Author + information
- Received January 16, 2013
- Revision received March 18, 2013
- Accepted March 22, 2013
- Published online August 1, 2013.
- Daniel Chamié, MD,
- Hiram G. Bezerra, MD, PhD,
- Guilherme F. Attizzani, MD,
- Hirosada Yamamoto, MD,
- Tomoaki Kanaya, MD,
- Gregory T. Stefano, MD,
- Yusuke Fujino, MD,
- Emile Mehanna, MD,
- Wei Wang, PhD,
- Ahmad Abdul-Aziz, BS,
- Matthew Dias, BS,
- Daniel I. Simon, MD and
- Marco A. Costa, MD, PhD∗ ()
- Division of Cardiology, Harrington Heart & Vascular Institute, University Hospitals Case Medical Center, Case Western Reserve School of Medicine, Cleveland, Ohio
- ↵∗Reprint requests and correspondence:
Dr. Marco A. Costa, Division of Cardiology, Harrington Heart & Vascular Institute, University Hospitals Case Medical Center, Case Western Reserve School of Medicine, 11,100 Euclid Avenue, Lakeside 3001, Cleveland, Ohio 44106.
Objectives This study sought to investigate the frequency, predictors, and detailed qualitative and quantitative assessment of optical coherence tomography (OCT)-detected stent edge dissections. Its impact on subsequent management and clinical outcomes were also investigated.
Background OCT is a high-resolution imaging modality that can lead to more frequent recognition and accurate assessment of vascular injuries during percutaneous coronary intervention (PCI).
Methods From September 2010 to June 2011, all patients with OCT post-PCI were enrolled. Edge dissections were defined as disruptions of the arterial lumen surface in both the 5-mm distal and proximal stent edges. Qualitative and quantitative analyses of all edges were performed at 0.2-mm intervals.
Results In total, 395 edges (249 lesions in 230 patients) were analyzed. The overall incidence of OCT-detected edge dissection was 37.8%, and most (84%) were not apparent on angiography. Independent predictors for OCT-detected dissections were presence of atherosclerotic plaque at stent edges, calcification angle, minimum fibrous cap thickness, thin-cap fibroatheromas, stent/lumen eccentricity, and vessel overstretching. Mean dissection length measured 2.04 ± 1.60 mm, 96.2% appeared as flaps, and 52.8% extended beyond the intima/atheroma layer. Additional stenting was performed in 22.6% of all dissections, which were longer, had bigger dimensions, and promoted deeper vascular injury. The 12-month major adverse cardiac event rate was similar between patients with (7.95%) and without (5.69%, p = 0.581) dissections.
Conclusions High rates of stent edge dissections were detected by OCT, usually related to the presence of atherosclerosis at stent edges and to PCI technique. Detailed OCT assessment of dissection severity was possible and affected the subsequent management of this complication. Non–flow-limiting, small, and superficial dissections left untreated proved benign.
Plaque fracture and dissection of the arterial wall are key mechanisms of lumen enlargement after balloon angioplasty (1,2). During stent implantation, unplanned vessel tearing may occur at the transition between the rigid stent struts and the adjacent arterial wall (3), which has been associated with increased risks of thrombosis and major adverse cardiac events (MACE) (4–8).
Coronary angiography (9) and intravascular ultrasound (IVUS) (3,10) have traditionally been used for the diagnosis of dissections at stent edges. Reported incidences of stent edge dissection range from 5% to 23% of the percutaneous coronary intervention (PCI) procedures as detected by IVUS (3). The superior resolution (12 to 18 μm) (11) of light-based optical coherence tomography (OCT) compared with IVUS (150 to 250 μm) may lead to more frequent recognition of vascular injury during PCI (12,13). Although physicians welcome better imaging and more knowledge of the vascular response to PCI, they face a clinical dilemma on the subsequent management of OCT findings. Recent reports provided early insights on the incidence and characteristics of stent edge dissections by OCT (14,15), but the inherent selection bias of retrospective, small studies have limited our understanding of the true incidence and clinical implications of OCT-detected stent edge dissections. Therefore, the goals of the present study were to: 1) determine the frequency of stent edge dissections by OCT in a large, unrestricted population representative of the daily clinical practice; 2) perform a detailed morphometric assessment of stent edge dissections and identify potential factors associated with its occurrence; 3) assess the impact of the OCT findings on the clinical management of edge dissections in real practice; and 4) examine the clinical outcomes of OCT-detected edge dissections.
Study population, study design, and PCI procedures
OCT was adopted as the standard intravascular imaging modality at our Institution after its approval by the U.S. Food and Drug Administration in May 2010. The University Hospitals Case Medical Center OCT Program is a prospective initiative approved by the local institutional review board, and comprises 2 distinct phases: Phase I (September to October 2010) was a quality initiative program designed to evaluate safety, feasibility, and clinical impact of OCT usage on patient management in the daily practice. During this phase, OCT was performed in all patients undergoing a coronary intervention or requiring an invasive evaluation of intermediate coronary lesions (16). There were no pre-defined exclusion criteria, thus representing an “all-comers,” unrestricted use of OCT; Phase II, from November 2010 onward, comprised the “real-world” clinical utilization of OCT per the operator's discretion, thus representing a “selective use” of OCT in routine practice. All patients submitted to OCT examination from September 2010 to June 2011 were prospectively entered into a dedicated database, and comprise the population included in the current analysis.
PCI was performed per standard practice. Patients were treated with a loading dose of aspirin 300 mg and clopidogrel 300 mg or 600 mg. At the beginning of the procedure, a weight-adjusted (100 IU/kg) bolus of unfractionated heparin or a bolus of bivalirudin (0.75 mg/kg) was administered. Additional heparin boluses or continuous intravenous infusion of bivalirudin 1.75 mg/kg/h were given in order to maintain an activated clotting time ≥250 s. Glycoprotein IIb/IIIa inhibitors were used as per the operator's discretion. OCT imaging was mandatory in Phase I and left at the operator's discretion during Phase II. No formal recommendations were given regarding the management of stent edge dissection identified by OCT, which allowed proper assessment of the operator reaction to OCT findings without bias.
All angiograms were analyzed off-line for the presence of stent edge dissections by 2 experienced interventional cardiologists blinded to the OCT findings. The target lesion was classified according to the American College of Cardiology/American Heart Association classification system (17). Severe calcification was characterized by the presence of radiopacities noted without cardiac motion before contrast injection, and generally compromising both sides of the arterial lumen (18). Severe angulation was defined as an artery bend >60° (17). Haziness was characterized as the presence of nonhomogeneous reduced contrast density within the arterial lumen, not fulfilling the criteria of definite thrombus or dissection (19,20). The step-up/step-down phenomenon was characterized by an abrupt transition of the vessel caliber between the proximal and distal stent borders to its respective references (21). Severity of stent edge dissections was classified according to the National Heart, Lung, and Blood Institute criteria as types A through F (9). Final coronary epicardial flow was classified according to the Thrombolysis In Myocardial Infarction criteria from 0 to 3 (22).
OCT image acquisition and analysis
OCT images were acquired with a commercially available Fourier-Domain OCT system (C7 XR OCT Intravascular Imaging System, St. Jude Medical, St. Paul, Minnesota) after intracoronary administration of 100 to 200 μg of nitroglycerin (unless clinically contraindicated) through conventional 6-F or 7-F guiding catheters. A 0.014-inch angioplasty guidewire was positioned distally in the target vessel, and the OCT catheter (C7 Dragonfly, St. Jude Medical) was advanced at least 10 mm distal to the distal stent edge. The entire length of the stented region was scanned using an integrated automated motorized pullback device at a speed of 20 mm/s. Continuous flush of iodine contrast medium (8 to 20 ml), through a power injector (3 to 7 ml/s at 300 psi), was performed to create a virtually blood-free environment and to trigger image acquisition. In cases with stented segments longer than 54 mm (the maximum scan length of the current OCT system), performance of another pullback was recommended to ensure visualization of the proximal stent edge. All OCT images were digitally stored and submitted to the institution's cardiovascular imaging core laboratory for off-line independent analysis.
All OCT runs were carefully screened for quality, and stent edges with <5 mm length or those with poor visualization due to artifacts or inadequate blood clearance were excluded. All images were analyzed at 0.2-mm (every-frame) intervals with dedicated automated contour-detection software (OCT system software B.0.1, St. Jude Medical).
Stent borders were defined as the first and last cross sections of the stented segment where struts could be seen in all 4 quadrants of the image. The peristent region consisted of the first frame immediately following the stent border (0.2 mm apart). Stent edges were defined as the 5-mm regions immediately adjacent to the stent borders, both distally and proximally. Edge dissection was defined as disruptions of the arterial lumen surface in the stent edge segment seen in at least 2 consecutive cross-section images. Its longitudinal length was calculated by multiplying the number of consecutive cross sections containing dissection by the slice thickness (pullback speed [mm/s] × frame rate [s]).
Dissections were morphologically classified as flaps and cavities. Specific morphometric parameters for each flap (Fig. 1A) and cavity (Fig. 1B) were calculated to estimate the magnitude of dissections.
The severity of each dissection was further estimated according to a qualitative assessment of the depth of vessel injury as: 1) intimal (limited to the intima layer or atheroma); 2) medial (extending into the media layer); and 3) adventitial (extending through the external elastic membrane) (23) (Fig. 1C). Presence of intramural hematoma was defined as an accumulation of flushing medium (or blood) within the medial space, displacing the internal elastic membrane inward and the external elastic membrane outward (23). The intramural hematoma area was determined by planimetry, and its longitudinal length calculated (Fig. 1D). Intraluminal thrombus was characterized as a mass attached to the luminal surface or floating within the lumen (23) (Fig. 1D).
Presence of atherosclerotic disease at the stent edges was investigated (Fig. 2). A “normal” vessel was characterized by the presence of a 3-layered architecture comprising the evidence of intimal, medial, and adventitial layers (23) with an intimal thickness <250 μm (24). Whenever atherosclerotic disease was present, plaques were classified according to previously described criteria as fibrotic, fibrocalcific, or lipid-rich (23,25) (Fig. 2B). Circumferential distribution of calcium and calcium-to-lumen distance were determined for all calcified plaques (Fig. 2C). In lipid-rich plaques, the entire extension of the overlying fibrous cap (FC) was semiautomatically segmented using a custom-built software, and the FC surface area and its minimum and mean thicknesses were automatically calculated (26) (Fig. 2C). Thin-cap fibroatheroma (TCFA) was defined as a lipid-rich plaque with minimum FC thickness ≤65 μm (27).
At the stent border, the stent area was measured, and the stent eccentricity index was calculated as: (maximum stent diameter − minimum stent diameter)/maximum stent diameter. Lumen eccentricity index was determined as (maximum lumen diameter − minimum lumen diameter)/maximum lumen diameter. Lumen area stenosis was calculated as: ([mean lumen area − minimum lumen area]/mean lumen area × 100). Stent diameter/lumen diameter ratio and stent area/lumen area ratio were calculated at the peristent region and the 5-mm edges to investigate the degree of vessel stretching by the stent.
Patient follow-up and clinical outcomes
Patients were followed up by telephone or hospital visit at 1, 6, and 12 months. MACE was recorded in a cumulative and hierarchical fashion, comprising the composite of death, nonfatal myocardial infarction (MI), and target lesion revascularization. Unless otherwise documented, all deaths were considered of cardiac origin. Occurrence of MI was defined according to the Joint ESC/ACCF/AHA/WHF Task Force for Redefinition of Myocardial Infarction (28). Target lesion revascularization was defined as any revascularization procedure (either PCI or coronary artery bypass graft) of the target lesion in the presence of angiographic restenosis and signs and symptoms of ischemia. Target vessel revascularization was defined as any repeat revascularization (either PCI or coronary artery bypass graft) of the target vessel, including upstream and downstream branches and the target lesion itself. Stent thromboses were classified according to the Academic Research Consortium criteria (29).
All statistical analyses were performed using SAS software version 9.2 (SAS Institute, Cary, North Carolina), and statistical significance was assessed at the 0.05 level. Continuous variables are expressed as mean ± SD, and categorical variables are expressed as counts and percentages. For per-patient analyses, differences between the 2 groups were evaluated by the unpaired Student t test for continuous variables, and Fisher exact test for categorical variables. To estimate the differences between the 2 groups at the lesion or segment level, multilevel mixed models (which can address random effects at lesion, vessel, and subject levels) were used for binary and continuous outcomes comparison. Multivariate logistic regression models considering hierarchical data structure were developed, and odds ratios (ORs) were used to evaluate factors associated with dissection identification after adjusting significant univariate predictors.
Incidence of OCT-detected stent edge dissection
Overall, 321 patients (345 lesions in 339 vessels) underwent intracoronary OCT investigation between September 2010 and June 2011, and during this period, OCT was performed after stenting in 266 patients (290 lesions in 276 vessels). After excluding cases in which OCT analysis could not be performed in part or in all of the pre-defined 5-mm edge segments, there were available for the final analysis a total of 395 stent edges (from 249 lesions in 236 vessels of 230 patients). A flowchart with the reasons for exclusion and the number of cases included in each phase of the study is presented in Figure 3. Overall, the incidence of OCT-detected stent edge dissection was 39.1% per patient (90 of 230 patients), 38.6% per vessel (91 of 236 vessels), 37.8% per lesion (94 of 249 lesions), and 26.8% per stent edge (106 of 395 edges). Although not significant, OCT-detected edge dissections were slightly more frequent during the study's Phase II (42.6% [52 of 122] of the patients; 41.7% [53 of 127] of the vessels; 40.9% [56 of 137] of the lesions; and 28% [60 of 214] of the analyzed edges) than in Phase I (35.2% [38 of 108] of the patients; 34.9% [38 of 109] of the vessels; 33.9% [38 of 112] of the lesions; and 25.4% [46 of 181] of the edges; p = 0.280 for patient-level, p = 0.369 for vessel-level, p = 0.371 for lesion-level, and p = 0.542 for edge-level comparisons). Dissections were more frequently located at distal edges (67.9%; 72 of 106 edges) than at proximal edges (32.1%; 34 of 106 edges). Nine of 90 patients (10%) presented dissections at both distal and proximal edges of the same lesion, and 4 patients (4.4%) presented at least 1 dissection in more than 1 treated lesion in the same vessel.
Baseline clinical, angiographic, and procedural characteristics
Baseline clinical characteristics are presented in Table 1. The population with stent edge dissection was older, had higher rates of hypertension, and had a lower incidence of smokers in comparison to the group without dissection. Acute coronary syndromes were responsible for 68.7% of all PCI indications. Baseline angiographic characteristics are presented in Table 2. The left anterior descending artery was the most frequently treated vessel in both groups. In comparison to the group without dissection, patients with edge dissection had more complex (American College of Cardiology/American Heart Association type B2/C) lesions (79.8% vs. 63.2%, p = 0.034), higher prevalence of severe vessel angulation (14.9% vs. 5.2%, p = 0.041), and severe calcification (42.6% vs. 14.3%, p = 0.002), smaller reference vessel diameters (2.61 ± 0.67 vs. 2.93 ± 0.76, p = 0.011), and smaller minimal lumen diameters (0.69 ± 0.47 vs. 0.88 ± 0.55, p = 0.027).
Only 17 (16%) of the total 106 edge dissections identified by OCT were also visualized by angiography, of which 14 (82.4%) were deemed type A, and 3 (17.6%) type B. Lesions with OCT-detected edge dissection had a higher prevalence of haziness (2.1% vs. 5.7%, p = 0.084) and step-up/step-down (8.8% vs. 19.8%, p = 0.001) by angiography. There were no cases of false-positive dissections indicated by angiography. No significant angiographic differences were observed between distal and proximal dissected edges (Fig. 4).
OCT qualitative assessment of stent edges
Qualitative assessment of stent edges by OCT is presented in Table 3. Overall, a high frequency of atherosclerotic plaques was seen at stent edges, particularly at OCT-detected dissected edges (95.3% vs. 75.4%, p < 0.001). Fibrotic plaques were more frequent in nondissected edges (50.0% vs. 35.6%, p = 0.029). Edges with dissection had wider (125.76 ± 77.94° vs. 65.39 ± 44.30°, p < 0.001) and more superficial calcifications (calcium-to-lumen distance: 0.07 ± 0.07 mm vs. 0.15 ± 0.12 mm, p = 0.002) than nondissected edges. Likewise, lipid plaques in dissected edges showed a trend towards a larger circumferential distribution, and had smaller minimum FC thickness (65.90 ± 38.33 μm vs. 103.07 ± 44.63 μm, p = 0.018). TCFAs were significantly more frequent at dissected edges than at nondissected edges (58.3% vs. 19%, p = 0.037).
Atherosclerotic disease was similarly encountered at both distal and proximal dissected edges. Although fibrotic plaques were more prevalent at distal edges (47.1% vs. 12.1%, p = 0.022), lipid plaques had significantly thinner FCs at proximal edges (60.83 ± 30.90 μm vs. 70.97 ± 45.41 μm, p = 0.044) (Online Table 1).
OCT quantitative assessment of stent edges
Quantitative OCT assessment is also presented in Table 3. Despite higher stent eccentricity in edges with dissection (0.11 ± 0.08 vs. 0.08 ± 0.05, p < 0.001), similar stent areas were observed in both dissected and nondissected edges. Mean lumen area and minimal lumen area were significantly smaller in dissected edges, with higher lumen eccentricity in comparison to nondissected edges. After PCI, percent diameter stenosis and percent area stenosis were also significantly higher in dissected edges.
Dissected edges had significantly higher stent/artery ratios. Interestingly, in comparison to distal dissected edges, proximal dissected edges presented significantly smaller stent diameter/lumen diameter (1.04 ± 0.13 vs. 1.13 ± 0.11, p = 0.005) and stent area/lumen area (1.07 ± 0.24 vs. 1.27 ± 0.25, p = 0.006) ratios, suggesting that mechanisms beyond vessel overstretching may contribute to dissections at proximal stent edges (Online Table 1).
Predictors of OCT-detected stent edge dissection
Key independent predictors for OCT-detected edge dissections were the presence of atherosclerotic plaque at stent edges, angle of calcification in fibrocalcific plaques, minimum FC thickness in lipid-rich plaques, presence of TCFAs, stent and lumen eccentricity, stent-to-lumen diameter, and area ratios (Fig. 5).
Qualitative and morphometric analysis of stent edge dissections
Overall, longitudinal dissection length measured 2.04 ± 1.60 mm, and the majority of dissections appeared as flaps (96.2%). Forty-one dissected edges (38.7%) had more than 1 flap, and in 9 (8.5%) edges, flap and cavity coexisted. In most dissections, the depth of vascular injury was beyond the intimal/atheroma layer (48.1% medial and 4.7% adventitial). Intramural hematoma was detected in 10 (9.4%) dissected edges, with 9 of them located at distal edges. Only 1 intramural hematoma (10%) was located at a nondiseased edge, and 9 (90%) were located at sites with eccentric plaque. Intraluminal thrombus was detected in 4 (3.8%) dissected edges, with equal distribution between distal and proximal edges (Table 4). Dissections concomitantly identified by OCT and angiography (n = 17) were longer, had bigger flap dimensions, and had deeper vessel injury than dissections seen only by OCT (n = 89). All dissections seen by both OCT and angiography extended beyond the intimal/atheroma layer (Online Table 2).
Management of stent edge dissections
Of all 106 dissections, 24 (22.6%) were treated with additional stents. Treated dissections were longer and had more flaps per dissection, bigger flap dimensions, and deeper degrees of vascular injury in comparison to the dissections not treated (Table 5). Incidence of intramural hematoma was similar between the groups, and a trend was observed towards more intraluminal thrombus in treated dissections. Of all 24 dissections treated, 13 (54.2%) had been identified by both OCT and angiography, representing 76.5% of all 17 angiography-detected dissections. The rate of dissection treatment was not significantly different between the study's Phase I (9 of 46 [19.5%]) and Phase-II (15 of 60 [25%], p = 0.641), and no significant differences were observed regarding the morphometric parameters of dissections between the 2 study phases (Table 5). Factors identified by the multilevel univariate logistic regression analysis that predicted the decision to treat an OCT-detected dissection were the concomitant identification of dissection by angiography, dissection longitudinal length, and specific morphometric parameters by OCT (Table 6).
Clinical follow-up was available for 211 (91.74%) of the total 230 patients included in the study (88 of 90 patients [97.7%] with OCT-detected edge dissections and 123 of 140 patients [87.8%] without dissections) (Table 7). The mean follow-up time was 11.49 ± 1.63 months for patients with dissection and 11.02 ± 2.48 months for patients without dissection (p = 0.120). There were a total of 14 MACE (7 of 88 [7.95%] patients with dissection vs. 7 of 123 [5.69%] patients without dissection [p = 0.581]). In the dissection group, there were 3 deaths (1 of cardiac origin) and 4 nonfatal MIs, from which only 1 was related to the target lesion requiring revascularization. Of note, all events occurred more than 30 days after the index procedure, and happened in patients in whom the dissections were only seen by OCT, as the majority of dissections concomitantly seen by OCT and angiography were treated. Edge dissections in patients with and without MACE had similar morphometric characteristics (Online Tables 3 and 4).
In the group of patients without dissection, there were 5 deaths (4 of cardiac origin) and 2 nonfatal MIs, from which 1 was due to severe in-stent restenosis and required revascularization. There were no stent thrombosis in the edge dissection group and only 1 probable stent thrombosis in a patient without dissection who had a cardiac death in the first week after the index procedure (Table 7).
We report the first large comprehensive OCT evaluation of arterial wall injury at stent edges in the daily clinical practice. The present study led to the following observations: 1) edge dissections after stenting are frequently detected by OCT, with a “true” incidence of 33.9% when OCT was performed in all treated vessels (Phase I) or a “real-world” rate of 40.9% when OCT was selectively used per the operator's discretion (Phase II); 2) most OCT-detected edge dissections (84%) were not apparent on angiography; 3) atherosclerotic disease was frequent (95.3%) at dissected edges, with the odds of having a dissection increasing by more than 6 times in the presence of a plaque; 4) circumferential calcium angle, FC thickness over lipid-rich plaques, and vessel overstretching by the stent were independent predictors for the occurrence of edge dissections; 5) a relatively high frequency (22.6%) of OCT-detected dissections were treated with additional stenting, which was driven by the concomitant presence of dissection on angiography in addition to OCT findings; and 6) patients with untreated OCT-detected edge dissections had similar 12-month outcomes compared with patients without dissection.
Incidence of stent edge dissection
Edge dissections post-stenting have been observed in 5% to 23% of cases examined by IVUS (3,30). A recent OCT investigation reported the presence of stent edge dissections in 25% of 73 patients (14). During the first phase of the present study, edge dissections were detected in 33.9% of 116 treated lesions examined by OCT. Although these figures may represent the “true” incidence of OCT-detected stent edge dissection, the expected incidence of dissections in everyday practice may be relatively higher (40.9%), likely because of more selective use of OCT. Our findings further confirm that dissections are twice more likely to occur in distal edges in comparison to proximal edges (67.9% vs. 32.1%) (7,14).
Predictors for stent edge dissection
Incomplete lesion coverage has been associated with increased risk of procedural complications and poor clinical outcomes (31,32). The somewhat expected, yet alarming, high prevalence of “untreated disease” found at stent margins is in part related to PCI technique and limitations of angiography to properly determine disease severity and extension. Only 4.7% of all dissected edges were free of disease by OCT, contrasting to 24.6% of the nondissected edges (p < 0.001). The odds of having an edge dissection were 6 times higher when the stent was placed over a diseased region than over a normal vessel (OR: 6.15, 95% confidence interval: 2.09 to 18.11, p = 0.001). These results are in line with previous IVUS observations (30,33) and underscores a modifiable factor, that is, judicious PCI technique to minimize longitudinal geographic miss, that may further improve PCI results.
Beyond the presence of disease at stent edges, the type of plaque may also play an important role in the pathogenesis of edge dissections. Previous work utilizing IVUS to understand angiographic haziness at stent edges demonstrated a significantly higher prevalence of echolucent “soft” plaques in patients with peristent injury than in controls (20). In the current study, the majority of plaques were fibrotic or fibrocalcific in both dissected and nondissected edges, and OCT-defined lipid-rich plaques represented only 23.8% of plaques at dissected edges. Differences between OCT and IVUS properties to detect lipid may explain these somewhat discordant results on the association of lipid-rich plaques and stent dissection. Our study, however, expands upon these prior observations and shows that morphometric aspects of FC overlying lipid/necrotic core pose significant risk for edge dissections. The incidence of TCFA was higher in dissected edges than in nondissected edges, and minimum FC thickness over lipid-rich plaques was significantly smaller. Furthermore, FC thickness was identified as an independent predictor for the occurrence of edge dissections. Based on an exploratory receiver-operating characteristic curve analysis, a minimum FC thickness ≤80 μm was the best cutoff to predict stent edge dissection on a lipid-rich plaque, with a sensitivity and specificity of 73.9% and 72.5%, respectively.
Presence of calcification has been traditionally identified as an independent predictor of coronary dissection after either balloon angioplasty or stent implantation (7,34,35). Gonzalo et al. (14) assessed 24 stent edge dissections with OCT and found fibrocalcific plaques (43.8%) to be more frequent at distal dissected edges than were fibrotic plaques (10%, p = 0.009). In the current study, we found a similar frequency of fibrocalcific plaques in dissected edges (40.6%), but that was not significantly different from nondissected edges (30.7%, p = 0.137). Notwithstanding, the circumferential extension, but not the simple presence of calcification, was identified as an independent predictor of stent edge dissection (OR: 1.02 for every 1° increase; 95% CI: 1.00 to 1.03, p = 0.017). OCT enabled assessment of calcium depth, but this was not associated with occurrence of dissection. An exploratory receiver-operating characteristic curve analysis revealed an angle of calcification ≥72° to be the best cutoff to predict calcium-related edge dissections, with a sensitivity of 71.1% and specificity of 71.2%. These results suggest that loss of circumferential vessel compliance might be more important in the pathogenesis of calcium-related dissections rather than the presence or depth of calcification.
Mechanical factors, such as vessel overstretching by an oversized stent relative to the reference vascular dimensions, have also been suggested as an important cause of dissection at stent edges. In the present study, we found significantly greater stent/artery ratios in dissected edges than in nondissected edges. For every 1% increase in stent diameter/lumen diameter or stent area/lumen area ratios, the odds of having a dissection increased 22% and 12%, respectively. As anticipated, vessel overstretching was more pronounced in the distal compared with the proximal dissected edges. Of note, the mean stent/lumen diameter and area ratios of 1.04 and 1.07 indicate an adequate stent sizing at proximal edges, suggesting that other factors such as plaque type, FC thickness, and circumferential calcification play a more significant role in proximal edge dissections.
OCT assessment of stent edge dissections and management
Of all 106 dissections, 24 (22.6%) were treated with additional stents. As expected, treated dissections were longer, more complex, and deeper in comparison to the dissections left untreated. The presence of intramural hematoma did not impact the decision to treat; only 3 of the 10 dissections with intramural hematoma were treated. Operators were more prone to treat dissections observed on both OCT and angiography. Importantly, the time period of the study was not identified as a predictor for the decision to treat, indicating that the learning curve of interpreting OCT images did not influence the decision to further intervene upon dissected stent edges.
The arterial wall usually responds to mechanical injury through a series of cellular and molecular mechanisms implicated in neointimal hyperplasia formation and vessel remodeling, which ultimately may lead to restenosis (36). In addition, plaque disruption with exposure of its prothrombotic milieu may promote acute/subacute thrombosis (6,7). In fact, stent edge dissections have been associated with increased short- and mid-term incidence of MACE and stent thrombosis (4–7,37). In the present study, the 12-month MACE rate was relatively low in patients with and without edge dissections detected by OCT. Furthermore, OCT morphometric characteristics of untreated dissections in patients who developed MACE were not significantly different than dissections in patients who had uneventful outcomes. Virtually all dissections presented with TIMI flow grade 3 at the end of the procedure and were classified as National Heart, Lung, and Blood Institute type B or less. These observations suggest that angiographically “silent” dissections seen only by OCT have minimal impact on outcomes after stent deployment in modern PCI. However, these findings should be interpreted with caution. The more severe OCT-detected dissections were treated. As a result, edge dissections left untreated were smaller in size, had shorter longitudinal extension, and were usually more superficial, with their depths limited to the intimal/atheroma layer (Table 6). Nevertheless, the present results are reassuring in that minor, non–flow-limiting OCT-detected dissections have a benign course and are not associated with an increased incidence of adverse outcomes (34,38–41).
The present study is insufficient to support specific criteria to guide clinical decision making on which OCT-detected dissection should be treated, but it indicates that dissections with longitudinal length ≤1.75 mm, with <2 concomitant flaps, flap depth ≤0.52 mm, flap opening ≤0.33 mm, and not extending deeper than the media layer have favorable outcomes and can be left untreated. Future studies would be required to understand whether the more complex characteristics of the treated dissections in the present study would be associated with poor outcomes if left untreated. However, such study will be difficult to be conducted due to the inherently ethical limitations of leaving potentially life-threatening vascular complications untreated. Another potential limitation is that the absence of serial OCT examination precludes conclusions on the mechanisms of stent edge dissection healing and determination of predictors for dissection persistence over time, aspects that need to be investigated in future works.
Utilization of OCT in daily practice reveals a high rate of stent-edge dissections. Dissections were associated with atherosclerosis at the stent edge and PCI technique. A small minority of dissections was angiographically evident, though with detailed OCT assessment, subsequent management was altered. Non–flow-limiting, short, and superficial dissections proved to be benign when left untreated.
For supplementary tables, please see the online version of this paper.
Dr. Bezerra has received honoraria and research funding from St. Jude Medical. Dr. Attizzani has received consultant fees from St. Jude Medical. Dr. Stefano has served as a consultant for St. Jude Medical. Dr. Simon has served on the advisory boards of Medtronic Vascular, Cordis/Johnson & Johnson, Merck, and Janssen/Johnson & Johnson. Dr. Costa has received honoraria and research funding from St. Jude Medical, and is on their Speakers' Bureau. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- fibrous cap
- intravascular ultrasound
- major adverse cardiac event(s)
- myocardial infarction
- optical coherence tomography
- odds ratio
- percutaneous coronary intervention
- thin-cap fibroatheroma
- Received January 16, 2013.
- Revision received March 18, 2013.
- Accepted March 22, 2013.
- American College of Cardiology Foundation
- Waller B.F.
- Farb A.,
- Virmani R.,
- Atkinson J.B.,
- Kolodgie F.D.
- Cutlip D.E.,
- Baim D.S.,
- Ho K.K.,
- et al.
- Cheneau E.,
- Leborgne L.,
- Mintz G.S.,
- et al.
- Biondi-Zoccai G.G.,
- Agostoni P.,
- Sangiorgi G.M.,
- et al.
- Choi S.Y.,
- Witzenbichler B.,
- Maehara A.,
- et al.
- Holmes D.R. Jr..,
- Holubkov R.,
- Vlietstra R.E.,
- et al.
- Bezerra H.G.,
- Costa M.A.,
- Guagliumi G.,
- Rollins A.M.,
- Simon D.I.
- Bouma B.E.,
- Tearney G.J.,
- Yabushita H.,
- et al.
- Gonzalo N.,
- Serruys P.W.,
- Okamura T.,
- et al.
- Ellis S.G.,
- Vandormael M.G.,
- Cowley M.J.,
- et al.
- Mintz G.S.,
- Popma J.J.,
- Pichard A.D.,
- et al.
- Gibson C.M.,
- de Lemos J.A.,
- Murphy S.A.,
- et al.
- Tearney G.J.,
- Regar E.,
- Akasaka T.,
- et al.
- Yabushita H.,
- Bouma B.E.,
- Houser S.L.,
- et al.
- Thygesen K.,
- Alpert J.S.,
- White H.D.,
- et al.,
- Joint ESC/ACCF/AHA/WHF Task Force for the Redefinition of Myocardial Infarction
- Cutlip D.E.,
- Windecker S.,
- Mehran R.,
- et al.,
- Academic Research Consortium
- Liu X.,
- Tsujita K.,
- Maehara A.,
- et al.
- Lemos P.A.,
- Saia F.,
- Ligthart J.M.,
- et al.
- Fitzgerald P.J.,
- Ports T.A.,
- Yock P.G.
- Costa M.A.,
- Simon D.I.
- van Werkum J.W.,
- Heestermans A.A.,
- Zomer A.C.,
- et al.
- Leimgruber P.P.,
- Roubin G.S.,
- Anderson H.V.,
- et al.
- Bourassa M.G.,
- Lesperance J.,
- Eastwood C.,
- et al.
- Hermans W.R.,
- Rensing B.J.,
- Foley D.P.,
- et al.