Effect of Eicosapentaenoic Acid/Docosahexaenoic Acid on Coronary High-Intensity Plaques Detected Using Noncontrast T1-weighted Imaging: The AQUAMARINE EPA/DHA Randomized Study

Kazuhiro Nakao; Teruo Noguchi; Hiroyuki Miura; Yasuhide Asaumi; Yoshiaki Morita; Satoshi Takeuchi; Hideo Matama; Keniciro Sawada; Takahito Doi; Hayato Hosoda; Takahiro Nakashima; Satoshi Honda; Masashi Fujino; Shuichi Yoneda; Shoji Kawakami; Toshiyuki Nagai; Kensaku Nishihira; Tomoaki Kanaya; Fumiyuki Otsuka; Michio Nakanishi; Yu Kataoka; Yoshio Tahara; Yoichi Goto; Kengo Kusano; Haruko Yamamoto; Katsuhiro Omae; Hisao Ogawa; Satoshi Yasuda

doi:10.5551/jat.64063

Abstract

Aim: Omega-3 fatty acids have emerged as a new option for controlling the residual risk for coronary artery disease (CAD) in the statin era. Eicosapentaenoic acid (EPA) is associated with reduced CAD risk in the Reduction of Cardiovascular Events with Icosapent Ethyl-Intervention trial, whereas the Statin Residual Risk with Epanova in High Cardiovascular Risk Patients with Hypertriglyceridemia trial that used the combination EPA/docosahexaenoic acid (DHA) has failed to derive any clinical benefit. These contradictory results raise important questions about whether investigating the antiatherosclerotic effect of omega-3 fatty acids could help to understand their significance for CAD-risk reduction.

Methods: The Attempts at Plaque Vulnerability Quantification with Magnetic Resonance Imaging Using Noncontrast T1-weighted Technic EPA/DHA study is a single-center, triple-arm, randomized, controlled, open-label trial used to investigate the effect of EPA/DHA on high-risk coronary plaques after 12 months of treatment, detected using cardiac magnetic resonance (CMR) in patients with CAD receiving statin therapy. Eligible patients were randomly assigned to no-treatment, 2-g/day, and 4-g/day EPA/DHA groups. The primary endpoint was the change in the plaque-to-myocardium signal intensity ratio (PMR) of coronary high-intensity plaques detected by CMR. Coronary plaque assessment using computed tomography angiography (CTA) was also investigated.

Results: Overall, 84 patients (mean age: 68.2 years, male: 85%) who achieved low-density lipoprotein cholesterol levels of ＜100 mg/dL were enrolled. The PMR was reduced in each group over 12 months. There were no significant differences in PMR changes among the three groups in the primary analysis or analysis including total lesions. The changes in CTA parameters, including indexes for detecting high-risk features, also did not differ.

Conclusion: The EPA/DHA therapy of 2 or 4 g/day did not significantly improve the high-risk features of coronary atherosclerotic plaques evaluated using CMR under statin therapy.

See editorial vol. 31: 117-118

Introduction

Available evidence suggests that lowering low-density lipoprotein (LDL) cholesterol levels with statins reduces the risk of cardiovascular disease¹⁾. However, many clinical trials have reported a significant residual risk of cardiovascular events, even in the setting of optimal LDL cholesterol reduction with statins²⁾. Thus, establishing strategies beyond lowering LDL cholesterol levels with statins to reduce the risk of cardiovascular events is needed.

The Japan EPA Lipid Intervention Study (JELIS) showed the efficacy of 1.8 g of eicosapentaenoic acid (EPA) in preventing coronary artery disease (CAD) in patients with hypercholesterolemia receiving statins³⁾. The Reduction of Cardiovascular Events with Icosapent Ethyl-Intervention (REDUCE-IT) Trial demonstrated that icosapent ethyl (high-purity prescription form of EPA; 4 g/day) reduced cardiovascular events⁴⁾. In contrast, the Long-Term Outcomes Study to Assess Statin Residual Risk with Epanova in High Cardiovascular Risk Patients with Hypertriglyceridemia (STRENGTH) demonstrated that 4 g of carboxylic acid formulation of EPA and docosahexaenoic acid (DHA) therapies resulted in no significant difference in a composite outcome of major adverse cardiovascular events compared to control⁵⁾. Considering these contradictory results, investigating the antiatherosclerotic effect of omega-3 fatty acids could help to understand their significance for CAD-risk reduction.

We previously reported that coronary high-intensity plaques detected using noncontrast T1-weighted (T1WI) cardiac magnetic resonance (CMR) imaging were associated with computed tomography angiography (CTA), intravascular ultrasound (IVUS), and optical coherence tomographic (OCT)-derived features of high-risk plaques^{6, 7)}. High-intensity plaques were able to be quantitatively assessed using the plaque-to-myocardium signal intensity ratio (PMR)⁶⁾, and their existence was significantly associated with future coronary events⁸⁾. In the Attempts at Plaque Vulnerability Quantification with Magnetic Resonance Imaging Using Noncontrast T1-weighted Technic (AQUAMARINE) study, intensive lipid–lowering therapy with pitavastatin reduced the PMR by approximately 25%⁹⁾. However, coronary high-intensity plaques remained positive at PMR ≥ 1.0, and this indicates the statin-residual risk on CMR. Therefore, this study aimed to investigate the effect of EPA/DHA by examining changes in the PMR of coronary high-intensity plaques using CMR in patients with CAD on statin therapy.

Methods

Study Design

The study design and rationale for the AQUAMARINE EPA/DHA study have been previously published¹⁰⁾. In brief, this was a single-center, triple-arm, randomized, controlled, open-label trial that examined the effect of 12 months of EPA/DHA (Lotriga^®; Takeda Pharmaceutical Co. Ltd., Osaka, Japan) on coronary high-intensity plaques detected using CMR in patients with CAD treated with statins. A 1-g EPA/DHA formulation comprised over 90% of N−3 PUFAs, including EPA (465 mg) and DHA (375 mg). Eligible participants were randomly assigned to the 2-g/day EPA/DHA group, 4 g/day EPA/DHA group, and no- treatment (statin only) group (allocation rate 1:1:1). Dynamic allocation was performed at the registration/allocation center at the start of therapy with respect to age, gender, and the presence or absence of type II diabetes. The allocation table and the most recent breakdown of allocated subjects in each group were used for dynamic allocation at the registration/allocation center. Data collection was performed at the National Cerebral and Cardiovascular Center.

Study Population

Patients aged ≥ 20 years with CAD who were receiving statin therapy were eligible for the study. CAD was diagnosed as having ＞25% coronary stenosis as measured using CTA or coronary angiography. Participants had at least one coronary plaque with PMR ≥ 1.0 on noncontrast T1WI during the screening examination. It was necessary for eligible participants to achieve optimal LDL-C levels (LDL-C ＜100 mg/dL) under statin treatment. Patients were excluded if they were scheduled for revascularization.

Study Outcomes

The primary endpoint was prespecified as the change in the PMR of the primary coronary lesion measured using noncontrast T1W cardiac imaging from baseline (pretreatment) to follow-up (posttreatment). We evaluated several other endpoints: change in the PMR of all lesions; absolute change and percent change in Hounsfield units (HU), plaque volume, vessel area, and plaque area of the primary lesion or all lesions measured using CTA from baseline to follow-up; and change in lipid parameters.

CMR Scanning and Analysis

Noncontrast T1WI was performed on a 3 tesla magnetic resonance imaging system with a 32-channel cardiac coil (MAGNETOM Verio; Siemens AG Healthcare Sector, Erlangen, Germany). The procedures used to acquire MR images in our study have been previously described¹¹⁾. Briefly, coronary plaque imaging was performed using an inversion recovery–prepared 3D T1W turbo-fast low-angle shot sequence with an electrocardiographic trigger, navigator-gated free breathing, and fat suppression. Trans-axial sections covered the entire heart (inversion time, 650 ms; field of view, 280×228 mm; acquisition matrix, 256×187; reconstruction matrix, 512×374; acquisition slice thickness, 1.0 mm; reconstruction spatial resolution, 0.6×0.5×0.6 mm; repetition time/echo, 4.7 ms/2.13 ms; flip angle, 12°; Generalized Autocalibrating Partial Parallel Acquisition factor, 2; navigator gating window, ±1.5–2.5 mm; data acquisition window, 84–120 ms). The trigger delay and acquisition window were based on the duration of minimal right coronary artery motion as determined using cine MR imaging. The PMR was defined as the ratio of the signal intensity of the coronary plaque to the signal intensity of the nearby left ventricular myocardium. The highest signal intensity detected in each plaque was considered as PMR value for that plaque in the segment-based analysis. In the patient-based analysis, the highest PMR among the coronary plaques was defined as the PMR for that participant. A coronary plaque with PMR ≥ 1.0 was defined as a high-intensity plaque. To confirm that the location of an observed high-intensity plaque corresponded to the presence of a coronary plaque, crosssectional and reconstructed curved multiplanar CTA images were used^7-9). All imaging, including CMR and CTA measurements, was analyzed by physicians who were blinded to the patient’s treatment status.

CTA Scanning and Analysis

Coronary CTA was performed using a dual-source computed tomography (CT) scanner (SOMATOM Definition Flash; Siemens Healthcare, Erlangen, Germany). Coronary images were acquired using a 128×0.6-mm slice collimation, 280-ms gantry rotation time, 120-kV tube voltage, and 280-mAs quality reference current–time product. A retrospective electrocardiogram (ECG)-gated spiral scan with tube current modulation or a prospective ECG-triggered high-pitch spiral scan was selected depending on the heart rate. The images were reconstructed with a 0.6-mm slice thickness in 0.3-mm increments with a medium-smooth convolution kernel (B26f). Quantitative analysis was performed using software that facilitates plaque volume measurement (Ziostation2; Ziosoft, Tokyo, Japan). The parameters assessed included (1) HU, (2) plaque volume, (3) low-attenuation plaque (LAP) volume, (4) vessel area, and (5) plaque area. The fixed HU cutoff values that were used for the classification were ＜30 for LAP, 30–150 for intermediate attenuation plaques, and ≥ 500 for calcified plaques. The CT density of the plaques for at least three points was assessed, and the average was recorded.

Statistical Analyses

Based on the results from our previous study⁹⁾, a clinically meaningful change from baseline in the PMR was assumed to be −0.15 in the 2- and 4-g EPA/DHA groups, and no change was observed in the no-treatment group, with a common standard deviation of 0.2 among the groups. The sample size per group of 46 achieved 90% power to detect the assumed difference in changes using a two-group, one-sided t-test with a 1.25% significance level, where the overall significance level of the one-sided test for two comparisons was 2.5%. Based on this estimation, 50 participants per group were planned to be recruited. Patient enrollment was slow, with fewer patients meeting the inclusion criteria than expected; after the extension of the enrollment period, the enrollment was prematurely terminated on December 31, 2017, after enrolling 84 patients in 43 months. The analysis was based on an intent-to-treat basis. Continuous variables are presented as mean±standard deviation for normally distributed variables; they were compared using analysis of variance. Nonnormally distributed variables are presented as medians (interquartile ranges). They were compared using the Kruskal–Wallis test. Categorical baseline variables were compared using the Fisher’s exact test or chi-square test, as appropriate. Testing for significant differences in each parameter between baseline and the 12-month follow-up was performed using the paired Student’s t-test or Wilcoxon signed-rank sum test. This test was performed only on participants whose pre- and postvalues were available. A one-way analysis of variance or the nonparametric Kruskal–Wallis test was used to assess the difference in change among three counties based on the distribution of the variables. All analyses were performed using JMP statistics, version 11 (SAS Institute Inc., Cary, NC, USA) and STATA15 (Stata Corp. LLC, College Station, TX, USA).

Results

Patients’ Characteristics

Between May 2014 and December 2017, 84 patients were randomly assigned to the no-treatment, 2-g EPA/DHA, and 4-g EPA/DHA groups. Excluding 3 cases (2 cases of consent withdrawal and 1 case of protocol violation), 81 cases were analyzed. Of the 27, 28, and 26 patients assigned to the no-treatment, 2-g, and 4-g groups, respectively, 26, 24, and 25 patients completed the 12-month follow-up, respectively (Supplementary Fig.1). The number of serial examinations that had sufficient quality to be evaluated was 26, 21, and 24 in CMR and 24, 20, and 23 in CTA. The characteristics of patients at baseline were well balanced among the trial groups (Table 1). Three patients developed stable CAD, one developed small intestinal bleeding, and one developed thalamic bleeding in the EPA/DHA group during the study period. The relationship between these diseases and drug administration cannot be completely denied.

Supplementary Fig.1.

This chart shows the enrollment, randomization, and exclusion of participants.

DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid

Table 1. Patients’ characteristics

Baseline variable	No-treatment group (n = 27)	2 g EPA/DHA group (n = 28)	4 g EPA/DHA group (n = 26)	P-value
Age, years^＊	63.8 (10.5)	67.6 (9.1)	69.0 (8.1)	0.855
Male, n (%)	23 (85.2%)	24 (85.7%)	33 (84.6%)	0.994
Current smoker, n (%)	7 (26%)	4 (14.2%)	3 (11.5%)	0.365
BMI, kg/m^{2 ＊}	25.0±3.2	24.4±3.5	23.5±3.4	0.240
Hypertension, n (%)	18 (66.7%)	13 (46.4%)	15 (57.7%)	0.316
Diabetes, n (%)	9 (33.3%)	10 (35.7%)	8 (31.0%)	0.929
Previous ischemic heart disease, n (%)	26 (96.3%)	25 (89.3%)	35 (96.2%)	0.467
Medications
Aspirin, n (%)	23 (85.2%)	23 (82.1%)	25 (96.2%)	0.538
Beta-blockers, n (%)	19 (70.4%)	17 (60.7%)	18 (69.2%)	0.708
Statins, n (%)	27 (100%)	28 (100%)	26 (100%)	-
ACE-I or ARB, n (%)	17 (63.0%)	16 (57.1%)	15 (57.7%)	0.890

^＊Values are presented as the mean±standard deviation, n (%).

ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; BMI, body mass index

Changes in Lipid Profile and Other Laboratory Parameters

The changes in lipid profile, fatty acid fraction, and other laboratory parameters at baseline and the 12-month follow-up are presented in Table 2. Serum LDL concentrations in the no-treatment, 2-g EPA/DHA, and 4-g EPA/DHA groups were lower than 80 mg/dL and showed no significant difference. No significant difference was observed in Δ (changes from baseline to follow-up) LDL among the groups. Serum triglyceride concentrations among the three groups were 112.8±42.4, 142.0±134.8, and 123.4±71.5 mg/dL, respectively. There was no difference in triglyceride levels between baseline and follow-up in each group, and no significant difference was observed in the amount of change between the groups. Serum EPA concentrations did not change in the no-treatment group, whereas they increased in the 2- and 4-g groups. ΔEPA levels were different for each group: 2.2±31.8, 81.2±58.0, and 126.3±55.9 mg/dL, respectively. ΔDHA levels were −7.5±36.0, 35.3±66.6, and 30.6±62.7 mg/dL in the no-treatment, 2-g, and 4-g groups, respectively, and showed significant differences.

Table 2. Baseline, Follow-up and Changes in Lipid Profiles and other parameters

	No-treatment group			2g EPA/DHA group
	B/L	F/U	p Value	B/L	F/U	p Value
TC, mg/dL	136.4±20.1	141.7±27.4	0.087	140.1±28.8	150.8±30.9	0.017
LDL, mg/dL	71.0±13.7	73.4±18.0	0.335	69.9±18.7	70.7±22.1	0.729
HDL, mg/dL	41.8±12.5	45.5±11.8	0.015	46.5±13.8	48.8±11.8	0.104
TG, mg/dL	112.8±42.4	114.9±76.9	0.863	142.0±134.8	198.8±276.8	0.107
RLP-CHO, mg/dL	4.9±2.9	4.6±2.4	0.583	5.8±5.3	8.1±13.8	0.215
LPL, mg/dL	49.3±17.1	57.0±16.5	0.007	49.1±21.5	52.0±17.6	0.278
EPA, μg/ml	57.1±28.6	59.3±25.7	0.728	60.1±30.3	141.3±64.5	＜0.001
DHA, μg/ml	118.9±43.0	111.3±28.7	0.295	125.4±51.8	160.8±73.7	0.016
AA, μg/ml	184.0±57.4	181.3±54.8	0.659	198.9±55.0	191.6±68.5	0.385
EPA/AA ratio	0.33±0.17	0.35±0.17	0.638	0.32±0.18	0.80±0.36	＜0.001
hs-CRP, mg/dL	0.14±0.16	0.09±0.09	0.104	0.08±0.10	0.08±0.11	0.841
Cre, mg/dL	0.92±0.24	0.91±0.19	0.753	0.87±0.16	0.88±0.16	0.523
Leptin, ng/ml	10.4±8.2	13.7±9.7	＜0.001	10.0±2.0	14.0±10.4	0.004
Adiponectin, μg/mL	8.0±4.1	8.9±5.3	0.069	6.9±3.2	6.5±3.0	0.191
	4g EPA/DHA group			Changes form B/L to F/U
	B/L	F/U	p Value	No-treatment group	2g EPA/DHA group	4g EPA/DHA group	p Value
TC, mg/dL	146.0±17.2	142.1±16.9	0.209	5.3±15.2	10.1±19.4	−3.9±15.0	0.014
LDL, mg/dL	74.2±12.7	69.3±16.5	0.0498	2.4±12.6	0.8±11.6	−4.9±11.8	0.083
HDL, mg/dL	47.3±11.0	46.3±11.0	0.567	3.7±7.3	2.3±6.6	−1.0±8.6	0.081
TG, mg/dL	123.4±71.5	117.2±74.4	0.742	2.1±61.7	56.8±158.1	−6.2±92.5	0.109
RLP-CHO, mg/dL	5.4±3.0	4.8±4.2	0.512	-0.3±2.7	2.3±8.9	−0.6±4.5	0.173
LPL, mg/dL	58.0±52.4	56.4±52.4	0.867	7.7±13.4	2.9±12.7	−1.7±48.3	0.537
EPA, μg/ml	75.0±48.4	194.0±67.3	＜0.001	2.2±31.8	81.2±58.0	126.3±55.9	＜0.001
DHA, μg/ml	135.7±52.2	166.3±65.8	0.026	−7.5±36.0	35.3±66.6	30.6±62.7	0.017
AA, μg/ml	202.4±51.6	160.6±24.6	＜0.001	−2.7±30.3	−7.3±40.4	−41.8±42.7	＜0.001
EPA/AA ratio	0.39±0.25	1.21±0.39	＜0.001	0.02±0.22	0.48±0.29	0.87±0.25	＜0.001
hs-CRP, mg/dL	0.06±0.06	0.08±0.12	0.536	−0.05±0.14	0.01±0.14	0.02±0.14	0.226
Cre, mg/dL	0.89±0.17	0.89±0.17	0.750	−0.01±0.17	−0.01±0.11	−0.01±0.14	0.989
Leptin, ng/ml	8.1±3.6	9.7±4.6	0.055	3.3±4.4	4.1±6.2	1.6±3.8	0.211
Adiponectin, μg/mL	8.3±4.2	9.1±6.1	0.252	0.9±2.5	−0.4±1.3	0.8±3.2	0.142

Values are presented as the mean±standard deviation.

DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid; hs-CRP, high-sensitivity C-reactive protein; HDL, high-density lipoprotein; LDL, low- density lipoprotein; RLP-CHO, remnant-like particle cholesterol; LPL, lipoprotein lipase; TC, total cholesterol; TG, triglycerides. B/L, baseline; F/ U, 12-month follow-up.

Results of CMR and CTA Analysis

The PMR was positively correlated with plaque volume detected by CTA (Spearman correlation coefficient r=0.36, P＜0.01) (Supplementary Fig.2). The changes in the PMR of the primary lesion from baseline to the 12-month follow-up are presented in Fig.1a–c. After 12 months, the PMR significantly decreased in all groups. In the primary lesion analysis, ΔPMR from baseline to follow-up showed no difference among the three groups (Fig.1d). %ΔPMR also showed no difference (Supplementary Table 1). In the total lesion analysis (Fig.2a–c) evaluating all lesions of high-intensity plaques, including the primary lesion, the PMR was significantly lower at follow-up than at baseline in all groups. However, there were no differences in ΔPMR or %ΔPMR among the three groups (Fig.2d and Supplementary Table 1).

Supplementary Fig.2.

Association of PMR and Plaque volume evaluated CTA (Spearman correlation coefficient)

Fig.1. Comparisons of change in the PMR (primary lesion analysis)

Primary lesion analysis was performed on lesions defined as the primary lesion with the highest PMR in each patient. Individual changes in the PMR from baseline to 12 months are shown in the no-treatment (A), 2-g EPA/DHA (B), and 4-g EPA/DHA groups (C). Each line represents one patient. The comparisons of the group data (D) show that there were no significant differences in absolute changes in the PMR in the primary lesion analysis.

B/L, baseline; F/U, 12-month follow-up; PMR, plaque-to-myocardium signal intensity ratio

Supplementary Table 1. Difference of change in %ΔPMR from baseline to follow-up

	No treatment group (Primary lesion N = 26, Total lesion N = 59)	2g EPA/DHA group (Primary lesion N = 21, Total lesion N = 48)	4g EPA/DHA group (Primary lesion N = 24, Total lesion N = 60)	p-value
Primary lesion
%ΔPMR	-19.3 (-30.8, -1.8)	-19.2 (-33.3, 4.9)	-19.1 (-31.0, -3.8)	0.912
Total lesion
%ΔPMR	-7.8 (-24.4, 8.5)	-5.2 (-24.9, 5.4)	-12.0 (-24.7, 5.6)	0.901

Values are medians (interquartile range).

Delta (Δ) means a change from the baseline to 12-month follow-up.

PMR, plaque-to-myocardium signal intensity ratio

Fig.2. Comparisons of change in the PMR (total lesion analysis)

A total lesion analysis was performed on all lesions with coronary artery stenosis detected at ＞25%. Individual changes in the PMR from baseline to 12 months are shown in the no-treatment (A), 2-g EPA/DHA (B), and 4-g EPA/DHA groups (C). Each line represents a lesion. The comparisons of the group data (D) show that there were no significant differences in absolute changes in the PMR in the total lesion analysis.

In the CTA plaque assessment, no significant change was detected in each parameter (HU, plaque volume, LAP volume, vessel area, and plaque) from baseline to 12 months in each group of primary lesions. No significant difference in absolute change or percentage (%) change of these atherosclerosis indexes between the groups was detected. In the total lesion analysis, a significant difference was found only in the %vessel area (Table 3 and Supplementary Table 2).

Table 3. CTA parameters between Baseline and Follow-up in each group

	No-treatment group			2 g EPA/DHA group
	B/L	F/U	P-value	B/L	F/U	P-value
Primary lesion	N= 24			N= 20
HU^＊	49.1 (17.7, 73.7)	51.0 (30.9, 81.5)	0.693	44.6 (28.0, 80.8)	60.2 (34.4, 74.4)	0.701
Plaque volume, mm³	163.8 (89.0, 254.4)	154.4 (103.5, 213.6)	0.317	122.1 (74.2, 190.8)	121.7 (100.4, 200.3)	0.737
LAP Volume, mm³	32.2 (28.3, 60.9)	35.5 (25.5, 57.0)	0.170	27.8 (15.6, 41.2)	22.5 (13.4, 37.5)	0.296
vessel area, mm²	19.1 (14.0, 23.9)	18.4 (16.2, 25.7)	0.271	14.1 (11.8, 17.5)	13.6 (11.5, 17.7)	0.455
plaque area, mm²	11.7 (7.9, 17.6)	12.0 (8.8, 15.8)	0.954	9.1 (5.8, 12.1)	8.6 (6.5, 10.3)	0.433
Total lesion	N= 56			N= 46
HU^＊＊	59.9 (34.4, 79.0)	67.8 (35.6, 91.3)	0.018	58.9 (43.4, 86.8)	60.5 (42.3, 83.2)	0.611
Plaque volume, mm³	147.6 (89.0, 223.5)	147.7 (99.8, 207.5)	0.257	141.3 (91.1, 225.5)	143.5 (112.5, 213.2)	0.814
LAP Volume, mm³	31.1 (21.6, 49.3)	29.5 (18.0, 39.8)	0.048	28.9 (18.8, 47.3)	28.3 (17.0, 41.3)	0.291
vessel area, mm²	16.8 (11.8, 21.8)	17.0 (12.5, 22.7)	0.079	16.3 (12.6, 21.0)	15.4 (12.6, 20.8)	0.397
plaque area, mm²	8.8 (7.2, 15.3)	9.9 (7.6, 14.9)	0.955	10.5 (7.2, 12.6)	9.2 (6.8, 12.8)	0.505
	4 g EPA/DHA group			Changes form B/L to F/U
	B/L	F/U	P-value	No-treatment group	2g EPA/DHA group	4g EPA/DHA group	p Value
Primary lesion	N= 23
HU^＊	57.6 (35.6, 87.3)	68.4 (43.5, 95.4)	0.191	0.2 (-13.9, 17.3)	2.9 (-23.1, 29,3)	10.6 (-9.2, 25.5)	0.811
Plaque volume, mm³	105.6 (79.1, 184.6)	95.8 (73.4, 188.0)	0.447	-5.3 (-17.3, 12.6)	-5.6 (-20.0, 18.3)	1.65 (-10.0, 3.9)	0.818
LAP Volume, mm³	26.2 (12.6, 44.0)	27.7 (10.7, 44.9)	0.761	-2.2 (-8.6, 3.2)	-1.1 (-9.4, 3.3)	-1.4 (-5.0, 4.2)	0.694
vessel area, mm²	12.2 (8.5, 16.4)	12.1 (8.3, 17.0)	0.513	0.0 (-1.6, 3.8)	-0.1 (-3.1, 1.7)	-0.4 (-1.8, 1.2)	0.335
plaque area, mm²	8.1 (5.1, 10.3)	7.5 (5.9, 10.3)	0.903	0.6 (-2.7, 2.4)	-0.1 (-2.1, 1.3)	-0.3 (-1.0, 1.1)	0.781
Total lesion	N= 60
HU^＊＊	69.7 (42.5, 79.6)	72.7 (58.8, 95.2)	0.059	7.0 (-6.0, -19.0)	2.9 (-20.3, 29.3)	7.6 (-7.1, 17.7)	0.654
Plaque volume, mm³	117.0 (81.6, 153.4)	106.8 (78.3, 145.0)	0.918	0.23 (-16.7, 9.9)	1.4 (-26.0, 21.0)	2.1 (-9.7, 5.6)	0.619
LAP Volume, mm³	23.3 (12.4, 41.7)	22.3 (12.0, 36.7)	0.346	-2.3 (-5.6, 2.9)	-0.2 (-9.9, 5.2)	-0.6 (-4.2, 3.5)	0.601
vessel area, mm²	13.4 (9.3, 16.1)	12.2 (9.3, 16.6)	0.200	0.5 (-1.4, 2.5)	-0.4 (-2.2, 1.4)	-0.2 (-1.5, 0.8)	0.082
plaque area, mm²	7.8 (5.6, 10.4)	7.8 (6.1, 10.4)	0.971	0.3 (-2.1, 1.6)	-0.1 (-1.7, 1.4)	-0.1 (-1.2, 1.4)	0.836

Values are medians (interquartile range).

Delta (Δ) means a change from the baseline to 12-month follow-up.

CTA, computed tomography angiography; HU, Hounsfield unit; LAP, low attenuation plaque

^＊No-treatment group N = 23, 2 g EPA/DHA group N = 13, 4 g EPA/DHA group N = 15

^＊＊No-treatment group N = 48, 2 g EPA/DHA group N = 33, 4 g EPA/DHA group N = 33

Supplementary Table 2. Difference of change in CT parameters from baseline to follow-up

	No treatment group (Primary lesion N = 24, Total lesion N = 56)	2g EPA/DHA group (Primary lesion N = 20, Total lesion N = 46)	4g EPA/DHA group (Primary lesion N = 23, Total lesion N = 60)	p-value
Primary lesion
%ΔHU^＊	0.52 (-24.9, 147.7)	-1.1 (-53.9, 73.8)	-4.4 (-14.3, 47.7)	0.765
%Δplaque Volume	-5.5 (-10.7, 9.6)	-4.8 (-14.5, 15.0)	0.7 (-10.7, 5.0)	0.911
%ΔLAP Volume	-6.0 (-23.8, 8.4)	-9.6 (-36.3, 21.9)	-3.7 (-23.1, 26.7)	0.694
%Δvessel area	0.2 (-8.4, 26.5)	-1.6 (-13.1, 11.3)	-3.2 (-13.0, 8.4)	0.196
%Δplaque area	5.4 (-17.8, 26.2)	-0.4 (-18.5, 16.2)	-5.1 (-12.9, 16.7)	0.697
Total lesion
%ΔHU^＊＊	12.0 (-15.2, 57.1)	-1.1 (-37.1, 48.5)	8.3 (-11.5, 29.8)	0.430
%Δplaque Volume	-0.2 (-11.6, 6.3)	1.0 (-16.0, 16.8)	1.8 (-9.4, 6.7)	0.765
%ΔLAP Volume	-8.7 (-21.6, 10.5)	-3.6 (-29.1, 24.3)	-3.8 (-21.7, 17.5)	0.900
%Δvessel area	2.8 (-6.5, 18.2)	-2.3 (-14.6, 14.5)	-1.9 (-12.9, 6.2)	0.026
%Δplaque area	3.9 (-13.0, 19.5)	-0.6 (-18.3, 14.0)	-1.7 (-16.8, 19.5)	0.506

Values are medians (interquartile range).

Delta (Δ) means a change from the baseline to 12-month follow-up.

CTA, computed tomography angiography; HU, Hounsfield unit; LAP, low attenuation plaque

^＊No-treatment group N = 23, 2 g EPA/DHA group N = 13, 4 g EPA/DHA group N = 15

^＊＊No-treatment group N = 48, 2 g EPA/DHA group N = 33, 4 g EPA/DHA group N = 33

Representative image findings at baseline and 12 months in a patient with a high triglyceride level are illustrated in Supplementary Fig.3a (decreased PMR) and Supplementary Fig.3b (increased PMR).

Supplementary Fig.3.

a) Representative images of pre- and post-EPA/DHA administration in a patient with high triglyceride levels. A high-intensity plaque with a PMR of 1.81 was observed in the proximal segment of the right coronary artery at baseline (A, dotted square). This plaque was a low-density coronary plaque with positive remodelling on CTA (B). Cross-sectional CTA images of this lesion show positive remodelling with LAP (ⅰ, ⅱ, and ⅲ). PMR decreased to 1.42 after 12 months of intensive statin treatment with EPA/DHA (C, dotted square). CTA findings at follow-up indicate partial regression with the coronary plaque (D, ⅳ, ⅴ, and ⅵ). Triglyceride levels decreased from 210 mg/dL (baseline) to 116 mg/dL (follow-up). b) Representative images at baseline and at the 12-month follow-up in a patient with high triglyceride levels in whom EPA/DHA was not administered. The PMR was 0.77 for a lesion in the proximal segment of the right coronary artery (A, dotted square), which corresponds to a coronary plaque on CTA (B). Cross-sectional CTA images of this lesion show relatively minor remodelling (ⅰ, ⅱ, and ⅲ). A PMR increase to 1.19 for the same lesion after 12 months (C, dotted square) and CTA findings at follow-up indicate partial progression with the coronary plaque (D, ⅴ). Triglyceride levels increased from 192 mg/dL (baseline) to 238 mg/dL (follow-up)

Discussion

The AQUAMARINE EPA/DHA randomized study was the first prospective and serial CMR study to investigate the antiatherogenic effect of EPA/DHA therapy in patients with CAD on statin treatment by examining the change in the PMR of coronary high-intensity plaques. Our major findings are as follows: (1) 12-month treatment with EPA/DHA under statin treatment did not show an additional effect on reductions of the PMR in the primary lesions. (2) There was also no additional effect in the analysis of all lesions. (3) The plaque component of high-risk features evaluated by CTA did not change by the treatment of EPA/DHA.

The REDUCE-IT and JELIS trials demonstrated the value of high-dose EPA for the preventive treatment of atherosclerotic cardiovascular disease^{3, 4)}. However, the recently published STRENGTH study showed inconsistent results with no additional preventive effect with 4-g EPA/DHA on cardiovascular events under statin therapy. Regarding this contradiction in results between these studies, there has been much debate^12-15), such as differences in the action of EPA and EPA/DHA, differences in the severity of participants between these studies, and the difference in oil used as a comparator between these two. To date, there is no consensus on the significance of the cardiovascular preventive effects of omega-3 fatty acids. To understand this contradiction in results, it is necessary not only to clarify the factors attributable to these clinical studies themselves but also to examine the antiatherosclerotic effect of omega-3 fatty acids in the mechanism of action on CAD-risk reduction.

It has been investigated that some biological actions beyond their ability to reduce triglyceride levels of omega-3 fatty acids, such as anti-inflammatory and antithrombotic activities and membrane-stabilizing effects, could contribute to the suppression of cardiovascular events^{16, 17)}. Indeed, in an experimental study with apolipoprotein E-deficient (ApoE^−/−) mice, EPA/DHA significantly attenuated the development and destabilization of atherosclerotic plaques¹⁸⁾. Several clinical studies also suggested antiarteriosclerotic effects evaluated by IVUS^{18, 19)} or OCT^{20, 21)}. The EVAPORATE study showed the inhibitory effect of the 4-g icosapent ethyl on plaque volume evaluated using serial CTA²²⁾. In contrast, there are a few studies suggesting the inhibitory effect of EPA/DHA on coronary arteriosclerotic plaques except for the following two studies: one suggested improvement of thin fibrous cap atheroma detected using OCT in a limited number of patients²¹⁾ and the other suggested a decrease in fibrous plaque volume using coronary CT in patients under the use of low dose statin²³⁾.

High-intensity plaque on noncontrast T1W MRI has been reported to show a strong correlation with high-risk plaque features such as positive remodeling and low attenuation observed on CTA or IVUS^{6, 9, 24)}. An OCT study showed that high-intensity plaque was strongly associated with a healed plaque rupture and a large lipid core⁷⁾. A recent postmortem pathological study revealed that a short T1 signal was associated with a large necrotic core and intraplaque hemorrhage in coronary atherosclerotic plaques²⁵⁾. On the basis of these morphological high-risk features, high-intensity plaques on CMR have been demonstrated to be a significant predictor of major adverse cardiac events in patients with CAD. (4) The AQUAMARINE pilot study⁹⁾ showed that statin treatment significantly reduced the PMR of high-intensity plaque; hence, T1W CMR imaging could be a useful technique for repeated quantitative assessment of plaque composition. It should be noted that coronary high–intensity plaques remained positive at PMR ≥ 1, that is, statin-residual risk on CMR. Therefore, the AQUAMARINE EPA/DHA study was designed to investigate the effect of EPA/DHA on the PMR of coronary high–intensity plaques.

Our study evaluating the effect of omega-3 fatty acids on arteriosclerotic plaques using T1WI CMR did not show significant differences in ΔPMR among the three groups. This result indicates that the addition of omega-3 fatty acids to patients treated with statins had no additional effect on high-risk arteriosclerotic plaques detected using T1WI CMR. The previous study suggested that statin therapy decreases the PMR⁹⁾. In this study, all participants were taking statins and were well controlled (serum LDL concentration: approximately 70–75 mg/dL), which might have contributed to the consistent decrease in the PMR in all groups during the study period. In addition, some lifestyle improvements due to participation in this study may also contribute to this change. In the present CTA analysis, there was no difference in plaque volume or high-risk indicators such as LAV and HU. Indeed, a subanalysis of a previous study assessed by CTA demonstrated that the inhibitory effect of high-dose EPA/DHA on coronary plaques found under low-dose statin therapy (serum LDL concentration: approximately 80.1 mg/dL) was attenuated in the therapy with high-dose statin (serum LDL concentration: approximately 76.1 mg/dL)²³⁾.

This is the first study using T1WI CMR to examine the effect of high-dose EPA/DHA on the coronary atherosclerotic plaque, and this study adds new insights into this important field. Regarding the CTA analysis, this study was still the second RCT to evaluate 4-g EPA/DHA on atherosclerotic plaques using CTA, as far as we know.

Our trial has some points that differ from STRENGTH. In our study, 1) eligible criteria were not limited to patients with a high triglyceride level; 2) no clear effect of lowering triglyceride levels was observed; and 3) EPA concentration in the EPA/DHA groups was higher than that in STRENGTH at 1 year (AQUAMARINE EPA/DHA 194 µg/mL vs. STRENGTH 89.6 µg/mL). Although subanalysis analysis targeting patients who met the STRENGTH eligible criteria may add findings, our study lacked the power to conduct them.

Clinical studies to evaluate the effect of high-dose EPA/DHA on atherosclerotic plaques are still insufficient. A multicentre and double-blind trial with a larger number of patients with residual cardiovascular risk despite treatment with statins compared between EPA/DHA and control or EPA/DHA and EPA should be investigated with the standardized CMR analysis. Such studies are expected to clarify the significance of the effects of each type of omega-3 fatty acid on the arteriosclerotic plaque, and the accumulation of such scientific studies may provide a better understanding of the cardiovascular protective effects of omega-3 fatty acids.

Our study has several limitations. First, this was a single-center trial with a relatively small sample size in a nonblinded manner. Because the number of subjects in this study did not reach sufficient statistical power to evaluate the primary endpoint, we cannot deny the possibility of underestimating the effect of EPA/DHA on high-risk plaques detected using CMR. Second, the endpoint of this study was a surrogate marker that evaluates atherosclerotic plaque using MRI or CT. Third, it is unknown whether 1 year of use is sufficient to determine the effect of EPA/DHA on atherosclerotic plaque. However, there are several studies that have been able to evaluate the effect of drugs on changes in plaque features over a period of equal or less than 1 year by IVUS, OCT, or MRI^{9, 21, 26)}. Fourth, we do not have information on the use of ezetimibe or PCSK9 inhibitors.

Conclusions

The AQUAMARINE EPA/DHA study, a single-center, triple-arm, randomized, controlled, open-label trial, demonstrated that the EPA/DHA therapy of 2 or 4 g/day did not significantly improve the high-risk features of coronary atherosclerotic plaques evaluated by serial CMR and CTA in patients with CAD under statin therapy.

Acknowledgement of Grant Support

This work was supported by Takeda Pharmaceutical Co., Ltd. Funding No. OME-IIT-001 (142CR1-002).

Conflicts of Interest

Dr. Yasuda S. receiving lecture fees or honoraria from Bayer Yakuhin, Ltd, Daiichi Sankyo Co. Ltd., research funding from NEC Solution Innovators, Ltd., Bayer Yakuhin, Ltd, Daiichi Sankyo Co. Ltd., scholarship grants from Abbott Medical Japan LLC, Amicus Therapeutics Co. Ltd. Otsuka Pharmaceutical Co. Ltd., Kowa Company. Ltd., Sumitomo Dainippon Pharma Co. Ltd., Roche Diagnostics K.K. Courses endowed by companies from Abbott Medical Japan LLC, Terumo Co., Medtronic Japan Co. Ltd., Tesco Co. Ltd., Nihon Kohden Co., Japan Lifeline Co. Ltd., Sound Wave Innovation Co. Ltd., ONO PHARMACEUTICAL CO. LTD., SHIONOGI ＆ CO. LTD., ZEON MEDICAL INC., Nippon Shinyaku Co. LTD., Nippon Boehringer Ingelheim Co. Ltd., Takeda Pharmaceutical Co. Ltd., Mochida Pharmaceutical Co. Ltd., BIOTRONIC JAPAN Co. Ltd., KANEKA MEDICAL PRODUCTS, FUKUDA DENSHI Co. Ltd. and Philips Japan Ltd.

Dr. Ogawa H. receiving lecture fees or honoraria from Bayer Yakuhin, Ltd

Dr. Asaumi Y. receiving scholarship grants from Abbott Medical Japan LLC and Terumo Co.

The remaining authors have nothing to disclose.

Ethics Statements

The ethical principles guiding the study have their origins in the Declaration of Helsinki. This trial was approved by the institutional review boards of Osaka University Hospital and the National Cerebral and Cardiovascular Centre.

Clinical Study Registration Number

This clinical trial was registered on the University Hospital Medical Information Network Clinical Trials Registry (UMIN 000015316) and Japan Registry of Clinical Trials (jRCTs051180125). Written informed consent was obtained from patients.

Authorship Statement

SY is the principal investigator of this study; he was primarily responsible for surveillance conducting the clinical trial, acquiring funding, designed the study, writing the manuscript, and supervising the study. TN designed and acquired ethics approval of the study and was in charge of participant recruitment. KN drafted and wrote the manuscript and was in charge of participant recruitment. YM and HM were responsible for imaging using CMR and CTA. KN, HM, and TN performed statistical analyses. KO supervise statical analysis. KK. YG, HY, HO supervised the study. YA, KN, TN, HM, HM, KS, TD, HH, SY, SK, TN, KN, TK, FO, MN, YK, YT and YK oversaw recruitment and follow-up. All authors read and approved the final manuscript.

References

責任著者(Corresponding author)

J-STAGEへの登録はこちら（無料）