Atherogenesis in Native Coronary Segments and In-Stent Neoatherogenesis Beyond Three Years After First-Generation Drug-Eluting Stent Implantation: Angiographic and Optical Coherence Tomography Study

Objectives. The mechanisms underlying the development of neoatherosclerosis following stent implantation remain to be further elucidated. The aim of this study was to investigate the association between subclinical in-stent neoatherosclerosis (NA) and atherosclerosis progression of native coronary segments in patients with chronic coronary syndrome 3 and 9 years after first-generation drug-eluting stent implantation. Methods. This is a prespecified analysis of the prospective cohort study evaluating long-term neointimal healing in consecutive patients undergoing elective percutaneous coronary intervention with sirolimus-eluting stent (SES) or paclitaxel-eluting stent (PES) implantation. Quantitative coronary angiography (QCA) was evaluated in non-stented coronary segments. Results. Forty-three patients underwent optical coherence tomography (OCT) and QCA at 3 years and 21 patients at 3 years and 9 years after SES or PES implantation. NA was identified in 44.2% at 3 years and in 66.7% at 9 years after the index procedure. NA at 3 years was more frequently observed in patients with atherosclerosis progression in native coronary segments than without (66.7% vs 15.8%; P<.01). Higher low-density cholesterol level (93 mg/dL vs 77 mg/dL; P=.04), greater maximal neointimal thickness (0.74 mm vs 0.37 mm; P<.001), and presence of peristrut low-intensity areas (57.9% vs 20.8%; P=.01) were more frequent in patients with NA. NA progression (P=.01) along with greater neointimal growth (P<.01) were detected in serial analysis between 3-year and 9-year OCT assessments. Conclusions. OCT-confirmed NA formation after first-generation drug-eluting stent implantation was associated with QCA-defined atherosclerosis progression in non-stented segments between 0 and 3 years. NA and neointimal proliferation continued between 3 and 9 years.

J INVASIVE CARDIOL 2021;33(9):E738-E747. Epub 2021 August 25.

Key words: drug-eluting stent, neoatherosclerosis, optical coherence tomography

The mechanisms underlying neoatherosclerosis (NA) formation have yet to be elucidated.¹ While atherogenesis develops in native coronary arteries over several years, NA occurs within months to years following percutaneous coronary intervention (PCI) and can result in very-late stent thrombosis or in-stent restenosis.^2,3 This accelerated process has been shown to occur earlier and more frequently after drug-eluting stent (DES) implantation than after bare-metal stent (BMS) implantation.^4-6 Delayed endothelialization, along with poorly formed cell junctions and rapid infiltration of macrophages, has been proposed as one of the causes of in-stent NA.^7,8 The pathogenic mechanisms of NA might be similar to those involved in atherosclerosis of native coronary segments. In the era of second-generation DES devices, which have reduced the prevalence of late and very-late stent thrombosis, the frequency of NA did not differ significantly from first-generation devices.^3,9 Therefore, long-term serial follow-up data after first-generation DES implantation will be valuable in the future. We sought to investigate the association between subclinical NA and atherosclerosis progression of native coronary segments in patients with chronic coronary syndrome 3 and 9 years after first-generation DES implantation.

Study population. This is a prespecified analysis of a prospective registry including 43 patients with chronic coronary syndrome who underwent PCI with implantation of sirolimus-eluting stent (SES) (Cypher; Cordis) or paclitaxel-eluting stent (PES) (Taxus; Boston Scientific) between January 2003 and November 2005 and exhibiting no stent failure.^10,11

The inclusion criteria included single DES implantation in clinically and angiographically relevant stenosis of a native vessel, implantation of a stent between 2.5-3.5 mm in diameter, at least 36 months of uneventful follow-up after the index PCI, informed consent to participate in the intravascular imaging follow-up, and adherence to dual-antiplatelet therapy continuation for the 12 months following the index procedure according to the recommendations applicable at the time when the study was conducted.¹⁰

The exclusion criteria included a history of target-vessel revascularization (TVR), myocardial infarction (MI), and stroke in the period between the index PCI and planned quantitative coronary angiography (QCA) and optical coherence tomography (OCT) examination, left main as a culprit vessel, lesions located <10 mm from the native vessel ostium (due to lack of possibility to perform OCT measurement with proximal balloon occlusion), PCI of a chronic total occlusion of a native artery, and chronic kidney disease with baseline estimated glomerular filtration rate <30 mL/min/1.73 m².¹⁰

Patients underwent clinical assessment, angiography, and OCT examination of the coronary segments with previously implanted SES or PES at 3 years and 9 years. OCT images were acquired after diagnostic coronary angiography. QCA and OCT examinations were assessed by an independent observer. Informed consent was obtained from each patient. The study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in a priori approval by our institution’s human research committee.

QCA and OCT analysis. Coronary segments were defined according to the modified American Heart Association/American College of Cardiology (AHA/ACC) coronary segment classification.¹²  QCA was performed in non-stented coronary segments ≥2.5 mm in diameter. Standard biplane angiographic views were recorded in at least 2 orthogonal views to match baseline, 3-year, and 9-year images. Coronary atherosclerosis progression was defined as the presence of at least 1 of the following criteria: (1) increase of ≥10% diameter stenosis (DS) of at least 1 pre-existing lesion with ≥50% DS,(2) increase of ≥30% DS of a pre-existing lesion with <50% DS; and (3) progression of a lesion to total occlusion from baseline to angiographic follow-up.¹³ Mean difference of variables was derived from each non-stented segment (follow-up – baseline). In-stent restenosis (ISR) was defined as angiographically documented DS >50% within the stented segment.¹⁴

OCT images of the stented segments at 3-year follow-up were acquired using a time-domain OCT imaging system (M2 System; LightLab Imaging) in accordance with manufacturer recommendations. At 9 years, OCT images of stented segments were obtained with a commercially available frequency domain OCT imaging system (C7XR System with Dragonfly image catheters; LightLab Imaging) using the non-occlusive flushing technique. For quantitative measurements, cross-sectional OCT images were analyzed at 0.1 mm intervals. Lumen area, diameter, neointimal thickness, stent area, and DS were determined using LightLab offline analytical software based on expert consensus documents.¹⁵NA was defined as new lipid and/or calcific plaques, and/or macrophages, and/or neovascularization seen within the neointima.¹⁶ The lipid was recognized as a signal-poor region with poorly delineated borders covered by a fibrous cap.¹⁷Calcium was defined as a signal-poor or heterogeneous region with a sharply delineated border.¹⁷ Macrophages were identified as signal-rich punctate regions with high attenuation.¹⁸Neovascularization was defined as the presence of vessels within the intima and was determined as sharply delineated signal-poor voids followed in multiple contiguous frames.¹⁷ The thin-capped fibroatheroma (TCFA) was defined as a necrotic core with an overlying fibrous cap thickness <65 µm.¹⁷Peristrut low-intensity area (PLIA) was defined as a region surrounding stent struts characterized with a homogenous low-intensity appearance by OCT without significant signal attenuation.¹⁹Stent malapposition was defined as >150 µm distance from the vessel wall in a segment not overlying a sidebranch. Stent struts are deemed uncovered struts if no evidence of tissue could be visualized above the struts.¹⁷

Statistical analysis. Continuous variables are presented as median with interquartile range (IQR) when non-normally distributed and as mean value ± standard deviation when normally distributed. The type of variable distribution was assessed using the Kolmogorov-Smirnov test. Categorical data were presented as absolute values and percentages. Differences between non-normally and normally distributed continuous variables were calculated with the Mann-Whitney U-test and Student’s t-test, respectively. Differences between categorical variables were calculated with Fisher’s exact test. Continuous related variables were compared using the paired Student’s t-test or Wilcoxon signed-rank tests and categorical related variables were compared by McNemar’s test. Univariable logistic regression analyses were performed to determine the predictors of NA. All statistical tests were two-sided and a P-value <.05 was considered statistically significant. Statistical analysis was performed using SPSS Statistics, version 26 (IBM).

Of 156 consecutive patients with chronic coronary syndrome who underwent PCI with implantation of a first-generation DES, 43 patients (16 SES and 27 PES) met the inclusion criteria and were assessed with OCT at 3 years and 21 patients (7 SES and 14 PES) were assessed at 9 years post index procedure. Baseline clinical characteristics of patients with and without atherosclerosis progression in native coronary segments and with and without NA in stented segments after 3 years are summarized in Table 1. No significant differences were recorded, with the exception of higher low-density lipoprotein cholesterol (LDL-C) level in patients with NA formation after 3 years (93 mg/dL in the NA group vs 77 mg/dL in the no-NA group; P=.04).

A total of 487 untreated native coronary artery segments at baseline were matched with the corresponding segments at 3-year follow-up and 227 segments at 9-year follow-up, allowing the assessment of longitudinal changes over time (Table 2 and Supplemental Table S1; Part 1, Part 2). Atherosclerosis progression in at least 1 segment was more frequent in the NA group (84% vs 33% in the no-NA group; P<.01) (Table 2). A reduction in minimal lumen diameter (MLD) and DS was observed in all patients irrespective of NA formation or meeting the criteria for significant progression of atherosclerosis after 3 years (Table 2). Mean difference of MLD between baseline and 3 years was greater in the NA group (-0.43 mm vs -0.15 mm in the no-NA group; P<.001) (Table 2). Mean difference in reference vessel diameter (RVD) was also higher in the NA group (-0.22 mm vs -0.07 mm in the no-NA group; P<.01) (Table 2). Mean differences in RVD, MLD, DS between baseline and 9 years, and atherosclerosis progression did not differ between the 2 groups (Supplemental Table S1; Part 1, Part 2).

A total of 4137 OCT frames after 3 years and 4407 frames after 9 years were analyzed. NA formation was observed in 44.2% of patients after 3 years and in 66.7% after 9 years (P=.02) with an increase in frequency of lipid accumulation (28.6% at 3 years vs 66.7% at 9 years; P<.01) (Table 3). There was a significant decrease in minimum lumen area (MLA) (4.27 mm² at 3 years vs 3.86 mm² at 9 years; P<.01), increase in maximal neointimal thickness (0.51 mm at 3 years vs 0.67 mm at 9 years; P<.01), and increase in area stenosis (26.8% at 3 years vs 38.3% at 9 years; P<.01) between 3 and 9 years (Table 3). NA formation at 3 years was detected in 66.7% of patients with significant atherosclerosis progression in native coronary segments and in 15.8% without atherosclerosis progression (P<.01) (Table 4). In patients with NA formation after 3 years, maximal neointimal thickness was greater than in patients with no NA (0.74 mm in the NA group vs 0.37 mm in the no-NA group; P<.001) (Table 4). The majority of plaques were composed of lipid (39.6%), and less frequently macrophages (16.3%) and calcium (14.0%). The other findings of NA were less frequent, with neovascularization in 7.0% and TCFA in 4.7% (Table 4 and Figure 1). The presence of PLIA was more frequent in patients with NA (57.9% vs 20.8% in the no-NA group; P=.01) (Table 4). The prevalence of uncovered stent struts was high, with 76.7% after 3 years and 61.9% after 9 years. Per-frame analysis confirmed a higher number of frames with signs of NA in patients with atherosclerosis progression than without after 3 years (22.2% vs 8.7%, respectively; P<.01) (Supplemental Table S2). At 9 years, there was no difference in NA formation in patients with vs without atherosclerosis progression between 0-9 years (78.6% vs 42.9%, respectively; P=.10) (Table 5). PLIA at 9 years was detected only in patients with co-occurrence of NA (42.9% vs 0%; P=.04) (Table 5). OCT per-frame analysis at 9 years is presented in Supplemental Table S3. Atherosclerosis progression in native coronary segments was the only predictor of NA plaques in univariate logistic regression analysis (odds ratio, 10.7; 95% confidence interval, 2.4-47.7) at 3 years, but not at 9 years (Supplemental Table S4).

The main findings of the present study can be summarized as follows: (1) the progression of native coronary atherosclerosis was associated with in-stent NA; (2) higher LDL-C level and presence of PLIA were more frequent in patients with NA; (3) NA was observed in 19 of 43 patients (44.2%) at 3 years and in 14 of 21 patients (66.7%) at 9 years after first-generation DES implantation; and (4) NA (P=.01) and neointimal growth (P<.01) progressed continuously between 3 years and 9 years after first-generation DES implantation.

NA has been described as one of the causes of very-late stent thrombosis and restenosis,^2,6,20 yet the specific mechanism of NA formation remains unclear. Our study indicates that factors associated with atherosclerosis progression in native coronary arteries might correlate with the process of in-stent NA. In this longitudinal invasive imaging study, we have demonstrated that the more pronounced reduction in MLD and RVD was detected between baseline and 3 years in the NA group compared with the no-NA group. Significant atherosclerosis progression in at least 1 segment was more frequent in the NA group.

Possible mechanisms responsible for NA are believed to be multifactorial.¹ Studies aimed to determine the predictors of NA showed that apart from stent type and stent age, patient characteristics, including current smoking, chronic kidney disease, and angiotensin-converting enzyme inhibitors/angiotensin II receptor blockade, were associated with the presence of NA.²¹ In the present study, atherosclerosis progression in native coronary segments was the only predictor of NA plaques. Lee et al has identified that chronic kidney disease, LDL-C >70 mg/dL, and stent age were all independent predictors for NA, whereas the type of DES (first-generation vs second-generation) was not.³ The accelerated process of NA with DES use in comparison with BMS suggests that the risk factors related to stent design might predispose the patient to NA. One of the suggested mechanisms is a delayed endothelialization induced by the drug eluted from the DES. Poor cell-to-cell junctions, reduced expression of antithrombotic molecules, platelet aggregation, and decreased nitric oxide production favor a greater permeability of LDL and macrophages into the neointima.^7,8,22

The other proposed mechanism underlying NA is chronic inflammation promoted by the DES polymer. Accumulation of eosinophils, lymphocytes, and giant cells around the DES struts may contribute to NA.^8,23,24 High LDL-C is known to be associated with atherosclerosis progression and cardiovascular adverse events.²⁵ In our study, patients with NA had higher levels of LDL-C vs those without NA lesions after 3 years (93 g/dL vs 77 g/dL; P=.04). Although all patients were on statin therapy, the currently recognized guideline-recommended goals of LDL-C reduction were not achieved in this cohort. Incomplete and delayed re-endothelialization in our study, in combination with prolonged exposure to even a modest level of LDL-C, may potentially lead to NA. Kuroda et al have reported that apart from LDL-C, the C-reactive protein level was also independently associated with NA.²⁰ The identification of inflammatory response after stent deployment is still limited. Among intravascular imaging methods widely used in clinical practice, only OCT generates images with resolution comparable to histological studies. PLIA, defined as a low-intensity area surrounding struts, corresponds with peristrut giant cell accumulation (23%), peristrut leukocyte accumulation (13%), peristrut neovascularization (36%), or peristrut calcification (18%) in histopathological studies.¹⁸ High-intensity, high-attenuation pattern revealed a predominance of foam-cell accumulation (68%), superficial elastic fibers without foamy macrophages (12%), and neointimal calcification (11%).¹⁸ Despite a wide variety of histological differential diagnoses, the surrogate imaging parameter of inflammation had clinical significance in long-term follow-up. Superficial location of macrophages with co-presence of MLA <4 mm² and fibrous cap thickness <75 µm are validated features of plaque vulnerability.²⁷ Observational OCT studies of patients receiving was associated with an increased rate of target-lesion revascularization.²⁶ In the present study, the presence of PLIA was more frequent in patients with NA 3 years after stent placement (57.9% vs 20.8% in the no-NA group; P=.01). At 9-year follow-up, PLIA was detected only in patients with NA (42.9% vs 0% in the no-NA group; P=.04) (Table 4 and Table 5).

The frequency of in-stent NA is higher in the present study than in previously published in vivo studies using OCT. Usui et al reported 28.1% NA at 5-10 years after first-generation DES implantation and 59.1% NA at >10 years.²⁸ In 5-year observational studies after first-generation DES implantation, the prevalence of NA differs from 21% to 40.9%.^13,29 Serial 3-year and 9-year assessments showed NA of 23.1% and 36.4%, respectively.¹¹ The differences may be explained by various NA definitions. In the present study, apart from detection of fibroatheroma or fibrocalcific plaque, we also included macrophages and neovascularization as findings related to NA.³⁰ In an autopsy study, the frequency of NA (defined as the presence of either peristrut foamy macrophage clusters, fibroatheromas, TCFA, or plaque ruptures with thrombosis) was 31% in the DES group and 16% in the BMS group >30 days after implantation.⁴ Otsuka et al reported similar NA prevalence in first-generation and second-generation DES options (29% in EES, 35% in SES, and 19% in PES) 30 days to 3 years after stent implantation.⁹ In the present study, the risk of NA progression increased and neointima continuously grew between 3 and 9 years (Table 3). These data may explain the risk of very-late stent failure even after a long period of time following PCI. Usui et al also have confirmed continuous development of NA with higher frequency of NA >10 years in comparison with 5-10 years after SES implantation.²⁸

These results reinforce the need for secondary prevention, including intensive lipid-lowering and anti-inflammatory therapy, in high-risk patients to decrease the incidence of NA along with atherosclerosis progression in native coronary segments. In addition to secondary prevention, improvement in stent design is still necessary. Thinner struts, biodegradable polymers, and more precise molecular targeting of the mTOR complex may improve re-endothelialization and reduce peristrut inflammation and fibrin deposition.^31-35

Study limitations. This was a small, single-center, non-randomized study. Selection bias cannot be excluded. The study did not assess the potential relationship of NA with clinical events because of the limited number of patients. The prevalence of NA has been investigated in first-generation DESs that are no longer used in clinical practice. However, the frequency of NA did not differ significantly from second-generation stents^3,9 and few long-term follow-up data are available. The ability to identify macrophages in OCT is limited. Bright spots were correlated with a variety of plaque components that cause sharp changes in the index of refraction (macrophages, cellular fibrous tissue, interfaces between calcium and fibrous tissue, calcium and lipids, fibrous cap, and lipid pool).³⁶ Although OCT is the most accurate imaging method to detect in vivo neointimal healing pattern, we are aware of misdiagnosis and over-reporting of NA.

OCT-confirmed NA formation after first-generation DES implantation was associated with QCA–defined atherosclerosis progression in non-stented segments between 0 and 3 years, with continuous expansion of NA and neointimal growth between 3 and 9 years. Higher LDL-C levels and PLIA appearance were more prevalent in patients with NA. Further studies are required to more accurately characterize in vivo NA and its impact on clinical outcomes.