Design of an Embolic Protection Filter for Carotid Artery ...



Redesign of a Distal Protection Filter

for Carotid Artery Angioplasty

By:

Sandeep Devabhakthuni

Chenara Johnson

Daphne Kontos

Perry Tiberio

Professor Mark Garter

Bioeng 1160 Bioengineering Design

26 April 2004

Introduction

The first percutaneous carotid angioplasty was performed in 1980, and since then, there has been a continuous influx of new interventional technologies1. The main safety issue arising from this procedure is the risk of cerebral embolization. This occurs when the angioplasty balloon is inflated to clear the stenosis, resulting in the release of debris into the blood stream. This debris (emboli) travels to the brain, potentially causing a stroke. Hence a new technology was developed to increase the safety of the procedure, the neuroprotection device.

Neuroprotection devices are designed to capture and remove emboli released during carotid angioplasty. There are two main types: balloon occlusion devices and distal protection filters (DPF). Since the balloon occlusion inhibits blood flow during the procedure and does not allow for continuous angiographic feedback, the distal protection filter is favored. The filters permit continuous blood flow during the procedure while still providing distal cerebral protection. The main weaknesses of the filters, however, are their pore size. Since the blood is composed of particles ranging from 2 to 120 µm2, determining the most efficient pore size is critical to the design of a distal protection filter. Currently, the smallest pore diameter in filter devices is about 40 µm, with the majority of filters having an 80 to 100 µm pore diameter. The largest pore diameter is about 190 µm3. In addition to pore size, an inadequate fit to the artery allows emboli to pass the filter.

The three filters that minimize these complications are the FilterWire EX (Boston Scientific, Natick, MA, USA), the AccuNet (Guidant Corporation, Indianapolis, IN) and the AngioGuard (Cordis Corporation, Miami Lakes, FL). These filters are considered to be the three best distal protection filters.

The FilterWire EX (see Fig. 1) polyurethane filter has one of the largest filter volumes (15 mm long), allowing it to capture more embolic particles. The 110µm-pore filter bag is aligned at an angle to the vessel wall. The main benefits of this simple design are its filter volume and large proximal opening. Also, its nitinol ring support allows the filter to mold against the vessels’ topography. Some complications with this design include incorrect placement in the vessel due to the filter’s angled orientation in the vessel.

[pic]

Figure 1: FilterWire EX (Boston Scientific, Natick, MA)5

The AccuNet (see Fig. 2) polyurethane filter has a tubular-shape consisting of 4 pre-shaped nitinol wire struts with one opening at the proximal end to allow perfusion. Just like the FilterWire EX, the AccuNet also has a large filter volume (20 mm in length) allowing for maximum capture. The 115µm-pore filter bag is deployed perpendicular to the vessel and its nitinol struts expand a lie flush to the vessel wall.

[pic]

Figure 2: The AccuNet with a centrally mounted filter3

The Angioguard (see Fig. 3) has a hemispherical polyurethane filter basket supported by 8 pre-shaped nitinol wire struts. The filter has one large proximal opening but has a small filter volume, unlike FilterWire EX and AccuNet. The 100µm-pore filter bag is deployed perpendicular to the vessel. Because of the filter’s shallow basket, embolic particles overfill the basket and may escape into the vessel lumen during retrieval.

[pic]

Figure 2: The AngioGuard (Cordis Corporation, Miami Lakes, FL)3

Based on the research on these current filters, our goal is to redesign the distal embolic protection filter device to be more efficient in capturing embolic particles by decreasing the pore size and adding a skirt to prevent passage of emboli. Furthermore, we propose that a nylon/nitinol filter material be used in our design to maintain the mechanical properties, while retaining viability of these two materials. This will be tested through comparison of results of in-vitro studies that mimic in vivo conditions under which the device will be operated. When the experiment is performed and the analytical results are calculated, we hope to determine that the redesign of the distal embolic protection filter is providing protection at a greater efficiency than other filters, is no more difficult to use than other filters, will be used in future studies, and be marketed as the top distal protection device.

Methods

Through analyzing previous designs, we determined that the optimal pore size should be between 70 and 80 µm. As shown in Figure 3, the redesign of the distal protection filter included a radio plaque nitinol elliptical ring with the 8.51mm being the short diameter. It had a cylindrical filter bag (28mm long) mounted on the side. The nitinol ring was supported by one nitinol strut in order to support the filter and position it at a 20° angle to the vessel. The filter had a nylon skirt (roughly 0.1 mm) extending outward from the nitinol ring. The filter will be delivered by pushing the strut to collapse the nitinol ring and deployed by releasing the strut. The ring will default to the manufactured 20° angle.

[pic]

Figure 3: Picture of Redesigned Distal Protection Filter

This preliminary design provided all the safety precautions necessary for embolic protection during the procedure. The filter was redesigned using the SolidWorks program. The SolidWorks files were then used as a guideline for the fabrication of a testable prototype. To build the model above, the following materials were required: nylon for the basket (Sefar Holding Inc, Freibach, Thailand), nitinol, polyimide tubing and stainless steel for the frame (New England Precision Guiding Inc, Holliston, MA.).

Before testing the new filter model, the flow loop used in previous studies was redesigned in order to obtain more efficient results7. The flow loop previously used utilized continuous flow. The new flow pump induced peristaltic movements, simulating vascular smooth muscle contractions.

Figure 4: Insertion points for catheter8

The flow loop length represented the path from the groin to the carotid artery (Figure 4). Also, the flow rate, in addition to being peristaltic, was set to 150 mL/min which was the approximate flow rate in the internal carotid artery. The material of the flow loop was changed from silicon to vinyl, a more flexible and compliant material, that more accurately imitated the arterial walls. Changing the in vitro set up to model in vivo conditions provided more accurate results in testing. Using this modified flow loop design, the efficiency of the filter redesign was examined and compared to the previously tested filters: AngioGuard, AccuNet and FilterWire.

Silica artificial embolic microspheres (Bangs Laboratory, Fischers, IN) were premanufactured to simulate the emboli released during angioplasty. Collection of the particles took place in an inline filter distal to the protection filter deployment point. The size of the microspheres ranged from 150µm to 300µm, and an average weight of about 85mg was used in each trial. Weight was measured using an analytical and toploading balance by Fisher Science Education, Hanover Park, IL. The microspheres were used to establish capture efficiency by quantifying the amount of emboli captured while the filter was in place. The fluid used to model blood in the flow loop was a solution consisting of glycerol and water at a ratio of 9:16 resulting in a viscosity of 3.5 cP in order to have a viscosity similar to blood.

After the prototype was manufactured and the in vitro flow loop was completed, training was implemented for the use of the delivery of the filter and use of the flow loop. If improperly used, the results would have been poor due to human error. In addition, the placement of the filter was crucial in gaining precise results. Therefore, five trial runs with the filter in the flow loop were performed before the actual experiment is conducted.

After proper training of filter use, the new filter design was tested and results were compared to previously tested filters to determine if the new design was more efficient in capturing the embolic particles and if it was a competitive new product. Ten subsequent runs, with only 5 successful trial runs, were used for analysis. Collection of the embolic microspheres occurred at one point within the filter during each trial. While the filter was deployed within the vessel walls, the in-line filter collected the missed particles to determine efficiency while deployed. The in-line filter had a pore size of less than 100µm.

[pic]

Figure 5. Redesigned flow loop schematic

The flow loop is detailed in Figure 5. Per trial, 28 liters of liquid (10 L of glycerol and 18 L of water) originated in the reservoir and was pumped through the circuit back to the same reservoir. Before the flow pump was turned on, the filter was inserted through the one-way valve and deployed. Then, the peristaltic flow pump (ISMATEC, Glattbrugg, Switzerland) was set to 150 mL/min. Once the flow was stabilized, microspheres that functioned as embolic particles were injected through a one-way valve to allow for continuous flow for each trial. After the microspheres were inserted, the trial ran for five minutes before the flow pump was shut off. Once the filter was retrieved, the mass of the particles removed from the filter was weighed. This completed the use of the flow loop and the process was repeated for the next trial

Once this experiment was completed, data analysis was performed. To analyze the data, the percentage device efficiency was calculated and compared with the other previously tested filters. The overall device efficiency will be determined using Equation 1.

[pic] [1]

In addition to overall device efficiency, statistical analysis including ANOVA and paired t-tests (assuming normality) determined the significance of the data obtained from the experiments using a p-value of 0.05. ANOVA was used for between-group comparison (i.e. how the new design compares with the AngioGuard, AccuNet, and Filterwire), and the paired t-test was used to determine in-group differences, (e.g. the amount of emboli captured vs. total amount of emboli). ezANOVA, a computerized analysis program, was used.

Results

After each trial was completed, the mass of the particles captured by the filter was weighed. Using Equation 1, the efficiency was determined for each of the trials. This efficiency represented the number of embolic particles that the filter was able to capture relative to the number of embolic particles injected. The results of these calculations were are tabulated in Table A.1 (Appendix A). Next, the device efficiencies of all ten trials were plotted with error bars, depicting the variation between trials. Shown in Figure 6, the graphical representation of the efficiencies indicated a significant amount of variation.

[pic]

Figure 6. Graphical representation of the capture efficiency for all trials

As Figure 6 shows, there were data points that were significantly low. These data points were determined to be a result of device or human error and were excluded. Out of then ten trials that were performed, five trials resulted in significantly low efficiencies due to device failure. Because the results obtained for these trials were not accurate, these results were excluded in order to determine the overall efficiency of the device more accurately. As shown in Figure 7, the modified data was plotted for the five successful trials. The plot indicates that the capture efficiency decreased as the filter use increased.

[pic]

Figure 7. Graphical representation using the modified data for capture efficiency

Once the overall device efficiency was calculated for each of these trials, the efficiency of our filter was then compared to the efficiency of predicate devices: AngioGuard, AccuNet, and FilterWire. Data for each of these three filters collected using an old flow loop tested in previous studies was used for the comparison7. The overall device efficiency for each of these filters was calculated and the results of these calculations are shown in Table A.2 (Appendix A). The average and standard deviation was calculated for all of the filters in order to make the comparison. Table 3 showed the comparison of the device efficiency of our filter with the average device efficiency of the other filters. As the table indicates, our filter was slightly lower in the overall device efficiency with a higher standard deviation as compared to the other filters.

Table 3. Comparison of Overall Device Efficiency for the Four Filters

|  |AngioGuard |AccuNet |FilterWire |Our Filter |

|Capture Efficiency |96.81+4.78 |97.73+3.94 |90.86+12.57 |82.8+14.5 |

However, when ezANOVA was used to analyze the variations between the four filters, the analysis revealed that there was no significant difference between our filter and each of the predicate filters (p > 0.05 for all the comparisons between our filter and each of the predicate filters). In addition, a paired t-test using Microsoft Excel® was performed for our filter in order to compare the initial weight of particles and the final weight of the particles that were captured in the filter. This test indicated that there was no significant difference between these two measurements (p < 0.0784).

Discussion

Although the overall device efficiency of our filter was comparable to the efficiencies of the filters on the market, the insignificant difference between all four filters indicated that our filter did not have a significant advantage over the other filters. However, this conclusion may be inaccurate because of the variance associated with our filter’s performance and because of several problems that occurred during the testing of the filter. First, the filter was designed to be used only once. This accounts for the highest efficiency that occurred during the first trial. After the first trial, there was a significant decrease in the efficiency as the filter use increased. Second, there were unexpected failures that arose as the testing proceeded. For example, the nitinol ring/polyimide tubing cracked in one of the later trials, which made it difficult for the filter to capture emboli effectively. In addition, the back end of the filter tore, causing a gap between the vessel lumen and the filter. This gap allowed the emboli particles to escape the filter, resulting in lower capture efficiency.

The reason for these unforeseen failures was due to the human error involved in the testing and device failure. As Figure 6 indicated, there was a significant amount of variation when comparing the efficiencies for all ten trials. This significant amount was considered to be caused by excessive user error during some of the trials. Because the error was preventable, the data for these trails were not relevant in determining the overall device efficiency of the filter. As a result, the data for the five unsuccessful trials were excluded, and the device efficiency of the filter was determined by the modified data shown in Figure 7.

In addition to these unexpected occurrences, there were other possible limitations that accounted for poor results. For instance, the placement of the filter was not consistent for all the trials. After each trial was performed, the flow loop had to be disassembled and the filter had to be removed in order to retrieve the captured particles for weighing. Once the filter was inserted again and the flow loop was reassembled, the filter placement was changed. Therefore, this inconsistent filter placement may have affected results. Moreover, when the filter was being retrieved, the filter was not fully collapsible. This made it easier for the particles to escape the filter. Since this retrieval mechanism was not efficient, this resulted in lower capture efficiency. With that stated, our filter also suffered from mechanical failures associated with manufacturing techniques. Consequently, these limitations generated poor results that did not accurately assess the device efficiency of our filter.

Due to these limitations and user error during testing, several modifications to the design were recommended for higher capture efficiency. First, a more manageable retrieval mechanism is needed in order to minimize the loss of emboli particles when the filter is in movement. A possible improvement would be to collapse the filter completely, which would minimize the loss of particles. Second, instead of using nylon as the material for the filter basket, using polyurethane would be more efficient because polyurethane is more elastic leading to greater flexibility. This material property would be beneficial because it can aid in the capture of the particles and prevent loss of particles as the filter is being retrieved. Finally, using nitinol tubing for the ring would allow more flexibility for the filter, which would allow the filter to collapse easily. While the current materials used are all biocompatible as demonstrated by their use in other medical devices that contact blood, using materials that are seen in the predicate devices would give our filter the necessary mechanical properties. These modifications to the design would allow the filter to be more efficient in both capturing the particles and retrieving the particles.

References

1. Liistro, Di Mario. Carotid Artery Stenting. Heart 2003; 89:944-948

2. Freitas RA. Nanomedicine, Volume I: Basic Capabilities. Landes Bioscience, Georgetown,

TX, 1999.

3. Kasirajan K, Schneider P, and Kent K. “Filter devices for cerebral protection during

carotid angioplasty and stenting.” Journal of Endovascular Therapy. 2003;10:1039-

1045.

4. Ohki, Takao. The Dark Side of Embolic Protection Devices. Endovascular Today

September 2003.

5. Fasseas P, Orford J, Denkta A, Berger P. “Distal protection devices during percutaneous

coronary carotid interventions.” Current Controlled Trials in Cardiovascular Medicine. 2001;2:286-291.

6. Reimers B, Corvaja N, Moshiri S, Sacca S, Albiero R, Di Mario C, Pascotto P, Colombo A.

“Cerebral Protection With Filter Devices During Carotid Artery Stenting.” Circulation. 2001;104:12.

7. Azemi E, Gray H, Miner E. “Comparison and Redesign of Distal Protection Filters for

Angioplasty Stenting Procedures.” University of Pittsburgh. 2003.

8. FreePatentsOnline. 14 Dec 04.

9. VascularWeb. 14 Dec 04.

Appendix A: Data Collection

Table 1. Measurements of the Weight of the Particles Before and After Trials

|  |Initial Weight (mg) |Filter Particle Weight (mg) |

|Trial 1 |79.6 |78.1 |

|Trial 2 |84.7 |37.4 |

|Trial 3 |92.7 |42.6 |

|Trial 4 |85.4 |82.2 |

|Trial 5 |87.3 |28.1 |

|Trial 6 |84.7 |22.7 |

|Trial 7 |88.0 |70.4 |

|Trial 8 |108.4 |83.0 |

|Trial 9 |80.7 |15.6 |

|Trial 10 |82.0 |51.9 |

Table 2. Overall Device Efficiencies for Three Filters Currently on the Market

|  |AngioGuard |AccuNet |FilterWire |

|Trial 1 |100 |100 |100 |

|Trial 2 |100 |100 |75.47 |

|Trial 3 |89.29 |97.73 |100 |

|Trial 4 |94.74 |100 |78.85 |

|Trial 5 |100 |90.91 |100 |

-----------------------

Inline filter

Glycerin/Water solution (9:16)

Point of Filter Deployment

Insertion Point of Filter (one-way valve)

Length = 20 cm

Peristaltic Flow Pump

Reservoir

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download