It is widely recognised that overall morbidity and mortality are lower in vegetarians compared with omnivores(Reference Messina and Burke1). The dietary patterns of vegetarians as well as their healthful lifestyle practices are thought to at least partly explain these differences. One notable difference relates to the type and amount of fat in the diet. Vegetarian diets are slightly lower in total fat than omnivorous diets(Reference Davis and Kris-Etherton2). However, vegetarians eat about one-third less saturated fat (vegans about one-half) and about one-half as much cholesterol (vegans consume none) as omnivores(Reference Janelle and Barr3). In addition, vegetarian diets are rich in fibre, Mg, Fe3+, folic acid, vitamins C and E, n-6 PUFA, phytochemicals and antioxidants(Reference Li4).
However, vegetarian diets are low in Na, Zn, Fe2+, vitamins A, B12 and D, and especially n-3 PUFA(Reference Li4). A low intake of total fat, SFA and Na and a high intake of fibre, phytochemicals and antioxidants in vegetarians are associated with low blood pressure and BMI, which are known to reduce the risk of CVD(Reference Li4). Vitamin B12, which is mainly from seafood, animal meats, eggs and liver and not found in plant foods, plays an important role in homocysteine (Hcy) metabolism. Vitamin B12 deficiency directly leads to high plasma Hcy, which is an independent risk factor for CVD(Reference Mann, Li and Sinclair5).
With respect to essential fatty acid intake and balance, vegetarian diets appear to offer no advantages over omnivorous dietary patterns. Some have suggested that vegetarians could be at a significant disadvantage, as consumption of α-linolenic acid (18 : 3n-3) is low relative to linoleic acid (18 : 2n-6), resulting in the limited conversion of α-linolenic acid to EPA (20 : 5n-3) and DHA (22 : 6n-3)(Reference Lee, Woo and Chen6). In addition, vegetarians consume very little EPA and DHA(Reference Davis and Kris-Etherton2).
Our previous studies reported that platelet/plasma phospholipid (PL) n-3 PUFA were negatively associated with plasma Hcy in middle-aged and geriatric hyperlipaemia patients(Reference Li, Yu and Xie7) and in healthy male subjects(Reference Li, Mann and Sinclair8). Our animal study suggested that DHA decreases plasma Hcy concentration by regulating critical gene expression and enzyme activity(Reference Huang, Wahlqvist and Li9). Our population studies found that dietary fatty acids interact with methylenetetrahydrofolate reductase (MTHFR) and methionine adenosyltransferase I, α (MAT1A) genetic variants in determining plasma Hcy concentration(Reference Huang, Tucker and Lee10, Reference Huang, Tucker and Lee11).
To date, limited research comparing plasma PL fatty acid composition of omnivores and vegetarians suggests that vegetarians have lower serum and/or platelet levels of DHA(Reference Sanders12). We previously also reported that n-3 PUFA were lower in vegetarians than in omnivores(Reference Li, Ball and Bartlett13). However, no study has reported the associations of plasma PL fatty acids with plasma Hcy concentration in Chinese vegetarians.
The purpose of the present study is to investigate the status of plasma PL fatty acids and to examine the potential relationship between PL fatty acids and plasma Hcy levels in Chinese vegetarians and omnivores.
Materials and methods
Subjects
The present study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects/patients were approved by the Ethics Committee, College of Biosystem Engineering and Food Science, Zhejiang University, China. Written informed consent was obtained from all subjects/patients.
A total of 103 male vegetarians (aged 40 (sd 10) years) were recruited in Linyin Temple, Hangzhou. A vegetarian was defined as someone who ate no red meat, consumed fish or chicken less than once per week, and had been following this diet for at least 6 months before the study. A total of 128 male omnivores (aged 44 (sd 8) years) were recruited through a health check programme during the period of October 2010 to March 2011 in the Zhejiang Hospital. An omnivore was defined as someone who ate meat at least five times per week.
Blood collection
Subjects visited the Zhejiang Hospital in the morning following an overnight fast. Subjects were allowed to sit relaxed for 10 min, and the subject's weight, height, waist:hip ratio and blood pressure were then measured. Subsequently, venous blood was taken in plain and EDTA vacuum tubes with a twenty-one-gauge needle (Longhe). Within 1 h after blood collection, plasma samples were prepared quickly by using a refrigerated centrifuge, and aliquoted into separate tubes and stored at − 20°C until analysis.
Laboratory measurements
Plasma total Hcy was determined by polarised fluorescence immunoassay in an AxSYM system (Abbott Laboratories)(Reference Vizcaino, Diez-Ewald and Herrmann14). Hyperhomocysteinaemia (HHcy) was defined as a Hcy level above the 95th percentile (14 μmol/l)(Reference Cattaneo, Martinelli and Mannucci15). Plasma folate and vitamin B12 were measured using immulite chemiluminescent kits according to the manufacturer's instructions (Diagnostic Products Corporation/Siemens). Plasma lipids were determined using an autoanalyser (Olympus AU2700; Olympus France), via commercially available kits (Olympus). Fasting glucose was measured by standard methods as described previously(Reference Li, Xu and Takase16). Total lipid content of plasma was extracted with solvents, the PL fraction was separated by TLC, and fatty acid methyl esters were prepared and separated by GLC as described previously(Reference James17).
Statistical analysis
Data analyses were performed using SAS for Windows, version 9.1 (SAS Institute). All continuous variables were examined for a normal distribution. We categorised plasma PL fatty acids into quantiles (median) using the SAS program (PROC RANK). Differences between the two groups for each outcome were analysed using ANOVA. The associations between plasma PL fatty acid composition and Hcy were determined by partial correlation, controlling for potential confounding factors (age, sex, vitamin B12 and folate). Population medians for plasma PL fatty acids such as total SFA, MUFA and PUFA, and n-3:n-6 were used as cut-offs to dichotomise these variables. All data are expressed as means and standard deviations. Differences between the groups were considered to be statistically significant at P< 0·05.
Results
Compared with omnivores, body weight was significantly lower in vegetarians (P= 0·044). Systolic blood pressure was significantly lower in vegetarians (P= 0·001), while diastolic blood pressure was significantly higher in vegetarians (P= 0·005). In addition, plasma total cholesterol, total TAG, HDL-cholesterol, LDL-cholesterol, glucose and vitamin B12 were also significantly lower in vegetarians. However, plasma Hcy and folate were significantly higher in vegetarians than those in omnivores (Table 1).
Demographic and biochemical measurements in vegetarians and omnivores (Mean values and standard deviations)
SBP, systolic blood pressure; DBP, diastolic blood pressure; TC, total cholesterol; HDL-C, HDL-cholesterol; LDL-C, LDL-cholesterol; ALB, albumin; Hcy, homocysteine.
* P< 0·05 indicates significant difference between the groups (t test).
Plasma PL fatty acid composition was significantly different between vegetarians and omnivores. Compared with omnivores, plasma PL n-3 PUFA (P< 0·001), 22 : 6n-3 (P< 0·001), 22 : 5n-6 (P= 0·021), 22 : 2n-6 (P< 0·001) and SFA (P= 0·017) were significantly lower in vegetarians, while plasma PL n-6 PUFA (P= 0·007) and total PUFA (P< 0·001) were significantly higher in vegetarians. Plasma PL MUFA (P= 0·094) and n-3:n-6 PUFA (P= 0·064) were not significantly different between the two groups (Table 2).
Plasma phospholipid (PL) fatty acid composition in vegetarians and omnivores (Mean values and standard deviations)
* P< 0·05 is considered to indicate a significant difference.
† The difference between vegetarians and omnivores is determined by using the general linear model controlled for potential confounding factors (age and BMI).
In vegetarians, plasma PL 22 : 6n-3 (r − 0·257, P= 0·046) was significantly negatively associated with plasma Hcy. In addition, plasma PL n-6 PUFA (r 0·249, P= 0·045) were significantly positively associated with plasma Hcy. In omnivores, plasma PL 18 : 1n-7 (r 0·237, P= 0·030) was significantly positively associated with plasma Hcy in omnivores. Plasma PL 22 : 1n-9 (r − 0·228, P= 0·037) and 22 : 6n-3 (r − 0·217, P= 0·048) were negatively associated with plasma Hcy in omnivores (Table 3). There were no further partial correlations between any fatty acid and Hcy in either vegetarians or omnivores.
Partial correlations between plasma phospholipid (PL) fatty acid compositions and plasma homocysteine*
* The associations of plasma PL fatty acid compositions with plasma homocysteine concentration were tested by using Pearson's partial correlation after controlling for confounding factors (age, BMI, vitamin B12 and folate).
Multiple regression analysis adjusted for age, sex, BMI, vitamin B12, folate, glucose, lipids, insulin, vegetarian diet and fatty acids was created to identify potential independent correlates of plasma Hcy. The only significant correlates of Hcy were age (P= 0·001), vitamin B12 (P= 0·009), folate (P= 0·015), height (P= 0·010), glucose (P= 0·007), SFA (P= 0·040) and 22 : 4n-6 (P= 0·024). These variables explained 49 % of the variance in plasma Hcy (Table 4).
Independent predictors of plasma homocysteine after adjustment for other clinical characteristics and risk factors
We further examined the prevalence of HHcy in vegetarians and omnivores. We observed that the prevalence of HHcy in vegetarians (26·47 %) was significantly higher than that in omnivores (13·28 %) (P< 0·01; Fig. 1).
Prevalence of hyperhomocysteinaemia (HHcy) in vegetarians (n 103) and omnivores (n 128). The prevalence of HHcy in vegetarians was significantly higher than that in omnivores (P< 0·01; χ2 test).
We further examined the associations of PL fatty acids with the prevalence of HHcy. The results showed that PL SFA were significantly positively associated with the prevalence of HHcy in vegetarians (P= 0·024) and omnivores (P= 0·018) (Fig. 2). In addition, PL 22 : 4n-6 and 18 : 2n-6 were also significantly positively associated with the prevalence of HHcy in vegetarians (Figs. 3 and 4, respectively). We also found that PL 22 : 6n-3 was significantly negatively associated with the prevalence of HHcy in vegetarians (P= 0·017; Fig. 5).
Association of plasma phospholipid SFA with the prevalence of hyperhomocysteinaemia (HHcy) in vegetarians (lower SFA, n 34; higher SFA, n 36) and omnivores (lower SFA, n 48; higher SFA, n 46). Phospholipid SFA were significantly positively associated with the prevalence of HHcy in vegetarians (P= 0·024; χ2 test) and omnivores (P= 0·018; χ2 test).
Association of plasma phospholipid 22 : 4n-6 with the prevalence of hyperhomocysteinaemia (HHcy) in vegetarians (lower 22 : 4n-6, n 38; higher 22 : 4n-6, n 39) and omnivores (lower 22 : 4n-6, n 50; higher 22 : 4n-6, n 54). Phospholipid 22 : 4n-6 was significantly positively associated with the prevalence of HHcy in vegetarians (P= 0·004) and omnivores (P= 0·017).
Association of plasma phospholipid 18 : 2n-6 with the prevalence of hyperhomocysteinaemia (HHcy) in vegetarians (lower 18 : 2n-6, n 40; higher 18 : 2n-6, n 45) and omnivores (lower 18 : 2n-6, n 57; higher 18 : 2n-6, n 47). Phospholipid 18 : 2n-6 was significantly positively associated with the prevalence of HHcy in vegetarians (P= 0·008) and omnivores (P= 0·164).
Association of plasma phospholipid 22 : 6n-3 with the prevalence of hyperhomocysteinaemia (HHcy) in vegetarians (lower 22 : 6n-3, n 40; higher 22 : 6n-3, n 38) and omnivores (lower 22 : 6n-3, n 51; higher 22 : 6n-3, n 53). Phospholipid 22 : 6n-3 was significantly negatively associated with the prevalence of HHcy in vegetarians (P= 0·017) and omnivores (P= 0·264).
Discussion
In the present study, we found that the prevalence of HHcy in vegetarians (26·47 %) was significantly higher than that in omnivores (13·28 %). Plasma PL n-3 PUFA, 22 : 6n-3 and SFA were significantly lower in vegetarians, while plasma PL n-6 PUFA and total PUFA were significantly higher in vegetarians than in omnivores. Plasma PL 22 : 6n-3 was significantly negatively associated with plasma Hcy. In addition, plasma PL SFA were significantly positively associated with the prevalence of HHcy in vegetarians and omnivores. To our knowledge, this is the first report; however, the mechanism at this stage is not known.
The risk for many chronic diseases, including CVD, is influenced by dietary fatty acid intake(Reference Kris-Etherton, Harris and Appel18). A higher degree of incorporation of n-3 PUFA into myocardial membranes reduces deaths following myocardial ischaemia(Reference Leaf, Kang and Xiao19). The n-3 PUFA content in erythrocyte membranes reflects that in cardiac membranes(Reference Harris, Sands and Windsor20). A low erythrocyte 20 : 5n-3 and 22 : 6n-3 percentage content has been identified as a risk indicator for death from CVD(Reference Harris and Von Schacky21). Because fatty acid profiles of platelet and plasma/serum PL reflect an individual's type of dietary fat intake(Reference Li, Yu and Xie7), in addition, the database on the individual fatty acid content of all foods is not available for Chinese populations, no method has been developed to accurately estimate the dietary intake of individual fatty acids since it is rare for people to prepare all meals at home. The amount of fat intake is highly variable depending on the ingredients used to prepare dishes in restaurants or canteens. Thus, compositions of platelet PL FA were used as a surrogate marker of dietary intake of FA(Reference Huang, Bhulaidok and Cai22). In the present study, we observed a higher level of plasma PL 20 : 4n-6 but lower levels of n-3 PUFA, 20 : 5n-3 and 22 : 6n-3, and the n-3:n-6 ratio in vegetarians than those in omnivores, which may tend to promote thrombotic risk. However, lower concentrations of plasma PL SFA in vegetarians may provide beneficial effects on CVD risk(Reference Li, Ball and Bartlett13).
Previous studies and the present study demonstrated that n-3 PUFA and 22 : 6n-3 in plasma or erythrocytes were significantly lower in vegetarians. Sanders(Reference Sanders12) reported that the proportions of 22 : 6n-3 in plasma, erythrocytes, breast milk and tissues are substantially lower in vegans and vegetarians compared with omnivores. Kornsteiner et al. (Reference Kornsteiner, Singer and Elmadfa23) also demonstrated that vegetarians and vegans, who do not eat meat or fish, tend to have very low or negligible intakes of 20 : 5n-3 as well as 22 : 6n-3. Fokkema et al. (Reference Fokkema, Brouwer and Hasperhoven24) investigated the PUFA status of Dutch vegans and omnivores in erythrocyte membranes. They did not find significant differences in total n-3 PUFA in the erythrocytes; however, 20 : 5n-3 and 22 : 6n-3 were significantly reduced in vegans compared with omnivores. On the other hand, the results of Dutch vegans showed a higher 22 : 5n-3 content compared with Dutch omnivores(Reference Fokkema, Brouwer and Hasperhoven24), which is consistent with the present results. Rosell et al. (Reference Rosell, Lloyd-Wright and Appleby25) investigated n-3 PUFA in the plasma of British meat-eating, vegetarian and vegan men, and found that 20 : 5n-3, docosapentaenoic acid (22 : 5n-3) and 22 : 6n-3 were markedly decreased. Li et al. (Reference Li, Ball and Bartlett13) demonstrated a decreased content of 20 : 5n-3, 22 : 6n-3 and total n-3 PUFA in Australian vegetarian females. Sanders et al. (Reference Sanders, Ellis and Dickerson26) showed that erythrocytes from vegans contained lower proportions of 20 : 5n-3, 22 : 5n-3 and 22 : 6n-3 and higher proportions of 18 : 2n-6, 20 : 2n-6 and 22 : 4n-6. Kornsteiner et al. (Reference Kornsteiner, Singer and Elmadfa23) documented that the imbalance in the n-6:n-3 ratio and the limited dietary sources of 20 : 5n-3 and 22 : 6n-3 in vegans and vegetarians led to reductions in 20 : 5n-3, 22 : 5n-3, 22 : 6n-3 and n-3 PUFA in phosphatidylethanolamine, phosphatidylcholine and phosphatidylserine compared with omnivores and semi-omnivores. Animal studies and human intervention studies have demonstrated that a high n-3 PUFA intake increases the level of n-3 PUFA in tissues(Reference Kornsteiner, Singer and Elmadfa23, Reference Huang, Sinclair and Shen27). Therefore, the low content of plasma PL n-3 PUFA reflects the limited n-3 PUFA dietary intake in vegetarians(Reference Li4).
Ovo-lacto-vegetarians consume minimal amounts of 20 : 5n-3 and varying amounts of 22 : 6n-3 from eggs, milk and dairy products. Vegans consume negligible amounts of long-chain n-3 PUFA (20 : 5n-3 and 22 : 6n-3) and rely entirely on the in vivo biosynthesis of n-3 PUFA from the precursor 18 : 3n-3, but the conversion via desaturation and elongation, especially to 22 : 6n-3, is not efficient(Reference Burdge and Wootton28). Previous studies with stable isotopically labelled 18 : 3n-3 have shown the conversion of 18 : 3n-3 to 20 : 5n-3 varying from 6–21 % to much lower values (0·1–0·2 %)(Reference Hussein, Ah-Sing and Wilkinson29). In one study, the conversion of 18 : 3n-3 to 22 : 6n-3 has been reported to range from 4–9 % to 0·04 %(Reference Simopoulos30), or with undetectable 22 : 6n-3 synthesis. The lack of 20 : 5n-3 and 22 : 6n-3 in vegetarian diets is reflected in reduced amounts of these fatty acids in platelets, erythrocytes and plasma(Reference Kornsteiner, Singer and Elmadfa23). Thus, the uptake of preformed 22 : 6n-3 from the diet may be critical for maintaining adequate membrane 22 : 6n-3 concentrations in vegetarians(Reference Kornsteiner, Singer and Elmadfa23).
In recent years, many epidemiological studies have investigated the associations of n-3 PUFA with CVD risk factors such as plasma Hcy. In the present study, we demonstrated that plasma Hcy concentration in vegetarians was significantly higher than that in omnivores. The prevalence of HHcy in vegetarians (26·47 %) was significantly higher than that in omnivores (13·28 %). This observation can be explained by the lower level of plasma vitamin B12 and n-3 PUFA in vegetarians. Previous studies suggested that n-3 PUFA play an important role in Hcy metabolism(Reference Huang, Wahlqvist and Li9–Reference Huang, Tucker and Lee11, Reference Huang, Wahlqvist and Li31). In the present study, plasma PL 22 : 6n-3 was significantly negatively associated with plasma Hcy. In our previous study, plasma Hcy concentration was negatively correlated with n-3 PUFA and the ratio of n-3:n-6 PUFA intake in Puerto Rican adults(Reference Huang, Tucker and Lee11). Li et al. (Reference Li, Yu and Xie7) also found that increased total n-3 PUFA and n-3 : n-6 PUFA in platelet PL are associated with decreased plasma Hcy in middle-aged and geriatric hyperlipaemia patients in Hangzhou, China. In addition, plasma PL 22 : 6n-3, n-3 PUFA and the n-3:n-6 PUFA ratio were also negatively correlated with plasma Hcy in healthy Australian males(Reference Li, Mann and Sinclair8). Over the past two decades, several intervention studies of small sample size and short duration have documented the effects of n-3 PUFA on plasma Hcy concentration(Reference Zeman, Zak and Vecka32–Reference Beavers, Beavers and Bowden34). However, these results were not consistent: some studies have not shown a significant decrease in plasma Hcy(Reference Piolot, Blache and Boulet35, Reference Carrero, Fonolla and Marti36), while other studies have documented a plasma Hcy-lowering effect after n-3 PUFA supplementation(Reference Zeman, Zak and Vecka32, Reference Carrero, Lopez-Huertas and Salmeron33, Reference Pooya, Jalali and Jazayery37). Therefore, we conducted a meta-analysis to increase the sample size and demonstrated that a high consumption of n-3 PUFA decreases plasma Hcy(Reference Huang, Zheng and Chen38).
The potential mechanisms by which n-3 PUFA decrease plasma Hcy have been investigated in our animal and population studies(Reference Huang, Wahlqvist and Li9–Reference Huang, Tucker and Lee11). In our animal study(Reference Huang, Wahlqvist and Li9), we found that plasma Hcy was significantly decreased by tuna oil rich in 22 : 6n-3. Methionine adenosyl transferase (MAT) activity was significantly increased and MAT mRNA expression was significantly up-regulated by 22 : 6n-3; cystathionine-γ-lyase mRNA expression was significantly up-regulated by 22 : 6n-3. We suggested that 22 : 6n-3 decreases the concentration of Hcy despite increasing MAT activity and up-regulating MAT mRNA expression through compensatory cystathionine-γ-lyase mRNA expression, both of which are involved in Hcy metabolism(Reference Huang, Wahlqvist and Li9). However, HHcy has multifactorial determinants. It reflects genetic and environmental factors or their interactions. Therefore, genetic variants involved in Hcy metabolic pathways may modify the effects of dietary fatty acids on plasma Hcy in humans. Our previous population studies(Reference Huang, Tucker and Lee10, Reference Huang, Tucker and Lee11) have shown that two functional MTHFR variants, 1298A>C and 677C>T, which are not in linkage disequilibrium in Boston Puerto Rican adults, are significantly associated with hypertension. Importantly, these variants exhibited significant interactions with intakes of total and n-6 PUFA and with the n-3:n-6 PUFA ratio of the diet in determining plasma Hcy concentration. Participants with combined genotypes of both SNP (677 TT with 1298 AC or CC) who consumed high levels of n-3 PUFA (>0·66 % energy) had lower plasma Hcy compared with those who had the same genotype and consumed low levels of n-3 PUFA ( ≤ 0·66 % energy). Therefore, it has been suggested that dietary PUFA intake modulates the effect of MTHFR variants on plasma Hcy(Reference Huang, Tucker and Lee11). Moreover, the genetic variant MAT1A 3U1510 displays a significant interaction with the dietary n-3:n-6 PUFA ratio in determining plasma Hcy. Homozygotes for 3U1510G have significantly lower plasma Hcy concentrations than those who are major allele homozygotes and heterozygotes (AA+AG) and when the n-3:n-6 ratio is >0·09. Also, two other MAT1A variants (d18777 and i15752) show significant interactions with different constituents of dietary fat in influencing Hcy concentration. Furthermore, haplotypes consisting of three variants display a strong interaction with the n-3:n-6 ratio influencing Hcy concentrations(Reference Huang, Tucker and Lee10).
Based on the present data, vegetarians should probably increase their relatively low dietary vitamin B12 and n-3 PUFA and decrease dietary SFA, and thereby reduce the risk of HHcy.
Acknowledgements
This study was supported by a grant from the National Natural Science Foundation of China (no. 30972464), the National Basic Research Program of China (973 Program: 2011CB504002), Zhejiang Science Foundation of Aging (2008ZJ003) and Zhejiang Natural Science Foundation (Y2101133). T. H., X. Y. and T. S. carried out the studies and drafted the manuscript; M. L. W. and D. L. participated in the project design and manuscript preparation. All authors read and approved the final manuscript. The authors have no financial/commercial conflicts of interest in relation to the present study.
