Percutaneous balloon-assisted aspiration thrombectomy of clotted hemodialysis access grafts

Combination of duplex ultrasound-guided manual declotting and percutaneous transluminal angioplasty in thrombosed native dialysis fistulas

BACKGROUND

To investigate the safety, feasibility, efficacy, and long-term patency rate of manual declotting under duplex ultrasound (US) guidance followed by percutaneous transluminal angioplasty (PTA) in thrombosed native arteriovenous fistulas (AVFs).

METHODS

Of 87 consecutive thrombosed AVFs evaluated by duplex US, 22 patients with 25 recently thrombotic events in 22 AVFs were suitable for manual declotting. PTA was performed following successful declotting, and long-term patency was assessed.

RESULTS

The procedure success rate of manual declotting was 80% (20 of 25), and a residual stenosis of 74+/-9% was identified by duplex US after declotting. PTA reduced the diameter stenosis to 25+/-6% and increased the lumen diameter from 1.33 +/-0.85 mm to 4.62+/-0.98 mm. Neither embolic nor bleeding complications were noted during the procedure. The average procedure time and the fluoroscopy time were 28.4+/-9.9 and 7.2+/-4.1 minutes, respectively. Primary patency rates at 1, 2, and 3 years were 47%, 35%, and 28%; assisted primary patency rates at 1, 2, and 3 years were 71%, 63%, and 63%; and secondary patency rates at 1, 2, and 3 years were 76%, 71%, and 63%, respectively, during a maximum follow-up period of 42 months.

CONCLUSION

The combination of duplex US-guided manual declotting and angioplasty of underlying stenosis is a safe and feasible method to treat recently thrombosed native AVFs in selected patients. It simplifies the interventional procedure, reduces cost and radiation exposure time, and extends life span of dialysis fistula with acceptable long-term patency rate.

A new mechanical device for declotting of hemodialysis access grafts: first clinical experience

PURPOSE

To review the early clinical experience with a new mechanical thrombectomy device for declotting polytetrafluoroethylene (PTFE) hemodialysis access (HDA) grafts.

METHODS

The XTD is a 6-F introducer-compatible device that pulverizes clot with a rotating curved tip driven by a flexible spiral shaft. Fragments are aspirated into a collection container. Fifty patients (32 women; mean age 66 years) with clotted 6-mm polytetrafluoroethylene HDA grafts were treated at 2 medical centers in a study spanning 22 months. In all, 59 procedures were performed on an outpatient basis under monitored conscious sedation. Ancillary techniques were used as needed to treat the underlying cause of the clot. Technical success was defined on a per-procedure basis as removal of sufficient thrombus to visualize the underlying disease without major device-related complications.

RESULTS

Technical success was 100%; no device-related adverse events occurred. Immediate clinical success was achieved in 47/59 (80%) cases. Continued clinical success in survivors was 52% (30/58) at 1 month and 46% (26/56) at 3 months. Eliminating from analysis a high-risk patient subgroup (reocclusions, chronically clotted or unsalvageable grafts, untreatable central venous stenosis, and venous anastomoses resistant to wire/catheter passage) increased the 3-month clinical success to 59% (22/37).

CONCLUSIONS

The XTD is a promising device capable of safely and effectively declotting PTFE grafts. Further investigation is warranted.

Effectiveness and safety of dialysis vascular access procedures performed by interventional nephrologists

BACKGROUND

The purpose of this report was to analyze the results obtained from a group of interventional nephrologists working in multiple centers performing basic procedures that are used routinely in the management of vascular access problems, with an effort toward establishing standards for evaluating success, complication rates, and acceptable times for procedure duration and fluoroscopy.

METHODS

Data on six basic procedures were analyzed-angioplasty of arteriovenous fistulas (AVF-PTA), angioplasty of synthetic grafts (graft-PTA), thrombectomy of arteriovenous fistulas (AVF declot), thrombectomy of synthetic grafts (graft declot), placement of tunneled dialysis catheters (TDC placement), and tunneled dialysis catheter exchange (TDC exchange). These data were examined both as a group and by individual physician operator. RESULTS. A total of 14,067 cases were performed under the six categories of procedure that were the subject of this report; 13,503 cases (96.18%) were successful. The overall complication rate for the combined group of procedures was 3.54%, with 3.26% falling within the minor category and 0.28% within the major. The number of cases performed in each individual category with success rates for each were as follows: TDC placement-1765 cases, 98.24% successful; TDC exchange-2262 cases, 98.36% successful, AVF-PTA-1561 cases, 96.58% successful; graft-PTA-3560 cases, 98.06% successful; AVF declot-228 cases, 78.10% successful; graft declot-4671 cases, 93.08% successful.

CONCLUSION

This study demonstrates that appropriately trained interventional nephrologists can perform these basic procedures in both a safe and effective manner.

Percutaneous treatment of complications occurring during hemodialysis graft recanalization

INTRODUCTION/OBJECTIVE

To describe and evaluate percutaneous treatment methods of complications occurring during recanalization of thrombosed hemodialysis access grafts.

METHODS AND MATERIALS

A retrospective review of 579 thrombosed hemodialysis access grafts revealed 48 complications occurring during urokinase thrombolysis (512) or mechanical thrombectomy (67). These include 12 venous or venous anastomotic ruptures not controlled by balloon tamponade, eight arterial emboli, 12 graft extravasations, seven small hematomas, four intragraft pseudointimal 'dissections', two incidents of pulmonary edema, one episode of intestinal angina, one procedural death, and one distant hematoma.

RESULTS

Twelve cases of post angioplasty ruptures were treated with uncovered stents of which 10 resulted in graft salvage allowing successful hemodialysis. All arterial emboli were retrieved by Fogarty or embolectomy balloons. The 10/12 graft extravasations were successfully treated by digital compression while the procedure was completed and the graft flow was restored. Dissections were treated with prolonged Percutaneous Trasluminal Angioplasty (PTA) balloon inflation. Overall technical success was 39/48 (81%). Kaplan-Meier Primary and secondary patency rates were 72 and 78% at 30, 62 and 73% at 90 and 36 and 67% at 180 days, respectively. Secondary patency rates remained over 50% at 1 year. There were no additional complications caused by these maneuvers.

DISCUSSIONS AND CONCLUSION

The majority of complications occurring during percutaneous thrombolysis/thrombectomy of thrombosed access grafts, can be treated at the same sitting allowing completion of the recanalization procedure and usage of the same access for hemodialysis.

Low-power transverse ultrasonic treatment of portal vein thrombosis in an animal model

PURPOSE

The authors have recently developed a small-diameter, thin, flexible ultrasonic catheter device that permits the removal of thrombus by low-power transverse ultrasonic cavitation energy. In this study, the authors sought to determine whether this device could be used to eliminate portal vein thrombosis in an animal model.

MATERIALS AND METHODS

In five anesthetized pigs, a total of six occlusions of the left portal vein were achieved with use of autologous clot with (n = 2) or without (n = 4) thrombin injection (250 U) introduced via a 7-F transhepatic catheter/sheath system. Angiographic examination documented complete occlusion of this vessel. The 75-cm-long, 21-gauge ultrasonic catheter (Resolution) was introduced into the clot under angiographic guidance via the transhepatic sheath. Transverse-wave ultrasonic energy was then delivered from the distal 5 cm of the probe at 3.5 W +/- 10% power for up to 6 minutes. Repeat angiographic studies were performed to document patency. After the procedure, gross and histopathologic examinations were performed.

RESULTS

Restoration of patency of the main left portal vein was documented in all cases at angiography, with no evidence of residual clot fragments in the major branches. However, side branches demonstrated small thrombotic plugs on pathologic examination. No complications such as perforation of the vessel adjacent to the active ultrasonic tip were encountered. Virtually all thrombolysis was documented to occur within the first minute of energy application. At gross pathologic examination, there was no evidence of damage to the portal vein, and histopathologic examination demonstrated minimal intimal disruption without damage to the media.

CONCLUSIONS

This preliminary animal study suggests the feasibility of a percutaneous transhepatic approach to the treatment of portal vein thrombosis with use of low-power ultrasonic cavitation energy. With further study, this method may have potential for the treatment of thrombotic disease, thereby offering novel therapy to patients with thrombotic vascular occlusions.

Alteplase for hemodialysis access graft thrombolysis

PURPOSE

To evaluate the efficacy and safety of alteplase, a recombinant tissue plasminogen activator, in hemodialysis access graft thrombolysis.

MATERIALS AND METHODS

From November 1999 to May 2001, 68 episodes of occlusion in 50 grafts (in 49 patients) were included in the study. Occlusion was treated with pulse-spray (n = 41) or lyse-and-wait (n = 27) thrombolysis with use of alteplase. Balloon angioplasty of all identified stenoses was performed. The arterial plug was mobilized with the Fogarty maneuver.

RESULTS

Procedural success was achieved in 64 of 68 episodes (94%) with a dose of 2-10 mg (mean = 4.13 mg) of alteplase, allowing successful hemodialysis within 24 hours. Failures (6%) were the result of PTA perforation (one of 68), nonnegotiable outflow occlusion (one of 68), delayed bleeding (one of 68), and balloon bursting and shearing becoming occlusive within the graft (one of 68). Primary and secondary patency rates were 72% and 87% at 30 days, 57% and 80% at 90 days, and 44% and 72% at 180 days, respectively. Arterial emboli (two of 68) were treated by Fogarty balloon retrieval and alteplase infusion locally over the course of 20 minutes. One of two PTA perforations was controlled by balloon tamponade.

CONCLUSION

Alteplase can be used successfully for hemodialysis graft thrombolysis.

Thrombosed hemodialysis grafts: lyse and wait with tissue plasminogen activator or urokinase compared to mechanical thrombolysis with the Arrow-Trerotola Percutaneous Thrombolytic Device

PURPOSE

To determine if the lyse and wait (L&W) technique with a 4-mg dose of alteplase (tissue plasminogen activator; tPA) is a safe and effective method of declotting dialysis grafts as compared to use of the Arrow-Trerotola Percutaneous Thrombectomy Device (PTD) or the L&W technique with use of urokinase (UK).

MATERIALS AND METHODS

Forty patients were randomized prospectively to undergo L&W declotting with use of 4 mg of tPA or mechanical thrombolysis with the PTD. The time interval to restored graft flow, total procedure time, hemostasis time, and anatomic success, clinical success, complications, and patency rates were analyzed. These were compared with historic results in 20 patients treated with the L&W technique with use of 250,000 U UK.

RESULTS

The immediate anatomic success rate was 95% in the tPA L&W and PTD groups. The mean in-room lysis time with restored flow was 10 minutes for L&W with tPA and 19 minutes for PTD (P = .002). The mean in-room procedure time was 39 minutes for L&W and 45 minutes for PTD (P = NS). Mean hemostasis time with use of manual compression was 44 minutes for L&W with tPA and 23 minutes for PTD (P = .057). The historic group of 20 patients who underwent L&W with UK had a 95% anatomic success rate, a mean of 14 minutes of lysis time, a mean of 34 minutes of procedure time, and a mean of 26 minutes of time to hemostasis. No bleeding complications occurred in the PTD group. Seven episodes of bleeding occurred in six patients given tPA; four were delayed 60-90 minutes after the procedure, one necessitated hospitalization, and two required additional therapies. Four of the 20 patients undergoing L&W with UK had minor puncture site bleeding during the procedure. The 3-month primary patency rates were 65%, 65%, and 60% for L&W with tPA, PTD, and L&W with UK, respectively (P = NS).

CONCLUSION

The 4-mg dose of tPA is effective but results in more bleeding complications and longer hemostasis times than mechanical thrombolysis with use of the PTD. Unlike in our experience with UK, bleeding complications with tPA were both major and delayed.

Short-term patency and safety of an expanded polytetrafluoroethylene encapsulated endoluminal device at the venous anastomosis of a canine arteriovenous graft model

PURPOSE

To determine the safety and short-term patency of a polytetrafluoroethylene (PTFE)-encapsulated carbon-lined endoluminal device (ED) deployed across the venous anastomosis of arteriovenous conduits.

MATERIALS AND METHODS

Arteriovenous grafts (n = 16) were created between femoral arteries and veins in eight female canines and allowed to mature 30 days +/- 5 (SD). Five were excluded before implantation because of thrombosis or intragraft stenosis. Deployment was conducted in the remaining 11 anastomoses. Fistulography and intravascular ultrasound (IVUS) were performed before and after the procedure and 1 month postimplantation. Stent migration, apposition, and stenosis were evaluated. The angle of the anastomosis was compared before and after deployment and at follow-up. Mural thrombus thickness, percentage of surface covering, and percentage of endothelialization within the device were measured histologically.

RESULTS

There was no significant migration. By explant, all devices were completely apposed. Stenosis occurred in three of nine grafts. The angle of the venous anastomosis decreased by 29.5 degrees (posteroanterior) and 32.4 degrees (oblique) after ED deployment. There was a further decrease of 6.1 degrees (posteroanterior) and 3.2 degrees (oblique) during the 4-week follow-up period. Hemostasis was difficult to achieve in this animal model. Five required more than 1 hour to achieve hemostasis manually. Six in three animals were closed with a Perclose device, achieving immediate hemostasis; however, three (one in each animal) re-bled intermittently 2 weeks after implantation for an average of 9.3 days. The puncture site of each graft that bled was radiographically shown abnormal.

CONCLUSION

The ED can be deployed without stent migration and is completely apposed and patent after 4 weeks. Although bleeding was a problem with this animal model, delayed bleeding complications associated with puncture site abnormalities were seen only in grafts closed with a percutaneous suturing device.

Early rethrombosis of clotted hemodialysis grafts: graft salvage achieved with an aggressive approach

OBJECTIVE

The purpose of this study was to assess the efficacy of repeated pulsed spray pharmacomechanical thrombolysis for salvage of early rethrombosis of hemodialysis grafts and to identify factors that predict successful outcome.

MATERIALS AND METHODS

Thirty-four patients with initial successful thrombolysis were referred for repeated thrombolysis because of early rethrombosis. Repeated thrombolysis occurred within 1 month of initial thrombolysis. Technical success and patency rates were calculated. Causes of graft thrombosis and procedural modifications were analyzed.

RESULTS

The 39 rethrombosed grafts were successfully treated using pharmacomechanical thrombolysis, and patients underwent subsequent hemodialysis. The underlying flow-limiting stenoses were treated with balloon angioplasty using a larger balloon (41%), a same-size angioplasty balloon (18%), stent placement (15%), or increased anticoagulation (5%). A new stenosis location was discovered in 18%. Mean primary patency was 80.9 days (2.6 months) and secondary patency was 235.4 days (7.8 months). With life table analysis, 1-, 3-, 6-, and 12-month primary patency rates were 72%, 31%, 23%, and 15%, and secondary patency rates were 77%, 62%, 51%, and 31%, respectively. Graft patency rates in our study were compared with our institutional historic graft patency rates, with no significant difference noted (p = 0.76). No major procedural complications occurred.

CONCLUSION

Adequate technical success and patency rates for pharmacomechanical thrombolysis occur even for hemodialysis grafts that rethrombose within 1 month. After thrombolysis, aggressive search for and treatment of additional stenoses are warranted.

Recanalization of thrombosed dialysis access with use of a rotating mini-pigtail catheter: follow-up study

PURPOSE

To evaluate the feasibility, efficacy, and safety of mechanical thrombectomy of occluded dialysis access with use of a rotating mini-pigtail catheter.

MATERIALS AND METHODS

Thrombus was fragmented by mechanical action of the rotating pigtail tip (5-mm diameter), while the guide wire exited a sidehole at the pigtail curvature and served as a fixed rotation axis. Twenty-six procedures were performed in 22 patients (12 men, 10 women; mean age, 55.5 years). Native fistulas were treated in 15 instances, polytetrafluoroethylene (PTFE) grafts were treated in 11 instances. Average occlusion time was 20 hours +/- 13 (range, 5-46 hours), average occlusion length was 25.6 cm +/- 10.1 (range, 6-45 cm). Thrombus fragmentation was followed by balloon angioplasty of underlying stenoses.

RESULTS

In all 26 procedures, the dialysis access was successfully declotted with subsequent dialysis using the access (clinical success rate, 100%). Handling of the mini-pigtail catheter was simple and rapid, regardless of whether a graft or a native fistula was treated. Average duration of the intervention was 118 minutes +/- 30. Mean primary patency was 165 days +/- 167. Primary patency rate was 82% at 30 days, 65% at 3 months, and 47% at 6 months. There was no evidence of complications due to the thrombus fragmentation procedure.

CONCLUSION

The results suggest that declotting of occluded dialysis grafts and fistulas with the mini-pigtail catheter is as effective and safe as other more established percutaneous therapies. It may serve as an easy-to-handle, low-budget alternative to current thrombectomy devices.

Treatment of failed native arteriovenous fistulae for hemodialysis by interventional radiology

BACKGROUND

We studied the feasibility, technical problems, safety, and effectiveness of percutaneous declotting of thrombosed native arteriovenous fistulae for hemodialysis.

METHODS

Between 1992 and 1998, 93 declotting procedures were performed in 73 consecutive upper limb native fistulae (forearm 56 and upper arm 17), and 162 procedures were performed in 78 prosthetic grafts using manual catheter-directed thrombo-aspiration, with or without previous urokinase infusion. Detection of restenosis by clinical surveillance led to redilation or stent placement. Rethrombosis in four forearm and six upper arm fistulae were treated by 20 further declottings by aspiration.

RESULTS

The initial success was 93% in the forearm and 76% in the upper arm (99% in grafts). The complications included one pulmonary embolism, one acute pseudoaneurysm, and one blood depletion requiring transfusion. Primary patency rates at one year were 49% in the forearm and 9% in the upper arm (14% in grafts). Secondary patency rates were 81 and 50% at one year, respectively (83% in grafts). Reinterventions were necessary every 19.6 months in the forearm and every 5.7 months in the upper arm (every 6.4 months in grafts, P < 0.05). Stents were placed in 11% of forearm fistulae and in 41% of upper arm fistulae (45% of grafts) for treatment of acute rupture (5 out of 19), stenosis recoil (6 out of 19), and early (< 6 months) recurring stenosis (8 out of 19).

CONCLUSIONS

The percutaneous declotting of forearm fistulae by manual catheter-directed thrombo-aspiration was effective in more than 90% of cases and yielded 50% primary and 80% secondary patency rates at one year. The results were poorer in upper arm fistulae. The need for maintenance reinterventions was three times smaller in forearm fistulae than in upper arm fistulae and grafts.

Retrospective comparison of the Amplatz thrombectomy device with modified pulse-spray pharmacomechanical thrombolysis in the treatment of thrombosed hemodialysis access grafts

PURPOSE

To retrospectively evaluate the Amplatz thrombectomy device (ATD) in the treatment of thrombosed hemodialysis grafts and compare it with modified pulse-spray pharmacomechanical thrombolysis (PPT).

MATERIALS AND METHODS

During a 4-month period, 79 patients presented with 126 episodes of graft occlusion. Percutaneous recanalization was performed by using the ATD (n = 57) or the modified PPT technique (n = 69). Evaluation included the technical success, complications, and primary patency rates.

RESULTS

Technical success was achieved in 93% (53 to 57) of the cases treated with the ATD and in 96% (66 of 69) of the cases treated with modified PPT (P = .70). Complications occurred in 6% (four of 69) of modified PPT procedures and 16% (nine of 57) of ATD procedures. This difference was not statistically significant (P = .08); however, there were significantly more local complications in the ATD group (P = .04). The primary patency rates at 30, 90, and 180 days were 65% and 65%, 36% and 50%, and 26% and 33% for modified PPT and ATD, respectively. Survival curves were found not to differ significantly (P = .49).

CONCLUSION

The ATD and modified PPT were similarly successful in the recanalization of thrombosed hemodialysis access grafts and achieved comparable primary patency rates. The higher rate of local complications and technical difficulties encountered with use of the 8-F ATD limit its usefulness for this indication.

Mechanical thrombectomy in hemodialysis access shunts using a 5F pigtail rotation catheter. In vitro and in vivo investigations

RATIONALE AND OBJECTIVES

To evaluate the feasibility of mechanical thrombectomy in occluded hemodialysis access shunts by using a newly developed 5F pigtail rotation catheter.

METHODS

Thrombosed hemodialysis access shunts were simulated by clotted bovine blood in silicone tubing (diameter 6 mm). After retrograde and antegrade sheath placement (6F), mechanical fragmentation was performed using a 5F rotatable pigtail device. Average tube length was 27 to 47 cm; average thrombus weight was 5 to 11.9 g (8.2 +/- 1.59). Clinical application involved six patients with fresh shunt occlusions (three Brescia-Cimino shunts, three Gore-Tex shunts).

RESULTS

Using the in vitro setup, the device was able to restore a continuous lumen within 10 minutes with no remaining wall-adherent thrombi. The average amount of particles in the effluent was 3.0 g (2.0 to 3.9) for particles < or = 1.0 mm and 0.67 g (0.44 to 0.96) for particles > or = 0.2 mm wet weight; (compared with initial thrombus weight, 30.7% and 1.1%, respectively). Clinically, all six hemodialysis access shunts were successfully recanalized. Technical problems did not occur. There were no clinical symptoms indicating pulmonary embolism in any of the treated patients.

CONCLUSIONS

In our experimental setup as well as under clinical conditions, effective treatment of occluded hemodialysis access sites was achieved. The pigtail rotation device is an easy-to-handle, inexpensive alternative to mechanical thrombus fragmentation in occluded hemodialysis access shunts. The rate of emboli in the effluent vein of approximately one third of the initial thrombus weight must be taken into consideration in frequent intraindividual use of this technique.

Manual thromboaspiration and dilation of thrombosed dialysis access: mid-term results of a simple concept

PURPOSE

To report the feasibility, safety, and effectiveness of manual thromboaspiration as a single means of declotting dialysis access.

MATERIALS AND METHODS

Between April 1994 and December 1996, 59 consecutive conduits (43 polytetrafluoroethylene [PTFE] grafts, 16 native fistulas) were declotted with 8-F or 7-F angulated catheters. Unmasked stenoses were dilated. Clinical and paraclinical nephrologic surveillance (poor flow, palpation, difficulties with cannulation, increased compression times, increasing venous pressures) led to redilations and stent placements. Rethromboses were treated with further declotting by aspiration. The results are presented according to the life-table method.

RESULTS

The initial success of 43 of 43 PTFE grafts (mean procedure time, 119 min +/- 29 [standard deviation]) led to a primary patency rate of 85% +/- 5% (SE) at 1 month, 33% +/- 8% at 6 months, and 24% +/- 12% at 1 year. A graft was ligated 6 days after declotting for acute bleeding in one patient given high-dose warfarin. The secondary patency rates were 86% +/- 7% at 6 months and 86% +/- 9% at 1 year, with a mean duration of patency of 5.7 months between two radiologic interventions performed to maintain or to restore patency, and 19 stents were placed at a mean follow-up of 3 months. The success rate was 81% for native fistulas, with primary patency rates of 81% +/- 10% at 1 month, 74% +/- 14% at 6 months, and 60% +/- 27% at 1 year; secondary patency rates of 81% +/- 12% at 6 months and 81% +/- 18% at 1 year.

CONCLUSION

Thromboaspiration is a safe and effective method for declotting dialysis access, yielding a low rethrombosis rate during the first month. Overall radiologic management with reintervention on average every 6 months results in high secondary patency rates at 1 year (81%-86%).

Dialysis grafts arterial plug: retrieval using the tulip sheath device in vitro

The "arterial plug" is a resistant thrombus that frequently persists at the arterial anastomosis of clotted hemodialysis grafts following thrombolytic therapy. We studied the physical and morphological characteristics of the plug and determined the feasibility of transcatheter removal in vitro using the tulip compression thrombectomy system. Sixteen thrombus plugs were recovered during surgical thrombectomy of clotted human dialysis grafts. The physical and gross physical characteristics of all plugs were analyzed. Eight specimens were evaluated microscopically. Transcatheter compression thrombectomy of eight plugs was attempted in vitro. Each plug was embedded in a polyvinyl tube filled with newly clotted blood and connected to a flow circuit. First, balloon-assisted aspiration thrombectomy (BAT) of soft thrombus was performed, while sparing the distal-most segment containing the plug. The tulip sheath was then introduced facing the "arterial end" of the tube. The thrombus segment containing the plug was pulled back into the tulip mesh using either a 3 Fr Fogarty balloon catheter or a self-expanding rake. The tulip was closed to compress and remove the trapped plug. Near-complete thrombectomy of soft clot was achieved in all tested tubes. Compression and retrieval of the entire arterial plug was successful in all except one, where only partial compression of the plug occurred, presumably due to fibrotic changes. No fragmentation or embolization occurred in the remaining procedures. Spongy consistency was noted in 94% of the specimens. Microscopic evaluation showed organized layered thrombus with compaction in five plugs. Transcatheter removal of a thrombus plug is feasible in vitro using the tulip compression-thrombectomy system.