Introduction
Acute carotid stent occlusion (CSO) is a rare cause of acute ischemic stroke that is associated with significant morbidity and mortality [1]. CSO occurs in 0.05% to 0.8% of patients with the internal carotid artery (ICA) or common carotid artery stents and is caused by antiplatelet medication noncompliance or discontinuation, antiplatelet medication resistance, overlapping stent placement, or intrinsic prothrombotic disorders [1-4]. In addition, procedural events and complications, such as dissection, atheroma perturbation, or ICA kinking after stent placement, may predispose a stent to occlusion [5].
CSO treatments include conservative medical therapy, endovascular treatment (ET), surgical stent explantation, carotid endarterectomy, or a combination of these approaches [1,3,5-7]. ET of CSO presents challenges for neurointerventionalists, and the risk of revascularization techniques must be balanced against the risk of clot propagation, carotid stent damage, and reperfusion injury [3,5]. Described ET techniques for CSO include intra-arterial (IA) thrombolysis often with tissue plasminogen activator (tPA) or glycoprotein IIb/IIIa receptor inhibitors and aspiration thrombectomy with or without mechanical thrombectomy [5,8-12]. Angioplasty has been described for in-stent stenosis and intraprocedural expansion of incompletely secured stents with thrombus formation, but not for CSO [3,5,13-14].
We present the first report of combined angioplasty and aspiration thrombectomy for the treatment of acute CSO. All patients underwent angioplasty to promote thrombus disruption and restore antegrade flow through the occluded carotid stent. Residual in-stent thrombus was removed using aspiration thrombectomy. This technique is described in detail, and its effectiveness is compared to the literature for ET of CSO.
Materials and methods
Patient information
The study was approved by the Institutional Review Board (IRB) and complied with the Health Insurance Portability and Accountability Act. The need for informed consent was waived the IRB. We retrospectively reviewed our neurointerventional database to identify consecutive patients who underwent ET for acute CSO treatment between January 2008 and March 2018. Patient demographics, endovascular treatment details, and outcome data were determined from the electronic medical record.
Among patients who underwent pre-interventional perfusion imaging, automated post-processing was performed using RApid processing of PerfusIon and Diffusion (RAPID) software (iSchemaView, Menlo Park, CA). Core infarct and penumbral volumes (defined as the volume of tissue with time-to-maximum (Tmax) >6 seconds) were determined using RAPID. Patients had pre-interventional computed tomography angiography (CTA) or magnetic resonance angiography (MRA). In one patient, the CTA was non-diagnostic due to technical issues. In this patient, a virtual CT angiogram was reconstructed from the perfusion source images.
All patients who undergo carotid stent placement at our institution undergo surveillance CTA at 3, 6, and 12 months to evaluate for in-stent stenosis. However, patients who present with CSO before these follow-up appointments were not screened for an in-stent stenosis. Therefore, only one patient in this series underwent follow-up imaging due to delayed presentation (patient four). Antiplatelet resistance in patients with verified medication compliance, and no other identified cause of CSO was verified using thromboelastography, with secondary testing by a hemostasis platelet function assay-100 (PFA-100) system (Siemens, Tarrytown, NY).
Endovascular treatment of CSO
All patients underwent ET in a biplane neuroangiography suite (Axiom Artis, Siemens) under either monitored anesthesia care or general anesthesia. Common femoral artery access was obtained using standard techniques, and an 8- or 9-French sheath was placed in the descending thoracic aorta. Access into the common carotid artery of the affected hemisphere was obtained with a 5-French Berenstein angiographic catheter (Cordis, Milpitas, CA), which was placed through a 6-French shuttle sheath (Cook Medical, Bloomington, IN). Selective digital subtraction angiography (DSA) of the affected common carotid artery was performed prior to ET.
The affected common carotid artery was accessed (Figure 1A), and an aspiration catheter was advanced and withdrawn through the thrombosed stent into the more distal cervical ICA under continuous aspiration to at least partially recanalize the CSO (Figure 1B). Next, an embolic protection device (Accunet, Abbott, Abbott Park, IL) was placed distal to the stent over a guidewire (Figure 1C) in three patients. The embolic protection device could not be navigated through the partially occluded stent in the fourth patient. A balloon catheter was then advanced within the partially recanalized stent and step-wise angioplasty was performed to macerate the residual thrombus against the stent wall (Figure 1C). Different balloon catheters were used for each patient: a 4x12mm monorail balloon catheter (Boston Scientific, Marlborough, MA), a non-compliant Trek 5x15mm balloon microcatheter (Abbott, Abbott Park, IL), a Trek rapid exchange 5x12mm balloon catheter (Abbott, Abbott Park, IL), and a Viatrac 6x20mm balloon catheter (Abbott, Abbott Park, IL). Post-angioplasty DSA was performed demonstrating improved stent caliber with minimal residual thrombus (Figure 1D).
Figure 1
Schematic of endovascular treatment of CSO
Schematic for endovascular recanalization of CSO. After obtaining access to the affected common carotid artery (A), an aspiration catheter is advanced through the thrombosed stent under continuous aspiration and removed (B). An embolic protection device is placed distal to the stent, and angioplasty is performed to macerate the residual thrombus against the stent wall (C). Post-angioplasty DSA demonstrates stent recanalization with minimal residual thrombus (D).
CSO, carotid stent occlusion; DSA, digital subtraction angiography
Tandem intracranial occlusions were treated with a combination of IA-tPA, aspiration, and mechanical thrombectomy. Detailed ET descriptions for each patient are presented in the Appendix Text.
Technique effect and outcome metrics
The primary outcome was successful CSO recanalization and cerebral vascular reperfusion (modified thrombolysis in cerebral infarction, (mTICI) score IIB-III). Secondary outcome measures were National Institutes of Health Stroke Scale (NIHSS) shift from presentation to discharge, modified Rankin Scale (mRS) score at 3 months and mortality [15-16].
Literature review
A comprehensive review of PubMed and Embase was performed to identify all eligible acute CSO treatment studies published since January 2008, with the last update of literature review on July 2019. Search terms used included: “carotid stent” AND “occlusion”, “carotid artery stent” AND “occlusion”, “carotid stent” AND “thrombosis”, “carotid artery stent” AND “thrombosis”, or “carotid” AND “stent” AND “occlusion”. Inclusion criteria were: English language, acute onset of CSO (i.e. presentation of acute stroke symptoms leading to discovery of the CSO) irrespective from time of original stent placement, and publication dates January 2008 - July 2019. In total, 65 non-duplicate studies were identified and 51 were excluded due to: absence of complete CSO, non-acute CSO, no surgical or ET intervention performed or described, incomplete or absent characterization of CSO onset, etiology, and/or outcome (Figure 2). Two authors reviewed each study that met all inclusion criteria and did not meet exclusion criteria.
Figure 2
PRISMA flow diagram of database search and literature review
PRISMA, preferred reporting items for systematic reviews and meta-analyses; ET, endovascular therapy; CSO, carotid stent occlusion
Results
Endovascular treatment of CSO
In our institutional neurointerventional database, we identified four patients who underwent ET for symptomatic acute CSO. Patient demographic and treatment data are described in Table 1. The mean presentation NIHSS score was 14.5 ± 3.9. Three patients underwent perfusion imaging prior to treatment, and, in these patients, the mean infarct core was 0.7 ± 0.7 ml and penumbra (Tmax > 6) was 110.5 ± 59.7 ml. Three patients were treated with aspiration thrombectomy and angioplasty, and the fourth underwent angioplasty alone for CSO treatment. Procedural details are described in the above Materials and Methods section, with full technical treatment details for each patient described in the Appendix Text. Successful CSO recanalization was achieved in all four patients (100%) (Figures 3 & 4; Appendix Figures 5 & 6). Tandem intracranial occlusions were present in three patients, and two of these patients had evidence of these occlusions on pre-ET imaging (Figure 3, Appendix Figure 6). All intracranial occlusions were successfully revascularized with mTICI IIB-III reperfusion (100%; Table 1).
Table 1
Acute carotid stent occlusion patient clinical characteristics, stent information, and outcomes after endovascular therapy (ET).
ET for all patients was aspiration thrombectomy, followed by angioplasty, except patient 4 who was treated with angioplasty only.
ET, endovascular treatment; CSO, carotid stent occlusion; NIHSS, National Institutes of Health Stroke Scale; mTICI, modified thrombolysis in cerebral infarction; ICA, internal carotid artery; HTN, hypertension; DM, diabetes mellitus; HLD, hyperlipidemia; CAD, coronary artery disease; OSA, obstructive sleep apnea; MI, myocardial infarction; PVD, peripheral vascular disease; DAPT, dual antiplatelet; SAPT, single antiplatelet; IA-tPA, intra-arterial tissue plasminogen activator.
| Patient | Past Medical History | Carotid Stent Side | Pre-Stenting Stenosis | Pre-Stenting Anti-thrombotic(s) | Stent Type | Post-Stent & Pre-ET NIHSS | Latency to Acute Thrombosis (CSO) | Anti-thrombotic Regimen at time of CSO | CSO Etiology | Symptoms & NIHSS at Presentation | Time from last known normal | mTICI Score / Tandem Treatment | Post ET NIHSS | Mortality at 3 months | mRS at 3 months |
| 1 | Tonsil cancer with resection and x-ray therapy, CAD, HTN, HLD, Hypothyroidism, DM | Left | 90% of distal bulb and proximal ICA | Aspirin 81 mg, Clopidogrel 75 mg | Acculink carotid stent system 6-8 mm taper x 30 mm | 0 | 7 days | Aspirin 81 mg, Clopidogrel 75 mg | Clopidogrel resistance, inconsistent clopidogrel compliance, radiation vasculopathy | Right arm hemiparesis, NIHSS 4 | 0.5 hr | III No tandem occlusions | 0 | Alive | 1 |
| 2 | Atrial fibrillation, CAD, HTN, HLD, OSA, bilateral carotid stenosis with remote carotid endarterectomy | Right | > 90% proximal ICA | Clopidogrel 300 mg Coumadin, INR at 2-3 | Xact stent 9 x 40 mm | 0 | 8 hours | Protamine 100 mg, Clopidogrel 75 mg | Protamine reversal after heparin in the setting of atrial fibrillation | Left hemiparesis, NIHSS 13 | 0.25 hr | IIB Aspiration thrombectomy, Solitaire Retriever | 3 | Alive | 3 |
| 3 | CAD with MI, HLD, DM, smoker, PVD, chronic right ICA occlusion | Left | 80% ICA | Aspirin 81 mg, Clopidogrel 75 mg | Xact stent 8 x 40 mm | 0 | 6 days | Aspirin 81 mg, Clopidogrel 75 mg | Noncompliance with DAPT (not taking any antiplatelet) | Aphasia and right hemiparesis, NIHSS 20 | 12 hr | IIB IA-tPA | 18 | Deceased | 6 |
| 4 | CAD with MI | Left | >70% of ICA reconstruction | Clopidogrel 75 mg | Xact stent 9 x 30 mm | 0 | 18 months | Aspirin 325 mg, Clopidogrel 75 mg | Noncompliance with DAPT (not taking any antiplatelet) | Aphasia and right hemiparesis, NIHSS 21 | 7 hr | IIB Merci Retriever | 42 | Deceased | 6 |
Figure 3
Patient 1 with left CSO seven days post stent placement
(A) Coronal maximum-intensity-projection images following a CTA demonstrates a thrombus within the left internal carotid artery stent (arrow). Inset shows magnified region of stent with thrombus (arrow). (B) Left: Maximum intensity projection of the Circle of Willis CTA demonstrates no intracranial large vessel tandem occlusions. Right: CT perfusion imaging shows a perfusion deficit (Tmax >6 seconds) in the left middle cerebral artery territory (dashed outline) secondary to the carotid stent thrombus. (C-G) Left common carotid artery DSA images. There is occlusion of the left carotid stent (c, arrow) with no antegrade filling of the cervical left ICA (c) and poor filling of the intracranial left ICA (d, arrow), largely via left external carotid artery collaterals (d, dashed arrow). (E) CSO treatment by aspiration thrombectomy resulted in antegrade filling of the left ICA stent with residual non-occlusive thrombus within the stent (arrow; inset, arrow) and improved filling of the more distal cervical ICA (dashed arrow). (F) Subsequent angioplasty resulted in minimal residual thrombus within the stent (arrow). (G) Robust antegrade filling of the left anterior circulation was present after CSO treatment.
CTA, computed tomography angiography; ICA, internal carotid artery; DSA, digital subtraction angiography; CSO, carotid stent occlusion
Figure 4
Patient 2 with right CSO eight hours post stent placement with tandem anterior cerebral artery and middle cerebral artery occlusions
(A) Due to technical limitations, a CTA was not obtained, thus CT perfusion source data was reformatted into a 5mm MIP from the peak vascular enhancement series using manual bone masking at the skull base to demonstrate the pre-intervention tandem occlusions of the right A2 (arrow) and M2 (dashed arrow) segments of the anterior and middle cerebral arteries. (B) CT perfusion imaging shows a perfusion deficit (Tmax >6 seconds) in the right middle and anterior cerebral artery territory (dashed outline) secondary to the carotid stent thrombus. (C-H) Right common carotid artery DSA images. (C) There is occlusion of the right carotid stent (arrow) with no antegrade filling of the cervical right ICA (arrow). (D) Balloon angioplasty of the right carotid stent (arrow) was performed. (E) After aspiration thrombectomy and angioplasty, there is residual thrombus in the carotid stent (arrow) with additional non-occlusive thrombus in the more distal right ICA (dashed arrow). (F) After additional aspiration thrombectomy, there is minimal residual thrombus in the carotid stent (arrow) and no residual thrombus in the more distal right ICA (dashed arrow). (G) Following CSO treatment, there is poor filling of the right middle cerebral artery inferior divisions (arrow), as well as the A2 segment of the right anterior cerebral artery. (H) After combined cerebral aspiration and mechanical thrombectomy, there is excellent reperfusion of the right anterior circulation (arrow).
CT, computed tomography; CTA, computed tomography angiography; ICA, internal carotid artery; DSA, digital subtraction angiography; CSO, carotid stent occlusion
Following ET, one patient expired after a cerebral reperfusion hemorrhage. At discharge, the mean NIHSS shift in the remaining three patients was -5.3 ± 7.2. By three months after ET, one additional patient had expired from a myocardial infarction, and the two surviving patients had 3-month mRS scores of 1 and 3 (Table 1).
Literature review of acute CSO treatment
Acute CSO is most commonly secondary to antiplatelet resistance, intolerance, or noncompliance (Table 2). Previously described ET techniques for CSO treatment apply methods used for non-stented carotid occlusions, and include IA thrombolysis, IA aspiration, heparinization, and mechanical thrombectomy [5,8-12,14,17-24]. A total of 23 patients were described, with most achieving recanalization (approximately 85-90%). However, outcomes reported range from a general single classification of either improved or not improved, or a spectrum of significant new functional deficits (Table 2). The nonuniform reporting of TICI score, mRS, and mortality data limits precise comparative analytics between the studies and their techniques.
Table 2
Literature review of acute carotid stent occlusion etiology, treatment, and outcome (2008-2019)
Twenty-three patients were described; however, the success rates of ET as measured by mTICI score nor mRS are not uniformly reported.
CSO, carotid stent occlusion; IA, intraarterial; IV, intravenous; tPA, tissue plasminogen activator; CEA, carotid endarterectomy; ET, endovascular treatment; mTICI, modified thrombolysis in cerebral infarction; STA-MCA, superficial temporal artery-middle cerebral artery; DAPT, dual antiplatelet; DSA, digital subtraction angiography; mRS, modified Rankin score
[5], [11-12], [14], [19-27]
| Author | Year | N | Time to Acute CSO (patients separated by “,”) | CSO Etiology | Treatment Method | mTICI Score | Outcome | Tandem Occlusion |
| Toljan et al. | 2019 | 1 | 2 hours | Clopidogrel resistance | Aspiration thrombectomy, IA-tPA, IV eptifibatide | - | Thrombus resolved, deficits not described at 3 months | None |
| Hu et al. | 2018 | 1 | 2 minutes | Incomplete stent adherence poor expansion, baseline stenosis remained | Heparin, Intrathrombus tPA, salvage angioplasty to re-expand stent | - | Thrombus resolved, no deficits | Not reported |
| Moulakakis et al. | 2018 | 2 | 30 minutes, 1 hour | Plaque protrusion through sent, not reported | Surgical stent explantation, IA tPA, then surgical stent explantation | - | Complete resolution of symptoms in both patients | Not reported |
| Moulakakis et al. | 2017 | 4 | 1 hour, 2 hours, 3 days, 4 days | Dissection, two overlapping stents, two overlapping stents and malignancy, two overlapping stents | Aspiration thrombectomy with CEA and stent explantation, IA urokinase and aspiration thrombectomy with additional stent placement, Tinzaparin, Nadroparin with Aspirin and Clopidogrel | - | Mechanical ET with full recanalization, IV only treatment with no recanalization, all with mild residual hemiparesis or speech impairment | Not reported |
| Koklu et al. | 2015 | 1 | 1 day | Aspirin and Clopidogrel resistance | Ticlopidine, Heparin | - | Improved dysarthria, stable right hemiplegia | Not reported |
| Munich et al. | 2014 | 1 | Intraprocedural | Embolic protection device thrombus | Aspiration with penumbra 4Max | - | 25-30% residual stenosis, no new deficits | Not reported |
| Kim et al. | 2013 | 3 | Intraprocedural (all) | Embolic protection device thrombus | Forced arterial suction thrombectomy with Penumbra | III | All improved | Not reported, no indication intracranial thrombectomy was required and all TICI 3 |
| Kanemaru et al. | 2013 | 1 | 6 days | Hypercoagulability of malignancy | Aspirin, Clopidogrel, Cilostazol, Coumadin | - | Thrombus resolved, no new deficit | Not reported, no DSA |
| Markatis et al. | 2012 | 1 | 2 days | Noncompliance with DAPT | Heparin, surgical exploration, thrombectomy with stent removal | - | Sensory loss on right hand | Not reported |
| Choi et al. | 2012 | 2 | 4, 9 days | Not known, Aspirin and Clopidogrel resistance | STA-MCA anastomosis | - | Improved hemiparesis and dysarthria improved | |
| Iancu et al. | 2010 | 2 | Intraprocedural | Carotid dissection, balloon rupture | Intrathrombus Streptokinase, Intrathrombus Tenecteplase and stent secured and expanded with angioplasty | - | Thrombus resolved, no new neurological deficits | Not reported |
| Naito et al. | 2010 | 2 | 2 months, 7 days | DAPT discontinued for surgery and hypercoagulability of malignancy, noncompliance | Aspiration thrombectomy, Aspiration thrombectomy and Urokinase | Thrombus resolved, no noted new deficit | Not reported | |
| Dhall et al. | 2010 | 1 | Intraprocedural | - | IA Urokinase and Abciximab, aspiration thrombectomy | - | Thrombus resolved, no new neurological deficit | Not reported |
| Seo et al. | 2008 | 1 | Intraprocedural | Distal stent filling defect | IV Tirofiban | III | Complete recanalization |
Discussion
We describe a novel method of effective ET for symptomatic CSO using a combined aspiration thrombectomy and angioplasty technique. Our technique resulted in the successful recanalization of the CSO in 100% of patients described here and has become the standard intervention for CSO at our institution.
The etiology of CSO in our series was due to inadequate medical platelet inhibition secondary to poor medication compliance, intrinsic antiplatelet resistance, or unintentional anticoagulation reversal, which is similar to prior studies [1-4]. These etiologies most likely led to acute in-stent thrombus formation, which was amenable treatment using our technique. Furthermore, the variable timing of CSO in our series suggests that even carotid stents that are likely well endothelialized are at risk of CSO in the absence of adequate antiplatelet or anticoagulation protection.
Symptomatic acute CSO necessitates emergent intervention regardless of the interval between initial stent placement and presentation as recanalization is the single modifiable outcome predictor for carotid occlusions [18,28]. ET techniques for CSO treatment include IA-tPA, aspiration, and mechanical thrombectomy and remain the most common approach for the treatment of CSO [5,8-12,14,19-24,26]. These techniques result in CSO recanalization in most, but not all (approximately 85-90%), of cases (Table 2), which is lower than the 100% efficacy of our technique. The nonuniform reporting of mTICI score, mRS, and mortality limits outcome comparison between studies using different ET techniques.
Aspiration thrombectomy for CSO has been described as an effective treatment [12,23]. However, we found aspiration thrombectomy alone resulted in insufficient CSO recanalization in all three patients who underwent this technique before angioplasty. In our cohort, aspiration thrombectomy created a channel through the in-stent thrombus that allowed for the passage of a balloon catheter for subsequent angioplasty.
Combining other ET techniques with aspiration thrombectomy may have a good effect on full recanalization. Angioplasty for acute CSO has not been described, and we have applied this technique as an additional revascularization technique. Angioplasty is often reserved for placing and re-expanding stents that were not already fully secured and were associated with a thrombus [5,14]. Alternative techniques that combine mechanical and aspiration thrombectomy for CSO have been described as similarly effective [10-11]. However, pulling a stent retriever through a recently placed, non-endothelialized carotid stent may risk stent retriever detachment from the pusher wire.
An in-stent stenosis secondary to intimal hyperplasia may predispose to CSO. However, in our series, there was an easy passage of the balloon microcatheter and aspiration catheter through the stent and unremarkable serial surveillance CTA and CT perfusion scans, which suggests an absence of significant in-stent stenosis. In 75% of patients in our series, an embolic protection device was deployed to reduce the theoretical risk of secondary thromboembolism during angioplasty, though no debris was observed within the embolic protection devices at procedure end. Protection against emboli from the stent may be obtained with the use of a balloon-guide catheter, although these catheters are not routinely used at our institution.
Surgical management of CSO by stent explantation, carotid endarterectomy, or carotid artery bypass grafting is considered an alternative in patients with verified dual anti-platelet resistance or as a rescue procedure for failed ET [6,22]. As endovascular techniques continue to improve, we anticipate that the need for surgical treatment of CSO will be further reduced.
The frequency of tandem intracranial occlusions in the setting of CSO remains poorly described, and we found tandem occlusions in 3 of 4 of our patients. Of these patients, 2 had evidence of these occlusions prior to ET, and the third patient developed new tandem occlusions within the A2 segment of the left anterior cerebral artery and distal M2 segment of the left middle cerebral artery despite the use of a distal embolic protection device (Appendix Figure 4). This new embolic occlusion may have occurred prior to, during placement of the embolic protection device, or following angioplasty. In the absence of frequent interval intracranial imaging during ET, it is challenging to conclude if a component of the ET itself was causal for the distal emboli.
Tandem intracranial occlusions were successfully revascularized in all patients in our series using standard endovascular techniques [18,28]. Of note, IA-tPA is most commonly used for acute occlusions at the time of stent placement, for which it is an effective treatment [9]. IA-tPA was used in two patients in our series beyond six hours' time since last seen normal. The clinical benefit of IA-tPA for CSO treatment before or after six hours since last seen normal remains unclear, as in other types of stroke from large vessel occlusion in the form of carotid occlusion, it may reduce mortality but not change functional outcome [5,17-18].
A single patient in our series experienced a fatal reperfusion hemorrhage, which is a risk inherent to all endovascular reperfusion therapies. Our series did note a high 50% mortality rate at 3 months, which partially reflects both the symptomatic nature of the CSO and the severe medical co-morbidities of our cohort. The odds of a poor outcome or death would likely be higher if the recanalization of the occluded stents was not achieved. Non-intervention for CSO has not been studied, but it is likely to result in poor outcomes; comparison to patients who do not undergo thrombectomy for large vessel occlusions supports this hypothesis [29]. A recent systematic literature review found in nearly 60% of patients treated for CSO, that there was either no improvement after therapy or outcomes were not reported [30]. The same is true for nonuniform reporting of outcomes in the literature reviewed here. Notably, some studies that reported more favorable outcomes characterized patients with an asymptomatic CSO.
No procedural complications or damage to the previously placed carotid stents occurred in our series. A meta-analysis or multicenter experience will likely be required to definitively describe the procedural risk of ET for CSO given the rarity of this event.
Limitations
Our study is limited by its retrospective design, single-center experience, and small sample size. Despite these limitations, this series is, to our knowledge, the largest reported series of effective endovascular treatment of CSO. While variability in the time of presentation, and likely varying degrees of stent endothelization exist within our cohort, it is representative of the spectrum of onset for CSO. We find that the heterogeneity in presentation reflects the reality of CSO occurrence and increases the generalizability of our findings to most patients with an acute CSO.
Conclusions
Combined aspiration thrombectomy and angioplasty is a viable technique for the neurointerventional treatment of CSO and results in high rates of stent recanalization and cerebral reperfusion, which are both improved from that of previous techniques. Patient symptoms as measured by mean NIHSS shift also improved; however, the severe non-neurological comorbidities within our cohort led to high mortality by 3-month follow-up. Further multicenter studies are required to risk-stratify patients for specific ET interventions.
Copyright © 2020, Murray et al. This work is published under http://creativecommons.org/licenses/by/3.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.