With change in the western human diet over the past 100 years, consumption of food rich in n-6 PUFA has increased(Reference Kris-Etherton, Taylor and Yu-Poth1). n-6 PUFA are needed for many physiological functions of the human system and are well known for their protective effects against CVD. For these reasons, the American Heart Association Nutrition Subcommittee has recently recommended a consumption of at least 5–10 % of energy from n-6 PUFA(Reference Harris, Mozaffarian and Rimm2). However, some researchers recommend to reduce dietary n-6 PUFA intake in order to prevent adverse effects on health, in particular pro-inflammatory response(Reference Simopoulos3–Reference Ailhaud5). In parallel, some national recommendations for daily n-6 PUFA intake are already based on low figures (i.e. 4 % of total energy in France), considering the risk of inflammation or obesity linked with significant n-6 PUFA consumption as non-negligible. Fixing an upper limit to n-6 PUFA consumption for healthy populations is a recurrent issue in the international scientific agenda for lipids intake guidelines. To get an in-depth analysis of the benefits/risks balance of n-6 PUFA intake, the present work reviews studies on the link between n-6 PUFA and CVD risks factors: dyslipidaemia, hypertension, thrombosis susceptibility, oxidisability of lipoproteins, obesity and a pro-inflammatory response. Based on this analysis, recommendation and comments regarding minimal v. optimal n-6 PUFA intake, and the need for an upper limit, are discussed.
Dietary n-6 PUFA
A carbon chain that contains two or more cis double bonds with the first double bond located between the sixth and seventh carbon atom from the methyl end of the fatty acid (n-6 position) characterises n-6 PUFA. The main dietary n-6 PUFA is linoleic acid (LA; 18: 2n-6), which can be found in vegetable oils such as soyabean, safflower, maize and rapeseed oils. LA cannot be synthesised by human subjects and other mammals and, as such, is provided by dietary intake only(Reference Holman6, Reference Williard, Nwankwo and Kaduce7). The average LA intake in USA is 14·8 g/d (6·7 % of energy)(Reference Moshfegh, Goldman and Cleveland8). In France, according to the SUpplementation en VItamines et Mineraux AntioXydants study, LA intake is 10·6 g/d in men and 8·1 g/d in women, representing 4·2 % of energy intake(Reference Astorg, Bertrais and Laporte9). These values reflect well the variability of daily LA consumption across countries(Reference Poudel-Tandukar, Nanri and Matsushita10–Reference Baylin, Kim and Donovan-Palmer12).
Another n-6 PUFA provided by the diet, but in a lower amount, is the arachidonic acid (AA; 20 : 4n-6) which can be found in meats, poultry and eggs. Dietary AA intake accounts for an average of 0·08 % energy in France with a daily consumption of 0·22 g/d in men and 0·16 g/d in women(Reference Astorg, Bertrais and Laporte13), similar to the world intakes(Reference Calder14–Reference Williams, Baylin and Campos16). AA can also be synthesised by the conversion of LA after successive desaturation and elongation reactions occurring in the endoplasmic reticulum of the cell(Reference Sprecher, Luthria and Mohammed17). The δ-6-desaturase enzyme, which catalyses the conversion of LA to γ-linolenic acid (18 : 3n-6), seems to be the rate-limiting step in the n-6 PUFA metabolism(Reference Nakamura and Nara18). Results from human subjects and animal studies report that the rate of this initial conversion is low(Reference Cunnane, Keeling and Thompson19–Reference Emken, Adlof and Gulley21). Then, γ-linolenic acid is in turn elongated to dihomo-γ-linolenic acid (20 : 3n-6). In a third step, dihomo-γ-linolenic acid is desaturated to AA by δ-5 desaturase. It is worth mentioning that the overall conversion of LA to AA is extremely low, certainly below 0·5 %(Reference Hussein, Ah-Sing and Wilkinson22). This might explain why variations in dietary n-6 intake, including LA, have little effect on AA levels in serum cholesterol esters, erythrocyte and platelet membranes(Reference Sarkkinen, Agren and Ahola23).
n-6 PUFA and CVD risk factors
n-6 PUFA and blood lipids
Abnormal blood lipid levels such as elevated LDL-cholesterol (LDL-C) are major risk factors for atherosclerosis and CVD(24, Reference Briel, Ferreira-Gonzalez and You25). These risks can be reduced by dietary intervention and, in particular, by change in fat composition of the diet. Indeed, a decrease in dietary SFA can induce a significant lowering of plasma LDL-C levels(Reference Katan, Zock and Mensink26). Diets rich in PUFA are well known for their hypocholesterolaemic action(Reference Nichaman, Sweeley and Olson27, Reference Shepherd, Packard and Grundy28). When the dietary proportion of SFA remains constant and n-6 PUFA replace carbohydrates, a decrease in LDL-C plasma levels is observed(Reference Mensink, Zock and Katan29). A meta-analysis of sixty controlled trials reported that the replacement of carbohydrates with PUFA (largely n-6) was predictive of the largest change in the total cholesterol:HDL-cholesterol (HDL-C) ratio, and in LDL-C concentrations, compared with other types of fatty acids(Reference Mensink, Zock and Kester30). Replacement of 1 % of energy of carbohydrates by saturated fats increases LDL-C serum level by approximately 0·03 mmol/l, whereas replacement by n-6 PUFA decreases this level by 0·02 mmol/l.
Replacing dietary SFA by PUFA, which are mainly n-6 PUFA, is also efficient in decreasing plasma concentration of cholesterol. In a meta-analysis of seventy-two metabolic ward studies of solid food diets in healthy volunteers, Clarke et al. (Reference Clarke, Frost and Collins31) observed that replacement of 5 % energy as SFA by PUFA led to a − 0·39 mmol/l change in total blood cholesterol. In the work of Hodson et al. (Reference Hodson, Skeaff and Chisholm32), replacing 6·4 % energy as SFA by n-6 PUFA while keeping total fat content at 30–33 % of energy led to a 22 % decrease in plasma LDL-C ( − 0·63 mmol/l) and a 14 % decrease in HDL-C. The decrease in HDL-C, although not systematically observed in all n-6 PUFA interventional studies, could be challenged because a low concentration of HDL-C has been associated with a higher risk for CHD(Reference Franceschini33). However, the total cholesterol:HDL-C ratio is considered as a better predictor of CVD than HDL-C alone(Reference Yusuf, Hawken and Ounpuu34). Analysis of the Framingham Heart Study's data by Siguel(Reference Siguel35) reported that a total cholesterol:HDL-C ratio decreased when the percentage of plasma PUFA increased, either in response to a diet enriched in PUFA or across different subjects. To conclude, replacing SFA by n-6 PUFA (or eating a diet enriched with n-6 PUFA) leads to a substantial reduction in total and LDL-C cholesterol, as well as a reduction of the total cholesterol:HDL-C ratio, and may reduce the risk of CVD(Reference Jakobsen, O'Reilly and Heitmann36, Reference Hodson, Skeaff and Chisholm37).
n-6 PUFA and blood pressure
In a cross-sectional study by Salonen et al. (Reference Salonen, Salonen and Ihanainen38), LA intake (average of 10 g/d) assessed by a 4-d dietary recall by household measures in 722 men was not correlated with the mean resting blood pressure. In contrast, Oster et al. (Reference Oster, Arab and Schellenberg39) found a strongly significant negative correlation between LA content of adipose tissue and systolic as well as diastolic blood pressure (Pearson's r − 0·16, P < 0·001 and r − 0·12, P < 0·001, respectively) in a cohort of 650 healthy men. In an other observational study in a large population of 4033 healthy men, a 2-sd increase in plasma levels of LA was associated with a 1·9 (95 % CI 1·0, 2·8) mmHg decrease in systolic blood pressure(Reference Grimsgaard, Bonaa and Jacobsen40). This is in line with the results obtained in control subjects of the multiple risk factor intervention trial, where plasma levels of LA were inversely associated with systolic and diastolic blood pressures ( − 3·02 (95 % CI − 5·26, − 0·77) and − 1·62 (95 % CI − 2·83, − 0·41), respectively)(Reference Simon, Fong and Bernert41). Results of several interventional trials reported by Iacono et al. (Reference Iacono, Dougherty and Puska42) showed that consumption of a diet with PUFA:SFA ratio at about 1·0 led to a significant decrease in blood pressure in normotensive or mildly hypertensive healthy subjects, compared with an usual diet, regardless of the level of fat energy of the intervention diets (25 or 44 % of energy). As highlighted by a recent review of cross-sectional studies, an increase in dietary n-6 PUFA intake is often associated with a decrease in blood pressure, which is in favour of a reduced risk of CVD(Reference Hall43).
n-6 PUFA and thrombus susceptibility
Three decades ago, dietary LA was administered as a natural precursor of PGE1 in order to reduce platelet aggregation, a well-known risk factor for atherogenesis and thrombogenesis. Hornstra et al. (Reference Hornstra, Chait and Karvonen44) reported that a polyunsaturated-rich diet (PUFA:SFA ratio = 1·60) was associated with a significant decrease in platelet aggregation compared with a saturated-rich diet (PUFA:SFA ratio = 0·25) in men. Since then, other interventional studies have investigated the effect of LA on various haemostatic parameters but the results are not consistent(Reference Knapp45). In the recent randomised cross-over controlled trial of Thijssen et al. (Reference Thijssen, Hornstra and Mensink46), forty-five healthy subjects consumed three different diets for 5 weeks each. Diets contained 38 % energy as fat and differed by 7 % of energy from stearic acid, oleic acid or LA. Consumption of LA relative to stearic acid was related to an increase in ex vivo platelet aggregation time in men (P < 0·036) suggesting an antithrombogenic effect of LA. However, the three diets had no effect on in vitro whole-blood platelet aggregation variables, on factor VIIam activity and on fibrinolytic activity. Overall, the results from human studies are not conclusive, and further investigation is needed to clarify the role of n-6 PUFA in susceptibility to thrombus.
n-6 PUFA and oxidative stress
PUFA are particularly vulnerable substrates to oxidative stress because reactive oxygen species can easily remove hydrogen atom from their numerous double bonds and generate toxic peroxide species(Reference Halliwell and Chirico47). This lipid peroxidation leading to pro-inflammatory oxidised LDL and HDL is highly suspected of contributing to atherosclerosis pathogenesis(Reference Steinberg, Parthasarathy and Carew48, Reference Navab, Ananthramaiah and Reddy49). Interventional studies investigating the link between dietary PUFA intake and atherogenesis produced mixed results. Several works have shown that dietary supplementation rich in n-6 PUFA increases the extent of LDL oxidation in vitro compared with a diet enriched in MUFA(Reference Abbey, Belling and Noakes50–Reference Bonanome, Pagnan and Biffanti52). In contrast, markers related to LDL-C oxidation in vitro or LDL levels of malondialdehyde were not correlated with the n-6 PUFA intake in a group of healthy volunteers(Reference Kleemola, Freese and Jauhiainen53). Furthermore, a controlled double-blind 2 × 2-factorial, 8-week intervention in a cohort of healthy men showed that fish oil consumption combined with a high LA intake (21 g/d) did not raise the plasma level of oxidised LDL compared with the same fish oil consumption but combined with a low level of LA(Reference Damsgaard, Frokiaer and Andersen54).
n-6 PUFA and inflammation
n-6 PUFA have long been considered as pro-inflammatory molecules because they are the main precursors of eicosanoids, a family of mediator molecules which are involved in immune and inflammatory response such as Prostaglandin E2, thromboxane A2 and leukotriene B4(Reference Calder and Grimble55). However, both the level and the nature of the prostanoids synthesised from AA change dramatically according to the amplitude of the inflammatory response and the course of this response(Reference Tilley, Coffman and Koller56–Reference Serhan58). It is also argued that higher dietary intake of n-6 PUFA may lead to a competition between n-6 and n-3 metabolism resulting in a reduced production of anti-inflammatory molecules from n-3 PUFA(Reference Simopoulos3). In human subjects, higher intakes of n-6 fatty acids do not appear to be associated with elevated levels of inflammatory markers. A study in a large US adult population reported that n-6 fatty acids did not inhibit the anti-inflammatory effects of n-3 fatty acids. In addition, combination of both types of fatty acids (higher percentile of EPA+DHA intake of 1·12 % of energy among men and 0·471 % among women; α-linolenic acid ranging from 0·46 to 0·52 % and LA ranging from 4·3 to 5·4 %) was associated with the lowest levels of inflammation, assessed by C-reactive protein, IL 6 and soluble TNF receptors 1 and 2 plasma levels(Reference Pischon, Hankinson and Hotamisligil59). In the In CHIANTI (Invecchiare in Chianti, ageing in the Chianti area) study, in the context of a mean PUFA intake of 7 g/d, higher plasma levels of n-6 PUFA (mainly AA) and n-3 PUFA (mainly DHA) were independently associated with lower levels of serum pro-inflammatory markers(Reference Ferrucci, Cherubini and Bandinelli60). Nevertheless, Thies et al. (Reference Thies, Miles and Nebe-von-Caron61) reported that a dietary supplementation with moderate amounts of long-chain n-6 or n-3 PUFA did not significantly affect inflammatory cell numbers or neutrophil and monocyte responses.
n-6 PUFA and obesity
Obesity is a cardiovascular risk factor which is linked with substantial increases in incidence of type 2 diabetes mellitus, systemic hypertension and dyslipidaemia, which are all known risk factors for CVD(Reference Lewis, McTigue and Burke62). Normal weight obesity (normal BMI and high body fat content) is also associated with a high prevalence of cardiometabolic abnormalities and CVD risk factors(Reference Romero-Corral, Somers and Sierra-Johnson63). Adipose tissue obesity is thought to depend on both hypertrophy of preexisting adipocytes and hyperplasia due to adipogenesis(Reference DiGirolamo, Fine and Tagra64–Reference Tchoukalova, Sarr and Jensen66). It has been proposed that n-6 PUFA may be involved in the differentiation of preadipose cells to adipocytes(Reference Cleary, Phillips and Morton67, Reference Massiera, Saint-Marc and Seydoux68). To date, no firm conclusion can be drawn on AA role in the differentiation of preadipose cells from available in vitro studies(Reference Gaillard, Negrel and Lagarde69–Reference Serrero, Lepak and Goodrich72), and animal studies investigating the effect of a diet enriched in n-6 PUFA on adipose tissue have produced conflicting results(Reference Cleary, Phillips and Morton67, Reference Massiera, Saint-Marc and Seydoux68, Reference Matsuo, Takeuchi and Suzuki73, Reference Okuno, Kajiwara and Imai74). There are few interventional studies which have investigated the relationship between a diet enriched in n-6 PUFA and adiposity as well as body weight. A recent work on fatty acids composition of adipose tissues in extremely obese patients (BMI>40 kg/m2) has found significant negative associations between n-6 PUFA and metabolic risk factors such as cholesterol and HDL-C(Reference Hernandez-Morante, Larque and Lujan75). These data suggest that n-6 PUFA from different sources, i.e. plasma as well as subcutaneous and visceral adipose tissues, may protect extreme obese patients against metabolic alterations. Overall, the role of n-6 PUFA in adipogenesis and obesity is still unclear as no firm conclusion can be drawn from available epidemiological or experimental data.
n-6 PUFA and CVD epidemiological studies
Prospective cohort studies on dietary PUFA intake and prevalence of CHD events are listed in Table 1. CHD events taken into account vary between studies, but myocardial infarction, sudden cardiac death and acute coronary syndrome were generally cited. One study found a correlation between dietary PUFA level of the initial examination and coronary deaths observed 19 years later(Reference Shekelle, Shryock and Paul76). Of eleven studies, five examined LA or total n-6 PUFA specifically: a significant negative association between n-6 PUFA and risk of CHD events or mortality was observed in three of them. In the nurse's health study, Hu et al. (Reference Hu, Stampfer and Manson77) found that PUFA intake (no details about n-6 or LA figures) was inversely associated with CHD risk, with the highest quintile corresponding to a daily intake of 6·4 % of energy. The authors concluded that replacing 5 % of energy from SFA with energy from unhydrogenated MUFA and PUFA would reduce CHD risk by 42 %, and would be more effective in preventing CHD than reducing overall fat intake. When extending the follow-up time from 14 to 20 years, Oh et al. (Reference Oh, Hu and Manson78) reported an inverse association between the highest quintile of PUFA intake (7·4 % of energy, no details about n-6 or LA figures) and the risk of CHD, with a stronger association in women under 65 years or overweight. A spline regression analysis showed a linear relationship between the dietary LA intake and the relative risk of CHD, with the highest proportion of LA intake (7·0 % of energy) corresponding to the lowest risk. In the Kuopio IHD risk factor study, men with the highest daily intake of LA (12·9 g/d) were up to 61 % less likely to die of CVD than their counterparts whose intake was in the lower third (6·5 g/d). Dietary PUFA intake was also positively associated with a lower risk of CVD mortality. The associations of serum-esterified fatty acids proportions with CVD mortality were equivalent to those of dietary fatty acids: higher thirds of esterified PUFA proportions (44 % of serum fatty acids), esterified LA (32 % of serum fatty acids) and esterified n-6 fatty acids (38 % of serum fatty acids) were associated with lower CVD mortality compared with lower third proportions (37, 24 and 31 %, respectively). Overall, men with the highest levels of serum LA were up to three times less likely to die of CVD(Reference Laaksonen, Nyyssonen and Niskanen79).
M, male; F, female; LA, linoleic acid; RR, relative risk.
A meta-analysis of 25 case–control studies was carried out by Harris et al. (Reference Harris, Poston and Haddock80) in 2007 in order to assess the association between n-3 and n-6 tissue content and CHD events. When all studies were combined, LA values were significantly lower in cases relative to controls (effect size: Hedges g = − 0·28, P = 0·02, 95 % CI − 0·04, − 0·53) and were inversely associated with non-fatal coronary events (Hedges g = − 0·21, P < 0·01, 95 % CI − 0·06, − 0·36). When studies were stratified by phospholipid- v. TAG-rich tissues, the AA content of adipose tissue was higher in cases relative to control (Hedges g = 0·47, P = 0·01, 95 % CI 0·83, 0·1) but, overall, the tissue AA content was not associated with CHD events. Similar results were found in the study by Block et al. (Reference Block, Harris and Reid81) investigating the link between acute coronary syndrome and the fatty acid content of whole-blood cell membranes. The authors found that a 1-sd decrease in LA was associated with more than three times the odds for having a acute coronary syndrome (OR 3·23, 95 % CI 2·63, 4·17). Results with AA were more complex to interpret because both very low and very high levels were associated with increased risk of acute coronary syndrome.
Randomised controlled trials
Several small randomised trials have investigated the impact of dietary PUFA on CVD. Most of them were based on the replacement of SFA, and not carbohydrates or protein, in populations of patients presented with CHD. Dietary intervention varied between studies with a PUFA intake ranging from 5·7 % of total energy in the study of Frantz et al. (Reference Frantz, Dawson and Ashman82) to 20·6 % in the study of Leren(Reference Leren83), and included reduction of total fat in some cases (see Table 2). A first major result is that all these dietary interventions resulted in substantial reduction of serum cholesterol (from 11 % in Woodhill et al. (Reference Woodhill, Palmer and Leelarthaepin84) up to 25 % in Rose et al. (Reference Rose, Thomson and Williams85)). The effect of dietary PUFA intake on CVD mortality and incidents differed between studies. Of nine studies listed in Table 2, four showed a decrease in CHD events in both men and women(Reference Frantz, Dawson and Ashman82, Reference Leren83, Reference Turpeinen, Karvonen and Pekkarinen86, Reference Miettinen, Turpeinen and Karvonen87). Miettinen et al. (Reference Miettinen, Turpeinen and Karvonen87) found a lower CHD events' incidence in the test group (25 %) compared with the control (39·4 %) after a diet rich in PUFA (36·8 % of total fat). When data are stratified by age, it seems that this type of diet is more efficient in preventing CHD in younger subjects than in older ones(Reference Leren83, Reference Dayton, Pearce and Goldman88).
NA, not applicable; M, male; LA, linoleic acid; F, female.
Many of the intervention trials that have examined the impact of PUFA on health outcomes have done so by replacing SFA in the diet with PUFA. Furthermore, in many of the epidemiological studies, it is likely that a PUFA-rich diet may be associated with a lower SFA intake. The respective PUFA and SFA weights in terms of CVD benefits/risks balance is therefore a recurrent topic. Interestingly, a recent meta-analysis of twenty-one prospective studies has been concluded on an absence of significant evidence for an association of dietary saturated fat with an increased risk of CHD or CVD(Reference Siri-Tarino, Sun and Hu89). In parallel, the cholesterol-lowering effect of n-6 PUFA and, especially, LA is well established from human trials, as well as observational studies generally suggest a benefit of n-6 PUFA intake on CHD risk. In the nurse health study, the inverse association found between polyunsaturated fat and relative risk of CHD (relative risk 0·62; 95 % CI; for each increase of 5 % of energy) appears stronger compared with the positive association linked to saturated fat (relative risk 1·17; 95 % CI; for each increase of 5 % of energy). Overall, data converge to an SFA-independent benefit of n-6 PUFA intake on CVD risk.
International recommendations for n-6 PUFA intake
Despite large variations across the world, increasing quantities and changing qualities of fat consumed, in particular SFA and trans-fatty acids, give rise to serious concerns in the maintenance of good public health. There is a consensus for an upper limit of saturated fat intake at 10 % of total energy intake, but this is not the case for n-6 PUFA.
Recommendations on PUFA intake in healthy adults are listed in Table 3. PUFA intake values range from 3·6(Reference Kris-Etherton, Innis and Ammerican90) to 12 % of total energy(Reference Trumbo, Schlicker and Yates91). The same disparity holds true for dietary n-6 PUFA amounts which vary from 3(Reference Kris-Etherton, Innis and Ammerican90) to 10 % of total energy(92, Reference Sasaki93). Beyond these recommendations, controversial issues exist on the relevance of setting an upper limit for n-6 consumption and recommending an optimal n-6:n-3 ratio(Reference Simopoulos3–Reference Ailhaud5).
FA, fatty acid; LA, linoleic acid.
Currently, there is no mention of an upper n-6 PUFA value in the Eurodiet core report(92, 94, 95). The 2005 advisory committee on dietary guidelines for America documents the upper limit of LA intake assimilated to n-6 PUFA intake on three lines of evidence: the actual dietary intake of North American populations, absence of epidemiological data on health consequences of a greater intake, and finally the pro-oxidant state produced by high intakes of LA that could promote CHD and cancer(95). In contrast, in the nutrient reference values report for Australia and New Zealand, no upper level of intake for either linoleic or α-linolenic acids is set because ‘there is no known level at which adverse effects may occur’(92, 95). These different positions reflect the current worldwide debate on the relevance of an upper limit in dietary n-6 PUFA intake and highlight the need for further in vivo investigations. The n-6:n-3 ratio issue has been debated in detail by Stanley et al. (Reference Stanley, Elsom and Calder96) and Harris(Reference Harris97), which concludes that using this ratio is not relevant when setting up recommendations.
n-6 PUFA are critical to many physiological functions of the organism, and their derivatives are involved in complex molecular pathways. Dietary n-6 PUFA intake from 5 to 20 % of the energy intake lowers LDL-C blood levels, and this may explain why n-6 PUFA (in particular LA) consumption is associated with a decreased risk of CHD(Reference Jakobsen, O'Reilly and Heitmann36, Reference Katan98). No adverse effect of n-6 PUFA intake on blood pressure, inflammatory markers or haemostatic parameters has been observed, even with intake up to 15 % of total energy. In addition, there is no evidence for a causal link between n-6 PUFA intake and obesity in human subjects. From all available investigations, the most effective replacement of saturated fat in regards to CHD outcome are PUFA, especially LA(Reference Harris, Mozaffarian and Rimm2, Reference Tunstall-Pedoe99). The body of data supports the recommendation for n-6 PUFA intake above 5 %, and ideally about 10 % of total energy. The cardiovascular benefit could be even stronger when combined with a recommendation on reducing SFA intake. Finally, whatever the objective might be (i.e. to limit the risk of developing inflammatory or obesity diseases), recommending n-6 PUFA consumption below the current lowest values (i.e. 4 % of total energy in France) is not supported.
S. C. has received honoraria from Lesieur, E. B. has received honoraria from Danone and Unilever, and D. T. has received honoraria from Unilever.