Fatty Acid Profile and Cardiometabolic Markers in Relation with Diet Type and Omega-3 Supplementation in Spanish Vegetarians

Plant-based diets are becoming increasingly popular, and scientific information concerning the nutritional status in this population is needed. This study determined the fatty acid profile of Spanish lacto-ovo vegetarians (LO-vegetarians) and vegans. Participants were 104 healthy adults, LO-vegetarians (n = 49) and vegans (n = 55). Lifestyle habits and consumption of food and omega-3 supplements were estimated by questionnaires. BMI, blood pressure, and abdominal and body fat were determined. Serum was collected to analyze fatty acids, glucose, lipids, homocysteine, insulin, and leptin. Volunteers were classified according to serum omega-6 to omega-3 (n-6/n-3) ratio into three groups: n-6/n-3 < 10, n-6/n-3 ≥ 10 to 20, and n-6/n-3 > 20. Results showed low cardiovascular risk and high insulin sensitivity with negligible differences between diet types. Linoleic acid (C18:2n-6) was the major serum fatty acid, followed by oleic (C18:1n-9) and palmitic (C16:0) acids. In contrast, serum eicosapentaenoic acid (EPA, C20:5n-3) and docosahexaenoic acid (DHA, C22:6n-3) were (median, interquartile range) 0.27, 0.18% and 1.59, and 0.93%, respectively. Users of n-3 supplements (<10% of total vegetarians) had significantly higher EPA than non-users, while frequent consumption of flax-seeds was associated with increased α-linolenic acid (C18:3n-3). However, neither n-3 supplementation nor food consumption affected DHA levels in this vegetarian population.


Introduction
Diets exclusively based on plant foods have become progressively common in developed countries. However, the scientific evidence concerning the nutritional status of this population is limited. Vegetarian diets are usually associated with the prevention of some diseases such as ischemic heart disease, obesity and type 2 diabetes [1]. Nevertheless, they may result in insufficient intake of several nutrients such vitamin B12, vitamin D, and essential fatty acids [2,3]. There are two essential polyunsaturated fatty acids (PUFAs) for human nutrition: linoleic acid (LA, C18:2n-6) and α-linolenic acid (ALA, C18:3n-3), of the n-6 and n-3 PUFAs pathways respectively. These fatty acids are substrates for long-chain PUFAs which exert important functions. LA is a precursor of arachidonic acid (AA, C20:4n-6), and ALA of eicosapentaenoic acid (EPA, C20:5n-3) and docosahexaenoic acid (DHA, C22:6n-3) [3]. In addition, the eicosanoids derived from the n-6 or n-3 pathways have pro-inflammatory or anti-inflammatory activities, respectively, [4] and these should be balanced for adequate health maintenance.

Blood Sampling and Cardiometabolic Markers Determinations
Blood samples were collected in Vacuette Z Serum Sep Clot Activator tubes (Greiner Bio-One GmbH, Frickenhausen, Germany) at the Human Nutrition Unit (UNH) of the Institute of Food Science, Technology and Nutrition (ICTAN-CSIC, Madrid, Spain) from 8:00 h to 8:30 h, after a 12 h fasting overnight period. Serum was isolated by centrifugation in a Jouan CR-312 centrifuge (Jouan Ltd, Ilkeston, UK) at 1000 g for 15 min and samples were then stored at −30 • C for fatty acid analysis. Fatty acids were measured in serum because serum fatty acid levels are correlated with recent intake and reflect the dietary habits of the subjects at the time of the study [10,11]. Moreover, elaidic acid (ELA, C18:1n-9t) was analyzed as an unsaturated trans-fatty acids status indicator because it is the major trans-fatty acid found in hydrogenated vegetable oils.

Serum Fatty Acid Analysis
Fatty acids analysis was performed by gas chromatography (GC) using a Varian CP-3800 gas chromatograph with a flame ionization detector (Varian, Inc., Palo Alto, CA, USA). Fatty acid methyl esters (FAMEs) were obtained based on the method developed by Lepage and Roy [12]. Two mL of a 4:1 (v/v) methanol:benzene solution were added to 200 µL of serum samples plus 25 µL of methyl heptadecanoate, the internal standard. After mixing carefully, 200 µL of acetyl chloride were slowly incorporated while stirring. Tubes were firmly closed and subjected to methanolysis at 100 • C for 1 h. Once tubes were cooled, 5 mL of potassium carbonate solution (6% w/v) were added to stop the reaction and to neutralize the mixture. Tubes were stirred and centrifuged (1000 g, 10 min) to separate the two resulting phases. The upper phase was transferred to GC vials and injected into the gas chromatograph. The separation of the FAMEs was carried out on a BPX70 capillary column (30 × 0.25 × 0.25) from SGE Analytical Science (Melbourne, Australia). Fatty acid peak areas and concentrations were determined using the Empover 3 software (Waters Corporation, Milford, MA, USA). For statistical analyses, data were expressed as percentage of total fatty acids.

Statistical Analysis
The distribution of the variables was analyzed by the Kolmogorov-Smirnov test, and natural log-transformed data were used for statistical analysis when it was possible. However, several variables could not be normalized: STE, ELA, OA, MUFA, n-6, and the ratios OA/STE, LA/OA, ALA/LA, GLA/LA, EPA/ALA, DHA/AA, PUFA/SFA, and PUFA/MUFA. The data were expressed as median (interquartile range), except for the proportion of subjects in each gender; age; n-3 supplementation; smoking habits; education level; and physical activity groups, which were expressed as n (%).
Differences between diet type groups were studied by independent samples t-test or Mann-Whitney's U test for normally distributed and non-normally distributed variables, respectively.
The studied subjects were classified into three groups according to the following cut-off levels: n-6/ n-3 ≤ 10; n-6/n-3 > 10 and ≤20; n-6/n-3 > 20. Differences among them were studied by χ 2 tests or Kruskal-Wallis tests. The significance level was set at p < 0.05, and all the statistical analyses were performed with SPSS 24.0 for Windows (IBM, Armonk, NY, USA).

Characteristics of the Participants
Characteristics of the studied population are described in Table 1. The study was completed by 104 volunteers, of which 49 were LO-vegetarians (11 men and 38 women) and 55 were vegans (12 men and 43 women). The majority of the studied individuals were young, with a median age of 30 years, non-smokers, had a bachelor's degree, and exhibited moderate physical activity. Only nine volunteers were n-3 supplement users (four LO-vegetarians and five vegans). There were no significant differences between LO-vegetarians and vegans in terms of BMI, waist circumference, body fat, abdominal fat, glucose, insulin, leptin, Hcy, TC, HDLc, LDLc, TAG, and cardiometabolic indexes (TC/HDLc, LDLc/HDLc. TAG/HDLc, HOMA-IR, QUICKI, and HOMA-B). SBP was similar in both groups, although DBP was slightly lower in vegans than in LO-vegetarians (p = 0.047).

Serum Fatty Acids and Main Cardiometabolic Indexes
Total serum fatty acids for the whole population (median, interquartile range) were 2005.8 and 607.6 µg/mL, without significant differences between women and men. Table 2 presents the percentiles of serum fatty acids percentage. Major serum fatty acids were LA, OA and PAL. In contrast, the trans-fatty acid status indicator ELA was very low. Table 3 shows the differences of the types of vegetarian diets and n-3 supplementation on serum fatty acids. Vegans compared with LO-vegetarians had significantly higher OA (p = 0.010), OA/STE (p = 0.047), MUFA (p = 0.018), and AA/(EPA+DHA) (p = 0.026), but lower ELA (p = 0.003) and n-3 (p = 0.048) levels. All other parameters were not significantly different between the two vegetarian diet groups.

LO-vegetarians and Vegans Classification According to n-6/n-3 Ratio
Volunteers were classified into three groups according to their n6/n3 ratio: <10, ≥10 to 20, and >20 (Table 4). Gender distribution neither varied among groups nor were there any differences in physical activity or anthropometric and body fat parameters. There were significantly higher proportions of LO-vegetarians and n-3 supplement users (p = 0.032 and 0.017, respectively, χ 2 test) in the lowest and middle n-6/n-3 groups compared to the n-6/n-3 > 20 group. Concerning the cardiometabolic markers, no significant differences among n-6/n-3 groups were observed for glucose, lipid and hormone levels (insulin, leptin), but serum Hcy was significantly lower in the n-6/n-3 < 10 group than in the other groups (p = 0.003). The level of the monounsaturated fatty acid POA was higher in the n6/n3 < 10 group (p = 0.024) while the n-6 LA and the n-3 fatty acids ALA, EPA and DHA were the main contributors to the differences among the n-6/n-3 groups.  The analysis of the FFQ data confirms that participants in the LO-vegetarian group consumed milk, eggs and their derived products while vegans did not declare any consumption of these products. Concerning the culinary fat, most of the vegetarians consumed olive oil several times per day (Table 5). Table 5 shows the frequency of consumption of nuts, seeds, oils, and other fat-rich foods of the subjects classified according to n-6/n-3 groups. Flax-seeds were more frequently consumed in the n-6/n-3 < 10 group but the differences were marginally significant (p = 0.056) and no other differences were observed. Figure 1 presents serum ALA and DHA according to frequency of consumption of flax-seeds. Serum ALA increases as flax-seeds intake increases while no variation in serum DHA was observed in relation to the consumption of this food item. Table 5. Frequency of consumption of foods according to n-6/n-3 ratio. Results are expressed as median (interquartile range). Considered categories were: 0 (never), 1 (two to four times a month), 2 (two to three times a week), 3 (four to six times a week), 4 (once a day), and 5 (≥2 times/day). * Pearson's χ 2 tests (Flax-seeds p = 0.056).
Nutrients 2019, 11, x FOR PEER REVIEW 8 of 14 The analysis of the FFQ data confirms that participants in the LO-vegetarian group consumed milk, eggs and their derived products while vegans did not declare any consumption of these products. Concerning the culinary fat, most of the vegetarians consumed olive oil several times per day (Table 5). Table 5 shows the frequency of consumption of nuts, seeds, oils, and other fat-rich foods of the subjects classified according to n-6/n-3 groups. Flax-seeds were more frequently consumed in the n-6/n-3 < 10 group but the differences were marginally significant (p = 0.056) and no other differences were observed. Figure 1 presents serum ALA and DHA according to frequency of consumption of flax-seeds. Serum ALA increases as flax-seeds intake increases while no variation in serum DHA was observed in relation to the consumption of this food item. Table 5. Frequency of consumption of foods according to n-6/n-3 ratio. Results are expressed as median (interquartile range). Considered categories were: 0 (never), 1 (two to four times a month), 2 (two to three times a week), 3 (four to six times a week), 4 (once a day), and 5 (≥2 times/day). * Pearson´s χ 2 tests (Flax-seeds p = 0.056).

Discussion
In this study, serum fatty acids profiles of Spanish vegetarians are presented for the first time. Participants were healthy young adults, with similar sample sizes of LO-vegetarians and vegans, and less than 10% of the volunteers consumed n-3 supplements. Lifestyle habits, body composition, physical activity, and cardiometabolic markers indicate that this population has low cardiovascular risk and high insulin sensibility.
The major serum fatty acid was LA, followed by OA and PAL; while levels of n-3 PUFA, except for ALA, were low in comparison with the data reported for general population [13]. These results clearly show an inadequate n-6/n-3 ratio, indicating a metabolic imbalance. While the results of LA and n-3 levels are consistent with other reports of vegetarians [5,14], OA levels were remarkably high in comparison to other countries, representing about 20-25% of the total serum fatty acids. For instance, OA values of vegetarians in China, South India and Australia were around 1%, 7%, and 10%, respectively [15][16][17]. This is certainly due to the high olive oil consumption characteristic of the Mediterranean diet in Spain. Consistently, this oil was consumed by the volunteers several times per day (Table 5) and our OA levels were similar to those of the Spanish PREDIMED study at baseline point [13,18].
Moreover, vegans presented higher OA (and consistently higher MUFA and OA/STE) and lower ELA, the trans isomer of OA, compared to LO-vegetarians. This trans fatty acid is a marker of ultraprocessed food, e.g., margarine, chocolate, potato flakes, potato fries, breakfast cereals, etc. [19], which would suggest a more unfavorable diet in LO-vegetarians than vegans. However, the values obtained in our vegetarian population, either LO-vegetarians or vegans, were much lower than those of other European populations. In this regard, plasma ELA was about three times higher in the EPIC cohort [20,21]. The results of the FFQ of the present study confirm that the frequency of consumption of industrial processed food was similar in LO-vegetarians and vegans, except for food items that contained eggs and milk whose consumption was declared null by the vegans. Natural ELA is present in ruminant's milk, and thus the consumption of milk and dairy products may have contributed to the slightly higher ELA in LO-vegetarians in comparison with vegans [22].

Discussion
In this study, serum fatty acids profiles of Spanish vegetarians are presented for the first time. Participants were healthy young adults, with similar sample sizes of LO-vegetarians and vegans, and less than 10% of the volunteers consumed n-3 supplements. Lifestyle habits, body composition, physical activity, and cardiometabolic markers indicate that this population has low cardiovascular risk and high insulin sensibility.
The major serum fatty acid was LA, followed by OA and PAL; while levels of n-3 PUFA, except for ALA, were low in comparison with the data reported for general population [13]. These results clearly show an inadequate n-6/n-3 ratio, indicating a metabolic imbalance. While the results of LA and n-3 levels are consistent with other reports of vegetarians [5,14], OA levels were remarkably high in comparison to other countries, representing about 20-25% of the total serum fatty acids. For instance, OA values of vegetarians in China, South India and Australia were around 1%, 7%, and 10%, respectively [15][16][17]. This is certainly due to the high olive oil consumption characteristic of the Mediterranean diet in Spain. Consistently, this oil was consumed by the volunteers several times per day (Table 5) and our OA levels were similar to those of the Spanish PREDIMED study at baseline point [13,18].
Moreover, vegans presented higher OA (and consistently higher MUFA and OA/STE) and lower ELA, the trans isomer of OA, compared to LO-vegetarians. This trans fatty acid is a marker of ultra-processed food, e.g., margarine, chocolate, potato flakes, potato fries, breakfast cereals, etc. [19], which would suggest a more unfavorable diet in LO-vegetarians than vegans. However, the values obtained in our vegetarian population, either LO-vegetarians or vegans, were much lower than those of other European populations. In this regard, plasma ELA was about three times higher in the EPIC cohort [20,21]. The results of the FFQ of the present study confirm that the frequency of consumption of industrial processed food was similar in LO-vegetarians and vegans, except for food items that contained eggs and milk whose consumption was declared null by the vegans. Natural ELA is present in ruminant's milk, and thus the consumption of milk and dairy products may have contributed to the slightly higher ELA in LO-vegetarians in comparison with vegans [22]. Nevertheless, the generally low ELA levels obtained in the present study suggest a low consumption of ultra-processed food in both LO-vegetarians and vegans.
Vegans presented lower n-3 levels than LO-vegetarians, which can be mainly attributed to their lower percentage of DHA; but the fatty acid profiles did not markedly differ between these groups. In fact, and consistently with other studies in vegetarians [5], PUFAs were near 50% of all fatty acids, and LA (C18:2n6) was the major PUFA independently of subject's classification. Therefore, there was an imbalance in n-6/n-3, except for the subjects who consumed supplements of n-3 fatty acids ( Table 3).
The metabolic changes induced by the elimination of the main sources of n-3 PUFA, fish and their respective products, from the diet should be briefly commented. It has been demonstrated that the n-6 and n-3 pathways are interconnected. The essential fatty acids LA and ALA are precursors of the n-6 and n-3 PUFA families, respectively. These fatty acids compete for the same enzymes ∆6 and ∆5 desaturases and elongases, and the ∆6 desaturase constitutes the rate-limiting enzyme, presenting higher affinity for ALA than LA [23]. However, under the high LA consumption conditions observed in the present study, the n-6 pathway predominates in detriment of the n-3 pathway [24]. Consistently, the observed ALA/LA ratios indicate that ALA levels (n-3) were approximately 100 times lower than those of LA (n-6). Moreover, results of EPA/ALA and DHA/ALA confirm the low intake of both EPA and DHA, typical of vegetarian diets, suggesting that the production of these fatty acids from ALA was very limited. In agreement, the overall calculated efficiency of conversion from ALA to DHA is lower than 1% or even 0.01%, depending on the used models [23,25,26]. Interestingly, consumption of n-3 supplements (by less than 10% of the total volunteers) was associated with an increase of the EPA levels and the EPA/ALA and EPA/AA ratios without a parallel rise in DHA. This can be explained by the poor enzymatic efficiency in the DHA production from EPA and ALA, and by the composition of the n-3 supplements consumed. In this line, Brenna et al. summarized the change in blood n-3 PUFAs after the consumption of supplements containing ALA or EPA [27], and concluded that the intake of these fatty acids cannot increase serum DHA. Therefore, present results reflect an inadequate supplementation practice. We observed that the n-3 supplements used were very diverse, ranging from supplements containing DHA specially designed for vegetarians to EPA+DHA combinations manufactured from fish or mixtures of n-3, n-6, n-9, and n-7. Indeed, when volunteers were classified according to the n-6/n-3 ratio, only one third of the subjects included in the n-6/n-3 < 10 were n-3 supplements users.
The cardiometabolic risk markers (glucose, lipids, insulin, the insulin resistance/sensitivity indexes, leptin and Hcy) studied according to the n-6/n-3 classification, confirm that this vegetarian population exhibits a low cardiometabolic risk, in agreement with other authors [28,29]. Moreover, Hcy, which constitutes an independent factor for cardiovascular disease [30], was lower in the n-6/n-3 < 10 group compared to the others. This result cannot be attributed to vitamin B12 deficiency, since we confirmed that 75% of the participants consumed vitamin B12 supplements, as previously described [3,31], and vitamin B12-supplemented subjects were equally distributed among the n-6/n-3 groups. In fact, an inverse relationship between n-3 and Hcy has been observed in different studies [17,32]. Bertasd et al. [33] observed that self-reported fish intake was negatively associated with Hcy plasma concentrations in a large cohort of Norwegian people, and another study found that n-3 supplement consumption (3,6 g/day) during 3 months, decreased Hcy in treated dyslipemic diabetics in comparison with a placebo [34].
The higher POA in the group n-6/n-3 < 10 may be related to a higher intake of this fatty acid or may also reflect the ∆9 desaturase activity that converts PAL into POA [35]. In this group, higher ALA, EPA and DHA were shown together with lower n-6 PUFA LA, as expected. Concerning the results of the different fatty acid ratios, lower serum EPA/AA has been related with increased risk of coronary heart disease in a Japanese cohort of adults [36]. In another study, a ratio of EPA/AA lower than 0.40 was associated with adverse cardiac events in coronary patients [37]; and although our population was healthy, the observed results are much lower than this cut-off value. Other ratios such as the DHA/AA appear to be less valuable as markers of cardiovascular risk [38]. Therefore, we conclude that the vegetarians of the present study were at low cardiometabolic risk but most of them have an imbalance n-6/n-3, and low levels of EPA and particularly DHA.
As the group with n-6/n-3 < 10 included consumers and non-consumers of n-3 supplements, the influence of any specific food item intake on serum long-chain n-3 fatty acids was evaluated. We observed similar consumption patterns of nuts and oils in the three n-6/n-3 groups while among the seeds, results suggest an improvement of the n-6/n-3 ratio associated with higher frequency of consumption of flax-seeds. In fact, these seeds are rich in ALA, and consistently, results showed higher serum ALA concentration with higher intake of flax-seeds ( Figure 1). However, there was no rise in serum DHA as flax-seed intake increased. This is according to studies reporting that the intake of ALA-rich foods is insufficient to increase serum DHA [23,26,27]. Moreover, this coincides with the results discussed above regarding n-3 supplementation effects on DHA. Altogether, results suggested that the n-3 supplementation increased EPA but not DHA, and the intake of specific foods was insufficient to increase DHA serum levels.
There are many aspects of the relationship between plant-based diets and health that remain unknown. A recent meta-analysis on the effect of vegetarian diets on major cardiovascular outcomes, found very low evidence of associations of these diets with reduction of coronary heart disease incidence (atherosclerosis and/or myocardial infarction) and mortality, but no association with overall cardiovascular disease mortality and stroke mortality [39]. The specific effects of n-3 fatty acids on heart and circulatory disease was evaluated in a Cochrane systematic review. The review provides moderate-and high-quality evidence that increasing the intake of long-chain n-3 fatty acids, mainly from supplements, does not benefit heart health or reduce the risk of stroke or mortality from any cause, and found low-quality evidence suggesting that increasing plant-based ALA may be slightly protective for some heart and circulatory diseases [40]. However, although studies on vegetarians were not excluded in this review, the available information was very limited.
This opens new research lines, as long-chain n-3 fatty acids are relevant in lipid metabolism, lipogenesis and β-oxidation and may be implicated in reducing cardiovascular events. Therefore, further studies are needed to establish if preformed DHA supplements should be widely recommended to vegetarians. At present, ALA and LA are the two fatty acids accepted as essential, but perhaps this should be re-evaluated for the vegetarian and whole population. On one hand, more research should be done on the design of safe and bioavailable DHA sources of plant origin, i.e., marine algae; on the other hand, measurements of cardiometabolic markers should be complemented with functional markers, such as blood hemodynamic, visual and brain function tests, to be used in controlled and well-designed interventions in vegetarian and non-vegetarian volunteers.
This study presents several limitations. The volunteers were LO-vegetarians and vegans but no control group of meat and fish eaters was included. Dietary assessment was carried out by an FFQ and details of quantities of food or nutrients ingested are not known. Apart from that, this is the first study that evaluates lifestyle habits, body composition, cardiometabolic risk markers, and serum fatty acid profile of Spanish vegetarians.

Conclusions
Fatty acid profile of Spanish LO-vegetarians and vegans is characterized by high levels of LA and OA, as olive oil is widely consumed in this Mediterranean population. However, levels of long-chain n-3 fatty acids are very low, and consumption of n-3 supplements is associated with an increase in EPA but not DHA. Similarly, frequent consumption of ALA-rich seeds is reflected in serum ALA levels, but further conversion to EPA and DHA is undetected.