The Penumbra Aspiration system (Penumbra Inc., Alameda, CA) is designed to remove intracranial occlusions by debulking, aspiration, and extraction [24]. It is comprised of an aspiration platform containing multiple devices that are size matched to the specific neurovascular anatomy (Fig. 6). A 2.8–5 F reperfusion catheter is connected to the aspiration pump, generating a suction force of −700 mmHg. Because the device works on the proximal surface of the occlusion, it optimizes safety and eliminates the need for navigation beyond the clot. Early clinical experience suggests that the Penumbra can be applied for successful revascularization in patients with acute stroke [24]. In an early study, Grunwald et al. [28] used the Penumbra to treat two of three patients aged 7–16 years with AIS; the remaining child was treated with the pCR. Partial recanalization was achieved with no adverse events. This prompted a larger, prospective, single-arm trial by the same group [29] including 23 patients with angiographically verified occlusion of a treatable intracranial vessel who presented within 8 h of symptom onset. Twenty-one target vessels in 20 patients were treated with the Penumbra system. All were revascularized to TIMI grade 2 or 3 (100 % success rate). At the 30-day follow-up, 45 % of patients showed an increase of 4 or more points on the NIHSS or an mRS score of ≤2. The mortality rate was 45 %, which was considered good in this cohort in which 70 % of patients had a baseline NIHSS score of >20 or BA occlusion [9]. The Penumbra was approved by the FDA in 2008.

The same year, Bose et al. [9], from the University of Saarland, reported the results of a 3-year prospective, nonrandomized, single-center controlled study of the clinical and functional outcome of 29 patients with stroke in the posterior circulation, including acute occlusions of the distal ICA (T-lesion), MCA–M1, or BA. Mechanical thrombectomy with the Penumbra was applied in patients who were ineligible for IV tpA or failed IV tPA administered within 3 h of symptom onset. Mean baseline NIHSS score was 20 for the whole cohort and 33.5 in the patients with BA strokes. Complete revascularization (TIMI 3) was achieved in 72.4 % of patients, and partial revascularization (TIMI 2) in 13.8 %. Revascularization failed in 13.8 %. The subgroup with BA strokes had a partial revascularization rate of 83.3 %. The NIHSS score improved by 4 points or more in 65.5 % of the cohort, and an mRS score of ≤2 was documented in 37.9 %. There were no device-related adverse events. Seven percent of patients had sICH [9].

Results of the Penumbra Pivotal Stroke Trial [71] were presented at the International Stroke Conference in February 2008 and published shortly thereafter. A prospective, multicenter, single-arm, nonrandomized design was used. The sample consisted of 125 patients with neurological deficits (NIHSS ≥8) at presentation and angiographically proven occlusion (TIMI 0 or 1) of a treatable large intracranial vessel (total, 125 target vessels). Patients were either ineligible for or refractory to tPA therapy or presented beyond the 3-h window (between 3 and 8 h of symptom onset). At 90 days after treatment, revascularization was successful (TIMI 2–3) in 81.6 % of treated vessels. Eighteen intraprocedural events were reported in 16 patients (12.8 %); in 3 patients (2.4 %), the events were considered serious. Intracranial hemorrhage was detected on 24-h CT scan in 35 patients (28 %), of whom 14 (11.2 %) were symptomatic. All-cause mortality was 32.8 % at 90 days, with 25 % of patients achieving an mRS score of ≤2. The authors concluded that the Penumbra is safe and effective for revascularization in patients with AIS secondary to large-vessel occlusive disease if administered within 8 h of symptom onset. Although the primary endpoint was met, this study was criticized for the low rate of 90-day functional independence. Goyal et al. [25] attempted to explain the enigma of the high revascularization rate on the one hand and lower-than-expected good-outcome rate on the other. Analysis of the pretreatment CT scans revealed that an ASPECTS score of >7 (small infarct) at baseline was associated with a higher likelihood of achieving functional independence after revascularization than a score of ≤7 (50 % vs. 15 %; p = 0.0001), and a total of 64 % of patients in the trial presented with large infarcts. This finding, along with other emerging data [95], supports the use of advanced imaging selection to guide the use of IA therapy.

Thereafter, 7 international centers published their collective post-marketing experience with the Penumbra Aspiration System in a retrospective review of 157 consecutive patients (Penumbra POST) [86]. Mean patient age at baseline was 65 years, and mean NIHSS score, 16. The rate of MCA occlusions (58 %) was lower than in the Pivotal trial, and the rate of vertebral artery occlusions (20 %) was higher, but neither of these differences was statistically significant (p < 0.05). Revascularization was documented in 87 % of the treated vessels: to TIMI 2 in 54 % and to TIMI 3 in 33 %, compared to 82 % in the Pivotal trial. In addition, 32 % of patients received adjunctive IA lytic treatment during the procedure, 18.5 % received IV lytic treatment prior to the procedure (refractory to tPA), and 23 % received both. Although the IA tPA group tended to have a higher rate of sICH, adjunctive lytic therapy did not materially affect the revascularization rate or the clinical outcome. There were 9 serious procedural events (5.7 %): 2 dissections, 2 perforations, one access site hematoma, one peripheral hemorrhage, and one cardiac arrest. Three device malfunctions occurred, including 2 fractures of the 032 reperfusion catheter at ~40 cm from the tip, and one breakage of the 041 separator at ~6 mm from the distal end of the device. The breakage was due to a hard clot that required the operator to pull back on the separator with considerable force after penetrating the thrombus. The tip of the separator stayed within the thrombus and was not retrievable. None of these malfunctions led to a serious adverse event or patient death. The all-cause mortality rate was 20 %. Forty-one percent of patients had an mRS score of ≤2 at 90 days compared to 25 % in the Pivotal trial. Although the postmarketing Penumbra data were not significantly different from the Pivotal trial data, the proportion of patients with a good functional outcome was higher than expected. A review of the Penumbra POST trial database did not reveal a definitive reason for the observed difference in outcome between the studies. However, it raised the intriguing possibility that it was at least partly attributable to an inherent variability in patient stroke pathophysiology at presentation.

The 2011 Penumbra SPEED study [20] assessed the safety and efficiency of the 054 catheter, which was added to the existing 3 catheters of the Penumbra system in 2009. It features a larger, 0.064 in., proximal internal diameter for enhanced aspiration efficiency. The interim results (n = 73) yielded a median aspiration time of 18 min compared with 45 min for the original catheters used in the Penumbra Pivotal trial (p < 0.001). The time from puncture to final revascularization was 53.5 min in the SPEED study and 97 min in the Pivotal Trial. The 054 catheter was associated with a high revascularization rate (TIMI 2–3) of 92 % when used as frontline therapy, prior to any adjunctive treatment.

The Solitaire (ev3 Inc., Plymouth, MN) is a self-expanding stent initially designed for stent-assisted coiling in remodeling of brain aneurysms (Fig. 7). It is currently offered in two versions: aneurysm bridging (AB) and flow restoration (FR). The Solitaire is unique in that it can be fully deployed and then completely retrieved if not detached [35, 87] once recanalization is established. Although detachment of the device is not within the primary scope for the treatment of ischemic stroke, it is a viable option if thrombectomy fails and thrombus compression by the deployed stent restores flow in the target vessel [72]. Other advantages include a closed-cell design, high comparative radial force, high cell deformation resistance, and good vessel-wall adaptation. Furthermore, the device is applicable multiple times and can be used even in small peripheral vessel branches (e.g., M2 segments). The FR variant was CE-marked for mechanical thrombectomy in July 2009.

The results of the first (pilot) study of the Solitaire AB stent in mechanical thrombectomy were published in 2010 in Stroke. Castaňo et al. [12] prospectively evaluated the intraprocedural and postprocedural findings in 20 patients with large artery occlusions of the anterior circulation attending a single tertiary center. All were refractory to or ineligible for IV tPA. Additional inclusion criteria were age ≤80 years, NIHSS score ≥8 (or lower if there was a fluctuating neurological deficit), <8 h between symptom onset and presentation, absence of large signs of ischemia (defined as an ASPECT score of <7 or magnetic resonance diffusion-weighted lesion in >50 % of the MCA territory), and proven occlusion of the anterior circulation on angiography. The treatable vessels included the terminus ICA, tandem proximal ICA/MCA, and MCA-M1 and -M2. The median NIHSS score was 19 (interquartile range, 15–23). An 8-Fr balloon guide-catheter was inserted via a femoral approach, so that the device was positioned in the proximal ICA. A heparinized saline solution was continuously perfused through the catheter during the procedure. With the balloon of the guide-catheter deflated, a 0.014-in. guide wire was advanced coaxially over a Rebar 18 microcatheter (ev3 Inc., Irvine, CA) within the occluded intracranial vessel and navigated distal to the clot. The microcatheter was then advanced over the wire through the clot, and the guide-wire was exchanged for the Solitaire AB embolectomy device. The device was advanced and deployed, with its distal portion placed a few millimeters distal to the clot. If the clot was shorter than the stent (20 mm), flow distal to the thrombus was restored immediately. The stent was maintained in place for 1 or 2 min, followed by inflation of the guiding catheter balloon to arrest the flow. The microcatheter and embolectomy device were gently aspirated simultaneously through the guiding catheter. Successful revascularization, defined as TICI 2b/3, was achieved in 90 % of treated vessels; flow was immediately restored in 80 %. The mean number of passes for maximal recanalization was 1.4, and the median (quartiles) time from groin puncture to recanalization was 50 min (range 38–71 min). In no case was adjuvant therapy required after deployment of the embolectomy device. There were no significant adverse procedural events. Distal embolization to the MCA territory occurred in two patients, but there were no cases of thrombus migration to a different territory. There were also no cases of arterial rupture or dissection. Two patients (10 %) had sICH: one fatal parenchymal hemorrhage, and one subarachnoid hemorrhage. Four patients (20 %) died during the 90-day follow-up period. Forty-five percent of patients showed a good functional outcome at 3 months (mRS ≤2). Age and NIHSS score were predictors of good outcome, in agreement with other reports on thrombectomy [82, 84]. The mean number of passes was lower than in the MERCI studies (2.9 passes), and the mean time from groin puncture to recanalization was shorter. This study suggested for the first time that the Solitaire AB device can rapidly, safely, and effectively retrieve clots from the MCA terminus and ICA within 8 h of symptom onset.

In other small case series, Miteff et al. [61] evaluated 26 consecutive patients, Nayak et al. [67] 7 patients, and Möhlenbruch et al. [63] 25 patients. Recanalization (TIMI grade ≥2) was achieved with Solitaire thrombectomy as monotherapy in 61.5, 100, and 88 % of patients, respectively. Corresponding rates of sICH were 7.7, 0, and 12 %. In the series of Miteff et al. [61], a favorable clinical outcome (mRS ≤2) was documented in 3 of 5 patients (79 %) with MCA occlusion, 6 of 11 patients (55 %) with ICA occlusion, and 2 of 10 (20 %) with BA occlusion; in the series of Nayak et al. [67], an improvement of more than 4 points on the NIHSS was documented in 5 patients (20 %), of whom 4 (57 %) had a 30-day mRS score of ≤2. No data on clinical outcome were available for the series of Möhlenbruch et al. [63]. Additional single-center experiences with the Solitaire for AIS are summarized in Table 1.

The two most comprehensive series of temporary endovascular bypass with the Solitaire were those of Liebig et al. [50] (50 patients and 53 occlusions) and Costalat et al. [17] in the Rescue, Combined Management and Stand-Alone Thrombectomy (RECOST) study (100 patients and 103 large-vessel occlusions). In the study of Liebig et al. [50], repeated stent manipulation was required in 64 % of patients; only six procedures (11 %) resulted in stent deployment. There were four adverse events, none device related. Immediate reperfusion (TICI ≥2a) was achieved in 73.6 % of procedures, extending the theory that stenting leads to relatively quick recanalization. The final reperfusion (TICI 2–3) rate was 86.6 %. The RECOST study used three different thrombectomy strategies based on time of symptom onset and location of vessel occlusion (MCA-M1 or -M2, 40 % of patients; ICA, 28 %; BA, 32 %). (1) Rescue therapy (24 % of cases) was applied in patients with MCA-M1 or -M2 occlusion who presented within 4.5 h of symptom onset and failed early IV tPA treatment (defined as NIHSS ≥8 60 min after IV fibrinolysis or neurological deficit significantly impacting quality of life). (2) Combined (bridging) therapy (56 % of cases) was used in patients with terminal ICA or tandem occlusions who presented within 4.5 h of symptom onset and patients with BA occlusions who presented within 24 h. No second neurological assessment was obtained before urgent thrombectomy, and IV fibrinolysis was continued during the endovascular procedure. (3) Stand-alone thrombectomy (20 % of cases) was used in patients who presented beyond the IV fibrinolysis therapeutic window (4.5–6 h for MCA and terminal ICA occlusions) and patients with well-accepted contraindications to IV fibrinolysis. Mean patient age was 67.6 years. All patients underwent clinical and magnetic resonance imaging (MRI) assessment. Mean NIHSS score at presentation was 14.7, and mean MRI ASPECT score was 6. Mean recanalization time from symptom onset was 377 min, with an overall recanalization rate (TICI 3) of 84 %. NIHSS score at discharge was 6.5; 60 % of patients had an NIHSS score of 0–1 or an improvement of >9 points. The symptomatic complication rate was 10 %. At 3 months, 54 % of patients had mRS score ≤2; the overall mortality rate was 12 %. The authors concluded that patient selection by MRI ASPECT score contributed to the low complication rates because it prevented futile and dangerous interventions [17].

The preliminary results of the ongoing Solitaire FR With the Intention For Thrombectomy (SWIFT) Study [80], the first to compare the Solitaire self-expanding stent with the MERCI retriever as initial therapy, were reported in 2012 at the International Stroke Conference. A multicenter, randomized, active comparator, noninferiority study design was used. Between January 2010 and December 2011, 58 patients were randomized to the Solitaire FR arm, and 55 to the MERCI arm. Primary endpoints were successful recanalization (TIMI 2 or 3 in all treated vessels), no sICH, and no rescue therapy (tPA, IIb/IIIa antagonists, etc.). Secondary endpoints were either good clinical outcome (mRS ≤2 or equal to the prestroke mRS if the prestroke mRS was higher than 2 or improvement by ≥10 in NIHSS score). A core lab blinded to the treatment assignment provided independent angiographic, CT, and MRI evaluations. Compared to the MERCI arm, the Solitaire FR arm had significantly higher rates of successful recanalization without sICH (core-lab-assessed) (60.7 % vs. 24.1 %, p = 0.0001), end-of-procedure successful recanalization (core-lab-assessed) (80.4 % vs. 57.4 %, p = 0.013), and good neurological outcome (mRS ≤2 or improvement in NIHSS score by ≥10) at 90 days (58.2 % vs. 33.3 %, p = 0.017), and a significantly lower 90-day mortality rate (17.2 % vs. 38.2 %, p = 0.02).

The sponsored multicenter prospective Solitaire FR Thrombectomy for Acute Revascularization (STAR) study [26] is the first “real-life” investigation of use of the Solitaire for acute stroke. The enrollment of 206 patients from 18 centers in Canada and Australia has just been completed. Inclusion criteria are presentation within 8 h of symptom onset and contraindications for IV tPA or failed IV tPA after bridging therapy or thrombectomy as initial treatment.

A search of PubMed until December 2012 using the combined keywords “Solitaire” and “stroke” yielded 43 publications of related devices to the Solitaire system. Early followers are discussed below, including the Trevo (Concentric Medical), the ReVive (Micrus Endovascular), and the Aperio (Acandis). Others include the IRIIS Plu and OptiCell (MindFrame, CA), the Pulse (Penumbra), and the pREset (Phenox) [72]. They differ in several features, but their intended clinical use is essentially similar.

The Trevo (Concentric Medical Inc., Mountain View, CA) is designed for mechanical removal of occlusive thrombi in vessels ranging from 1.5 to 3.5 mm in diameter [59]. It measures 44 mm in length and 4 mm in diameter and has a hydrophilic coating to reduce friction during use. The device consists of a 180-cm long, proximal 0.018-in. core wire, with a 75-cm transition area at the proximal end and a closed-cell stent-like section at the distal end (Figs. 8 and 9). The proximal 10 mm are tapered for easy resheathing, and the distal 10 mm are tapered for smooth transition from the tip to the active area of the device. The distal tip is soft and floppy for safe and accurate deployment, and radiopaque to facilitate fluoroscopic visualization. The higher radial force across the active area allows for placement of the device in more distal and smaller vessels and more efficient clot retrieval and incorporation. A shaft marker indicates the proximity of the device to the microcatheter tip. By contrast to stents used to treat intracranial aneurysms (neck bridging) or intracranial atherosclerosis, in which the broader portion of the struts is in contact with the blood vessel wall in order to optimize coverage and wall apposition, in the Trevo device, the broader portion of the struts has an endoluminal orientation to optimize thrombus incorporation.

Nogueira et al. [69] recently published their preclinical experience with the Trevo device in confirmatory animal models (2 swine, 1 canine) of arterial thrombo-occlusive disease induced by autologous thrombin-generated thrombi. Device performance was evaluated by angiographic response and clot incorporation by the device; high-resolution flat-panel three-dimensional CT was performed to further define the in vivo device-thrombus-vessel interaction. In addition, samples from vessels treated with 6 passes of the device were explanted for histopathological analysis. Sixteen clots of variable hardness and consistency were retrieved from the internal maxillary, lingual, and forelimb arteries (swine models) as well as from the external carotid and vertebral arteries (canine model). TIMI 2–3 reperfusion was achieved in all cases immediately after device deployment. One pass was required for 15 clots and 2 passes for the remaining clot. Histopathological analysis demonstrated severe disruption of the intima but no hemorrhage of the media or adventitia. The authors concluded that the Trevo device is highly effective at achieving immediate reperfusion of occluded arteries without causing any clinically significant disruption of vascular integrity.