OUP user menu

Fatty acids and cardiovascular disease

Jean-Michel Lecerf
DOI: http://dx.doi.org/10.1111/j.1753-4887.2009.00194.x 273-283 First published online: 1 May 2009


Fatty acids have been classified into “good” or “bad” groups according to their degree of unsaturation or whether they are “animal fat” or “vegetable fat”. Today, it appears that the effects of fatty acids are complex and vary greatly according to the dose and the nature of the molecule. Monounsaturated fatty acids are still considered as having a “neutral” status, but any benefits may be related to the chemical environment of the source food or the associated overall food pattern. Controversy surrounds omega-6 polyunsaturated fatty acids, because even though they lower LDL cholesterol levels, excessive intakes do not appear to be correlated with cardiovascular benefit. The omega-3 fatty acids are known to exert cardiovascular protective effects. Dairy fat and its cardiovascular impact are being evaluated. This review examines the existing literature on the relationships between the different fatty acids and cardiovascular disease.

  • cardiovascular risk
  • dairy fat
  • saturated fatty acids
  • unsaturated fatty acids


Cardiovascular disease (CVD) is the leading cause of mortality in the Western world. Even though CVD encompasses a broad range of pathologies with varied etiologies, atherosclerosis is currently the main pathogenic process leading to both chronic conditions (coronary failure, etc.) and acute phenomena (myocardial infarction, sudden death, etc.). This widespread disease is most often polyarterial and multifactorial. Beyond the related morbidity and the debilitating nature of the final damage (heart failure), sudden death accounts for 65–75% of all mortality in CVD. Stroke is also a major manifestation of CVD, whether thrombotic or hemorrhagic. Typically, arterial hypertension is the main risk factor for stroke and, setting aside the negative effect of salt and the positive effect of fruits and vegetables, much of the data in this field concerns fats. Nutrition plays a major, determinant role in atherosclerosis and its complications (thrombosis, rhythm disorders, etc.). Fatty acids are involved at all levels in these processes.


Atherosclerosis is an inflammatory, degenerative affection of the arterial wall of the large and medium arteries and is characterized by 1) the accumulation of lipids (both saturated and unsaturated fatty acids) and cholesterol in the subendothelial space, 2) infiltration of inflammatory cells and fibrous elements into the intima, and 3) the proliferation of smooth muscle cells in the media.

The deposition of cholesterol and fatty acids has led to the physiopathology of atherosclerosis being schematized as being simply related to the increase in lipoproteins (which transport fatty acids and other lipids). Underpinning this conclusion is epidemiological data identifying a number of risk factors: atherogenic dyslipidemia, elevated cholesterol, hypertension, diabetes, smoking, and so on. The underlying causes, however, are often hidden: genetic factors, abdominal obesity (a risk factor for hypertension, diabetes, and dyslipidemia), dietary imbalance (lack of fiber or antioxidants, imbalance in the omega-6/omega-3 ratio, etc.). Furthermore, the detailed mechanisms involve other processes in addition to the oxidation of LDL particles (required for accumulation of the latter in macrophages), such as inflammation, endothelial dysfunction, platelet aggregation and insulin-resistance, which are all impact sites at which fatty acids (but also other nutritional factors) intervene.1

Sudden death is a complication of infarction and thus of thrombosis, generally via rhythm disorders, although the phenomenon can occur in the absence of a thrombotic accident.


Classification of fatty acids only from a biochemical point of view is no longer relevant.

Saturated fatty acids (SFAs) must be considered individually according to their chain length, because lauric (C12:0), myristic (C14:0), palmitic (C16:0), and stearic (C18:0) acids all have different effects. Likewise, within the group of unsaturated fatty acids, oleic acid, monounsaturated (C18:1n-9) fatty acids, and the omega-6 and omega-3 polyunsaturated fatty acids also have to be examined individually. Then, within the omega-6 and omega-3 fatty acid family, one must draw a distinction between the precursors (linoleic and alpha-linolenic acids, respectively) and the very long chains (arachidonic acid and eicosapentaenoic and docosahexaenoic acids, respectively).

Dairy fat contains 60% SFAs, including the specific myristic acid primarily esterified to glycerol in the sn-2 position but also SFAs with an odd number of carbons; 30% monounsaturated fatty acids, 2–3% linoleic acid, 1% alpha-linolenic acid, some natural trans fatty acids, such as vaccenic acid, and some conjugated linoleic acids, such as rumenic acid. Overall, modern chromatography is capable of identifying up to 406 different fatty acids in dairy fat.


Trans fatty acids and conjugated linoleic acids will not be discussed here.

Saturated fatty acids

Epidemiological studies.  Observational studies started with the Seven Countries Study, which revealed the harmful role of very excessive consumption of SFAs and a parallel elevation of plasma cholesterol levels.2 The Japanese NI HON SAN migration study3 and the Israel Heart Study4 looked at geographical origin and showed a coronary morbidity-mortality gradient associated with an increase in the percentage of SFAs in the calorie intake. A proportion below 10% was associated with a low risk.

The Ireland Boston Diet Heart Study5 compared an Irish population with American migrants of Irish descent and showed that the consumption of SFAs was associated with higher cardiovascular risk when the polyunsaturated fatty acid intake was low, suggesting an interaction between the two types of fatty acids. The LRC 12-year follow-up study6 revealed an effect of SFAs that was independent of plasma cholesterol. A prospective study of American nurses only showed a weak correlation between SFAs and the risk of coronary disease,7 which agrees with the fact that seven other, smaller studies had also not demonstrated a link between these two variables.7

Last and most important, after 25 years of follow-up in the Seven Countries Study, it appears that plasma cholesterol is not associated with an increase in coronary mortality in the Mediterranean countries or in Japan, which is doubtless due to other associated, protective nutritional factors.8

Intervention studies have also been conducted but most have involved a decrease in the SFA intake and a simultaneous increase in the unsaturated fatty acid intake. These studies have generally (but not always) seen a primary reduction in the incidence of coronary disease in hypercholesterolemic subjects. The most spectacular study was the Oslo Diet Heart Trial, carried out by Hjermann et al.9 Although the initial SFA intakes in subjects having confirmed hypercholesterolemia were very high, the intervention yielded a simultaneous decrease in SFA intake and an increase in monounsaturated fatty acid intake. It is, however, noteworthy that 20 years after the end of the study in question, a recently published retrospective evaluation shows that the intervention group and the control group did not differ in terms of plasma lipids or smoking status, but they did differ in terms of the intakes of total fats and saturated and polyunsaturated fatty acids.10 The incidence of myocardial infarction in the intervention group was lower after 5 years and 10 years but not after 16 years.11

Impact sites for saturated fatty acids.  Above a certain threshold, SFAs cause LDL cholesterol (LDLC) to increase, especially if the unsaturated fatty acid (linoleic acid) intake is low,12 but they also produce a very clear increase in HDL cholesterol (HDLC).13 The physiopathological significance of this increase is not well established: it may not necessarily be favorable because a decrease in the activity of lecithin cholesterol acyl transferase14 and the formation of “pro-inflammatory” HDL15 have been observed.

The effect of the SFAs, however, cannot be dissociated readily from that of unsaturated fatty acids. Furthermore, the LDLC/HDLC ratio is more favorably influenced by replacing SFAs with unsaturated fatty acids than by reducing SFAs alone.16 Likewise, and despite stable, high, total lipid levels, we have shown that partial (and moderate) replacement lowers LDLC in subjects with moderate hypercholesterolemia.17 The composition and size of the LDL particles must also be taken into consideration. A low-fat and low-saturated-fatty-acids diet results in smaller, denser LDLs18 and leads to an increase in the number of subjects who acquire an LDL B phenotype19 (small, dense LDLs are known to be atherogenic, partly because they are more readily oxidized and penetrate more easily into the subendothelial space).20

Finally, it is important to underline the heterogeneity of the SFAs and their effects on risk factors.21 For example, stearic acid is not a hypercholesterolemic agent.22 Myristic acid is the most hypercholesterolemic but only when intakes are very high; moderate intakes lack a negative effect.23

It is also thought that SFAs are thrombogenic, based on platelet aggregation work (notably that performed by Serge Renaud24 in the Moselle region of eastern France on subjects consuming significant amounts of SFAs), even though the omega-6/omega-3 ratio was not taken into account in these older studies. Here again, there is heterogeneity – stearic acid is supposedly less thrombogenic than palmitic, myristic, and lauric acids. Some studies, however, indicate that SFAs (particularly stearic acid) activate the factor VIIa less and have a slightly more favorable effect on postprandial fibrinolysis, confirming that in this particular (postprandial) situation, stearic acid is less thrombogenic than the unsaturated trans fatty acids contained in partially hydrogenated fats.25

Monounsaturated fatty acids

Epidemiological studies.  Few observational, epidemiological studies (and not a single intervention study) have looked at monounsaturated fatty acids, which have long been considered as neutral and thus devoid of interest. Indeed, most people wrongly consider studies of the Mediterranean diet or of olive oil to be studies of monounsaturated fatty acids. In fact, the Mediterranean diet is much more complex, since it involves a set of complex, nested nutritional factors (notably fruit and vegetable intake). Furthermore, olive oil also includes an unsaponifiable fraction, which is of major significance.

The Seven Countries Study, however, indicated there may be an inverse relationship between monounsaturated fatty acids and coronary mortality. The American prospective studies of healthcare professionals and nurses have shown an estimated reduction of 19% in the risk of coronary disease when the intake in MUFAs was increased by 5% (as a percentage of the total energy intake).26

Impact sites for monounsaturated fatty acids.  A literature review on this question has recently been published.27 It notably established that MUFAs barely lower cholesterol levels, since they decrease LDLC but increase HDLC – especially when compared with an identical, carbohydrate-derived energy intake. The effect of MUFAs is very similar to that of PUFAs, although the latter decrease triglycerides a little more.

Monounsaturated fatty acids have other effects: when compared with PUFAs, MUFAs decrease LDL's susceptibility to oxidation. Other mechanisms may testify to an antiatherogenic effect: an improvement in endothelial function and processes leading to monocyte homing and adhesion and a reduction in inflammation marker levels and platelet aggregation. Part of these effects, however, could be due to minor compounds in the olive oil tested in the experimental protocols. Finally, studies suggest that MUFAs may have a modest antihypertensive effect and could improve insulin sensitivity.

Omega-6 polyunsaturated fatty acids

Epidemiological studies.  In the Seven Countries Study, PUFA intake did not appear to be linked to cardiovascular risk: given that omega-3 intakes are usually low, the fact that no distinction was made between omega-6 and omega-3 fatty acids probably did not influence the statistical analysis.

Many other epidemiological studies on omega-6 PUFAs have been conducted. Some ecological studies have shown a positive correlation between apparent PUFA consumption and cardiovascular risk28 but this type of study has limited value, except when investigating two populations that only differ in terms of their diet. Most observational studies29 have shown an inverse relationship between coronary risk and omega-6 PUFA intake when the latter is expressed as a percentage (with an intake of 6.4% versus 2.9% in the Nurses' Health Study, for instance) but for others only when it is expressed in grams but not in percentages (e.g., the Puerto Rico and Honolulu Heart Program studies). In other studies, the intake expressed in grams is favorable at levels above 6.8 g/day (the NHLBI Health Study).

Other studies have revealed a positive relationship between omega-6 PUFA intake and risk of infarction on the basis of an indirect estimation of the intake via plasma and erythrocyte levels of fatty acids (linoleic acid and dihomogammalinolenic acid) in the Edinburgh Artery Study30 and via the fatty acid content of the adipose tissue (arachidonic acid) in the Jerusalem Acute MI Registry.31 In the latter study, however, the calculated PUFA intake accounted for over 6% of the total energy consumption in 90% of the subjects, which is considerable. The Cholesterol Lowering Atherosclerosis Study reported that an increase in PUFA intakes was associated with new arterial lesions.32 However, genetic factors can express individual susceptibility. Hence, in the Los Angeles Atherosclerosis Study, an increase in arachidonic acid intake increased the thickness of the intima media and doubled C-reactive protein levels in some subjects with genotypic variants of 5-lipoxygenase; in contrast, an increase in omega-3 fatty acid intake decreased the impact of the genotype. Thus, depending on genetic factors, omega-3 fatty acids enhance or inhibit leukotriene-mediated inflammation and lead to atherosclerosis – a good example of a gene-nutrition interaction.33

Intervention studies with omega-6 PUFAs, which are generally quite dated (see review of Kris-Etherton et al.29), have all obtained a 13–16% reduction in total cholesterol and, generally, a reduction in coronary events and recurrences – this was particularly the case in the Finnish Mental Hospital study, with a polyunsaturated to saturated (P/S) ratio of 1.5 versus 0.25 (very low) – but in these studies, the distinction between primary prevention and secondary prevention was not always made. Rose's corn oil study, however, revealed an increase in coronary events, and the Minnesota Coronary Survey (despite an identical drop in cholesterol) did not achieve a favorable change in coronary events, but the follow-up period only lasted a year. Lastly, Strandberg's study (with a significant increase in the P/S ratio) led to a significant (RR = 2.4) increase in cardiac deaths and overall mortality 10 years after the end of the study. Furthermore, in some studies, non-cardiovascular mortality increased (in the Los Angeles Veteran Study, for example, where the P/S ratio changed from 0.5 to 2.1), and overall mortality increased (in the Anti-Coronary Club of New York City, for example).

Hence, the data on omega-6 PUFAs are somewhat contradictory, but these compounds appear to be harmful when the intake level and/or the P/S ratio are too high.

Impact sites for omega-6 polyunsaturated fatty acids.  It is clear that omega-6 PUFAs decrease LDLC in the absence of a significant effect on HDLC (or, at most, a small decrease) and lower postprandial lipemia (known to be potentially thrombogenic). It has been suggested, however, that high intakes34 have a harmful effect, notably on the inflammatory process – they lead to amplification of oxygenase-mediated hydroperoxide production and thus formation of eicosanoids and increased oxidative stress. Omega-6 PUFAs also do the following: 1) stimulate endothelial activation via induction of the κB nuclear transcription factor (leading to the production of monocyte adhesion molecules), 2) induce the production of inflammatory cytokines (IL-1, -6, and -8), and 3) increase C-reactive protein and serum amyloid A levels.35 Hence, these effects result in low plaque stability, which might explain a negative effect on fragile plaques in general and in secondary prevention in particular. These plaque effects may be even more marked when the omega-6/omega-3 ratio is extremely high, as reported notably by Thies et al.36

Omega-3 polyunsaturated fatty acids

Epidemiological studies.  It is important to distinguish between studies on alpha-linolenic acid and those looking at long-chain PUFAs (LC-PUFAs) such as eicosapentanoic acid (EPA) and docosahexanoic acid (DHA).37

Alpha-linolenic acid

Almost all the observational studies have revealed an inverse relationship between alpha-linolenic acid intake and cardiovascular and coronary morbidity/mortality, regardless of whether the design was prospective (the MRFIT study,38 the Health Professional Follow-up Study,39 and the Nurses Study – except for non-fatal myocardial infarction40), case-controlled with tissue markers of alpha-linolenic acid intake (Baylin's Costa Rica study41 and the Edinburgh Artery Study30), or cross-sectional (the National Heart Lung and Blood Institute Family Heart Study).42,43 Only the Zutphen Elderly Study44 showed a negative result, but this work has been criticized in view of the absence of data on the tissue alpha-linolenic acid content and the small difference in intake between the control group and the case group.

Three intervention studies have been performed to date, i.e., the Lyon study45 and two studies by Singh.46,47 Although these studies were positive, interpretations should be tempered by the following facts: 1) the Lyon study looked at a “Franco-Mediterranean” diet rather than alpha-linolenic acid per se, and 2) doubt has been cast on the authenticity of the Singh studies.

Lastly, a gene-nutrition interaction related to a polymorphism in the delta-6 desaturase gene (which turns ALA into EPA) has been studied recently, but the FADS2 deletion in the delta-6 desaturase gene did not modify the relationship between adipose tissue alpha-linolenic acid content and the risk of non-fatal infarction.48

Omega-3 long-chain polyunsaturated fatty acids

There have been many coherent studies of LC-PUFAs,49 based on the estimated intake through fish consumption, or direct analysis of intake, or use of an indirect marker (e.g., tissue fatty acid content).

Dyerberg et al.50 and then Kromann et al.51 initiated these studies at the end of the 1970s, with populations of Greenland Eskimos. Other ecological studies have shown the same cardiovascular protection in high-level consumers of fatty fish and sea mammals.37,49 With the exception of the Euramic Study,52 all of the case-control studies37 measuring tissue fatty acids have shown an inverse correlation between EPA, DPA, and DHA on the one hand and ischemic heart disease on the other. Prospective studies37,49 have mainly focused on fish consumption and all have shown the benefit of eating fish more than once a week and the lack of a significant additional effect of more than two fish meals a week. We know, however, that prospective studies are not decisive, since (even after adjustment) they can incorporate the influence of different lifestyles or other confounding factors.

Several observational case-control and prospective studies37,49 of LC-PUFAs have investigated sudden death (which represents the majority of cardiovascular deaths) and have shown a significant protective effect, with up to an 80% decrease in the risk of sudden death.53,54

There have been four large, intervention, secondary prevention studies. The DART study55 in which fatty fish was consumed twice a week or 340 mg of EPA was consumed per day achieved a 32% drop in deaths due to ischemic heart disease; the daily amount was calculated using the mean intake of fish twice a week in addition to the mean supplement intake. The lack of observed benefit 10 years after the end of the study may be due to the fact that compliance with dietary advice had become mediocre as early as year 2 after the end of the study.56

The GISSI study57,58 involved food supplementation with 880 mg of ethylic esters of EPA-DHA and resulted in a 30% drop in cardiovascular mortality and a 53% fall in the sudden death rate. The study performed in India by Singh et al.46 also achieved spectacular results, but doubt has been cast on the authenticity of this work.59 The second study by Burr et al.60 is the only one in which no benefit was observed, but the subjects had angina pectoris and no previous myocardial infarction.

Impact sites for omega-3 polyunsaturated fatty acids.  Intervention studies with omega-3 LC-PUFAs and the Lyon study have been characterized by an extremely rapid effect (within 2–3 months), an absence of changes in cholesterol levels, and a major effect of LC-PUFAs on sudden death. In contrast, LC-PUFAs may not have any effect on the progression of atherosclerosis (Burr et al.60) in cases of stable angina. This argues in favor of an effect on thrombosis via platelet aggregation and/or endothelial dysfunction,61 a cytokine-mediated anti-inflammatory effect62 (and thus an effect on plaque stability) and/or an anti-arrhythmic effect.63

The anti-arrhythmic effect of the omega-3 fatty acids is very well documented, both in the laboratory and in the clinic. It may occur via QT lengthening,63 an increase in the cardiac frequency variability and a membrane stabilizing effect49 due to incorporation of the LC-PUFAs in the cardiomyocyte phospholipids at the expense of arachidonic acid (omega-6).64 As with plaque stability, the omega-6/omega-3 ratio may well be decisive.36,37

These findings do not, however, rule out an LC-PUFA effect on the metabolism of lipoproteins in general and triglyceride-rich lipoproteins in particular.37 Furthermore, LC-PUFAs may have a positive effect on insulin resistance65 and weight control.66,67


It is generally acknowledged that fats have a role in weight gain (a major factor in dyslipidemia and other metabolic anomalies), especially when the weight gain is abdominal. Likewise, a reduction in fat intake is frequently construed as a recommendation in cardiovascular disease prevention. However, the question is now increasingly subject to debate.68

In the 3-year ERA study, Mozaffarian et al.69 monitored 235 menopausal women with coronary heart disease (≥1 stenosis ≥30%) who were not receiving hormone replacement therapy and who had a low fat intake (25%). By stratifying them into SFA intake quartiles, the author showed that the women with the lowest SFA intakes had the lowest total fat intakes and the highest carbohydrate intakes. The lower the intake of SFAs and of total fats (18%), the greater the change in mean coronary lumen diameter (as a marker of disease progression). This paradoxical effect could be linked to induced lipoprotein modifications. Additionally, in men with a stable body weight, Lefevre et al.70 showed that in comparison with a 37% fat diet, STEP I (28% fat) and STEP II (24% fat) diets led not only to a decrease in LDLC but also to an equivalent decrease in the HDLC and an increase in triglycerides – an atherogenic profile characteristic of metabolic syndrome, with small, dense LDLs. The reduction in LDLC was less intense in heavier subjects with insulin resistance syndrome. Hence, a low-fat diet does not appear to be advisable, since it leads to a decrease in HDLC, prompted by both a high carbohydrate intake and a low saturated fatty acid intake.71 More recently still, Krauss et al.72 found in moderate overweight men (BMI 26–35) that a low-carbohydrate, low-saturated-fatty-acid diet (which nevertheless had a high lipid content) induced a greater reduction in triglyceride and apo B levels and a greater increase in HDL levels and LDL size than a high-carbohydrate, low-fat diet. A low-carbohydrate, high-fat and high-saturated-fatty-acid diet, however, led to the same lipidemic profile as a low-carbohydrate, low-saturated-fatty-acid diet (except for the LDLC levels, due to an increase in the number of large LDLs). Hence, maintenance of a high fat intake and, potentially, a high saturated fatty acid intake is not associated with a harmful lipid profile, as long as carbohydrate intake is low. In contrast, in the second part of the trial, weight loss was achieved through a reduction in energy intake: in this case, the most favorable lipid profile (except in terms of LDL size) is obtained with a low-fat diet.

In the Women's Health Initiative study,73 during which 48,835 menopausal women between the ages of 50 and 79 years received nutritional advice in order to reduce their fat intake to 29% instead of 38%, with reductions in SFA (9.5%), MUFA (10.8%), and PUFA levels (6.1%) over 8 years, no benefit was observed in terms of coronary, cardiovascular, and cerebrovascular disease, except in subjects who had been able to further reduce their SFA intakes (<6.1%).74 In contrast, the women with a history of cardiovascular events suffered an aggravation of their CVD risk, which argues against an overly intense reduction in fat in the context of underlying coronary disease. It should be noted, however, that the absence of benefit for the participants as a whole was associated with an absence of weight loss. Thus, reducing fat intake is not beneficial in the absence of weight loss and dietary advice (as in the Women's Health Initiative study), such as increasing fish consumption and reducing salt intake.

The Women's Healthy Lifestyle Project75,76 enrolled 535 premenopausal women of normal weight into two randomized groups, with a view to preventing post-menopausal weight gain. A reduction in the fat and SFA intake and an increase in physical activity resulted in lower increases in LDLC, triglycerides, and glycemia, as well as a decrease in body weight and waistline (compared with the control group) and slower progression of atherosclerosis over 5 years. Thus, when weight loss is accomplished through physical exercise, a reduction in fat intake results in a favorable outcome.


Epidemiological data

Africa's Maasai semi-nomadic cattle herders (who display intense physical activity, low cholesterolemia, and very low cardiovascular mortality) are heavy consumers of dairy products and have very high SFA intakes.77,78 The Swedish research team of Ravnskov et al.79 also pointed out this contradiction of the standard dogma through a broad range of epidemiological studies. It has been suggested that, in addition to the complexity of the confounding factors related to lifestyle and nutritional interactions, genetic factors may be involved. For instance, the apo CIII polymorphism can modulate the lipid response to SFAs: when compared with a high-SFA diet, a low-SFA diet is only correlated with a favorable lipid profile in subjects carrying the 455T-625T polymorphism in the apo CIII promoter.80 These confounding factors, which can also impact the converse approach, have been mentioned by Renaud et al.81 who showed that the negative correlation between dairy fat and coronary disease disappeared when wine consumption was included in the statistical analysis.

Despite their high SFA content, however, dairy products truly appear to be associated with a lower cardiovascular risk. It is, of course, necessary to draw a distinction between butter and other dairy products since butter is generally considered less healthy.

A 2004 literature review analyzed 10 prospective (cohort) studies and 2 retrospective studies.82,83 All but one showed a protective effect of dairy product consumption on coronary risk, and all revealed a lower risk of stroke, with the mean odds ratios calculated to be 0.87 and 0.83, respectively, representing around a 15% risk reduction. Two more recent studies84,85 confirmed an inverse relationship between first infarction and the consumption of dairy fat as measured by a characteristic dairy fatty acid (the C15:0 pentadecanoic acid) in adipose tissue or in phospholipids; however, in one of the two studies, the correlation disappeared after adjusting for standard risk factors.

This is in agreement with other studies that have shown an inverse relationship between fat and SFA intake on the one hand and the risk of ischemic or hemorrhagic stroke on the other – not only in Asian populations but also in the Framingham study86 and in male US healthcare professionals.87,88 Conclusions should not be drawn too hastily, however, since these are only observational studies that (especially for prospective studies) may express case-control differences in lifestyle or behavior. If there is a benefit, however, the nutritional factors involved and the impact sites remain to be identified.

Impact sites for dairy fats

It is clear that SFAs increase not only LDLC89 but also HDLC.13 In the context of a carbohydrate-rich diet, a reduction in SFA intake resulted in a lipid profile that was similar to that seen in metabolic syndrome.1872 Other fatty acids could be involved: minor fatty acids such as C15:0 or C17:0, (although C15:0 increases cholesterol), the trans vaccenic acid or certain conjugated linoleic acids such as rumenic acid, the effects of which are not yet understood but could be positive.84,85

The increase in LDLC is reduced, however, in the presence of calcium; hence, cheese is less hypercholesterolemic than butter,90 since the calcium forms soaps with the fatty acids liberated by hydrolysis in the intestine, i.e. the fatty acids situated in positions 1 and 3 on glycerol.91 Furthermore, several studies have shown that calcium could have a hypocholesterolemic effect, although our group did not find one with calcium-rich water.92 The type of food associated with the triglycerides doubtless also plays a role.

Several studies have also revealed a moderately hypocholesterolemic effect of fermented milks,93,94 which may occur through deconjugation and thus excretion of bile acids.95 Elevated HDLC has also been observed during the consumption (over 6 months) of fermented milks.96

Other risk factors are influenced by the consumption of dairy products: a decrease in arterial blood pressure and a lower prevalence of metabolic syndrome has notably been reported in several studies including the Monica study,97 the Cardia Study,98 the DESIR study,99 and a study performed in an Iranian population.100 Calcium could play a role in the observed decrease in blood pressure101 and a large body of work argues in favor of a modest role of this mineral in weight loss.102 Hence, the benefit of dairy products may also involve factors other than cholesterol or fatty acids, e.g., calcium, probiotics, and certain functional peptides.


The cardiovascular disease/atherosclerosis/SFAs/plasma cholesterol equation is misguided because it is too simplistic (Table 1). In this respect, one can cite the Honolulu Heart Program study, which never revealed a link between diet and atherosclerosis in deceased, autopsied ethnic Japanese CVD subjects living in Hawaii.103 It is certain that with a Western diet and lifestyle, a very high (and thus excessive) SFA intake is associated with a higher cardiovascular risk and an increase in cholesterol levels. Reduction of the lipid intake and SFA intake, however, does not provide any benefit (and can even be harmful) in the absence of weight loss and physical activity in subjects with atherogenic dyslipidemia, coronary disease or simply insulin resistance, a metabolic syndrome, or overweight. Moreover, it appears that in coronary heart disease, the omega-6 fatty acid intake must not be too high, since it can increase plaque instability.

View this table:
Table 1

Summary of the effects of fatty acids on cardiovascular risk factors and of related epidemiological studies.


In the active, non-overweight subject with no risk factors, the diet can be either high-carbohydrate or high-fat, with a predominance of monounsaturated fatty acids and without excessive reductions in SFAs. In the overweight subject (regardless of the metabolic syndrome status) who is on a lower-calorie diet but is sedentary (and thus in the absence of long-term weight loss), it is preferable to reduce the quantity of carbohydrates and maintain a balanced fat intake with sufficient MUFAs, PUFAs, and SFAs, without cutting the latter too drastically. In the coronary patient (again, regardless of the metabolic syndrome status), it is advisable to limit the omega-6 PUFA intake in favor of MUFAs and, above all, to increase the omega-3 fatty acid intake (alpha-linolenic acid and especially LC-PUFAs), particularly in the context of a previous myocardial infarction.

There are significant interactions between fatty acids, so the P/S ratio must doubtlessly fall in the range between 0.5 and 1 and the omega-6/omega-3 ratio between 2 and 5. In addition, nutrition is a complex interaction and it is not appropriate to stigmatize a particular nutrient or food product. Finally, it is important not to oversimplify: we shall increasingly be able to consider each type of fatty acid individually rather than lumping them all together. It is already clear that dietary advice must be individualized and this will be increasingly true in the future as our knowledge of specific genetic factors improves.


The author thanks Elise Clerc and Marie-Claude Bertière for their helpful comments as well as Yvette Soustre and Helen Bishop MacDonald for their kind re-reading.


View Abstract