Eicosapentaenoic acid therapy is associated with decreased coronary plaque instability assessed using optical frequency domain imaging

Abstract Background The relationship between eicosapentaenoic acid (EPA) therapy and coronary plaque stability assessed by optical frequency domain imaging (OFDI) has not been thoroughly described. Hypothesis EPA therapy is associated with decreased plaque instability in patients undergoing percutaneous coronary intervention (PCI) using OFDI. Methods Data on coronary artery plaques from 121 patients who consecutively underwent PCI between October 2015 and July 2018 were retrospectively analyzed. Of these patients, 109 were untreated (no‐EPA group), whereas 12 were treated with EPA (EPA group). Each plaque's morphological characteristics were analyzed using OFDI. Results We used 1:4 propensity score matching for patients who received or did not receive EPA therapy before PCI. Baseline characteristics were balanced between both groups (age, sex, body mass index, diabetes mellitus, hypertension, dyslipidemia, chronic kidney disease, smoking, previous PCI or coronary artery bypass grafting, previous myocardial infarction, prior statin use, acute coronary syndrome, hemoglobin A1c level, low‐density lipoprotein cholesterol concentration, triglyceride concentration, and high‐density lipoprotein cholesterol concentration). OFDI data from 60 patients were analyzed in this study. The EPA group had significantly lower mean lipid index (818 ± 806 vs 1574 ± 891) and macrophage grade (13.5 ± 5.9 vs 19.3 ± 7.4) but higher mean minimum fibrous cap thickness (109.2 ± 55.7 vs 81.6 ± 36.4 μm) than the no‐EPA group (P = 0.010, 0.019, and 0.040, respectively). Multiple logistic regression analyses showed that prior EPA use was independently associated with lower lipid index and macrophage grade (P = 0.043 and 0.024, respectively). Conclusion This OFDI analysis suggests that EPA therapy is associated with decreased plaque instability in patients undergoing PCI.


| INTRODUCTION
Eicosapentaenoic acid (EPA) is a member of a group of n−3 polyunsaturated fatty acids derived from fish oil. Epidemiologic data suggest that long-term intake of n−3 polyunsaturated fatty acids plays an important role in reducing adverse cardiovascular events. 1 Moreover, some clinical studies suggested that administration of purified EPA for secondary and primary prevention lowers the total and ischemic cardiovascular mortality. 2,3 Optical coherence tomography (OCT) and optical frequency domain imaging (OFDI) are helpful intravascular imaging modalities that use the reflection of near-infrared light to create images. Recently, many OCT and OFDI studies have reported the characteristics of unstable plaques, including plaque rupture, 4-6 lipidrich plaques, 7 thin-cap fibroatheroma (TCFA), 8,9 cholesterol crystals, 10,11 macrophage accumulation, 12,13 and microchannels. 8,14 Several coronary computed tomography studies have shown that coronary plaques had lower plaque burden in patients receiving EPA therapy than in those not receiving EPA therapy. 15,16 Furthermore, intravascular ultrasound studies have indicated smaller coronary atherosclerotic plaque or lipid volume in patients on EPA therapy than in those not on EPA therapy. 17,18 However, no study has comprehensively examined whether EPA therapy is associated with decreased plaque instability in patients undergoing percutaneous coronary intervention (PCI) using OFDI. This study aimed to investigate the association between EPA therapy and coronary plaque instability in patients undergoing PCI using OFDI. This study was approved by the ethics committee of Hokkaido Cardiovascular Hospital and was conducted in accordance with the ethical principles of the Declaration of Helsinki for medical research involving human subjects. The committee approved the image analysis, using only the data which were gotten when performing PCI.

| Coronary angiography
Coronary angiograms were analyzed by offline quantitative coronary angiography (GE ver. 5.10.1; Pie Medical Imaging BV, Maastricht, the Netherlands). Reference diameter, minimum lumen diameter, diameter stenosis, and lesion length were measured.

| OFDI and analysis
An OFDI catheter (FastView; Terumo Corporation, Tokyo, Japan) was advanced using a 0.014-in. guide wire with the help of a 6-or 7-Fr guiding catheter, and the imaging core was placed at the distal site of the lesion. Images obtained using OFDI were obtained using a continuous flush of contrast media at a rate of 4 mL/s, and the OFDI wire was pulled back at a rate of 20 to 40 mm/s. Although OFDI was generally performed without dilation using a balloon catheter, the lesion was dilated using a small-sized balloon if the OFDI catheter could not pass through the lesion because of severe stenosis. Aspiration thrombectomy was performed before OFDI for patients with acute coronary syndrome who did not show spontaneous recanalization. The plaque morphology of culprit lesions was studied. Following identification of the most stenotic lesion, 5-mm proximal and distal lesions (total length: 10 mm) were examined. Cross-sectional images were analyzed at every 1 mm for each OFDI parameter. OFDI analysis was performed by two independent investigators (Takao Konishi and R. K.) who were blinded to each patient's clinical course. In case of discordance between the investigators, a consensus reading was performed.

| OFDI definitions
OFDI analysis indicated the presence of plaque rupture, plaque erosion, calcified nodule, lipid-rich plaque, TCFA, macrophage accumulation, microvessels, cholesterol crystals, and calcification in the plaque. Plaque rupture was defined as the presence of fibrous cap discontinuity leading to a communication between the necrotic core and lumen ( Figure 1A). 19,20 Plaque erosion was defined as a lesion with a thrombus but without fibrous cap disruption. 21 Thrombus was defined as a well-delineated, high-signal mass attached to the luminal surface or floating within the lumen. 20 Lipid-rich plaques were defined as lesions with a lipid arc of more than 180 ( Figure 1B). 20 Lipid arc was measured within a lipid-rich plaque, and the maximum value was recorded ( Figure 1B). Lipidcore length was defined as the length of lipid plaque and was measured in the longitudinal view. Lipid index was defined as the maximum lipid arc multiplied by the lipid-core length. 22 Cholesterol crystals were defined as thin, linear regions of high signal intensity within the lipid plaque, without backscattering. 20 TCFA was defined as an atheroma with a fibrous cap thickness (FCT) < 65 μm. 20 FCT was defined as the minimum thickness of a signalrich layer from the coronary artery lumen to the inner border of the underlying lipid in the culprit lesion ( Figure 1C). 20 Macrophage accumulation was defined as increased signal intensity within the fibrous cap, accompanied by heterogeneous backward shadows ( Figure 1D). 20 Semiquantification of macrophage accumulation was performed as follows based on axial and circumferential distribution: grade 0, no macrophage; grade 1, localized macrophage accumulation; grade 2, clustered accumulation in <1 quadrant; grade 3, clustered accumulation in ≥1 quadrant and < 3 quadrants; and grade 4, clustered accumulation in ≥3 quadrants (Supporting Information Figure S1). 23 The possible range of macrophage score was 0-40, which is a summation of grades 0-4 across all 10 slices.
Microchannels were defined as small vesicular or tubular structures with diameters of 50-300 μm within the intima ( Figure 1E). 20 The number of microchannels was counted across all 10 slices at every 1 mm. Calcification was defined as well-delineated, lowbackscattering heterogeneous regions ( Figure 1F). 20

| Statistical analysis
Propensity score-matched analysis was performed to compare the two groups. 24 The propensity of being in the EPA group was estimated using a logistic regression model with the following covariates in our study database: age, sex, body mass index, diabe-  F I G U R E 1 Representative coronary artery plaque images for optical frequency domain imaging (OFDI) analyses. A, Plaque rupture (white arrow) with fibrous cap discontinuity, leading to a communication between the LC and the lumen. LC, lipid core. B, Measurement of lipid arc in a lipid-rich plaque. Maximum lipid arc was measured within a lipid-rich plaque (yellow lines). C, Measurement of fibrous cap thickness (FCT). Minimum FCT was measured thrice at the thinnest part, and the three measurements were subsequently averaged (white arrows). D, Macrophage accumulation was defined as increased signal intensity within the fibrous cap, accompanied by heterogeneous backward shadows (arrowheads). E, Microchannels (yellow arrows) were defined as small vesicular or tubular structures with diameters of 50 to 300 μm within the intima. F, Calcification was defined as well-delineated, low-backscattering heterogeneous regions 3 | RESULTS

| Clinical characteristics
After the propensity score matching, OFDI images from 60 patients were analyzed in this study. The clinical characteristics of the two patient groups are compared in Table 1. Although EPA/arachidonic acid ratios were available only in 31 patients (10 in EPA group and 21 in no-EPA group, the mean EPA/arachidonic acid ratio was significantly higher in the EPA group than in the no-EPA group (1.63 ± 0.46 vs 0.48 ± 0.21, P < 0.001). The mean LDL cholesterol concentration was 94 ± 29 mg/dL in the no-EPA group and 97 ± 36 mg/dL in the EPA group (P = 0.807). Furthermore, the mean triglyceride concentration was 161 ± 92 mg/dL in the no-EPA group and 153 ± 114 mg/dL in the EPA group (P = 0.803). Other characteristics including medications and concomitant diseases were similar in both groups. The type and doses of EPAs administered in 12 patients in the EPA group were as follows: ethyl icosapentate, 1800 mg/day (n = 5); ethyl icosapentate, 2700 mg/day (n = 1); and omega-3-acid ethyl esters, 2 g/day (n = 6).

| Angiographic findings
Plaque location and angiographic data are shown in Table 2. Diameter stenosis was similar between the no-EPA and EPA groups (85.9 ± 11.5 vs 86.4 ± 16.5, P = 0.885). No significant differences in plaque location and other angiographic data were observed between both groups.

| Plaque characteristics assessed using OFDI
The results of the qualitative and semiquantitative analysis of various coronary plaque characteristics using OFDI are compared in Table 3.

| Multiple logistic regression analyses for lipid index, TCFA, and macrophage invasion
Multiple logistic regression analyses were performed to assess the risk factors for lipid index, TCFA, and macrophage invasion (n = 60).
Because LDL cholesterol to HDL cholesterol ratio was negatively cor-  (Table 4), whereas HDL cholesterol concentration was associated with TCFA (Table 5). Moreover, CKD and prior EPA use were associated with macrophage infiltration (Table 6).

| Correlations of lipid index, TCFA, and microchannels with macrophage grade
Significant correlations were observed between lipid index, FCT, and macrophage grade. Lipid index was positively correlated with macrophage grade (r = 0.803, P < 0.001; Figure S2). In contrast, minimum FCT was negatively correlated with macrophage grade (r = −0.650, P < 0.001; Figure S3). The correlation between macrophage grade and the number of microchannels was not statistically significant (r = 0.096, P = 0.467). However, in 31 patients in whom the maximum lipid arc was <270 , a weak but positive correlation between macrophage grade and the number of microchannels was observed (r = 0.421, P = 0.018; Figure S4).

| Observer variabilities
Images obtained using OFDI were analyzed by two independent observers. The inter-observer reliabilities and intra-observer repro-  This study suggests that EPA therapy itself is effective for coronary plaque stabilization. As shown in Table 3, we observed that patients who received EPA therapy had lower lipid burden, higher FCT, and less macrophage accumulation than those who did not receive EPA therapy (P = 0.010, 0.040, 0.019, respectively). Nonetheless, patient background (including mean LDL cholesterol and triglyceride concentrations), except for EPA/arachidonic acid ratio, was not statistically different between the two groups (P = 0.803, Table 1).
Watanabe et al. showed that lipid volume and plaque volume reductions with EPA therapy were independent of decreases in LDL cholesterol and triglyceride concentrations, 17 which is consistent with our results.

| Lipid-rich plaque
An important mechanism of plaque rupture is a large lipid core, which mechanically enhances the tension of fibrous cap covering the lipid  26 Ferguson et al.
reported that EPA attenuated the inflammatory activation of in vitro human adipocytes and reduced lipogenesis. 27 In the process of atherosclerotic development, lipid-core enlargement is accelerated by apoptotic macrophage accumulation and elevated chemokine expression followed by intimal recruitment of circulating monocytes. 28 In the present study, macrophage grade was significantly lower in the EPA group than in the no-EPA group (Table 3) and was positively correlated with lipid index ( Figure S2). Therefore, less macrophage accumulation might contribute to decreased lipid core in the EPA group.

| FCT
TCFA is one of the most important characteristics of unstable plaques in the coronary and carotid arteries. 29,30 Several mechanisms could elucidate the higher FCT in the EPA group than in the no-EPA group. First, EPA therapy inhibits the ability of macrophages to secrete matrix metalloproteinase (MMP) and monocyte chemotactic protein (MCP)-1. 31 As shown in Figure S3, minimum FCT was inversely correlated with macrophage grade. Because the MMPs released by macrophages induce the thinning of fibrous caps of atherosclerotic plaques via collagen breakdown, 32 less macrophage accumulation in the plaque might contribute to the higher FCT in the EPA group (Table 3). Second, EPA lowered the levels of pentraxin 3, an inflammatory marker associated with TCFA. 33 Although the mean minimum FCT was higher in the EPA group than in the no-EPA group (Table 3), EPA therapy was not an independent predictor of TCFA, probably because of the small sample size (Table 5).

| Macrophage accumulation
There are several mechanisms for reduced macrophage accumulation by EPA. First, EPA is incorporated into atherosclerotic plaques, and a higher EPA content in the plaques is associated with a reduced number of foam cells and T cells and less inflammation, leading to increased stability. 36 Second, Yamada et al reported that EPA acts as an anti-inflammatory agent that reduces the production of pro-inflammatory eicosanoid mediators from arachidonic acid and decreases adhesion molecule expression. 37 Furthermore, they showed the in vivo and in vitro inhibition of monocyte adhesion to aortic endothelial cells by EPA. Third, monocyte recruitment into atherosclerotic plaques is also reduced by EPA via favorable alteration of monocyte subsets independent of effects on plasma cholesterol. 38 Fourth, macrophages extravasating from intraplaque microvessels might be inhibited by EPA via reduction in vascular endothelial growth factor and Flk-1 receptor expression. 39,40 In the present study, the mean number of microchannels was lower in the EPA group than in the no-EPA group, albeit without statistically significant difference (Table 3). These mechanisms might explain the association between EPA therapy and lower macrophage grade.
T A B L E 2 Angiographic and procedural characteristics

| Other observations
Prior statin use was not significantly associated with plaque stability in the current study (Tables 4-6). JCAD study showed that if three or more major coronary risk factors are present, the risk for cardiovascular events is 1.3-fold higher compared to having two or fewer risk factors. 41

| Study limitations
The present study has some limitations. First, this retrospective, cross-sectional study had a small sample size and was conducted at a single medical center within Japan, where people generally have a higher omega-3 dietary intake in comparison to many other countries.
Its results need to be confirmed by a study that enrolls a larger number of patients. Second, there is an inherent discrepancy between characteristics assessed using OFDI and real histopathological findings, as shown in a previous comparison study. 45

| CONCLUSIONS
This OFDI analysis suggests that EPA therapy is associated with decreased plaque instability in patients with CAD undergoing PCI.
Patients with CAD who are at high risk for cardiovascular events should receive EPA therapy as intensive medical management for the stabilization of coronary atherosclerotic plaques.