Mechanical thrombectomy in MCA-mainstem occlusion in patients with low NIHSS scores

Abstract

Introduction

While effectiveness of mechanical thrombectomy (MTE) in M1 segment occlusion is approved for patients with severe impairment, there is a lack of evidence for a potential benefit of MTE in patients with minor to moderate symptoms. The purpose of this study was to evaluate neurological outcome and occurrence of periprocedural complications after MTE in patients with low National Institutes of Health Stroke Scale (NIHSS) scores.

Materials and methods

A retrospective analysis of 1081 consecutive patients with anterior circulation ischemic stroke due to M1 occlusion detected by computed tomography angiography and treated with MTE at our hospital between February 2012 and November 2017 was performed. NIHSS, Barthel Index (BI) and modified Rankin Scale (mRS) scores between admission and discharge were compared with paired Mann–Whitney test, and recanalization rate and complications were assessed in patients with NIHSS ≤ 5 at admission.

Results

Thirty patients were included with a median NIHSS score of 4. NIHSS score (median: 4 vs. 1; p < 0.001), BI (median: 43 vs. 80; p < 0.001) and mRS (2 vs. 1; p < 0.001) showed significant improvement from admission to discharge after MTE. Recanalization rate was Thrombolysis in Cerebral Infarction (TICI) 2b to 3 in 29 of 30 patients (96.7%). One case of an intracerebral reperfusion hematoma (ECASS: PH2) required surgical treatment.

Conclusion

MTE might lead to a significantly improved clinical outcome also for patients with low NIHSS score due to M1 segment occlusion. Periprocedural complications appeared infrequently. These results encourage further evaluation of the benefit-risk profile of MTE compared to standard treatment in patients with low NIHSS scores in future randomized trials.

Introduction

Patient populations of previously published larger randomized trials, proving the efficacy of mechanical thrombectomy (MTE) in large-vessel occlusion (LVO) of the anterior circulation, predominantly included patients with severe stroke symptoms. Although the study protocol of the Multicenter Randomized Clinical Trial of Endovascular Treatment for Acute Ischemic Stroke in the Netherlands (MR CLEAN) and the Extending the Time for Thrombolysis in Emergency Neurological Deficits-Intra-Arterial (EXTEND-IA) study also involved patients with minor neurologic deficit (National Institutes of Health Stroke Scale (NIHSS) 2–42 MR CLEAN, NIHSS 0–42 EXTEND-IA), the median NIHSS score for patients who underwent acute endovascular thrombectomy was approximately 15 to 17 in all major stroke trials.^1–5 There is a paucity of data from randomized clinical trials on the efficacy and safety of MTE in patients with lower NIHSS as a main study population. However, there is no clear definition of minor or moderate stroke. Also in case of low NIHSS scores in underlying LVO, patients can have a severe functional deficit like aphasia, hemianopia or monoplegia.⁶ On the contrary many patients with low NIHSS scores have relatively mild clinical symptoms, and thus the question arises whether the risk of an endovascular procedure is justified in these patients.

There are some considerations as to why MTE in patients with low NIHSS scores might be useful. Despite a low NIHSS score, an LVO of the middle cerebral artery (MCA) might be found, with a probability of recanalization by intravenous thrombolysis (IVT) of < 1% when thrombotic clot length exceeds 8 mm.⁷ Nine percent to 10% of patients with initially minor stroke may be affected by early progression of symptoms and infarct growth in association with LVO.^8–10 In a retrospective analysis of 7621 patients with NIHSS score ≤ 5 treated with IVT, absence of recovered independent ambulatory ability was observed in only 30.3% at discharge.¹¹ However, procedural risk of MTE in low-NIHSS patients has to be critically balanced against the benefit, because the advantage of MTE compared to pharmacological treatment has not yet been proved. Based on the major MTE trials, the American Heart Association guidelines provided level 1a evidence for MTE for patients with NIHSS scores of 6 or more, so that our selected NIHSS range ≤ 5 includes a patient collective about whom less is known. Some current studies already reported good clinical outcome after MTE in patients with mild symptoms harboring LVO in the anterior circulation.^12,13 In this single-center study, we intend to evaluate whether these results can be confirmed in patients with M1 occlusion and low baseline NIHSS scores at our institution.

Material and methods

Patient selection

We retrospectively assessed all patients with an NIHSS score ≤ 5 who were treated by MTE between February 2012 and November 2017 at the University Hospital Duesseldorf. Data evaluation was approved by the local ethics committee. For each patient, neurological examination was performed at admission by the attending neurologist in the emergency department, including detailed assessment of NIHSS, modified Rankin Scale (mRS) and Barthel Index (BI) scores. Standardized stroke imaging at our institution included non-contrast-enhanced cranial computed tomography (NECCT), CT-perfusion (CTP) and CT angiography (CTA). In admitted patients who received stroke imaging in external hospitals prior to admission, CTP imaging was absent in some cases. In patients with unknown symptom onset, magnetic resonance (MR) stroke imaging including diffusion-weighted imaging (DWI), fluid-attenuation inversion recovery, and susceptibility-weighted imaging sequences were performed. Extensive early ischemic signs or hemorrhage could be excluded in pre-interventional imaging by CT- and DWI-Alberta Stroke Program Early CT score (ASPECTS) assessment.

Eligibility for MTE in patients with acute M1 occlusion established in CTA or time-of-flight-magnetic resonance angiography (TOF-MRA) was determined individually for each patient in consensus between neurologists and neurointerventionalists, depending on estimated procedural risk, probability of recanalization by IVT, contraindication for IVT, social and medical pre-stroke conditions, and comorbidities. IVT was applied as bridging therapy according to national and international guidelines in eligible patients when treatment was initiated within 4.5 hours after symptom onset.

Endovascular treatment

Stent retriever thrombectomy (4.5 × 40 mm or 3.5 × 28 mm Aperio; Acandis, Pforzheim, Germany) combined with local thromboaspiration via a 5-Fr or 6-Fr intracranial intermediate catheter (Sofia 5-Fr; MicroVention or 6-Fr Navien Intracranial Support Catheter; formerly the ReFlex Intracranial Catheter; Covidien Vascular Therapies, Mansfield, MA, USA) at the MCA mainstem was performed. During stent retrieval, aspiration was applied through the intermediate catheter via a 20 ml syringe. Arterial punctures were closed by vascular closure devices (Angio-Seal VIP; St. Jude Medical, Minnetonka, MN, USA). MTE was performed under permanent monitoring of vital parameters and analgesia (local anesthesia in the groin, 1 g of metamizole as short infusion), and without general anesthesia or sedation in all cases. In case of vital deterioration, stand-by support by intensive care physicians was obtainable. After intervention, patients were admitted to the stroke unit of our hospital and treated according to in-house standard operating procedures. Patients who received acute stenting of extracranial arteries in addition to endovascular thrombectomy were treated with the gpIIb/IIIa antagonist tirofiban (1.250 mg bolus during intervention followed by a continuous infusion of 0.1 µg/kg body weight/minute) from time of acute stenting until a switch to oral aspirin and clopidogrel was performed, usually within 12–24 hours, with 12-hour overlap. Follow-up imaging CCT was routinely performed 6–24 hours after treatment. Postinterventional NIHSS, mRS and BI were assessed by the treating neurologist at discharge.

Imaging and procedural data collection

Time from symptom onset to stroke imaging, to start of angiography, and to recanalization were captured. The location of the occlusion and thrombus length were assessed on CTA or TOF-MRA images. Collateral supply of the occluded MCA territory from pre-interventional CTA scans was scored on the basis of the collateral grading system of Tan et al. on a scale of 0–3.¹⁴ A score of 0 corresponds to absent collateral supply, a score of 1 corresponds to collateral supply filling <50% but >0%, and a score of 2 corresponds to collateral supply filling >50% but <100% of the occluded MCA territory. One hundred percent collateral supply of the occluded MCA territory was scored as 3. Devices and medication used during the interventional procedures, procedural duration, and intra-procedural complications were evaluated from the treatment protocols. Angiographic outcome was graded by the modified Thrombolysis in Cerebral Infarction (TICI) scale.¹⁵ Pre- and post-treatment cerebral infarction, according to the ASPECTS and post-treatment intracerebral hemorrhage, according to the European Cooperative Acute Stroke Study classification (ECASS¹⁶), were assessed by routine follow-up imaging as described above.

Statistical analysis

Statistical analysis was performed by using R, version 3.4.3. Stroke severity as measured by NIHSS, clinical outcome according to mRS, and BI at admission and discharge were compared by paired Mann–Whitney test. Correlation between ASPECTS and time to successful recanalization (TTSR), and ASPECTS and collateral score was assessed by Spearman correlation. Results were considered statistically significant at a level of p < 0.05.

Results

A total of 1081 patients with M1 occlusion were treated by MTE between February 2012 and November 2017. Of these, 30 patients presented with a pre-therapeutic NIHSS score of 1 to 5. Thrombus location in all included patients is summarized in Table 1. Twenty-four patients received intravenous (i.v.) bridging lysis. All patients were treated by stent retriever thrombectomy under local aspiration. In five patients carotid artery stent implantation because of arteriosclerotic carotid artery stenosis for intracranial access was necessary. CTA-based collateral status assessment was available in 21 patients (score 0, n = 0; 1, n = 2; 2, n = 5; 3, n = 14). In nine patients, preinterventional cranial imaging was performed by MR and thus were not eligible for collateral score assessment.

Table 1.

Baseline characteristics, treatment details and clinical outcome of patients with low NIHSS scores due to M1-segment occlusion who received mechanical thrombectomy.

Participant sample n = 30
Sociodemographic characteristics
Age (years), mean ± SD	71.5 ± 11
Female	n = 17	(56.7%)
Coronary artery disease	n = 6	(20%)
Myocardial infarction	n = 2	(6.7%)
Atrial fibrillation	n = 11	(36.7%)
Arterial hypertension	n = 24	(80%)
PAOD	n = 0	(0%)
Diabetes mellitus	n = 3	(10%)
Dyslipidemia	n = 11	(36.7%)
COPD	n = 4	(13.4%)
Coumarin treatment	n = 2	(6.7%)
Previous stroke	n = 1	(3.3%)
Baseline ASPECTS median (IQR)	10	(10–10)
Treatment details
Admission NIHSS median (IQR)	4	(2–5)
Admission mRS median (IQR)	2	(2–3)
Admission BI median (IQR)	43	(25–75)
Occlusion location
- M1 isolated	n = 16	(53.4%)
- M1 + M2	n = 5	(16.7%)
- Cervical ICA + M1	n = 5	(16.7%)
- Carotid T (including M1)	n = 4	(13.4%)
Thrombus length (mm) median (range)	14	(5–22)
IV rtPA	n = 24	(80%)
IA rtPA	n = 15	(50%)
Collateral status median (IQR)	3	(2–3)
Median time (minutes) median (mean ± SD)
- from symptom onset to imaging	79	(146 ± 175)
- from symptom onset to groin puncture	203	(243 ± 164)
- to final recanalization M1	244	(300 ± 167)
Treatment and clinical outcome
TICI 2b–3	n = 29	(96.7%)
TICI 2a	n = 1	(3.3%)
TICI 2b	n = 9	(30%)
TICI 2c/TICI 3	n = 20	(66.7%)
Discharge NIHSS median (IQR)	1	(0–2)
Discharge mRS (0–2)	n = 28	(93.4%)
Discharge mRS median (IQR)	1	(1–2)
Discharge BI median (IQR)	80	(60–100)
Symptomatic intracranial  hemorrhage (ECASS:PH2)	n = 1	(3.3%)
Post-treatment ASPECTS median (IQR)	9	(7–10)

ASPECTS: Alberta Stroke Program Early CT score; BI: Barthel Index; COPD: chronic obstructive pulmonary disease; ECASS: European Cooperative Acute Stroke Study classification; IQR: interquartile range; IV: intravenous; mRS: modified Rankin Scale; NIHSS: National Institutes of Health Stroke Scale; PAOD: peripheral arterial occlusive disease; rtPA: recombinant tissue plasminogen activator; TICI: Thrombolysis in Cerebral Infarction.

Overall median thrombus length, measured by CTA, was 14 mm (range: 5 to 22 mm), and 14 mm for patients receiving i.v. bridging lysis as well.

A favorable recanalization rate (TICI 2b–3), was achieved in 29 patients, with TICI 2b in nine, TICI 2c in four, and TICI 3 in 16 patients. Embolization to other vascular territories was not observed. In the one patient with unsuccessful recanalization (TICI2a) of a right-sided M1 occlusion, a good collateralization could be confirmed by digital subtraction angiography. Despite a pre- and postinterventional ASPECTS of 5, this patient’s NIHSS improved from 5 to 0 and mRS from 3 to 1 at discharge. The patient with TICI2a was excluded from the following subgroup analysis because of poor statistical significance.

The median NIHSS score of the 30 included patients was 4 before MTE. At discharge, median NIHSS score changed from 4 to 1 (p < 0.001), median mRS decreased from 2 to 1 (p < 0.001), and median BI improved from 43 to 80 (p < 0.001). Patients were discharged or moved to another hospital after 2 to 25 days (median: 9). At that time clinical outcome according to the mRS scale in detail was 0: n = 5, 1: n = 13, 2: n = 10, 3: n = 1, 4: n = 0, 5: n = 1. In one patient (3.3%) a reperfusion hematoma (ECASS: PH2) of the basal ganglia occurred after M1 revascularization (TICI 3) and additional carotid stenting for severe carotid stenosis.

Small or confluent petechial infarction without space-occupying effect and without clinical deterioration was found in 5/30 patients (16.7%). No deaths occurred. Median ASPECTS of the 30 patients was 10 in the preinterventional NECCT imaging and 9 at the 24-hour follow-up imaging.

Subgroup analysis between TICI2b and TIC2c/3 patients regarding pre- and post-treatment ASPECTS differences was performed. Median pretreatment ASPECTS was 10 in both groups. Median post-treatment ASPECTS was 10 in TICI2b patients and 8 in TICI2c/3 patients without statistically significant differences in pre- to post-treatment ASPECTS shift between the groups (p = 0.23).

Changes between pre- and post-treatment ASPECTS did not correlate to TTSR (ρ = –0.12, p = 0.51; after removal of one outlier, ρ = –0.02, p = 0.89, Figure 1).

Figure 1.

Alberta Stroke Program Early CT score (ASPECTS) in relation to time to successful recanalization, ASPECTS_post-pre = differences between pre- and postinterventional ASPECTS.

The relation between ASPECTS difference and collateralization was assessed in a correlation analysis. There was a significant correlation between the range of pre- compared to post-treatment ASPECT scores and collateralization score, i.e. a higher collateral score resulted in a lower decrease of posttreatment ASPECTS compared to baseline ASPECTS (ρ = 0.53, p < 0.005, Figure 2).

Figure 2.

Alberta Stroke Program Early CT (ASPECT) score in relation to collateral score, ASPECTS_post-pre = difference between pre- and postinterventional ASPECT score.

Discussion

MTE is the standard of care for ischemic stroke patients with LVO of the anterior circulation in addition to IVT or as first-line therapy for patients who are not eligible for IVT.^20–22 Thrombectomy results in significantly better clinical outcomes compared to IVT only in patients with acute occlusion of the intracranial M1 segment of the MCA.^1–5 While MTE is considered a highly effective procedure with a low risk of complications for LVO in the anterior circulation, there are only a few non-randomized study series for MTE in patients initially presenting with mild stroke symptoms (Table 2).^{12,13,17,18,19}

Table 2.

Studies including patients with low NIHSS due to LVO of the anterior circulation.

Study	Description	NIHSS	TICI (2b–3)	Outcome	Complications
Bhogal et al. 201613	MTE n = 41	≤5	87.8%	90-day mRS (0–2): 75%	sICH: 4.9%
Pfaff et al. 201612	MTE n = 33	≤8	78.7%	90-day mRS (0–2): 63.6%	sICH: 6% Mortality: 9.1%
Urra et al. 201417	MTE vs. IVT (n = 78); MTE, n = 34 vs. IVT, n = 44	≤5	MTE, 91.2% vs. IVT, 63.4% (p = 0.006)	90-day mRS (0–1): MTE, 58.8% vs. IVT, 68.2% (p = 0.393)	sICH: MTE,11.8% vs. IVT, 0% (p = 0.003) Death: MTE,11.8% vs. IVT, 4.5% (p = 0.395)
Haussen et al. 201718	MTE vs. IVT (n = 32); MTE, n = 10 vs. IVT, n = 22	≤5	MTE, 100% IVT, NA	90-day mRS (0–2): MTE, 100% vs. IVT, 77% (p = 0.15)	sICH: n = 0 Mortality: MTE, 0% vs. IVT, 14% (p = 0.38)
Messer et al. 201719	MTE vs. IVT (n = 54); MTE-I, n = 8 vs. delayed MTE-END, n = 6 vs. IVT, n = 40	≤5	MTE-I, 75% vs. MTE-END, 100% (NA) IVT, NA	mRS (0–1): MTE-I, 75% vs. MTE-END, 33.3% vs. IVT, 55%	sICH: n = 0 Mortality: n = 3 (overall group)

LVO: large-vessel occlusion; MTE: mechanical thrombectomy; IVT: intravenous thrombolysis; MTE-I: MTE immediately; MTE-END: MTE early neurologic deterioration; sICH: symptomatic intracranial hemorrhage; mRS: modified Rankin Scale score; NIHSS: National Institutes of Health Stroke Scale; TICI: Thrombolysis in Cerebral Infarction; NS: not statistically significant; NA: not available.

Substantiating the encouraging results from these reports, we could confirm a favorable clinical outcome in patients with low NIHSS scores after MTE in the present study (mRS 0–2 93.4%) that are in accordance with these previous studies. Successful recanalization (modified TICI 2b–3) in 96.7%, no embolization to new territories, and symptomatic intracranial hemorrhage rate of 3.3%, in our study complied with the suggested threshold levels for recanalization and complication rates from the current guidelines.²³ To date there is little information about the effectiveness of endovascular treatment in patients with anterior LVO compared to IVT (Table 2) but there are some reasons why MTE in low-NIHSS score patients should be considered.^17–19 Previous studies showed that patients with mild stroke and LVO who did not receive acute recanalization therapy experienced a poor clinical outcome at 90 days.^24,25 This might be explained by the observation that thrombus length exceeding 8 mm has a limited potential of thrombolysis.^7,26 These findings are in accordance with our own results. In all 24 patients who received i.v. bridging lysis, a persistent thrombus (median thrombus length: 14 mm) was observed in subsequent catheter angiography. The above-mentioned studies and our own results demonstrate limited effectiveness of IVT in broad M1 segment thrombus. Also patients with mild clinical symptomatology are at risk for neurological deterioration because of a decrease in their collateralization status.¹⁸ For instance, there were two patients in our sample with M1 segment thrombus and good collateralization proven by CTA showing nearly complete regression of neurological symptoms after IVT, but who presented with neurological deterioration after several hours due to persistent thrombus. Both patients had successful recanalization and favorable clinical outcome.

Campbell et al. demonstrated that deterioration in collateral status between baseline and subsequent imaging was strongly associated with infarction volume growth in patients without recanalization.²⁷ Correlation between good collateral status and favorable clinical outcome has been shown in several clinical trials.^28–30 But good collateralization alone without recanalization might not prevent an unfavorable outcome. The role of successful recanalization in acute ischemic stroke patients with LVO was demonstrated in a study by Miteff et al. Among patients with good collaterals, favorable outcome was achieved in those who received MTE, whereas in patients without recanalization, only 38% had a favorable outcome.²⁸ Our findings that better collateralization scores were associated with less ASPECTS worsening between pre- and post-treatment ASPECTS supports the presumption that good collateralization is related to favorable treatment outcome. In the current study the extent of infarction assessed by differences between pre- and post-treatment ASPECTS was independent from TTSR. Although time is still obviously very important and the earlier recanalization is achieved the better the results, we have shown in our case series that time should not be the only factor that determines endovascular therapy decisions. Moreover, these findings let us assume that the so-called “time window” is not stationary. Instead it might be a dynamic period of time that is determined by several factors including the complex interplay between LVO, blood pressure, collateral status, and individual cerebral ischemic tolerance.

In our sample, MTE was often initiated after a prolonged period after symptom onset for different reasons. First, many patients were referred from external hospitals to our institution, which could delay initiating of endovascular treatment. Some patients showed fluctuating symptoms and were readmitted to the hospital after a prolonged time. Because of clinical deterioration after fluctuating symptoms, MTE and carotid stent implantation were initiated after a prolonged time period in one patient with a very poor outcome (mRS 5) because of a symptomatically reperfusion hematoma. Although procedural-related complications in MTE like reperfusion hematoma, angiographically apparent vessel perforation, symptomatic subarachnoid hemorrhage with an appearance of 0.5%–4.9%, and embolization to new territories with a reported incidence of 4.9%–8.6% are serious complications, this might be an acceptable complication rate compared to the expected stroke severity as a consequence of a persistent LVO.^1–5 On the contrary, although we assume impairment with a high probability in patients with moderate symptoms due to LVO, it is not evidenced that the natural course of disease or standard treatment would not result in a comparable outcome like endovascular treatment. Therefore, in patients with a low NIHSS score, a markedly lower procedural complication rate than recommended in guidelines should be targeted.

Sufficient stroke imaging might improve patient selection. CTA or MRA are required for proof of LVO and thrombus extent. If there is a flowed-around short-range thrombus, IVT has a high feasibility of recanalization.⁷ Evaluation of anatomic complexity of supra-aortic access is necessary to estimate risk of procedural complications. The collateral status may give indications to brain condition and outcome. Optimally, stroke imaging should include CT perfusion imaging to estimate penumbra and severity of perfusion restriction. An obvious limitation of this study is the lack of a control group with low-NIHSS score patients treated by standard of care with or without IVT.

This was a non-randomized, retrospective study with a small case series that naturally limits the statistical significance of results. Clinical follow-up was ascertained at a nine-day median, and there is a lack of long-term outcome. On the other hand, the good short-term results indicate the effectiveness of MTE in this trial. The patient number was too low for subgroup analysis to define valid lower NIHSS score thresholds for treatment decisions.

However, the limited current state of research for MTE in patients with mild to moderate stroke symptoms in LVO makes it challenging to determine if MTE as rescue therapy after neurological deterioration or IVT is the most beneficial therapy for this subgroup. The results from previous studies and the current study indicate a benefit of early MTE with regard to good clinical outcome without increased procedural risk. The need for randomized trials still remains, to select patients with mild stroke symptoms who mostly benefit from immediate MTE.

Conclusion

In patients presenting with low NIHSS score due to M1 occlusion, MTE could be performed safely, technically very successfully, and with a favorable outcome. Deterioration in M1 segment occlusion after i.v. lysis can be observed, despite low NIHSS score. Therefore, although data from randomized controlled trials are still lacking, early endovascular recanalization has to be taken into consideration in these patients.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.