I still experimental, but their use as a bridge to heart

Transcription

1 Preperitoneal Insertion of the HeartMate 1000 IP Implantable Left Ventricular Assist Device Patrick M. McCarthy, MD, Nan Wang, MD, and Rta Vargo, MSN, RN Department of Thoracic and Cardiovascular Surgery, The Cleveland Clinic Foundahon, Cleveland, Ohio During a 16-month period, 12 consecutive patients underwent insertion of a HeartMate 1000 IP left ventricular assist device in a preperitoneal pocket that separated the device from the abdomen. All patients were in cardiogenic shock awaiting heart transplantation. Preoperatively, the mean cardiac index was 1.6 L - min- * m-, with 11 patients on intraaortic balloon pump support and 2 on extracorporeal membrane oxygenation. The pocket was formed below the rectus abdominis and internal oblique muscles and above the posterior rectus sheath. The pump fit easily in all patients. One patient died of progressive multiorgan failure. Four patients are still on support. Seven patients underwent successful transplantation after a mean duration of 55 days of support (mean pump index was 3.1 L - min- - m- during support). One patient had driveline revision because of an exit site infection but had successful transplantation. The pump was explanted without difficulty at each transplant operation. All patients having transplantation are alive and well. Preperitoneal insertion of the HeartMate left ventricular assist device can be safely performed and may avoid problems posed by intraabdominal left ventricular assist device insertion. (Ann Thoruc Surg 2994;57:63&8) mplantable left ventricular assist devices (LVADs) are I still experimental, but their use as a bridge to heart transplantation is becoming more common. The implantable LVAD was designed so that the body can accommodate the pump for eventual permanent insertion as an alternative to heart transplantation. The I-IeartMate LVAD (Thermo Cardiosystems, Inc, Woburn, MA) has been placed inside the abdomen, and occasional complications (and theoretical problems) have been described, including bowel perforation or obstruction, diaphragmatic hernia after LVAD removal, and wound dehiscence [l-31. However, the Novacor left ventricular (LV) assist system (Novacor Co, Oakland, CA), which is a larger pump, has typically been placed in a preperitoneal pocket behind the rectus muscle [4, 51. The preperitoneal pocket has several potential advantages over the abdomen as an implantable site. We describe the surgical implantation technique and clinical results in 12 consecutive patients who had HeartMate LVAD pumps inserted into a preperitoneal pocket. Material and Methods The 12 patients were all listed status I for heart transplantation and fulfilled Food and Drug Administration inclusion and exclusion criteria for HeartMate LVAD implantation. Table 1 outlines the clinical condition of the patients immediately before LVAD insertion. The group included 10 men, and the mean age was 50 years (range, 35 to 62 years). Mean body surface area was 1.95 m2, and height ranged from 1.50 to 1.80 m. None of the patients had had a previous abdominal operation. Five had had Accepted for publication May 7, 1993 Address reprint requests to Dr McCarthy, The Cleveland Clinic Foundation, 9500 Euclid Ave, F-25, Cleveland, OH previous coronary artery bypass grafting. All patients were on a regimen of inotropic drugs, and 11 had an intraaortic balloon pump in place. Two patients were also supported by extracorporeal membrane oxygenation. Nine were intubated for pulmonary edema. The mean cardiac index before pump insertion was 1.6 L - min- - rn- (range, 1.2 to 1.9 L. min- * rn- ). The LVAD implant operations were performed from February 1992 to April A midline sternotomy was extended to the umbilicus. The sternum was divided with a saw. The left anterior rectus sheath was opened medially, and electrocautery was used to create a pocket behind the rectus muscle. The dissection was extended laterally, and care was taken to minimize bleeding. The pocket was formed between the posterior rectus sheath and transversalis fascia underneath and the rectus abdominis and internal oblique muscles above (Fig 1). The pericardium was opened and reflected laterally to expose the LV apex. The peritoneum wasi dissected away from the diaphragm. In patients with previous coronary artery bypass grafting, the medial diaphragm was dissected away from the inferior chest wall to allow easy insertion of the LV inflow cannula into the LV apex. In others, a 3-cm2 opening was made in the diaphragm to pass the inflow cannula from the pocket into the pericardium. Meticulous hemostasis was obtained in the pump pocket before heparinization and pump insertion (Fig 2A). Cardiopulmonary bypass was established, and intermittent cold blood cardioplegia was administered antegrade and retrograde. The LV apex was removed with a special coring knife. Care was taken to remove an apical thrombus found in 2 patients. The LVAD driveline (pneumatic system) was passed from the inferior aspect of the pocket through the subcutaneous tissue and to the medial side of the left iliac crest. The 0.56-kg pump was easily by The Society of Thoracic Surgeons /94/$7.00

2 Ann Thorac Surg 1994; McCARTHY ET AL 635 I'REI'ERITONEAL HEARTMATE LVAD INSERTION Table 1. Summary of Patient Data Preimplantation Characteristics Postimplantation Characteristics Duration of Patient Cardiac Index Body Surface LVAD Pump Index No. Diagnosis (L * min-'. m-') Area (m') Support (d) (L. min-'. m-*) Outcome 1 ICM Tx, well 2 ICM (CABG) Tx, well 3 ICM Tx, well 4 DCM Tx, well 5 ICM (CABG) Tx, well 6 ICM Died 7 ICM Tx, well 8 ICM (CABG) 1.9" Tx, well 9 ICM (CABG) 1.4" Ongoing 10 ICM n g o i n g 11 ICM (CABG) Ongoing 12 DCM Ongoing a This was the result when patient was temporarily weaned from extracorporeal membrane oxygenator support. CABG = previous coronary artery bypass grafting; DCM = dilated cardiomyopathy; ICM = ischemic cardiomyopathy; LVAD = left ventricular assist device; Tx = transplantation. seated inside the pump pocket in all patients (Fig 2B). The LV inflow cannula was inserted into the LV apex and secured. The outflow graft was sewn to the ascending aorta. The final connection of the outflow graft to the outflow valved conduit completed pump insertion (Fig 3). All patients were weaned from cardiopulmonary bypass with good pump flow using inotropic agents and pulmonary vasodilators as needed to maximize right ventricular output. A small preperitoneal pocket was made behind the right rectus muscle. A variable portion of the outflow graft and outflow conduit was placed in this smaller pocket (depending on the patient's body habitus and orientation of the pump). In all patients, the LV inflow cannula and conduit were carefully placed so that flow into the device was laminar and without turbulence caused by kinking of the inflow cannula. Two large closed-suction drains (Jackson-Pratt drains; Baxter Healthcare Corp, Irving, CA) were directed to the most dependent portion of the left pump pocket. In patient 6, the lower 3 cm of sternum on the right was removed to prevent compression of the A it \e / Rectus abdominis // rectussheath / w LVAD 1 \ Internal oblique \Transversus a b d o m i n i s Fig 1. The left ventricular assist device (LVAD) pump was placed in a pocket below the left rectus muscle but above the posterior rectus sheath, thereby excluding the pump from the abdomen. The outflow valved conduit frequently was situated in a small pocket behind the right rectus muscle. B Fig 2. (A) The completed pocket allows ample space for pump insertion and easy access to the left ventricular apex. (B) The pump fits well in this pocket, even though the patient was only 1.50 m tall. The inflow cannula should be placed carefully to avoid kinking and to allow laminar flow from the mitral valve to the inflow cannula.

3 636 MCCARTHY ET AL Ann Thorac Surg 1994; Fig 3. Completed HeartMate left ventricular assist device system showing the pump inflow from the left ventricular apex and the outflow into the ascetzding aorta. outflow graft. No other patient had compression of the outflow graft. The chest incision was closed in all patients, and the abdominal incision was closed using fascia retention sutures. The intraaortic balloon pump was removed immediately after operation. Results Patient 6, who had multiorgan failure and heparininduced antiplatelet antibodies, experienced a severe co- agulopathy requiring the infusion of large volumes of blood and clotting factors. Twenty hours after LVAD insertion, this patient's pulmonary vascular resistance had risen from 3.6 Wood units before implantation to 10 Wood units, and pump flow dropped. A centrifugalpump right ventricular assist device was placed with improvement in LVAD pump flow, but the patient died of progressive multiorgan dysfunction and an intracerebral hemorrhage on the third postoperative day. The other 11 patients made steady progress. Four patients are still on LVAD support as of May 1, 1993: 2 have had an implant for more than a month and are awaiting a donor, and 2 have had an LVAD for less than 3 weeks. Seven patients have undergone transplantation. Their mean cardiac index rose to 3.1 L * min-' - m-* (range, 2.8 to 3.5 L * min-' m-*). All 7 were active and exercising regularly on a treadmill before transplantation. Patient 5, who was overweight and diabetic, had development of a driveline infection from Sfuphylococcus pureus. He responded initially to vancomycin, but fever returned when the antibiotic was stopped. On the 65th day after LVAD placement, he underwent driveline revision (Fig 4). The pump pocket was opened because of erythema at the incision. The pump pocket was copiously irrigated, drains were placed, the pocket was reclosed, and then it was irrigated with 5% povidone-iodine for 5 days. The patient underwent heart transplantation 76 days after LVAD insertion, and vancomycin was given for 4 weeks after transplantation. The posttransplantation course was uneventful. At 5 months after operation, there had been no further infection. For the 7 patients receiving transplants, the mean duration of LVAD support was 55 days (range, 22 to 101 days). In general, recovery after transplantation was rapid, and we attribute this in part to their excellent physical condition before transplantation. All 7 patients $ I I 5~ I I A B C Fig 4. The infected driveline in patient 5 was "moved" to a new site by (A) excising the inflammatory tissue around the old driveline, (B) extending the incision down to the fascia and to a level 10 m inferiorly, and (C) closing the incision over the driveline to create a "new" driveline exit site in uninfected tissue.

4 Ann Thorac Surg 1994:5763&8 MCCARTHY ET AL 637 are alive and well. There were no technical difficulties during device explantation and heart transplantation. The pump was easily removed from the preperitoneal pocket without any danger from intraabdominal adhesions. A closed-suction drain was left in the pump pocket for 2 to 3 days after transplantation. No late infections or complications developed in the evacuated pocket. Comment The Novacor pump has generally been implanted in a preperitoneal pocket as described here. The first Novacor implant was complicated by colon perforation after device explantation [6], and the outflow graft had been placed intraabdominally and anastomosed to the supraceliac aorta. Complications from Novacor insertion in the preperitoneal pocket have been few and limited primarily to infection and fascia1 dehiscence [4, 71. The HeartMate LVAD, however, has typically been placed in the abdomen, except for one early preperitoneal implant (0. Frazier, Houston, TX: personal communication). The inflow cannula of the HeartMate pump is much shorter than the Novacor inflow graft, and the feasibility of placing the inflow and upper portion of the HeartMate pump under the costal margin had not been well tested. For early preperitoneal implants, we used a model of a HeartMate pump to verify good fit of the pump under the costal margin before we proceeded with the pump implantation. The preperitoneal positioning of the LVAD pump has several attractive features. First, it eliminates the problems posed by intraabdominal adhesions that can complicate pump insertion if there has been a previous abdominal operation, obstruct the bowel, or complicate device explantation and contribute to possible bowel perforation. Second, the preperitoneal location decreases the risk of device erosion into the stomach, colon, liver, or other viscera. Third, a diaphragmatic hernia into the pericardium is avoided. Fourth, pocket infections can be more efficiently handled with antibiotic irrigation than intraabdominal infections. An intraabdominal LVAD pump implant may be preferable for patients who are thin and in whom erosion of the pump through the skin (as implantable defibrillators can do [S]) is a concern. Also, a thin patient may not have much subcutaneous fat to allow tunneling of the Dacron-coated driveline, which allows for tissue ingrowth and is a barrier to infection. The implantable LVAD has been criticized because the design, weight, and anatomical location of the device caused serious and life-threatening patient morbidity [9]. Although we do not want to be criticized for ill-founded optimism, [9] we think that the preperitoneal LVAD implant will avoid some of the problems encountered with intraabdominal LVAD placement. With further experience, we may adopt a flexible approach; some patients will best be served by a preperitoneal pump implant and others, by an intraabdominal implant. Portable electrically powered LVADs, already in clinical use [lo], will be offered as an alternative to heart transplantation in the foreseeable future. Painful lessons from current bridge-totransplant applications do not mean that the firstgenerqtion implantable LVADs are irrevocably flawed; we may simply have to modify our approach. References 1. Frazier OH, Rose EA, Macmanus Q, et al. Multicenter clinical evaluation of the HeartMate 1000 IP left ventricular assist device. Ann Thorac Surg 1992;53: Phillips WS, Burton NA, Macmanus Q, Lefrak EA. Surgical complications in bridging to transplantation: the Thermo Cardiosystems LVAD. Ann Thorac Surg 1992;53: Parnis SM, McGee MG, Igo SR, Dasse K, Frazier OH. Anatomic considerations for abdominally placed permanent left ventricular assist devices. ASAIO Trans 1989;35: McCarthy PM, Portner PM, Tobler HG, Starnes VA, Ramasamy N, Oyer PE. Clinical experience with the Novacor ventricular assist system. J Thorac Cardiovasc Surg 1991;102: Portner PM, Oyer PE, Pennington G, et al. Implantable electrical left ventricular assist system: bridge to transplantation and the future. Ann Thorac Surg 1989;47: Portner PM, Oyer PE, McGregor CGA, et al. First human use of an electrically powered implantable ventricular assist system. Artif Organs 1985;9(A): Kormos RL, Borovetz HS, Gasior T, et al. Experience with univentricular support in mortally ill cardiac transplant candidates. Ann Thorac Surg 1990;49: Wunderly D, Maloney J, Edel T, McHenry M, McCarthy PM. Infections in implantable cardioverter defibrillator patients. PACE 1990;13: 136M. 9. Hill JD. Invited commentary [2]. 10. Frazier OH. Chronic left ventricular support with a vented electric assist device. Ann Thorac Surg 1993;55: INVITED COMMENTARY Despite the maximal efforts of transplant organ banks, the number of heart donors has not significantly increased during the last 4 years [l]. As a result, the number of patients with end-stage heart failure who can be treated through transplantation is limited, and the wait for a donor heart continues to grow. The longer the wait, the more likely a transplant candidate will become sicker and require hospitalization in an intensive care unit [l, 21. In 1991, candidates who were in the intensive care unit at the time of transplantation required an average stay of 49 days [3]. Such care can add more than $200,000 to hospital costs before the transplant operation. To compound this problem, the shortage of donors has been particularly critical in the subgroup of patients having type 0 blood and weighing more than 90 kg: in 1991, their median wait for a heart was 590 days [3]. Clearly, we must not only

5 638 McCARTHY ET AL Ann Thorac Surg 1994; seek ways to maintain the health of transplant candidates before operation but also provide alternatives to transplantation for select patients. In 1986, the first clinical implantation of the HeartMate (Thermo Cardiosystems, Inc, Woburn, MA) left ventricular assist device (LVAD) took place at Texas Heart Institute. Since then, more than 15 years worth of cumulative patient experience has been obtained using this technology as a bridge to transplantation at 20 Food and Drug Administration-approved centers in the United States. This LVAD has worked efficiently, safely, and reliably, with minimal complications. We must now consider using this LVAD technology as it was originally intended. The HeartMate is a derivative of the National Heart, Lung, and Blood Institute program in the early 1970s for the development of an implantable pump to serve as an alternative to transplantation. A battery power source, which can allow patients to be discharged from the hospital, is bringing us closer to that goal [4]. Thus, new protocols should be designed to determine which patients will benefit most from longterm LVAD treatment. In this important article, McCarthy and co-workers report their success with preperitoneal implantation of the HeartMate pump. This experience represents an important step toward broadening the use of this LVAD for long-term support. In the future, this implant technique could increase the ease of exchanging pumps should the need arise during long-term support. 0. H. Frazier, MD Transplant Service Cullen Cardiovascular Research Laboratories PO Box Texas Heart Institute Houston, TX References Reemtsma K, Berland G, Merrill J, et al. Changing patterns of patient selection and costs in heart transplantation. J Thorac Cardiovasc Surg 1992;104: Reedy JE, Pennington DG, Miller LW, et al. Status I heart transplant patients: conventional versus ventricular assist device support. J Heart Lung Transplant 1992;11:24&52. UNOS OPTN data as of June 9, Richmond, VA: United Network for Organ Sharing. Frazier OH. Chronic left ventricular support with a vented electric assist device. Ann Thorac Surg 1993;55:27%5.