A proposed modern standardized technical approach for symptomatic chronic carotid total occlusion management

Abstract

Chronic carotid total occlusion (CCTO) is a known cause of ischemic stroke and transient ischemic attack. Symptomatic CCTO is associated with up to 30% risk of recurrent ischemic stroke, despite optimal medical treatment. Notably, a randomized controlled trial reported that previous surgical management did not improve the overall prognosis of these patients. Endovascular treatment of CCTO has been proposed as a feasible strategy to re-establish cerebral perfusion in symptomatic patients. However, its use is controversial and not supported by evidence from randomized clinical trials. Recently, a meta-analysis reported a reasonably high procedural success without an excess periprocedural complication rate, but several steps are needed before the procedure is mature enough to be tested in randomized controlled trials. This review highlights the developments in the endovascular recanalization of CCTO and emphasizes key steps towards standardizing the procedure.

Introduction

Chronic carotid total occlusion (CCTO) may be benign in the absence of neurological symptoms. ¹ However, once symptomatic, it is associated with up to a 5–6% annual risk of stroke, despite the use of the best medical therapy.^2–4 Recurrent stroke risk is higher in patients with symptomatic CCTO and associated hypoperfusion. Previous retrospective studies estimated the recurrence risk to be up to 30%, while medically treated patients in the United States carotid occlusion surgery study (COSS) trial showed a 22.7% recurrent stroke risk.^1,5–7

The COSS trial was a randomized controlled trial comparing surgical extracranial-intracranial (EC-IC) bypass to best medical therapy in a high-risk sub-group of patients with hypoperfusion demonstrated with positron emission tomography imaging. After this study failed to show the superiority of the revascularization strategy, from a therapeutic point of view, symptomatic CCTO became a condition without an established hemodynamic treatment.^7–9 The COSS trial reduced the incidence of recurrent events after successful surgery when the periprocedural stroke was not considered. However, the trial was stopped early due to a high periprocedural stroke and death rate that offset the long-term benefit of re-establishing adequate cerebral perfusion. ⁷ If the periprocedural strokes (12 out of 14 in the first 30 days occurred during the first 48 h) had been prevented, the trial would have been largely positive, with a rate of recurrent stroke of 9% in the surgical group and 22.7% in the non-surgical group. ¹⁰ These data support the hypothesis that reversing hypoperfusion could ameliorate long-term prognosis.

Several groups have proposed endovascular and hybrid surgery therapies for symptomatic CCTO.^11–13 A meta-analysis investigating the feasibility and safety of these approaches showed that successful recanalization is achieved in around 70% of cases, with a 5% morbidity and a 2% mortality rate. ¹⁴ Moreover, early ischemic events were encountered in 9% of patients and hemorrhagic events in 5%. This compares favorably to a peri-operative stroke rate of 15% in the COSS trial and 12% in the international EC-IC Bypass Study.^7,9 However, most of these series are retrospective single-center studies, probably subject to selection bias and underreporting of complications. Minimizing the acute complication rate is imperative if we are to realize any potential long-term benefit from the procedure. ⁷ While reporting lower acute complication rates, long-term patency is still a problem as endovascular reconstruction faces up to a 25% re-occlusion rate compared to a more than 95% patency rate for bypass surgery.^10,15

The long-term patient benefit derived from these procedures is still unknown. However, centers that perform endovascular recanalization of CCTO have perfected and improved their technique in recent years. Further pursuing randomization may be a necessary next step, but the caveat is to avoid repeating the COSS trial experience. For this reason, increasing the intervention’s periprocedural safety and reducing the constructs’ re-occlusion rates should be the focus of groups tackling CCTO. We aim to present a detailed, didactic, stepwise approach encompassing our center’s endovascular therapy management protocol of CCTO patients.

Patient selection

Endovascular recanalization of CCTO is an emerging procedure, but different articles report heterogeneous selection criteria. In general, operators have selected symptomatic patients with a combination of recurrent symptoms and demonstrated hypoperfusion measured with either single-photon emission computed tomography (SPECT), perfusion computed tomography (CT), or perfusion magnetic resistance imaging.^16–18 However, the procedure was also offered to asymptomatic patients demonstrating hypoperfusion. ¹⁹ As no clear benefit of the procedure exists by limiting the indication to symptomatic patients, proposing a 21-day delay after ischemic stroke may be reasonable, based on lessons learned from intracranial stenosis trials. Moreover, there are some hints that the procedure seems to have worse outcomes if performed as an emergency.^15,16,20,21

Our patient selection criteria were based on previously published reports and align with the typical criteria presented in the literature.^13,15 The usual presentation is a minor stroke or transient ischemic attack in the setting of a carotid occlusion with severe hypoperfusion. The 3-week antithrombotic preparation phase described below may lead to recanalizing some occlusions, as previously described.^6,21 Obtaining this kind of result may transform the procedure into a more limited and safer carotid stenting.^22,23 Furthermore, it may reduce the risk of intracranial emboli during the endovascular procedure.^6,21

Pre-treatment work-up

Careful identification of suitable patients for this complex endovascular procedure is mandatory. Ideally, pre-planning of the intervention should be based on a diagnostic digital substraction angiography showing collateral circulation, stump anatomy, and occlusion length, which are prognostic factors for treatment success and can guide patients’ expectations. ¹⁷ The success of the occlusion is higher if a proximal stump is present, which facilitates the finding of the true lumen. Collateral circulation enhances success if the distal internal carotid artery (ICA) is kept patent by ECA/ PcomA collaterals. Pre-visualization of the collateral status permits planning of the procedure if dual arterial access to obtain visualization of the re-entry zone is needed. ¹⁷ The extent of hypoperfusion should be proved before treatment by using CT or MR perfusion, which are correlated to SPECT imaging and more readily available in clinical practice. ²⁴

While there are no established cut-off values for hypoperfusion prompting the intervention, a volume of hypoperfused brain defined by Tmax >6 s of at least 40 ml was associated with a higher risk of recurrence in a monocentric retrospective study. ⁶ Patient selection should be based on clinical symptoms, while the area and extent of hypoperfusion are to be judged as a prognostic factor for potential future cognitive decline. Of note, there is a sub-group of patients who have symptoms without baseline perfusion abnormalities. In these patients, acetazolamide challenge perfusion imaging may demask normal cerebral blood flow impairment by reducing vascular resistance secondary to autoregulatory vasodilation. ²⁵ Finally, pre-intervention cognitive testing should be performed in all patients as this may identify otherwise “asymptomatic patients” given that cognitive functions are proven to improve after reestablishing adequate perfusion. ²⁶

Pre-treatment antithrombotic preparation

Once the decision to pursue endovascular treatment is taken, oral anticoagulation (usually apixaban 5 mg twice daily) and dual antiplatelet therapy (DAPT) with a 100 mg dose of acetylsalicylic acid associated with a P2Y12 inhibitor is started for up to 3 weeks before the procedure. Single antiplatelet therapy may be used initially with anticoagulation based on infarct size. This is usually ticagrelor 90 mg twice daily but prasugrel with dose adjustment can also be used (10 mg or 5 mg once daily). Clopidogrel 75 daily is also an option, but is generally avoided because of its variability of action. ²⁷ Loading doses are usually administered, depending on the context of the presentation. However, despite using antithrombotic drugs with low-resistance rates, ²⁸ a P2Y12 actual PRU cut-off value is systematically performed before the intervention, and a platelet inhibition rate of more than 60% is targeted. Symptomatic carotid occlusions are associated with microembolization in up to 40% of cases. ²⁹ Reducing the potential thrombotic burden with an aggressive and efficacious antithrombotic preparation may reduce the procedural embolic complication rate in cervical stenting procedures and facilitate initial recanalization. ³⁰ In our experience and according to the literature, this antithrombotic preparation phase may lead to recanalizing some occlusions that were probably acute.^6,21 This kind of result may transform the procedure into a more limited and safer carotid stenting procedure by reducing the need to dissect the occluded artery.^22,23

Cases requiring an emergency endovascular intervention are usually rare and associated with a worse prognosis. ¹⁵ However, in cases where a P2Y12 inhibitor was not administered or an optimal platelet inhibition rate of more than 60% before the procedure was not achieved, a bail-out intravenous cangrelor strategy (5 μg/kg bolus + 1 μg/kg/min of drug infusion and subsequent PRU testing at 10 min for protocol adjustment) is recommended. ³¹ The patient is kept under intravenous cangrelor up to 1 h after a 180 mg loading dose of ticagrelor can be administered. In these cases, a 250 mg intravenous loading dose of aspirin is administered on the operating table.

Procedure

Phase I. Intimal dissection

Femoral puncture is always performed using a 9-French (Fr) femoral short sheet under ultrasound guidance. While recent randomized trial evidence does not show a significant added safety of performing ultrasound-guided punctures, these data were mainly restricted to 6-Fr femoral sheets (around 80% of inclusions) and patients were not usually on aggressive antiplatelet treatment (less than 40% on clopidogrel and less than 10% on ticagrelor). ³² Previously published meta-analyses have shown that ultrasound-guided puncture is easy and faster.^32,33 Given the vital need to continue aggressive antithrombotic medication in these patients, any precaution should be taken to reduce puncture site complications.

Once femoral access is gained, a 9-Fr balloon guide catheter (BGC) (Merci, Stryker, Fremont, CA) is placed in the common carotid artery (CCA). An angiogram is performed to visualize the opacification of the proximal and distal ICA stump, and 3D imaging may help identify an entry point inside the proximal stump (Figure 1). In the event that the target distal ICA stump is not visualizable through external carotid artery collaterals, a 5-Fr radial access is usually used to catheterize the contralateral ICA or the dominant vertebral artery with an additional diagnostic catheter that offers intracranial control. This dual angiography technique is mandatory and has been shown in chronic coronary total occlusions to lead to better outcomes. ³⁴ The BGC is inflated in the CCA. Under roadmap guidance, a 125 cm 5-Fr vertebral catheter (Impress, Merit, UT, USA) is advanced inside the proximal stump to support the dissection-entry point. Once in position, the dissection is started using a 0.014″ guidewire (Asahi Chikai Black, Asahi Intecc. Co. Ltd, Japan) and a 0.017″ 150 cm microcatheter (Headway, Microvention, Austin, Tustin, CA).

Figure 1.

Endovascular recanalization of a short-cervical lesion. (a) Significant perfusion abnormalities on the side of the occlusion; (b) DSA showing the initial stump. (c) Intimal dissection phase, notice the orientation of the vertebral catheter in the direction of the stump and the negative roadmap of the microcatheter, proving early re-entry in the petrous segment, the BGC is inflated. (d) Initial predilation with inflated BGC; (e) final construct with three subsequent Wallstents, no need for further coverage as there was normal, undissected lumen; (f) normalization of perfusion abnormalities immediately after the procedure; (g) 1 month OCT shows good endothelialization of the implant across its entire length. DSA: digital substraction angiography; BGC: balloon guide catheter.

Phase II. Catheterization of the ICA lumen

The goal of the dissection phase is to reach the normal arterial lumen, which should always be visible on the roadmap after initiating the dissection. While blindly progressing through the supposed ICA trajectory and always inside the ICA fascial sheath, there are usually two options: (a) the wire becomes rapidly free and torquable and the previous ICA lumen is easily achieved; and (b) the lumen has been collapsed for a long time and the dissection continues through the wall of the ICA and re-entry is attempted at the level of the distal stump (usually cavernous segment). The normal ICA lumen may be more easily reached if the 5-Fr is used to orient the dissection in the anterior plane due to the tendency of carotid plaques to be located posteriorly. ³⁵

Advancement of the wire should be attempted with gentle torquing maneuvers and as far as possible without knuckling the guidewire. While the knuckling technique has been proven helpful in crossing chronic total occlusions, it should be used as a last resort as it extends the dissection flap. ³⁶ When the wire seems to torque freely in the true lumen, the best control is to remove the guidewire and try to aspirate with a 3 ml syringe once the microcatheter is in the desired position. Good backflow usually means that the true lumen has been catheterized and a small 1–2 ml puff can be performed via the microcatheter to control the position. However, frequent puffs are discouraged because they enhance the dissection flap. There are several caveats to this phase. The 5-Fr vertebral catheter may dissect and create a channel up to the petrous segment to gain support. This technique is helpful, especially if the microcatheter guidewire stumbles in the same false lumen.

The most dangerous part is the re-entry phase. While coming from the artery wall towards the true lumen, one can extend the dissection flap to cover the collateral circulation (frequently given by the ophthalmic artery). Avoiding injections when one is very near to this collateral branch, as well as careful re-entry, is critical and procedures may be abandoned without incurring any deficits to the patient if one considers that closing down the collateral circulation while attempting re-entry is highly likely. Re-entry is usually sought by slowly torquing and advancing the guidewire towards the distal stump. It is common to change the guidewire due to damage during the dissection phase, and two or three guidewires may be required in some cases. Due to different tip strengths and constructions, different wires may be suitable for different parts of the procedure. While there is an abundant cardiologic literature on the subject, dedicated neurointerventional devices are lacking, and operator experience and habit remain the critical determinants of guidewire choice. ³⁷ Sometimes reversing the guidewire and gently pushing the back tip toward the wall with the microcatheter creates a javelin-like effect and may facilitate re-entry by increasing penetrating power. However, this technique requires careful control since the back ends of 0.014″ guidewires are barely visible under fluoroscopy.

Phase III. Anterograde reconstruction

After re-entry and true lumen confirmation, the middle cerebral artery (MCA) is catheterized with the 0.017″ microcatheter, and the 2 meters 0.014″ guidewire is retrieved. Due to a larger caliber, a J-shaped formed 300 cm rigid exchange guidewire is inserted and advanced toward the distal end of the dominant M2 division (Luge, Boston Scientific, Natick, MA). ³⁸ In short occlusions involving the cervical and petrous segment, it is unnecessary to catheterize the M2 segment for adequate support and the guidewire tip can be retained in the distal ICA. At this point, the microcatheter and 5-Fr vertebral catheter are gently retrieved under continuous fluoroscopy in order not to lose the true lumen and to prevent inadvertent guidewire perforations. Two technical considerations merit discussion at this phase. First, the exchange operator usually has to concentrate only on the exchange. The second operator will keep constant pressure on the microcatheter to limit rapid retrieval, which may inadvertently lead to the displacement of the guidewire tip. Second, using 150 cm length microcatheters is essential. Even with these, the valve usually must be removed. It may be necessary to finish the exchange maneuver by gently pushing saline with a 5 ml syringe while retrieving the last few cm.

Once the exchange maneuver is performed, the procedure continues under a continuous flow arrest. The BGC is inflated in the CCA and the flush is cut-off to perform the procedure in constant flow arrest and reversal. ³⁹ Initially, the exchange length guidewire is used to mount 4 × 20 rapid exchange non-compliant balloons (Ultrasoft, Boston Scientific, Natick, MA) and perform several anterograde angioplasties in the cervical ICA. They are followed by placing several long self-expanding carotid stents (Wallstent) with a generous overlap (ideally more than 10–15 mm) up to the horizontal part of the petrous segment. Once the stents are in place, a pump aspiration is performed on the BGC and after the balloon is deflated and advanced inside the first stent, where it is reinflated, and the three-way valve is opened to monitor for retrograde flow.

Theoretically, with the petrous-cavernous segment still occluded, there is no retrograde or anterograde flow through the stents. There are two main reasons for starting with a small anterograde reconstruction: (a) it assures that once the retrograde reconstruction is finished and the artery is open, we prevent the chance that an accidental loss of guidewire position in the true lumen due to a cervical loop, may prevent proximal reconstruction; and (b) if for any reason there is a problem during the procedure, it can be stopped without exposing the patient to an embolic risk as the distal part of the chronic occlusion is still closed.

Phase IV. Retrograde reconstruction

With the BGS inflated in the proximally reconstructed stents, a second retrograde angioplasty phase with a 4 × 20 mm non-compliant balloon (Ultrasoft) descending from the true lumen to the cervical stents is initiated. Until this point, no angiography runs were made and the opening of the ICA can be assessed by an adequate black flow through the three-way valve. For the retrograde reconstruction phase, it is helpful to advance the BGS to the petrous segment. This maneuver can be done using the inflated angioplasty balloon as an anchor, which also prevents potential emboli during the quick deflation of the balloon guide. Once in position, the balloon is reinflated. If adequate backflow is present, a contralateral run through a diagnostic catheter while aspirating through a pump on the balloon guide will retrogradely opacify the ICA without risking embolic events. If the patient's anatomy does not show too much tortuosity, further reconstructions with smaller diameter braided carotid stents (5 mm) may be tempted (Wallstent); if not, one or two long cobalt-chromium flow-diverting stents (Pipeline, Medtronic, Irvine, CA) are used to reconstruct the distal ICA retrogradely (Figure 2).

Figure 2.

Long-lesion recanalization with four subsequent Wallstents in a 60-year-old patient. (a) Right ICA injection showing no-stump and retrograde filling of the cavernous segment. (b) Initial dissection with 0.0140″ guidewire and microcatheter through 5-Fr Vertebral. (c) Exchange wire placed in the angular artery. (f) Successful recanalization with four subsequent Wallstents (Boston Scientific), notice significant overlap for the distal proximal two stents with less overlap for the high-cervical and petrous stents. (d,e) OCT follow-up imaging at 1 month shows good endothelialization at the level of proximal stents with small decoaptation of the stents and intraluminal struts at the level of the cervical stents.

Regardless of what kind of stents is used, the distal landing zone of the first deployed stent should be in an undissected normal lumen. The first carotid stents can usually be advanced on their own or, if difficulties are encountered, a quadriaxial technique using a 150 cm 0.021″ microcatheter, a 130 cm 4Fr distal access catheter (Stryker) and a robust 115 cm 6 Fr distal access catheter (Fargomax, Balt, Montmorency, France, or Navien, Medtronic, Irvine, US) may be used. Thus, once the 6 Fr distal access catheter is in place, the microcatheter and the 4 Fr are removed, the stent is advanced inside the 6 Fr catheter, and an unsheath maneuver is used to perform the deployment in the desired location. This technique is useful for bridging the petrous segment as it prevents the dislocation of stents during forceful catheterization maneuvers. Adequate overlap between stents should be sought. If the reconstruction is performed using flow-diverter stents, a 0.027″ microcatheter is advanced over the 3-meter 0.014″ guidewire in the MCA. The deployment is then started in the desired location, assuring a good overlap in the petrous segment.

The full metal jacket is post-dilated using 6 × 40 mm non-compliant balloons (Ultrasoft). During this maneuver, the BGS is deflated, pulled back in the CCA and reinflated, while an inflated non-compliant balloon protects the flow. The construct is then post-dilated, with careful inflations at the apposition zones of stents, and a last aspiration on the inflated BGC is performed. The BGS is deflated and a first anterograde run is completed. A vaso-CT of the cervical construct finishes the procedure to ensure optimal overlap and apposition between stents. The femoral puncture site is closed using a closure device with care to prevent groin-hematoma complications, which could lead to the stopping of antithrombotic medication and result in essential complications.

Reducing periprocedural hemorrhagic complication rates

Cerebral hyperperfusion syndrome with associated reperfusion hemorrhage is one of the most feared complications after CCTO treatment. It can be observed in up to 5% of patients compared to 1% in carotid stenting cohorts.^14,15,40 Several strategies should be put in place to prevent intracerebral hemorrhage complications. First, as previously mentioned, adequate catheterization of the dominant MCA under roadmap guidance and J-shaped careful exchanges are mandatory, given the high antithrombotic load of patients. Second, the prevention of post-procedural reperfusion hemorrhage is imperative. After successful reconstruction, heparin should be reversed and patients should wake up with careful blood pressure control. The optimal systolic blood pressure target should be <120 mmHg. Connecting them to intravenous drips of fast-acting antihypertensive agents such as nicardipine or urapidil may help control blood pressure during this phase. Furthermore, these patients should be closely followed for at least 7 days in a dedicated ward with trained nurses to follow strict blood pressure targets and protocols.

Short- and long-term reductions of re-occlusion rates

The goal of future developments in endovascular CCTO reconstruction is to facilitate the procedure's short- and long-term viability. To overcome the massive thrombogenicity and biological reaction created by implanting a high metal load in theoretically non-functional endothelium, we advocate for a short-term 1-month course of triple therapy with a non-vitamin K oral anticoagulant associated with DAPT.

After this initial medical course, we have adopted a personalized approach to long-term antiplatelet therapy by controlling the viability of our implant, at least in the cervical segment. OCT studies performed in CCTO that were endovascular recanalized showed delayed endothelial stent coverage and a high incidence of malposition associated with stent thrombosis. Both require more prolonged DAPT administration to prevent stent thrombosis.^41,42 Therefore, we control the endothelization of our implant with OCT using a previously described technique and guide antithrombotic therapy based on OCT findings. ⁴³ Moreover, we may also add carotid stents to cover an area of loss of overlap between two stents or intraluminal protrusion of stent struts.

Choosing the right materials

Dedicated materials for CCTO procedures are lacking. Most groups reporting series of CCTO reconstruction have favored self-expanding carotid stents for cervical segment reconstruction and coronary stents for intracranial segment reconstruction. ⁴⁴ Our group favored the initial use of closed-cell rigid self-expandable carotid stents for cervical segment reconstruction (Xact, Abbot, Park, IL) and coronary stents for the intracranial reconstruction of the occluded artery.^15,45 However, more recently, we have switched to a braided stent design across the whole length of the occluded vessel and have previously described several benefits of this approach. ¹⁵ There are two important caveats for obtaining good clinical results with endovascular CCTO recanalization after immediate treatment success: (a) long-term viability of the implant; and (b) easily recrossing the implant when needed.

Long-term viability of the implant may be threatened by cervical stent problems such as stent fracture and loss of overlap after the vessel has distended to normal diameter (Figure 3). Stent fractures were reported in up to 5% of single carotid artery stents implanted in patients with symptomatic carotid stenosis. These are considered risk factors for vessel occlusion and are usually regarded as benign.^46,47 Nevertheless, when long lesions are reconstructed with several overlapping stents, the risk of stent fracture probably increases due to angulation and torsional neck movements with a fixed metal construct inside the immobile petrous part of the carotid artery. In single carotid stenting procedures, the unstented segments accommodate increased flexion and torsion, leading to friction at both ends of the stent. ⁴⁸ This strain caused by stents struts in a biologically active vessel responded by increased collagen deposition, marked destruction of elastin, and persistent inflammation, which likely contribute to vessel lumen reduction and long-term high occlusion rates. ⁴⁹ To address this problem, we switched to a more flexible braided, self-expandable carotid stent design (Wallstent). However, the high torsion resistance of this stent creates problems with the initial overlap of the stents, when no generous overlap between the cervical stents is accomplished. ⁵⁰

Figure 3.

(a) DSA shows a long-term complication of restenosis in a recanalized CCTO; (b) magnified unsubstracted imaging shows stent fracture at the occlusion site; (c) OCT shows intimal hyperplasia due to stents struts malaposition. The patient was successfully treated with the implantation of another stent and angioplasty at this level. CCTO: Chronic carotid total occlusion; DSA: digital substraction angiography.

When the artery is not stenosed by an intracranial plaque, we prefer reconstructing the artery with braided chrome-cobalt flow-diverter stents intracranialy. The rationale for this relies in the easy recrossing and the more anatomical reconstruction of the artery, which may facilitate adequate endothelization and thus lower overall thrombogenicity despite higher metal content. Moreover, theoretically, when not fighting severe atherosclerotic plaque, remodeled vessels with braided intracranial stents may be sufficient.^15,51,52 Previously, frequent occlusions due to inadequate bridging in the petrous cervical segment have been described, either due to coronary stent or open-cell self-expanding carotid stent implantation. ¹⁵ Inward strut movement and buckling of coronary stents combined with chronic recoil and lower conformability may lead to vessel lumen narrowing due to neointimal hyperplasia and associated increased thrombogenicity.^53,54 This might explain the frequent occlusions observed in these patients. It is reasonable to cover all the dissected and diseased parts of re-opened artery to permit endothelization from metabolically normal distal to metabolically normal proximal endothelium.

Limitations and further directions

Despite all of the recently published studies of CCTO endovascular therapy, it remains an unproven procedure performed as a bail-out in experienced centers. Due to the aforementioned technical aspects, operator experience and procedure standardization are mandatory before future randomized trials can be initiated. Without careful patient selection, identifying those most likely to benefit, and without a standard technique, any randomized controlled trial may be prone to repeat the COSS experience.

This review summarizes a reproducible technique in our center by different operators. A minimum number of future changes might be expected, such as using sirolimus-coated balloons (to reduce subsequent hyperplasia) and using more flexible braided stents in the cervical segment. Adapting existing materials from cardiology and peripheral vascular interventions to the neurovascular field might reduce the incidence of re-occlusions and provide more chances of obtaining long-lasting stroke rate reductions.

Conclusion

Experience with endovascular recanalization of CCTO has rendered the intervention feasible. However, to progress to a phase of randomized controlled trials, the procedure must be standardized, and a focus should be placed on dramatically reducing periprocedural complications and increasing the longevity of the construct to prevent a repeat of the tale of failed bypass surgery trials.