New Gianturco-Grifka Vascular Occlusion Device
Initial Studies in a Canine Model
Background Transcatheter closure of cardiovascular defects remains a challenge. Several occlusion devices are available, but each device has limitations. The purpose of this study was to evaluate the new Gianturco-Grifka vascular occlusion device (GGVOD) in a canine model.
Methods and Results A total of 26 GGVODs were implanted as part of short- and long-term studies. In the short-term study, 1 GGVOD was implanted in each of 11 systemic arteries from 3.2 to 9.0 mm in diameter. All 11 arteries were occluded immediately. In the long-term study, an aortopulmonary shunt was placed in 10 dogs (9, Gore-tex graft; 1, subclavian artery) followed by GGVOD implantation; additionally, a GGVOD was implanted in 5 subclavian arteries. The dogs were boarded for 3 to 6 months, then recatheterized and euthanatized. Immediately after implantation, the 5 subclavian arteries and 9 Gore-tex shunts were occluded completely; the 1 subclavian artery shunt had a small residual leak. At recatheterization, all 10 shunts and 5 subclavian arteries were occluded completely. Necropsy revealed all shunts to be occluded, with the aortic and pulmonic orifices covered with a neointimal layer. The mean fluoroscopic time needed for GGVOD implantation was 9 minutes (range, 3 to 22 minutes).
Conclusions (1) In a canine model, the GGVOD is effective for transcatheter occlusion of arteries and aortopulmonary shunts from 3 to 9 mm in diameter. Possible indications in children include aortopulmonary collateral vessels, long patent ductus arteriosus, systemic-pulmonary shunts, AV malformations, and arteries supplying tumors. (2) GGVOD implantation requires a short fluoroscopic time.
Transcatheter occlusion of congenital heart defects has become an increasingly important therapeutic option.1 2 Devices have been developed for the transcatheter occlusion of several defects, including patent ductus arteriosus (PDA), atrial septal defects, and aortopulmonary collateral vessels.3 4 5 6 7 8 However, these occlusion devices have several limitations, including complicated delivery systems, device embolizations, and residual leaks. Therefore, improved occlusion devices are a welcome addition to the transcatheter armamentarium. The purpose of this study was to develop and evaluate a new transcatheter vascular occlusion device that will overcome many of the limitations of the present devices.
The Gianturco-Grifka vascular occlusion device (GGVOD) consists of a nylon sack attached to an end-hole catheter. A modified spring guide wire is advanced through the end-hole catheter and into the sack. Once inside the sack, the wire coils, which expands the sack, occluding the vessel. Then, the coil-filled sack is released from the catheter.
The sack is constructed from two pieces of tightly woven nylon. The edges of the nylon pieces are thermally sealed, affording a circular sack enclosure. The sack is attached to the sack catheter, which is a 4.5F end-hole catheter with an everted flare on the distal tip (Fig 1⇓). The sack fits tightly over the flared catheter tip and is secured by a radiopaque metal tie string. A 5.5F release catheter is used to push the sack off the sack catheter. The sack is delivered to the vessel through a modified 8F 70-cm-long delivery sheath.
Two separate wires are used in the device, a floppy filler wire and a stiff pusher wire (Fig 2⇓). The filler wire is a standard 0.025-in spring guide wire with three modifications: the stiff inner core removed, internal threads on the proximal end, and a J-curve on the distal tip. The filler wire internal threads screw onto the external threads of the pusher wire. The pusher wire forces the filler wire through the sack catheter into the sack. Once inside the sack, the filler wire coils, which serves two functions: (1) filling the sack, thus occluding the vessel lumen, and (2) providing transmural pressure to maintain the sack position in the vessel. To remove the pusher wire, its crank handle is rotated counterclockwise, which unscrews and separates it from the filler wire.
At this point, the sack is filled with the filler wire, but the sack remains attached to the sack catheter. To release the sack, the release catheter is advanced up against the sack and held firmly in position (Fig 3⇓). The sack catheter is withdrawn firmly (into the fixed release catheter), which inverts the flared tip, allowing the catheter to pull out of the sack.
The GGVOD design allows for repositioning of the device before release. After the filler wire is inserted and before the pusher wire is unscrewed, an angiogram can be performed to assess device position and vessel occlusion. If needed, the filler wire can be pulled out of the sack, the device repositioned, and the filler wire reinserted into the sack. The filler wire can be removed and reinserted numerous times until the appropriate device position is obtained. If the GGVOD does not occlude the vessel, the wire can be pulled out of the sack and the sack pulled carefully back into the sheath and replaced with a different GGVOD (the same size, larger, or smaller) inserted through the same sheath.
To have devices that will allow a wide range of blood vessel diameters to be occluded, we evaluated sacks varying in diameter from 3 to 9 mm. Initially, the sack was manufactured in three sizes: 3-, 6-, and 9-mm diameter. However, this resulted in a range of vessel diameters too large for each size GGVOD. Currently, the GGVOD sack is manufactured in four sizes: 3-, 5-, 7-, and 9-mm diameter; all four GGVOD sizes are delivered through the same 8F sheath. Sacks of each size have filler wires of appropriate lengths to completely fill the sack. To provide sufficient transmural pressure, the sack diameter should be 1.0 to 1.5 mm larger than the diameter of the blood vessel to be occluded.
A short-term study was performed in a canine model using mongrel dogs weighing from 22 to 32 kg. All studies were performed according to institutional guidelines and the principles of the American Physiological Society. In 6 dogs, GGVODs were implanted in various subclavian, carotid, and renal arteries ranging from 3 to 9 mm in diameter. Thirty minutes before surgery, each dog received an intramuscular injection (acepromazine 0.5 mg/kg and atropine 0.05 mg/kg), had a peripheral intravenous line placed, then received general anesthesia (isoflurane 0.5% to 4%, nitrous oxide 5% to 50%, oxygen 25% to 50%) and was intubated. The femoral areas were shaved, prepared, and draped. By percutaneous technique, a 7F introducer sheath with hemostasis valve (Argon Inc) was inserted into the femoral artery. A 7F NIH Cardiomarker catheter (USCI Inc) was advanced through the sheath into the aorta. In the lateral projection, angiograms were obtained in the subclavian, carotid, and renal arteries. The artery diameter was calculated from the 1-cm calibration marks on the Cardiomarker catheter. The NIH catheter was removed and replaced with an end-hole catheter (Cook Inc), which was advanced into the corresponding subclavian, carotid, or renal artery. A 0.038-in stiff exchange wire (Cook Inc) was inserted through the end-hole catheter into the artery distal to the measured area. The catheter and sheath were removed, with wire position maintained. The modified 8F long sheath and dilator (Cook Inc) were inserted over the exchange wire into the artery and advanced to the measured area. The dilator and wire were removed. Contrast medium was injected through the sheath to confirm position. The GGVOD sack (Cook Inc) was inserted into the long sheath and advanced out the distal end. The (empty) sack was positioned in the artery at the desired point of occlusion. The pusher wire was advanced into the sack catheter, filling the sack with the filler wire. After the pusher wire was inserted completely, the crank handle was rotated counterclockwise, and the pusher wire was removed. The release catheter was advanced up to the sack, then the sack catheter was pulled firmly, which released the sack. To confirm vessel occlusion, the sheath was withdrawn 20 mm and used to perform an angiogram. If a residual leak occurred, a repeat angiogram was performed 5 minutes later. In 5 dogs, a second GGVOD was placed in a separate artery supplying a different vascular location. After device implantation, the dogs received intravenous euthanasia solution (Beuthanasia D solution, 10 to 20 mL), and the arteries containing the devices were removed at necropsy.
With the experience and results obtained from the short-term study, a long-term animal model was created to evaluate the GGVOD. Ten dogs (weight, 23 to 32 kg) underwent placement of an aortopulmonary shunt; nine Gore-tex shunts were placed from the descending aorta to the main pulmonary artery, and one shunt was a classic Blalock-Taussig shunt. These aortopulmonary shunts facilitated GGVOD implantation by providing a rather straight catheter course from the femoral vessels to the shunt, possibly an easier catheter course than some aortopulmonary shunts in children.
Under general anesthesia and by a standard left thoracotomy approach, the left lung was retracted. The anterior aspect of the descending aorta was dissected cleanly, as was the superior aspect of the main pulmonary artery. Heparin (100 U/kg) was administered. An aortopulmonary shunt was created by use of nonringed Gore-tex tubing (W.L. Gore & Assoc Inc) 5 or 6 mm in diameter and 5 cm long. The thoracotomy incision was closed in layers. The dogs remained intubated; catheterization and GGVOD implantation were performed. In each shunt, a GGVOD equal to or 1 mm larger than the shunt diameter was implanted. In 5 of the 10 dogs, a second GGVOD was implanted in a subclavian artery.
A 7F sheath was placed in both the femoral artery and vein. A 7F pigtail catheter (Cook Inc) was advanced into the arterial sheath to the level of the shunt. Angiography was performed to evaluate shunt patency and location; the best angiographic projection was either straight posteroanterior or shallow (20°) left anterior oblique angulation. A 7F end-hole catheter was advanced into the femoral venous sheath prograde through the right heart and into the main pulmonary artery. A 0.035-in angled Terumo Glidewire (Medi-Tech Inc) was advanced through the end-hole catheter, into the shunt, and down the descending aorta; this wire did not damage the recent surgical shunt anastomoses. The end-hole catheter was advanced over the wire. The wire was removed and replaced with a 0.038-in stiff exchange wire, then the catheter and sheath were removed. The 8F long sheath and dilator were advanced over the wire into the middle of the shunt. The GGVOD was advanced through the sheath into the shunt. As described previously, the filler wire was inserted into the sack, and the GGVOD was implanted. After device implantation, three angiograms were performed: one using the sheath in the main pulmonary artery to ensure that there was no obstruction to pulmonary blood flow and two in the aorta to assess the effectiveness of the GGVOD occlusion. In several dogs, the pigtail catheter was exchanged for a Judkins right coronary artery catheter (Cook Inc), which was positioned in the aortic end of the shunt for an additional angiogram.
The catheters and sheaths were removed, and hemostasis was obtained. The dogs awoke from anesthesia, were observed closely for several days, then were boarded for 3 to 6 months. Repeat catheterization was performed, with angiograms done in both the aorta and main pulmonary artery. If a second GGVOD was placed in a subclavian artery, an angiogram was performed proximal to this device. The animals received intravenous euthanasia solution, then the GGVODs were harvested. The shunts and arteries were evaluated macroscopically, followed by thin sectioning of the GGVOD. The sections were stained (hematoxylin and eosin, Masson’s trichrome, and toluidine blue), then examined microscopically.
A total of 11 GGVODs were implanted in 6 dogs: 5 devices in subclavian arteries, 4 devices in carotid arteries, and 2 devices in renal arteries. GGVODs of five different sizes were implanted (two 3 mm, three 5 mm, two 6 mm, three 7 mm, and one 9 mm) in arteries ranging from 3.0 to 9.0 mm in diameter. Postimplant angiography revealed that each GGVOD resulted in immediate, complete occlusion of the artery (Fig 4⇓). No arteries were damaged and no devices embolized. At necropsy, the arteries were distended slightly by the GGVOD, but there was no disruption of the vessel wall or extravascular hemorrhage.
Ten dogs had aortopulmonary shunts created: 9 Gore-tex shunts (three 5 mm, six 6 mm) and 1 Blalock-Taussig shunt. A GGVOD was implanted successfully in each shunt. The GGVODs remained in position and resulted in immediate, complete occlusion of all 9 Gore-tex shunts (Figs 5⇓, 6⇓, and 7⇓). The Blalock-Taussig shunt had a small residual leak immediately after GGVOD implantation. However, postcatheterization review of the cineangiograms revealed that the subclavian artery diameter was 0.2 mm larger than the occlusion device. Three to 6 months after device implantation, all 10 dogs underwent a follow-up catheterization. At recatheterization, all GGVODs remained well positioned, and every aortopulmonary shunt was occluded completely, including the Blalock-Taussig shunt.
The 5 additional GGVODs, which were implanted in subclavian arteries, resulted in immediate complete subclavian artery occlusion. At recatheterization, each subclavian artery remained occluded completely, the GGVOD position was unchanged, and the vessels were intact.
Necropsy evaluation revealed that both ostia of the shunt were occluded, covered by a layer of uninterrupted endothelium (Figs 8⇓ and 9⇓). Beneath the endothelial layer was a neointima composed of dense fibrous connective tissue; this showed a transition to thrombus undergoing organization in the region of the GGVOD. There was a minimal inflammatory response to the nylon sack (Fig 10⇓).
Fluoroscopic time was recorded during GGVOD implantation. For device implantation in a subclavian artery, 2 to 6 minutes (mean, 4 minutes) of fluoroscopy was needed. For device implantation in an aortopulmonary shunt, 3 to 22 minutes (mean, 9 minutes) of fluoroscopy was needed. A majority of this time was used manipulating the wires and catheters carefully across the recent shunt anastomoses.
One complication occurred during this investigation. During device implantation in a Gore-tex shunt, a sack dislodged from the delivery catheter and embolized to the distal left pulmonary artery. A second device was implanted successfully in the shunt. The embolization was caused by two factors: a kinked sheath and an insufficient flare on the delivery catheter. The sheath material has been improved, and the flaring mechanism has been corrected. Subsequently, eight devices were implanted without an embolization.
Although there are several components to the complete GGVOD system (sheaths, catheters, wires), no other technical malfunctions occurred. Also, when the long sheath was allowed to fill with blood (from the high-pressure aorta) before the sack was inserted, air could not enter the sheath, eliminating the possibility of an air embolism.
Transcatheter occlusion devices, as treatment for congenital heart defects, have been used successfully since 1974, when Portsmann et al9 occluded a PDA, and 1979, when King and Mills10 closed an atrial septal defect. The original devices required very large delivery catheters and were difficult to implant. With recent design improvements, new occlusion devices have been developed.11 In ongoing clinical trials, several of these devices have shown promising results for treating specific defects.12 13 14 15 16 This has resulted in improved management of several congenital heart defects by providing a variety of catheter-delivered occlusion devices for different defects. However, each occlusion device has intrinsic limitations, including complicated delivery mechanisms, residual leaks, and device embolizations. Recently, a detachable balloon occlusion device, which did work for specific defects, was removed from clinical use because of concerns about deleterious effects of the balloon material (silicone).17
Each occlusion device is effective for a specific defect. Problems arise when patients have other cardiovascular defects that require occlusion, and the present devices are technically suboptimal for treatment of these defects. When an occlusion device is used for such a defect, the procedure is more complicated and the results are often less successful than desired. Improved occlusion devices, including both improved old devices and totally new devices, will broaden the spectrum of cardiovascular defects that can be treated without surgery. With this as our goal, we have been developing and evaluating the GGVOD.18 19
Animal Studies Using the GGVOD
In the short-term study, the GGVOD was implanted in various systemic arteries. Each GGVOD remained in position, all vessels were occluded completely, and there were no device embolizations. However, since each artery had a gradually tapering diameter, if the GGVOD were to embolize, it could only migrate distally several millimeters. Nonetheless, this model demonstrated the feasibility of GGVOD implantation. Also, it confirmed the hypothesis that if the sack was not in the desired position (due to catheter manipulation, wire insertion, patient movement, etc) before release, the wire could be withdrawn and the sack repositioned. None of the occlusion devices currently available afford this margin of safety.
The long-term study addressed several additional questions. Over a 3- to 6-month period, the nylon sack retained its integrity, and the device did not migrate from the implantation site. This is encouraging information, since in the aortopulmonary shunts, there was a persistent high pressure gradient against the GGVOD. Thrombus formed on both sides of the GGVOD, ensuring vessel occlusion and device position. As anticipated, a smooth neointimal layer covered the thrombus. The one shunt with the small residual leak closed spontaneously and also developed thrombus formation and neointimal covering. Since GGVOD implantation in Gore-tex grafts did not provide information about “native” vessel reaction to the device, a GGVOD was implanted in subclavian arteries; each GGVOD remained well positioned in the subclavian artery, the arterial walls were intact, and the vessel remained occluded.
Comparison of GGVOD With Other Occlusion Devices
(1) The GGVOD is inserted through an 8F sheath, which is smaller than the sheath size required for many other occlusion devices. (2) The nylon sack is flexible and will conform readily to the intravascular contour of any vessel. (3) If the wire-filled sack is not in the appropriate position, the wire can be withdrawn from the sack and the sack repositioned. This repositioning can be repeated multiple times until the desired device position is obtained. (4) Once it is inserted into the vessel, if the GGVOD is too small, the wire and sack may be pulled out of the long sheath and replaced with a larger device. Conversely, if the GGVOD is too large, it can be removed and replaced with a smaller sack. (5) Since the GGVOD position and degree of occlusion can be evaluated before device release, this should result in a decreased incidence of residual leaks and device embolizations. (6) Short fluoroscopic times are needed to implant the GGVOD. For GGVOD implantation in children and adults, fluoroscopic times will be reduced further, since device implantation will be performed with a high-resolution cineangiographic unit and will not be implanted immediately after surgery, as was performed in this study.
(1) Although the GGVOD is inserted through an 8F sheath (see “Advantages”), if an arterial approach is required, this sheath size may limit GGVOD use in small infants. (2) As with all occlusion devices, to implant the GGVOD, a long delivery sheath must be positioned in the desired vessel. Sheath placement may be difficult if the vessel has a tortuous course or has an acute-angle bend (which occurs in some modified Blalock-Taussig shunts). To minimize this problem, an improved sheath is used. This sheath is made of a high-quality blue Teflon (Cook Inc) that is very resistant to kinking and can be curved manually in a hot water bath. (3) For the sack to maintain sufficient transmural pressure on the vessel, we recommend a sack diameter 1.0 to 1.5 mm larger than the vessel diameter. Although during several successful implantations, the sack and vessel diameters were equal, theoretically this could result in incomplete vessel occlusion or device embolization. Since veins are more compliant than arteries, a larger sack diameter may be needed for venous structures. Also, large vessels (7 to 9 mm) may require a sack more than 1.0 to 1.5 mm larger to provide sufficient transmural pressure. Studies are continuing to determine the optimal diameter for the sack relative to the vessel diameter. (4) To safely implant the GGVOD, a minimal length of vessel is required. In this study, a vessel length of at least twice the sack diameter was used (eg, a 5-mm-diameter sack implanted in a 10-mm-long vessel). As more experience is gained with the GGVOD, the minimal vessel length may be reduced to 1.0 to 1.5 times the vessel diameter.
We conclude that the GGVOD was effective in completely occluding peripheral vessels and surgically created aortopulmonary shunts in a canine model. We speculate that the GGVOD may prove useful in humans for the occlusion of both tapered and nontapered tubular vessels, including long PDA, aortopulmonary collateral vessels, systemic–to–pulmonary artery shunts, AV malformations, and arteries supplying tumors.
The authors would like to thank the Lillie Frank Abercrombie Fund of Texas Children’s Hospital and Cook Inc for grants to support this study. Also, special thanks to Drs Sang Park (Children’s Hospital of Pittsburgh) and G. Wesley Vick (Texas Children’s Hospital) for helping prepare the figures.
Consulting editor for this article was Eric Topol, Department of Cardiology, Cleveland Clinic Foundation, Cleveland, Ohio.
- Received July 19, 1994.
- Revision received October 5, 1994.
- Accepted October 30, 1994.
- Copyright © 1995 by American Heart Association
Rashkind WJ. Transcatheter treatment of congenital heart disease. Circulation. 1983;67:711-716.
Rashkind WJ, Cuaso CC. Transcatheter closure of a patent ductus arteriosus: successful use in a 3.5 kg infant. Pediatr Cardiol. 1979;1:3-7.
Sato K, Fujino M, Kozuka T, Naito Y, Kitamura S, Nakano S, Oyama C, Kawashima Y. Transfemoral plug closure of patent ductus arteriosus: experiences with 61 consecutive cases treated without thoracotomy. Circulation. 1975;51:337-341.
Bash SE, Mullins CE. Insertion of patent ductus occluder by transvenous approach: a new technique. Circulation. 1984;70(suppl II):II-285-II-288.
Rashkind WJ, Mullins CE, Hellenbrand WE, Tait MA. Nonsurgical closure of patent ductus arteriosus: clinical application of the Rashkind PDA Occluder® System. Circulation. 1987;75:583-592.
O’Laughlin MP, Nihill MR, Mullins CE. Patent ductus arteriosus occlusion: results in 205 procedures. Circulation. 1990;82(suppl III):III-582. Abstract.
Lock JE, Cockerham J, Keane J, Finley JP, Wakely PE Jr, Fellows KE. Transcatheter umbrella closure of congenital heart defects. Circulation. 1987;75:593-599.
Lock JE, Block PC, McKay RG, Baim DS, Keane JF. Catheter closure of postinfarction/postoperative ventricular defect: initial experience. Circulation. 1987;76(suppl IV):IV-28-IV-33.
Magal C, Wright KC, Duprat G Jr, Wallace S, Gianturco C. A new device for transcatheter closure of the patent ductus arteriosus: a feasibility study in dogs. Invest Radiol. 1989;24;272-276.
Grifka RG, O’Laughlin MP, Mullins CE. Initial studies with the Gianturco vascular occlusion device in a canine model. J Am Coll Cardiol. 1992;7:117A. Abstract.