Difference of Tissue Characteristics Between Early and Late Restenosis After Second-Generation Drug-Eluting Stents Implantation　― An Optical Coherence Tomography Study ―

Abstract

Background: The mechanism and time course of in-stent restenosis (ISR) after implantation of second-generation DES have not yet been fully elucidated. We sought to evaluate the differences in tissue characteristics between the different phases of ISR after second-generation DES implantation using optical coherence tomography (OCT).

Methods and Results: From June 2010 to December 2015, 324 consecutive patients with 337 ISR lesions underwent OCT. Of these, we analyzed 53 lesions in 53 patients who had their first ISR after second-generation DES implantation and underwent OCT before any procedures. According to the timing of ISR, the patients were divided into the early group (within 1 year: E-ISR, n=30) and late group (beyond 1 year: L-ISR, n=23). Quantitative parameters and qualitative characteristics of the neointima were evaluated. In the minimum lumen area site analysis, the E-ISR group had more frequently homogeneous intima than the L-ISR group (26.7% vs. 4.4%, P=0.02). The frequencies of neointima with lipid-laden, thin-cap fibroatheroma, neovascularization and macrophage infiltration were significantly higher in the L-ISR group than in the E-ISR group (30.0% vs. 69.6%, P<0.01; 0.0% vs. 26.1%, p <0.01; 6.7% vs. 26.1%, P=0.049; 3.3% vs. 26.1%, P=0.01, respectively).

Conclusions: Neointimal tissue characteristics differed between E-ISR and L-ISR after second-generation DES implantation. E-ISR was mainly caused by neointimal hyperplasia, whereas neoatherosclerosis was the main mechanism of L-ISR.

First-generation drug-eluting stents (DES), which inhibit intimal hyperplasia, have dramatically reduced the rate of in-stent restenosis (ISR) and subsequent target lesion revascularization (TLR) within the first year of stent implantation compared with bare-metal stents (BMS).^1,2 However, in real-world practice, late adverse events such as very late stent thrombosis (VLST) and late TLR beyond 1 year have emerged as unsolved issues after first-generation DES implantation.³

Previous optical coherence tomography (OCT) studies have reported that the mechanism and temporal course of ISR differ between BMS and first-generation DES.^4–7 Although in-stent neoatherosclerosis has been suggested as a cause of late adverse events with both BMS and first-generation DES, pathological studies demonstrated that the earliest atherosclerotic change with foamy macrophage infiltration began at 4 months after first-generation sirolimus eluting-stent (SES) implantation, whereas the same change after BMS implantation was not seen before 2 years and remained a rare finding until 4 years.^8,9

Second-generation DES have further improved the safety and similar efficacy compared with first-generation DES.^10,11 Recently, Otsuka et al reported from an autopsy study in human coronary arteries that cobalt chromium everolimus-eluting stents (CoCr-EES) yielded a lower inflammation score without any hypersensitivity and less fibrin deposition than first-generation SES and paclitaxel-eluting stents, while the frequency of neoatherosclerosis was comparable among them.¹² Those findings indicate that in-stent neoatherosclerosis remains an unsolved issue even in the era of second-generation DES. To date, however, the mechanism and temporal course of ISR after second-generation DES implantation have not yet been fully evaluated, so we used OCT in the present study to evaluate the differences in tissue characteristics in relation to the timing of ISR after second-generation DES implantation.

Methods

Study Population

From June 2010 to December 2015, 324 consecutive patients with 337 ISR lesions underwent OCT. Inclusion criteria were as follows: (1) lesions with first ISR after implanting second-generation DES (XienceV^TM, Xience Prime^TM, Xience Alpine^TM, Xience Xpedition^TM, Abbot Vascular, Santa Clara, CA, USA; Promus^TM, Promus Element^TM, Promus Premier^TM, Boston Scientific, Natick, MA, USA) for de novo lesion and (2) OCT performed before any procedure. Exclusion criteria were hemodynamic instability, bypass graft lesions, poor imaging quality and acute myocardial infarction.

According to the time of diagnosed ISR, we divided the patients into 2 groups: (1) early ISR (E-ISR), defined as the first ISR observed within the first year after the procedure, and (2) late ISR (L-ISR), defined as the first ISR observed beyond the first year and not detected by follow-up angiography within the first year. All study patients gave written informed consent for the procedure and the follow-up protocol, which was approved by the institutional review board of Kokura Memorial Hospital.

Quantitative Coronary Angiographic Analysis

Coronary angiography (CAG) was performed after intracoronary administration of 0.2 mg nitroglycerin. Quantitative CAG (QCA) was performed before any procedure in the respective ISR phase using a guiding catheter to calibrate the magnification and a validated automated edge detection algorithm (CASS 5.11, Pie Medical Imaging, Eindhoven, The Netherlands). Follow-up CAG was scheduled 6–12 months after the initial procedure, regardless of clinical symptoms. Patients who underwent unscheduled follow-up angiography for clinical reasons within the 12 months were included in the angiographic analysis. The analyses were performed independently by 2 experienced observers in the Angiographic Core Laboratory, Kokura Memorial Hospital, and the observers were blinded to the clinical information. ISR was defined as a percent diameter stenosis >50% within the stent. Angiographic restenosis was classified according to Mehran et al.¹³

OCT Image Acquisition

OCT images were evaluated before any procedures. We performed OCT using one of the following systems: M2 OCT system (Light Lab Imaging, Westford, MA, USA; C7XR Fourier-Domain System (St Jude Medical, St Paul, MN, USA; LUNAWAVE (Terumo, Tokyo, Japan). Using the M2 OCT system, the occlusion balloon catheter (Helios occlusion balloon catheter, LightLab Imaging) was advanced proximal to the implanted stent with a 0.014-inch guidewire during angiography, and then the guidewire was exchanged for an OCT imaging wire, which was then positioned distal to the stent. During image acquisition, lactated Ringer solution was continuously flushed through the inner lumen of the balloon occlusion catheter. Motorized pullback OCT imaging was performed at a rate of 1.0 mm/s through the stent. Images were acquired at 15.6 frame/s and digitally archived. The C7XR and LUNAWAVE systems used a conventional wire to cross the segment of interest. The OCT imaging catheter (Dragonfly, St Jude Medical; FastView, Terumo) was then advanced distally to the stented lesion. Pullback was performed during continuous injection of contrast media through the guide catheter by injection pump. OCT images were acquired automatically at a pullback rate of 20 mm/s (100 frame/s) and that of optical frequency domain imaging (OFDI) at a pullback rate of 20 mm/s (160 frame/s).

OCT Image Analysis

OCT pullback images were analyzed offline using LightLab OCT imaging proprietary software (LightLab Imaging) or Terumo software. For quantitative and qualitative analyses, cross-sectional OCT images were analyzed at 1-mm intervals within the stented lesion. The lumen and stent were manually traced and the area of neointimal hyperplasia (NIH=stent area−lumen area) was calculated for each frame. The percent NIH area was also calculated as: (NIH area/stent area)×100. Bifurcation cross-sections with side branches were excluded from the analysis. All OCT and OFDI images were analyzed by 2 investigators (H.J. and S.K.) who were blinded to the patient’s information.

OCT Quantitative and Qualitative Evaluations

Neointimal tissue was classified as follows: (1) homogeneous neointima was defined as a uniform signal-rich band without focal variation or attenuation; (2) heterogeneous neointima had focally changing optical properties and various backscattering patterns; (3) lipid-laden neointima had a diffuse border, signal-poor region with marked attenuation.^4,14,15 In addition, thin-cap fibroatheroma (TCFA)-containing intima was defined as a fibrous cap thickness at the thinnest part <65 µm and an angle of lipidic tissue >180 degrees.^14,16 Intimal disruption was defined as a discontinuity of the lumen border with a visible cavity.^14,17 Calcification was defined as a well-delineated, signal-poor region with sharp borders.⁶ Furthermore, in-stent neoatherosclerosis was defined as a lesion with lipid-laden neointima, neointima with calcification within the neointima, a TCFA-like neointima or neointimal rupture.¹⁸ Macrophage infiltration was defined as linear, strong OCT images on the plaque surface accompanied by high attenuation.^19,20 Neovascularization was defined as a small vesicular or tubular structure with a diameter <200 µm.²¹ Thrombus was defined as a mass protruding into the lumen and a dimension >250 µm) and thrombus was categorized by the presence of shadowing.¹⁴ The peri-strut low intensity area (PLIA) was defined as a region around the stent struts with homogeneous lower intensity than the surrounding tissue on OCT images without signal attenuation.^22,23 Representative OCT images of ISR are shown in Figure 1.

Figure 1.

Representative optical coherence tomography images of ISR after second-generation DES implantation. (A) homogeneous neointima, (B) heterogeneous neointima, (C) lipid-laden neointima, (D) thin-cap fibroatheroma, (E) intimal disruption, (F) macrophage infiltration (blue arrows), (G) neovascularization (blue arrowheads), (H) thrombus (red arrows), and (I) peri-stent low intensity area (yellow arrowheads). DES, drug-eluting stent; ISR, in-stent restenosis.

When the qualitative analysis by the 2 investigators differed, a consensus was reached and used as the final decision. To test for interobserver variability of the qualitative OCT analysis, a total of 100 cross-sections within the restenotic lesions from 10 patients by each 10 cross-sections were randomly selected and analyzed.

Statistical Analysis

Data are presented as values and percentages, mean±SD, or median (interquartile range). Categorical variables were compared between the 2 groups with the chi-square test or Fisher’s exact test, as appropriate. Continuous variables were compared between groups using the unpaired t test or the Mann-Whitney U test, based on the data distribution. Intra- and interobserver variabilities for qualitative variables were assessed with a k-test.

We estimated the group means or proportions of cross-section-level variables using mixed-effects models and the GLIMMIX procedure in SAS version 9.4 (SAS Institute Inc.). For continuous data, we fitted linear mixed-effects models, which included the intervention groups (E-ISR and L-ISR) as fixed effects, and lesions as random intercepts that followed a normal distribution with a mean 0 and group-specific variances. For binary (presence/absence) data, logistic random-intercept models were fitted as in the case of continuous data. Proportions in each group and odds ratios (ORs) were estimated by transforming fixed-effects contrasts. However, it should be noted that parameters in logistic models represent conditional effects given fixed (groups) and random effects (lesions), rather than marginal predicted probability averaged across patients within a group. The estimated proportions are therefore the estimates of conditional proportions that correspond to random effects set to 0, and conditional ORs for them. Mixed-effects logistic models were estimated by maximizing the Laplace approximation of marginal likelihood, which is reported as most preferable among the available optimization routines in current computing resources.²⁴ Robust variance estimators were used for constructing 95% confidence intervals.

A 2-sided P value <0.05 was considered to indicate statistical significance. Statistical analyses were performed by a physician (H.J.) and a statistician (T.S.), using JMP version 10.1 (SAS Institute Inc., Cary, NC, USA) and SAS version 9.4 (SAS Institute Inc.) for multilevel modeling.

Results

Clinical and Lesional Characteristics

Among 324 patients with 337 lesions, a total of 53 patients with 53 lesions met the inclusion and exclusion criteria and were enrolled in the present study (Figure 2). These patients were divided into the 2 groups: E-ISR (30 patients with 30 lesions) and L-ISR (23 patients, 23 lesions). The patients’ characteristics were comparable between the 2 groups except for age, hyperlipidemia and stent type (Table 1). The lesional characteristics are listed in Table 2. No significant difference was found between groups without culprit vessels.

Figure 2.

Study flow chart. DES, drug-eluting stent; ISR, in-stent restenosis; OCT, optical coherence tomography.

Table 1. Baseline Characteristics of Study Patients With ISR

	Overall (n=53)	E-ISR (n=30)	L-ISR (n=23)	P value
Age, years	66.0±9.5	63.1±9.5	69.9±8.3	0.02
Male	41 (77.4)	24 (80.0)	17 (73.9)	0.60
Risk factors
Hypertension	43 (81.1)	22 (73.3)	21 (91.3)	0.09
Hyperlipidemia	36 (67.9)	17 (56.7)	19 (82.6)	0.04
Diabetes mellitus	33 (62.3)	18 (60.0)	15 (65.2)	0.69
Current smoker	12 (22.6)	7 (23.3)	5 (21.7)	0.89
Past medical history
Previous MI	12 (22.6)	8 (26.7)	4 (17.4)	0.42
Previous CABG	9 (17.0)	7 (23.3)	2 (8.7)	0.15
Cerebrovascular disease	1 (1.9)	0 (0.0)	1 (4.4)	0.19
EF on admission, %	59.1 (50.8–66.8)	57.4 (50.0–66.7)	62.6 (52.3–67.0)	0.54
Reason for stenting				0.28
Stable angina	52 (98.1)	29 (96.7)	23 (100.0)
Unstable angina	1 (1.9)	1 (3.3)	0 (0.0)
Acute MI	0 (0.0)	0 (0.0)	0 (0.0)
Clinical presentation				0.71
Stable	50 (94.3)	28 (93.3)	22 (95.7)
Unstable	3 (5.7)	2 (6.7)	1 (4.4)
Stent type				<0.01
CoCr-EES	42 (79.3)	20 (66.7)	22 (95.7)
PtCr-EES	11 (20.8)	10 (33.3)	1 (4.4)
Elapsed time from index procedure, days	336.0 (200.0–508.5)	212.0 (189.0–288.8)	632.0 (412.0–1,427.0)	<0.01
Imaging technique				0.15
Optical coherence tomography	44 (83.0)	23 (76.7)	21 (91.3)
Optical frequency domain imaging	9 (17.0)	7 (23.3)	2 (8.7)

Values are mean±SD, median (interquartile), or number (%). CABG, coronary artery bypass graft; CoCr, cobalt chromium; EES, everolimus-eluting stent; EF, ejection fraction; E-ISR, early in-stent restenosis; ISR, in-stent restenosis; L-ISR, late in-stent restenosis; MI, myocardial infarction; PtCr, platinum chromium.

Table 2. Baseline Lesion Characteristics of the Study Patients With ISR

	Overall (n=53)	E-ISR (n=30)	L-ISR (n=23)	P value
Vessel				0.04
RCA	27 (50.9)	19 (63.3)	8 (34.8)
LAD	22 (41.5)	8 (26.7)	14 (60.9)
LCX	4 (7.6)	3 (10.0)	1 (4.4)
Mehran classification				0.41
Focal	44 (83.0)	25 (83.3)	19 (82.6)
Diffuse	8 (15.1)	5 (16.7)	3 (13.0)
Multifocal	1 (1.9)	0 (0.0)	1 (4.4)
Reference diameter, mm	2.75±0.57	2.86±0.53	2.61±0.59	0.18
Diameter stenosis rate, %	69.6±9.7	70.8±10.2	68.0±9.0	0.31
Minimum lumen diameter, mm	0.84±0.30	0.83±0.33	0.84±0.26	0.69
Lesion length, mm	9.9±4.6	9.8±5.2	10.1±3.8	0.37

Values are expressed as mean±SD or number (%). LAD, left anterior descending artery; LCX, left circumflex artery; RCA, right coronary artery. Other abbreviations as in Table 1.

OCT Findings

Entire Stent Analysis OCT analysis data for the entire stent are summarized in Table 3. The quantitative analysis was not significantly different between the 2 groups. In the qualitative analysis, neointima with TCFA-like and lipid-laden characteristics was more frequently observed in the L-ISR group than in the E-ISR group (P=0.01 and P<0.01, respectively). Furthermore, neoatherosclerosis was significantly higher in the L-ISR group (weighted estimated of 15.7%, 95% CI: 9.15–25.5) than in the E-ISR group (2.82 %, 95% CI: 1.16–6.71%), with an OR between groups of 0.16 (95% CI: 0.05–0.47, P<0.01). Frequency of thrombus, intimal disruption and calcification were very low in both groups. In addition, PLIA was less frequently observed in the L-ISR group (2.07 %, 95% CI: 0.67–6.22%) than in the E-ISR group (8.97 %, 95% CI: 4.30–17.8%), with an OR between groups of 4.67 (95% CI: 1.16–18.7, P=0.03).

Table 3. OCT Analysis of Entire Stent

	Mean or proportion (95% CI)		Mean difference or OR (95% CI)	P value
	E-ISR	L-ISR	E-ISR vs. I-ISR
Total no. of cross-sections	573	441
Quantitative analysis
Lumen area, mm2	4.33 (3.76–4.90)	3.56 (3.00–4.12)	0.77 (−0.02 to 1.57)	0.06
Stent area, mm2	6.34 (5.63–7.04)	5.81 (5.05–6.56)	0.53 (−0.50 to 1.57)	0.31
NIH area, mm2	2.04 (1.64–2.44)	2.26 (1.74–2.78)	−0.22 (−0.88 to 0.44)	0.51
Percent NIH, %	0.33 (0.29–0.37)	0.38 (0.32–0.45)	−0.05 (−0.13 to 0.03)	0.18
Qualitative analysis
Homogeneous intima, %	74.0 (60.7–83.9)	61.1 (41.9–77.4)	1.81 (0.67–4.86)	0.24
Heterogeneous intima, %	20.2 (12.0–32.0)	13.7 (7.43–23.8)	1.60 (0.64–4.01)	0.31
Lipid-laden, %	2.82 (1.16–6.71)	13.9 (7.64–24.0)	0.18 (0.06–0.55)	<0.01
Neoatherosclerosis, %	2.82 (1.16–6.71)	15.7 (9.15–25.5)	0.16 (0.05–0.47)	<0.01
TCFA-like pattern, %	0.00 (0.00–0.09)	0.79 (0.05–11.1)	0.00 (0.00–0.25)	0.01
Intimal disruption, %	0.00 (0.00–0.10)	0.22 (0.00–15.4)	0.01 (0.00–2.99)	0.11
Calcification, %	0.00 (0.00–0.00)	0.00 (0.00–0.02)	NE	NE
Macrophage infiltration, %	0.03 (0.00–5.09)	2.63 (0.79–8.40)	0.01 (0.00–2.29)	0.10
Neovascularization, %	0.21 (0.00–15.7)	8.06 (4.76–13.4)	0.02 (0.00–2.20)	0.10
Thrombus, %	0.00 (0.00–0.02)	0.01 (0.00–0.22)	0.19 (0.00–8.53)	0.39
Thrombus with shadow, %	0.00 (0.00–0.00)	0.0 (0.00–0.00)	NE	NE
Thrombus without shadow, %	0.00 (0.00–0.00)	0.00 (0.00–0.05)	NE	NE
PLIA, %	8.97 (4.30–17.8)	2.07 (0.67–6.22)	4.67 (1.16–18.7)	0.03
Malaposition, %	0.31 (0.00–69.0)	0.11 (0.00–87.8)	2.72 (0.00–153,185.0)	0.86

Values are expressed as mean, percentage or OR (95% CI). NE, not estimated; NIH, neointimal hyperplasia; OCT, optical coherence tomography; OR, odds ratio; PLIA, peri-low intensity area; TCFA, thin-cap fibroatheroma. Other abbreviations as in Table 1.

Minimum Lumen Area Site Analysis At the site of minimum lumen area, homogeneous neointima was more frequently observed in the E-ISR group than in the L-ISR group (26.7% vs. 4.4%, P=0.02) (Table 4). Conversely, the incidence of neointima with TCFA, macrophage infiltration, neovascularization and lipid-laden characteristics was significantly higher in the L-ISR group than in the E-ISR group (0.0% vs. 26.1%, P<0.01, 3.3% vs. 26.1%, P=0.01, 6.7% vs. 26.1%, P=0.049, 30.0% vs. 69.6%, P<0.01, respectively). Furthermore, neoatherosclerosis was more frequently observed in the L-ISR group than in the E-ISR group (30.0% vs. 73.9%, P<0.01). As shown in Figures 3 and 4, the incidence of neoatherosclerosis increased over time as opposed to the incidence of homogeneous neointima.

Table 4. OCT Analysis of Minimum Lumen Area Site

	Overall (n=53)	E-ISR (n=30)	L-ISR (n=23)	P value
Quantitative analysis
Lumen area, mm2	1.4±0.9	1.5±1.0	1.3±0.7	0.65
Stent area, mm2	5.6±2.0	5.7±2.1	5.4±1.7	0.86
NIH area, mm2	4.2±1.9	4.2±2.1	4.1±1.7	0.89
Qualitative analysis
Homogeneous intima	9 (13.2)	8 (26.7)	1 (4.4)	0.02
Heterogeneous intima	19 (35.9)	13 (43.3)	6 (26.1)	0.19
Lipid-laden	25 (47.2)	9 (30.0)	16 (69.6)	<0.01
Neoatherosclerosis	26 (49.1)	9 (30.0)	17 (73.9)	<0.01
TCFA-like pattern	6 (11.3)	0 (0.0)	6 (26.1)	<0.01
Intimal disruption	2 (3.8)	1 (3.3)	1 (4.4)	0.85
Calcification	0 (0.0)	0 (0.0)	0 (0.0)	NE
Macrophage infiltration	7 (13.2)	1 (3.3)	6 (26.1)	0.01
Neovascularization	8 (15.1)	2 (6.7)	6 (26.1)	0.049
Thrombus	5 (9.4)	3 (10.0)	2 (8.7)	0.87
Thrombus with shadow	3 (5.7)	3 (10.0)	0 (0.0)	0.06
Thrombus without shadow	2 (3.8)	0 (0.0)	2 (8.7)	0.06

Values are expressed as mean±SD or number (%). Abbreviations as in Tables 1,3.

Figure 3.

Incidence of neoatherosclerosis in different phases of in-stent restenosis (ISR) after implantation of second-generation drug-eluting stents.

Figure 4.

Incidence of homogeneous intima in different phases of in-stent restenosis (ISR) after implantation of second-generation drug-eluting stents.

Reproducibility of Qualitative OCT Analysis

Intra- and interobserver variabilities (k values) for the qualitative OCT assessment were as follows: homogeneous neointima (0.95/0.82), heterogeneous intima (0.89/0.81), lipid-laden neointima (0.85/0.85), TCFA-like neointima (1.00/1.00), macrophage infiltration (0.90/0.83), neovascularization (0.82/0.77), and PLIA (0.91/0.82).

Discussion

The main findings of our study are as follows: (1) in-stent neoatherosclerosis was more frequently observed in the L-ISR group than in E-ISR group, and (2) the E-ISR group more frequently had homogeneous neointima than the I-ISR group.

The mechanism of ISR after first-generation DES implantation has been demonstrated in previous studies.^7,9,25,26 Recently, it was demonstrated by intravascular ultrasound (IVUS) that first-generation DES-ISR was mainly caused by NIH and stent underexpansion.^25,26 Furthermore, a pathological study demonstrated that neoatherosclerosis occurred earlier with first-generation DES, which may be related to the presence of durable polymer.⁹ Compared with first-generation DES, second-generation DES are equipped with biocompatible or biodegradable polymer to improve the long-term safety and efficacy.^10,11 Nevertheless, second-generation DES, as well as first-generation DES, are not immune to ISR and subsequent TLR.¹¹ More recently, Goto et al reported from their IVUS study that both NIH and stent underexpansion emerged as the main causes of ISR even in the second-generation DES era.²⁵ However, there were a number of limitations to that study, such as very few cases of L-ISR after second-generation DES implantation and use of gray-scale IVUS. Considering those findings, the mechanism of ISR at the different phases after implanting second-generation DES with partial improvement of the problem, which first-generation DES had, has not yet been fully evaluated.

In-stent neoatherosclerosis was suggested as one of the causes of late adverse events after both first-generation DES and BMS implantation.⁹ Recently, Takano et al reported that the presence of lipid-laden neointima and intimal disruption in the late phase (>5 years) increased compared with that in the early phase (<6 months) after BMS implantation.¹⁴ Furthermore, Habara et al reported that TCFA-like intima and intimal disruption increased from the early (<1 year) to late (1–3 years) and very late phases (>3 years) after first-generation DES implantation.⁷ In the present study, the incidence of neoatherosclerosis was more frequent in the late phase (>1 year) than in the early phase (<1 year) and increased over time. Interestingly, Otsuka et al reported that the observed frequency of neoatherosclerosis did not differ significantly between first-generation DES and CoCr-EES in a human autopsy study.¹² These findings suggest that neoatherosclerosis is the main cause of L-ISR after second-generation DES implantation and that it increases over time.

A pathological study demonstrated that NIH of BMS-ISR was mainly composed of vascular smooth muscle cells, whereas that of DES-ISR was composed of rich proteoglycan extracellular matrix and fibrin.^22,27–29 An OCT study demonstrated that homogeneous neointima was more frequent in the early phase (<1 year) of BMS-ISR, whereas heterogeneous neointima was more frequent in the very late phase (>4 years).⁵ In contrast, at the MLA site, heterogeneous intima was mainly observed in the early, late and very late phases of DES-ISR and the incidence of heterogeneous intima was not significantly changed among the 3 phases.⁷ In the present study, however, homogeneous neointima was more frequently observed in the early phase than in the late phase. Although it remains unclear why there is this discrepancy, it might be related to the difference in vascular response between first- and second-generation DES. Second-generation DES decrease abnormal vascular responses such as fibrin deposition and inflammatory cell infiltration compared with first-generation DES.¹² Nevertheless, second-generation DES have not been immune to ISR in the real world. As one of the mechanisms, second-generation DES occasionally cannot inhibit the proliferation of smooth muscle cells in the early phase, resulting in a high rate of homogeneous neointima; in the late phase, neoatherosclerosis gradually occurs over time, even with second-generation DES as well as BMS and first-generation DES, although the timing and frequency may differ among them.

The present study demonstrated that homogeneous neointima was predominately observed within the entire stent in both the E-ISR and L-ISR groups, although the dominant neointima was different at the MLA site in the 2 groups. Recently, Habara et al reported their OCT findings of first-generation DES in the 3 phases (early, <1 year; late, 1–3 years; and very late, >3 years): homogeneous neointima was predominately observed within the entire stent in all phases (early, 72.8%; late, 68.1%; and very late, 58.2%), whereas heterogeneous neointima, including TCFA-like neointima, was more frequently observed at the MLA site (early, 65.1%; late, 63.6%; and very late, 66.7%).⁷ Similarly, the present study demonstrated that homogeneous neointima was predominately observed within the entire stent in both the E-ISR and L-ISR groups, whereas neoatherosclerosis was the dominant neointima at the MLA site in the L-ISR group. These findings indicated that the vascular response to the stent may be different between the MLA and non-MLA sites and that the underlying mechanism of ISR may change over time.

Pathologic studies have demonstrated that organized thrombus could be a component of restenosis in both BMS and first-generation DES.^30–32 Recently, Kang et al reported that the incidence of thrombosis was 58% in patients with DES-ISR.⁶ Furthermore, Habara et al reported that thrombosis at the MLA site was frequently observed in 28.6% of cases of first-generation DES-ISR in the very late phase (>3 years).⁷ In the current study, it was intriguing that the incidence of thrombosis at the MLA site was very low. Although that might be related to the lower incidence of unstable angina compared with previous studies, it suggested that thrombosis was unlikely to be associated with ISR after second-generation DES implantation.

PLIA was considered to be a reflection of delayed arterial healing after DES implantation and represented the presence of fibrinogen or extracellular matrix in an in vitro study.^22,23 Recently, Otake et al reported an association between PLIA and neointimal thickening after first-generation DES implantation.²³ With first-generation DES, the incidence of PLIA was reported to decrease during long-term follow-up.^7,33 Interestingly, the current study showed that PLIA was less frequently observed in the L-ISR group than in the E-ISR group. These findings suggested that the mechanism of E-ISR after second-generation DES implantation was partially associated with delayed vascular healing, which gradually improved over time.

Clinical Implications

Recent pivotal clinical trials and network meta-analyses have demonstrated the improved safety and efficacy of second-generation DES compared with first-generation DES.^10,11 Nevertheless, late adverse events such as late TLR beyond 1 year and VLST have emerged as unsolved issues even with second-generation DES. Therefore, it is very important for clinicians to elucidate the mechanism of ISR after second-generation DES. In the present study, E-ISR was mainly caused by NIH, but neoatherosclerosis was the main mechanism of L-ISR. As shown in previous studies, neoatherosclerosis is of interest as an important mechanism of DES failure.^6,9 Our results support that we should tackle it to obtain better clinical outcomes in the second-generation DES era. To date, however, it remains unclear how we can prevent the occurrence of neoatherosclerosis. Further studies are required to investigate new approaches to preventing in-stent neoatherosclerosis.

Study Limitations

There are several limitations to the present study. First, this was a retrospective single-center study that included a relatively small study population. Therefore, selection bias may exist and have biased the conclusions. Second, there was little data regarding the correlation of OCT images of DES-ISR tissue and histological findings. Therefore, OCT findings of intimal tissue should be interpreted with caution. Third, there was a significant difference in the stent type between the 2 groups. Therefore, it might have some influence on our results. Fourth, this study was not designed as a serial analysis. Therefore, it remains unclear whether the differences in OCT and OFDI images at different time points are related to differences in terms of the ISR mechanism or the time course of neointimal tissue formation. Finally, this study included both OCT and OFDI analysis. Therefore, the difference between OCT and OFDI findings of intimal tissue might have influenced the results of the present study.

Conclusions

Neointimal tissue characteristics differed between E-ISR and L-ISR groups of patients after second-generation DES implantation. E-ISR was mainly caused by NIH, but in-stent neoatherosclerosis was the main mechanism of L-ISR.

Acknowledgments

The authors thank Naoka Katsumi, Yukie Ochi and Miho Hasegawa for their assistance with this work.

Disclosures

No conflict of interest.

References

• 1.   Moses JW, Leon MB, Popma JJ, Fitzgerald PJ, Holmes DR, O’Shaughnessy C, et al. Sirolimus-eluting stents versus standard stents in patients with stenosis in a native coronary artery. N Engl J Med 2003; 349: 1315–1323.
• 2.   Stone GW, Ellis SG, Cox DA, Hermiller J, O’Shaughnessy C, Mann JT, et al. A polymer-based, paclitaxel-eluting stent in patients with coronary artery disease. N Engl J Med 2004; 350: 221–231.
• 3.   Kimura T, Morimoto T, Nakagawa Y, Kawai K, Miyazaki S, Muramatsu T, et al. Very late stent thrombosis and late target lesion revascularization after sirolimus-eluting stent implantation: Five-year outcome of the j-Cypher Registry. Circulation 2012; 125: 584–591.
• 4.   Gonzalo N, Serruys PW, Okamura T, van Beusekom HM, Garcia-Garcia HM, van Soest G, et al. Optical coherence tomography patterns of stent restenosis. Am Heart J 2009; 158: 284–293.
• 5.   Habara M, Terashima M, Nasu K, Kaneda H, Inoue K, Ito T, et al. Difference of tissue characteristics between early and very late restenosis lesions after bare-metal stent implantation: An optical coherence tomography study. Circ Cardiovasc Interv 2011; 4: 232–238.
• 6.   Kang SJ, Mintz GS, Akasaka T, Park DW, Lee JY, Kim WJ, et al. Optical coherence tomographic analysis of in-stent neoatherosclerosis after drug-eluting stent implantation. Circulation 2011; 123: 2954–2963.
• 7.   Habara M, Terashima M, Nasu K, Kaneda H, Yokota D, Ito T, et al. Morphological differences of tissue characteristics between early, late, and very late restenosis lesions after first generation drug-eluting stent implantation: An optical coherence tomography study. Eur Heart J Cardiovasc Imaging 2013; 14: 276–284.
• 8.   Kimura T, Abe K, Shizuta S, Odashiro K, Yoshida Y, Sakai K, et al. Long-term clinical and angiographic follow-up after coronary stent placement in native coronary arteries. Circulation 2002; 105: 2986–2991.
• 9.   Nakazawa G, Otsuka F, Nakano M, Vorpahl M, Yazdani SK, Ladich E, et al. The pathology of neoatherosclerosis in human coronary implants bare-metal and drug-eluting stents. J Am Coll Cardiol 2011; 57: 1314–1322.
• 10.   Kedhi E, Joesoef KS, McFadden E, Wassing J, van Mieghem C, Goedhart D, et al. Second-generation everolimus-eluting and paclitaxel-eluting stents in real-life practice (COMPARE): A randomised trial. Lancet 2010; 375: 201–209.
• 11.   Stone GW, Rizvi A, Newman W, Mastali K, Wang JC, Caputo R, et al. Everolimus-eluting versus paclitaxel-eluting stents in coronary artery disease. N Engl J Med 2010; 362: 1663–1674.
• 12.   Otsuka F, Vorpahl M, Nakano M, Foerst J, Newell JB, Sakakura K, et al. Pathology of second-generation everolimus-eluting stents versus first-generation sirolimus- and paclitaxel-eluting stents in humans. Circulation 2014; 129: 211–223.
• 13.   Mehran R, Dangas G, Abizaid AS, Mintz GS, Lansky AJ, Satler LF, et al. Angiographic patterns of in-stent restenosis: Classification and implications for long-term outcome. Circulation 1999; 100: 1872–1878.
• 14.   Takano M, Yamamoto M, Inami S, Murakami D, Ohba T, Seino Y, et al. Appearance of lipid-laden intima and neovascularization after implantation of bare-metal stents extended late-phase observation by intracoronary optical coherence tomography. J Am Coll Cardiol 2009; 55: 26–32.
• 15.   Kuramitsu S, Sonoda S, Yokoi H, Iwabuchi M, Nishizaki Y, Shinozaki T, et al. Long-term coronary arterial response to biodegradable polymer biolimus-eluting stents in comparison with durable polymer sirolimus-eluting stents and bare-metal stents: Five-year follow-up optical coherence tomography study. Atherosclerosis 2014; 237: 23–29.
• 16.   Yabushita H, Bouma BE, Houser SL, Aretz HT, Jang IK, Schlendorf KH, et al. Characterization of human atherosclerosis by optical coherence tomography. Circulation 2002; 106: 1640–1645.
• 17.   Tanaka A, Imanishi T, Kitabata H, Kubo T, Takarada S, Tanimoto T, et al. Morphology of exertion-triggered plaque rupture in patients with acute coronary syndrome: An optical coherence tomography study. Circulation 2008; 118: 2368–2373.
• 18.   Lee SY, Shin DH, Mintz GS, Kim JS, Kim BK, Ko YG, et al. Optical coherence tomography-based evaluation of in-stent neoatherosclerosis in lesions with more than 50% neointimal cross-sectional area stenosis. EuroIntervention 2013; 9: 945–951.
• 19.   MacNeill BD, Jang IK, Bouma BE, Iftimia N, Takano M, Yabushita H, et al. Focal and multi-focal plaque macrophage distributions in patients with acute and stable presentations of coronary artery disease. J Am Coll Cardiol 2004; 44: 972–979.
• 20.   Tearney GJ, Yabushita H, Houser SL, Aretz HT, Jang IK, Schlendorf KH, et al. Quantification of macrophage content in atherosclerotic plaques by optical coherence tomography. Circulation 2003; 107: 113–119.
• 21.   Rathore S, Matsuo H, Terashima M, Kinoshita Y, Kimura M, Tsuchikane E, et al. Rotational atherectomy for fibro-calcific coronary artery disease in drug eluting stent era: Procedural outcomes and angiographic follow-up results. Catheter Cardiovasc Interv 2010; 75: 919–927.
• 22.   Teramoto T, Ikeno F, Otake H, Lyons JK, van Beusekom HM, Fearon WF, et al. Intriguing peri-strut low-intensity area detected by optical coherence tomography after coronary stent deployment. Circ J 2010; 74: 1257–1259.
• 23.   Otake H, Shite J, Ikeno F, Shinke T, Teramoto T, Miyoshi N, et al. Evaluation of the peri-strut low intensity area following sirolimus- and paclitaxel-eluting stents implantation: Insights from an optical coherence tomography study in humans. Int J Cardiol 2012; 157: 38–42.
• 24.   Kim Y, Choi YK, Emery S. Logistic regression with multiple random effects: A simulation study of estimation methods and statistical packages. Am Stat 2013; 67: 171–182.
• 25.   Goto K, Zhao Z, Matsumura M, Dohi T, Kobayashi N, Kirtane AJ, et al. Mechanisms and patterns of intravascular ultrasound in-stent restenosis among bare metal stents and first- and second-generation drug-eluting stents. Am J Cardiol 2015; 116: 1351–1357.
• 26.   Kang SJ, Mintz GS, Park DW, Lee SW, Kim YH, Whan Lee C, et al. Mechanisms of in-stent restenosis after drug-eluting stent implantation: Intravascular ultrasound analysis. Circ Cardiovasc Interv 2011; 4: 9–14.
• 27.   Ramcharitar S, Garcia-Garcia HM, Nakazawa G, Kukreja N, Ligthart J, Virmani R, et al. Ultrasonic and pathological evidence of a neo-intimal plaque rupture in patients with bare metal stents. EuroIntervention 2007; 3: 290–291.
• 28.   Kume T, Okura H, Kawamoto T, Akasaka T, Toyota E, Watanabe N, et al. Images in cardiovascular medicine: Fibrin clot visualized by optical coherence tomography. Circulation 2008; 118: 426–427.
• 29.   Grewe PH, Deneke T, Machraoui A, Barmeyer J, Muller KM. Acute and chronic tissue response to coronary stent implantation: Pathologic findings in human specimen. J Am Coll Cardiol 2000; 35: 157–163.
• 30.   Glover C, Ma X, Chen YX, Miller H, Veinot J, Labinaz M, et al. Human in-stent restenosis tissue obtained by means of coronary atherectomy consists of an abundant proteoglycan matrix with a paucity of cell proliferation. Am Heart J 2002; 144: 702–709.
• 31.   Komatsu R, Ueda M, Naruko T, Kojima A, Becker AE. Neointimal tissue response at sites of coronary stenting in humans: Macroscopic, histological, and immunohistochemical analyses. Circulation 1998; 98: 224–233.
• 32.   van Beusekom HM, Saia F, Zindler JD, Lemos PA, Swager-Ten Hoor SL, van Leeuwen MA, et al. Drug-eluting stents show delayed healing: Paclitaxel more pronounced than sirolimus. Eur Heart J 2007; 28: 974–979.
• 33.   Kozuki A, Shinke T, Otake H, Shite J, Nakagawa M, Nagoshi R, et al. Temporal course of vessel healing and neoatherosclerosis after DES implantation. JACC Cardiovasc Imaging 2013; 6: 1121–1123.