Association between Serum Oxysterols and Coronary Plaque Regression during Lipid-Lowering Therapy with Statin and Ezetimibe: Insights from the CuVIC Trial

Abstract

Aim: Several clinical trials using intravascular ultrasound (IVUS) evaluation have demonstrated that intensive lipid-lowering therapy by statin or a combination therapy with statin and ezetimibe results in significant regression of coronary plaque volume. However, it remains unclear whether adding ezetimibe to statin therapy affects coronary plaque composition and the molecular mechanisms of plaque regression. We conducted this prospective IVUS analysis in a subgroup from the CuVIC trial.

Methods: The CuVIC trial was a prospective randomized, open, blinded-endpoint trial conducted among 11 cardiovascular centers, where 260 patients with coronary artery disease who received coronary stenting were randomly allocated into either the statin group (S) or the combined statin and ezetimibe group (S＋E). We enrolled 79 patients (S group, 39 patients; S＋E group, 40 patients) in this substudy, for whom serial IVUS images of nonculprit lesion were available at both baseline and after 6–8 months of follow-up.

Results: After the treatment period, the S＋E group had significantly lower level of low-density lipoprotein cholesterol (LDL-C; 80.9±3.7 vs. 67.7±3.8 mg/dL,p=0.0143). Campesterol, a marker of cholesterol absorption, and oxysterols (β-epoxycholesterol, 4β-hydroxycholesterol, and 27-hydroxycholesterol) were also lower in the S＋E group. IVUS analyses revealed greater plaque regression in the S＋E group than in the S group (−6.14% vs. −1.18% for each group,p=0.042). It was noteworthy that the lowering of campesterol and 27-hydroxycholesterol, but not LDL-C, had a significant positive correlation with plaque regression.

Conclusions: Compared with statin monotherapy, ezetimibe in combination with statin achieved significantly lower LDL-C, campesterol, and 27-hydroxycholesterol, which resulted in greater coronary plaque regression.

See editorial vol. 30: 859-860

Clinical trials registry: UMIN000005597

Introduction

Despite major advances in cardiovascular science, including interventional cardiology, coronary artery disease (CAD) remains as one of the leading causes of mortality and morbidity. Intensive lipid-lowering therapy with 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitor (statin) has been established as a cornerstone of cardiovascular medication, especially for secondary prevention. However, patients under aggressive statin treatment hold substantial residual risk for recurrent coronary events. There have been some reports indicative of the hypothetical concept that inhibition of cholesterol synthesis by statin treatment results in enhanced absorption of cholesterol, which is associated with higher risk for coronary events ^{1 - 3)} . Ezetimibe is an inhibitor of the cholesterol transporter Niemann-pick C1-like 1, which is expressed in the small intestine. It blocks absorption of dietary and biliary cholesterol and potently lowers serum cholesterol level. A recent study clearly demonstrated that the addition of ezetimibe to statin treatment led to a significantly lower incidence of cardiovascular event in patients with acute coronary syndrome (ACS) ⁴⁾ . Coronary plaque burden assessed with intravascular ultrasound (IVUS) is a predictor of future cardiovascular events and has been commonly used as a surrogate marker in several clinical trials ^{5 , 6)} . Recently, combination therapy with a statin and ezetimibe showed coronary plaque regression by IVUS in the PRECISE-IVUS trial ⁷⁾ . Although this finding might be attributed to aggressive lipid-lowering by inhibition of cholesterol absorption, the molecular mechanisms driving the additional benefit of ezetimibe on plaque progression need to be further investigated ⁸⁾ .

Oxysterols, oxidized products of cholesterol, are enzymatically generated during cholesterol catabolism in mammals (e.g., 27-hydroxycholesterol is formed enzymatically through sterol 27-hydroxylase especially expressing in the liver) or nonenzymatically generated depending on oxidative stress in vivo or oxidation during cooking (e.g., 7-ketocholesterol) ⁹⁾ . An oxysterol in food was absorbed from the intestine, which was inhibited by the treatment with ezetimibe in humans ¹⁰⁾ . Oxysterols are found in the serum as well as in the plaques ⁹⁾ .

We previously reported that dietary oxysterols accelerated atherosclerotic plaque destabilization in a hypercholesterolemic mouse model and ezetimibe monotherapy ameliorated the plaque destabilization associated with decreases in low-density lipoprotein cholesterol (LDL-C), oxidized LDL, and oxysterol levels ¹¹⁾ . In addition, we recently reported a result of the CuVIC trial (Effect of Cholesterol Absorption Inhibitor Usage on Target Vessel Dysfunction After Coronary Stenting), in which ezetimibe in combination with statin ameliorated coronary endothelial dysfunction (CED) associated with significant reductions in oxysterol levels in patients after coronary stenting ¹²⁾ . We hypothesized that coronary plaque regression seen in serial examinations with IVUS under the treatment with statin with or without ezetimibe is associated with the degrees of reductions in oxysterols.

Aim

In this substudy of the CuVIC trial, we evaluated the impact of ezetimibe in combination with statin on coronary plaque burden in nontarget lesions in patients who had undergone coronary stenting and examined the correlation between the plaque burden and the changes in lipid profile, including oxysterols.

Methods

The present study is a post-hoc analysis of the CuVIC trial. The design of the CuVIC main study has been described previously ¹²⁾ . Briefly, 260 patients with CAD who underwent coronary stenting at 11 cardiovascular centers were randomly allocated into two arms [statin monotherapy (S) vs. ezetimibe 10 mg/day＋statin combination therapy (S＋E)]. In both treatment groups, we set the target LDL-C value at 100 mg/dL or less. Participating physicians were allowed to titrate statin doses to achieve that goal. The primary endpoint of the main study was target vessel dysfunction defined as the composite of target vessel failure during the follow-up period of 6–8 months and CED determined by intracoronary injection of acetylcholine at the follow-up coronary angiography (CAG). Among those patients, we enrolled the patients for whom serial IVUS images of nonculprit lesion were available in this IVUS substudy. Nonculprit lesions were defined as a lesion with plaque burden greater than 25% at the proximal or distal to the stented site on the target vessel.

Biomarker Assessment

Blood samples were collected at baseline and follow-up. LDL-C levels were calculated using the Friedewald equation. In addition to routine laboratory tests, including lipid profiling at each participating center, samples were measured for high-sensitivity C-reactive protein (hs-CRP) and malondialdehyde-modified LDL (MDA-LDL) levels at Medical and Biological Laboratories Co., Ltd., Nagoya, Japan. The noncholesterol sterols (campesterol, sitosterol, and lathosterol) were measured at SRL, Inc., Tokyo, Japan, using gas chromatography (GC-2010; Shimadzu Co., Kyoto, Japan). Oxysterols were quantified using gas chromatography mass spectrometry (GC/MS QP2010; Shimadzu Co.) equipped with an SPB-1 fused silica capillary column (60 m×0.25 mm, 0.25 µm phase thickness; Supelco Inc., Bellefonte, PA, USA).

IVUS Image Acquisition and Analyses

Serial IVUS was performed at baseline and follow-up. We used 40 MHz IVUS transducer (Intrafocus, Terumo, Japan) to acquire the image of the target vessel. The IVUS catheter was inserted as distally as possible, and the image was acquired during mechanical pullback at a speed of 0.5 mm/s. The target segment to be analyzed had to be at least 5 mm distal or proximal to the stented segment. An independent observer who was unaware of the treatment allocation analyzed the acquired images. For volumetric measurement of plaque, the surface of the intima and the external elastic membrane (EEM) were manually traced in every 30th frame (0.5 mm apart). The change in plaque burden was evaluated by the net change in % plaque burden defined as (plaque burden)/(EEM volume)×100 over the follow-up period ^{5 , 6)} . Integrated backscatter intravascular ultrasound (IB-IVUS) parameters were obtained in the same frames to evaluate plaque composition. The IB-IVUS system (IB-IVUS, YD Co., Ltd., Nara, Japan) demonstrated close agreement with the histological assessment of plaque characteristics ¹³⁾ . The plaque components were classified into the following four categories: (1) lipid pool (blue), (2) fibrosis (green), (3) dense fibrosis (yellow), and (4) calcification (red).

Statistical Analysis

Statistical analysis was performed using the JMP software (SAS Institute Inc., Cary, NC, USA). Continuous variables are presented as medians and ranges or interquartile ranges, and categorical variables are presented as numbers and percentages. Categorical variables were analyzed using a χ² test, and continuous variables were analyzed using a paired t-test and unpaired Student’s t-test or an analysis of covariance. A two-way analysis of variance was used to assess the differences between treatment arms. A p-value ＜0.05 was considered significant.

Results

Study Flow and Baseline Characteristics

In the CuVIC main study, 260 patients who underwent successful percutaneous coronary intervention were randomly allocated to statin monotherapy or ezetimibe plus statin combination therapy. Among these patients, serial IVUS images of nonculprit lesions were available in 102 patients. We excluded 23 patients because of insufficient image quality. Overall, 79 patients (39 in the S group and 40 in the S＋E group) were enrolled in this substudy ( Fig.1 ) . Table 1 shows the baseline characteristics of enrolled patients in both treatment groups. More male patients were included in the S＋E group than in the S group, and angiotensin-converting enzyme inhibitor (ACE-I) was prescribed more often in the S＋E group than in the S group. Age, body weight, coronary risk factors, medication (excluding ACE-I), and patients with ACS were similar between the two treatment groups.

Fig.1. Study flowchart.

IVUS=intravascular ultrasound, CAG=coronary angiography, TVF=target vessel failure, f/u=follow-up, S=statin group, S＋E=statin and ezetimibe group.

Table 1. Clinical characteristics and medication use of subjects.

	S (n= 39)	S+ E (n= 40)	P-value S vs S+ E
Male	25 (64%)	36 (90%)	0.0061
Age	69 [63-74]	65 [58.8- 72.8]	0.1911
BW (kg)	62.5 [55.8-73.2)	64.3 [58.3-78.2]	0.1075
BMI (kg/m2)	24.3 [22.6-26.9]	24.2 [22.1- 26.9]	0.7618
Metabolic syndrome	18 (46%)	14 (35%)	0.3127
Hypertension	31 (79%)	25 (63%)	0.0966
Diabetes	16 (41%)	23 (58%)	0.1431
Dyslipidemia	37 (95%)	35 (88%)	0.2490
Family History	10 (26%)	6 (15%)	0.2394
Smoking	14 (36%)	15 (38%)	0.8825
Aspirin	39 (100%)	40 (100%)	-
Thienopyridine	39 (100%)	40 (100%)	-
Other antiplatelets	0 (0%)	1 (3%)	0.3204
Anticoagulant	3 (8%)	4 (10%)	0.7182
Statin	19 (49%)	22 (55%)	0.5764
β-blocker	21 (54%)	29 (73%)	0.8550
ACE inhibitor	5 (13%)	13 (33%)	0.0371
ARB	19 (49%)	13 (33%)	0.1421
Ca-blocker	24 (62%)	16 (40%)	0.0556
Nitrate	13 (33%)	11 (28%)	0.5730
Oral antidiabetics	10 (26%)	17 (43%)	0.1142
Insulin	7 (18%)	3 (8%)	0.1626
Acute coronary syndrome	10 (26%)	13 (33%)	0.5023

Data are expressed as medians and interquartile ranges for continuous variables and as percentages for categorical variables. BW = body weight, BMI = body mass index, ACE = angiotensin-converting enzyme inhibitors, ARB = angiotensin II receptor blockers, S = statin group, S＋E = statin and ezetimibe group.

Lipid Profile and Cholesterol Synthesis/Absorption Markers

Blood samples were obtained both at baseline and after 6–8 months of follow-up. Although both treatment groups achieved significant reduction of total cholesterol (T-Chol) and LDL-C during the treatment period, greater reduction was observed in the S＋E group than in the S group. As a result, T-Chol and LDL-C were significantly lower in the S＋E group than in the S group at follow-up (S group vs. S＋E group; T-Chol, p=0.0092; LDL-C, p=0.0156) ( Table 2 ) . We also examined campesterol and sitosterol as cholesterol absorption markers and lathosterol as a cholesterol synthesis marker. Consistent with a previous report, statin treatment significantly increased campesterol and sitosterol levels, suggesting enhanced cholesterol absorption (campesterol, 4.1±2.1–5.1±2.1 µg/mL, p=0.0014; sitosterol, 2.1±1.0–2.7±1.1 µg/mL, p=0.0004). Conversely, we found significant inhibition of cholesterol absorption in the S＋E group (campesterol, 3.7±1.4–2.2±0.9 µg/mL, p＜0.0001; sitosterol, 2.0±1.0–1.4±0.6 µg/mL, p＜0.0001) ( Table 2 ) . Lathosterol level did not change in the S group, but modestly increased in the S＋E group (1.1±0.4–1.3±0.4 µg/mL, p=0.0446) ( Table 2 ) . The detailed lipid profiles in patients with ACS and patients with stable CAD are described in Supplemental Tables 1 and 2 . As reported previously ⁷⁾ , the plaque regression effect was greater in patients with ACS in both groups ( Supplemental Table 3 ) .

Table 2. Changes of lipid profiles, cholesterol absorption/synthesis markers, oxidized LDL, and high-sensitivity C-reactive protein at baseline and F/U in both groups.

	S (n=27-39)			S+E (n=30-40)			P-value (t-test) S vs. E+S at Baseline	P-value (t-test) S vs. E+S at F/U
	Baseline	F/U	P-value (paired t-test)	Baseline	F/U	P-value (paired t-test)
T-Chol (mg/dL)	172 [150-195]	152 [131-173]	0.0013	151 [130-188)	132 [123.3-153.8)	0.0007	0.0961	0.0092
TG (mg/dL)	123 [98.3-169.3)	118 [78-141)	0.1589	121 [90.8-157]	106.5 [76.3-197]	0.9354	0.7712	0.2158
HDL-C (mg/dL)	43.5 [38-54.5]	45 [38-52.3]	0.2888	38.5 [31.3-44.8)	41 [35.3-47.8)	0.0136	0.0091	0.0414
LDL-C (mg/dL)	96 [82.5-112.3)	83.5 [65.8-100)	0.0009	81 [73-118]	62.5 [54-77]	0.0001	0.1470	0.0156
Campesterol (µg/mL)	3.6 [2.8-4.4]	4.7 [3.6-6.5]	0.0014	3.5 [2.6-4.5]	1.9 [1.4-2.8]	＜0.0001	0.7399	＜0.0001
Sitosterol (µg/mL)	1.8 [1.3-2.5]	2.4 [1.8- 3.2]	0.0004	1.6 [1.2-2]	1 [1-1.5]	＜0.0001	0.6061	＜0.0001
Lathosterol (µg/mL)	1 [1-1]	1 [1-1]	0.7076	1 [1-1]	1.1 [1-1.3]	0.0856	0.8626	0.0446
Campesterol/ Lathosterol	3.4 [2.45-4.3]	4.5 [3.2-6.1]	＜0.0001	3.5 [2.5-4.5]	1.5 [1.2-2.1]	＜0.0001	0.8812	＜0.0001
MDA-LDL (U/L)	63 [51-77)	60.5 [49.5-81.3]	0.9330	66.5 [45.5-94.3]	63.5 [45.5-75.8]	0.1146	0.4596	0.5867
CRP (mg/dL)	0.1 [0.06-0.37]	0.09 [0.03-0.17]	0.2617	0.12 [0.07-0.79]	0.1 [0.04-0.3]	0.0524	0.325	0.5725
hsCRP (mg/dL)	0.32 [0.073-0.66)	0.062 [0.031-0.125]	0.0002	0.31 [0.12-0.75]	0.077 [0.037-0.22]	0.0028	0.5978	0.1971

Data are expressed as medians and interquartile ranges. T-Chol = total cholesterol, TG = triglyceride, HDL-C = high-density lipoprotein cholesterol, LDL-C = low-density lipoprotein cholesterol, F/U = follow-up, S = statin group, S＋E = statin and ezetimibe group.

Supplemental Table 1. The lipid profile in ACS and stable CAD of the statin group.

	ACS (n = 10)			Stable CAD (n = 29)
	Baseline	F/U	P-value Paired t-test	Baseline	F/U	P-value Paired t-test
T-Chol (mg/dL)	180.5 [154.5-210]	162 [127.5-180.8]	0.0655	166 [147.5-190.5]	151 [135.5-167]	0.0065
TG (mg/dL)	113.5 [105.5-135.5]	123.5 [62.3-177.3]	0.8454	128 [94.3-182]	118 [81.5-140.5]	0.030
HDL-C (mg/dL)	46.5 [33-56]	52 [38.5-63.3]	0.1925	43.5 [39.3-51.5]	43.5- [38-50]	0.7687
LDL-C (mg/dL)	105 [91.8-126.3]	73.5 [63.5-104.3]	0.0346	93 [79-114]	84 [67.3-99.5]	0.0075
Campesterol (µg/mL)	3.3 [2.6-6.2]	4.6 [3.2-8.2]	0.3081	3.9 [2.7-4.4]	4.7 [3.7-6.0]	0.0898
Sitosterol (µg/mL)	1.8 [1.2-3.0]	2 [1.5-4.5]	0.1987	1.7 [1.3-2.5]	2.4 [2.0-3.1]	0.0497
Lathosterol (µg/mL)	1 [1-1.3]	1 [1-1]	0.8580	1 [1-1]	1 [1-1	0.7572
Dietary	687.7 [441.8-824.3]	493.7 [396.5-786.0]	0.6443	691.9 [525.6-1098.6]	688.6 [459.9-1265.5]	0.9844
β-epoxy-C	105.3 [69.4-132.5]	92.6 [58.7-149.2]	0.9674	112.0 [69.0-194.4]	104.9 [63.0-187.9]	0.5645
7-keto-C	350.0 [228.7-385.4]	244.2 [175.2-432.9]	0.8143	328.5 [215.5-590.6]	339.9 [224.5-638.3]	0.4949
Intrinsic	552.6 [491.7-750.1]	614.5 [427.9-688.7]	0.3694	448.1 [389.9-592.1]	474.1 [379.4-638.2]	0.8130
4β-OH-C	94.3 [55.9-114.7]	69.0 [46.2-86.1]	0.0100	81.4 [52.7-100.7]	72.1 [49.7-128.2]	0.5712
27-OH-C	493.8 [395.4-610.0]	506.6 [347.6-587.6]	0.5864	344.7 [276.1-469.1]	364.3 [303.9-432.1]	0.4747
Total oxysterol	1453.8 [1215.4-1646.0]	1341.6 [1064-1764.1]	0.8657	1344.8 [1094.2-1636.0]	1300.0 [1049.5-2155.1]	0.7198

Data are expressed as medians and interquartile ranges. T-Chol = total cholesterol, TG = triglyceride, HDL-C = high-density lipoprotein cholesterol, LDL-C = low-density lipoprotein cholesterol, β-epoxy-C =β-epoxy-cholesterol, 7-keto-C = 7-keto-cholesterol, 4β-OH-C = 4β-hydroxy- cholesterol, 27-OH-C = 27-hydroxy-cholesterol, F/U = follow-up, ACS = acute coronary syndrome, CAD = coronary artery disease.

Supplemental Table 2. The lipid profile in ACS and stable CAD of the statin＋ezetimibe group.

	ACS (n = 13)			Stable CAD (n = 27)
	Baseline	F/U	P-value Paired t-test	Baseline	F/U	P-value Paired t-test
T-Chol (mg/dL)	170 [125.5-195.5]	147 [129-161]	0.1563	148.5 [132.3-185.8]	130 [123-148]	0.0007
TG (mg/dL)	144 [102-208.5]	122 [89.5-208.5]	0.5930	109 [85-153]	105 [67-191]	0.4553
HDL-C (mg/dL)	33 [27.5-47.5]	39 [35-53.5]	0.0125	39 [34-45]	41 [35-46]	0.3552
LDL-C (mg/dL)	83 [69.5-125]	61 [46.5-91.5]	0.0771	80 [73-118]	63 [54-75]	0.0003
Campesterol (µg/mL)	3.1 [2.35-3.95]	1.7 [1.2-2.6]	0.0081	3.5 [2.7-4.9]	2 [1.6-2.9]	< 0.0001
Sitosterol (µg/mL)	1.4 [1.15-1.95]	1 [1-1.4]	0.1291	1.9 [1.2-2.7]	1 [1-1.6]	0.0007
Lathosterol (µg/mL)	1 [1-1]	1.1 [1-1.4]	0.7407	1 [1-1]	1.1 [1-1.4]	0.1196
Dietary	512.2 [390.6-1833.3]	348.3 [254.7-1104.9]	0.1539	621.8 [491.8-844.4]	584.5 [456.3-896.7]	0.9746
β-epoxy-C	117.2 [62.3-216.9]	66.7 [52.6-129.3]	0.0701	112.3 [76.5-138.3]	105.2 [65.1-135.8]	0.4222
7-keto-C	258.3 [162.8-960.8]	130.5 [107.6-575.8]	0.2219	308.1 [216.2-447.9]	288.4 [192.7-403.3]	0.5023
Intrinsic	527.3 [395.5-681.9]	450.0 [294.7-527.5]	0.0049	471.4 [404.1-562.0]	408.8 [319.7-508.3]	0.0006
4β-OH-C	81.6 [44.8-120.2]	50.7 [35.3-84.3]	0.0123	68.2 [52.2-96.5]	58.0 [43.8-85.0]	0.0696
27-OH-C	417.6 [321.3-480.8]	355.2 [246.6-444.0]	0.0578	375.0 [323.6-462.8]	316.8 [241.6-411.7]	0.0002
Total oxysterol	1217.1 [984.3-2857.3]	952.3 [784.8-1870.4]	0.1231	1407.8 [11.45.3-1667.5]	1281.0 [1026.2-1630.4]	0.5715

Data are expressed as medians and interquartile ranges. T-Chol = total cholesterol, TG = triglyceride, HDL-C = high-density lipoprotein cholesterol, LDL-C = low-density lipoprotein cholesterol, β-epoxy-C =β-epoxy-cholesterol, 7-keto-C = 7-keto-cholesterol, 4β-OH-C = 4β-hydroxy- cholesterol, 27-OH-C = 27-hydroxy-cholesterol, F/U = follow-up, ACS = acute coronary syndrome, CAD = coronary artery disease.

Supplemental Table 3. The plaque burden at baseline and F/U in ACS and stable CAD.

A. Statin group
	ACS (n = 10)			Stable CAD (n = 29)
	Baseline	F/U	P-value (paired t-test)	Baseline	F/U	P-value (paired t-test)
Plaque Burden (%)	48.6 [46.9-54.8]	46.0 [44.5-49.0]	0.0028	45.8 [38.5-51.0]	46.0 [38.6-53.8]	0.7517
B. Statin＋Ezetimibe group
	ACS (n = 13)			Stable CAD (n = 27)
	Baseline	F/U	P-value (paired t-test)	Baseline	F/U	P-value (paired t-test)
Plaque Burden (%)	47.6 [34.2-59.6]	40.23 [30.1-52.7]	0.0053	41.1 [35.3-55.1]	41.8 [35.2-54.5]	0.0071

Data are expressed as medians and interquartile ranges. ACS = acute coronary syndrome, CAD = coronary artery disease, F/U = follow-up.

Change in Oxidized LDL, hs-CRP, and Oxysterols

We also analyzed oxidized LDL, hs-CRP, and oxysterols as exploratory endpoints. MDA-LDL, a representative oxidized LDL, tended to decrease only in the S＋E group, but the difference was not statistically significant (S group: 66±24–66±25 U/L, p=0.9330; S＋E group: 70±28–63±23 U/L, p=0.1146) ( Table 2 ) . Hs-CRP significantly decreased in both treatment groups (S group: 0.4±0.3–0.1±0.2 mg/dL, p=0.0002; S＋E group: 0.4±0.3–0.2±0.2 mg/dL, p=0.0028) ( Table 2 ) . Total oxysterol tended to decrease in the S＋E group, although it did not reach statistical significance ( Table 3 ) In contrast, total oxysterol did not change in the S group. We observed a significant decrease in the levels of β-epoxycholesterol, 4β-hydroxycholesterol, and 27-hydroxycholesterol only in the S＋E group (β-epoxycholesterol, 131±110–98±49 ng/mL, p=0.0475; 4β-hydroxycholesterol, 86±49–64±29 ng/mL, p=0.0042; 27-hydroxycholesterol, 407±107–339±106 ng/mL, p＜0.0001) ( Table 3 ) .

Table 3. Changes of the oxysterol at baseline and F/U in both groups.

	S (n = 38)			S＋E (n = 39)			S vs. S＋E
	Baseline	F/U	P-value Paired t-test	Baseline	F/U	P-value Paired t-test	P-value F/U t-test
Dietary	689.8 [515.8-912.8]	610.4 [442.8-1155.8]	0.8855	588 [441.5-854.4]	574.8 [348.3-914.6]	0.1619	0.0987
β-epoxy-C	107.2 [70.7-186.9]	102.7 [63.5-176.4]	0.5736	112.6 [67.8-150.4]	99.2 [56.8-130.6]	0.0475	0.0986
7-keto-C	339.3 [217.1-521.9]	306.0 [210.5-581.2]	0.4844	260.3 [175.5-451.9]	248.3 [134.9-441.7]	0.2815	0.1553
Intrinsic	473.9 [412.6-654.7]	486.9 [389.7-648.9]	0.5193	505.5 [405.3-603.4]	416.5 [304.6-510.84]	＜0.0001	0.0055
4β-OH-C	86.1 [53.4-104.4]	71.1 [49.1-98.5]	0.9921	73.2 [52.3-109.4]	57.3 [38.9-83.3]	0.0042	0.0306
27-OH-C	372.2 [293.8-499.2]	382.4 [317.1-512.3]	0.7990	384.2 [324.0-480.3]	325.6 [243.7-422.7]	＜0.0001	0.0094
Total oxysterol	1348.5 [1099.2-1627.4]	1313.1 [1053.5-2042]	0.8940	1361.4 [1066.9-1813.9]	1163.9 [863.7-1577.2]	0.0972	0.0665

Data are expressed as medians and interquartile ranges. β-epoxy-C =β-epoxy-cholesterol, 7-keto-C = 7-keto-cholesterol, 4β-OH-C = 4β-hydroxy-cholesterol, 27-OH-C = 27-hydroxy-cholesterol, F/U = follow-up, S = statin group, S＋E = statin and ezetimibe group.

Change in Plaque Burden Assessed with IVUS

We assessed the coronary plaque burden of nonculprit lesion before and after the treatment period. Plaque burden significantly decreased from 47% to 44% in the S＋E group, whereas it did not change in the S group (from 47% to 46%) ( Fig.2A , Table 4 ) . The percent change in plaque burden was significantly greater among patients in the S＋E group than in the S group (−6.14% vs. −1.18% for each group, p=0.042) ( Fig.2B ) . There was no change in vessel area and lumen area over time in both groups; the vessel remodeling did not contribute the coronary plaque regression ( Table 4 ) . Next, we analyzed plaque composition as defined by IB-IVUS. According to the IB value, each plaque component was classified into four categories: lipid pool, fibrosis, dense fibrosis, or calcification. In this substudy, the proportion of lipid pool and fibrosis did not change between the S and S＋E groups ( Fig.3 ) .

Fig.2. Coronary plaque burden of nonculprit lesion.

(A) Plaque burden significantly decreased in the S＋E group, whereas it did not change in the S group.

(B) The percent change in plaque burden was significantly greater in the S＋E group than in the S group.

f/u=follow-up, S=statin group, S＋E=statin and ezetimibe group.

Table 4. The vessel area, lumen area, plaque area, and plaque burden at baseline and F/U in both groups.

	S (n = 39)			S＋E (n = 40)			P-value (ANCOVA) S vs. S＋E
	Baseline	F/U	P-value (paired t-test)	Baseline	F/U	P-value (paired t-test)
Vessel area (mm2)	13.9±5.4	13.7±5.4	0.1638	15.0±6.1	14.8±5.7	0.5165	0.6488
Lumen area (mm2)	7.3±3.0	7.4±3.2	0.5610	7.6±3.0	8.0±3.0	0.0172	0.1574
Plaque area (mm2)	6.6±3.0	6.3±2.8	0.0370	7.4±4.6	6.9±4.3	0.0050	0.4448
Plaque burden (%)	47±11	46±10	0.3054	47±14	44±14	＜0.0001	0.0532

Data are expressed as mean±SD. F/U = follow-up, S = statin group, S＋E = statin and ezetimibe group.

Fig.3. Analysis of plaque composition defined by IB-IVUS.

The proportion of lipid pool and fibrosis did not change between the S and S＋E groups.

f/u=follow-up, S=statin group, S＋E=statin and ezetimibe group

Correlation between Plaque Burden and Biomarkers

We evaluated the correlation between the lipid biomarkers and the plaque regression. In the current study, there was no significant correlation between LDL-C reduction and plaque burden regression ( Fig.4A ) . Additionally, there was no correlation between high-density lipoprotein cholesterol (HDL-C) and percent change of plaque burden in both groups ( Supplemental Fig.1 ) . In contrast, there was a significant correlation between campesterol reduction and plaque regression (p=0.0028, correlation coefficient=0.336) ( Fig.4B ) . These results suggested that inhibition of cholesterol absorption by ezetimibe rather than LDL-C lowering may contribute to the greater plaque regression in the S＋E group than in the S group. Furthermore, the current sub-analysis showed a correlative trend between lowering of oxysterols and plaque regression (p=0.0520) ( Fig.5A ) . It was noteworthy that there was a significant correlation between 27-hydroxycholesterol reduction and plaque regression (p=0.0004, correlation coefficient=0.392) ( Fig.5B ) . Additional multivariate adjusting analysis using factors that were significantly correlated and therapeutic interventions showed that the decrease of 27-hydroxycholesterol significantly correlated with the reduction of coronary plaque burden (p=0.0130) ( Supplemental Table 4 ) . These results suggest that 27-hydroxycholesterol reduction would be one of the novel molecular mechanisms of plaque regression in the addition of ezetimibe to statin therapy.

Fig.4. The correlation between lipid biomarkers and plaque regression.

(A) There was no significant correlation between LDL-C reduction and plaque burden regression.

(B) There was a significant correlation between campesterol reduction and plaque regression.

Supplemental Fig.1. The effects of therapeutic interventions and interactions between lipids.

There was no significant correlation between HDL-C and plaque burden regression. HDL-C=high-density lipoprotein cholesterol.

Fig.5. The correlation between oxysterol and plaque regression.

(A) There was a correlative trend between lowering of oxysterols and plaque regression.

(B) There was a significant correlation between 27-hydroxycholesterol reduction and plaque regression.

Supplemental Table 4. The multivariate analysis using factors that were significantly correlated and therapeutic interventions.

	β	P
Therapeutic interventions	0.0065	0.9971
Sex	0.66	0.6768
ACE inhibitor	-2.03	0.1984
LDL-C	-0.031	0.4328
Campesterol	0.51	0.5110
Total oxysterol	3.92	0.3110
27-OH-C	0.032	0.0130

ACE inhibitor = angiotensin-converting enzyme inhibitor, LDL-C = low-density lipoprotein cholesterol, 27-OH-C = 27-hydroxy-cholesterol.

Discussion

The novel findings of this subgroup analysis of the CuVIC trial include the following: 1) ezetimibe in combination with statin significantly reduced the coronary plaque burden compared with statin monotherapy and 2) suppression of campesterol and decrease of oxysterols, especially 27-hydroxycholesterol, correlated with the reduction of coronary plaque burden.

Inhibition of cholesterol synthesis by statin treatment results in enhanced absorption of cholesterol. Ezetimibe is a cholesterol absorption inhibitor that prevents the absorption of dietary and biliary cholesterol from the small intestine. Dual inhibition of cholesterol synthesis and absorption led to further reduction in LDL-C levels and produced better clinical outcomes than statin monotherapy ⁴⁾ . In addition, combination therapy with statin and ezetimibe produced coronary plaque regression compared with statin monotherapy. Our sub-analysis reported similar results to those previous studies ( Table 2 ) . Although this finding might be attributed to the larger lipid-lowering by inhibition of cholesterol absorption, the molecular mechanisms driving the additional benefit of ezetimibe on plaque regression were previously unknown.

To explore the beyond-cholesterol effects of the combination of statin plus ezetimibe, we analyzed detailed biomarkers (such as oxidized LDL, hs-CRP, and oxysterols) as exploratory endpoints. Oxysterols in the diet are absorbed from the intestine and incorporated into chylomicrons and chylomicron remnants in humans ¹⁴⁾ , and ezetimibe inhibits their absorption ¹⁰⁾ . We have shown that oxysterols in the serum were decreased by 7-month treatment with ezetimibe in patients with CAD ¹²⁾ . We also have previously shown that dietary oxysterols accelerate atherosclerotic plaque destabilization in hypercholesterolemic mice, which was associated with increased monocyte infiltration/activation, monocyte chemoattractant protein-1 expression, and matrix metalloproteinase activity ¹¹⁾ . Furthermore, ezetimibe monotherapy ameliorated plaque destabilization associated with decrease in serum LDL-C, oxidized LDL, and oxysterol levels ¹¹⁾ . In a rabbit model of arterial balloon injury, high cholesterol diet and infusion of angiotensin II increased serum oxysterols, and ezetimibe monotherapy decreased oxysterols and enhanced re-endothelialization. Specific oxysterols, 7-ketocholesterol and 7α-hydroxycholesterol, inhibited the proliferation of vascular endothelial cells; 7-ketocholesterol and 27-hydroxycholesterol induced tissue factor in vascular smooth muscle cells, which were blocked by a reactive oxygen species scavenger ¹⁵⁾ .

To the best of our knowledge, there are no reports that evaluated the correlation with oxysterols and coronary plaque burden in humans. We observed a significant decrease in the levels of β-epoxycholesterol, 4β-hydroxycholesterol, and 27-hydroxycholesterol only in the S＋E group and found a significant positive correlation between oxysterols and plaque regression ( Table 3 , Fig.5 ) . To our knowledge, the present study demonstrates for the first time that the decrease of 27-hydroxycholesterol correlated with the reduction of coronary plaque burden in patients treated with ezetimibe and statins. It is reported that 27-hydroxycholesterol is an abundant oxysterol metabolized by CYP7B1 and a competitive antagonist of estrogen receptor that regulates endothelial nitric oxide synthase expression in the vascular endothelium ¹⁶⁾ , and that increases of 27-hydroxycholesterol via cyp7b1 deletion promoted atherosclerosis in apoe^−/− mice without altering lipid status ¹⁷⁾ . The potential mechanisms of the association between 27-hydroxycholesterol and plaque progression may include an interference of estrogen receptor that mediates atheroprotective effects of estrogen or other yet unknown mechanisms. Although more research is needed to elucidate the mechanism, plaque regression by lipid-lowering therapy is associated with improved prognosis, and the findings of this study are clinically important ¹⁸⁾ . Contrary to our expectations, there were no differences on plaque composition that might represent plaque stabilization assessed by IB-IVUS between these treatment groups. In the present study, patients with ACS accounted for less than 30% of the study population. According to a previous report, plaque regression was greater in patients with ACS than in patients with stable angina pectoris ⁷⁾ . Furthermore, the follow-up period was relatively short (i.e., 7 months), and longer follow-up may be necessary to evaluate the impact of the combination therapy on coronary atherosclerotic plaque composition.

Note that this study has limitations. First, this is a sub-analysis of the CuVIC trial, in which the serial examinations with IVUS were left for the decision of physicians. Second, the small number of study patients might limit the power to clarify the additional benefit of ezetimibe on the proportion of plaque component and unrevealed factor for plaque regression.

Conclusions

In conclusion, in patients with CAD, ezetimibe in combination with statin achieved significantly lower LDL cholesterol, campesterol, and several oxysterols than statin monotherapy. The lowering of campesterol and 27-hydroxycholesterol, but not LDL cholesterol, had a significant positive correlation with plaque regression.

Acknowledgements and Notice of Grant Support

Y. Nakano, T. Matoba (Department of Cardiovascular Medicine, Kyushu University), and J. Kishimoto (Center for Clinical and Translational Research, Kyushu University) had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. We appreciate Maki Kimura and Hiroko Watanabe for their excellent data management. This work was supported by JSPS KAKENHI Grant Number JP21H02916 for TM.

Conflict of Interest

TM has received personal fees from Abbott, Bayer Yakuhin, and MSD; research funding from Amgen, Bayer Yakuhin and Kowa outside the submitted work. HT has received lecture fees (Kowa, Teijin Pharma, Nippon Boehringer Ingelheim, Mitsubishi Tanabe Pharma, Pfizer Japan, Ono Pharmaceutical, Daiichi Sankyo, Novartis Pharma, Bayer Yakuhin, Otsuka Pharmaceuticalm, and AstraZeneca); manuscript fees (Nippon Rinsho); research funding (Mitsubishi Tanabe Pharma, Nippon Boehringer Ingelheim, IQVIA Services Japan, MEDINET, Medical Innovation Kyushu, Kowa, Daiichi Sankyo, Johnson & Johnson, and NEC Corporation) and scholarship funds (Abbott Medical Japan, Otsuka Pharmaceutical, Boston Scientific Japan, Ono Pharmaceutical, Bayer Yakuhin, Nippon Boehringer Ingelheim, St.Mary’s Hospital, Teijin Pharma, Daiichi Sankyo, and Mitsubishi Tanabe Pharma), outside the submitted work. The other authors report no conflicts.