- Email: [email protected]
Hunter vena cava balloon: Rationale and results
H u n t e r vena cava balloon: Rationale and results James A. H u n t e r , M.D., and Giacomo A. DeLaria, M.D., Chicago, Ill. The majority of patients with venous thromboembolism are successfully managed with anticoagulation therapy, but certain patients require inferior vena cava (IVC) interruption. Traditional open operations on the IVC have important disadvantages, among which are significant morbidity and mortality. Twenty years ago work began to develop a transvenous method to interrupt the IVC. As the requirements for a safe and effective method were defined, it became apparent that the best approach would be with a catheter-delivered detachable balloon secured by hyperinflation in the distensible IVC. The concept of a "filter" was discarded because of predicted problems with thrombosis, induced embolism, and device migration. We have treated 135 patients with the Hunter IVC balloon. Sick patients tolerate the procedure, and it is highly effective in preventing pulmonary embolism. Leg morbidity is acceptable, parallels the extent of the phlebitis, and is lessened by leg care and simultaneous anticoagulation therapy. Long-term results with follow-up to 13 years are excellent. (J VASC SURG 1984; 1:491-7.)
Venous thromboembolism is an important clinical problem. It may be the most common cause o f hospital death.l Most patients with venous thromboembolism are successfully managed with anticoagulation therapy, but certain patients require mechanical protection against pulmonary embolism. In the latter situation ligation or plication of the inferior vena cava (IVC) has been the accepted procedure. Twenty years ago we recognized that traditional open surgical procedures on the IVC had two major drawbacks. First, because the patients are often old or obese and have compromised hearts and lungs and serious underlying disease, general anesthesia and major surgery resulted in significant mortality. Second, such surgery required that anticoagulation therapy be discontinued--even if such treatment was not otherwise contraindicated. We believed that this was the price to pay if one elected operation, since continued administration o f anticoagulants had a favorable effect on controlling the phlebitic process in the legs and perhaps favored resolution o f the clots already embolized to the lungs, With these concerns in mind and recognition o f the anatomic truth that the IVC is readily accessible from the periphery, we determined to develop a transvenous method of IVC interruption that could be done in the conscious anticoagulated patient. From the Department of Cardiovascular Surgery, Rush Medical College, and Rush-Presbyterian-St. Luke's Medical Center. Reprint requests: JamesA. Hunter, M.D., 1725 W. Harrison St., Ste. 850, Chicago, IL 60612.
REQUIREMENTS
FOR A
TRANSVENOUS METHOD Two decades ago there was no recorded experience or published thought on a transvenous method. Our first responsibility, therefore, was to establish requirements for a technique that would be both safe and effective. After careful review of the opportunities and limitations o f a transvenous approach, six requirements evolved. Early in our study it became apparent that the anatomic and physiologic considerations of a transvenous method were entirely different from those of open surgery and that a new set of principles applied. In particular it became evident that the age-old debate o f ligation vs. plication, done at open surgery, was not germane to the new technique. The first requirement for a transvenous method was that the approach be transjugular and use local anesthesia. Among the several advantages of the cervical approach over the femoral is that it avoids the potentially clot-filled channels o f the legs and pelvis. The second requirement was that the intracaval device accommodate any size o f IVC. Studies had shown that the IVC varies greatly in dimension from patient to patient. A third requirement was that the instrument have a "built in" capability for venography. This would allow precise definition o f individual anatomy and permit confirmation of proper placement of the intracaval device before it was detached and left in place. 491
Journal of VASCULAR SURGERY
492 Hunter and DeLaria
Hg. 1. Body cross section showing relationship oflVC to aorta, duodenum, and right ureter. (From Hunter JA, DeLaria GA, Goldin MD, Javid H, Najafi H, Serry C. Requirements for a method of transvenous inferior vena cava interruption. Arch Surg 1980; 115:1324-30.) The fourth requirement was that the technique must permit simultaneous administration o f anticoagulants. In this regard we recognized that the caval device should not compromise or perforate the IVC wail since this could result in retropcritoneal bleeding. The fifth requirement was that the method should produce complete occlusion o f the IVC. With no reference to any preference for ligation or plication, as is done at open surgery, this decision was dictated solely by the considerations inherent in a transvenous method. The "filter" concept was discarded for two reasons. First, it was illogical to suppose that a flow-through filter device placed in the low-pressure IVC of a patient with a hypercoagulate state would remain free of thrombus. Clot formation on or above a filter device could result in a pulmonary embolism. It is well known that even in the high-flow high-pressure aortic root, intravascular devices (prosthetic cardiac valves) have as a principal complication clot formation on the device with resulting embolism. A second reason for discarding a filter design was that such a device would require a rigid or semirigid structure with penetration o f the venous wall for retention. This would result in transmural migration o f the device and make it unacceptable. A final requirement was that the device not have sharp pins, points, or hooks to perforate the venous wall for retention. The IVC is a thin-walled structure in immediate contact with the aorta, duodenum, and right ureter (Fig. 1). A device requiring
sharp edges for retention wotfld inevitably migrate into the retroperitoneal space. Not only would such migration threaten important adjacent structures, but it would render the device deficient in its intracaval purpose. The above-stated judgments about the specifications for a safe and effective technique of transvcnous IVC interruption have been previously published. 2 It was on the basis of these considerations that the vcna cava balloon evolved. THE H U N T E R B A L L O O N The above-described requirements for a method having been accepted as valid and logical, research was begun to create a technique and equipment th~ ~ would embody them. We decided that the best ap -~ proach would be to develop a transjugular catheter-delivered detachable balloon that could be inflated to any size needed and rendered secure in the distensible IVC by hyperinflation. The concept of an occluding balloon was not an objective in itself but rather an effort to develop a method within the framework of the opportunities and limitations of the transvenous approach as defined. With the participation o f a biomedical engineer, Robert Sessions, a device was designed and laboratory work began. Years of such study preceded clinical application. 3 The method was first used in a critically ill patient in July 1970. 4 Subsequently, years o f personal clinical use resulted in modification and improvement o f the device. With clear documentation that the method produced excellent results, 5-7
Volume 1 Number 3 May 1984
H u n t e r v e na cava balloon
493
Fig. 2. Photograph oflVC balloon occlusion device. Side stopcocks on handle permit intravenous infusion and venography with fluid exiting from ports in catheter just proximal to balloon. Stopcock at end of handle permits balloon distention through inflation needle inside outer
Fig. 3. X-ray films showing procedure for balloon placement. Left to right, top to bottom: Catheter is seen in right iliac vein; venogram shows anatomy and left iliac thrombus; balloon is round on initial inflation; balloon elongates and is molded by IVC; venogram confirms adequate balloon size and proper placement; catheter separates andis removed, (From Hunter JA, Dye WS, Javid H, Najafi H, Goldin MD, Serry C. Permanent transvenous balloon occlusion of the inferior vena cava. Ann Surg 1977; 186:491-9.) Procedures are done under local anesthesia in the Radiology Department with the participation of a radiologist and anesthesiologist. The balloon is readily secured by hyperinflation just below the visualized renal veins (Fig. 3). Proof of adequate inflation is shown by demonstrating that there is well-developed elongation o f the balloon as it is molded by the cylindric IVC, that the balloon is measured to be wider than the IVC above, and that firm traction on the inserting catheter shows the balloon to be fixed and immobile. Designed for placement through the right internal jugular vein, the balloon can be inserted through the femoral vein when there is a technical obstacle to the cervical approach. The balloon deflates over a period of approximately 2 years (Fig. 4). The venous wall in contact with the balloon undergoes fibrosis and gradually strictures down to contain the remnant. ~-~ The end result is similar to that formerly accomplished by IVC ligature.
Journal of VASCULAR SURGERY
494 Hunter and DeLaria
Fig. 4. Serial x-ray films show gradual deflation of balloon. Rounding of balloon occurs on day 4 as IVC accommodates to lateral pressure of balloon. Venogram at 6 months shows collapsing balloon and patent IVC above. (From Hunter JA, Dye WS, Javid H, Najafi H, Goldin MD, Serry C. Permanent transvenous balloon occlusion of the inferior vena cava. Ann Surg 1977; 186:491-9.)
CLINICAL EXPERIENCE Since the first Hunter balloon was placed over 13 years ago, 135 patients have been treated at Rush-Presbyterian-St. Luke's Medical Center and Rush Medical College in Chicago. All patients were entered in a prospective study with hospital care and long-term follow-up supervised by senior cardiovascular-thoracic surgeons. Patients were referred for IVC interruption from virtually every service of a 1000-bed tertiary care university teaching hospital. The cause of venous thromboembolism in the majority of patients was evident. About one third had cancer and another third cardiopulmonary disease; approximately another third had recently undergone an operation. A disproportionate number were obese. Several had venous thromboembolism without apparent cause and later were shown to have neoplastic disease. There was an equal number o f men and women. The ages of the patients ranged from 17 to 88 years (average 55 years). Approximately half the patients had balloon occlusion because of emboli occurring in spite of anticoagulation therapy and the other half because of Contraindications to or complications from anticoagulation therapy. A few patients had IVC interruption because o f the presence o f a large thrombus in the iliac vein or vena cava. 6 Over the years we have become increasingly convinced that such thrombus
in the larger proximal veins poses a special threat and should often be treated by IVC interruption. Initially balloon occlusion was used only to treat critically ill patients who could not conceivably tolerate general anesthesia and open surgery. As the technique proved safe and highly effective, criteria were liberalized. Nevertheless, most patients were very sick at the time of caval interruption, with over 80% having compromised cardiopulmonary function. Even patients who were intubated and on respirators tolerated balloon occlusion. Two patients exhibited a drop in blood pressure with balloon inflation. In the latter cases the balloon was deflated, colloid was given through the device, and then-with blood volume augmented--IVC interruption was tolerated. Hospital death occurred in 15% of patients and was due to progression of the underlying disease that caused the thromboembolism. In several instances it was due to anticoagulation therapy complications that had developed prior to IVC interruption. When death occurred, it averaged 14 days after IVC interruption and was unrelated to it. No patient who has had balloon occlusion has died of a pulmonary embolism. It is likely that no patient had a pulmonary embolism in the hospital following IVC interruption. Two patients are suspected of having had late emboli 2 and 3 years, re,
Volume 1 Nttmber 3 May 1984
Fig. 5. Illustration showing thrombus on vena cava filter. Patient was suspected of having recurrent pulmonary embofism. (From Adelson J, Steer ML, Glotzer DJ, Skillman JJ, Simon M, Salzman EW. Thromboembolism after insertion of the Mobin-Uddin caval filter. Surgery 1980; 87:184-9.) spectively, after discharge. Both patients recovered. Late results show that although patients have superb immediate and long-term protection from emboli, the late prognosis is often compromised by underlying disease. For example, a recent review of our patients shows that almost one third of them had neoplastic disease as the apparent cause of their venous thromboembolism. 9 Sixty-eight percent of those with cancer and venous thromboembolism eventually died of their neoplastic disease (average ) 3 + 4.7 months after hospital discharge). Even these results, however, represent worthwhile palliation and an important improvement over published data regarding cancer patients managed with anticoagulation therapy and without IVC interruption. 1° There have been no instances of malfunction of the catheter-delivered balloon that resulted from this study. Hundreds of follow-up x-ray films show that the radiopaque balloon always slowly collapses and remains absolutely stable and unchanged in its position. Late autopsy examination in patients dying, usually of cancer, show the balloon remnant to be contained in a fibrous capsule. ~'7 Migration, transmural or otherwise, has not occurred. L O W E R EXTREMITY M O R B I D I T Y
Most patients having IVC balloon occlusion ~ave had little or no difficulty with lower extremity
H u n t e r vena cava balloon
495
Fig. 6. Illustration showing retroperitoneal migration of Greenfield filter and contact with transverse duodenum. (From Berland LL, Maddison FE, Bernhard VM. Radiologic follow-up of vena cava filter devices. Am J Roentgenol 1980; i34:1047-52.) swelling. Seventy percent of the patients with late follow-up had no problem with edema or controlled mild swelling simply by periodically elevating their legs. Disability from swelling did not occur except in isolated instances where the patient had had marked edema prior to IVC interruption. The late sequelae very closely paralleled the severity of the edema and phlebitis present before balloon occlusion. There are reasons why our patients' lower extremities have done so well. First, about 50% of the IVC interruptions have been done while the patients were fully anticoagulated. This combination of two modalities of treatment (IVC interruption and anticoagulation therapy) was not possible with open surgery and has produced excellent results. We have shown that the morbidity of caval interruption can be greatly reduced if it is done in the heparinized patient. A second reason for the good results relates to the fact that we take special care with patients' legs. Essentially providing optimal management of any patient with phlebitis (with or without IVC interruption), we are strict about lower extremity elevation after balloon placement. Patients are ambulatory but are prohibited from unnecessary sitting; all transport is by cart (never wheelchair), and the foot of the bed is kept elevated. At discharge detailed advice is given on leg care with the admonition that swelling is not to bc tolerated. Custom-measured knee-length elastic stockings are prescribed only if needed to prevent edema. Patients who have venous thromboembolism, whose legs appear normal or near normal, and who
Journal of VASCULAR SURGERY
496 Hunter and DeLaria
have balloon occlusion while on anticoagulation therapy seldom have significant leg morbidity. On the other hand, patients who have marked swelling and cannot be anticoagulated are likely to have postphlebitic problems. BALLOONS VS. FILTERS
It is not possible to address the subject oftransvenous IVC interruption without comparing the Hunter balloon with the Mobin-Uddm and Greenfield filters. The physical characteristics, underlying concepts, and results of use are very different. The Hunter balloon is a sophisticated device having more than 30 parts in the balloon catheter assembly. It evolved after careful definition of the limitations and opportunities of a transvenous method. The filter devices have a basis contrary to these principles as previously defined. The problems and complications of filter devices can be traced to this conflict with basic principles and in this sense were predictable. The literature documents the many problems of vena cava filters. These relate to difficulties in insertion, misplacement, thrombosis, and migrafion.6,11-~4 Filter thrombosis has been a significant problem, especially with the Mobin-Uddin umbrella. The occurrence o f clot on a flow-through intracaval device exposes the patient to device-induced pulmonary embolism 15,~G(Fig. 5). Both the Mobin-Uddin and Greenfield devices have a high incidence of migration into the sensitive retroperitoneal space. The most dramatic migration has probably occurred with the Greenfield device, which depends on steel h o o k - bearing struts placed under tension against the thin venous wall (Fig. 6). Photographs recorded in the literature show the Greenfield device to be in good part outside the vena cava. 1J-14 Case reports in the literature record penetration and damage to retroperitoneal structures by the Mobin-Uddin device. 6 The literature records "deep penetration" into the retroperitoneal space by Greenfield's device with "indentation of the duodenum, m4 Such migration not only jeopardizes immediately adjacent important structures but compromises device prophylaxis against pulmonary embolism. Phillips et al. ~2 point out that when the Greenfield filter migrates, the "IVC is no longer 'compartmentalized' into six equal cross-sectional areas." This defeats the device design and "one or more compartments of the IVC may now be large enough to allow passage of significant thromboemboil." Reporting on a series of 66 filter patients (33
Mobin-Uddin and 33 Greenfield), Berland et al. 14 conclude that "neither of the devices is fully safe or effective in preventing pulmonary emboli. ''14 The IVC balloon is designed for ease of insertion and precise placement with confirmation by venography before detachment. This is not true of vena cava filters. Passage of the larger and leading portion of the filter catheter assembly that contains the device into the venous system under the clavicle has been described as a problem. Neither filter device has provision for intrinsic venography, placement with venography confirmation, and the ability to be relocated as needed before detachment. The literature describes the difficulties of filter insertion and proper placement. Dr. Greenfield has recommended placement of a second device if his filter is improperly positionedY Greenfield has actually developed a second and separate instrument to remc:~ his filter when the position is unsatisfactoryY Vena cava filters were developed as an expression of preference, for vena cava fenestration as done at open surgery. The objective is IVC patency. The shortcoming of such thinking relates to the fact that there is little in common between plicating or clipping the IVC as done at open surgery and the transvenous placement of a sharp-edged flowthrough intraluminal device. Filters migrate into the retroperitoneal space, and the patency of devices no longer contained in the vena cava is not germane, The positioning of sharp-edged foreign bodies in the retroperitoneal space is at best a controversial' practice. SUMMARY The vena cava IVC balloon has evolved from nearly two decades of laboratory and clinical inv¢ tigation. The technique was based on careful assessment of the limitations and opportunities of a transvenous method, which are entirely different than those of open surgery. We have shown that even very ill conscious patients tolerate IVC occlusion, that simultaneous administration of anticoagulants is beneficial and safe, and that lower extremity morbidity is acceptable. Because the balloon has no sharp edges or hooks that penetrate the venous wall for retention, there has been no device migration. The Hunter balloon is highly effective in preventing pulmonary emboli. Long-term results are excellent, with patient follow-up to over 13 years. REFERENCES
1. Sasahara A. Therapy for pulmonary embofism. JAMA 1974; 229:1795-8.
Volume 1 Number 3 May 1984
2. Hunter JA, DeLaria GA, Goldin MD, Javid H, Najafi H, Serry C. Requirements for a method of transvenous inferior vena cava interruption. Arch Surg 1980; 115:1324-30. 3. Hunter JA, Sessions R, Buenger R. Experimental balloon obstruction of the vena cava. Arch Surg 1970; 171:315-20. 4. Hunter JA. Surgery of venous and thromboembolic disease. Surg Clin North Am 1971; 151:99-110. 5. Hunter JA, Sessions R, Petasnick J. Therapeutic balloon occlusion of the inferior vena cava. JAMA i975; 234:1034-7. 6. Hunter JA, Dye WS, Javid H, Najafi H, Goldin MD, Serry C. Permanent transvenous balloon occlusion of the inferior vena cava. Ann Surg 1977; 186:49i-9. 7. Hunter JA. Inferior vena cava interruption with the HunterSessions occtuder. Contemp Surg 1979; 14:11-23. 8. Hunter JA. Vena cava interruption--Yale Vascular Seminar. Contemp Surg 1982; 20:43-68. 9. DeLaria GA, Hunter JA, Serry C, Goldin MD. Venous thromboembolism and cancer: Treatment with the Hunter balloon. J Vasc SURG. (In press.) 10. Moore FD, Osteen RT, Karp DD, Steele Jr G, Wilson RE. . Anticoagulants, venous thromboembofism and the cancer patient. Arch Surg 1981; 116:405-7.
H u n t e r vena cava balloon
497
11. Wingerd M, Bernhard VM, Maddeson F, Towne JB. Comparison of caval filters in the management of venous thromboembolism. Arch Surg 1978; i13:1264-71. 12. Phillips MR, Widricb WC, Johnson WC. Perforation of the inferior vena cava by the Kim-Ray Greenfield filter. Surgery 1980; 87:233-5. I3. Cimochowski GE, Evans RH, Zarins CK, Lee C-T, De Meester TR. Greenfield filter versus Mobin-Uddin umbrella. J Thorac Cardiovasc Surg i980; 79:358-65. i4. Berland LL, Maddison FE, Bernhard VM. Radiologie follow-up ofvena cava filter devices. Am J Radiol 1980; I34: 1047-52. 15. Adelson J, Steer ML, Glotzer DJ, Skillman JJ, Simon M, Salzman EW. Thromboembolism after insertion of the Mobin-Uddin caval filter. Surgery 1980; 87:184-9. i6. Safiani B, Denning D. Recurrent pulmonary embolism following Kim Ray Greenfield filter insertion. Contemp Surg 1982; 21:86-88. 17. Greenfield LJ, Peyton R, Crute S, Barnes R. Greenfield vena caval filter experience. Arch Surg I98i; i16:1451-6. 18. Greenfield LJ, Crute SL. Retrieval of the Greenfield vena caval filter. Surgery 1980; 88:719-22.
Venous thromboembolism is an important clinical problem. It may be the most common cause o f hospital death.l Most patients with venous thromboembolism are successfully managed with anticoagulation therapy, but certain patients require mechanical protection against pulmonary embolism. In the latter situation ligation or plication of the inferior vena cava (IVC) has been the accepted procedure. Twenty years ago we recognized that traditional open surgical procedures on the IVC had two major drawbacks. First, because the patients are often old or obese and have compromised hearts and lungs and serious underlying disease, general anesthesia and major surgery resulted in significant mortality. Second, such surgery required that anticoagulation therapy be discontinued--even if such treatment was not otherwise contraindicated. We believed that this was the price to pay if one elected operation, since continued administration o f anticoagulants had a favorable effect on controlling the phlebitic process in the legs and perhaps favored resolution o f the clots already embolized to the lungs, With these concerns in mind and recognition o f the anatomic truth that the IVC is readily accessible from the periphery, we determined to develop a transvenous method of IVC interruption that could be done in the conscious anticoagulated patient. From the Department of Cardiovascular Surgery, Rush Medical College, and Rush-Presbyterian-St. Luke's Medical Center. Reprint requests: JamesA. Hunter, M.D., 1725 W. Harrison St., Ste. 850, Chicago, IL 60612.
REQUIREMENTS
FOR A
TRANSVENOUS METHOD Two decades ago there was no recorded experience or published thought on a transvenous method. Our first responsibility, therefore, was to establish requirements for a technique that would be both safe and effective. After careful review of the opportunities and limitations o f a transvenous approach, six requirements evolved. Early in our study it became apparent that the anatomic and physiologic considerations of a transvenous method were entirely different from those of open surgery and that a new set of principles applied. In particular it became evident that the age-old debate o f ligation vs. plication, done at open surgery, was not germane to the new technique. The first requirement for a transvenous method was that the approach be transjugular and use local anesthesia. Among the several advantages of the cervical approach over the femoral is that it avoids the potentially clot-filled channels o f the legs and pelvis. The second requirement was that the intracaval device accommodate any size o f IVC. Studies had shown that the IVC varies greatly in dimension from patient to patient. A third requirement was that the instrument have a "built in" capability for venography. This would allow precise definition o f individual anatomy and permit confirmation of proper placement of the intracaval device before it was detached and left in place. 491
Journal of VASCULAR SURGERY
492 Hunter and DeLaria
Hg. 1. Body cross section showing relationship oflVC to aorta, duodenum, and right ureter. (From Hunter JA, DeLaria GA, Goldin MD, Javid H, Najafi H, Serry C. Requirements for a method of transvenous inferior vena cava interruption. Arch Surg 1980; 115:1324-30.) The fourth requirement was that the technique must permit simultaneous administration o f anticoagulants. In this regard we recognized that the caval device should not compromise or perforate the IVC wail since this could result in retropcritoneal bleeding. The fifth requirement was that the method should produce complete occlusion o f the IVC. With no reference to any preference for ligation or plication, as is done at open surgery, this decision was dictated solely by the considerations inherent in a transvenous method. The "filter" concept was discarded for two reasons. First, it was illogical to suppose that a flow-through filter device placed in the low-pressure IVC of a patient with a hypercoagulate state would remain free of thrombus. Clot formation on or above a filter device could result in a pulmonary embolism. It is well known that even in the high-flow high-pressure aortic root, intravascular devices (prosthetic cardiac valves) have as a principal complication clot formation on the device with resulting embolism. A second reason for discarding a filter design was that such a device would require a rigid or semirigid structure with penetration o f the venous wall for retention. This would result in transmural migration o f the device and make it unacceptable. A final requirement was that the device not have sharp pins, points, or hooks to perforate the venous wall for retention. The IVC is a thin-walled structure in immediate contact with the aorta, duodenum, and right ureter (Fig. 1). A device requiring
sharp edges for retention wotfld inevitably migrate into the retroperitoneal space. Not only would such migration threaten important adjacent structures, but it would render the device deficient in its intracaval purpose. The above-stated judgments about the specifications for a safe and effective technique of transvcnous IVC interruption have been previously published. 2 It was on the basis of these considerations that the vcna cava balloon evolved. THE H U N T E R B A L L O O N The above-described requirements for a method having been accepted as valid and logical, research was begun to create a technique and equipment th~ ~ would embody them. We decided that the best ap -~ proach would be to develop a transjugular catheter-delivered detachable balloon that could be inflated to any size needed and rendered secure in the distensible IVC by hyperinflation. The concept of an occluding balloon was not an objective in itself but rather an effort to develop a method within the framework of the opportunities and limitations of the transvenous approach as defined. With the participation o f a biomedical engineer, Robert Sessions, a device was designed and laboratory work began. Years of such study preceded clinical application. 3 The method was first used in a critically ill patient in July 1970. 4 Subsequently, years o f personal clinical use resulted in modification and improvement o f the device. With clear documentation that the method produced excellent results, 5-7
Volume 1 Number 3 May 1984
H u n t e r v e na cava balloon
493
Fig. 2. Photograph oflVC balloon occlusion device. Side stopcocks on handle permit intravenous infusion and venography with fluid exiting from ports in catheter just proximal to balloon. Stopcock at end of handle permits balloon distention through inflation needle inside outer
Fig. 3. X-ray films showing procedure for balloon placement. Left to right, top to bottom: Catheter is seen in right iliac vein; venogram shows anatomy and left iliac thrombus; balloon is round on initial inflation; balloon elongates and is molded by IVC; venogram confirms adequate balloon size and proper placement; catheter separates andis removed, (From Hunter JA, Dye WS, Javid H, Najafi H, Goldin MD, Serry C. Permanent transvenous balloon occlusion of the inferior vena cava. Ann Surg 1977; 186:491-9.) Procedures are done under local anesthesia in the Radiology Department with the participation of a radiologist and anesthesiologist. The balloon is readily secured by hyperinflation just below the visualized renal veins (Fig. 3). Proof of adequate inflation is shown by demonstrating that there is well-developed elongation o f the balloon as it is molded by the cylindric IVC, that the balloon is measured to be wider than the IVC above, and that firm traction on the inserting catheter shows the balloon to be fixed and immobile. Designed for placement through the right internal jugular vein, the balloon can be inserted through the femoral vein when there is a technical obstacle to the cervical approach. The balloon deflates over a period of approximately 2 years (Fig. 4). The venous wall in contact with the balloon undergoes fibrosis and gradually strictures down to contain the remnant. ~-~ The end result is similar to that formerly accomplished by IVC ligature.
Journal of VASCULAR SURGERY
494 Hunter and DeLaria
Fig. 4. Serial x-ray films show gradual deflation of balloon. Rounding of balloon occurs on day 4 as IVC accommodates to lateral pressure of balloon. Venogram at 6 months shows collapsing balloon and patent IVC above. (From Hunter JA, Dye WS, Javid H, Najafi H, Goldin MD, Serry C. Permanent transvenous balloon occlusion of the inferior vena cava. Ann Surg 1977; 186:491-9.)
CLINICAL EXPERIENCE Since the first Hunter balloon was placed over 13 years ago, 135 patients have been treated at Rush-Presbyterian-St. Luke's Medical Center and Rush Medical College in Chicago. All patients were entered in a prospective study with hospital care and long-term follow-up supervised by senior cardiovascular-thoracic surgeons. Patients were referred for IVC interruption from virtually every service of a 1000-bed tertiary care university teaching hospital. The cause of venous thromboembolism in the majority of patients was evident. About one third had cancer and another third cardiopulmonary disease; approximately another third had recently undergone an operation. A disproportionate number were obese. Several had venous thromboembolism without apparent cause and later were shown to have neoplastic disease. There was an equal number o f men and women. The ages of the patients ranged from 17 to 88 years (average 55 years). Approximately half the patients had balloon occlusion because of emboli occurring in spite of anticoagulation therapy and the other half because of Contraindications to or complications from anticoagulation therapy. A few patients had IVC interruption because o f the presence o f a large thrombus in the iliac vein or vena cava. 6 Over the years we have become increasingly convinced that such thrombus
in the larger proximal veins poses a special threat and should often be treated by IVC interruption. Initially balloon occlusion was used only to treat critically ill patients who could not conceivably tolerate general anesthesia and open surgery. As the technique proved safe and highly effective, criteria were liberalized. Nevertheless, most patients were very sick at the time of caval interruption, with over 80% having compromised cardiopulmonary function. Even patients who were intubated and on respirators tolerated balloon occlusion. Two patients exhibited a drop in blood pressure with balloon inflation. In the latter cases the balloon was deflated, colloid was given through the device, and then-with blood volume augmented--IVC interruption was tolerated. Hospital death occurred in 15% of patients and was due to progression of the underlying disease that caused the thromboembolism. In several instances it was due to anticoagulation therapy complications that had developed prior to IVC interruption. When death occurred, it averaged 14 days after IVC interruption and was unrelated to it. No patient who has had balloon occlusion has died of a pulmonary embolism. It is likely that no patient had a pulmonary embolism in the hospital following IVC interruption. Two patients are suspected of having had late emboli 2 and 3 years, re,
Volume 1 Nttmber 3 May 1984
Fig. 5. Illustration showing thrombus on vena cava filter. Patient was suspected of having recurrent pulmonary embofism. (From Adelson J, Steer ML, Glotzer DJ, Skillman JJ, Simon M, Salzman EW. Thromboembolism after insertion of the Mobin-Uddin caval filter. Surgery 1980; 87:184-9.) spectively, after discharge. Both patients recovered. Late results show that although patients have superb immediate and long-term protection from emboli, the late prognosis is often compromised by underlying disease. For example, a recent review of our patients shows that almost one third of them had neoplastic disease as the apparent cause of their venous thromboembolism. 9 Sixty-eight percent of those with cancer and venous thromboembolism eventually died of their neoplastic disease (average ) 3 + 4.7 months after hospital discharge). Even these results, however, represent worthwhile palliation and an important improvement over published data regarding cancer patients managed with anticoagulation therapy and without IVC interruption. 1° There have been no instances of malfunction of the catheter-delivered balloon that resulted from this study. Hundreds of follow-up x-ray films show that the radiopaque balloon always slowly collapses and remains absolutely stable and unchanged in its position. Late autopsy examination in patients dying, usually of cancer, show the balloon remnant to be contained in a fibrous capsule. ~'7 Migration, transmural or otherwise, has not occurred. L O W E R EXTREMITY M O R B I D I T Y
Most patients having IVC balloon occlusion ~ave had little or no difficulty with lower extremity
H u n t e r vena cava balloon
495
Fig. 6. Illustration showing retroperitoneal migration of Greenfield filter and contact with transverse duodenum. (From Berland LL, Maddison FE, Bernhard VM. Radiologic follow-up of vena cava filter devices. Am J Roentgenol 1980; i34:1047-52.) swelling. Seventy percent of the patients with late follow-up had no problem with edema or controlled mild swelling simply by periodically elevating their legs. Disability from swelling did not occur except in isolated instances where the patient had had marked edema prior to IVC interruption. The late sequelae very closely paralleled the severity of the edema and phlebitis present before balloon occlusion. There are reasons why our patients' lower extremities have done so well. First, about 50% of the IVC interruptions have been done while the patients were fully anticoagulated. This combination of two modalities of treatment (IVC interruption and anticoagulation therapy) was not possible with open surgery and has produced excellent results. We have shown that the morbidity of caval interruption can be greatly reduced if it is done in the heparinized patient. A second reason for the good results relates to the fact that we take special care with patients' legs. Essentially providing optimal management of any patient with phlebitis (with or without IVC interruption), we are strict about lower extremity elevation after balloon placement. Patients are ambulatory but are prohibited from unnecessary sitting; all transport is by cart (never wheelchair), and the foot of the bed is kept elevated. At discharge detailed advice is given on leg care with the admonition that swelling is not to bc tolerated. Custom-measured knee-length elastic stockings are prescribed only if needed to prevent edema. Patients who have venous thromboembolism, whose legs appear normal or near normal, and who
Journal of VASCULAR SURGERY
496 Hunter and DeLaria
have balloon occlusion while on anticoagulation therapy seldom have significant leg morbidity. On the other hand, patients who have marked swelling and cannot be anticoagulated are likely to have postphlebitic problems. BALLOONS VS. FILTERS
It is not possible to address the subject oftransvenous IVC interruption without comparing the Hunter balloon with the Mobin-Uddm and Greenfield filters. The physical characteristics, underlying concepts, and results of use are very different. The Hunter balloon is a sophisticated device having more than 30 parts in the balloon catheter assembly. It evolved after careful definition of the limitations and opportunities of a transvenous method. The filter devices have a basis contrary to these principles as previously defined. The problems and complications of filter devices can be traced to this conflict with basic principles and in this sense were predictable. The literature documents the many problems of vena cava filters. These relate to difficulties in insertion, misplacement, thrombosis, and migrafion.6,11-~4 Filter thrombosis has been a significant problem, especially with the Mobin-Uddin umbrella. The occurrence o f clot on a flow-through intracaval device exposes the patient to device-induced pulmonary embolism 15,~G(Fig. 5). Both the Mobin-Uddin and Greenfield devices have a high incidence of migration into the sensitive retroperitoneal space. The most dramatic migration has probably occurred with the Greenfield device, which depends on steel h o o k - bearing struts placed under tension against the thin venous wall (Fig. 6). Photographs recorded in the literature show the Greenfield device to be in good part outside the vena cava. 1J-14 Case reports in the literature record penetration and damage to retroperitoneal structures by the Mobin-Uddin device. 6 The literature records "deep penetration" into the retroperitoneal space by Greenfield's device with "indentation of the duodenum, m4 Such migration not only jeopardizes immediately adjacent important structures but compromises device prophylaxis against pulmonary embolism. Phillips et al. ~2 point out that when the Greenfield filter migrates, the "IVC is no longer 'compartmentalized' into six equal cross-sectional areas." This defeats the device design and "one or more compartments of the IVC may now be large enough to allow passage of significant thromboemboil." Reporting on a series of 66 filter patients (33
Mobin-Uddin and 33 Greenfield), Berland et al. 14 conclude that "neither of the devices is fully safe or effective in preventing pulmonary emboli. ''14 The IVC balloon is designed for ease of insertion and precise placement with confirmation by venography before detachment. This is not true of vena cava filters. Passage of the larger and leading portion of the filter catheter assembly that contains the device into the venous system under the clavicle has been described as a problem. Neither filter device has provision for intrinsic venography, placement with venography confirmation, and the ability to be relocated as needed before detachment. The literature describes the difficulties of filter insertion and proper placement. Dr. Greenfield has recommended placement of a second device if his filter is improperly positionedY Greenfield has actually developed a second and separate instrument to remc:~ his filter when the position is unsatisfactoryY Vena cava filters were developed as an expression of preference, for vena cava fenestration as done at open surgery. The objective is IVC patency. The shortcoming of such thinking relates to the fact that there is little in common between plicating or clipping the IVC as done at open surgery and the transvenous placement of a sharp-edged flowthrough intraluminal device. Filters migrate into the retroperitoneal space, and the patency of devices no longer contained in the vena cava is not germane, The positioning of sharp-edged foreign bodies in the retroperitoneal space is at best a controversial' practice. SUMMARY The vena cava IVC balloon has evolved from nearly two decades of laboratory and clinical inv¢ tigation. The technique was based on careful assessment of the limitations and opportunities of a transvenous method, which are entirely different than those of open surgery. We have shown that even very ill conscious patients tolerate IVC occlusion, that simultaneous administration of anticoagulants is beneficial and safe, and that lower extremity morbidity is acceptable. Because the balloon has no sharp edges or hooks that penetrate the venous wall for retention, there has been no device migration. The Hunter balloon is highly effective in preventing pulmonary emboli. Long-term results are excellent, with patient follow-up to over 13 years. REFERENCES
1. Sasahara A. Therapy for pulmonary embofism. JAMA 1974; 229:1795-8.
Volume 1 Number 3 May 1984
2. Hunter JA, DeLaria GA, Goldin MD, Javid H, Najafi H, Serry C. Requirements for a method of transvenous inferior vena cava interruption. Arch Surg 1980; 115:1324-30. 3. Hunter JA, Sessions R, Buenger R. Experimental balloon obstruction of the vena cava. Arch Surg 1970; 171:315-20. 4. Hunter JA. Surgery of venous and thromboembolic disease. Surg Clin North Am 1971; 151:99-110. 5. Hunter JA, Sessions R, Petasnick J. Therapeutic balloon occlusion of the inferior vena cava. JAMA i975; 234:1034-7. 6. Hunter JA, Dye WS, Javid H, Najafi H, Goldin MD, Serry C. Permanent transvenous balloon occlusion of the inferior vena cava. Ann Surg 1977; 186:49i-9. 7. Hunter JA. Inferior vena cava interruption with the HunterSessions occtuder. Contemp Surg 1979; 14:11-23. 8. Hunter JA. Vena cava interruption--Yale Vascular Seminar. Contemp Surg 1982; 20:43-68. 9. DeLaria GA, Hunter JA, Serry C, Goldin MD. Venous thromboembolism and cancer: Treatment with the Hunter balloon. J Vasc SURG. (In press.) 10. Moore FD, Osteen RT, Karp DD, Steele Jr G, Wilson RE. . Anticoagulants, venous thromboembofism and the cancer patient. Arch Surg 1981; 116:405-7.
H u n t e r vena cava balloon
497
11. Wingerd M, Bernhard VM, Maddeson F, Towne JB. Comparison of caval filters in the management of venous thromboembolism. Arch Surg 1978; i13:1264-71. 12. Phillips MR, Widricb WC, Johnson WC. Perforation of the inferior vena cava by the Kim-Ray Greenfield filter. Surgery 1980; 87:233-5. I3. Cimochowski GE, Evans RH, Zarins CK, Lee C-T, De Meester TR. Greenfield filter versus Mobin-Uddin umbrella. J Thorac Cardiovasc Surg i980; 79:358-65. i4. Berland LL, Maddison FE, Bernhard VM. Radiologie follow-up ofvena cava filter devices. Am J Radiol 1980; I34: 1047-52. 15. Adelson J, Steer ML, Glotzer DJ, Skillman JJ, Simon M, Salzman EW. Thromboembolism after insertion of the Mobin-Uddin caval filter. Surgery 1980; 87:184-9. i6. Safiani B, Denning D. Recurrent pulmonary embolism following Kim Ray Greenfield filter insertion. Contemp Surg 1982; 21:86-88. 17. Greenfield LJ, Peyton R, Crute S, Barnes R. Greenfield vena caval filter experience. Arch Surg I98i; i16:1451-6. 18. Greenfield LJ, Crute SL. Retrieval of the Greenfield vena caval filter. Surgery 1980; 88:719-22.