Outcomes of Percutaneous Coronary Intervention in Patients with Spontaneous Coronary Artery Dissection

Objectives To compare outcomes of percutaneous coronary intervention (PCI) in spontaneous coronary artery dissection (SCAD) patients versus conservative therapy. Background SCAD is an important cause of myocardial infarction (MI) in young-to-middle-aged women. Percutaneous coronary intervention (PCI) is often pursued, but outcomes compared to conservative therapy are unclear. Methods 403 nonatherosclerotic SCAD patients were enrolled between 2011 and 2017 and prospectively followed up in our Vancouver General Hospital registries. Detailed baseline, hospital, PCI, and outcomes were recorded. We explored the outcomes of SCAD patients who underwent PCI during their initial presentation. Results PCI was performed in 75 patients, the average age was 48.9 ± 10.1 yrs, and 94.7% were women. All presented with MI; 50.7% STEMI, 49.3% NSTEMI, and 13.3% had VT/VF. PCI was successful in 34.7%, partially successful in 37.3%, and unsuccessful in 28.0%. Stents were deployed in 73.3%, 16.0% had balloon angioplasty alone, 10.7% had wiring attempts only, and 5.3% required bailout surgery. Major adverse cardiovascular event rates (MACE) were significantly higher with the PCI group in hospital (29.3% versus 2.8%, p < 0.001), and at median follow-up of 3.7 yrs (58.7% versus 22.6% (p < 0.001) compared to the non-PCI group. Conclusion PCI in SCAD patients was associated with high failure rate and MACE in hospital and at long-term follow-up. These findings support the recommendation of conservative therapy as first-line management unless high-risk features are present.


Introduction
Spontaneous Coronary Artery Dissection (SCAD) is defined as a spontaneous, nontraumatic, and noniatrogenic dissection of the coronary arterial wall, not related to atherosclerosis [1]. SCAD is characterized by the development of intramural hematoma with or without intimal disruption, leading to variable degrees of luminal narrowing causing ischemic myocardial injury. e origin of the intramural hematoma within the arterial wall could be from intimal tear leading to dissection or from bleeding due to ruptured vasa vasorum in the absence of an intimal tear [2,3]. e true prevalence of SCAD is unknown, mostly because it remains an underdiagnosed condition and underreported cause of acute coronary syndrome (ACS). Recent studies showed that SCAD was the cause of 0.1-4% of all ACS cases [1,4]. It is more prevalent in young-to-middle-aged women. It was reported to cause 24-35% of ACS among women under fifty years of age [5][6][7] and was the most common cause of ACS during pregnancy (43%) [8].
e optimal treatment strategy for SCAD remains controversial, although conservative therapy is generally the preferred strategy, particularly for stable patients with no further evidence of ischemia. Revascularization is typically pursued for unstable patients, such as those with ongoing ischemia, hemodynamic instability, or left main involvement [9,10]. Percutaneous coronary intervention (PCI) is the preferred revascularization modality, and coronary artery bypass grafting (CABG) is relegated as a bailout strategy for failed PCI or for patients with left main dissection. One of the key rationales for conservative management is derived from the observation that spontaneous angiographic healing occurs in the vast majority of cases after about a month [11]. Furthermore, small retrospective series had shown that PCI of SCAD lesions can be challenging and associated with poor outcomes [1].
ere has been no randomized controlled trial performed to date comparing conservative therapy to PCI in patients presenting with acute SCAD, and published observational series assessing PCI outcomes with SCAD were small and lack long-term follow-up. erefore, the objective of our study is to assess the acute and long-term outcomes in a relatively large cohort of SCAD patients who underwent PCI and compare the outcomes to conservatively managed SCAD patients.

Methods
We retrospectively analyzed our cohort of nonatherosclerotic SCAD patients who were enrolled in our Vancouver General Hospital SCAD registries and prospectively followed up at the Vancouver SCAD Clinic. We included patients who underwent PCI and compared their outcomes to patients treated conservatively. All patients provided informed consent for our SCAD registries approved by the University of British Columbia Research Ethics Board. Patients were interviewed and completed detailed questionnaires on potential predisposing and precipitating stressors, gynecologic history, clinical symptoms, and family history. Detailed baseline demographics, medical history, clinical presentation, laboratory results, angiographic findings, and PCI procedural details were recorded. In-hospital and long-term cardiovascular events were collected. Screening for extracoronary fibromuscular dysplasia (FMD) in 3 arterial territories (renal, iliac, and cerebrovascular) was performed with catheter angiography or CT/ MR angiography. All patients were followed up at least annually at the SCAD clinic or by telephone follow-up.
All coronary angiograms were reviewed by two experienced cardiologists for SCAD diagnosis and angiographic classification as previously described [12][13][14]. Type 1 angiographic SCAD describes the classic appearance of contrast dye staining of arterial wall with multiple radiolucent lumens. Type 2 angiographic SCAD describes diffuse smooth narrowing that can vary in severity; variant 2A has normal arterial segments proximal and distal to the dissection, whereas variant 2B has dissection that extends to the distal tip of the artery. Type 3 angiographic SCAD describes focal or tubular stenosis that mimics atherosclerosis, and usually requires optical coherence tomography (OCT) or intravascular ultrasound (IVUS) to prove intramural hematoma or double lumen. e coronary artery segment dissected was defined by the Bypass Angioplasty Revascularization Investigation Classification [2,15]. Lesion length and stenosis severity were measured by quantitative coronary analysis (QCA).
PCI outcomes were defined as follows: (a) successful PCI was defined as angioplasty or stenting of the dissection with TIMI 3 flow and no residual dissection ( Figure 1); (b) partially successful PCI was defined as angioplasty or stenting with residual dissection or stenosis ≤50% of lumen diameter, and with final TIMI 3 or improved flow ( Figure 2); and (c) unsuccessful PCI was defined as angioplasty or stenting with residual dissection or stenosis >50% of lumen diameter or worsened TIMI flow compared to baseline or extension of dissection requiring bailout CABG ( Figure 3). Spontaneous angiographic healing at follow-up angiography was defined as angiographic resolution of the coronary dissection with residual stenosis <50% and no further evidence of multiple lumen or contrast wall staining.
In-hospital major adverse events (MAEs) were defined as a composite of all-cause mortality, stroke, reinfarction, cardiogenic shock, congestive heart failure, severe ventricular arrhythmia (requiring defibrillation or antiarrhythmic agents), repeat revascularization (or unplanned revascularization), and cardiac transplantation. Long-term major adverse cardiovascular events (MACEs) included a composite of all-cause mortality, stroke, recurrent MI (including recurrent SCAD), congestive heart failure, and revascularization. SCAD extension was defined as proximal or distal extension of the original SCAD lesion. Recurrent SCAD was defined as de novo recurrent spontaneous dissection with new recurrent MI symptoms and cardiac biomarkers elevation, which did not involve extension of dissection of the original SCAD lesion.

Statistical Analysis.
Descriptive statistics were used to summarize baseline characteristics. Continuous variables were reported as mean ± SD or median and interquartile range. Categorical variables were reported as frequency and percentage. Categorical data were compared with the chisquare or Fisher's exact tests. Continuous data were compared using Student's t-test or Mann-Whitney test. Twosided p values of <0.05 were considered significant. e logrank test was performed to compare groups in survival analysis. Statistical analyses were performed with the SPSS software (IBM SPSS version 23, New York).

Results
Four hundred and three SCAD patients enrolled between 2011 and 2017 were included in this analysis; 75 (18.6%) underwent PCI of the SCAD-affected artery, and 328 (81.4%) were treated conservatively during their initial SCAD hospitalization. Baseline characteristics, risk factors, predisposing/precipitating factors, and hospital presentation of the patients are shown in Table 1. Patients in both groups were predominantly female and Caucasian. e mean age was lower in the PCI group (48.9 ± 10.1 yrs) compared to the non-PCI group (53.1 ± 9.6 yrs), p � 0.001. e PCI group also had a lower prevalence of hypertension (25.3% versus 38.7%, p � 0.033) and FMD (46.7% versus 64.0%, p � 0.008) and were less likely to be postmenopausal (50.7% versus 67.7%, p � 0.009). No patients were lost to follow-up in the trial period.

Discussion
We performed a retrospective analysis comparing the clinical outcomes of those who underwent PCI at the index SCAD event compared to those treated conservatively. Patients who underwent PCI had higher-risk baseline characteristics, including more STEMI presentations, worse LVEF, higher troponin elevation, longer lesion length, greater stenosis severity, larger-diameter arteries dissected, and more left main and proximal artery dissections. PCI procedures were successful or partially successful in 72% of cases and unsuccessful in 28%. e majority (73.3%) were treated with stent placement. Both in-hospital MAE and overall MACE at long-term follow-up were higher in the PCI group, primarily driven by rates of repeat revascularization and overall MI.
Acute management of SCAD remains challenging mainly due the lack of randomized trials and the fact that this disease is still underrecognized and underreported. Conservative therapy is recommended as first-line management based on expert consensus from observational and retrospective studies [9,10]. However, SCAD patients with high-risk characteristics such as ongoing ischemia, hemodynamic instability, or left main dissection often require urgent revascularization to relieve ischemia and provide myocardial salvage. Unfortunately, PCI of SCAD-affected arteries can often be challenging and carries a high risk of failure and suboptimal outcomes [1]. Reported success rates of PCI with SCAD lesions ranged from 47%-91% in small retrospective series [1,12,13,16,18,19]. SCAD-affected arteries are more susceptible to iatrogenic catheter-induced  dissection [20] and propagation of the dissection induced by wire manipulation, angioplasty, or stenting, which can lead to antegrade or retrograde extension of dissections during PCI. SCAD also usually involves long segments of arteries requiring long stents, which can increase the risk of stent restenosis and thrombosis. Furthermore, undersizing of the stents in the setting of intramural hematoma may result in malapposition of the struts after resorption of the hematoma, with increased risk of late in-stent thrombosis [17]. erefore, PCI for SCAD lesions should be reserved for patients with clinical high-risk features, and a suggested algorithm on the PCI strategy with SCAD was recently described [21].
Previous retrospective studies have reported on the suboptimal outcomes with SCAD PCI. Our current series differs in being one of the largest series and with details on PCI rationale and strategies and comparative long-term outcomes to a large cohort of patients managed conservatively. In our series, only a small and selected proportion of patients (18.6%) underwent PCI in the overall cohort, with the most common rationale for PCI being ongoing ischemia/ symptoms, hemodynamic instability, severe stenosis, proximal artery involvement, and TIMI 0 or 1 flow. Inhospital MAE was higher in the PCI group, with higher recurrent MI, repeat revascularization, and stroke. After discharge, repeat revascularization and admission for unstable angina remained higher with the PCI group. In the Mayo Clinic series of 187 SCAD patients, almost half of their patients (87/187, 46.5%) were treated with PCI, suggesting a much lower threshold for intervention in their series (rationale for PCI not provided). eir procedural success rate was only 47%, which is lower than our reported rate, and this may be related to differences in definition of PCI success used. As well, the PCI strategy (i.e., proportion receiving stents or angioplasty alone) in this study was not reported. A higher proportion (13%) of their patients required bailout CABG. ey found no difference in the 5-year rates of target vessel revascularization and recurrent SCAD in PCI versus conservative therapy groups (30% versus 19%, p � 0.06; 23% versus 31%, p � 0.7, respectively) [13]. In contrast, we found higher rates of repeat revascularization both in hospital and following discharge in the PCI group. It is difficult to compare the differences in repeat revascularization in these 2 studies, since there may be differences in PCI strategies and subjective indications for reintervention.
In the series by Lettieri of 134 SCAD patients, 51 underwent PCI (41.8%) and 5 (3.7%) underwent CABG. Successful PCI was achieved in 72.5%, and patients treated   [16]. In a meta-analysis of 11 studies published by Martins, among 631 SCAD patients in these nonrandomized studies, 253 were treated with PCI or CABG as the initial strategy. ey found no difference in mortality, MI, or SCAD recurrence with revascularization versus conservative therapy. However, revascularization was associated with an estimated additional risk of target vessel revascularization of 6.3% [22]. ese findings are concordant with our study. In our study, the majority of patients who underwent PCI were treated with stents (73.3%), and angioplasty alone was performed in 16.0%. In 10.7%, wiring was unsuccessful. e optimal approach with PCI in SCAD lesions is unclear at this point, with several strategies that are feasible including balloon angioplasty alone, cutting balloon fenestration, sequential stenting, stenting edges first before middle, and use of bioabsorbable stents [21]. e selected PCI approach should be individualized and may be guided according to the anatomic appearance and extent of dissection. Nevertheless, our study supports the current recommendations that conservative therapy should be first-line, unless patients have high-risk features. Furthermore, studies that assessed repeat angiography for conservatively managed SCAD patients also showed that spontaneous healing occurs in the vast majority of cases (95% after 30 days) [11]. In the current series, ∼86% of conservatively managed patients had spontaneous healing on repeat angiography. e American Heart Association SCAD Scientific Statement recommended conservative management for clinically stable patients with no high-risk anatomy; PCI or CABG should be considered for patients with active or ongoing ischemia or hemodynamic instability. In stable patients with left main or severe proximal 2-vessel dissection, CABG or conservative management may be considered [9]. In the European Society of Cardiology SCAD position paper, it was highlighted that revascularization was associated with an increased risk of complications, and a conservative approach was recommended in the absence of ongoing ischemia [10].

Limitations
Our study is retrospective and observational and, therefore, subject to bias of patient selection for PCI versus conservative therapy. Selection of a treatment strategy was at the discretion of the operator, and provision of the rationale for PCI indicated that more high-risk patients were selected for PCI. Indeed, patients who underwent PCI had larger MIs, worse LVEF, and more ominous geographic SCAD anatomy. us, the worse outcomes observed with PCI may be related to these differences in baseline demographics, in addition to the high procedural failure. Interestingly, although overall MACE was higher with the PCI group, this was primarily driven by repeat revascularization and in hospital MI, with no significant difference in long-term mortality, recurrent MI, or recurrent SCAD. is suggests that the dominant consequence of PCI in SCAD arteries is related to mechanical interventional issues, as opposed to systemic effects, and does not affect future risk of SCAD.

Conclusions
In our large SCAD series, patients who underwent PCI had higher baseline and angiographic risk characteristics. PCI was associated with low procedural success and higher inhospital complications of recurrent MI, repeat revascularization, and stroke, and also long-term risk of repeat revascularization, compared to conservative therapy. Our study supports the current societal expert recommendations of conservative therapy as first-line treatment for SCAD patients. Ideally, a randomized study comparing PCI to conservative therapy should be performed; however, the logistics of treating high-risk SCAD patients (e.g., left main dissection, ongoing ischemia) conservatively or treating lowrisk SCAD patients with PCI (where most heal spontaneously) are fraught with ethical challenges and are against current practice recommendations. erefore, although our study is observational and retrospective, this is currently the best-available evidence to evaluate revascularization indications and strategies with SCAD. Further studies are required to assess contemporary mechanical strategies to improve SCAD PCI outcomes.

Data Availability
All data used in the study are part of the Canadian SCAD Registry and are subject to registry data storage protocols.

Conflicts of Interest
e authors declare that they have no conflicts of interest.