Article Text

Download PDFPDF

Original research
Mechanical thrombectomy with the Trevo ProVue device in ischemic stroke patients: does improved visibility translate into a clinical benefit?
  1. Christoph Kabbasch1,
  2. Anastasios Mpotsaris1,
  3. De-Hua Chang1,
  4. Sonja Hiß1,
  5. Franziska Dorn1,
  6. Daniel Behme2,
  7. Oezguer Onur3,
  8. Thomas Liebig1
  1. 1Department of Radiology and Neuroradiology, University Hospital of Cologne, Cologne, Germany
  2. 2Department of Neuroradiology, University Medical Center Goettingen, Goettingen, Germany
  3. 3Department of Neurology, University Hospital of Cologne, Cologne, Germany
  1. Correspondence to Dr T Liebig, Department of Radiology and Neuroradiology, University Hospital of Cologne, Kerpener Str 62, Cologne 50937, Germany; thomas.liebig{at}


Purpose To investigate the efficacy and safety of the Trevo ProVue (TPV) stent retriever in stroke patients with large artery occlusions, with particular attention to the full structural radiopacity of the TPV.

Materials and methods Case files and images of TPV treatments were reviewed for clinical and technical outcome data, including revascularization rates, device and procedure related complications, and outcome at discharge and after 90 days.

Results 76 patients were treated with TPV. Mean National Institutes of Health Stroke Scale (NIHSS) score was 18 and 68% had additional intravenous thrombolysis. 63 occlusions were in the anterior circulation: 44 M1 (58%), 8 M2 (11%), 8 internal carotid artery-terminus (11%), 2 internal carotid artery- left (3%), 1 A2 (1%), and 13 vertebrobasilar (17%). 58 of 76 (76%) were solely treated with TPV; the remainder were treated with additional stent retrievers. Mean number of passes in TPV only cases was 2.2 (SD 1.2). In rescue cases, 3.2 (SD 2.2) passes were attempted with the TPV followed by 2.6 rescue device passes (SD 2). TPV related adverse events occurred in 4/76 cases (5%) and procedural events in 6/76 cases (8%). Mean procedural duration was 64 min (SD 42). Thrombolysis in Cerebral Infarction (TICI) 2b/3 recanalization was achieved in 69/76 patients (91%), including 50% TICI 3. Of 56 survivors (74%), 37 (49%) showed a favorable outcome at 90 days (Solitaire With the Intention for Thrombectomy trial criteria), statistically associated with age, baseline NIHSS, onset to revascularization time, and TICI 2b–3 reperfusion. TPV radiopacity allowed for visual feedback, changing the methodology of stent retriever use in 44/76 cases (58%).

Conclusions Neurothrombectomy with TPV is feasible, effective, and safe. The recanalization rate compares favorably with reported data in the literature. Improved structural radiopacity may facilitate neurothrombectomy or influence the course of action during retrieval.

  • Stroke
  • Thrombectomy
  • Intervention
  • Angiography

Statistics from


Since their introduction in 2008, self-expanding stent retrievers have been studied for use in mechanical thrombectomy.1 ,2 After encouraging early results of single center series and non-randomized multicenter trials,3–13 the Solitaire and Trevo devices have proven to increase good functional outcome compared with the Merci device14 ,15 in multicenter registries and in post marketing studies.16–20 In 2015, five randomized multicenter trials of thrombectomy versus conservative treatment demonstrated the clinical benefit of mechanical thrombectomy (Multicenter collaboration for endovascular treatment of acute ischemic stroke in The Netherlands (MRCLEAN), Endovascular Treatment for Small Core and Proximal Occlusion Ischemic Stroke (ESCAPE), Extending the Time for Thrombolysis in Emergency Neurological Deficits-Intra-Arterial (EXTEND-IA), Solitaire With the Intention For Thrombectomy as Primary Endovascular Treatment (SWIFT-PRIME), and Endovascular Revascularization With Solitaire Device Versus Best Medical Therapy in Anterior Circulation Stroke Within 8 Hours (REVASCAT)). In all trials, the vast majority of patients were treated with a stent retriever.21–25

The Trevo ProVue stent retriever

The underlying concept of stent retriever neurothrombectomy makes use of a self-expanding mesh that interacts with the clot, entangling it, so that both can be withdrawn from an occluded artery. The Trevo ProVue (TPV) device consists of a flexible tapered nitinol core wire with a shaped section at the distal end for retrieving the clot. The distal shaped section is tapered for delivery to the target vessel location, holding a platinum marker for fluoroscopic control of its distal most position. In addition, integration of platinum wires into the stent struts itself allows for visualization of the complete device (figure 1). To date, this feature remains unique and enables the operator to view the placement and strut behavior of the whole device throughout the entire procedure. The TPV can be used with lesional aspiration through an intermediate catheter as well as with flow arrest and proximal aspiration (eg, through a balloon occulsion guide catheter (balloon occulsion guide catheter (BOGC)).

Figure 1

Trevo (upper) and Trevo ProVue (lower) devices under subtracted fluoroscopy.


The purpose of this study was to evaluate the technical feasibility and safety of the TPV as well as its angiographic and clinical effectiveness in the revascularization of large vessel occlusions in acute ischemic stroke. The study was a high volume single center assessment to determine whether the structural radiopacity of the TPV (1) influences the progress of thrombectomy and (2) if this translates into a procedural or even clinical advantage. This advantage could be reflected in the number of passes needed, final recanalization result, incidence of adverse events, and clinical outcome.

Materials and methods

The TPV retriever has been available at our institution since January 2012. Prospectively, during all treatments that involved the use of the TPV, a note was kept in a data file as to whether the novel feature of full structural radiopacity influenced therapeutic decision making. Our prospectively kept institutional database was then retrospectively screened for patients with acute ischemic stroke due to proximal large artery occlusion who underwent mechanical thrombectomy with the TPV. All cases were reviewed by two experienced neurointerventionists, and the images were re-evaluated applying the Thrombolysis in Cerebral Infarction (TICI) scale and respective vessel occlusion site definitions.26 Device and procedure related complications were extracted from the case files. All neurological assessments and scores (eg, National Institutes of Health Stroke Scale (NIHSS) score and modified Rankin Scale (mRS) score) during the acute phase and at follow-up after 90 days were performed independently by the treating neurologists.

Patient selection

The local ethics committee granted approval for inclusion and for data analysis of all patients. Patients were eligible for mechanical thrombectomy in the anterior circulation if they had symptoms of a stroke and a corresponding occlusion of a proximal large artery on CT or MRI that also served to indicate salvageable brain parenchyma (eg, CT perfusion mismatch or diffusion weighted imaging and apparent diffusion coefficient map suggestive of acute stroke, without a correlating FLAIR abnormality).27 ,28 Pertaining to the posterior circulation, all patients presenting with a brainstem syndrome within 24 h from onset were eligible. Patients with cerebral hemorrhage or infarct demarcation greater than one-third of the middle cerebral artery territory on pre-interventional imaging were excluded from mechanical thrombectomy. There was no general age limitation for inclusion. Patients were included in the present analysis whenever the TPV was utilized as a firstline device. This was at the discretion of the operator. Patients eligible for intravenous thrombolysis (IVT) were treated according to the national neurological guidelines, independent of the decision to perform mechanical thrombectomy.

Treatment technique

All endovascular procedures were performed under general anesthesia. The standard setup for endovascular treatment comprised a triaxial system. In anterior circulation stroke, a short 8 F sheath was used to navigate an 8 F guide catheter into the internal carotid artery. In vertebrobasilar occlusions, the guide catheter was 6 F. In all cases, intermediate catheters were used for lesional aspiration immediately adjacent to the clot with the stent retriever being pulled back into the intermediate catheter (DAC, Concentric Medical, Mountain View, California, USA): 0.057 inch for the anterior circulation and 0.044 inch for the posterior circulation. Then the 0.018 inch microcatheter of the Trevo system was navigated distal to the clot over a standard 0.014 inch guidewire that was exchanged for the TPV after verifying intra-arterial positioning by careful microcatheter contrast injection. The TPV was left in place for at least 2 min; the process of expansion and thus presumably clot integration was monitored fluoroscopically with more time allowed if expansion was visibly delayed. The retriever was then withdrawn under manual aspiration using a 20 mL syringe applied to the intermediate catheter via a rotating hemostatic valve. This maneuver was repeated until the final revascularization result. The final angiographic result was documented in anteroposterior and lateral projections covering all of the brain. Making use of its unique radiopacity, withdrawal of the TPV was interrupted and restarted in those cases where incomplete clot integration could be seen as a persisting deformity of the stent after deployment that remained locally constant with regard to the arterial segment (ie, it did not move with the stent retriever). If, on the other hand, indentation of the stent structure moved along with it, this was judged as an indicator of simultaneous retrieval of stent and clot.

Clinical and imaging assessment

NIHSS and mRS scores were assessed on admission and at discharge. All relevant periprocedural time points were recorded and the respective intervals calculated. The number of passes with each stent retriever and the sequential order of the applied devices were recorded. Procedure related adverse events, such as device failure or thromboembolic occlusions of previously unaffected territories caused by clot migration, were assessed.

Twenty-four hours post-interventional CT and/or MRI scans were evaluated for symptomatic intracerebral hemorrhage (sICH) based on European Cooperative Acute Stroke Study III (ECASS III) criteria.29

The effect of increased radiopacity was judged on a simple three point scale as ‘unnecessary’, ‘nice to have but did not alter the course of action’, and ‘important, with impact on the course of action’. The rating was performed in consensus but with the performing interventionist had the casting vote.

Statistical analysis

Continuous study parameters were compared between patients using Welch's t test for normally distributed data, or the Mann–Whitney U test for non-normal or ordinally distributed data. Categorical variables were compared using Fisher's exact test. All statistical analyses were performed using GraphPad Prism software V.6.1 (GraphPad Software Inc, La Jolla, California, USA). Logistic regression was executed using MedCalc software (MedCalc, Ostend, Belgium). The significance level for all tests was set at α=0.05

The subgroups with favorable and unfavorable clinical outcomes were analyzed for correlations with age, gender, baseline NIHSS, bridging IVT, onset to revascularization time, procedural duration, TICI reperfusion score, number of stent retriever passes, and discharge NIHSS. Stepwise logistic regression was calculated for those variables that were significantly different between groups with a favorable and unfavorable clinical outcome.



Seventy-six patients were treated with the TPV between January 2012 and March 2014; 40 women (53%) and 36 men (47%), with a mean age of 74.4 years (SD 12.9). The target occlusions were distributed as expected in a clinical setting, with the majority being middle cerebral artery occlusions (tables 1 and 2). Clinically, patients presented with a median baseline NIHSS score of 18 (range 5–38).

Table 1

Patient characteristics

Table 2

Territorial distribution of occlusions, stroke severity on admission, and angiographic outcome

Procedural timings

In 62/76 cases (82%), the time of symptom onset was known, resulting in an average symptom onset to groin puncture time interval of 230 min (SD 100 min). Mean total procedure time was 64 min (SD 42 min). It took an average of 298 min (SD 111 min) from stroke onset to obtain the final recanalization result.

Periprocedural aspects and complications

Fifty-two of 76 patients (68%) received IVT before thrombectomy in a bridging approach. In total, 187 retrieval maneuvers were performed with TPV retrievers in 76 patients (mean 2.46, SD 1.57; IQR 1–9). In 58/76 patients (76%), the TPV was used as the sole device. In 17 patients (22%) a Solitaire FR was used as a rescue device and in 1 case a pREset was used (Phenox, Bochum, Germany). In these 18 rescue cases, successful revascularization was achieved in 14 (78%). The mean number of passes was 2.2 (SD 1.24) in the TPV only cases. In the 18 unsuccessful TPV attempts, the mean number of passes with the TPV was 3.2 (SD 2.2); the rescue device was used for another 2.6 passes (SD 2.0). For comparison, the number of passes in other studies are as follows: SWIFT, 1.7±0.9 (mean); SWIFT-PRIME, not mentioned; Mendoca et al, 2 (1–3) (median (range)); San Roman et al, maximum 3; TREVO, 2.62±1.37 (mean); TREVO 2, 2.4±1.4 (mean). Concomitant lesional aspiration via the intermediate catheter was performed in all cases. The device was deployed successfully in 74/76 cases (97%). A delivery problem was observed in two distinct cases: the TPV got stuck in the delivery microcatheter and could not be deployed in the target vessel. Peripheral embolization into a new territory was seen in two cases (3%). These were not treated separately and did not account for any clinical sequelae.

Angiographic results

The initial angiographic TICI score was 0 in 75/76 patients (99%) and TICI 1 in 1/76 patients (1%). Overall, complete or near complete recanalization (TICI ≥2b) was achieved in 69/76 patients (91%), including 38 cases (50%) with full reperfusion (TICI 3). When the TPV was used as the sole device, TICI ≥2b was achieved in 55/58 cases (95%), comprising 32/58 (55%) TICI 3 recanalization.

Clinical patient outcome

Mean NIHSS score was 18 (5–38) on admission and 6 (1–36) at discharge. Fifty-six patients survived the inhospital phase (74%) and 37 patients (49%) showed a good clinical outcome (mRS 0–2) on follow-up at 90 days, based on the SWIFT study criteria (mRS 0–2 or NIHSS improvement ≥10 points).30 Clinical outcome was correlated with the following parameters in a univariate analysis: age, baseline NIHSS, onset to reperfusion, TICI 0–2a versus 2b–3, and discharge NIHSS (see table 3). Forward step logistic regression identified baseline NIHSS and onset to successful reperfusion as predictive factors of clinical outcome. There was no difference between the anterior and posterior circulation with regard to median mRS score at 90 days (mRS=3), but the rate of mRS 0–2 at 90 days in the posterior circulation was lower than in the anterior circulation (31% vs 38%).

Table 3

Clinical patient outcome data

Radiographic visualization

The procedural impact of the improved radiographic visualization of the TPV compared with other systems was rated on a simple three point scale as ‘unnecessary’, ‘nice to have but did not alter the course of action’, and ‘important with impact on the course of action’ (an example of the latter is displayed in figure 1). Here, initially the TPV could be seen to not open, even after 5 min. During retrieval, the TPV collapsed while passing the clot that consequently did not integrate into the TPV and the vessel remained occluded despite an initial displacement and bypass effect with distal contrast in the M2.

Possible alterations in the course of action included the following:

  • TPV did not open initially as a sign of non-integration of clot into the stent structure. In such cases, the time allowed for clot integration was chosen to be longer than the usual 2 min.

  • During retrieval, the TPV collapses distal and reopens proximal to the clot, indicating that the clot did not move together with the retriever, making successful retrieval highly unlikely. In such cases, withdrawal was halted and restarted with the TPV left in place and with the intention to allow extra time for clot integration.

  • Alteration of the course of the vessel, such as straightening, indicated by straightening of the retriever, was judged to be a sign of stretching and potential damage to the selected vessel. In these cases, partial resheathing with the microcatheter was applied to reduce the amount of force transmitted to the vessel.

Overall, the radiopacity of the TPV was judged to have influenced the course of action of the intervention in 44/76 cases (58%). In 22 cases (29%), increased radiopacity did not play a role, and in only 10 cases (13%) was it judged ‘nice to have’ but did not influence the course of the intervention, usually because of first pass recanalization success.


Previous studies on stent retriever based thrombectomy have shown that beneficial revascularization, defined as TICI 2b–3, can be achieved in up to 87% of cases in core laboratory controlled trials, such as EXTEND-IA.22 The standard Trevo has been evaluated in two multicenter trials with similar success rates: 86% TICI 2 or greater was achieved in the TREVO 2 trial.14 ,31 In our series, the TPV showed clinically beneficial TICI 2b–3 recanalization rates of 91% (total), 95% in TPV only cases, with 50% TICI 3. Final reperfusion success was achieved after an average of 2.2 passes. These numbers are within the range of other single center stent retriever studies.3 ,9 ,32–34 Our clinical results of 49% mRS 0–2 at 90 days or improvement of 10 or more on the NIHSS appear to be comparable with both TREVO trials (mRS 0–2 of 55% in TREVO and 40% in TREVO 2). The overall rate of procedural adverse events was 8% (6/76) and the rate of sICH was 5% (4/76), which is in line with 11% overall and 5% sICH reported in a comparable single center study by Behme et al35 but higher than in REVASCAT (1.9%) and in the stent retriever group of SWIFT-PRIME (0%).24 ,25 While being aware of the uncertainties of a single center study without core laboratory control, we assume that the performance of the TPV is at least on a par with other stent retrievers with regard to reperfusion success as well as the average time to achieve the final reperfusion result.

The goal of this study was to assess the safety and effectiveness of the TPV retriever for mechanical thrombectomy in acute stroke. The TPV was designed to achieve improvements in neurothrombectomy, not only by geometrical optimization (such as perpendicular orientation of the struts and relatively wide cells that should facilitate clot integration) but also by implementing full structural radiopacity as a distinct technical feature that would allow the physician to witness the TPV application and actual strut behavior throughout the entire procedure. This feature has not been addressed in previous studies. In contrast with endovascular implants that demand full radiographic control of the positioning process, such as platinum coils or guidewires, stent retrievers in thrombectomy are applied in a technically simple fashion with sole control of the distal position followed by deployment through backward motion of the delivery catheter. Therefore, visualization of just the distal markers of the stent retriever would be technically sufficient to perform this maneuver. On the other hand, clot integration and the interaction between the clot and stent are obscured from the interventionist, with the only exception being pre-retrieval contrast application to control for a potential bypass effect that will however not reliably distinguish between clot integration and mere clot displacement as long as the stent structure itself is not visible. During retrieval, there may be (very limited) tactile feedback but certainly no visual control of the clot–stent interaction and thus this maneuver is completely arbitrary or based on previous experience and the manufacturer’s guidelines for use.

The TPV is the first stent retriever that allows visualization of the clot–strut interaction during both deployment and retrieval. This enables the physician to adapt or modify the procedure to account for signs of potential failure, such as obvious non-integration of the clot with the stent retriever just sliding past it or visible straightening of the target vessel without relative movement of the stent retriever that may indicate potential structural damage, such as side branch rupture, and can be addressed with partial resheathing.

Even though the range of actions that may arise from the additional visual information are limited compared with other devices or implants, we felt that it is a helpful feature as long as it does not interfere with or hamper mechanical retrieval itself. While our reperfusion results virtually rule out negative effects on the mechanical aspects of the design of the TPV on retrieval itself, we found in our series that in more than half of the cases (58%) radiopacity was felt to have influenced the course of action by either allowing more or additional time for clot integration (40 cases) or by prompting partial resheathing (4 cases). Our results with the TPV compare favorable with a previously published series from our own institution with the Solitaire. In 108 cases, the overall rate of TICI 2b–3 reperfusion was 79%,3 although this series was acquired earlier and thus increased experience over time may be a factor here. The decision to use the Solitaire as a rescue device was made in 16 treatments and the two cases of device failure (TPV stuck in the catheter) and was based on the assumption that differences in cell size and the additional ability of the Solitaire to fold over on itself might be beneficial in cases where the TPV did not extract the clot within a reasonable number of attempts. The maximum number was nine in a case of basilar artery occlusion, where a Solitaire reached a TICI 3 after just two additional passes. This does not prove potential superiority of any available stent retriever but shows that differences in design and specific features may prove helpful at times and that potentially there is no one for all solution. Also, as shown in table 4, the results presented here compare favorably with the relevant literature with regard to the number of passes, time to recanalization, and recanalization success, even though the differences are marginal and, since there are so many additional confounders that contribute to clinical outcome, are not reflected in patient outcome. We however believe that with growing evidence and widespread use of mechanical thrombectomy with the now proven efficacy of stent retrievers, even such seemingly minor effects, such as an increase in first pass success or reduction of the total number of passes needed, will play a role. This has been proved by various groups (histologically from in vivo36 as well as in vitro37 work) who showed that stent retrievers may cause significant endothelial damage. The interaction of a stent retriever and the endothelium will thus influence the rate of secondary re-occlusion due to activation of extrinsic coagulation or it may cause delayed wall changes that could lead to vascular stenosis.

Table 4

Comparison with previous studies and case series

Beyond the scope of this evaluation, it was generally felt that radiopacity adds valuable feedback during thrombectomy, especially by less experienced physicians and those in training that were just watching the procedures. While intensive work has been done by X-ray equipment manufacturers to enhance stent visibility in other fields,38–40 nitinol stents with overlying bony structures usually remain almost invisible.41 Recently, the potential clinical benefit of conscious sedation over general anesthesia for neurothrombectomy has been emphasized.21 ,22 ,42 ,43 In such a setting, where safety, accuracy, and speed are paramount, clear and reliable visual feedback of the stent retriever device that goes beyond just a few markers would be very helpful for the operator.


The TPV has been shown to yield a high rate of successful recanalization and appears to be on a par with what has previously been published with the original Trevo stent retriever. The safety profile is acceptable for an emergency procedure. Full structural visibility of the TPV allows for procedural adaptation to the actual stent–clot or stent–vessel interaction. In more than half of the cases in our series, its visual feedback influenced the progress of the intervention, even though this was not reflected by improved reperfusion success or clinical outcome. Due to the limited range of procedural alterations in our series, the effect may only translate statistically into a clinical benefit in larger series.


View Abstract


  • Contributors CK: manuscript writing, data collection, and patient treatment. AM: manuscript editing and data collection. DB: statistics. FD: patient treatment and data contribution. D-HC: manuscript editing. SH: data collection. OAO: data collection, and patient physical and neurological examination. TL: project development, manuscript writing, and patient treatment.

  • Competing interests AM: proctor and consultant for Penumbra Inc and Sequent Medical. TL: proctor and consultant for Sequent Medical; consulting fees from Stryker and Acandis.

  • Ethics approval The study received local ethics committee approval for retrospective analysis of stroke related patient data.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement All data can be made available in an anonymized manner on request.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.