| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Developments |
1 From the Departments of Radiology (S.M.H., J.G., M.B., M.H.), Medical Physics (C.L.), and Medical Statistics (J.H.), University Hospital, Arnold-Heller-Strasse 9, 24105 Kiel, Germany. Received June 6, 2001; revision requested July 23; final revision received March 11, 2002; accepted March 25. Address correspondence to S.M.H. (e-mail: muehue@rad.uni-kiel.de).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
© RSNA, 2002
Index terms: Carotid arteries, stenosis or obstruction, 172.721, 904.721, 905.721 • Interventional procedures, complications • Interventional procedures, experimental studies, 172.1268, 904.1268, 905.1268
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
The concept of cerebral protection during carotid arterial stent placement to prevent cerebral embolization was first proposed by Theron et al (8,9). They initially used a homemade catheter system to perform temporary balloon occlusion of the distal internal carotid artery (ICA) during percutaneous transluminal angioplasty of the carotid artery. A system later became available that allowed balloon occlusion of the distal ICA during stent implantation in the carotid artery. At the time this article was written, the most commonly used protection devices were based on a filter or balloon blockade system. A filter maintains flow, whereas a balloon interrupts flow in the ICA for capturing emboli; before deflation, the balloon emboli have to be aspirated, and the filter has to be retrieved together with the captured emboli. However, to our knowledge, only preliminary experience and data regarding effectiveness exist, and the first clinical results are controversial (8–15).
The purposes of this study were to determine and compare the efficacy of the Angioguard (AG) and GuardWire (GW) cerebral protection systems and to evaluate whether prevention of central particle embolization can be enhanced by combining protection devices with additional aspiration techniques.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
To study the risk of peripheral embolism, the effluent was passed through polyethylene filters with a 100-µm mesh width (Schimmel, Nordheim, Germany) that were placed 100 cm behind the ECA and ICA. The weights of the particles and filters were determined prior to use. After completion of the procedure, the filters together with the captured particles were dried at 20°C for 4 days and weighed again. Each of the protection devices was tested separately.
Emboli
Commercially available polyvinyl alcohol (PVA) particles (Contur; Target, Fremont, Calif) suspended in 10 mL of saline were used to simulate dislodged particles. Particles weighed an average of 5.3 mg and were used in three sizes: 150–250 µm ("small," about 500 particles), 250–355 µm ("medium," about 200 particles), and 710–1,000 µm ("large," about 80 particles). After positioning the respective protection devices, embolization during carotid angioplasty was simulated by carefully injecting PVA particles into the ICA within a 30-second period by way of a multipurpose 8-F guiding catheter with a 0.088-inch inner diameter and 55-cm length (Cordis, Rhoden, the Netherlands). The guiding catheter was placed 1 cm in front of the protection devices, 4 cm from the bifurcation. After 5 minutes, with stent placement and final carotid angioplasty taken into account, additional aspiration techniques or device retrieval was performed. At the end of the intervention, flow was maintained for a further 30 seconds after removal of the protection device.
The following protocol was used to test the different protection devices in combination with the three PVA particle sizes: The occurrence of emboli was measured separately for each PVA particle size. First, small particles were injected, and the particles that passed the tested protection device were captured in the effluent filter of either the ICA or the ECA. The filters were then removed and the particles weighed. (The amount of particles captured is given as weight in milligrams and as a percentage of the weight of the originally injected particles.) The experiment was repeated 10 times, after which a different protection system was tested. With five possible systems and modifications (see subsequent Protection Devices section), 50 experiments were performed by using small particles. The experiment was then repeated with medium and large particles, such that 150 experiments were performed in total. Particles that passed the protection device and were captured by the ICA or ECA effluent filter were defined as emboli. The time from placement to removal (including particle removal) of the protection device and additional modification was measured and termed procedure time. It did not include particle delivery.
Protection Devices
The devices (Fig 2) were positioned into the ICA via the CCA by using a 9-F sheath and guiding catheter according to the manufacturers’ recommendations. Two systems were evaluated:
|
|
Once the guide wire is placed across the (hypothetical) lesion, the filter is expanded in an umbrella-like fashion by removing the 5-F outer deployment sheath in the arterial lumen. Following treatment of the lesion, the 5.1-F capture sheath is threaded over the proximal end of the guide wire, thereby closing the filter. After removal of the filter, the expanded basket was flushed with saline and the captured particles weighed by one of the authors (M.B.).
GW.—The GW (PercuSurge, Sunnyvale, Calif) (Fig 2b) is a temporary occlusion and aspiration system. It consists of a steerable 190-cm-long, 0.014-inch-diameter guide wire with a distal elastomeric lubricated occlusion balloon and a 3.5-cm-long flexible atraumatic tip. The balloon has a diameter of 3–6 mm, depending on the inflation volume, and was filled with a mixture of one part contrast medium (iopromide, Ultravist 300; Berlex, Wayne, NJ) and two parts heparinized saline. The volume of the balloon is controlled by using an inflator combined with a microseal adaptor (PercuSurge). An aspiration catheter that was 5.4 x 3.5 F at the tip and 4.6 F at the proximal shaft, with a monorail design (Export; PercuSurge), was used to aspirate debris three times with a 20-mL syringe prior to balloon deflation.
Once the protection device passed the lesion, the microseal adapter was attached to the guide wire for inflation of the occlusion balloon. After removal of the adapter with the balloon remaining inflated, the monorail aspiration catheter was advanced toward the lesion. When aspiration was completed, the occlusion balloon was deflated and removed.
Additional Aspiration Technique
To improve the aspiration of particles captured with either the occlusion balloon or the filter basket, a 6-F 100-cm–working length hydrodynamic catheter (AngioJet, Xpeedior 100; Possis, Minneapolis, Minn) was threaded over the wire of the protection device and then activated close to the protection device. The mechanism of the AngioJet catheter is based on debris aspiration following a recirculation effect that is caused by a negative pressure gradient in the vicinity of the catheter tip. The catheter is described in more detail elsewhere (16–18).
The second additional aspiration technique was aimed at increasing the effectiveness of the AG catheter. This was achieved by using the aspiration catheter, which was applied before closing the filter. The rationale was to aspirate particles from gaps between the wall of the silicone tube and the filter. Again, debris was aspirated three times by using a 20-mL syringe.
Statistical Analysis
Data are presented as means of particle weights ± SDs. A P value less than .01 was considered to indicate a significant difference. Statistical tests on the rates of embolization into the ICA and ECA (as well as their sum, to evaluate efficacy for prevention of embolization in general) and on procedure times were based on multivariate analyses of variance (MANOVA) (multivariate Wilks-
and univariate between-subject effects). In cases of significant differences, multiple comparison procedures, such as Scheffé and Games-Howell tests, were performed (BMPD; BMPD, Berkeley, Calif); these procedures were restricted to four comparisons to account for multiple testing on the same data set. The level of significance was set to 1% and did not require additional adjustment. Four groups were evaluated: AG versus GW, AG versus AGAJ, GW versus GWAJ, and AG versus AGAC, where AGAJ is AG plus AngioJet, GWAJ is GW plus AngioJet, and AGAC is AG plus the aspiration catheter.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Time for Protection, Device Placement, and Retrieval
The time for protection varied among the devices. The fastest was for AG (43 seconds ± 3), and the slowest was for GWAJ (207 seconds ± 6). GW (137 seconds ± 3) was significantly more time-consuming than AG (P < .001). Use of the modifications for the AG and GW devices prolonged the time for protection; AGAJ required 100 seconds ± 6, while additional aspiration with AGAC required 120 seconds ± 4.
Embolization and Debris Capture
None of the tested devices completely prevented embolization, with some of the particles getting past either the balloon or the basket. The mean weights of the captured particles in the AG basket, when used without modifications, were 4.86 mg ± 0.33 (small), 4.74 mg ± 0.31 (medium), and 4.90 mg ± 0.29 (large). Hence, a mean of 91% ± 6 of particles was captured. With additional aspiration, the captured particles were completely removed from the basket. The results for both the ICA and ECA are summarized in Table 1. According to MANOVA (Table 2), there were no statistically significant effects of particle size on the ICA and ECA (P = .259). There was, however, a significant effect from tested devices (P < .001).
|
|
ECA Findings
When capturing either small, medium, or large particles, the different protection devices performed equally. When the weights for all embolus sizes were added, significant differences were demonstrated for AG (0.48 mg, 3.0%) versus GW (1.2 mg, 7.6%, P < .001).
ICA and ECA Findings
With addition of the weights of the particles captured in the ICA and ECA filters, AG (1.28 mg, 8.0%) was more effective than GW (2.3 mg [14.6%], P < .001).
Statistical comparisons are summarized in Table 3.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
There is some controversy about the clinical importance of cerebral microemboli. It has not yet been determined whether the occurrence of cerebral emboli after carotid balloon angioplasty and stent placement is directly related to neurologic events. In a study of 14 patients undergoing balloon angioplasty and endarterectomy, Crawley et al (19) did not find a correlation between the number of transcranial emboli depicted with Doppler ultrasonography (US) and the periprocedural stroke rate (19). In spite of this uncertainty, it has to be assumed that fewer emboli may be beneficial to the patient. The lack of correlation between emboli and clinical sequelae has also led to the suggestion that a majority of emboli detected during balloon angioplasty are either gaseous or small platelet aggregates smaller than 200 µm in diameter, all of which correlate with a more benign outcome (19–21). Still, determination of the minimum size of emboli that need to be stopped to prevent stroke remains an active area of research.
Currently, the two main methods of embolus capture and prevention are balloon occlusion (eg, with the GW device) and temporary filter protection (eg, with the AG device). It was difficult to evaluate these devices in a controlled setting, as no animal models were available at the time this article was written. Thus, a flow model consisting of CCA, ECA, and ICA made from silicone tubing was constructed to simulate the capture of emboli during carotid angioplasty and stent placement. At the time of the experiments, only the GW and AG devices were available in Europe. To our knowledge, only limited clinical experience exists for the GW device (10,13), which has so far been used successfully during angioplasty and stent placement in diseased saphenous vein coronary bypass grafts (10). The AG system was evaluated in one recent study in patients during carotid arterial stent placement (22). Our current in vitro results indicate a benefit of the AG system over the GW system for reduction of PVA embolization into the ECA, accompanied by a significant reduction of time for device placement and retrieval.
Emboli during carotid angioplasty have been reported by a number of authors: Theron et al (11) analyzed aspirated blood after angioplasty under cerebral protection and found cholesterol crystals 600–1,300 µm in diameter in 17 of 21 cases (11); Martin et al (21) found cholesterol crystals 3.7–500.0 µm and lipoid masses 26–600 µm in diameter. It was therefore adequate to choose PVA particles 150–1,000 µm in diameter for our in vitro study. However, it should be mentioned that PVA articles are not necessarily of the sizes indicated by the manufacturer, as there are numerous smaller and some larger particles when suspended in saline (23). We will subsequently limit the discussion to "overall emboli," that is, the sum of the three groups of particle sizes.
The number of particles captured in the ICA and ECA filter was low for all tested devices and modifications. The percentage range of particles that got past the separately tested protection devices was from 3.2% (AGAC) to 7.0% (GW) in the ICA and from 3.1% (AGAC) to 7.6% (GW) in the ECA, a total of 6.3% (AGAC) to 14.6% (GW). For all particle sizes, the combination of filter basket and final aspiration seemed to be effective, with the most apparent difference of embolization into the ECA with use of a filter basket (AG, 3.0%) and balloon device (GW, 7.6%; P < .001). All of the effluent is directed to the ECA, except the volume in the ICA proximal to the balloon occlusion. In addition, increased ECA territory embolization, with the disadvantage of greater risk of complications from ECA to ICA collateral vessels, might be explained by the fact that balloon occlusion causes turbulence in the occluded ICA segment. In theory, balloon occlusion or filtration closer to the bifurcation would result in a smaller ICA volume containing emboli and possibly in even more embolization in the ECA.
However, our modifications of the protection devices did not improve their efficacy. Specifically, the additional use of the AngioJet as a modification to the AG did not result in a reduction of particles caught in the filters. This may be explained by the turbulent flow created by the AngioJet, which could have been responsible for embolus mobilization from the ICA into the ECA. Furthermore, one might even postulate that a collapsing filter basket was a result of the negative pressure zone around the tip of the AngioJet. Using the AngioJet together with the GW device did not improve efficacy in the ICA or ECA (P = .039).
The overall percentage of particles that got past the protection devices without modifications is similar to that in the Ohki et al study (24). Contrary to the model used by these researchers (real plaque, steady flow), our model made from elastic silicone tubes was driven by using a peristaltic pump, thus producing a physiologically more relevant pulsatile flow. Testing a NeuroShield filter device (MedNova, Horsham, UK) by using a human carotid arterial stent placement model, Ohki et al (24) found that 88% of particles 360–1,000 µm in diameter were captured. Carotid arterial plaques obtained from carotid endarterectomy were used to mimic the scenario in patients as much as possible. The data of Ohki et al (24) suggest that while protection devices may reduce emboli, absolute prevention is not possible. These observations correlate with findings of two embolic complications with use of temporary balloon occlusion in patients (25). In the first case, flushing debris into the ECA resulted in monocular blindness due to an unrecognized connection between the middle meningeal and ophthalmic arteries; in the second case, flushing resulted in proximal reflux of debris from the right common carotid artery into the right vertebral artery (25). Henry et al (15), using the Theron et al (25) distal balloon technique in 32 of 163 cases during carotid arterial stent placement, observed that two of three major strokes occurred in conjunction with use of this protection device (15). It should be pointed out that anomalous or retrograde supply to the ophthalmic artery is not common; still, the surgeon should be aware of this possibility, especially when treating a high-grade stenosis in the presence of an occluded contralateral ICA. In the presented model, aspiration alone was performed for removal of emboli. During our testing, no flushing was performed to prevent embolism into the ECA or aortic arch. This could occur in vivo, as there may be anastomoses of anomalous origin between the ECA and ophthalmic artery (12,26).
In another clinical study (27), interruption of blood flow in the ICA led to complications in 3.2% of patients undergoing balloon occlusion. In addition, the method of balloon protection might be inadvisable in patients with contralateral ICA occlusion or a diseased circle of Willis and poor collateral circulation (5). In these patients, an umbrella-like filtration device allowing residual blood flow should be considered instead; however, the possibility of hemolysis might be a disadvantage of such filter devices. In general, with use of either system, damage to the intima due to unexpected movement of the inflated balloon or deployed filter basket might occur.
Balloon or filter positioning, balloon inflation, debris flushing or aspiration, and balloon deflation may take longer in the clinical environment than in our model system because of factors such as patient compliance, anatomy, degree of stenosis, et cetera. Thus, the time required for in vivo intervention will always be longer than that in our in vitro experiments, such that results obtained for the time of protection in the model system should be extrapolated to the clinical situation with care.
Nevertheless, it is reasonable to mention the in vitro times, as they indicate the ease of handling of the devices. The shortest mean procedure time was achieved for AG (43 seconds) and the longest for GW with AngioJet (204 seconds); in clinical practice, about 120 seconds should be added for stent placement and angioplasty. These figures would still be lower than the mean in vivo occlusion time of 542 seconds ± 243 (53 treated ICA lesions) with use of the GW system (15).
There are a few shortcomings associated with the design of the model system. First, it was made from silicone tubing, such that possible endothelial damage may not be evaluated. In addition, the model did not include any lesions, such that embolization during catheter passage past a lesion, the most common complication during the procedure (24), did not occur.
One problem that may occur during percutaneous transluminal angioplasty and/or stent placement is embolization caused by debris hidden in the crevices between the balloon and lumen or between the filter and lumen. In our model, when performing balloon occlusion, we observed a small number of particles in the ICA (GW, 1.1%; GWAJ, 0.8%) that were dislodged after balloon deflation. However, in in vivo studies (11,14), embolization due to trapped debris in the crevices was not observed. The reason for this may be the different properties of the vessel wall in vivo and the silicone tubing in vitro—for instance, a difference in the elastic properties.
This simple model offers a method to reproducibly test the efficacy of protection devices for capture and/or retrieval of embolized particles under pulsatile flow conditions; the acquired data indicate the performance of the various protection devices that may be expected in vivo. Additional aspects that are difficult to realize in vitro may have to be considered for future in vivo studies: For instance, the composition of emboli is completely different from that of PVA particles, and the behavior of a native calcified and stenosed vessel differs from that of a silicone tube. Furthermore, the number of PVA particles used in our model greatly exceeded the amount of debris dislodged in a typical in vivo setting. Thus, while this study clearly demonstrates differences in the performance of the various devices, these results should be extrapolated to the clinical situation with care.
In conclusion, none of the devices and modifications can completely prevent embolization in vitro. No conclusion can be drawn about the superiority of the AG or GW cerebral protection devices on the basis of embolization into the ICA; however, in the ECA, a greater percentage of particles was found when using the GW device, as compared with that when using the AG system. Further investigation in vitro in an improved model and in vivo in patients is required to evaluate the devices further.
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, S.M.H., J.G.; study concepts and design, S.M.H.; literature research, S.M.H., J.G., M.B.; experimental studies, S.M.H., M.B.; data acquisition and analysis/interpretation, S.M.H., M.B.; statistical analysis, S.M.H., C.L., J.H.; manuscript preparation, S.M.H., C.L.; manuscript definition of intellectual content, S.M.H., C.L., J.H.; manuscript editing, S.M.H.; manuscript revision/review, all authors; manuscript final version approval, S.M.H., M.H.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. Zahn, T. Ischinger, B. Mark, S. Gass, U. Zeymer, W. Schmalz, K. Haerten, K. E. Hauptmann, E.-R. von Leitner, W. Kasper, et al. Embolic Protection Devices for Carotid Artery Stenting: Is There a Difference Between Filter and Distal Occlusive Devices? J. Am. Coll. Cardiol., June 7, 2005; 45(11): 1769 - 1774. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Muller-Hulsbeck, P. Stolzmann, C. Liess, J. Hedderich, F. Paulsen, T. Jahnke, and M. Heller Vessel Wall Damage Caused by Cerebral Protection Devices: Ex Vivo Evaluation in Porcine Carotid Arteries Radiology, May 1, 2005; 235(2): 454 - 460. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Taylor, F. Alcocer, and W. D. Jordan Jr Controversies in Carotid Stenting Vascular and Endovascular Surgery, March 1, 2003; 37(2): 79 - 87. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
