The eicosapentaenoic acid:arachidonic acid ratio and its clinical utility in cardiovascular disease

ABSTRACT

Eicosapentaenoic acid (EPA) is a key anti-inflammatory/anti-aggregatory long-chain polyunsaturated omega-3 fatty acid. Conversely, the omega-6 fatty acid, arachidonic acid (AA) is a precursor to a number of pro-inflammatory/pro-aggregatory mediators. EPA acts competitively with AA for the key cyclooxygenase and lipoxygenase enzymes to form less inflammatory products. As a result, the EPA:AA ratio may be a marker of chronic inflammation, with a lower ratio corresponding to higher levels of inflammation. It is now well established that inflammation plays an important role in cardiovascular disease. This review examines the role of the EPA:AA ratio as a marker of cardiovascular disease and the relationship between changes in the ratio (mediated by EPA intake) and changes in cardiovascular risk. Epidemiological studies have shown that a lower EPA:AA ratio is associated with an increased risk of coronary artery disease, acute coronary syndrome, myocardial infarction, stroke, chronic heart failure, peripheral artery disease, and vascular disease. Increasing the EPA:AA ratio through treatment with purified EPA has been shown in clinical studies to be effective in primary and secondary prevention of coronary artery disease and reduces the risk of cardiovascular events following percutaneous coronary intervention. The EPA:AA ratio is a valuable predictor of cardiovascular risk. Results from ongoing clinical trials will help to define thresholds for EPA treatment associated with better clinical outcomes.

KEYWORDS:

• Eicosapentaenoic acid
• arachidonic acid
• cardiovascular disease
• coronary artery disease
• omega 3 fatty acids
• long-chain polyunsaturated omega-3 fatty acids
• inflammation

1. Introduction

Omega-3 fatty acids, including eicosapentaenoic acid (EPA), docosahexaenoic acid (DHA), α-linolenic acid, stearidonic acid, and docosapentaenoic acid are long chain, polyunsaturated fatty acids, most commonly found in fish oil [1,2]. Supplementation with purified EPA has been shown to have pleiotropic beneficial effects on the development and progression of atherosclerotic plaque (Figure 1) [3]. EPA has been shown to lower triglycerides and to reduce the levels of atherogenic lipoproteins, including remnant lipoprotein cholesterol, inflammatory markers (including high-sensitivity C-reactive protein [hsCRP]), and inflammatory cytokines [4–6]. EPA also has beneficial effects on endothelial function, oxidative stress, foam cell formation, plaque formation/progression and rupture, platelet aggregation, and thrombus formation [4]. However, the major dietary polyunsaturated fatty acid is linoleic acid, a precursor to arachidonic acid (AA) [1].

Figure 1. Atherosclerosis is a multistep process ranging from endothelial dysfunction to plaque development, progression, and rupture, leading to thrombus formation and cardiovascular events. EPA has been reported to have beneficial effects on many of these steps. CRP, C-reactive protein; EPA, eicosapentaenoic acid; LDL, low-density lipoprotein; ox-LDL, oxidized low-density lipoprotein; RLP, remnant-like particle. Adapted from [3] with permission from Elsevier.

AA is a metabolic precursor for many prostaglandins, thromboxanes, leukotrienes, and other oxidized derivatives, and thus is considered to be predominantly pro-inflammatory (Figure 2) [7–9]. It is converted into these inflammatory mediators by enzymes such as cyclooxygenase (COX) and lipoxygenases (LOXs) [7]. Among the most important AA-derived inflammatory and aggregatory mediators are prostaglandin E₂ (PGE₂), leukotriene B₄ (LTB₄), and thromboxane A₂ (TXA₂) [7]. PGE₂ increases production of interleukin 6 (IL-6), induces fever, and increases vascular permeability, vasodilation, and pain; and LTB₄ increases production of tumor necrosis factor (TNF), IL-1β, and IL-6 while also increasing vascular permeability, enhancing local blood flow, inducing release of lysosomal enzymes, and acting as a chemotactic agent for leukocytes; TXA₂ is a vasoactive agent that affects blood pressure and blood flow and promotes activation and aggregation of platelets [10]. Although AA is predominantly pro-inflammatory, it does have some anti-inflammatory and pro-resolving effects. AA has also been shown to have an essential role in the central nervous system, especially the brain and cognitive functions, as well as in skeletal muscle and immune system function [11]. EPA can act as a competitive inhibitor to AA for COX and LOX enzymes [7], resulting in compounds that have slightly different structures to AA-derived compounds, and importantly, have less potent inflammatory effects [2,9,12,13]. In addition, AA and EPA give rise to different specialized pro-resolving mediators, with AA being a precursor for lipoxins LXA₄ and LXB₄ and EPA being a precursor for resolvins RvE₁ and RvE₂ [14]. EPA can also compete with AA for incorporation into membrane phospholipids [2]. As a result, the ratio of EPA to AA may play a role in regulating inflammatory processes and impacting the development and severity of inflammatory diseases, including atherosclerosis and other cardiovascular diseases [2,4,7]. The ratio of EPA:AA is therefore a measure of chronic inflammation, with a lower ratio corresponding to higher levels of inflammation. It may be worth noting that omega-6 fatty acids and AA are not entirely pro-inflammatory, and like EPA’s derivatives/metabolites, under certain conditions, those of AA can also contribute to the resolution of inflammation, albeit to a lesser degree [8]. However, in the context of cardiovascular disease as discussed herein, EPA is still the main driver of cardiovascular benefit.

Figure 2. Differential metabolism and metabolic products of EPA and AA. *Pro- and anti-inflammatory. ^†Anti-inflammatory and pro-resolving. AA: arachidonic acid; EPA, eicosapentaenoic acid; COXs: cyclooxygenases (COX-1 and COX-2); DHA: docosahexaenoic acid; HEPE: hydroxyeicosapentaenoic acid; HPEPE: hydroperoxyeicosapentaenoic acid; HPETE: hydroperoxyeicosatetraenoic acid; 12-HETE: 12-hydroxytetraenoic acid; 15-HETE:15(S)-hydroxyeicosatetraenoic acid; HPEPE: hydroperoxyeicosapentaenoic acid; LTs: leukotrienes; PGs: prostaglandins; PLA₂: phospholipase A₂; LOXs: lipoxygenases; TX: thromboxane. Adapted from [9] with permission of Oxford University Press.

The modern Western diet is rich in omega-6 fatty acids and deficient in omega-3 fatty acids, resulting in a pro-inflammatory state in many individuals [15]. Treatment with EPA, however, increases EPA levels in both plasma and tissues in a dose-dependent manner, thereby increasing the EPA:AA ratio [16–18]. EPA treatment also results in a corresponding decrease in plasma levels of AA and may, therefore, help reverse the EPA:AA ratio from being pro-inflammatory/pro-aggregatory to anti-inflammatory/anti-aggregatory (Figure 3) [17,19–24]. This review will focus on the role of the EPA:AA ratio as a potential prognostic marker and treatment target in cardiovascular disease. The clinical implications of the EPA:AA ratio and the effects of EPA treatment in key cardiovascular conditions are summarized in Table 1 [25–45].

Table 1. EPA:AA ratio in cardiovascular diseases.

CV condition	Key findings
CAD	EPA:AA ratio is associated with decreased risk of MACE in patients undergoing PCI [25] Patients with angina have significantly lower EPA:AA ratio [26] EPA treatment reduces risk of major coronary events [27]
ACS	Lower EPA:AA ratio is a strong independent risk factor for ACS [28,29] EPA treatment in ACS patients who underwent PCI resulted in reduced risk of future coronary events [30]
MI	PCI patients with lower EPA:AA ratio have a higher risk of periprocedural MI [31] EPA:AA ratio had no effect on the overall rate of MACE, mortality, cardiac failure, and bleeding, but a higher incidence of ventricular tachycardia/ventricular fibrillation while hospitalized is found in patients with lower EPA:AA ratio [32]
Stroke	Higher EPA:AA ratio appears to be protective [33] EPA treatment reduces risk of secondary stroke [34]
CHF	Higher EPA:AA ratio associated with decreased left ventricular wall thickness in diabetes [35] Lower EPA:AA ratio is associated with increased mortality in patients hospitalized for decompensated heart failure [36] EPA treatment improves LVEF [37]
PAD	Lower EPA:AA ratio is predictive of PAD diagnosis [38,39] Lower EPA:AA ratio is associated with major adverse limb effects (e.g. amputation, revascularization), risk of VTE, arterial stiffness in obese subjects, subclinical peripheral artery damage, and death [40–44] EPA treatment is associated with reduced risk of major coronary events [45]

CAD: coronary artery disease; ACS: acute coronary syndrome; MI: myocardial infarction; CHF: chronic heart failure; PAD: peripheral artery disease; MACE: major adverse coronary events; PCI: percutaneous coronary intervention; LVEF: left ventricular ejection fraction; VTE: venous thromboembolism.

Figure 3. EPA and AA in cardiovascular disease. EPA has anti-inflammatory, anti-aggregatory, and beneficial endothelial function effects, while AA generally has pro-inflammatory and pro-aggregatory effects. Shifting the balance to greater amounts of EPA than AA (shown in figure) increases the EPA:AA ratio and has beneficial cardiovascular effects. Lower EPA:AA ratios have been shown to be associated with increased cardiovascular disease risk. AA: arachidonic acid; EPA: eicosapentaenoic acid; IL: interleukin; LTB₄: leukotriene B₄; MCP-1: monocyte chemoattractant protein; PGE₂: prostaglandin E₂; RLP-C: remnant lipoprotein cholesterol; TNF-α: tumor necrosis factor-α; TXA₂: thromboxane A₂ [7,21–24].

2. Coronary artery disease (CAD)

Epidemiological studies have suggested that a low EPA: AA ratio may be associated with an increased incidence of CAD and angina. In an analysis of patients aged 30 to 69 years with and without angina, those with angina were found to have a significantly lower EPA: AA ratio than patients with no significant history of ill health who were not taking any medications (P < 0.05) [26]. In fact, this measure was found to be more predictive of angina than total cholesterol, low-density lipoprotein cholesterol (LDL-C), or triglycerides, indicating that the ratio is an important risk indicator for angina [26]. A more recent analysis found that a higher EPA:AA ratio, but not a higher DHA:AA ratio, was associated with a lower prevalence of CAD in patients with at least one CAD risk factor [46]. The EPA: AA ratio has also been found to be more closely associated with CAD pathophysiology than the DHA:AA ratio (odds ratio [OR], 0.0012 [P = 0.007] and 2.13 [P = 0.53], respectively) [47]. In addition, an analysis of patients who underwent percutaneous coronary intervention (PCI) also found that the serum EPA:AA ratio, but not the omega-3:omega-6 ratio or the DHA:AA ratio, was associated with a reduced incidence of major adverse cardiac events (MACE) [25].

The potential role of EPA treatment in increasing the EPA:AA ratio and reducing cardiovascular risk was suggested in the large, prospective, randomized Japan EPA Lipid Intervention Study (JELIS), which randomized hypercholesterolemic patients to EPA 1.8 g/day or no treatment on a background of statin therapy for primary prevention of cardiovascular events [27,48]. Treatment with EPA not only increased plasma EPA levels, but also increased the EPA:AA ratio approximately two-fold. In addition, the risk of major coronary events was significantly lower in patients with higher EPA:AA ratios than in those with lower EPA:AA ratios, suggesting that modulation of the EPA:AA ratio through treatment with high-purity EPA was mirrored by a reduction in cardiovascular risk [27].

A secondary prevention analysis from the JELIS study also found a significantly lower risk of sudden cardiac death or myocardial infarction (MI) in patients with a history of CAD and a high EPA:AA ratio (≥1.06) compared with those with a lower ratio (<0.55; hazard ratio [HR], 0.58, P = 0.038) [49]. In another study in which patients with stable CAD were randomized to 1.8 g/day EPA or control, EPA-treated patients achieved a significant reduction in serum non-high-density lipoprotein cholesterol (non-HDL-C) – an established risk factor for CAD – suggesting a reduced risk of future CAD events [50]. Notably, in that study, the achieved EPA:AA ratio rather than the absolute change in the EPA:AA ratio was the better independent predictor of a reduction in non-HDL-C at 6 months [50].

The relationship between the EPA:AA ratio and coronary plaque has been investigated in a number of studies using various imaging techniques, including angiography, angioscopy, optical coherence tomography, and intravascular ultrasound [51–56]. A low EPA:AA ratio has been associated with an increased risk of coronary plaque progression and vulnerability to rupture, while a higher ratio is associated with plaque stabilization, an increase in the fibrous cap thickness, and reduced lipid volume [52–56]. The EPA:AA ratio, therefore, has incremental value in predicting the presence of high-risk plaque over conventional coronary artery disease risk factors [55].

A lower EPA:AA ratio (<0.21) has also been significantly associated with the presence of complex coronary lesions in patients undergoing elective coronary angiography [57], and has been correlated with gray-scale median, a measure of plaque calcium, fibrous tissue, and lipid content using ultrasonography [58]. A recent study comparing treatment with EPA plus a statin versus statin alone in patients with stable angina pectoris and acute coronary syndrome (ACS) found that the lipid volume in plaque was significantly reduced only in those treated with EPA; the total atheroma volume was also significantly reduced in this group [59]. The EPA:AA ratio in this study almost doubled in the statin-plus-EPA group compared with no change in the statin-alone group (P < 0.001). A separate study suggested that addition of EPA to a strong statin stabilized coronary plaque within 6 months [60]. Taken together, these results suggest that an increase in the EPA:AA ratio driven by EPA intake reduces non-HDL-C and stabilizes coronary plaque leading to a reduced risk of CAD.

3. Acute coronary syndrome

ACS usually occurs secondary to atherosclerosis in the coronary arteries, either due to plaque rupture or plaque erosion and thrombus formation. As a result, the EPA:AA ratio may be a good biomarker in patients with ACS. Indeed, the EPA:AA ratio has been shown to be a critical risk factor for ACS. In a study of patients who underwent coronary angiography, a lower EPA:AA ratio was found to be a strong independent risk factor for ACS (odds ratio, 0.37; P = 0.023) [28]. In a second study, a lower EPA:AA ratio correlated with age at the time of ACS onset in patients who were admitted for PCI, suggesting that this ratio is a predictive risk factor for early onset of ACS [61]. The EPA:AA ratio has also been shown to be lower in patients in early phase ACS versus those with stable CAD and a history of ACS; those with stable CAD and no history of ACS had the highest ratio and were at lowest risk, suggesting a correlation between EPA:AA and plaque vulnerability [62]. However, it is possible that the decreased EPA:AA ratio in ACS is caused by coronary artery and myocardium inflammation in the acute phase following ACS, rather than being a relevant risk marker. In addition, multiple regression analysis found that EPA (P = 0.009) and the EPA:AA ratio (P = 0.023) but not DHA or the DHA:AA ratio were negatively associated with clinical profiles of ACS in CAD patients [62]. Another study of patients with confirmed CAD found that the EPA:AA ratio was significantly lower in ACS patients than in those with chronic CAD or chest pain syndrome (P < 0.01) [47]. Furthermore, in an observational study of more than 1000 patients with and without ACS, those with the lowest EPA:AA ratios (≤0.33) had a more than three times greater probability of ACS than those with higher ratios (≥0.55); no such relationship was observed for the DHA:AA ratio [29]. In a study of 61 consecutive patients (27 patients with ACS and 34 with stable CAD) who underwent virtual histology-intravascular ultrasound, the EPA:AA ratio was significantly lower in the ACS group (P = 0.00005) prior to PCI [63]. Under multiple logistic regression analysis, an EPA:AA ratio below the study median was associated with ACS (OR, 8.235; 95% confidence interval [CI], 1.436, 47.227; P = 0.018) but not DHA or the DHA:AA ratio below study medians. The EPA:AA ratio therefore is an independent factor for discrimination of ACS from CAD or chest pain syndrome.

Early EPA treatment plus a statin in patients with ACS who had undergone PCI has further been shown to increase the EPA:AA ratio and reduce the risk of future coronary events versus patients treated with statin alone [30]. As with other studies, EPA treatment resulted in a significantly higher EPA:AA ratio at follow-up compared with the control group (P < 0.001). Importantly, a subsequent cardiovascular event occurred in only 9.2% of EPA-treated patients compared with 20.2% in the control group (P = 40.02), indicating that changing the EPA:AA ratio through increased EPA intake was cardioprotective in this population. In addition, similar to JELIS, there were no significant differences in the changes in LDL-C between treatment arms. However, unlike in JELIS, there were no significant differences in the changes in triglycerides between treatment arms [30].

4. Myocardial infarction

An analysis of the Japanese Registry of Acute Myocardial Infarction Diagnosed by Universal Definition (J-MINUET) investigated whether the EPA:AA ratio had any impact on the risk of MACE following acute MI [32]. Although the overall rates of MACE, mortality, cardiac failure, and bleeding were similar for those with higher and lower EPA:AA ratios (≥0.4 vs. <0.4), the incidence of ventricular tachycardia/ventricular fibrillation while hospitalized was significantly higher in the low EPA:AA group (P = 0.008). A second study found that the EPA:AA ratio was predictive of MACE following a first MI; although using a slightly different measure of logAA:logEPA, a lower ratio was associated with better event-free survival [64]. This ratio, therefore, could possibly be used to identify those patients at increased risk of further major cardiac events following initial MI.

A low EPA:AA ratio has also been implicated in periprocedural MI following PCI. In a study of 302 consecutive patients who underwent PCI, those with a lower EPA:AA ratio were significantly more likely to experience subsequent periprocedural MI than those with a higher ratio (P = 0.006) [31]. The predictive value of the EPA:AA ratio in this study was independent of known predictors of MI such as unstable angina pectoris and longer stented length. Patients with prior MI and diabetes have also been shown to have a lower EPA:AA ratio than those without diabetes, suggesting that diabetic patients with a history of MI may benefit from EPA treatment [65]. By contrast, no such relationship was observed for patients with dyslipidemia or hypertension [65].

5. Stroke

A meta-analysis of nine independent cohorts demonstrated that regular fish intake was associated with a reduced risk of stroke, with a relative risk of 0.82 (95% CI, 0.72, 0.94) for fish intake 2 to 4 times per week compared with <1 time per month [66]. The JELIS study demonstrated the role of EPA in stroke prevention in statin-treated patients [34]. Although EPA had no significant effect on primary prevention of stroke, EPA was associated with a 20% relative risk reduction compared with no EPA treatment in secondary prevention of stroke (HR, 0.80; 95% CI, 0.64, 0.997). Although this analysis did not report the EPA:AA ratio, other reports from JELIS have noted the robust increase in this ratio in the EPA-treated patients in this study as described earlier.

The EPA:AA ratio was identified as a risk factor for ischemic stroke in a study in which 65 consecutive patients admitted for their first stroke were compared with 65 aged-matched controls [33]. In non-parametric testing, the EPA:AA ratio was a significant predictor of ischemic stroke, cardioembolism, and large artery atherosclerosis (all P < 0.001). However, under multivariate analysis, EPA:AA was independently associated with only cardioembolism (OR, 0.00012; 95% CI, 0.0000001, 0.143; P < 0.05). The receiver operating characteristic (ROC) area under the curve for the EPA:AA ratio was 0.470, 0.465, and 0.485 for ischemic stroke, cardioembolism, and large-artery atherosclerosis, respectively (all P < 0.001). Although data for the EPA:AA ratio is more sparse in stroke than other major cardiovascular events, it appears that a higher ratio is protective.

6. Chronic heart failure

In patients with chronic heart failure and dyslipidemia, treatment with EPA resulted in an improvement in left ventricular ejection fraction (LVEF) compared with controls who did not receive EPA; the EPA:AA ratio was also positively correlated with LVEF [37]. In this study, the EPA:AA ratio in the EPA-treated group was significantly greater than in the control group, with systolic blood pressure, heart rate, C-reactive protein, B-type natriuretic peptide, AA, and asymmetric dimethylarginine (ADMA) all being significantly lower. Of note, the EPA:AA ratio was significantly and negatively associated with levels of TNF-α, monocyte chemoattractant protein-1, and ADMA. A higher EPA:AA ratio has also been associated with a reduction in left ventricular wall thickness in diabetic subjects [35]. A lower EPA:AA ratio has further been associated with an increased risk of mortality in patients who were hospitalized for treatment of decompensated heart failure [36]. The results of these studies highlight how a higher EPA:AA ratio is associated with improvements in inflammation, endothelial function, and cardiac function in patients with chronic heart failure.

7. Peripheral artery and vascular disease

The EPA:AA ratio has implications in the peripheral vascular system as well as the coronary system. In a study of patients with peripheral artery disease (PAD) caused by femoropopliteal artery lesions, a lower EPA:AA ratio before endovascular therapy was associated with major adverse limb events (including amputation and target lesion revascularization) and death from any cause; patients with a low EPA:AA ratio (≤0.30) had a significantly worse outcome than those with a higher ratio (P < 0.001) [40]. A lower EPA:AA ratio was also found to be associated with subclinical peripheral artery damage in smokers with diabetes, as measured by ankle-brachial index [41].

Knowledge of the EPA:AA ratio may help to identify patients who may experience adverse clinical outcomes and might require adjunctive therapies in addition to endovascular therapy. Two studies of patients with and without PAD suggested that a low EPA:AA ratio was predictive of a PAD diagnosis [38,39]. In the case-control study by Gautam et al., PAD was independently associated under multivariate analysis with the EPA:AA ratio but not the DHA:AA ratio (OR, 0.38; 95% CI, 0.17, 0.86; P = 0.021) [39]. The ROC area under the curve for the EPA:AA ratio in this study was 0.58 (95% CI, 0.51, 0.64; P = 0.020), and the optimal cutoff value, based on the ROC curve, for the EPA/AA ratio was 0.55 (66.34% sensitivity, 47.18% specificity), suggesting more advanced PAD in patients with EPA:AA ≤0.55. Using a cutoff for the EPA:AA ratio of 0.49 in the second study, the ratio was diagnostic of PAD with 92.1% sensitivity and 50.0% specificity [38]. It has further been found that the EPA:AA ratio was significantly lower in patients with critical limb ischemia than in those with intermittent claudication (0.22 vs. 0.38; P = 0.0049); this may explain the higher incidence of cardiovascular events in patients with critical limb ischemia [67].

A small number of patients in the JELIS study (1.2%) had PAD [45]. In the control group, the incidence of major coronary events was significantly higher in patients with existing PAD and in newly diagnosed PAD than in those without PAD (HR, 1.97 [P = 0.039] and 2.88 [P = 0.030], respectively), indicating that the risk of coronary events is substantially higher in this group. Patients with PAD who received EPA had a 56% reduced risk of major coronary events compared with those receiving placebo (HR, 0.44; P = 0.041). Although this analysis of JELIS did not report the EPA:AA ratio, it is probable that EPA treatment in this PAD subgroup led to an increase in the ratio based on the robust ratio increases reported in other JELIS analyses [27].

A low EPA:AA ratio has been associated with accelerated arterial stiffness in obese patients compared with those of normal weight and those with a higher EPA:AA ratio [43]. Furthermore, an increase in the EPA:AA ratio following treatment with EPA in obese patients has been associated with corresponding improvements in arterial stiffness [42]. The EPA:AA ratio has also been shown to be predictive of venous thromboembolism (VTE), with patients experiencing VTE having a lower ratio than matched healthy controls [44]. Patients with abdominal aortic aneurysms also have lower EPA:AA ratios than the general population and ratios similar to patients with CAD, with a mean ratio of 0.44 [68]. In addition, the size of the aneurysm in this study appeared to be inversely correlated with the EPA:AA ratio.

8. Clinical utility of the EPA:AA ratio

The data discussed here identifies the EPA:AA ratio as a potentially useful, simple, sensitive, and reliable independent marker of cardiovascular risk [69]. This ratio has been shown to be at least as strong a marker of cardiovascular risk as total cholesterol, LDL-C, and triglycerides [26]. As such, the EPA:AA ratio may be useful for risk stratification of patients with cardiovascular disease [55]. A number of the studies identified in this review have found that the DHA:AA ratio is a less valuable marker of cardiovascular risk, suggesting that EPA may be superior to DHA for prevention of arteriosclerosis. Membrane models have suggested that EPA and DHA have different molecular locations and orientations resulting in distinct membrane structures which may confer different biological activities [70]. Differences in membrane interactions between EPA and DHA may result in the inhibition of cholesterol crystalline domains by EPA but not DHA, which may impact endothelial function [71]. Furthermore, EPA is more efficiently incorporated into HDL particles than DHA which prolongs and increases EPA’s ability to inhibit HDL oxidation compared with DHA [72]. In one clinical study, EPA incorporation into HDL particles correlated closely with the serum EPA:AA ratio and helped to restore the anti-oxidative and anti-inflammatory properties of HDL particles [73]. In subjects taking highly concentrated n-3 fatty acid ethyl esters, the proportion of EPA incorporated into carotid plaque phospholipids was significantly increased by an average of 100% (P < 0.001), but there was no significant similar increase in DHA content [74].

In general, measurement of EPA and DHA is preferred in red blood cell (RBC) membrane phospholipid, as it displays less variability than measurement in plasma due to limited exchange between plasma and cells [75,76]. However, measurement of whole blood EPA:AA provides similar results to that seen with the much more complex measurement of RBC membrane phospholipid content (R², 0.8675); EPA:AA also provides similar values to measurement of the omega-3:omega-6 ratio (R², 0.7114) [69]. Furthermore, a correlation between EPA:AA in serum and RBC membranes has been demonstrated [77]. Indeed, this correlation was found to be better than that for EPA, or DHA measured alone, suggesting that the ratio is a better marker of polyunsaturated fatty acid status. It is clear, however, that there is a need for standardization of assay protocols and identification of clinically relevant thresholds for practical application of this ratio, accounting for accuracy, practicality, and cost.

The omega-3 index is another increasingly recognized risk factor for death from coronary heart disease introduced by Harris and Von Schacky in 2004 [78], and is a possible clinical alternative to the EPA:AA ratio. The omega-3 index was defined as the content of EPA and DHA in RBC membranes. The rationale for the omega-3 index is that fatty acid composition of RBC phospholipids parallels the composition found in cardiac muscle phospholipids [79]. An omega-3 index ≤4% is associated with low cardioprotection, whereas an index ≥8% is associated with high cardioprotection. A recent analysis of the Framingham Offspring Cohort found that inclusion of the omega-6 fatty acid concentration as a covariate, or replacing the omega-3 index with the omega-6:omega-3 ratio had no effect on the results. The authors therefore concluded that measurement of omega-3 fatty acids alone is a sufficient marker for cardiovascular disease [80].

AA is a well-known theoretical precursor to a number of pro-inflammatory and pro-aggregatory mediators (Figure 2) including prostaglandin E₂, thromboxane A₂, and leukotriene B₄, and thus is often posited as a promoter of cardiovascular risk. A clear association between AA and atherosclerotic cardiovascular disease or events, however, is not persuasive. For instance, although studies of AA levels in adipose tissue suggested a correlation with cardiovascular risk [81,82], methodological and unaccounted comorbidities decrease the clinical relevance of these limited observations. Another potential negative association is that AA peroxidation products give rise to 9-hydroxytetraenoic acid (9-HETE) and F2-isoprostane, which have been shown to be independently associated with angiographic evidence of coronary artery disease, even after adjustment for Framingham risk score and C-reactive protein [83]. However, other studies have found no relationship between AA status and cardiovascular disease outcomes [84].

A consideration of the entire AA metabolome is required to appreciate AA not as a risk factor, but rather a fatty acid important for normal health [11]. Among some anti-inflammatory and anti-aggregatory AA metabolites are lipoxin A₄, and epoxyeicosatrienoic acids [85–87]. Higher plasma levels of omega-6 polyunstaurated fatty acids, mainly AA, have also been associated with lower pro-inflammatory markers, such as interleukin-6 and interleukin-1 receptor antagonists, and increased levels of anti-inflammatory markers, particularly transforming growth factor-β [88]. Consistent with this, Harris et al. have argued that AA supplementation does not generate a pro-inflammatory response, suggesting rather that a more inflammatory state is due to low EPA and DHA levels [89]. While these are important considerations, Harris’ conclusions were based on an analysis conducted in a large population from which patients with existing cardiovascular disease were excluded. However, multivariate analyses have found that the EPA:AA ratio and not EPA alone is predictive of the presence of clinically relevant high-risk plaque [55]. It is possible, therefore, that although low EPA is more important than high AA, inclusion of the pro-inflammatory AA is of importance at least for stratification of patients with existing CAD and the presence of active endovascular inflammation. There is no evidence that interventions to lower AA, rather than to increase EPA, have a role in atherosclerotic cardiovascular disease risk management.

A limitation of the EPA:AA ratio is that a clinically useful threshold for identification of at-risk patients or for intervention has yet to be identified in a large, prospective study. This is complicated by the fact that different cultures and regions have different EPA and AA content due to varying diets, which affects not only the EPA:AA ratio, but also cardiovascular risk [21]. The potential for such regional differences is highlighted by the fact that most studies of the EPA:AA ratio and EPA treatment have been conducted in Japanese patients. As a result, there is a dearth of data in Western populations. This, in turn, may determine the effectiveness of EPA treatment strategies in different populations.

Recently completed and ongoing international and US-based studies of high-purity omega-3s will help to determine these thresholds in Western populations, as well as establish the role of EPA treatment in patients at risk for atherosclerotic disease. The recently completed Reduction of Cardiovascular Events with Icosapent Ethyl–Intervention Trial (REDUCE-IT) [90] may help to clarify the relationship between EPA and/or the EPA:AA ratio and cardiovascular outcomes. This trial evaluated the potential benefit of icosapent ethyl 4 g/day vs. placebo on cardiovascular outcomes in 8179 randomized statin-treated patients with LDL-C controlled to between 41 and 100 mg/dL, elevated triglycerides 135 to 500 mg/dL, and either established cardiovascular disease (secondary prevention cohort) or diabetes and at least one other cardiovascular risk factor (primary prevention cohort). Findings from REDUCE-IT indicated that over a median follow-up time of 4.9 years, icosapent ethyl was associated with a statistically significant relative risk reduction of 25% (HR, 0.75; 95% CI, 0.68, 0.83; P < 0.001) in the primary composite endpoint of the first occurrence of MACE (cardiovascular death, non-fatal MI, non-fatal stroke, coronary revascularization, or unstable angina requiring hospitalization) and 26% (HR, 0.74; 95% CI, 0.65, 0.83; P < 0.001) in the key secondary composite endpoint (cardiovascular death, non-fatal MI, or non-fatal stroke). In the primary and key secondary endpoints, respectively, this represented an absolute between-group difference of 4.8% and 3.6% and a number needed to treat of 21 and 28 over 4.9 years. This was accompanied by a 20% reduction in cardiovascular death (HR, 0.80; 95% CI, 0.66, 0.98; P = 0.03), 31% reduction in MI (HR, 0.69; 95% CI, 0.58, 0.81; P < 0.001), and 28% reduction in stroke (HR, 0.72; 95% CI, 0.55, 0.93; P = 0.01). In the icosapent ethyl treatment group, EPA levels increased by 394% from a median of 26 µg/mL at baseline to 144 µg/mL at 1 year. Future analyses from REDUCE-IT may provide additional information on the association between the EPA:AA ratio and atherosclerotic cardiovascular disease. Regarding safety in REDUCE-IT, the overall adverse event rates were similar across treatment groups. There were numerically more serious adverse events related to bleeding in the icosapent ethyl group but the overall rates were low (2.7% for icosapent ethyl vs. 2.1% for placebo, P = 0.06), with no fatal bleeding observed and no significant increase in adjudicated hemorrhagic stroke or serious central nervous system or gastrointestinal bleeding. While the rates were low, a significantly higher rate of hospitalization for atrial fibrillation or flutter was observed in the icosapent ethyl group (3.1% for icosapent ethyl vs. 2.1% for placebo, P = 0.004). However, there was a relative risk reduction of stroke of 28% as noted earlier. Furthermore, the tertiary endpoints of sudden cardiac death and cardiac arrest were reduced by 31% (HR, 0.69; 95% CI, 0.50, 0.96) and 48% (HR, 0.52; 95% CI, 0.31, 0.86), respectively, in the icosapent ethyl group compared with placebo, suggesting that while there was a small increase in atrial fibrillation/flutter, there was a potential benefit in ventricular arrhythmias [90]. Because of the findings of REDUCE-IT, the American Diabetes Association recently issued a Level A recommendation that icosapent ethyl be considered for reducing CV risk in statin-treated patients with controlled LDL-C, TG 135–499 md/dL, diabetes, and atherosclerotic cardiovascular disease or other cardiac risk factors [91]. The positive results of this study raise the important clinical question as to whether the risk of cardiovascular events is lower for patients who have a naturally high EPA:AA ratio, or for those who have taken purified EPA. Future studies may help to address this important issue.

The US-based, randomized, double-blind, placebo-controlled Effect of Vascepa on Improving Coronary Atherosclerosis in People With High Triglycerides Taking Statin Therapy (EVAPORATE) trial is designed to assess the effect of icosapent ethyl 4 g/day on low-attenuation plaque volume in patients with coronary atherosclerosis and elevated triglycerides receiving stable statin treatment [92]. In this study, the EPA:AA ratio is one of the secondary endpoints [92]. Other important ongoing studies include the Randomized Trial for Evaluation in Secondary Prevention Efficacy and Combination Therapy–Statin and EPA (RESPECT-EPA), an open-label study in Japan of 1.8 g/day of high-purity EPA in nearly 4000 statin-treated patients with stable CAD; and the Statin Residual Risk Reduction With Epanova in High Cardiovascular Risk (STRENGTH), a cardiovascular outcomes trial investigating the effects of high-dose EPA+DHA [93]. Analysis of EPA:AA ratios from these studies will help to define thresholds for intervention with EPA as well as target ratios indicative of reduced risk for cardiac events.

9. Conclusions

The EPA:AA ratio has been shown to be a robust marker of future cardiovascular events in a number of clinical settings. Conversely, the DHA:AA ratio has little prognostic value, suggesting that treatment with EPA, rather than DHA or a mix of omega-3 fatty acids, is likely the best intervention for modulating this ratio and reducing cardiovascular risk. Critically, treatment with omega-3 fatty acids, such as high-purity EPA, can improve this ratio and has been associated with better clinical outcomes. Ongoing clinical outcomes studies will help to define potential thresholds for treatment targets. However, it will be important to be aware that such targets may differ depending on variations in the patient population, such as dietary intake of omega-3 and omega-6 fatty acids and baseline cardiovascular conditions.

Declaration of interest

J.R. Nelson serves on speaker’s bureaus for Boston Heart Diagnostic Laboratories, Amarin Pharma, Amgen, Kowa, Boehringer Ingelheim, Sanofi, and Regeneron; and is a consultant and advisor to and stock shareholder of Amarin Pharma and Amgen. S. Raskin serves on speaker’s bureaus for Amarin, Amgen, Boston Heart Diagnostics, Sanofi, and Regeneron. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties. A reviewer on this manuscript has advised Amarin in the past. A second reviewer operates a laboratory for fatty acid analyses and has received honoraria for consulting and/or speaking from BASF/Pronova, Huntsworth Medical, DSM, Marine Ingredients, and Norsan. A third reviewer has received remuneration for lectures from Amgen Astellas Biopharma K.K., Astellas Pharma Inc., AstraZeneca Pharma Ltd., GlaxoSmithKline K.K., Sanofi K.K., Shionogi Daiichi-Sankyo Pharmaceutical Company Ltd., Takeda Pharmaceutical Company Ltd., Terumo Corporation, Medtronic plc., Boehringer Ingelheim GmbH., Bayer AG., and MSD K.K., trust research and joint research funds from Kowa Pharmaceutical Company Ltd., Daiichi-Sankyo Company Ltd., Canon Medical Systems Corporation, Nihon Medi-Physics Company Ltd., Kaken Pharmaceutical Company Ltd., Kirin Holdings Company Ltd., Philips Respironics Ltd., Sanwa Kagaku Kenkyusho Company Ltd., and St. Jude Medical Japan Company Ltd., and scholarship fund from Astellas Pharma Inc., Abbott Laboratories, Otsuka Pharmaceutical Company Ltd., Sanofi K.K., Shionogi & Company Ltd., Sumitomo Dainippon Pharmaceutical Company Ltd., Boehringer Ingelheim GmbH., Bayer AG., Teijin Ltd., Daiichi-Sankyo Company Ltd., Takeda Pharmaceutical Company Ltd., MSD K.K., Tanabe Mitsubishi Pfizer Pharmaceutical Company Ltd., Philips Respironics Ltd., Bristol-Myers Squibb Company Ltd., Boston Scientific Ltd., St. Jude Medical Japan Company Ltd., and Actelion Pharmaceuticals Japan Ltd. They have been in affiliation with some endowed departments, including VitalAire Japan Company K.K., Philips Respironics Ltd., Fukuda Denshi Company Ltd., ResMed Ltd. and is a PI of the RESPECT-EPA study. Peer reviewers on this manuscript have no other relevant financial relationships or otherwise to disclose.

Additional information

Funding

This article was funded by Amarin Pharma Inc, Bedminster, NJ. Medical writing assistance was provided by Peloton Advantage, LLC, an OPEN Health company, Parsippany, NJ, and funded by Amarin Pharma Inc.