Omega-6 Fatty Acids and Cardiovascular Disease : Are We Getting Closer to the Truth?

Thomas A.B. Sanders, DSc


Circulation. 2019;139(21):2437-2439. 

Linoleic acid (18:2n-6; LA) is an essential polyunsaturated fatty acid provided mainly by vegetable oils, whole grains, nuts, and seeds. Fats in meat and milk from ruminant animals (cows, sheep, goats) are low in LA (2%–3% by weight of the total fatty acids) because microbes within the rumen convert LA into saturated and trans unsaturated fatty acids. The average US diet provides 6% to 7% of the dietary energy as LA (in the range of 14–18 g/d), which is in line with the US recommendation[1] (5%–10% energy). However, only 1% dietary energy is required to prevent essential fatty acid deficiency, and the World Health Organization and Food and Agriculture Organization[2] suggest an upper limit of 9% dietary energy for LA (about 20 g/d for women and 25 g/d for men). However, there have been repeated claims that such LA intakes might increase the risk of cardiovascular disease (CVD) through a variety of hypothetical mechanisms (Figure).


Relationship between increased intakes of linoleic acid (LA) with arachidonic acid and its metabolites, putative biological effects (both protective and harmful), and the association of a higher proportion of LA in blood and adipose tissue levels of LA with risk of cardiovascular disease (CVD) according to the current meta-analysis.3 Overall, higher blood/adipose tissue LA levels appear to be associated with lower CVD risks, although residual confounding may still occur. Convincing, possible, and not demonstrated (ND) refer to level of evidence. BP indicates blood pressure; CHD, coronary heart disease; IP, F2-isoprostanes; LDL-C, low-density lipoprotein cholesterol; LTB4, leukotriene B4; PGE2, prostaglandin E2; PGI2, prostacyclin I2; and TxA2, thromboxane A2.

In the 1950s, Ancel Keys[4] in well-controlled feeding trials showed that polyunsaturated fatty acids had a weaker but opposite effect on the blood cholesterol–raising effects compared with saturated fatty acids. In 1977, the McGovern Report therefore recommended that saturated fatty acids be replaced by polyunsaturated fatty acids (mainly LA) to help lower blood cholesterol to tackle the coronary heart disease (CHD) epidemic in the United States. The food industry responded by promoting vegetable oil high in LA such as corn oil, sunflower oil, and soybean oil, as well as margarine and shortenings made from these oils. Consequently, LA intakes have increased in the United States from 3% to 4% in the 1960s to around 6% to 7% dietary energy in the 1980s and have remained at this level since then, a finding corroborated by an increase in the proportion of LA in adipose tissue from about 8% in the 1960s to about 18% nowadays.[5] Similar increases in LA intake have occurred in the United Kingdom and other countries with the global expansion of vegetable oil production.

The wisdom of advocating an increased intake of polyunsaturated fatty acids began to be challenged in the 1990s when it was realized that the oxidative modification of low-density lipoprotein (LDL) was a key step in the atherogenic pathway. LDL enriched in LA is more susceptible to copper-induced oxidation ex vivo compared with LDL enriched with oleic acid.[6] It was postulated that this increase in LA might promote atherogenesis by increasing the amount of oxidized LDL taken up by tissue macrophages in the vascular endothelium to form foam cells. However, there is no evidence that higher intakes of LA increase lipid peroxidation in vivo as measured by F2-isoprostanes (8-iso-PGF)[7] or increase LDL oxidation in vivo. Furthermore, in nonhuman primates, LA-rich vegetable oils decrease atherogenesis compared with saturated or monounsaturated fatty acids.[1]

Other commentators[8,9] speculated that a higher intake of LA might increase CVD risk because some of its metabolites are involved in thrombosis and inflammatory pathways. LA is converted to arachidonic acid (20:4n-6n-6; AA), which is the primary precursor of eicosanoids (prostaglandins, thromboxane, prostacyclin, leukotrienes, anandamides) and other potentially biologically active metabolites. However, intakes of LA greater than those required to prevent essential acid deficiency do not augment AA levels in tissues. Moreover, AA metabolites produced in different locations by cyclo-oxygenase can have opposing effects. Nonsteroidal inflammatory drugs with a mode of action that inhibits cyclo-oxygenase have mixed effects on CVD risk.[10] For example, low-dose aspirin inhibits cyclo-oxygenase-1 in platelets and decreases thromboxane A2 and risk of CVD in high-risk patients, but cyclo-oxygenase-2 inhibitors reduce PGI2 and have adverse effects on cardiovascular risk. There is no evidence from human dietary trials that increasing LA intake has adverse effects on platelet aggregation, endothelial function,[11] or surrogate markers for inflammation such as high-sensitivity C-reactive protein.[12] Overall, it is difficult to predict what the net CVD effect of increasing dietary LA intake may be from the multiple proven or hypothesized biological actions of these derived metabolites; the best one can do (outside of randomized trials) is to examine the best evidence associating dietary intakes or blood levels of LA and AA with hard outcomes.

In terms of dietary evidence, Chowdhury et al[13] found no association between the dietary intake of omega-6 fatty acids (relative risk, 1.01 95% CI, 0.96–1.07) or the level of LA in circulating lipids (relative risk, 0.99 [95% CI, 0.77–1.28]) with risk of CVD mortality in a meta-analysis of observational studies. However, the effect of LA replacing saturated fatty acids was not considered. Farvid et al[14] addressed this issue in a larger meta-analysis including unpublished data from the Health Professionals' Study and the Nurses' Health Study, which had multiple assessments of dietary intake during follow-up rather than relying on a single baseline measurement of intake. They estimated that replacing 5% of the dietary energy as saturated fatty acids with LA was associated with a 9% lower risk of CHD events (relative risk, 0.91 [95% CI, 0.86–0.96]) and a 13% lower risk of CHD deaths (RR, 0.87 [95% CI, 0.82–0.94]).

The meta-analysis reported in this issue of the journal[3] used biomarkers of LA intake (adipose tissue and blood) and supports an association between higher intakes of LA and lower, not higher, risk of CVD and CVD mortality. Its strengths are the pooling of individual data, the larger number of participants and events, and the statistical adjustments for many confounding factors. Adipose tissue and total plasma total lipids were found to be the most sensitive biomarkers of intake. This report found no association between the proportion of AA in blood or tissue lipids and risk of CVD, in contrast to the report by Chowdhury et al[13] in which a higher proportion of AA was associated with decreased risk. However, an interesting incidental observation was a lower risk of CVD among carriers of the T allele of a common polymorphism (rs174547) of FADS1 who had higher levels of AA.

The top-line result from this meta-analysis suggests a 7% lower risk of incident CVD when those in the top and those in the bottom fifth of LA in blood or adipose tissue lipids are compared. However, the 22% lower risk of fatal CVD is much greater than expected. Exchanging 5% energy provided as saturated fatty acids with LA would lower LDL cholesterol by about 0.29 mmol/L2 and would be predicted to decrease risk of nonfatal and fatal CVD by 6% and 3%, respectively.[3] Because elevated LDL cholesterol is a much weaker risk factor for ischemic stroke than CHD, it is somewhat surprising that the risk of ischemic stroke was lower than that of CHD. It is, of course, conceivable that LA may have other cardioprotective effects; indeed, there is some evidence[2] that higher intakes of LA are associated with lower blood pressure. Furthermore, this may be an indirect effect because a higher intake of LA is associated with healthier dietary habits, particularly plant-based diets rich in whole grains and nuts such as the Dietary Approaches to Stop Hypertension and vegetarian diets,[15] which are associated with lower blood pressure, the main risk factor for stroke.

The authors acknowledge that residual confounding cannot be fully excluded in their findings. Although critics will no doubt call for a randomized, controlled trial to convincingly demonstrate a protective effect of LA, this is unlikely to happen owing to the enormous cost of recruiting >100 000 participants and maintaining them on their respective diets for several years. However, this latest meta-analysis at least provides a degree of certainty and, therefore, further comfort that an increase in LA intake within the range advocated by the American Heart Association[1] has no harms associated with it.