Ann Thorac Surg 2001;72:2051-2054
© 2001 The Society of Thoracic Surgeons
Original article: cardiovascular
Ventricular assist device use with mechanical heart valves: an outcome series and literature review
William B. Tisol, MDa,
Dale K. Mueller, MD*b,
Frederick B. Hoy, MDb,
Robert C. Gomez, MDb,
Barry S. Clemson, MDb,
Syed M. Hussain, MDa
a Division of Cardiovascular and Thoracic Surgery, University of Illinois College of Medicine, Peoria, Illinois, USA
b Illinois Cardiac Surgery Associates, Heart Care Midwest, and Downstate Heart Transplant Center at Order of St. Francis Medical Center, Peoria, Illinois, USA
Accepted for publication August 7, 2001.
* Address reprint requests to Dr Mueller, Illinois Cardiac Surgery Associates, 515 NE Glen Oak, Suite 202, Peoria, IL 61603, USA
Background. Management of postcardiotomy cardiogenic shock with a ventricular assist device (VAD) is a common and accepted therapeutic option. However, VAD use in patients with mechanical heart valves (MHVs) is thought to carry an increased risk of thromboembolus. We report a series of 7 patients with combined VAD-MHV and review the literature.
Methods. A retrospective review was performed on all patients who were supported with a ventricular assist device with a mechanical heart valve in place. A literature review was also performed from 1966 to 2000.
Results. Seven patients were identified from April 1988 to June 2000 as having VAD support with a MHV. One thromboembolic event was documented in the 7 patients (14%). Five of the 7 patients (71%) underwent VAD explantation. Overall survival rate was 3 of 7 (43%). Causes of death included heart failure, renal failure, multisystem organ failure, adult respiratory distress syndrome, and cerebral hypoxia. All patients who died had support withdrawn at the request of the family. All patients discharged are currently alive with length of survival of 3, 26, and 84 months.
Conclusions. This study suggests that this populations rate of survival to discharge and risk of thromboembolus compare favorably to that of the general VAD population. We believe that anticoagulation can be managed as with any MHV patient and that flow rates can be kept slightly lower, which may encourage valve washing.
Management of postcardiotomy cardiogenic shock with a ventricular assist device (VAD) is a common and accepted therapeutic option . However, VAD use in patients with mechanical heart valves (MHVs) is thought to carry an increased risk of thromboembolus . Nevertheless, the two devices are combined in select patient groups, usually as a result of failure to wean from cardiopulmonary bypass or as a bridge to transplantation. Infrequent case reports and one series have documented the outcome of patients with combined ventricular assist devices and mechanical heart valves . We report an additional series of 7 patients with combined VAD-MHV and review the literature. We hypothesize that this populations rate survival to discharge and risk of thromboembolus compare favorably to that of the general VAD population.
Material and methods
A retrospective review was performed on all patients of the Illinois Cardiac Surgery Associates (Peoria, IL) who were supported with a ventricular assist device and a mechanical heart valve. Data reviewed included the following: patient age and sex; operative procedure performed at the time that VAD support was initiated; the manufacturer, position and size of the mechanical valve; manufacturer and type of support (left ventricular or biventricular assist); flow rate and cannulation method; implant and explant date of the device; indication for VAD placement (failure to wean from cardiopulmonary bypass or bridge to transplant); method, measurement, and duration of anticoagulation; occurrence of a thromboembolic event; date of discharge or death; and cause of death. A review of the literature from 1966 to 2000 was performed using MEDLINE (United States National Library of Medicine, Bethesda, MD).
Seven patients were identified from April 1988 to June 2000 as having VAD support with a MHV. Table 1 summarizes patient demographics as well as operative procedure, valve type and size, and duration between valve placement and VAD implantation. Of the 7 patients, 4 were female (57%) and 3 were male (43%) with an age range of 45 to 77 years (mean 61.6 years). Four patients had mitral valve replacement, 2 had aortic valve replacement, and 1 had combined mitral and aortic valve replacement. Four of the 7 patients had their valve replacement combined with a coronary artery bypass graft procedure. Six valvular protheses were St. Jude (St. Jude Medical, St. Paul, MN) and in one (patient 1) was a Medtronic-Hall valve (Medtronic, Inc, Minneapolis, MN). Mitral valve size ranged from 27 to 31 mm; aortic valve size ranged from 23 to 25 mm. Six patients were placed on a VAD at the time of valve replacement, and 1 patient (patient 2) underwent VAD implantation while awaiting transplantation 7 years and 5 months after aortic valve replacement.
Six patients received Bio-Medicus centrifugal devices (Bio-Medicus Bio-Pump, Bio-Medicus, Eden Prairie, MN) and 1 (patient 2) received a Thoratec VAD (Thoratec Laboratories, Pleasanton, CA), as summarized in Table 2. Left ventricular support was provided in 5 patients, whereas 2 patients received biventricular support. Six patients received VADs secondary to failure to wean from cardiopulmonary bypass and 1 (patient 2) received the Thoratec VAD as a bridge to transplantation. Cannulation for the Bio-Medicus devices occurred from the left atrium to the aorta for left-sided support and from the right atrium to the pulmonary artery for right-sided support. The Thoratec VAD was cannulated from the left ventricle to the aorta. Flow rates averaged from 2.0 to 3.7 L/min for left-sided Bio-Medicus VADs and 1.7 to 3.5 L/min for right-sided Bio-Medicus VADs. These flow rates were limited to between 2 and 4 L/min as tolerated based on myocardial and end-organ recovery. The Thoratec VAD flows averaged from 4.0 to 6.9 L/min. In the 6 patients who received a VAD because of failure to wean from cardiopulmonary bypass, the duration of VAD support ranged from 1 to 6 days (mean 4.3 days). The patient using the VAD as a bridge to transplantation received support for 87 days.
Anticoagulation of these patients was variable, as summarized in Table 3. Five patients received heparin for initial postoperative anticoagulation. Patient 6 received enoxaparin because of a possible heparin allergy. Patient 3 was coagulopathic from the time of surgery until his death on postoperative day 1 and received no anticoagulation. Patient 4 received dextran in the early postoperative period after 7 days of heparin administration. All patients who survived were converted to warfarin. Patient 6 was being converted to warfarin at the time of death. Initiation of anticoagulation ranged from the day of surgery to postoperative day 3 (mean 1.5 days). Activated clotting time (ACT) was measured in 4 of the 7 patients with a range of 140 to 200 seconds. The ACT was not measured in patient 2 because a fixed rate of heparin was used. Patient 3 received no anticoagulation because of a coagulopathic postoperative state and thus ACT was not measured. It was also not measured in patient 6 because enoxaparin was used.
One thromboembolic event was documented in the 7 patients (14%) during the period of VAD support (Table 3). Patient 4 developed a thrombus in the left atrium near the site of the LVAD cannula insertion, as documented by routine transesophageal echocardiography (TEE) the day before VAD explantation. No thrombus was identified on the prosthetic mitral valve. This patient had the VAD explanted and was discharged from the hospital. No other thromboembolic events were documented in the other 6 patients.
The final outcomes are shown in Table 4. Five of the 7 patients (71%) underwent VAD explantation. Of these, 3 were discharged from the hospital (2 weaned, 1 transplanted) and 2 died before discharge. The remaining 2 patients died before explantation of the VAD. Overall survival rate was 3 of 7 (43%). Causes of death included heart failure, renal failure, multisystem organ failure, adult respiratory distress syndrome, and cerebral hypoxia. All patients who died had medical support withdrawn at the request of the family. All patients discharged are currently alive with length of survival of 3, 26, and 84 months.
Within the last decade, sporadic articles have been published of VAD-MHV use. Two case reports from Japan documented 2 survivors with combined VAD-MHV procedures. Kitamura and associates  reported the first case of combined VAD-MHV use in a 48-year-old woman with mitral regurgitation who failed to wean from cardiopulmonary bypass after mitral valve replacement with a 29-mm prosthesis (Duromedics, Baxter-Edwards Inc, Deerfield, IL). A left ventricular assist device was implanted and heparin was used for anticoagulation, keeping the ACT between 150 and 200 seconds. On postoperative day 5 the VAD was removed. The patient was successfully discharged from the hospital. Matsuwaka and associates  described a case of a 54-year-old woman with severe mitral regurgitation who had a 27-mm Björk-Shiley (Shiley, Inc, Irvine CA) mitral valve replacement. On postoperative day 1, left ventricular support was started after an episode of cardiac arrest. Heparin was infused directly into the left ventricle for anticoagulation, with an ACT of the systemic blood maintained between 150 to 200 seconds during full support and more than 200 seconds during the weaning period. After 6 days of support, the VAD was removed. No thrombus was observed by TEE. This patient was also discharged from the hospital.
In 1999, Swartz and associates  reported a series of 8 patients with mechanical heart valves who were supported with Thoratec VADs. Four of the patients received VAD support after failure to wean from cardiopulmonary bypass. The other four used the VAD as a bridge to transplantation. Anticoagulation was not started until after postoperative bleeding was controlled and hemostatic variables were normalizing. Heparin was started within 24 hours with a partial thromboplastin time maintained at 1.5 times control. Warfarin was started when the patient was tolerating oral intake and the International Normalization Ratio was kept between 3 and 3.5. Aspirin (81 mg/d) was administered on postoperative day 7 if the platelet count was more than 150,000 mm3 and stable. One of the patients had thrombus formation on the MHV but successfully underwent transplanted. No other thromboembolic events were recognized. Overall survival to discharge was 50% (1 patient was weaned and 3 underwent transplantation).
In our series, 7 patients were identified as having VAD support with a MHV over a 12-year period. Six of the 7 patients (86%) received VAD support after failure to wean from cardiopulmonary bypass. Of these 6 patients, 2 survived (33%). One additional survivor successfully underwent transplantion. The overall survival to discharge rate was 3 of 7 (43%).
The method of anticoagulation varied among patients. In general, patients were started on heparin late on postoperative day 1. Monitoring of anticoagulation was ordered to keep the ACT between 140 and 200 seconds. This was continued until the VAD was explanted, after which warfarin was begun. Warfarin was continued in the patients after discharge if they were not transplanted and the MHV was in place.
One thromboembolic event was identified among 7 patients. The thrombus was detected by TEE and was not clinically evident. This is similar to Swartz and associates  who also reported one in 8 patients with a thromboembolic event occurring on a MHV. Their thrombus occurred on a St. Jude aortic valve and was recognized at the time of transplantation. Overall, two thromboembolic events have been documented in the 16 patients (12.5%) reported in the literature. This compares favorably to an 11% occurrence of thromboembolus in the general postcardiotomy VAD supported population .
The management of flow rates and cannulation placement has been considered as possible methods of limiting thrombus formation. Swartz and associates  comment that placing cannulas so as to maximize flow across a MHV may reduce thrombus formation. They also state that flow rates are never limited so as to optimize VAD washing. Hagley and associates , in reviewing bioprosthetic valves and VADs, believe that manipulating VAD flow rates to maximize flow across the valve may guard against thrombosis. Mesana and associates , in reporting 2 cases of VAD use with bioprosthetic valves, hypothesize that whereas MHV are considered to be more thrombogenic than bioprosthetic valves, the presence of slight regurgitation with a MHV in a VAD setting may decrease thrombus formation when compared to bioprosthetic valves.
We normally will set flow rates between 4.0 and 5.0 L/min with a Bio-Medicus VAD in the postcardiotomy cardiogenic shock patient without a MHV for bridge to recovery . We also anticoagulate only when weaning is initiated after flows are less than 2.0 L/min . We believe that in patients with MHVs and a Bio-Medicus VAD that slightly reducing VAD flow rates (2.0 to 4.0 L/min) is the preferred method if planning to explant after recovery. This may allow sufficient flow across the MHV to encourage valve washing. However, enough circulatory support must be maintained to allow myocardial and end organ recovery. Heparin anticoagulation with a Bio-Medicus or Thoratec VAD and MHV should be performed with the ACT between 140 and 200 seconds after bleeding is controlled until the time of explantation of the VAD or cardiac transplantation. If long-term support is indicated for bridge to transplantation, conversion to warfarin should be instituted and flow rated maximized, as valve washing becomes unimportant. Given the limited data with these patients, only these general recommendations can be made.
If the patient populations from Kitamura and associates , Matsuwaka and associates , and Swartz and associates  are combined with our 7 patients, a total of 17 patients are documented as having had VAD support with an MHV, with an overall survival to discharge of 9 of 17 (53%). Overall survival to discharge for patients receiving VAD support for failure to wean from cardiopulmonary bypass is 5 of 12 (42%). The rate is similar to our survival rate of 44% in the general postcardiotomy cardiogenic shock patients who receive VAD support at our institution . Overall survival to discharge for patients receiving VAD support as a bridge to transplantation is 4 of 5 (80%).
This study suggests that this populations survival to discharge rate compares favorably to that of the general VAD population. We believe that the risk of thromboembolus is acceptable in this population and is comparable to that of the general VAD population. Finally, we believe that anticoagulation can be managed as with any MHV patient and that flow rates can be kept slightly lower, which may encourage valve washing while allowing for myocardial recovery.
Stevenson L.W., Kormos R.L. Mechanical cardiac support 2000: current applications and future trial design. J Am Coll Cardiol 2001;37:340-370.[Medline]
Holman WL. Ventricular assist: what, when and how. The Society of Thoracic Surgeons 1999 Annual Meeting on the Web; http://www.sts.org/doc/3194.
Kitamura S., Mizuguchi K., Kawachi K., Morita R., Nishii T., Koh Y. Successful recovery and long-term survival following postcardiotomy mitral valve replacement cardiogenic shock by means of a left ventricular assist device. J Jpn Assoc Thorac Surg 1989;37:1406-1412.
Matsuwaka R., Matsuda H., Kaneko M., Kadoba K., Kawashima Y. Benefit of heparin infusion through left ventricular catheter in patient undergoing mechanical valve replacement with the aid of a left ventricular assist device. J Thorac Cardiovasc 1991;102:804-805.[Medline]
Swartz M.T., Lowdermilk G.A., Moroney D.A., Mcbride L.R. Ventricular assist support in patients with mechanical heart valves. Ann Thorac Surg 1999;68:2248-2251.[Abstract/Free Full Text]
Anstadt M.P., Lowe J.E. Assisted circulation. In: Sabiston D.C., Spencer F.C., eds. Surgery of the chest. Philadelphia: WB Saunders, 1995:2008-2009.
Hagley M.T., Lopez-Candales A., Phillips K.J., Daily B.B., Kouchoukos N.T. Thrombosis of mitral valve bioprostheses 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. Thromboembolytic complications during circulatory assistance with a pump in patients with valvular prostheses. ASAIO Trans 1990;36:M525-M528.[Medline]
Hoy B.Y., Mueller D.K., Geiss D.M., et al. Bridge to recovery for postcardiotomy failure: is there still a role for centrifugal pumps?. Ann Thorac Surg 2000;70:1259-1263.[Abstract/Free Full Text]
- Mehmet C. Oz
Ann. Thorac. Surg. 2001 72: 2055.