Inferior vena cava interruption with the Hunter-Sessions balloon: Eighteen years' experience in 191 cases

Transcription

1 Inferior vena cava interruption with the Hunter-Sessions balloon: Eighteen years' experience in 191 cases lames A. Hunter, M_D, Giacomo A. DeLaria, MD, Marshall D. Goldin, MD, Cyrus Serry, MD, David O. Monson, MD, Michael J. DaValle, MD, and Hassan Najafi, MD, Chicago, Ill. Over a period of 18 years, 191 consecutive patients had interruption of the inferior vena cava with the Hunter-Sessions balloon for complications of deep venous thrombosis and pulmonary embolism. Causes of deep venous thrombosis and pulmonary embolism included the postoperative state (33%), cancer (32%), and stroke (11%). There were 93 females and 98 males; ages ranged from 17 to 90 years (average, 57 years). Indications for placement of the Hunter-Sessions balloon were as follows: contraindication to anticoagulants (33%), anticoagulant complications (24%), puhnonary embolism despite anticoagulants (45%), and others including inferior vena cava thrombus (12%). Sixty-eight percent had clinical phlebitis and 36% had positive venography results. Ihllmonary embolism had occurred in 165 patients (86%). It was diagnosed by ventilation-perfusion scanning (75%), angiography (23%), or on clinical grounds (2%) in patients with confirmed deep venous thrombosis. At the time of the procedure 52% were in significant cardiopulmonary distress, and 10% were intubated and on respirators. Transjugular placement was done in 188 patients, and transfemoral placement was performed in three. All tolerated inferior vena cava interruption. Thirty patients (15%) died while in the hospital an average of 21 days after balloon placement, which was unrelated to the deaths. Followup was 45 months. Ninety-four patients are dead, 95 are alive, and the status of two patients is unknown. Twenty-nine of 64 patients (45%) who died after they left the hospital died of cancer. At last follow-up, 75% of patients had legs free of edema and 25% had need for elastic stockings. No malfimction or migration has occurred with the device. No patient had a pulmonary embolism while in the hospital after insertion of the Hunter-Sessions balloon, and no patient died of pulmonary embolism. Late minor pulmonary embolism occurred in three patients. (J VASC SuR~ 1989;10:450-6.) During the past 18 years, 191 consecutive patients have had interruption of the inferior vena cava with the Hunter-Sessions balloon. All of the patients were treated personally by one of the authors, and all patients were followed at regular intervals. The stares of each patient was reviewed and made current for the purpose of making this report. PATIENT MATERIAL General information about the patients is provided in Table I. A disproportionate number were obese, with the heaviest patient weighing 450 pounds. From Rush Medical College, Rush-Presbyterian-St. Luke's Medical Center. Reprint requests: James A, Hunter, MD, Rush Medical College, Rush-Presbyterian-St. Luke's Medical Center, 1653 W. Congress Parkway, Chicago, IL / 1 / Table I. Patient profile No. of patients 191 Males 98 Females 93 Age range yr Average age 57 yr Obese patients 57 (30%) Over 200 pounds 39 (20%) Over 250 pounds 13 (07%) Over 300 pounds 05 (03%) The apparent causes of the venous thromboembolism are listed in Table II. The most common operations that were followed by the complications of deep venous thrombosis and pulmonary embolism were coronary artery bypass grafting (18 patients), colon cancer resection (i3), and craniotomy (seven). A total of 21 different neoplasms were associated with venous thromboembolism, the most common of which were lung cancer (13 patients), colon cancer

2 Volume I0 Number 4 October 1989 Vena cava interruption Hunter-Sessions balloon 451 Extraction wheel /withdraws inflation 6.3 mm inserting catheter Balloon ~ ~-~--- ~ Venogram ports ~ Balloon inflation stopcock Venogram--Heparin Stopcocks Balloon inflation needle Fig. 1. Diagram of Hunter-Sessions device for IVC interruption. The handle bears stopcoclcs for hcparin (H) infusion, venography (V), and balloon inflation (B). The 6.3 mm diameterinserting catheter is 75 cm long. The extraction wheel on the handle withdraws the inflation needle from the balloon separating it. Table II. Factors thought to contribute to cause of venous thromboembolism in 19i patients Postoperative state 64 patients (33%) Cancer 61 patients (32%) Stroke 21 patients (11%) Obesity 53 patients (28%) Other causes 28 patients (15%) No apparent cause 13 patients (06%) (13), and breast cancer (12). All patients with stroke (21) had major neurologic deficits. Other apparent causes of venous thromboembolism included congestive heart failure (12 patients), trauma (eight), and birth control drugs (eight). The preoperative state of the patients' legs, as judged at bedside examination and then as evaluated with the aid of laboratory study, is described in Table III. Nine percent of the 191 patients had thrombus extending into the inferior vena cava found on venography. Eighty-six percent of the 191 patients had pulmonary embolism. The method of diagnosis and principal symptoms are defined in Table IV. Indications for and technique of inferior vena cava interruption The four indications for vena cava interruption are listed in Table V. Some patients had more than one indication for balloon placement. The types of anticoagulant-induced bleeding that required such treatment be stopped are described in Table VI. Six of the patients with intracranial or pericardial bleeding required surgical decompression. Other indications for vena cava interruption included extension of the thrombus into the vena cava in 17 patients and intolerance to even the most minor additional embolism as the principal indication in five patients. All, except two patients, had vena cava interrup- Table III. Lower extremity findings as judged by bedside examination and laboratory data before vena cava interruption Clinical evaluation of legs Legs normal 32% Phlebitis with edema 57% Phlebitis without edema 11% Postphlebitic changes 04% Laboratory confirmation Of deep venous thrombosis Venography results positive 36% Abnormal radioisotope scan results 01% Abnormal legs and proven pulmonary embolism 39% Table IV. Pulmonary embolism in 191 patients undergoing vena cava interruption Incidence of pulmonary embolism 86% Method of diagnosis Abnormal ventilation-perfusion scan 75% Pulmonary angiogram 23% Clinical with proven DVT 02% Symptoms of pulmonary embolism Dyspnea 79% Chest pain 48% Hemoptysis 28% Cardiopulmonary arrest 06% tion performed in the radiology department. One patient who required simultaneous relief of cardiac tamponade and another with necd to control bleeding from a tense and friable jugular vein had balloon placement in the operating room. All patients had balloon placement under local anesthetic, except one patient who had seizures and required general anesthetic to complete the procedure. Placement of the balloon was successful in 188 patients via the right internal jugular vein. In three patients jugular placement could not be done and the right femoral vein was used. The surgeon was assisted in all cases by a radiologist and anesthesiologist. At the time of

3 452 Hunter et al. J'ournal of -vascular SURGERY Fig. 2. A venogram obtained before balloon separation proves proper position. The elongated balloon is wider than the vena cava above and is immobile when traction is placed on the inserting catheterl The balloon is positioned with a segment ofundistended vena cava between it and the renal veins. Table V. Indication for vena cava interruption Contraindication to anticoagulants 33% Complications of anticoagulants 24% Embolism despite anticoagulants 45% Other indications (as IVC thrombus) 12% vena cava interruption 52% of the patients were described by the surgeon as being in moderate or marked respiratory distress. Forty-nine percent were receiving oxygen; 10% were intubated and on a respirator. When not contraindicated, anticoagulant administration was continued during the procedure of balloon placement and gave rise to no technical difficulty. Thirty-eight percent of patients were receiving heparin, 3% were on only warfarin, and 14% were receiving both heparin and warfarin. A diagram of the supplied sterile device (Concept, Inc., Largo, Fla.) is shown in Fig. 1. By use of the fluoroscope and injection of contrast media through the device, the anatomy was defined with particular reference to the renal veins. The balloon was hyperinflated with 25% Hypaque (Winthrop Labs, New York, N. Y.) solution below the renal veins. A vcnogram was obtained before balloon separation to prove proper position (Fig. 2). The average inflation volume was I7 ml, with the range from 8 to 38 ml reflecting extremes of cava size. Structural detail of the device handle and the mechanism by which the balloon inflation needle is withdrawn from the balloon to separate it are illustrated in Fig. 3. The design of the catheter-balloon junction with detail of the inflation needle, collet, and double valve mechanism is shown in Fig. 4. The instrument consists of 37 separate parts, the balloons are pretested, and the inserting catheter disposable. Temporary holdup of the catheter, usually in the right renal vein, occurred in one third of the patients. This was overcome by partially inflating the balloon, which moved the catheter tip away from the venous wall into the central lumen. Special problems were occasionally encountered during balloon placement. Thrombus was removed from the jugular vein in a young patient with lung cancer to permit balloon placement. Two patients with retroperitoneal tumor had distortion of the cava, with temporary catheter holdup: A patient with a horseshoe kidney and iliac venous drainage had balloon placement above the vein, with no sequelae. The blood pressure was monitored through all procedures; in two instances, significant fall occurred

4 Volume 10 Number 4 October 1989 Vena cava interruption Hunter-Sessions balloon 453 < I balloon inflation needle out of balloon Luer for venography Luer for heparin drip Fig. 3. Diagram shows detail of the side stopcocks (V and H) and their continuity with the venogram ports via the inserting catheter. Stopcock at the end of handle (B) distends balloon via the inflation needle. Top diagram shows device as supplied and ready for balloon inflation. Lower diagram shows detail of how the inflation needle is withdrawn from the (inflated) balloon by turning the extraction wheel. with balloon inflation. In both cases, the balloon was deflated, colloid was given to augment blood volume, and then reinflation of the balloon was tolerated. RESULTS After vena cava interruption patients were observed in the recovery room and then returned to their hospital beds. Strict rules of treatment were enforced for the patient after vena cava surgery. In particular, the foot of the bed was constantly elevated, all transport was by cart (never wheelchair), and all sitting, except for eating, was prohibited. Patients were allowed to ambulate without restriction as their general condition permitted. Trained nurses gave each patient information about basic venous physiology and the care of his/her lower extremities. When not contraindicated anticoagulant therapy was conducted as in any patient with venous thromboembolism. Fifty-three percent of the patients were taking warfarin when they left the hospital. The postoperative balloon placement management was the same as that appropriate for patients with deep venous thrombosis treated without caval interruption. No patient experienced another pulmonary embolus in the hospital after balloon occlusion. There were no wound infections. Two patients taking anticoagulant medication had minor wound hematomas. Assessment of the postoperative status of the patients' legs showed that 39% were normal and free of edema. Eleven percent had more edema than be- Table VI. Types of bleeding that required stopping anticoagulant therapy in 45 patients undergoing vena cava interruption Gastrointestinal bleeding Retroperitoneal bleeding Subcutaneous wound hematoma Intrapericardial bleeding Hemothorax Intracranial bleeding 15 patients 13 patients 5 patients 5 patients 4 patients 3 patients fore balloon occlusion, 35% had less edema than before vena cava interruption, and 15% had the same amount of swelling as compared to the pretreatment state. A new clinical phlebitis developed after caval interruption in 5%. Patients' legs often improved after surgery because of bed elevation, method of transport, and avoidance of sitting: Thirty of the 191 patients died in the hospital after vena cava interruption from causes unrelated to it; an autopsy was performed in 17 patients. The hospital mortality rate (15%) was the same over the 18 years of the study; however, the average time of death after tteatment changed considerably, reflecthag a changing patient population. During the first 5 years only very critically ill patients were treated; the average time of death after vena cava interruption was 9 days. During the last 5 years nearly all patients requiring vena cava interruption, including those chronically ill, were treated; the average time of hospital death was 30 days after operation: Those patients with a tendency for edema were discharged with knee-length, custom-made and mea-

5 454.Hunter et al. Journal of VASCULAR SURGERY t t Fig. 4. Diagrams show the catheter-balloon junction with structural detail revealing how balloon is inflated, released, and balloon contents retained. Top, shows catheter-balloon assembly with the inflation needle (a) in balloon core and holding primary (b) and secondary (c) valves in open position. Balloon (d)can be inflated. Inflation needle flares collet (e), which locks catheter tip 09 to the balloon. Bottom, shows catheter-balloon assembly with the inflation needle (a) withdrawn from the balloon core. Spring ~q) forces primary and secondary valves (b and c) to close, retaining balloon contents. When the inflation needle is withdrawn, the collet (e) is no longer held in the expanded position and the balloon is detached. sured elastic stockings. All patients who were discharged were told how to care for their legs. It was emphasized that any swelling should be controlled, and the presence of edema was an indication to contact their physician. Insofar as possible, data about patients were accumulated in prospective fashion. All patients were personally followed indefinitely by the treating surgeons if the patient could and would come for regular examination and abdominal x-ray. In preparation for this report all patients known to be alive were located and their status was updated. Follow-up information was available on 189 of the 191 patients. Thirty patients died in the hospital, and 159 were discharged with follow-up. Sixty-four patients were found to have died after departing the hospital, leaving 95 patients for this evaluation. Overall, a total of 94 of the original patients are known to be dead. Information about the status and fate of the patients after leaving the hospital was acquired for this study in one of four ways: by one of the authors examining the patient (56%), from another physician (31%), from the patient's family (9%), or from the patient (4%), The average duration of follow-up from balloon placement until patient death or last evaluation was 3.7 years. The average duration of rife for those patients who died was 20 months and ranged up to 17 years. Cancer was the most common cause of late death and accounted for 45% of cases. Cancer was a major cause of venous thromboembolism in this study and was the factor that had the greatest potential to affect long-term survival. Forty-six of 61 patients originally treated are dead; average life span after balloon placement was 10 months. Thirty-five of these patients rived 6 months or less. The 15 patients who are alive have survived an average of 4.8 years, with seven surviving over 5 years after treatment. On follow-up examination patients were again told how to care for their legs, with emphasis on avoiding unnecessary sitting, elevating the legs periodically, and, if necessary, having the foot of the bed elevated at night. Detailed instruction was require& For example, it was observed that instruction to elevate the legs often led to the ineffectual use of

6 Volume 10 Number 4 October I989 Vena cava interruption Hunter-Sessions balloon 455 a footstool. Patients were told that leg elevation meant to place the "feet higher than the heart." Fiftyseven percent had swollen legs before vena cava interruption, and a significant incidence of venous insufficiency was anticipated. Patients were told that the point of successful management was to have extremities free of edema. At the time of latest followup, 75% of patients had extremities free of edema, and 25% required elastic stockings. Stockings were always knee-length and custom-fitted to their edemafree measurement. Four patients exhibited postphlebitic stasis changes; three of these patients were survivors of the original pretreatment group of seven patients who had chronic venous insufficiency with stasis changes. An attempt was made to relate the late status of the patients' legs to the presence or absence of edema at the time of vena cava interruption. One hundred forty-two patients (74%) survived long enough for late evaluation; 5% of those with normal legs before placement of the balloon (57 patients) and 27% with edematous legs before placement of the balloon (85 patients) later required elastic stockings. There has been no balloon migration or malfimction. Vena cava balloons gradually- lose their radiopaque contents over approximately 2 years' time. As the balloon slowly deflates, the metallic core is gradually surrounded and retained by a fibrous capsule. The result is a stable foreign body with no tendency to migration. Hundreds of x-ray films have been taken with the balloon remnant in the proper and unchanged position in all cases. A review was made to verify balloon stability for this writing. X-ray films were obtained by the authors, family physicians, and in some instances, by the patients. Analysis of the data for the entire series provided evidence of a stable balloon remnant: a recent abdominal x-ray (62%), autopsy (19%), or a favorable clinical course (18%). No information was available about two patients (2%). No patient has died of a pulmonary embolus. One patient, who died over a year after balloon placement, had later insertion of a transjugular access catheter and at autopsy had clot around the catheter in the superior vena cava and small pulmonary emboli, which was thought not to be the cause of death. Autopsy was done in 30% of the patients who left the hospital and later died. Three patients are thought to have had minor pulmonary emboli long after discharge from the hospital. DISCUSSION The choice of a balloon to completely interrupt the vena cava in the treatment of patients with venous thromboembolism was made after consideration of the limitations and oppommities for such a method and was not an expression of preference for occlusion versus fenestration. Studies had led to the conclusion that the requirements for a transvenous method were entirely different from those of open surgery. It became apparent that whether one chose to ligate or plicate the exposed vena cava had no relevance to this entirely new technique. Variations in venous anatomy required that a transvenous device have capability for venography to ensure proper placement. It was judged that the intracaval device should not depend on hooks or points that perforate the venous wall for retention because this would result in transmural migration. Migration of the device into the retroperitoneal space would compromise its intracaval purpose to prevent pulmonary embolism and would jeopardize adjacent structures. It was decided that the device should occlude the vena cava because a flow-through device placed in the low pressure vena cava in a patient with a hypercoagulate state could result in dotting and induced embolism. Our analysis led to the judgment that these requirements could be met only by use of a catheter-delivered detachable balloon that was positioned by intraoperative venography and retained by hyperinflation in the distensible vena cava. 2 After successful laboratory study 3 the first patient was treated during July In relation to our selection of a balloon to accomplish vena cava interruption, the questio n arises about the relative merit of a "filter" design. The "filter" concept is attractive if a device could be created that could be accurately positioned in the vena cava and would remain there without physiologic or anatomic complication until it could fulfill its purpose of preventing the passage of an embolus. Our assessment was that the creation of such a device was not possible. Vena cava filters have inherent complications relating to the fact that they are flowthrough intravascular and retroperitoneal foreign bodies that depend on sharp points perforating the venous wall for retention. The complications of the Greenfield filter, as reported in the literature, are illustrative and include difficulties in insertion, misplacement, migration, embolism, and small bowel perforation, s-12 In our review of this 18-year experience, we were surprised to discover that the most common operation followed by venous thromboembolism and vena cava balloon placement was coronary artery bypass grafting. Although the authors performed the bypass procedures, and thus there may be some special interest or awareness, the same surgeons do large

7 456 Hunter et al. Journ/of VASCULAR SURGERY numbers of vascular and other kinds of cardiac surgical procedures without finding a comparable incidence. A detailed review of the 18 patients with coronary disease who required postoperative vena cava interruption is underway, with preliminary findings that atrial fibrillation is a predisposing factor, and that traditional treatment of this postoperative group with anticoagulants alone is associated with more complications and embolic breakthrough than other patients. The radiology suite was preferred to the operating room for vena cava interruption because of its superior x-ray equipment and the participation of a radiologist. Exposure of the right internal jugular vein or, exceptionally, the right femoral vein was readily accomplished with the patient under local anesthetic. The presence of an anesthesiologist was essential in providing inhalation therapy, ventilating those ptients on respirators, and monitoring vital signs. Balloons were placed without difficulty below the renal veins and proper position verified before balloon separation. Very sick patients tolerated vena cava interruption under local anesthetic. Deaths from vena cava interruption did not occur. The mortality from the underlying disease is significant, with half of the patients having died during this 18-year study period. In particular, patients with cancer have a poorer prognosis; their selection for treatment must be made with great care. Of 61 patients with cancer treated, 46 are dead; 35 of these lived 6 months or less. Nevertheless, properly selected patients with cancer can be well served as evidenced by our I5 survivors who, on average have survived almost 5 years. The surgeon must consider the status and prognosis of the patient with cancer when deciding on the appropriateness of caval interruption. DeLaria et al.13 have reviewed our subset of patients with cancer and outlined differences in their behavior and management as compared to other kinds of patients with thromboembolism. All patients having deep venous thrombosis, with or without vena cava interruption, should be educated as to the nature of their problem, the fact that it may be chronic, and that continuing care of their lower extremities may be required. It is stressed that the point of successful management is to have legs free of edema. Our patients were carefully instructed, followed indefinitely by their treating surgeons, and told that if their legs were swollen something was wrong. The best results regarding the late status of the patients' legs after any method of vena cava interruption cannot be achieved unless these concepts are applied. The fact that 57% of this series of patients had edematous legs before vena cava interruption and only 25% had need on late follow-up for elastic stockings to control swelling reflects adherence to these principles. The most important factor influencing the late status of patients' legs is the presence of edema at the time of vena cava interruption. This study has shown that patients with swollen legs before vena cava interruption are five times more likely to need elastic stockings than patients who are free of edema at the time of treatment. We believe that this series of patients offers as severe a test of the efficacy of a method of caval interruption as any that can be described. No patient in the series has died from pulmonary embolism. No pulmonary embolism occurred in the hospital after vena cava interruption, and less than 2% of patients have had minor late embolism. It is concluded that this 18-year study has shown that interruption of the vena cava with the Hunter- Sessions balloon is highly effective in preventing pulmonary embolism. Results are excellent for the status of the patients' lower extremities. REFERENCES 1. 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: Hunter JA, DeLaria GA, Goldin MD, Javid H, Serry C. Requirements for a method of transvenous inferior vena cava interruption. Arch Surg 1980;115: Hunter JA~ Sessions R, Buenger R. Experimental balloon obstruction of the inferior vena cava. Ann Surg 1970; 171: Hunter JA. Surgery of venous and thromboembolic disease. Surg Clin North Am 1971;151: McAuley CE, Webster MW, Jarret F, Hirsch SA, Steed DL. The Greenfield filter as a source of recurrent pulmonary thromboernbolism. Surg 1984;96: Sidawy AN, Menzoian MD. Distal migration and deformation of the Greenfield vena cava filter. Surgery 1986;99: Wingerd M, Bernhard VM, Maddison F, Towne JB. Comparison of caval filters in the management of thromboembolism. Arch Surg 1978;113: Santiani B, Denning D. Recurrent pulmonary embolism following Kim-Ray Greenfield filter insertion. Contemp Surg 1982;21: Berland LL, Madison FE, Bernhard VM. Radiologic followup ofvena cava filter devices. Am J Radiol 1980;134: Philips MR, Widrich WC, Johnson WC. Perforation of the inferior vena cava by the IZim-Ray Greenfield filter. Surgery 1980;87: Cimochowski GE, Evans RH, Zarins CK, Lee CT, De- Meester TR. Greenfield filter versus Mobin-Uddin umbrella~ J Thorac Cardiovasc Surg 1980;79: Hirsch SB, Harrington EB, Miller CM, Estioko MR. Haimor M. Accidental placement of the Greenfield filter in the heart: report of two cases. J VAsc SURG 1987;6: DeLaria GA, Hunter JA~ Serry C, Goldin, MD. Thromboembolism and cancer: treatment with the Hunter balloon. J VAse SURG 1984;5:670-4.