Robert W. Regenhardt, MD, PhD

International Stroke Conference 2021
March 17–19, 2021
Session: Challenging EVT Decision Making: When, Where, and Who to Treat (Debate) (33, On Demand)

The session “Challenging EVT Decision Making: When, Where, and Who to Treat” (Debate) highlights some of the most difficult management decisions regarding EVT.

Dr. Sandra Narayanan built the case “Low NIHSS proximal occlusions should undergo thrombectomy.” She started by reviewing the magnitude of the question. An LVO is present in 18% of patients with NIHSS 0-4 and 39% of those with NIHSS 5-8. Furthermore, 15% of LVO stroke patients have minor symptoms. Deterioration can happen in early or delayed fashion; about 40% deteriorate early. Current guidelines suggest that treating patients with low NIHSS is reasonable. Indeed, several studies show a benefit. The Grady experience (JNIS 2017; 9:917-921) described 32 patients with NIHSS<6. Analyses of this cohort, while small, suggested a benefit of EVT. 22 were treated with medical management, of which 9 declined requiring EVT. The median time from arrival to deterioration was 5.2 hours. Subsequently, a larger study of 6 CSCs (Stroke 2018;49: 2391-2397) described 300 patients with NIHSS<6; 11.3% of those treated with medical management later declined. At 90 days, mRS 0-2 was observed in 84% of those treated with EVT, 70% of those with medical management, and 55% of those who underwent rescue EVT. Those who are allowed to deteriorate tend to have worse outcomes. The risks versus benefits should be carefully weighed up front because waiting more than 3 hours appears to impact outcomes. There is growing data that patients at risk for decline can be selected by collaterals, orthostatic challenges, perfusion imaging, and NIHSS eloquence/disability. Three randomized controlled trials are forthcoming: ENDOLOW, IN EXTREMIS, and TEMPO 2.

“Low NIHSS proximal occlusions should NOT undergo thrombectomy” was argued by Dr. Pooja Khatri. She began by emphasizing that we have no randomized data yet. In all the EVT trials to date, only 14 patients had NIHSS<6 (10 in MR CLEAN and 4 in EXTEND IA). However, there has been a recent propensity score matched analysis of the Swiss stroke registry (Neurology 2019;93:1618-1636). 108 patients were included in each group. Those in the EVT-treated group had a nonsignificant trend for worse outcomes. Another meta-analysis (JAMA Neurology 2020) showed no clear difference in outcomes but more ICH associated with EVT (5.6% vs 0.9%). A recent review (JNIS 2021) summarized several recent studies. Comparing EVT to best medical management, only 4/10 studies showed benefit from EVT and 6/10 showed more ICH with EVT. Comparing EVT to intravenous thrombolysis, 0/4 studies showed benefit from EVT. Khatri cautioned us to doubt non-randomized evidence due to the high risk of selection bias. She cited that among these patients with low NIHSS, between 60-80% won’t decline (Stroke 2020). Patients with low NIHSS likely represent a unique biology, which may be related to their collateral status. A study from 6 French centers (Stroke 2019) showed better collaterals were associated with more post thrombolysis recanalization before thrombectomy. Early recanalization was seen in 12% of those with poor/moderate collaterals, 15% of good collaterals, and 45% of excellent collaterals. Khatri added that these patients with good collaterals often have underlying ICAD, which is technically more challenging to treat with higher rates of re-occlusion. A recent survey showed half would treat patients with low NIHSS and half would monitor them (JSCVD 2020), further highlighting our present state of clinical equipoise.

Dr. Johanna Ospel made the point “Tandem occlusions: stent at the end or not at all.” She opened with an acknowledgement that we lack high-level evidence. However, we know tandem occlusions have worse outcomes compared to isolated intracranial occlusions. Her position was argued by three points: “time is brain,” “knowledge is power,” and “less is more.” Regarding “time is brain,” the intracranial occlusion should be prioritized. It requires an average of 15 minutes to deploy a carotid stent;  since 1.9 million neurons die per minute in the average LVO, that equates to around 29 million lost neurons. Regarding “knowledge is power,” Ospel contended that little is known at the start of the case. Going after the intracranial occlusion first provides key information. Will you be able to open the intracranial occlusion? This is important to avoid futile stent placement, which necessitates DAPT. What is the underlying etiology of the tandem occlusion? Opening the intracranial occlusion first allows more time to get information in the acute setting. Regarding “less it more,” Ospel emphasized that many times the carotid stenosis improves just by crossing it with a guide catheter or by using a balloon guide catheter inflated across the lesion. It may be possible to get by with angioplasty to avoid the DAPT required for stenting, which may be associated with more ICH in the acute setting. She asserted that it is always a good idea to avoid the permanent placement of materials in head unless truly needed. There are also reports of carotid stents causing balloon guide catheters, often used during intracranial EVT, to rupture, further supporting that stenting should be performed last or not at all. Poppe et al. (AJNR) summarized factors that favor stenting vs not stenting. It may be reasonable to avoid stenting if: there is a large infarct core, there is poor reperfusion, there are good collaterals, there is no active carotid re-occlusion, intravenous thrombolytic was used, the patient has a high bleeding risk, or if there is a known indication for long-term anticoagulation. There was a recent meta-analysis by Wilson et al. (JNIS) that suggested stenting tandem lesions may not be associated with benefit.

The counter point was argued by Dr. Johanna Fifi, “When tandem occlusions exist: Stent first! EVT after.” She highlighted that current guidelines suggest treatment of the cervical ICA may be reasonable. While there are no definitive trials delineating the optimum treatment approach, there is mounting retrospective data that supports stenting may be beneficial. A post hoc analysis of the ESCAPE trial included 54 patients with tandem lesions, and 30 were randomized to treatment (JNIS 2018; 10:429-433). Of these, 17 underwent carotid stenting (10 before and 7 after intracranial EVT). With stenting, there was better reperfusion (71% vs 46%), better 90-day outcomes (65% vs 54%), lower mortality (6% vs 15%), and lower rates of sICH (0% vs 8%). Furthermore, the TITAN retrospective analysis showed stenting led to higher rates of reperfusion, better 90-day mRS, less mortality, and similar ICH rates (Frontiers Neurology 2019; 10:206). Additional data in support of stenting was shown in an analysis of the STRATIS prospective registry (Stroke 2019; 50:428-433); 147 (15%) patients had tandem lesions, and 80% underwent acute stenting. There were similar rates of reperfusion, ICH, and mortality, but there were better 90-day mRS 0-2 rates among those stented. Fifi concluded that ample data exist in support of acute stenting, but less data are available about its timing (i.e., before or after intracranial EVT). Some have postulated that a proximal-to-distal stent first approach is safer and more efficient since the intracranial occlusion can offer protection from distal emboli during the proximal carotid stent placement. Furthermore, after opening the cervical carotid, anterograde flow can improve visualization and access to the distal intracranial occlusion. One retrospective single center study of 16 patients reported a proximal-to-distal approach had good outcomes in 50%, a higher rate than has been reported for patients with tandem lesions (JNIS 2015; 7:164-169). Another study utilizing a proximal-to-distal approach reported 54% had good outcomes (Clin Neuroradiol 2013; 23:207-215). Similarly, another reported good outcomes in 49% with no difference in time to recanalization (Neurosurgical Focus 2017 42:4).

Dr. Tudor Jovin, the final speaker, discussed “Intracranial dissections in acute stroke interventions.” He began by reminding us that intracranial dissections are very rare compared to extracranial dissections. They have an incidence of 3/100,000/year and are associated with pediatric and Asian populations. Intracranial dissection occurs most commonly in the anterior circulation among pediatric patients, while adults experience more posterior circulation intracranial dissections. It is not clear if the same risk factors for extracranial dissections apply to intracranial dissections. The pathophysiology involves interruption of the vessel intima with creation of intimal flap. Thrombus can form at the site of the flap and result in distal embolization. There can also be perforator occlusion at the site of injury. Furthermore, with the formation of mural hematoma and lumen narrowing, there can be flow limitation or occlusion. Pseudoaneurysms can also develop, leading to subarachnoid hemorrhage. On parenchymal imaging, combined evidence of ischemia and subarachnoid hemorrhage can be suggestive. In the case of stroke interventions, a key question to ask is whether the intracranial dissection was iatrogenic. This can be very difficult to determine, but there are some suggestions of intracranial dissections on noninvasive imaging. Features that should raise suspicion for intracranial dissection include intramural hematoma, irregular dilatation, double lumen, string of pearls, intimal flap, and any rapid change in morphology over time. Patient demographics are also helpful as dissections are more likely in younger patients without atherosclerosis risk factors.

Given the subtle features and difficult a priori recognition, Jovin suggested intracranial dissections are likely under diagnosed. After EVT, it can be especially challenging to know if the dissection was the initial etiology of the occlusion or iatrogenic related to treatment. One case series from France described 6 features associated with intracranial dissection (JNIS 2018). They described intracranial dissections in 13/391 patients (3%) who underwent EVT. Half of those were extension from extracranial dissections, and 62% had mRS 0-2 at 90 days. To diagnose intracranial dissections, they required 4/6 features: flame shape, mural hematoma, transverse or longitudinal flap, tubular or fusiform enlargement, normalization of caliber after stenting, and lack of thrombus after EVT. However, Jovin warned flames can be deceiving. Many times, flame sign can be seen in relation to slow flow from distal occlusions rather than true dissections.

Regarding treatment for intracranial dissections, there is growing evidence that stenting can benefit these patients. A small French study showed that performing only standard EVT had a high rate of residual stenosis (7/7 patients), re-occlusion, and repeat rescue EVT (4/7). However, after stenting there was a decreased rate of residual stenosis (2/7). Furthermore, Jovin highlighted that EVT can be difficult given the need to find the true lumen. He presented a case example in which retrograde access across the circle of Willis was utilized, showing sometimes creative approaches are necessary. Furthermore, Jovin recommended against using a balloon mounted stent, stating that self-expanding stents are adequate for intracranial dissections. Finally, he reminded us that the CADISS trial supports either antiplatelet or anticoagulant medical therapy as appropriate for dissections, so DAPT is very reasonable in the setting of stenting.