Rounding the corner on residual risk: Implications of REDUCE‐IT for omega‐3 polyunsaturated fatty acids treatment in secondary prevention of atherosclerotic cardiovascular disease

Abstract Patients with established atherosclerotic cardiovascular (CV) disease remain at increased risk of major adverse cardiovascular events even during optimal lipid‐lowering therapy. Recent studies using the methods of Mendelian randomization, as well as analyses of data from large statin trials, have concluded that elevated triglyceride (TG) levels contribute to that increased risk. Omega‐3 polyunsaturated fatty acids (omega‐3 PUFAs) from fish and shellfish (eicosapentaenoic acid [EPA] and docosahexaenoic acid [DHA]) reduce TG levels when added to the diet in sufficient amounts, and they have favorable effects on several other markers of CV risk. However, trials of omega‐3 PUFAs have had inconsistent findings regarding CV risk reduction. Recently, the REDUCE‐IT (Reduction of Cardiovascular Events with EPA‐Intervention Trial) trial reported that treatment of such high‐risk patients with icosapent ethyl, a purified and stabilized ethyl ester of EPA, reduced the risk of the trial's primary CV endpoint by 25% (95% confidence intervals [CI], 32%‐17%; P < .001). To appreciate the clinical implications of this result, it is important to understand how the REDUCE‐IT trial differed from previous trials, especially with regard to patient enrollment criteria and treatment dosing. We discuss these design features relative to other trials. TG lowering can account for only part of the risk reduction seen with icosapent ethyl; we also consider other potential contributory mechanisms.

To appreciate the clinical implications of this result, it is important to understand how the REDUCE-IT trial differed from previous trials, especially with regard to patient enrollment criteria and treatment dosing. We discuss these design features relative to other trials. TG lowering can account for only part of the risk reduction seen with icosapent ethyl; we also consider other potential contributory mechanisms. The trial demonstrated that treatment with a high dose (4 g per day) of the omega-3 polyunsaturated fatty acid (PUFA) icosapent ethyl significantly reduced the risk of ischemic cardiovascular (CV) events and CV death in patients who (a) were on a stable dose of statin, (b) had established CV disease or type 2 diabetes mellitus and at least one additional CV risk factor, and (c) had elevated triglyceride (TG) levels at baseline (>135 or 150 mg/dL). These results were remarkable because they followed a series of major TG-lowering trials that had failed to achieve reductions in CV outcomes in similar groups of patients. [2][3][4][5][6] Why was REDUCE-IT different? Are the results due solely to lowering of TG levels? What do the results mean for prevention of CV events in patients with established CV disease and elevated TG levels?

| RESIDUAL RISK
In recent decades, statins have become the mainstay of lipid-lowering therapy to reduce risk of CV events in patients with atherosclerotic cardiovascular disease (ASCVD). 5 Statin therapy, which primarily targets low-density lipoprotein cholesterol (LDL-C), has a firmly established role in reducing the risk of CV events by 25% to 35% in such patients. 5 The reduction in risk is related to the absolute reduction in LDL-C levels. 7 However, despite successful lowering of LDL-C levels with statins, patients with ASCVD remain at elevated risk of CV events. 8 PCSK9 (proprotein convertase subtilisin-kexin type 9) inhibitors further reduce LDL-C levels and CV event rates, but excess CV risk persists (eg, 3-year event rates of approximately 10-13 percent) even in patients with median LDL-C levels of 30 to 40 mg/dL achieved with the combination of a statin and a PCSK9 inhibitor. 9,10 Although non-lipid-related factors-such as smoking, obesity, and diabetes mellitus-account for some of this residual risk, 8 elevated TG levels or low high-density lipoprotein cholesterol (HDL-C) levels have been suspected, since they have been shown to be markers of residual risk in analyses of data from several large statin trials. [11][12][13][14][15] Other lipid and lipoprotein fractions-such as non-HDL-C, apolipoprotein B, and LDL particle number-are also markers of residual risk. 16 There is a strong inverse relationship between TG levels and HDL-C levels; individuals with low HDL-C levels generally have elevated TG levels and vice versa. 17 However, because observational studies have consistently found a link between low HDL-C and CV risk, [18][19][20][21][22][23] several clinical trials were conducted in an effort to reduce CV risk by administering treatments intended to increase HDL-C levels. [24][25][26][27][28] Although the treatments substantially increased HDL-C levels, only one trial reported a small decrease in major CV event rates, and that decrease was consistent with the degree of LDL-C lowering. 28 We now have a better understanding of why efforts to increase HDL-C levels have had such limited success in reducing CV event rates; recent studies using the method of Mendelian randomization (MR) have supported the conclusion that low HDL-C is not causally associated with increased CV risk. [29][30][31] These studies have revealed that genetic variants influencing HDL-C levels do not significantly affect CV risk. Thus, although a low HDL-C level is a marker of increased CV risk, there is now a strong body of evidence from both clinical trials and MR studies that a low HDL-C level is not causally related to an increased risk of CV events. 32 During the era just described, TG re-emerged as an independent CV risk factor. As recently as 2011, a scientific statement from the American Heart Association concluded that the role of TG as an independent causal factor for CV disease was "debatable." 17 Subsequently, MR studies provided strong evidence that elevated TG-rich lipoproteins (TRLs) are causally related to increased risk for CV events. 30,32-36 Furthermore, analyses of statin trials have found that patients with elevated TG levels after statin therapy are at particularly high risk for CV events. Thus, even though statins (as well as ezetimibe and PCSK9 inhibitors) reduce TG levels along with lowering LDL-C, 37 some patients continue to have elevated TG levels, which is now recognized as a risk factor for CV events. These developments have re-invigorated research efforts to identify and develop strategies to reduce TG levels and TRLs. 38 Several diet and lifestyle interventions-including weight loss, physical activity, moderation of alcohol consumption, and Mediterranean-style diet-are recommended for reducing elevated TG levels. 17 The mainstays of TG-lowering pharmacotherapies had been fibrates and niacin, despite the fact that these agents had very little success in lowering CV risk. More recently, attention has turned to the omega-3 fatty acids eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), which are found in high concentrations in oily fishes and are known to reduce blood TG levels. 39,40 Indeed, low consumption of EPA and DHA has been associated with increased risk for CV disease in some observational studies. 40 Beyond their effects on TG levels, omega-3 PUFAs have numerous biological effects that may influence CV risk. Therefore, before discussing clinical trials of omega-3 PUFAs for CV risk reduction, we will briefly review the biology of EPA and DHA.

| EPA and DHA: TG lowering and beyond
Essential fatty acids are those fats that cannot be synthesized by humans; they must be consumed. The two essential fatty acids in the human diet are linoleic acid (LA; an omega-6 FA) and alpha-linolenic acid (ALA; an omega-3 PUFA). 41 LA is the precursor for arachidonic acid, and ALA is the precursor for EPA (and minimally DHA). 42 The main dietary sources of ALA are plants and plant products (such as vegetable oils). In theory, humans possess enzymes necessary to synthesize EPA and DHA from plant-based ALA. In reality, the production of EPA and DHA from ALA is inconsequential. Making matters worse, omega-6 PUFAs, which are often overly abundant in the western diet, dampen the conversion of ALA to EPA and DHA. 42,43 Thus, many experts consider EPA and DHA to be essential fatty acids. 43 Key dietary sources of EPA and DHA are oily fishes, 42 and several epidemiologic studies have found that groups of people who consume large amounts of such fish have lower rates of CV events and CV death compared with other populations (reviewed in Nishizaki et al. 44 ). Other studies, however, did not find a significant protective effect from consumption of a diet high in fish or EPA or DHA. 45 These studies varied widely with regard to the amount and types of PUFAs consumed by the study population. In 2006, a pooled analysis of prospective studies and randomized trials concluded that consumption of fish or fish oil was associated with a 36% reduction in the relative risk of death from coronary heart disease. 46 Thus, at that time, the prevailing medical opinion was that omega-3 PUFAs from fish had a favorable effect on CV outcomes.
Further support for a role of EPA and DHA in CV risk reduction came from biomarker studies. A pooled analysis of cohort studies conducted around the world found that higher levels of plasma or phospholipid omega-3 PUFAs (EPA, DHA, docosapentaenoic acid [DPA], and plant-derived ALA) were associated with a modest (approximately 10%) but significantly lower risk of fatal CHD. 47 When evaluating both observational and interventional studies of omega-3 PUFAs, it is important, however, to consider whether the investigators have distinguished ALA, EPA, and, DHA. We now understand that there are substantial differences in their biological activities. 48 Thus, studies that combine the omega-3 PUFAs may fail to discern important health effects that are mostly due to one particular omega-3 PUFA.
Studies have consistently found that fish-oil consumption, as well as consumption of purified EPA or DHA, robustly reduce serum TG levels. [49][50][51][52][53] ALA, in contrast, has only small and inconsistent effects on TG levels. 49,50 In most of the studies that found TG-lowering effects of fish oil (or EPA/DHA), participants received 1 g to >4 g of the respective fatty acid per day. 52 The TG lowering effect is linearly dose-dependent, and individuals with higher baseline TG levels exhibit greater declines in absolute TG levels in response to omega-3 PUFA consumption. 40 Reported biological actions of EPA and DHA extend well beyond TG lowering and include effects expected to be cardioprotective. 48,[54][55][56][57] As summarized by Mozaffarian and Rimm, 46 and shown in Figure 1, fish oil omega-3 PUFAs have been linked with reduced blood pressure and heart rate, as well as decreased platelet aggregation. 40,[58][59][60][61] They have also been linked with antiinflammatory actions and stabilization of coronary plaque. [62][63][64][65] Antidysrhythmic actions have been observed in animal or cell models, but their significance in humans is controversial. 40 As shown in Figure 1, some of the observed cardioprotective effects of omega-3 PUFAs require comparatively high doses of EPA/DHA; for example, lowering of diastolic blood pressure seems to require doses of ≥2 g per day, 59 and antithrombotic effects appear to require even higher doses. 66 It is plausible that other biological effects of omega-3 PUFAs-such as contributions to the mediation of inflammation 67,68 -may contribute to improvements in CV endpoints, although those relationships need extensive additional testing.

| RANDOMIZED CONTROLLED TRIALS AND META-ANALYSES
Large-scale, randomized trials have examined three classes of treatments targeting TG levels: fibrates, niacin, and omega-3 PUFAs. These trials have been thoroughly discussed by other recent reviews, 5,17,44,69 so we will briefly summarize the state of the field up to 2018. Then we will discuss very recent studies, including three trials of omega-3 PUFAs that were published in 2018.

| Fibrates and niacin
As noted by prior reviews, 5,70 trials of fibrates or niacin added to statin therapy have failed to show a benefit of the added treatment in terms of CV event rates. 12,25,26,71,72 The ongoing PROMINENT trial, involving the novel fibrate pemafibrate, 73 may provide insight into this question (NCT03071692).

| Fish oil omega-3 PUFAs: Background of clinical studies
A 2011 review analyzed randomized controlled trials, prospective cohort studies, and meta-analyses conducted up to that time and concluded that consumption of fish or fish oil reduced coronary heart disease mortality in populations with and without established CV disease. 40 Three meta-analyses of randomized controlled trials of omega-3 PUFA supplements supported this conclusion. 46,74,75 However, another meta-analysis of randomized controlled trials published in 2012 concluded that omega-3 PUFA supplementation was not associated with significant reduction in the risk of all-cause death, cardiac death, or myocardial infarction. 76 That meta-analysis, however, used an unusually high threshold for statistical significance (α = 0.0063). It is remarkable, for example, that all 17 of the trials included in the analysis of all-cause death reported a relative risk <0.96 favoring omega-3 PUFA, and seven of those trials reported upper confidence intervals that were < 1.0, yet the meta-analysis  46 used with permission failed to find a significant benefit of omega-3 PUFAs for all-cause death. The authors noted that there was extensive heterogeneity across trials with regard to event rates, baseline CV disease risk, treatment setting (eg, primary prevention vs secondary prevention), coadministered therapies, baseline intake of omega-3 PUFAs, dose, and nature of omega-3 PUFA used for the intervention, and treatment adherence. Thus, heterogeneity was likely so high that a meaningful conclusion could not be reached. A more recent meta-analysis revealed the importance of such heterogeneity. That study 77  In the ASCEND and VITAL trials, there were no significant differences between active treatment groups and the placebo groups with regard to rates of non-fatal serious adverse events, including bleeding. 4,78 In REDUCE-IT, icosapent ethyl was associated with a small but significant increase in rates of hospitalization for atrial fibrillation or flutter vs placebo (3.1% vs 2.1%; P = .004) and a trend toward increased rates of serious bleeding (2.7% vs 2.1%; P = .06). 1 However, prior reviews of bleeding risk associated with pharmacotherapeutic use of omega-3 PUFAs (at daily doses of 1-6 g) concluded that there was no effect on the risk of clinically significant bleeding. 79  secondary CV prevention patients who are most likely to benefit from omega-3 PUFA treatment.
Another salient distinction between REDUCE-IT and prior trials was the dose of omega-3 PUFA administered during the treatment period (4 g per day vs ≤850 mg per day in other large trials other than JELIS). The efficacy of high-dose omega-3 PUFA in REDUCE-IT is understandable in the context of JELIS and a recent meta-analysis, which found that higher intakes of omega-3 PUFAs were associated with greater risk reductions. 77  Finally, it must be noted that icosapent ethyl is a pure, esterified form of EPA, and it was used in two of the three successful trials summarized in Table 1 (JELIS and REDUCE-IT). Most other trials used formulations combining EPA and DHA in various concentrations. It is unknown to what degree these differences in drug formulation account for some of the inconsistencies in CV outcomes across trials.
Although EPA and DHA share many biological effects, some differences have been reported. 54 Although more detailed analyses are needed, the currently available analyses support the conclusion that TG lowering is not the sole mechanism responsible for lowering CV risk among the patients who participated in REDUCE-IT. A similar conclusion was found in a post hoc analysis of the JELIS trial. 85 Previous trials of icosapent ethyl in patients with high TG levels (≥200 mg/dL) found that treatment led to improvements in levels of several atherogenic lipid particles and biomarkers, including apolipoprotein B, very low density lipoprotein cholesterol, lipoproteinassociated phospholipase A2, remnant-like particle cholesterol, and apolipoprotein C-III. 53,86,87 Furthermore, preclinical studies have described several other beneficial effects of omega-3 PUFAs that could translate into reductions in rates of clinical events-effects such as improving endothelial function and nitric oxide availability, reducing inflammatory cytokines and enzymes, reducing activation of platelets, and contributing to plaque stabilization (summarized in Ganda et al. 5 ).

| CONCLUSION
The REDUCE-IT trial found that icosapent ethyl (4 g per day), a pure and stable form of the omega-3 PUFA EPA, significantly reduced residual risk of cardiovascular events in patients with ASCVD and elevated TG levels (≥150 mg/dL) who were receiving a stable dose of statin at study entry. On superficial examination, the REDUCE-IT trial appears to be an outlier among trials of omega-3 PUFAs for secondary prevention of CV events. On closer inspection, however, its design and outcomes are consistent with lessons learned from previous studies. Baseline TG levels appear to be a marker of increased likelihood of benefiting from icosapent ethyl treatment, and treatment significantly lowers TG levels. However, currently available analyses suggest that reduction in TG levels likely do not fully account for the beneficial effects of treatment. Despite persistent questions regarding its full mechanism of action, icosapent ethyl significantly ameliorates residual CV risk in patients with ASCVD receiving statin therapy and has been associated with minimal adverse effects. It appears, therefore, that with the results of the REDUCE-IT trial, we have rounded the corner; we now have in our sights the ability to further reduce the risk of CV events in appropriately selected patients with residual CV risk on optimal statin therapy.