Ann Thorac Surg 1999;68:2248-2251
© 1999 The Society of Thoracic Surgeons
Original Articles
Ventricular assist device support in patients with mechanical heart valves
Marc T. Swartza,
Gregory A. Lowdermilk, MDa,
Debra A. Moroney, RNa,
Lawrence R. McBride, MDa
a Division of Cardiothoracic Surgery, Department of Surgery, St. Louis University, St. Louis, Missouri, USA
Address reprint requests to Mr Swartz, Department of Surgery, St. Louis University, 3635 Vista Ave at Grand Blvd, PO Box 15250, St. Louis, MO 63110-0250
e-mail: swartzmt{at}slu.edu
 |
Abstract
|
|---|
Background. Due to potential thromboembolic complications, mechanical valves within the native heart are often considered contraindications to ventricular assist device (VAD) support.
Methods. A retrospective review of VAD cases between June 1982 and March 1998 showed 8 patients with mechanical valves who were supported with Thoratec (Pleasanton, CA) VADs.
Results. There were 6 males and 2 females ranging in age from 20 to 69 years (mean 49.8 ± 5.6). Four patients were supported when they could not be weaned from cardiopulmonary bypass after reparative procedures and were thought to have reversible injuries. Four patients were supported as a bridge-to-cardiac transplantation. Two patients had mechanical mitral valves, 2 had aortic valve replacements, 1 had an aortic homograft and mechanical mitral valve, 2 had mechanical aortic and mitral prosthesis, and 1 patient had aortic, mitral, and triscupid valves. The types of valvular prostheses were St. Jude (5 patients) and Bjork-Shiley (3 patients). Duration of support ranged from 3.0 to 150 days (mean 34 days). Four patients were supported with biventricular assist devices and 4 had left VADs. Dextran and intravenous heparin anticoagulation were used in the shorter duration patients, with warfarin being used in the bridge patients. One patient received warfarin and aspirin. At the time of autopsy or device removal, only 1 of the 12 mechanical intracardiac valves showed any evidence of thrombosis, including the aortic valves in 2 patients supported for 2 and 5 months. There were no clinical thromboembolic events. Four patients (50%) were discharged (1 weaned, 3 transplanted).
Conclusions. The 50% (4 of 8) survival rate compares favorably with the 44% (41 of 92) overall survival rate for our Thoratec patients (bridge plus recovery) who did not have mechanical prosthetic valves. These data suggest that patients with mechanical intracardiac valves can be supported for short durations with some additional risk, which is yet to be determined.
|
Introduction
|
|---|
Patients with prosthetic mechanical heart valves were excluded from most ventricular assist device (VAD) studies based on the risk of thromboembolism due to a decrease in flow across the valve during VAD support. This was considered a significant risk even in patients who were anticoagulated. By the mid 1980s longer term postcardiotomy circulatory support, as well as bridging to cardiac transplantation, extended the durations of support for many VAD patients [1–3]. This raised the awareness and concern for supporting patients with mechanical prosthetic heart valves. This is reflected by the fact that at least one of the assist device manufacturers considered the presence of a mechanical aortic valve an exclusion criterion for VAD placement in its Investigational Device Exemption studies [4]. Other device manufacturers, while not considering mechanical prosthetic valves to be exclusionary, advised investigators of the additional risks [5]. Between 1981 and 1996 St. Louis University held an Investigational Device Exemption for the Thoratec VAD (Thoratec Laboratories, Pleasanton, CA). Since 20% of our elective cardiac surgical cases were valve replacements, many of these receiving mechanical valves, and since we received referrals for cardiac transplantation evaluation in patients with valvular cardiomyopathy, some of whom had previous mechanical valve replacements, we decided to include these patients in our protocol. While we considered the presence of a mechanical prosthetic valve to be an additional risk factor, no patient was excluded from VAD support due to the presence of a mechanical valve within the natural heart.
|
Material and methods
|
|---|
A retrospective review was performed on 100 patients who have been supported with the Thoratec VAD. Informed consent was obtained from all patients involved in this study. From March 1982 through January 1996 patients receiving the Thoratec VAD were enrolled in an Investigational Device Exemption study regulated by the Food and Drug Administration (FDA) and the Institutional Review Board of St. Louis University. The Thoratec VAD has had FDA clearance since January 1996. Table 1 shows the current anticoagulation protocol for Thoratec VAD patients with and without native heart mechanical cardiac valves. This protocol has been modified over the years. Anticoagulants or antiplatelet drugs were not started on any patient until postoperative bleeding was controlled (< 100 cc per hour for 3 consecutive hours) and until hemostatic variables were normalizing. In patients with mechanical prosthetic valves, the anticoagulation regimens were started earlier, and maintained at higher levels, than patients receiving the Thoratec VAD without a mechanical valve. Heparin was usually started within 24 hours postoperatively to maintain the partial thromboplastin time at 1.5 times control. Warfarin was initiated as soon as the patient was able to tolerate oral medications and the International Normalization Ratio (INR) was maintained at 3 to 3.5. Aspirin 81 mg/day was added at 7 days as long as platelet count was greater than 150,000 mm3 and stable. Data reviewed included preVAD history and status, position of valve, type of prosthesis, type of support (left ventricular assist device [LVAD], biventricular assist devices [BVAD]), method of cannulation, type of anticoagulation, and thromboembolic events. Patient status (weaned, transplanted, and survived) were also recorded. Other information collected included presence of thrombus within the VAD at the time of removal, transplant, or autopsy, and the duration of VAD support. Most of this information was obtained from a prospectively maintained database with only a small percentage obtained retrospectively.
|
Results
|
|---|
Patient demographics are shown in Table 2 along with the reparative operative procedure, type of valve, type of support, method of cannulation, and duration of VAD support. Seven of 8 patients had valve replacements with or without coronary artery bypass grafting (CABG) within 1 hour to 7 days of VAD insertion. One patient had BVADs inserted 22 years after aortic, mitral, and tricuspid valve replacements. Four patients (patients 1–4) were thought to have reversible myocardial damage and were considered recovery patients while 4 patients (patients 5–8) were believed to have irreversible myocardial dysfunction requiring cardiac transplantation. Recovery patients were supported for an average of 7.5 days versus 60.8 days for the bridge group.
Individual patient antithrombotic regimens are shown in Table 3. Three patients received dextran or heparin, and 5 patients received varying combinations of dextran, heparin, and warfarin. Only 1 patient received aspirin. There was no evidence of thrombi in any of the 12 VADs examined at the time of device removal. One patient (No. 5) did have thrombus on her mechanical aortic valve (St. Jude) which was discovered 11 days postVAD insertion at the time of cardiac transplantation. None of the remaining intracardiac mechanical valves (including two aortic valves in patients supported for 69 and 156 days) showed any evidence of thrombosis at the time of device removal. None of the patients had any clinical or pathological evidence of thromboembolism. One patient (No. 7) developed a late bleeding complication while receiving heparin during support with extracorporeal membrane oxygenation (ECMO).
The results are shown in Table 4. Two patients were weaned from VADs after their natural hearts recovered, with 1 surviving to discharge (No. 2). The 1 bridge patient who was not transplanted died of respiratory failure and hypoxia after being weaned from ECMO. Three of the 4 bridge-to-transplant patients were transplanted with all 3 discharged. One (No. 5) died 4 months posttransplant of a left main coronary artery occlusion and sudden death. Three patients are currently alive with length of survival of 10, 19, and 159 months. Two of the survivors are New York Heart Association functional class I and 1 is functional class II.
|
Comment
|
|---|
The presence of a mechanical cardiac valve adds considerable risk of morbidity to any acutely ill patient. This was thought to be particularly true in patients undergoing VAD support, where normal pressure gradients and blood flow across the prosthesis may be altered or interrupted. Several reports describing postcardiotomy patients receiving centrifugal assist device support have documented the risk of prosthetic valve thrombosis. Surprisingly, these reports detailed thrombosis of bioprosthetic valves in both the mitral and aortic positions [6, 7]. The Novacor (Novacor Division, Baxter Healthcare Corp, Oakland, CA) left ventricular assist system Investigational Device Exemption protocol included mechanical aortic valve as an exclusion factor. This was partially due to the risk of thromboembolism from a clotted nonopening valve in patients receiving left ventricular to aortic bypass. Also, if thrombosis occurred, it would be virtually impossible to determine whether the origin of the thrombus was from the LVAD or the mechanical aortic valve. No matter what the cause, if the LVAD was associated with thromboemboli, this would make it more difficult for the device manufacturer to obtain FDA clearance. Since we had no problems supporting patients with prosthetic mechanical cardiac valves for short durations early in our experience, we did not consider it an exclusion for bridging to transplantation. We recognized mechanical valves as additional risk factors and hoped to compensate by altering our protocols to anticoagulate earlier postoperatively and at higher levels.
All patients (4 LVAD, 4 BVAD) had interruption of flow across at least one mechanical cardiac valve (Table 2). Only 1 of these patients (No. 5) had thrombus on the prosthetic valve. This patient, however, showed no clinical signs of thromboemboli, was transplanted and survived. One BVAD patient with 22-year-old Bjork-Shiley AVR, MVR, and tricuspid valve replacement (TVR) had clean valves at transplant after 156 days of BVAD support.
We have supported 4 mechanical valve patients (3 MVR and 1 AVR) with RVADs using right atrial to pulmonary artery cannulation. In this group, operation of the RVAD did not interfere with normal pressure and flow across the left-sided prosthetic valves. At the time of implant we were not sure if these patients would require BVADs. In these cases also, however, there was no evidence of thrombus formation or thromboembolism. Two other patients with bioprosthetic valves (1 AVR, 1 MVR) with BVADs had no problems. At the time of device weaning or autopsy, all of the natural heart prosthetic valves in these 6 additional patients were clean.
The flow across the mechanical cardiac valve can sometimes be optimized depending on the type of VAD cannulation used. For example, patients with a prosthetic mitral valve, if possible, should have left ventricular to aortic rather than left atrial to aortic cannulation. Left ventricular cannulation allows normal blood flow across the mitral prosthesis. Prosthetic aortic valves with either left atrial to aortic or left ventricular to aortic cannulation will at best open intermittently. In these cases the likelihood of valve thrombosis should be considered and, in some cases, expected. What has yet to be determined is whether or not it is better to let the mechanical aortic valve open intermittently by limiting LVAD flow or to attempt to continuously unload the left ventricle, resulting in constant aortic valve closure. Clinically, it is often difficult to easily determine whether or not the native or prosthetic aortic valve is opening. We never limit VAD flows, hoping to optimize VAD washing.
While concerned about the presence of a mechanical cardiac valve in patients receiving VAD support, we have never considered it a contraindication in itself. Considerations have ranged from excluding this group to removing the mechanical prosthesis and replacing it with a bioprosthetic valve at the time of VAD implantation. Excluding this group may eliminate possible survivors while replacing the mechanical valve with a bioprosthetic valve could lengthen and complicate VAD implantation resulting in additional problems. In addition, previous experiences have shown that bioprosthetic valves will also clot [6, 7].
We understand that this small retrospective study provides a limited glance into the risks associated with mechanical cardiac valves and VAD support. Only 2 of 8 patients were supported more than 30 days, limiting the accuracy of predicting outcomes in patients supported for longer durations. Both of these longer-term patients had mechanical aortic valves. It is our hope that this information will provide at least a point of reference. Currently there is little information in the literature to assist clinicians in making this decision. Several VADs have been cleared by the FDA for commercial sale and in some respects should be considered conventional therapy. In this particular patient population, we have always given the individual the benefit of the doubt. Like reparative cardiac operations, cardiac transplantation, and other surgical procedures, mechanical circulatory support should attempt to expand its indications. It is doubtful whether a randomized trial evaluating prosthetic mechanical valves in VAD patients will ever be performed. In the future it is hoped that larger series will be published to verify or dispute these results.
We believe this information suggests that the presence of a mechanical prosthetic cardiac valve should be considered an additional risk factor based primarily on the greater risks of thromboembolism and bleeding. This additional risk must be weighed for each individual based on the clinical status, type, and position of valve, as well as the type of support required.
|
Acknowledgments
|
|---|
Supported in part by National Heart, Lung and Blood Institute Grant No. NOI-HV12909
|
References
|
|---|
-
Pennington D.G., McBride L.R., Swartz M.T., et al. Use of the Pierce-Donachy ventricular assist device in patients with cardiogenic shock after cardiac operations. Ann Thorac Surg 1989;47:130-135.[Abstract]
-
Pae W.E., Jr, Miller C.A., Matthews Y., Pierce W.S. Ventricular assist devices for postcardiotomy cardiogenic shock. J Thorac Cardiovasc Surg 1992;104:541-553.[Abstract]
-
Farrar D.J., Hill J.D. Univentricular and biventricular Thoratec VAD support as a bridge-to-transplantation. Ann Thorac Surg 1993;55:276-282.[Abstract]
-
Pennington D.G., Portner P.M., Swartz M.T. Clinical experience with the Novacor left ventricular assist system. In: Lewis T., Graham T.R., eds. Mechanical circulatory support. London: Edward Arnold Publisher, 1995:225-228.
-
Thoratec Ventricular Assist Device 14751 (REVB) 1995 Directions for Use. Thoratec Laboratories Corporation, Pleasanton, CA.
-
Hagley M.T., Lopez-Candales A., Phillips K.J., Daily B.B., Kouchoukos N.T. Thrombosis of mitral valve bioprothesis in patients requiring circulatory assistance. Ann Thorac Surg 1995;60:1814-1816.[Abstract/Free Full Text]
-
Mesana T., Monties J.R., Blin D., Goudard A., Mouly-Bandini A., Cornen A. Thromboembolic complications during circulatory assistance with a centrifugal pump in patients with valvular prostheses. ASAIO Trans 1990;36:M525-M528.[Medline]
Accepted for publication May 29, 1999.
Top
Abstract
Introduction
