pro®le may also account The presence of trans fatty acids (TFA) in our subjects for the difference in cardiovascular risk between Hong indicated the frequent consumption of TFA-containing Kong Chinese and Caucasians. Myristic acid (14:0) and foods, particularly processed food. Trans fatty acids are palmitic acid (16:0) concentrations have been shown to mainly produced during oil hardening in food processing. have a positive association with coronary heart disease risk In our study, a signi®cant positive correlation was found while polyunsaturated omega-3 fatty acids, especially doc- between serum TFA and serum triglyceride, consistent with osahexaenoic acid (DHA, 22:6n-3), have an inverse asso- the recent review that TFA increased LDL-C, total triglycer- ciation (Simon et al, 1995). Our ®ndings are compatible ide, lipoprotein(a), but decreased HDL-C (Ascherio et al, with the latter study, in that myristic and palmitic acids 1999). The net increase on the LDL-C to HDL-C ratio were signi®cantly associated with increased serum trigly- was suggested to be approximately double the effect of ceride, whereas polyunsaturated omega-3 eicosapentaenoic saturated fatty acids. acid (EPA, 20:5n-3) was associated with increased HDL-C. A high consumption of TFA (mainly from hydrogenated Supplementation of the diet with omega-3 fatty acids for 6 vegetable ) prediposes to coronary heart disease (Willett months has been shown to decrease endothelial cell mar- et al, 1993; Ascherio et al, 1994). A recent study on kers thrombomodulin (sTM) and von Willebrand factor replacing (high in saturated fat) by or (vMF), and increase the soluble forms of the cell adhesion shortening (high in TFA) showed a 4% increase in serum molecules E-selectin and vascular cell adhesion molecule 1 total cholesterol to HDL-C ratio relative to butter (Lichten- (VCAM-1) in hyperlipidaemic middle-aged male smokers stein et al, 1999). Avoidance of dietary TFA or food (Selje¯ot et al, 1998). A similar ®nding was observed in labelling of TFA content have been advocated (Ascherio patients with coronary heart disease after 4 weeks' treat- et al, 1999; Lichtenstein, 1997; Lichtenstein et al, 1999). ment (Johansen et al, 1999). Thrombomodulin, an endothe- Epidemiological studies showed that vegetarians have lial surface receptor for thrombin, is thought to be a speci®c lower incidence of chronic diseases including CHD (Key marker of endothelial injury, and vWF may be a marker of et al, 1996, 1999; Fraser, 1999), as the vegetarian diet has a both endothelial injury and activation. The decrease in better polyunsaturated=saturated fat (P=S) ratio (close to 1), thrombomodulin and vWF would indicate a reduced activ- and higher ®bre and antioxidants from vegetables, fruits ity towards endothelial damage and platelet activation and and nuts (Key et al, 1996; Fraser, 1999). However, the hence a lowering of haemostatic activity of the endothe- higher serum palmitic and oleic acid concentrations in our lium. Although an increase in E-selection and VCAM-1 vegetarian diets, and the high intake of carbohydrates may found after omega-3 fatty acids supplementation may induce hypertriglyceridaemia (Bonanome & Grundy, 1988; indicate an enhanced proin¯ammatory response, this Harris et al, 1984). This effect may be prevented by the may be due to oxidative stress induced by the excessively adequate intake of ®sh oil (Harris et al, 1984). On the other high concentration (5.1 g=day) of omega-3 fatty acids hand, the higher serum oleic and alpha-linolenic acid supplementation (Johansen et al, 1999). concentration in vegetarians may have an overall cardio- The vasodilatory effect of omega-3 fatty acids may also protective effect as discussed above. The lower total be bene®cial in relation to coronary heart disease. A 6 week omega-3 fatty acids in vegetarians is probably largely treatment with 1.8 g=day EPA showed that it signi®cantly explained by the omission of ®sh in their diet. In spite of augmented endothelial-dependent vasodilation in the fore- this, vegetarians are still at lower risk of cardiovascular arm of coronary disease patients (Tagawa et al, diseases, probably because of other favourable dietary 1999). The augmenting effect of EPA is due to its direct features such as lower intakes of saturated fats and trans effect on the endothelial vasodilator function, independent fatty acids, and higher intakes of PUFAs. Although the total of nitric oxide (NO)-mediated or non-NO-mediated omega-3 fatty acids is lower, vegetarians have twice the endothelial vasodilation, or to alterations in endothelial level of 18:3, but one-sixth the level of EPA and half the membrane properties such as changing receptor ± effector level of DHA compared with omnivores, suggesting a coupling or ¯uidity. The effects of omega-3 reduced formation of EPA and DHA from substrate in fatty acids on endothelium level may partly explain the vegetarians, perhaps as a result of higher linoleic acid level. lower incidence of coronary heart disease in Chinese, The lower proportion of DHA in the blood and tissue lipids Japanese and Eskimo populations, where ®sh consumption of vegetarians may adversely affect and cardiac is higher than among Caucasians. function (British Nutrition Foundation, 1992). Low The lower consumption of oleic acid in the local plasma EPA and DHA concentrations may also be asso- population compared with Caucasians as re¯ected in the ciated with the shorter bleeding times, higher blood platelet serum fatty acid pro®le may also be bene®cial with respect count and platelet aggregation observed in vegetarians

European Journal of Clinical Nutrition Dietary intake of Hong Kong Chinese HY Lee et al 773 (Mezzano et al, 1999; Li et al, 1999), which may predis- Institute of Health, Chinese Medical Science Institute (1992): Food pose to coronary thrombosis. Thus, ratio of linoleic acid to Composition Tables. Beijing, China: Peoples Health Press. Johansen O, Sleje¯ot I, Hostmark AT & Arnesen H (1999). The effect of alpha-linolenic acid in vegetarian diets of between 4:1 and supplementation with omega-3 fatty acids on soluble markers of 10:1 has been suggested (Sanders, 1999), and excessive endothelial function in patients with coronary heart disease. Arterio- intakes of linoleic acid should be avoided. A lower con- scler. Throm. Vasc. Biol. 19, 1681 ± 1686. centration of TFA was found in only 42% of the vegetar- Key TJA, Thorogood M, Appleby PN & Burr ML (1996): Dietary habits and mortality in 11000 vegetarians and health conscious people: results ians. TFA was found to be positively associated with serum of a 17 y follow up. Br. Med. J. 313, 775 ± 779. triglyceride in both omnivores and vegetarians in our study Key TJA, Fraser GE, Thorogood M, Appleby PN, Beral V, Reeves G et al (r ˆ 0.17, P ˆ 0.022). (1999): Mortality in vegetarians and nonvegetarians: detailed ®ndings In conclusion, there are some favourable features in the from collaborative analysis of 5 prospective studies. Am. J. Clin. Nutr. serum fatty acid pro®le of the local population with respect 70(Suppl), 516S ± 524S. Li D, Sinclair A, Mann N, Turner A, Ball M, Kelly F et al (1999): The to cardiovascular health compared with Caucasians. How- association of diet and thrombotic risk factors in healthy male vegetar- ever, the consumption of TFA is of concern and regular ians and meat-eaters. Eur. J. Clin. Nutr. 53, 612 ± 619. surveys of the serum fatty acid pro®le in the population are Lichtenstein AH (1997): Trans fatty acids, plasma lipids levels, and risk important in the monitoring of cardiovascular health. of developing cardiovascular disease: a statement for healthcare professionals from the American Heart Association. Circulation 95, 2588 ± 2590. References Lichtenstein AH, Ausman LM, Jalbert SM & Schaefer EJ (1999): Effects of different forms of dietary hydrogenated fat on serum lipoprotein Ascherio A, Hennekens CH, Buring JE, Master C, Stampfer MJ & Willett cholesterol levels. New Engl. J. Med. 340, 1933 ± 1940. WC (1994): Trans-fatty acids intake and risk of myocardial infarction. Mezzano D, Munoz X, Marinez C, Cuevas A, Panes O, Aranda E et al Circulation 89, 94 ± 101. (1999): Vegetarians and cardiovascular risk factors: hemostasis, Ascherio A, Katan MB, Zock PL, Stampfer MJ & Willett WC (1999): in¯ammatory markers and plasma homocysteine. Thromb. Haemost. Trans fatty acids and coronary heart disease. New Engl. J. Med. 340, 81, 913 ± 917. 1994 ± 1998. Nakamura T, Takebe K, Tando Y, Arai Y, Yamada N, Ishii M, Kikuchi H, Bonanome A & Grundy SM (1988): Effect of dietary stearic acid on Machida K, Imamura K & Terada A (1995): Serum fatty acid composi- plasma cholesterol and lipoprotein levels. New Engl. J. Med. 318, tion in normal Japanese and its relationship with dietary ®sh and 1244 ± 1248. contents and blood lipid levels. Ann. Nutr. & Metabol. British Nutrition Foundation (1992): Task Force on Unsaturated Fatty 39: 261 ± 270. Acids. London: Chapman and Hall. Nestel PJ, Pomeroy SE, Sasahara T, Yamashita T, Liang YL, Dart AM Carluccio MA, Massaro M, Bonfrate C, Siculella L, Maf®a M, Nicolardi G et al (1997): Arterial compliance in obese subjects is improved with et al (1999): Oleic acid inhibits endothelial activation: a direct vascular dietary plant n-3 fatty acid from ¯axseed oil despite increased LDL antiatherogenic mechanism of a nutritional component in the Mediter- oxidizability. Arterioscler. Thromb. Vasc. Biol. 17, 1163 ± 1170. ranean diet. Arterioscler. Thromb. Vasc. Biol. 19, 220 ± 228. Parkinson AJ, Cruz AL, Heyward WL, Bulkow LR, Hall D, Barstaed L & Davda RK, Stepniakowski KT, Lu G, Ullian ME, Goodfriend TL & Egan Connor WE (1994): Elevated concentrations of plasma omega-3 poly- BM (1995): Oleic acid inhibits endothelial nitric oxide synthase by a unsaturated fatty acids among Alaskan Eskimos. Am. J. Clin. Nutr. 59: protein kinase C-independent mechanism. Hypertension 26, 764 ± 770. 384 ± 388. De Lorgeril M, Salen P, Martin JL, Mamelle N, Monjaud I, Touboul P et al Popkin BM, Siega-Riz AM & Haines PS (1996): A comparison of dietary (1996): Effect of a Mediterranean type of diet on the rate of cardio- trends among racial and socioeconomic groups in the United States. vascular complications in patients with coronary artery disease. Insights New Engl. J. Med. 335, 716 ± 720. into the cardioprotective effect of certain nutrients. J. Am. Coll. Cardiol. Roshanai F & Sanders TA (1984): Assessment of fatty acid intakes in 28, 1103 ± 1108. vegans and omnivore. Hum. Nutr. Appl. Nutr. 38, 345 ± 354. Department of Health (1994): Public Health Report No. 1. A New Sanders TAB (1999): requirements of vegetarians Approach to Public Health. Hong Kong: Coronary Heart Disease and in pregnancy, lactation and infancy. Am. J. Clin. Nutr. 70(Suppl), Lung Cancer. 555S ± 559S. Fraser GE (1999): Associations between diet and cancer, ischemic Selje¯ot I, Arnesen H, Brude IR, Nenseter MS, Drevon CA & Hjermann I heart disease, and all-cause mortality in non-hispanic white (1998): Effects of omega-3 fatty acids and=or antioxidants on endothe- California Seventh-day Adventists. Am. J. Clin. Nutr. 70(Suppl), lial cell markers. Eur. J. Clin. Invest. 28, 629 ± 635. 532S ± 538S. Simon JA, Hodgkins ML, Browner WS, Neuhaus JM, Bernert JT & Hulley Friedewald WT, Levy RI & Frederickson DS (1972): Estimation of the SB (1995). Serum fatty acids and the risk of coronary heart disease. Am. concentration of low density lipoprotein cholesterol in plasma, without J. Epidemiol. 142, 469 ± 476. the use of the preparative ultra centrifuge. Clin. Chem. 18, 499 ± 502. Tagawa H, Shimokawa H, Tagawa T, Kuroiwa-Matsumoto M, Hirooka Y Harris WS, Connor WE, Inkeles SB & Illingworth DR (1984): Dietary & Takeshita A (1999): Long-term treatment with omega-3 fatty acids prevent carbohydrate-induced hypertriglycerid- augments both nitric oxide-mediated and non-nitric oxide-mediated emia. 33, 1016 ± 1019. endothelium-dependent forearm vasodilation in patients with coronary Holland B, Welch AA, Unwin ID, Buss DH, Paul AA & Southgate DAT artery disease. J. Cardiovasc. Pharm. 33, 633 ± 640. (1992): McCance and Widdowson's The Composition of Foods, 5th edn. Tsang YS & Fung W (1991): Diet and Health Guide and Composition of Food London: The Royal Society of and Ministry of Agriculture, in Southern China. China: Zhongshan University Press (in Chinese). Fisheries and Food. Willett WC, Stampfer MJ, Manson JE, Colditz GA, Speizer FE, Bernard Holman RT, Smythe L & Johnson S (1979): Effect of sex and age on fatty AR et al (1993): Intake of trans fatty acids and risk of coronary heart acid composition of human serum lipids. Am. J. Clin. Nutr. 32: 2390 ± disease among women. Lancet 341, 581 ± 585. 2399. Woo J, Leung SSF, Ho SC, Sham A, Lam TH & Janus ED (1997). Dietary Hunter DJ, Rimm EB, Sacks FM, Stampfer MJ, Colditz GA, Litin LB et al practices and liquid intake in relation to plasma lipid pro®le in Hong (1992): Comparison of measures of fatty acid intake by subcutaneous Kong Chinese. Eur. J. Clin. Nutr. 51, 467 ± 471. fat aspirate, food frequency questionnaire, and diet records in a free- Woo KS & Donnan SPB (1989): Epidemiology of coronary heart disease living population of US men. Am. J. Epidemiol. 135, 418 ± 427. in the Chinese. Int. J. Cardiol. 24, 83 ± 93.

European Journal of Clinical Nutrition