Clinical and technical evidence

A literature search was carried out for this briefing in accordance with the interim process and methods statement. This briefing includes the most relevant or best available published evidence relating to the clinical effectiveness of the technologies. Further information about how the evidence for this briefing was selected is available on request by contacting mibs@nice.org.uk.

Published evidence

Eight studies reporting on the clinical effectiveness of the devices are summarised in this briefing. Many systematic reviews were identified and there was substantial overlap in the trials summarised. As a result, only the most recent, highest-quality systematic reviews (that is, systematic reviews with multiple reviewers, quality assessment of studies, searches in multiple databases and reporting according to PRISMA guidelines) with meta-analyses were selected. Randomised controlled trials (RCTs) were selected only if they were not included in any of the systematic reviews and if they were recent and focused on the devices listed in the technology section. There is also other observational and registry based data on mechanical thrombectomy (MT) devices that has not been included in the MIB.

One systematic review with a meta-analysis and 7 RCTs (with a total of 2,718 patients) are summarised in this briefing. The systematic review included RCTs using mainly stent retriever devices. Six of the included RCTs compared MT devices with medical management against medical management alone. Of these, 3 mainly used stent retriever devices, 1 used only the Trevo stent retriever, 1 mainly used aspiration devices, and 1 did not report the type of devices used. Notably, 3 of these RCTs, as well as 1 of the trials included in the systematic review by Bush et al. (2016), were stopped early because of evidence showing that MT with medical management gave better clinical outcomes than medical management alone. A further RCT was a direct comparison of stent retrievers against aspiration devices with a superiority trial design.

Table 4 summarises the clinical evidence and its strengths and limitations. The main clinical outcomes assessed in the included studies are outlined below.

Functional independence

Functional independence was a primary outcome in all studies and was measured using modified Rankin Scale (mRS) score at 90 days of follow-up. In general, MT devices used with medical management gave substantially better mRS scores than medical management alone. However, 1 study comparing aspiration devices plus intravenous thrombolysis with intravenous thrombolysis alone did not find a statistically significant difference in mRS scores at 90 days (although the study may have been underpowered because of enrolment being stopped early). All of the studies comparing stent retriever, or mainly stent retriever, devices with medical management showed statistically significant improvements in mRS scores at 90 days.

When stent retrievers and aspiration devices were compared, they performed similarly with no significant differences in mRS scores. However, it is important to note that the study comparing the 2 types of device (Lapergue et al. 2017) did not have a non-inferiority design and so results from this trial cannot be taken to show equivalence.

Successful revascularisation

Revascularisation, measured using modified Thrombolysis in Cerebral Infarction (mTICI) scores, was reported in 9 of the 10 clinical studies. Using MT devices led to successful revascularisation rates of around 70% or higher. No difference was found in the studies that looked at a direct comparison between stent retrievers and aspiration devices.

Mortality and adverse events

Overall, there was very little difference in reported mortality rates between the intervention and control arms of the included studies. However, 1 study that compared MT devices (predominantly stent retrievers) with medical management alone found a significantly higher risk of mortality in the control group (Albers et al. 2018). Another study that compared predominantly aspiration devices used together with thrombolysis with thrombolysis alone reported a mortality rate more than 2 times greater in the thrombolysis-only group (Mocco et al. 2017).

There was little difference in the risk of intracerebral haemorrhage adverse events, when comparing stent retrievers and aspiration catheters against medical management alone or one another.

Overall assessment of the evidence

Several prospective RCTs and systematic reviews were identified. Stent retrievers were compared with other treatments commonly used for acute ischaemic stroke, including medical management alone (mainly thrombolysis with alteplase) and MT using aspiration catheters. A number of these studies focused on the Solitaire and Trevo stent retriever devices. Several studies also included a number of other stent retriever devices used in a few of the included patients, although no evidence was identified that explicitly assessed other stent retriever devices included in this briefing.

The evidence base for aspiration catheters is less well developed. One RCT made a direct comparison between aspiration devices and stent retrievers, while another compared aspiration devices used together with thrombolysis against thrombolysis alone. The studies included a number of the aspiration devices listed in this briefing, including Penumbra reperfusion catheters (90% in ASTER aspiration arm and in 96% of Therapy IAT arm) and a more limited number of Sofia and ARC devices.

A limitation of all of these studies was that, because of the nature of the intervention, participants and users could not be blinded and this presents some potential risk of bias. However, this risk was judged to be relatively low.

A large number of economic studies were identified and all compared MT with thrombolysis, which is a relevant comparison, and used appropriate sources for the data that informed the economic models (for example, pivotal phase III RCTs for effectiveness data). Only 2 of the studies were from the UK, which limits the generalisability of the evidence. Furthermore, the 3 studies that focused on the Solitaire device were funded by the company.

Table 4 Summary of selected clinical studies

Albers et al. (2018)

Study design and population

A randomised, open-label trial with blinded outcome assessment. Conducted in 38 centres in the US.

182 patients with proximal middle-cerebral-artery or internal-carotid-artery occlusion were randomised before the trial was stopped early because of the efficacy of the endovascular therapy.

Interventions

Intervention: Endovascular therapy plus standard medical therapy (n=92). Devices used were stent retrievers (n=74) and aspiration devices (n=25). Specific devices used were not specified.

A small number of patients also had cervical angioplasty and/or stent placement (n=13), intra-arterial thrombolysis (n=2), intracranial angioplasty or stent placement (n=3), or no endovascular therapy (n=2).

Comparator: Standard medical therapy alone, based on American Heart Association guidelines (n=90).

Results

Successful revascularisation (mTICI): TICI scores in the intervention arm: 10 (11%) patients had a TICI score of 0; 12 (13%) had a score of 2a; 52 (57%) had a score of 2b; 17 (19%) had a score of 3.

Functional independence (mRS): Distribution of mRS scores at 90 days was more favourable in the intervention arm: adjusted OR=3.36 (95% CI 1.96 to 5.77).

Percentage of patients who were functionally independent (mRS score of 0 to 2) at 90 days: 45% in the intervention arm compared to 17% in the control arm (risk ratio: 2.67, 95% CI 1.60 to 4.48).

Mortality: Mortality at 90 days: 14% in the intervention arm compared to 26% in the control arm.

ICH: Rates of S-ICH (increase ≥4 points on NIHSS) were 7% and 4%, with 5 and 2 resultant deaths, in the intervention and control arms respectively.

NIHSS Score: Reported in relation to specific adverse events (S‑ICH and thrombectomy-related complications).

ASPECTS Score: Not reported post-intervention.

Other

Adverse events: Serious adverse events were reported in 43% and 53% of patients in the intervention and control groups respectively. Parenchymal haematoma type 2 occurred in 9% and 3% of patients in the intervention and control groups respectively. Thrombectomy-related complications occurred in 2 patients, 1 of which led to neurological worsening (3‑point increase on NIHSS).

Imaging outcomes: Median growth of the volume of the infarct region was 23 ml in the intervention arm and 33 ml in the control arm. Reperfusion of >90% at 24 hours was 79% in the intervention arm and 18% in the control arm. Percentage of patients with complete recanalisation of the primary arterial occlusive lesion at 24 hours was 78% in the intervention group and 18% in the controls.

Strengths and limitations

Strengths: A multicentre RCT.

Limitations: none.

Bracard et al. (2016)

Study design and population

RCT in 26 centres in France.

412 patients with AIS and proximal cerebral artery occlusion were randomised before the trial was stopped early because results showed superiority of intravenous thrombolysis plus MT over thrombolysis alone.

Interventions

Intervention: Intravenous thrombolysis plus MT (n=204)

MT was performed using stent retrievers (n=108), including Solitaire, Trevo, pREset, and ReVive; aspiration alone (n=13), including the Penumbra System; or multiple systems (n=19).

59 patients randomised to the intervention did not have MT because of significant clinical improvement, partial/complete recanalisation, or violation of exclusion criteria.

Comparator: Intravenous thrombolysis alone (n=208)

8 (4%) patients randomised to the control arm eventually had MT.

Results

Successful revascularisation (mTICI): 95 (69%) patients in the intervention arm achieved an mTICI score of 2b or 3.

Functional independence (mRS): Percentage of patients with functional independence (mRS score 0 to 2) at 90 days: 53% (106/200) and 42% (85/202) in the intervention and control groups respectively (OR=1.55, 95% CI 1.05 to 2.30).

No difference was found between the 2 groups when mRS score at 90 days was considered as an ordinal variable in a regression model (OR=1.39, 95% CI 0.99 to 1.97).

Mortality: No significant differences were seen: 24 (12%) patients in the intervention arm and 27 (13%) patients in the control arm.

ICH: S‑ICH at 24 hours: 4 (2%) patients in the intervention arm and 3 (2%) patients in the control arm.

NIHSS Score: Median NIHSS score at discharge/7‑days post-intervention: 4 (IQR: 1 to 14) in the intervention arm; 8 (IQR: 2 to 16) in the control arm.

ASPECTS Score: Only reported at baseline and as a prognostic factor for acute stroke.

Other:

Adverse events: Thrombectomy-associated complications occurred in 9 (6%) of patients. No significant differences seen in the proportion of patients with adverse events at 3 months (31% and 27% in the intervention and control groups, respectively).

Activities of daily living: A higher proportion of patients in the intervention group had a Barthel index score of 95 to 100 at 3 months (61% versus 49%), although no significant difference was seen when the EQ-5D questionnaire was used.

Strengths and limitations

Strengths: A multicentre RCT with large number of patients.

Limitations: The vast majority of patients had anterior circulation strokes, so findings are only applicable in this group; protocol changes occurred during the trial; clinicians estimating mRS score were not blinded to treatment allocation.

Bush et al. (2016)

Study design and population

A systematic review and meta-analysis, incorporating results from 5 RCTs (NTR1804, NCT01778335, NCT01492725, NCT01657461 and NCT01692379) in various settings worldwide, including the UK, USA, Australia and several European countries.

1,287 patients with AIS. 634 patients (50.7%) were in the intervention group and 653 in the control group.

Interventions

Intervention: Stent retrievers used in 536 (85%) patients. Three of the included RCTs used Solitaire only, while 2 included a number of different devices (unnamed). 526 patients (83%) received IV t-PA in addition to thrombectomy.

Comparator: Medical management alone. 573 (88%) patients received IV t-PA.

Results

Successful revascularisation (mTICI): 373 patients (80.4%) in the intervention arm had mTICI scores of 2b or 3 and were reported as having achieved good reperfusion.

Functional independence (mRS): Patients in the intervention groups from all 5 RCTs had a pooled OR of 2.22 (95% CI 1.66 to 2.98) for improved mRS score at 90 days post-stroke compared to the control group (p<0.0001) with modest statistical heterogeneity across trials (I2=46.38%).

Mortality: Patients in the intervention groups from all 5 RCTs had a pooled OR=0.78 (95% CI 0.54 to 1.12) for mortality compared to the control group (p=0.1056) with no statistically significant heterogeneity across trials.

ICH: Patients in the intervention groups from all 5 RCTs had a pooled OR=1.19 (95% CI 0.69 to 2.05) for symptomatic ICH compared to the control group (p=0.5348) with no statistically significant heterogeneity across trials.

NIHSS Score: Patients in the intervention groups from 4 of the RCTs had a pooled OR=2.23 (95% CI 1.58 to 3.15) for NIHSS scores ≥17 at 90 days post-stroke compared to the control group (p<0.0001) with no statistically significant heterogeneity across trials.

ASPECTS Score: Patients in the intervention groups from 4 of the RCTs had a pooled OR=2.19 (95% CI 1.61 to 2.98) for high ASPECTS scores (≥8) at 90 days post-stroke compared to the control group (p<0.0001) with no statistically significant heterogeneity across trials.

Strengths and limitations

Strengths: High-quality systematic review and meta-analysis (multiple reviewers, quality assessment of included studies, statistical analysis, use of PRISMA guidelines).

Limitations: Includes 2 studies that do not specify the devices used.

Khoury et al. (2017)

Study design and population

Randomised care trial conducted in a single hospital in Canada.

77 patients with AIS were randomised. Randomised allocation was interrupted when results from other trials were published that showed the benefits of endovascular therapy.

Interventions

Intervention: Standard care plus MT (n=40). Of the 30 patients in whom MT was attempted, a stent retriever was used in 29. Specific devices used were not reported.

Comparator: Standard care alone (n=37), including intravenous thrombolysis when appropriate.

Results

Successful revascularisation (mTICI): Of the patients who had MT, 23 (76.7%) achieved TICI scores of 2b or 3.

Functional independence (mRS): mRS 0 to 2 at 90 days: 20 (50%) patients and 14 (37.8%) patients in the intervention and control arms respectively.

Mortality: Mortality at 3 months: 11 (27.5%) patients in the intervention arm and 9 (24.3%) patients in the control arm.

ICH: S-ICH occurred in 3 (7.5%) and 2 (5.7%) patients in the intervention and control arms respectively.

NIHSS Score: Of the patients with mRS scores of 0 to 2 at 90 days: 87.5% (14/16) and 62.5% (10/16) had NIHSS ≤16 and 25% (6/24) and 19% (4/21) had NIHSS >16 in the intervention and control arms, respectively.

ASPECTS Score: Not reported post-intervention.

Strengths and limitations

Strengths: Randomised trial design.

Limitations: Single-centre trial; relatively low number of patients because of trial interruption.

Lapergue et al. (2017)

Study design and population

A randomised, open-label, blinded endpoint clinical trial in 8 stroke centres in France.

381 adults were randomised. Patients were admitted with suspected ischaemic stroke secondary to occlusion of the anterior circulation.

Interventions

Intervention: Contact aspiration (n=192). Penumbra ACE and MAX reperfusion catheters were used in 90% of 'aspiration first' patients. Sofia and ARC devices were used in 17 patients and 1 patient respectively.

Comparator: Stent retrievers (n=189). Various devices were used, including Solitaire, Trevo, ReVive, ERIC and EmboTrap.

Results

Successful revascularisation (mTICI): Percentage of patients with successful revascularisation defined as mTICI score of 2b or 3 after all endovascular treatments: for the ITT analysis, 85.4% for the intervention group and 83.1% for the controls (OR 1.20; 95% CI 0.68 to 2.10); for the per protocol analysis, 91.5% for the intervention group and 84.9% for the controls (OR=1.91, 95% CI 0.93 to 3.91).

Functional independence (mRS): 45.3% patients in the intervention group achieved independence at 3 months, compared to 50% in the control group (OR=0.83, 95% CI 0.54 to 1.26). The median mRS score at 3 months was 3 for the intervention group and 2.5 for the controls (OR=0.76, 95% CI 0.53 to 1.10).

Mortality: Overall all-cause mortality at 3 months was 19.3% the intervention (35/181) and 19.2% (35/182) for the controls.

ICH: ICH at 24 hours was 46.3% (87/188) in the contact aspiration group and 46.2% (85/188) in the stent retriever group. S-ICH at 24 hours was 5.3% (10/188) and 6.5% (12/188) for the intervention and control groups respectively.

NIHSS Score: The mean change in NIHSS score at 24 hours was −4.8 (95% CI −6.1 to −3.6) for the intervention compared to −5.2 (95% CI −6.5 to −3.9) for the control group.

ASPECTS Score: The median score (at baseline) was 7 for both study arms.

Other: Procedure-related adverse events were 16.2% (31/192) and 15.9% (30/189) for the intervention and controls respectively.

Strengths and limitations

Strengths: RCT with a large number of patients; comparison of stent retrievers and contact aspiration.

Limitations: Trial was not designed to establish non-inferiority between aspiration devices and stent retrievers as first-line strategies; primary outcome was technical (successful revascularisation) rather than clinical (for example, mRS score); authors report that participating centres were highly experienced in the techniques, thus generalisability to stroke centres with less experience may be limited.

Mocco et al. (2017)

Study design and population

A multicentre, randomised concurrent controlled trial, conducted in 36 centres in the US and Germany.

A total of 108 patients were randomised. Patients were adults presenting with large-vessel ischaemic stroke.

Interventions

Intervention: Thrombolysis (intravenous alteplase) plus thrombectomy (IAT; n=55). Devices used were predominantly Penumbra aspiration catheters (n=43). A small number of patients received Solitaire or Trevo stent retrievers.

Comparator: Thrombolysis (intravenous alteplase) alone (IA; n=53).

Results

Successful revascularisation (mTICI): 73% (95% CI 58 to 85) patients achieved successful reperfusion, with mTICI scores of 2b or 3. For patients treated with the Penumbra system alone, this figure was 70%.

Functional independence (mRS): Patients achieving mRS scores of 0 to 2 at day 90: 38% in the IAT group compared to 30% in the IA group (OR=1.4, 95% CI 0.60 to 3.3).

These figures were 41% and 29% respectively for IAT and IA in the per protocol analysis (OR=1.6, 95% CI 0.64 to 4.2).

Mortality: 12% and 24% in the IAT and IA groups respectively (OR=2.3, 95% CI 0.8 to 6.8).

These figures were 7.3% and 24% respectively for IAT and IA in the per protocol analysis (OR=4.1, 95% CI 1.0 to 16.0).

ICH: The S-ICH (as treated) rates were 9.3% and 9.7% for the IAT and IA groups respectively (OR=1.0, 95% CI 0.3 to 3.9).

NIHSS Score: Not reported.

ASPECTS Score: 24-hour median ASPECTS score was 6 for IAT and 5 for the IA group (OR=1.9, 95% CI 0.93 to 1.7). ASPECTS worsening at 24 hours was 1 for IAT and 2 for the IA group (OR=2.4, 95% CI 1.2 to 4.9).

In the per protocol analysis, 24-hour median ASPECTS score was 7 for IAT and 5 for the IA group (OR=2.5, 95% CI 1.1 to 5.5). ASPECTS worsening at 24 hours was 1 for IAT and 2.5 for the IA group (OR=2.5, 95% CI 1.1 to 5.4).

Strengths and limitations

Strengths: A multicentre RCT.

Limitations: Small numbers of patients; trial was halted following external evidence of the added benefit of endovascular therapy to IA alone.

Muir et al. (2017)

Study design and population,

A multicentre randomised, controlled, parallel group clinical trial with blinded endpoint evaluation, conducted in 10 centres in the UK.

A total of 65 patients were randomised. Patients were adults presenting with acute supratentorial ischaemic stroke.

Interventions

Intervention: IVT+MT (IVT+MT) (n=33). MT Devices used were not specified Stent retrievers were used first in 68% procedures and aspiration devices in 32%.

Comparator: Intravenous thrombolysis alone (IVT; n=32).

Results

Successful revascularisation (mTICI): TICI 2b or 3 reperfusion at the end of MT procedure was achieved in 26 (87%) of 30 assessable immediate post-procedure angiograms.

Functional independence (mRS): Percentage of patients achieving mRS scores 0-2 at day 90: 51% in the intervention arm compared to 40% in the controls (OR=2.12, 95% CI 0.65 to 6.94). After adjustment for minimisation variables these figures were 57% and 35%, respectively (OR=4.92, 95% CI 1.23 to 19.7).

Complete functional recovery at day 90 (mRS 0-1): OR=7.63 (95% CI 1.56 to 37.22) in favour of IVT+MT.

Mortality: 7 deaths in the IVT+MT group compared to 4 in the IVT group (OR=1.56, 95% CI 0.29 to 8.40).

ICH: 3 cases of ICH were reported in each study arm. No S-ICH occurred.

NIHSS Score: No significant difference was found in early neurological improvement (improvement of ≥8 points on the NIHSS or an NIHSS of 0 or 1 24 hours after stroke) between groups.

ASPECTS Score: Not reported.

Other: The overall serious adverse events (including deaths) were 15 (45%) in the IVT+MT group and 11 (34%) in the IVT group.

Strengths and limitations

Strengths: A multicentre RCT conducted in the UK.

Limitations: Relatively low number of patients.

Nogueira et al. (2018)

Study design and population

A randomised controlled trial in 26 centres in the USA, Canada, Europe and Australia.

Patients were included who had acute stroke with a mismatch between salvageable brain tissue and volume of infarcted tissue. 107 patients were randomly assigned to the thrombectomy group and 99 to the control group.

Interventions

Intervention: MT using the Trevo device, plus standard medical care (n=107).

Comparator: Standard medical care alone (n=99).

Results

Successful revascularisation (mTICI): 90 (84%) patients in the thrombectomy group had mTICI scores of 2b or 3. Recanalisation was achieved in 82 (77%) patients in the thrombectomy group and 39 (39%) in the control group at 24 hours.

Functional independence (mRS): The utility-weighted mRS was 5.5 for the thrombectomy group compared to 3.4 for the control group at 90 days. Patients with functional independence at 90 days was 52 (49%) for the thrombectomy group and 13 (13%) for the controls.

Mortality: All-cause mortality at 90 days was 20 (19%) for the thrombectomy group and 18 (18%) for the control group (risk ratio: 1, 95% CI 1 to 2); stroke-related deaths at 90 days were 17 (16%) and 18 (18%) for the thrombectomy and control groups, respectively (95% CI 1 to 2).

ICH: Patients with symptomatic ICH at 24 hours was 6 (6%) and 3 (3%) (risk ratio: 2, 95% CI 1 to 7) for the thrombectomy and control groups respectively.

NIHSS Score: Patients classified as having an early response (a decrease in NIHSS Score of 10 points or more from baseline or an NIHSS Score of 0 or 1 on day 5, 6 or 7 post-intervention).

ASPECTS Score: Not assessed.

Strengths and limitations

Strengths: Multicentre RCT.

Limitations: Study was sponsored by device manufacturer (Stryker).

Abbreviations: AIS, acute ischaemic stroke; ASPECTS, Alberta stroke program early CT score; CI, confidence interval; IA, intravenous alteplase; IAT, intravenous alteplase plus thrombectomy; ICH, intracerebral haemorrhage; IQR, interquartile range; ITT, intention to treat; IVT, intravenous thrombolysis; IV t-PA, intravenous tissue plasminogen activator; mRS, modified Rankin Scale; mTICI, modified Thrombolysis in Cerebral Infarction; NIHSS, National Institutes of Health Stroke Scale; OR, odds ratio; PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses; PST, primary suction thrombectomy; RCT, randomised controlled trial; S-ICH, symptomatic intracerebral haemorrhage; SRT, stent retriever thrombectomy; TICI, treatment in cerebral infarction.

Table 5 Summary of the cost-effectiveness studies

Ganesalingham 2015

Outcome

IV t-PA + SR 1

IV t-PA

Difference

Total costs

£39,2742

£31,8152

£7,4322

QALYs

4.84

3.79

1.05

ICER

N/A

N/A

£7,065

Lobotesis 2017

Outcome

IV t-PA + SR (Solitaire)

IV t-PA

Difference

Total costs

£110,322

£143,512

−£33,190

Total QALYs

7.01

4.70

2.31

ICER

N/A

N/A

Dominant

Abbreviations: ICER, incremental cost-effectiveness ratio, IV t-PA, intravenous tissue plasminogen activator; SR, stent retriever; QALY, quality-adjusted life year.

1Pooled data from 5 trials: ESCAPE, EXTEND-IA, MR CLEAN, REVASCAT and SWIFT-PRIME.

2Values in the paper reported in US dollars and converted using an exchange rate of 0.61 (given in the paper).

Recent and ongoing studies

Five ongoing (currently recruiting) RCTs were identified:

Two completed RCTs were identified:

In addition to these RCTs, 1 single-arm study was identified:

4 observational studies were identified:

There are 7 other ongoing studies involving MT listed in which the device has not been specified: 3 RCTs, 3 observational and 1 non-randomised interventional study.