- Email: [email protected]
Right heart failure: best treated by avoidance
SESSION 7 :
LESSONS LEARNED
Right Heart Failure: Best Treated by Avoidance Clifford H. Van Meter, Jr, MD Division of Cardiothoracic Surgery and Transplantation, Alton Ochsner Medical Foundation, New Orleans, Louisiana
Right heart failure continues to affect our clinical success with left ventricular assist device support. The inability to consistently predict the probability of the onset of right heart dysfunction contributes to this problem. We have developed an aggressive approach to the manage-
ment of these patients in an attempt to decrease the incidence of this condition, which continues to carry a very high operative mortality. (Ann Thorac Surg 2001;71:S220 –2) © 2001 by The Society of Thoracic Surgeons
D
ity and the dp/dt [3]. Whatever the effect of the LVAD on the RV, the early dependence of the LVAD on RV function is clear. Mandarino and associates [4] and others demonstrated that LVAD mean filling rate correlates with RV stroke work. This followed previous work showing an improvement in RV ejection fraction from 21% to 32% [5] in a mixture of ischemic and dilated patients, probably due to after-load reduction [6], which improves the mechanical efficiency of the RV [7]. In spite of these elegant clinical and experimental studies, we are all aware of chronically supported LVAD patients who have demonstrated acceptable hemodynamics for prolonged periods with little or no RV function as in episodes of ventricular fibrillation. Reviews of clinical and hemodynamic perimeters have suggested that clinical characteristics may be more relevant in predicting the need for RVAD support [9]. Factors indicating clinical status include: preoperative inotropic requirements; fever without infection; edema on CXR, RV wall thickness; and postoperative transfusion requirements. These parameters do not include such clinical factors as ascites, elevated BUN/creatine, and others that have been associated with increased operative mortality in general. They do, however, reflect preimplant degree of illness and factors that result in impairment of pulmonary blood flow or reduced RV perfusion that may be more predictive than preimplant measures of RV function or other hemodynamic variables alone. Argenziano and associates [9] and others at Columbia University have refined and broadened the definition of RV failure to include decrease in right ventricular function or output requiring the insertion of an RVAD, administration of inhaled nitric oxide for elevated pulmonary vascular resistance (PVR), or pharmacologic inotropic support for more than 10 days. Based on this, they rely on central venous pressure, transpulmonary gradient [10] and pulmonary artery diastolic pressure as an assessment of RV compromise and need for support, as well as a guide to patient management and support weaning. In our effort to identify and treat right heart compromise in our LVAD population, we have chosen to separate risk factors into preoperative (pulmonary hypertension; acute, sustained elevation of central venous
espite an ever-growing experience with left ventricular assist devices (LVAD), the recurring problem of right heart failure (RHF) continues to complicate clinical results. Patient and device selection, surgical timing, perioperative management techniques, and our inability to always predict a patient’s clinical response lead to the persistence of this issue. It has been well recognized that the use of concomitant (RVAD) support is associated with an increased operative mortality [1]. This reality may affect clinical practice by postponing the inevitable and thereby confirming the obvious. The persistent complication of RHF in the hands of experienced teams suggests, however, that one major contributing factor is our inability to always predict its likely occurrence [1]. Efforts to predict the incidence of RHF after LVAD implant have included an analysis of hemodynamic parameters, clinical parameters, and the effect of the LVAD itself on right heart function. Clearly, all of these are important factors. Studies of the interplay between right ventricular (RV) function and LVAD influence offer misleading impressions if viewed in isolation. Some have explained RV dysfunction as due to a loss of direct left ventricular mechanical coupling or as a consequence of extreme biventricular geometric alterations. It has been demonstrated that in normal hearts, marked alterations in biventricular geometry accompanying left ventricular unloading by the LVAD do not significantly alter RV performance characteristics [2]. Is the circumstance of heart failure different in the fact that some right hearts suffer from myopathies and others suffer from an environment of pulmonary artery (PA) hypertension, volume overload, and poor pulmonary perfusion pressure, or both? To date, no clinical study has suggested a higher rate of RV failure in ischemic versus dilated myopathies. Other experimental studies during LVAD support have shown a leftward shift of the septum, resulting in a loss of interventricular balance, thereby increasing RV compliance, and decreasing RV contractilPresented at the Fifth International Conference on Circulatory Support Devices for Severe Cardiac Failure, New York, NY, Sept 15–17, 2000. Address reprint requests to Dr Van Meter, Division of Cardiothoracic Surgery and Transplantation, Ochsner Clinic, 1514 Jefferson Hwy, 08s-LT, New Orleans, LA 70121; e-mail: [email protected].
© 2001 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
0003-4975/01/$20.00 PII S0003-4975(00)02637-0
Ann Thorac Surg 2001;71:S220 –2
pressure (CVP); dilated right-sided chambers; rightsided pleural effusions; hyperbilirubinemia; ascites; compromised pulmonary function or pulmonary edema; long-term afterload reduction; and marked renal insufficiency) and intraoperative (pharmacologic measures needed to maintain adequate perfusion pressure onpump; prolonged clamp or ischemic time; compromise of RV protection; need for femoral-femoral cannulation; acidosis; and increased airway pressures). Many preoperative factors such as pulmonary edema, hyperbilirubenemia, and acute CVP elevation or pulmonary hypertension can be effectively treated or modified for surgery in some clinical circumstances. Additionally, many intraoperative factors, if recognized and appreciated, can lead to modifications in management. Repeated pharmacologic requirements, intraoperatively, to maintain perfusion pressure should prompt the use of Vasopressin or other pressors postoperatively to maintain the perfusion pressure vital to avoid right-sided compromise. The need for femoral cannulation should lead to anticipation of acidosis upon reperfusion of the lower extremity after decannulation. Aggressive ventilatory adjustments should be employed to manage increased airway pressures or compromise pulmonary compliance. Finally, there needs to be an appreciation of processes that lead to right-sided circulatory failure. Transfusion requirements can lead to coagulopathies and/or enhanced inflammatory responses. Prolonged cardiopulmonary bypass and the state of class IV congestive heart failure can lead to inflammatory responses and an enhanced humoral amplification system of fibrolysis, complement activation, and the Bradykinin-Kallekrien system. These can all lead to the overwhelming clinical complex we have all seen and appreciated as “metabolic anarchy”. Systemic vascular endothelial dysfunction caused by long-term ACE inhibition and the effects of prolonged cardiopulmonary bypass, transfusion-induced cytokine release, and the impaired release of nitric oxide by altered pulmonary vascular endothelium further disrupt homeostasis. Intraoperative and postoperative measures can be employed to help manage these phenomena. Intraoperative measures include: patient selection and timing; serine protease inhibition; right heart protective techniques; leukocyte depletion; maintenance of perfusion and mean pressures (Vasopressin); and right heart venting or bypass. Postoperative measures include: aggressive treatment of acidosis; hyperventilation (hypocarbia); nitric oxide; A-V synchrony; and early postoperative continuous veno-venous hemofiltration (CVVH). Serine protease inhibition can lessen transfusion requirements and markedly decrease inflammatory responses and lung injury, possibly by minimizing the effects of NFKB. Leukocyte depletion can be employed in line in the cardiopulmonary bypass circuit, as well as transfusion filters. As previously mentioned, Vasopressin can be used to maintain perfusion pressure without excessive pressor support and without marked compromise to the splanchnic and renal circulation. We have utilized a right heart bypass circuit to prime the LVAD [11] via the left atrium
CIRCULATORY SUPPORT VAN METER RHF AVOIDANCE
S221
Fig 1. Right heart protection algorithm. (Bili ⫽ bilirubin (serum); CVP ⫽ central venous pressure; TPg ⫽ transpulmonary gradient; RVAD ⫽ right ventricular assist device; Pa Systolic ⫽ systolic pulmonary artery pressure; AV ⫽ atrioventricular; HR ⫽ heart rate; NO ⫽ nitric oxide; CVVH ⫽ continuous veno-venous hemofiltration; RV ⫽ right ventricle.)
with warm oxygenated blood, thereby avoiding right heart overload and the pulmonary circulation while exceeding the minimal amount of flow needed to keep the VAD above the minimal threshold to avoid aspiration of air into the system through the inflow graft. We hyperventilate, aggressively treat presumed acidosis, and assure atrioventricular synchrony in a relatively high heart rate of 90 to 100 bpm. More recently, we have begun using nitric oxide (NO) routinely. Finally, we do not hesitate to begin CVVH if there is sluggish urine output postoperatively or if the response to diuretics is poor or unpredictable. We have found that the right heart can quickly become volume overloaded if urine output is not brisk when multiple infusions and frequent transfusions are being administered. Once the pattern of right heart overload begins, it can progress quite rapidly to failure. From 1996 through 1999, 35 LVADs were placed at the Ochsner Foundation Hospital, utilizing the management techniques that we have discussed to avoid right heart failure. There were seven operative deaths (within 30 days of implant), or 20% operative mortality. Two patients suffered intracranial hemorrhages, 1 developed multiple cerebral emboli from native ventricular thrombus, 1 had intractable fibrillation on day 12, and 1 patient died of progressive hepatic failure and hepatorenal syndrome 1 month postoperatively. A final patient developed sepsis from pneumonia and on postoperative day
S222
CIRCULATORY SUPPORT RHF AVOIDANCE
VAN METER
12 had a pulmonary-hypertensive crisis with acute right heart failure and death. This single patient represents the only patient in whom right failure could be identified as a cause of operative mortality. No patients received right ventricular device insertion after LVAD placement. Fifty percent of the patients received NO perioperatively, and 25% required long-term inotropic support of the right heart for 10 to 14 days postoperatively. During this same interval, 2 patients received biventricular support with Thoratec pumps (Thoratec Laboratories Corp, Berkeley, CA), 1 based on extremely low body surface area, less than 1.2 mm2, and 1 due to the clinical circumstance of cardiogenic shock with extracorporeal membrane oxygenation support as a bridge to VAD insertion. Consideration of these issues and aggressive institution of measures to manage right heart challenge cannot be started too late nor terminated too prematurely. Generally, we have initially weaned an intraaortic balloon pump first, if present, followed by NO, and finally, any pharmacologic support. Throughout this process we keep a watchful eye on peak airway pressures, central venous pressure, pulmonary artery diastolic pressure, and fluid balance. More recently, we have chosen to initially limit VAD flows based on cardiac index, because in the automatic mode with maximal right heart protective measures, flows can become unnecessarily excessive. The avoidance of right heart failure is one of the challenges that must be managed if we are to be successful in the expanded indication of devices as alternatives to transplant and if we will be able to offer newer devices to larger populations of patients through less invasive surgical procedures. Until we have a clearer understanding of who is at risk and when, consideration must be
Ann Thorac Surg 2001;71:S220 –2
given to a broad spectrum of preventative measures and maneuvers.
References 1. McCarthy PM. HeartMate implantable left ventricular assist device: bridge to transplantation and future applications. Ann Thorac Surg 1995;59:S146–51. 2. Elbeery JR, Owen CH, Savitt MA, et al. Effects of the left ventricular assist device on right ventricular function. J Thorac Cardiovasc Surg 1990;99:809–16. 3. Santamore WP, Austin EH, Gray L, Jr. Overcoming left ventricular failure with ventricular assist devices. J Heart Lung TX 1997;16:1122– 8. 4. Mandarino WA, Winowich S, Gorscan J, et al. Right ventricular performance and left ventricular assist device filling. Ann Thorac Surg 1997;63:1044 –9. 5. Charron M, Follansbee W, Ziady GM, Kormos RL. Assessment of biventricular cardiac function in patients with a Novacor left ventricular assist device. Transplant 1994;13: 263–7. 6. Morita S, Kormos RL, Mandarino WA, et al. Right ventricular/arterial coupling in the patient with left ventricular assistance. Circulation 1992;86(Suppl II):II316 –25. 7. Moon MR, Castro LJ, DeAnda A, et al. Effects of left ventricular support on right ventricular mechanics during experimental right ventricular ischemia. Circulation 1994;90:II92– 101. 8. Kormos RL, Gasior TA, Kawai A, et al. Transplant candidate’s clinical status rather than right ventricular function defines need for univentricular vs. biventricular support. J Thorac Cardiovasc Surg 1996;111:773– 83. 9. Argenziano M, Oz MC, Rose EA. The continuing evolution of mechanical ventricular assistance. Curr Probl Surg 1997: 321– 60. 10. Chen JM, Levin HR, Rose EA, et al. Experience with right ventricular assist devices for perioperative right-sided circulatory failure. Ann Thorac Surg 1996;61:305–10. 11. Van Meter CH, Robbins RJ, Ochsner JL. Technique of right heart protection and deairing during HeartMate vented electric LVAD implantation. Ann Thorac Surg 1997;63: 1191–2.
LESSONS LEARNED
Right Heart Failure: Best Treated by Avoidance Clifford H. Van Meter, Jr, MD Division of Cardiothoracic Surgery and Transplantation, Alton Ochsner Medical Foundation, New Orleans, Louisiana
Right heart failure continues to affect our clinical success with left ventricular assist device support. The inability to consistently predict the probability of the onset of right heart dysfunction contributes to this problem. We have developed an aggressive approach to the manage-
ment of these patients in an attempt to decrease the incidence of this condition, which continues to carry a very high operative mortality. (Ann Thorac Surg 2001;71:S220 –2) © 2001 by The Society of Thoracic Surgeons
D
ity and the dp/dt [3]. Whatever the effect of the LVAD on the RV, the early dependence of the LVAD on RV function is clear. Mandarino and associates [4] and others demonstrated that LVAD mean filling rate correlates with RV stroke work. This followed previous work showing an improvement in RV ejection fraction from 21% to 32% [5] in a mixture of ischemic and dilated patients, probably due to after-load reduction [6], which improves the mechanical efficiency of the RV [7]. In spite of these elegant clinical and experimental studies, we are all aware of chronically supported LVAD patients who have demonstrated acceptable hemodynamics for prolonged periods with little or no RV function as in episodes of ventricular fibrillation. Reviews of clinical and hemodynamic perimeters have suggested that clinical characteristics may be more relevant in predicting the need for RVAD support [9]. Factors indicating clinical status include: preoperative inotropic requirements; fever without infection; edema on CXR, RV wall thickness; and postoperative transfusion requirements. These parameters do not include such clinical factors as ascites, elevated BUN/creatine, and others that have been associated with increased operative mortality in general. They do, however, reflect preimplant degree of illness and factors that result in impairment of pulmonary blood flow or reduced RV perfusion that may be more predictive than preimplant measures of RV function or other hemodynamic variables alone. Argenziano and associates [9] and others at Columbia University have refined and broadened the definition of RV failure to include decrease in right ventricular function or output requiring the insertion of an RVAD, administration of inhaled nitric oxide for elevated pulmonary vascular resistance (PVR), or pharmacologic inotropic support for more than 10 days. Based on this, they rely on central venous pressure, transpulmonary gradient [10] and pulmonary artery diastolic pressure as an assessment of RV compromise and need for support, as well as a guide to patient management and support weaning. In our effort to identify and treat right heart compromise in our LVAD population, we have chosen to separate risk factors into preoperative (pulmonary hypertension; acute, sustained elevation of central venous
espite an ever-growing experience with left ventricular assist devices (LVAD), the recurring problem of right heart failure (RHF) continues to complicate clinical results. Patient and device selection, surgical timing, perioperative management techniques, and our inability to always predict a patient’s clinical response lead to the persistence of this issue. It has been well recognized that the use of concomitant (RVAD) support is associated with an increased operative mortality [1]. This reality may affect clinical practice by postponing the inevitable and thereby confirming the obvious. The persistent complication of RHF in the hands of experienced teams suggests, however, that one major contributing factor is our inability to always predict its likely occurrence [1]. Efforts to predict the incidence of RHF after LVAD implant have included an analysis of hemodynamic parameters, clinical parameters, and the effect of the LVAD itself on right heart function. Clearly, all of these are important factors. Studies of the interplay between right ventricular (RV) function and LVAD influence offer misleading impressions if viewed in isolation. Some have explained RV dysfunction as due to a loss of direct left ventricular mechanical coupling or as a consequence of extreme biventricular geometric alterations. It has been demonstrated that in normal hearts, marked alterations in biventricular geometry accompanying left ventricular unloading by the LVAD do not significantly alter RV performance characteristics [2]. Is the circumstance of heart failure different in the fact that some right hearts suffer from myopathies and others suffer from an environment of pulmonary artery (PA) hypertension, volume overload, and poor pulmonary perfusion pressure, or both? To date, no clinical study has suggested a higher rate of RV failure in ischemic versus dilated myopathies. Other experimental studies during LVAD support have shown a leftward shift of the septum, resulting in a loss of interventricular balance, thereby increasing RV compliance, and decreasing RV contractilPresented at the Fifth International Conference on Circulatory Support Devices for Severe Cardiac Failure, New York, NY, Sept 15–17, 2000. Address reprint requests to Dr Van Meter, Division of Cardiothoracic Surgery and Transplantation, Ochsner Clinic, 1514 Jefferson Hwy, 08s-LT, New Orleans, LA 70121; e-mail: [email protected].
© 2001 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
0003-4975/01/$20.00 PII S0003-4975(00)02637-0
Ann Thorac Surg 2001;71:S220 –2
pressure (CVP); dilated right-sided chambers; rightsided pleural effusions; hyperbilirubinemia; ascites; compromised pulmonary function or pulmonary edema; long-term afterload reduction; and marked renal insufficiency) and intraoperative (pharmacologic measures needed to maintain adequate perfusion pressure onpump; prolonged clamp or ischemic time; compromise of RV protection; need for femoral-femoral cannulation; acidosis; and increased airway pressures). Many preoperative factors such as pulmonary edema, hyperbilirubenemia, and acute CVP elevation or pulmonary hypertension can be effectively treated or modified for surgery in some clinical circumstances. Additionally, many intraoperative factors, if recognized and appreciated, can lead to modifications in management. Repeated pharmacologic requirements, intraoperatively, to maintain perfusion pressure should prompt the use of Vasopressin or other pressors postoperatively to maintain the perfusion pressure vital to avoid right-sided compromise. The need for femoral cannulation should lead to anticipation of acidosis upon reperfusion of the lower extremity after decannulation. Aggressive ventilatory adjustments should be employed to manage increased airway pressures or compromise pulmonary compliance. Finally, there needs to be an appreciation of processes that lead to right-sided circulatory failure. Transfusion requirements can lead to coagulopathies and/or enhanced inflammatory responses. Prolonged cardiopulmonary bypass and the state of class IV congestive heart failure can lead to inflammatory responses and an enhanced humoral amplification system of fibrolysis, complement activation, and the Bradykinin-Kallekrien system. These can all lead to the overwhelming clinical complex we have all seen and appreciated as “metabolic anarchy”. Systemic vascular endothelial dysfunction caused by long-term ACE inhibition and the effects of prolonged cardiopulmonary bypass, transfusion-induced cytokine release, and the impaired release of nitric oxide by altered pulmonary vascular endothelium further disrupt homeostasis. Intraoperative and postoperative measures can be employed to help manage these phenomena. Intraoperative measures include: patient selection and timing; serine protease inhibition; right heart protective techniques; leukocyte depletion; maintenance of perfusion and mean pressures (Vasopressin); and right heart venting or bypass. Postoperative measures include: aggressive treatment of acidosis; hyperventilation (hypocarbia); nitric oxide; A-V synchrony; and early postoperative continuous veno-venous hemofiltration (CVVH). Serine protease inhibition can lessen transfusion requirements and markedly decrease inflammatory responses and lung injury, possibly by minimizing the effects of NFKB. Leukocyte depletion can be employed in line in the cardiopulmonary bypass circuit, as well as transfusion filters. As previously mentioned, Vasopressin can be used to maintain perfusion pressure without excessive pressor support and without marked compromise to the splanchnic and renal circulation. We have utilized a right heart bypass circuit to prime the LVAD [11] via the left atrium
CIRCULATORY SUPPORT VAN METER RHF AVOIDANCE
S221
Fig 1. Right heart protection algorithm. (Bili ⫽ bilirubin (serum); CVP ⫽ central venous pressure; TPg ⫽ transpulmonary gradient; RVAD ⫽ right ventricular assist device; Pa Systolic ⫽ systolic pulmonary artery pressure; AV ⫽ atrioventricular; HR ⫽ heart rate; NO ⫽ nitric oxide; CVVH ⫽ continuous veno-venous hemofiltration; RV ⫽ right ventricle.)
with warm oxygenated blood, thereby avoiding right heart overload and the pulmonary circulation while exceeding the minimal amount of flow needed to keep the VAD above the minimal threshold to avoid aspiration of air into the system through the inflow graft. We hyperventilate, aggressively treat presumed acidosis, and assure atrioventricular synchrony in a relatively high heart rate of 90 to 100 bpm. More recently, we have begun using nitric oxide (NO) routinely. Finally, we do not hesitate to begin CVVH if there is sluggish urine output postoperatively or if the response to diuretics is poor or unpredictable. We have found that the right heart can quickly become volume overloaded if urine output is not brisk when multiple infusions and frequent transfusions are being administered. Once the pattern of right heart overload begins, it can progress quite rapidly to failure. From 1996 through 1999, 35 LVADs were placed at the Ochsner Foundation Hospital, utilizing the management techniques that we have discussed to avoid right heart failure. There were seven operative deaths (within 30 days of implant), or 20% operative mortality. Two patients suffered intracranial hemorrhages, 1 developed multiple cerebral emboli from native ventricular thrombus, 1 had intractable fibrillation on day 12, and 1 patient died of progressive hepatic failure and hepatorenal syndrome 1 month postoperatively. A final patient developed sepsis from pneumonia and on postoperative day
S222
CIRCULATORY SUPPORT RHF AVOIDANCE
VAN METER
12 had a pulmonary-hypertensive crisis with acute right heart failure and death. This single patient represents the only patient in whom right failure could be identified as a cause of operative mortality. No patients received right ventricular device insertion after LVAD placement. Fifty percent of the patients received NO perioperatively, and 25% required long-term inotropic support of the right heart for 10 to 14 days postoperatively. During this same interval, 2 patients received biventricular support with Thoratec pumps (Thoratec Laboratories Corp, Berkeley, CA), 1 based on extremely low body surface area, less than 1.2 mm2, and 1 due to the clinical circumstance of cardiogenic shock with extracorporeal membrane oxygenation support as a bridge to VAD insertion. Consideration of these issues and aggressive institution of measures to manage right heart challenge cannot be started too late nor terminated too prematurely. Generally, we have initially weaned an intraaortic balloon pump first, if present, followed by NO, and finally, any pharmacologic support. Throughout this process we keep a watchful eye on peak airway pressures, central venous pressure, pulmonary artery diastolic pressure, and fluid balance. More recently, we have chosen to initially limit VAD flows based on cardiac index, because in the automatic mode with maximal right heart protective measures, flows can become unnecessarily excessive. The avoidance of right heart failure is one of the challenges that must be managed if we are to be successful in the expanded indication of devices as alternatives to transplant and if we will be able to offer newer devices to larger populations of patients through less invasive surgical procedures. Until we have a clearer understanding of who is at risk and when, consideration must be
Ann Thorac Surg 2001;71:S220 –2
given to a broad spectrum of preventative measures and maneuvers.
References 1. McCarthy PM. HeartMate implantable left ventricular assist device: bridge to transplantation and future applications. Ann Thorac Surg 1995;59:S146–51. 2. Elbeery JR, Owen CH, Savitt MA, et al. Effects of the left ventricular assist device on right ventricular function. J Thorac Cardiovasc Surg 1990;99:809–16. 3. Santamore WP, Austin EH, Gray L, Jr. Overcoming left ventricular failure with ventricular assist devices. J Heart Lung TX 1997;16:1122– 8. 4. Mandarino WA, Winowich S, Gorscan J, et al. Right ventricular performance and left ventricular assist device filling. Ann Thorac Surg 1997;63:1044 –9. 5. Charron M, Follansbee W, Ziady GM, Kormos RL. Assessment of biventricular cardiac function in patients with a Novacor left ventricular assist device. Transplant 1994;13: 263–7. 6. Morita S, Kormos RL, Mandarino WA, et al. Right ventricular/arterial coupling in the patient with left ventricular assistance. Circulation 1992;86(Suppl II):II316 –25. 7. Moon MR, Castro LJ, DeAnda A, et al. Effects of left ventricular support on right ventricular mechanics during experimental right ventricular ischemia. Circulation 1994;90:II92– 101. 8. Kormos RL, Gasior TA, Kawai A, et al. Transplant candidate’s clinical status rather than right ventricular function defines need for univentricular vs. biventricular support. J Thorac Cardiovasc Surg 1996;111:773– 83. 9. Argenziano M, Oz MC, Rose EA. The continuing evolution of mechanical ventricular assistance. Curr Probl Surg 1997: 321– 60. 10. Chen JM, Levin HR, Rose EA, et al. Experience with right ventricular assist devices for perioperative right-sided circulatory failure. Ann Thorac Surg 1996;61:305–10. 11. Van Meter CH, Robbins RJ, Ochsner JL. Technique of right heart protection and deairing during HeartMate vented electric LVAD implantation. Ann Thorac Surg 1997;63: 1191–2.