Benefit of icosapent ethyl on coronary physiology assessed by computed tomography angiography fractional flow reserve: EVAPORATE-FFR_CT

Abstract

Aims

Icosapent ethyl (IPE) significantly reduced ischaemic events in statin-treated patients with atherosclerosis or diabetes and elevated triglycerides in REDUCE-IT, including large reductions in myocardial infarction and elective, urgent, and emergent coronary revascularization. However, the mechanisms driving this clinical benefit are not fully known. The EVAPORATE trial demonstrated that IPE significantly reduced plaque burden. No study to date has assessed the impact of IPE on coronary physiology. Fractional flow reserve (FFR) derived from coronary computed tomography angiography (CTA) data sets (FFR_CT) applies computational fluid dynamics to calculate FFR values in epicardial coronary arteries. Our objective was to assess the impact of IPE on coronary physiology assessed by FFR_CT using imaging data from EVAPORATE.

Methods and results

A total of 47 patients and of 507 coronary lesions at baseline, 9 months, and 18 months with coronary CTA and FFR_CT were studied in a blinded core lab. The pre-specified primary endpoint was the FFR_CT value in the distal coronary segment from baseline to follow-up in the most diseased vessel per patient using IPE compared with placebo. The pre-specified secondary endpoint was the change in translesional FFR_CT (ΔFFR_CT) across the most severe (minimum 30% diameter stenosis) coronary lesion per vessel. Baseline FFR_CT was similar for IPE compared with placebo (0.83 ± 0.08 vs. 0.84 ± 0.08, P = 0.55). There was significant improvement in the primary endpoint, as IPE improved mean distal segment FFR_CT at 9- and 18-month follow-up compared with placebo (0.01 ± 0.05 vs. −0.05 ± 0.09, P = 0.02, and −0.01 ± 0.09 vs. −0.09 ± 0.12, P = 0.03, respectively). ΔFFR_CT in 140 coronary lesions was improved, although not statistically significant, with IPE compared with placebo (−0.06 ± 0.08 vs. −0.09 ± 0.1, P = 0.054).

Conclusion

Icosapent ethyl demonstrated significant benefits in coronary physiology compared with placebo. This early and sustained improvement in FFR_CT at 9- and 18-month follow-up provides mechanistic insight into the clinical benefit observed in the REDUCE-IT trial. Furthermore, this is the first assessment of FFR_CT to determine drug effect.

Graphical Abstract

Graphical Abstract.

See the editorial comment for this article ‘Icosapent ethyl and plaque regression: insights from the EVAPORATE-FFRCT study’, by Robert Sykes and Richard McFarlane, https://doi.org/10.1093/ehjci/jead086.

Introduction

Patients with elevated triglyceride (TG) levels are at increased risk for ischaemic events.^1–5 Icosapent ethyl (IPE) significantly reduced the burden of first, subsequent, and total ischaemic events among statin-treated patients with cardiovascular disease or diabetes and elevated triglycerides in REDUCE-IT.^6–9 However, the mechanisms driving this marked clinical benefit are not fully known.

The CHERRY study demonstrated that IPE in addition to statin therapy significantly reduced coronary plaque volume assessed by intravascular ultrasound compared with statin therapy alone in a post-acute coronary syndrome population.¹⁰ However, intravascular ultrasound is an invasive technique employed in the catheterization laboratory. The Effect of Vascepa on Improving Coronary Atherosclerosis in People With High Triglycerides Taking Statin Therapy (EVAPORATE, NCT02926027) trial demonstrated that in statin-treated patients, IPE significantly reduced plaque burden measured by serial coronary computed tomography angiography (CTA) compared with placebo.¹¹ However, no study to date has assessed the impact of IPE on coronary physiology. Haemodynamic forces, plaque vulnerability, and the interaction between these factors may cause subsequent cardiovascular events.¹²

Among non-invasive anatomic tests for coronary artery disease (CAD), coronary CTA has emerged as a novel imaging modality that is capable of providing high-resolution images of coronary artery lesions.¹³ Over the past decade, there has been a strong interest in computing fractional flow reserve (FFR) non-invasively derived from coronary CTA data sets.¹⁴ FFR derived from coronary CTA data sets (FFR_CT) applies computational fluid dynamics to calculate FFR values in all epicardial coronary arteries without the need for additional medications, imaging, or change in imaging protocol.^15,16 The combination of high-resolution anatomic definition of CAD via CTA with FFR_CT in a single test provides a comprehensive non-invasive anatomic and functional assessment of epicardial CAD which is supported by the American College of Cardiology and American Heart Association guidelines.^17,18 We hypothesized that IPE would improve coronary physiology. This study assessed the impact of IPE compared with placebo on coronary haemodynamics assessed by FFR_CT using imaging data from the EVAPORATE clinical trial.

Methods

A total of 47 patients and 507 coronary lesions at baseline, 9 months, and 18 months with coronary CTA and FFR_CT were studied in core laboratories blinded to clinical information, symptom status, and outcomes (Lundquist/Harbor-UCLA, Los Angeles, CA, USA, and Loyola University Medical Center, Maywood, IL, USA).

Study endpoints

The purpose of this study was to assess whether IPE at 4 g/day, as an adjunct to diet and statin therapy, in patients with fasting triglycerides between 135 and 499 mg/dL at randomization affected coronary physiology at follow-up. The pre-specified primary endpoint was the FFR_CT value in the distal coronary segment from baseline to follow-up in the most diseased vessel per patient using IPE compared with placebo. If the evaluable distal FFR_CT segment was different, the proximal one was used as the reference for comparison.

The pre-specified secondary endpoint comparing baseline to follow-up using IPE compared with placebo was the change in translesional FFR_CT (ΔFFR_CT) across the most severe (minimum 30% diameter stenosis) coronary lesion per vessel.

Study population

The study population for EVAPORATE has been published previously.¹⁹ Briefly, EVAPORATE included individuals between 30 and 85 years with CAD (one or more angiographic stenoses with ≥20% narrowing), elevated fasting triglycerid levels (135–499 mg/dL), and low-density lipoprotein levels (LDL-C) ranging from 40 to 115 mg/dL. Patients were on stable statin therapy, with or without ezetimibe, and adhered to diet and exercise for ≥4 weeks prior to study entry.

Study design

The study design and rationale for EVAPORATE have been published previously.¹⁹ Briefly, EVAPORATE was a multicenter, randomized, double-blind, placebo-controlled trial that evaluated the effect of IPE 4 g/day on coronary plaque progression determined by coronary CTA compared with placebo. Patients were randomized 1:1 to IPE or placebo to evaluate plaque volume progression rates using coronary CTA. Participants underwent coronary CTA at baseline, 9 months, and 18 months.

The study was approved by the institutional review board at each site and was conducted in accordance with the principles of Good Clinical Practice, and the trial conformed to the principles outlined in the Declaration of Helsinki. All patients provided written informed consent prior to randomization.

Computation of FFR from coronary CTA

FFR_CT analysis was performed by HeartFlow Inc. (Mountain View, CA, USA) as previously described.¹⁵ First, semi-automated segmentation of the coronary arteries and determination of left ventricular mass were performed. Then, calculations of FFR_CT were achieved by computational fluid dynamic modeling and determined for all vessels of diameter >1.8 mm. ΔFFR_CT was derived by determining the presence of stenosis on the FFR_CT 3D model in multiple projections. The FFR_CT interactive model was co-localized with the coronary CTA data set using branch points and bifurcations as landmarks. The 3D model was used to pin proximal and distal reference points at regions immediately adjacent to the stenosis which appeared free of stenosis. ΔFFR_CT was defined as the difference of FFR_CT values between the proximal and distal points.²⁰ If the evaluable distal FFR_CT segment was different, the proximal one was used as the reference for comparison. An analysable baseline coronary CTA for FFR_CT was needed. In the event a vessel was not analysed for FFR_CT secondary to an artefact, when possible, another vessel with baseline and follow-up FFR_CT was used.

Statistical analyses

Baseline characteristics were compared between the two arms using Student’s t test and χ² statistic. Results are presented as counts and frequencies (per cent) for discrete variables and mean and standard deviations or median for continuous variables. FFR_CT changes between baseline, 9 months, and 18 months were compared between the two arms using a random effects model with baseline value, arm, visit, and the interaction between visit and arm as fixed effects and a random term for the intercept. Intra-observer and inter-observer reproducibility testing was performed by having both observers repeat measurements in random order for 10 randomly selected coronary vessels. Intra-observer reproducibility was reported using the coefficient of variance (COV), which measures relative dispersion, and intraclass correlation coefficient (ICC). Inter-observer variability was measured by inter-rater variability (kappa statistic).

All statistical analyses report two-sided P-values for the outcomes. A P < 0.05 was considered significant for the outcomes. No adjustment for multiplicity of testing was performed. All analyses were performed using SAS for Windows, version 9.4 (SAS Institute, Cary, NC), and intention to treat, with study patients analysed by treatment group assigned regardless of study drug adherence.

Results

Population characteristics

Of the 68 patients completing the 18-month visit and having a coronary CTA at baseline and the 18-month visit, a total of 47 patients had interpretable FFR_CT analyses at baseline and follow-up (Figure 1). Coronary CTA image quality was not acceptable for FFR_CT analysis due to motion artefact, calcification, and misregistration. The mean age of the participants was 57 ± 9 years, with 51% being male. Baseline characteristics of the study participants, stratified by arm (IPE group, n = 22, and placebo group, n = 25), are demonstrated in Table 1. Baseline demographics and risk factors (including age, BMI, hypertension, diabetes, hyperlipidaemia, and smoking prevalence) were similar between the IPE and placebo groups (Table 1). Bland–Altman plots for the assessment of intra-observer variability are shown in Figure 2. Overall, intra-observer reproducibility was excellent as the COV was 1.03% [confidence interval (CI): 0.69–1.38] and ICC was 99%. Limits of agreement (LOA) were 0.03 and 0.04 for the first and second observers, respectively. There was a very good inter-observer reproducibility as well, with a weighted kappa of 0.88 (CI: 0.81–0.95).

Figure 1.

Flow diagram.

Table 1.

Baseline characteristics of the EVAPORATE-FFR_CT cohort

Characteristic	Total	Icosapent ethyl	Placebo	P
	(n = 47)	(n = 22)	(n = 25)
	Mean (SD) or count (%)	Mean (SD) or count (%)	Mean (SD) or count (%)
Age, years	57.1 (9.3)	56.5 (9.9)	57.6 (8.9)	0.67
Sex, male	24 (51.1%)	11 (50.0%)	13 (52.0%)	0.89
Body mass index, kg/m2	32.0 (5.0)	31.8 (4.5)	32.2 (5.5)	0.78
Ethnicity, Hispanic	25 (53.2%)	11 (50.0%)	14 (56.0%)	0.68
Race, white	40 (85.1%)	19 (86.4%)	21 (84.0%)	0.88
Diabetes mellitus	29 (61.7%)	13 (59.1%)	16 (64.0%)	0.73
Family history	14 (29.8%)	6 (27.3%)	8 (32.0%)	0.72
Hypertension	32 (68.1%)	15 (68.2%)	17 (68.0%)	0.99
Past smoking	17 (36.2%)	8 (36.4%)	9 (36.0%)	0.98
Aspirin use	20 (42.6%)	9 (40.9%)	11 (44.0%)	0.83

Figure 2.

Bland–Altman plots for intra-observer reproducibility.

Changes with therapy

Baseline FFR_CT was similar for IPE compared with placebo (0.83 ± 0.08 vs. 0.84 ± 0.08, P = 0.55). Coronary physiology changes between groups were significantly different (Table 2). There was significant improvement in the primary endpoint as IPE improved distal segment FFR_CT at 9- and 18-month follow-up compared with placebo (0.01 ± 0.05 vs. −0.05 ± 0.09, P = 0.02, and −0.01 ± 0.09 vs. −0.09 ± 0.12, P = 0.03, respectively). Compared with placebo, IPE significantly improved the primary endpoint FFR_CT by 0.06 and 0.08 at 9- and 18-month follow-up [0.06 (95% CI: 0.01–0.11; P = 0.02) and 0.08 (95% CI: 0.01–0.15; P = 0.03), respectively]. ΔFFR_CT in 140 coronary lesions was improved with IPE compared with placebo (−0.06 ± 0.08 vs. −0.09 ± 0.1, P = 0.05; Table 3). Representative case examples comparing IPE to placebo on coronary physiology are illustrated in Figures 3 and 4.

Table 2.

FFR_CT change by treatment group

Characteristic	Overall (n = 47)	Icosapent ethyl (n = 22)	Placebo (n = 25)	Difference (95% CI)	P-value
Baseline FFRCT
Mean ± SD (n)	0.84 ± 0.08 (44)	0.83 ± 0.08 (20)	0.84 ± 0.08 (24)	−0.02 (−0.06, 0.03)	0.55
Median and range	0.86 (0.62–0.96)	0.86 (0.67–0.94)	0.86 (0.62–0.96)
IQR	0.79–0.89	0.76–0.89	0.8–0.9
FFRCT 9 months
Mean ± SD (n)	0.81 ± 0.11 (39)	0.83 ± 0.07 (19)	0.79 ± 0.13 (20)	0.04 (−0.03, 0.11)	0.23
Median and range	0.84 (0.5–0.93)	0.85 (0.73–0.93)	0.84 (0.5–0.92)
IQR	0.77–0.89	0.77–0.9	0.76–0.88
FFRCT 18 months
Mean ± SD (n)	0.78 ± 0.13 (38)	0.81 ± 0.11 (18)	0.75 ± 0.14 (20)	0.06 (−0.02, 0.14)	0.14
Median and range	0.79 (0.5–0.93)	0.81 (0.5–0.93)	0.78 (0.5–0.92)
IQR	0.74–0.89	0.76–0.9	0.68–0.87
FFRCT baseline change to 9 months
Mean ± SD (n)	−0.02 ± 0.08 (39)	0.01 ± 0.05 (19)	−0.05 ± 0.09 (20)	0.06 (0.01, 0.11)	0.02
Median and range	−0.01 (−0.28 to 0.11)	0.01 (−0.09 to 0.11)	−0.04 (−0.28 to 0.09)
IQR	−0.04 to 0.03	−0.01 to 0.04	−0.08 to 0.01
FFRCT baseline change to 18 months
Mean ± SD (n)	−0.06 ± 0.11 (38)	−0.01 ± 0.09 (18)	−0.09 ± 0.12 (20)	0.08 (0.01, 0.15)	0.03
Median and range	−0.03 (−0.41 to 0.12)	−0.01 (−0.25 to 0.12)	−0.05 (−0.41 to 0.04)
IQR	−0.1 to 0.01	−0.04 to 0.02	−0.13 to 0.01

FFR_CT, fractional flow reserve derived from coronary computed tomography angiography data sets.

Table 3.

Translesional FFR_CT (ΔFFR_CT) across the most severe coronary lesion per vessel

Characteristic	Overall (n = 140)	Icosapent ethyl (n = 78)	Placebo (n = 62)	Difference (95% CI)	P-value
ΔFFRCT
Mean ± SD (n)	−0.08 ± 0.09 (140)	−0.06 ± 0.08 (78)	−0.09 ± 0.1 (62)	0.0298 (−0.0006, 0.0601)	0.054
Median and range	−0.04 (−0.48 to 0)	−0.04 (−0.48 to 0)	−0.05 (−0.46 to 0)
IQR	−0.1 to 0.02	−0.09 to 0.02	−0.13 to 0.02

FFR_CT, fractional flow reserve derived from coronary computed tomography angiography data sets.

Figure 3.

Icosapent ethyl (IPE) group. Top: example of a patient from the IPE arm showing distal segment FFR_CT of a right coronary artery lesion at baseline, 9 months, and 18 months. Bottom: example of the same patient from the IPE arm showing ΔFFR_CT across the most severe right coronary artery lesion at baseline, 9 months, and 18 months. FFR_CT, fractional flow reserve derived from coronary computed tomography angiography data sets; ΔFFR_CT, translesional fractional flow reserve derived from coronary computed tomography angiography data sets.

Figure 4.

Placebo group. Top: example of a patient from the placebo arm showing distal segment FFR_CT of a left circumflex lesion at baseline, 9 months, and 18 months. Bottom: example of the same patient from the placebo arm showing ΔFFR_CT across the most severe left circumflex lesion at baseline, 9 months, and 18 months. FFR_CT, fractional flow reserve derived from coronary computed tomography angiography data sets; ΔFFR_CT, translesional fractional flow reserve derived from coronary computed tomography angiography data sets.

Discussion

The REDUCE-IT trial demonstrated that IPE significantly reduced ischaemic events in statin-treated patients with atherosclerosis or diabetes and elevated triglycerides, including large reductions in myocardial infarction (MI) and elective, urgent, and emergent coronary revascularization.^{6–9,21–24} However, the mechanisms driving this clinical benefit are not fully known. The EVAPORATE trial was designed to be a mechanistic study with similar inclusion and exclusion criteria to parallel REDUCE-IT and demonstrated that IPE significantly reduced plaque burden.¹¹ This is the first study to assess the impact of IPE on coronary physiology. We demonstrated significant benefits in coronary physiology with IPE compared with placebo. This early improvement in FFR_CT and ΔFFR_CT at 9 months was sustained at longer term 18-month follow-up and provides mechanistic insight into the clinical benefits observed in the REDUCE-IT trial. Moreover, this is the first assessment of FFR_CT to determine drug effect, which has potentially important implications in utilizing FFR_CT to predict treatment response.

Based on the evidence, the American Heart Association and American College of Cardiology, along with multiple other medical societies, elevated coronary CTA as a Class 1A non-invasive test and recognized FFR_CT as a Class 2a with Level B evidence in the Guideline for the Evaluation and Diagnosis of Chest Pain.^17,18 In fact, non-invasive coronary haemodynamics derived from coronary CTA and computational fluid dynamics have been shown to predict high-risk plaque and downstream clinical outcomes.

The international Assessing Diagnostic Value of Non-Invasive FFR_CT in Coronary Care (ADVANCE, NCT02499679) registry of over 5000 patients with CAD who underwent a diagnostic strategy of coronary CTA and FFR_CT demonstrated significantly lower cardiovascular death or MI in patients with higher FFR_CT values at 1-year follow-up.²⁵ Cardiovascular death or MI occurred more in patients with an FFR_CT ≤0.80 compared with patients with an FFR_CT > 0.80 (relative risk [RR]: 4.22; 95% CI: 1.28–13.95; P = 0.01), and major adverse cardiovascular events (MACE), all-cause death or MI, and cardiovascular death or MI increased as FFR_CT values decreased. Our findings demonstrate significant improvement in the primary endpoint as IPE improved distal segment FFR_CT at 9- and 18-month follow-up compared with placebo (0.01 ± 0.05 vs. −0.05 ± 0.09, P = 0.02, and −0.01 ± 0.09 vs. −0.09 ± 0.12, P = 0.03, respectively) which may shed light on the underlying mechanisms driving the marked clinical benefit observed in the REDUCE-IT trial, including early revascularization. Results from the REDUCE-IT REVASC study demonstrated significant benefit soon after randomization and achieved statistical significance by only 11 months.²¹

ΔFFR_CT has emerged as an imaging biomarker to predict cardiovascular outcomes.²⁶ The Exploring the Mechanism of the Plaque Rupture in Acute Myocardial Infarction (EMERALD, NCT02374775) multinational, multicenter trial assessed haemodynamic predictors of acute coronary syndrome (ACS) using computational fluid dynamics.²⁷ Compared with non-culprit lesions, culprit lesions had significantly lower FFR_CT and higher ΔFFR_CT, axial plaque, and wall shear stress. Non-invasive computational fluid dynamic–derived haemodynamic parameters were better at identifying culprit lesions causing ACS than diameter stenosis or adverse plaque characteristics. In fact, ΔFFR_CT was shown to be the greatest predictor of ACS.

Furthermore, in a study of 4730 patients from the ADVANCE registry, ΔFFR_CT remained an independent predictor for early revascularization [odds ratio per 0.05 increase (95% CI), 1.31 (1.26–1.35); P < 0.001] after adjusting for risk factors, stenosis features, and lesion-specific FFR_CT.²⁰ ΔFFR_CT improved the discrimination of patients who underwent early revascularization compared with a standard diagnostic strategy of coronary CTA with FFR_CT. In our study, ΔFFR_CT in coronary lesions was improved with IPE compared with placebo, which may partially explain the robust clinical benefit in ischaemic events observed in the REDUCE-IT trial.

FFR_CT has been associated with various clinical outcomes, such as the safe deferral of invasive coronary angiography, cardiovascular death or MI, and revascularization.^14,25 Considering this correlation between FFR_CT and clinical outcomes, assessing for improvement in coronary physiology using FFR_CT over time with pharmacologic therapy may serve as a surrogate for treatment response, cardiovascular outcomes, and drug efficacy and warrants further investigation. In addition to its high diagnostic performance and prognostic ability, this trial paves the way for FFR_CT to determine drug effect and has implications in utilizing FFR_CT and other non-invasive haemodynamic metrics derived from computational fluid dynamics to assess treatment response.

Limitations

The study’s primary limitation is the small sample size. Despite this, we demonstrated improvement in the primary and secondary endpoints with IPE compared with placebo. EVAPORATE did not assess for long-term outcomes; rather, it was designed as a mechanistic study to accompany the REDUCE-IT randomized outcome trial with similar study design, inclusion and exclusion criteria, and interventions. Notably, this is the first study to demonstrate improvement in coronary physiology and associate improved coronary haemodynamics with improvement in outcomes shown in a clinical trial. It would be interesting to study the impact of IPE on the microcirculation. However, with the current generation of FFR_CT utilized in this analysis, assessment of the microcirculation is not possible. Twenty-one of 68 patients did not have interpretable baseline and follow-up coronary CT scans for FFR_CT analysis. Guideline-directed coronary CTA acquisition methods, including adequate heart rate control and nitroglycerin administration, optimize image quality. Adherence to these methods may improve acceptance rates for FFR_CT.

Conclusion

Icosapent ethyl demonstrated significant benefits in coronary physiology. This early and sustained improvement in FFR_CT at 9- and 18-month follow-up provides mechanistic insight into the clinical benefit observed in the REDUCE-IT trial. Furthermore, this is the first assessment of FFR_CT to determine drug effect.

Funding

EVAPORATE was funded by Amarin Pharma, Inc. (Bridgewater, NJ, USA). As an investigator-initiated study, the company had no input in analysis, endpoint adjudication, or study performance or measures.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.