Partial Aortic Valve Fusion Induced by Left Ventricular Assist Device

Transcription

1 Partial Aortic Valve Fusion Induced by Left Ventricular Assist Device Alan G. Rose, MD, Soon J. Park, MD, Alan J. Bank, MD, and Leslie W. Miller, MD Department of Laboratory Medicine and Pathology, Division of Cardiovascular and Thoracic Surgery, and Division of Cardiology, University of Minnesota and Fairview-University Medical Center, Minneapolis, Minnesota Background. Left ventricular assist devices (LVADs) may be used (1) as a bridging device to cardiac transplantation, (2) for permanent replacement of left ventricular function, and (3) as a bridge to recovery of ventricular function, for example, in recoverable myocardial disease. In this third group of patients, it is important that the LVAD does not produce changes in the heart that will have a deleterious effect on cardiac function once the device is removed. Furthermore, if the LVAD fails, survival depends on optimal function of the diseased heart. Methods. All hearts with LVADs encountered as surgical specimens following heart transplantation or at autopsy at the Fairview-University of Minnesota Medical Center during the 5-month period August 1998 to January 1999 were examined for native valvular heart disease. The nature and extent of commissural fusion was noted and measured. Light microscopy was performed on any valve lesions. Results. Four of 6 patients with HeartMate (Thermo Cardiosystems, Inc, Woburn, MA) LVADs showed evidence of commissural fusion (acquired aortic stenosis). In 1 patient, this condition was caused by an organizing thrombus uniting a 14-mm length of the commissural region of the right coronary and noncoronary cusps of the aortic valve. Fibrous commissural fusion due to totally organized thrombus in the other 3 patients affected one aortic commissure (2 patients, 2 mm and 4 mm, respectively) and two commissures (1 patient, 2 mm and 5 mm). Partial cuspal fusion in each case was due to permanent closure of the native aortic valve induced by the LVAD s operating in its automatic setting. Mean length of commissural fusion was 5.4 mm (range, 2 to 14 mm; standard deviation [SD] 5.0 mm). Mean duration of implantation of the six LVADs was days (range, 26 to 689 days; SD days). The LVADs of the 3 patients with fibrous fusion of the commissures had been implanted for an average of days (range, 26 to 689 days; SD days). Conclusions. Normal function of the LVAD produces permanent closure of the native aortic valve. Stasis on the ventricular aspect of the aortic valve, combined with a low level of anticoagulation, favors thrombosis at this site. Thrombus organization leads to aortic stenosis of variable severity. This previously unsuspected complication was not detected clinically in any of our patients. Aortic stenosis may hold serious implications for patients in whom the LVAD acts as a bridge to recovery or in those in whom the LVAD fails. Prevention may be achieved by intermittently reducing LVAD pumping action. A built-in venting cycle would be of value in long-term implants. Thrombi on the aortic valve may also predispose patients to infective endocarditis, because bloodstream infection is common in patients with LVADs. (Ann Thorac Surg 2000;70:1270 4) 2000 by The Society of Thoracic Surgeons The left ventricular assist device (LVAD) has proved to be an effective bridge to transplantation [1 5]. It has recently been shown that the use of a LVAD has allowed some patients with dilated cardiomyopathy to recover sufficient native myocardial function to allow the pump to be removed [6, 7]. The University of Minnesota is one of the sites at which a clinical trial of LVAD (HeartMate; Thermo Cardiosystems, Inc, Woburn, MA) as an alternative to medical therapy for patients who are not transplant candidates is currently under way. Disadvantages of current mechanical devices [5, 8] include the following: a variable propensity for thromboembolic complications necessitating systemic anticoagulation (and its attendant Accepted for publication April 20, Address reprint requests to Dr Rose, Department of Laboratory Medicine and Pathology, University of Minnesota Medical School, MMC 76, Mayo Building, 420 Delaware St SE, Minneapolis, MN 55455; rosex031@tc.umn.edu. complications), infection of the transcutaneous connections, and possible device malfunction. A low thromboembolic risk without anticoagulation using advanceddesign LVADs has been reported [9]. Nevertheless, although procoagulant and fibrinolytic pathways are apparently balanced in such patients, the presence of the LVAD leads to significant thrombin generation and corresponding fibrinolysis. This underscores the potential for the development of bleeding or thrombosis in clinically relevant settings [10]. The purpose of this communication is to draw attention to the possible development of aortic valve stenosis in recipients of LVADs and to suggest a means of preventing this potentially serious complication. Material and Methods During the period January 1995 to January 1999, 26 HeartMate LVADs have been implanted at the Fairview by The Society of Thoracic Surgeons /00/$20.00 Published by Elsevier Science Inc PII S (00)

2 Ann Thorac Surg ROSE ET AL 2000;70: LVAD-INDUCED AORTIC STENOSIS 1271 Fig 1. Heart 1. (A) Organizing fibrin thrombus unites anterior aspect of commissure between right coronary cusp (RCC; left) and noncoronary cusp (NCC; right). (B) Aortic aspect of the aortic valve shown in (A) showing the fusing commissure between RCC (left) and NCC (right). A portion of the left ventricular assist device outflow anastomosis in the aorta distal to the aortic valve is present (top). (C) Organizing intercuspal fibrin-platelet thrombus (left) and portion of RCC (right); artifactual disruption has occurred during sectioning. (Hematoxylin and eosin; 70.) University Medical Center. All patients with implants received one aspirin daily for anticoagulation, and the LVAD pump was vented once per nursing shift. Left ventricular function stayed depressed in most patients on echo assessment. The native aortic valve stayed closed during the automatic pumping mode. All hearts bearing a HeartMate LVAD, including hearts explanted at cardiac transplantation and those encountered at autopsy in the Department of Laboratory Medicine and Pathology at Fairview-University of Minnesota Medical Center during the period August 1998 to January 1999, were examined for evidence of valvular heart disease. Careful note was made of any valvular alterations. The length and nature of commissural fusion was recorded. The valves were photographed, and histologic sections were made of any valvular lesions encountered. The commissures were sectioned horizontally to include the aortic wall as well as both cusps. All valvular tissue was processed by routine paraffin embedding. Sections were stained by hematoxylin and eosin and by Verhoeff s elastic stain with a van Gieson counterstain. Table 1. Clinicopathologic Data on 5 Patients With HeartMate LVAD Patient No. Age (y) Sex Cardiac Disease Duration of LVAD Placement (d) Aortic Valve 1 52 male IHD 152 RCC and NCC joined by 14- mm organizing thrombus 2 52 male IHD 118 Normal 3 53 female DCM mm fusion of LCC/NCC and 5-mm fusion of RCC/LCC 4 46 male DCM 55 Normal 5 68 female IHD 26 2-mm fusion of LCC/NCC 6 57 male IHD 42 4-mm fusion of RCC/NCC DCM dilated cardiomyopathy; IHD ischemic heart disease; LCC left coronary cusp; LVAD left ventricular assist device; NCC noncoronary cusp; RCC right coronary cusp.

3 1272 ROSE ET AL Ann Thorac Surg LVAD-INDUCED AORTIC STENOSIS 2000;70: Results We encountered six hearts with an implanted HeartMate LVAD; three were explanted recipient hearts following cardiac transplantation, and three were autopsy hearts. Five out of the six HeartMate LVADs (those in hearts 1, 2, 3, 4, and 6) were pneumatically powered (model 1000 IP devices), and one heart (heart 5) had the HeartMate VE (vented electrical) model. Clinicopathologic data are given in Table 1 and Figures 1, 2, and 3. Morphologic evidence of acquired aortic valvular disease (Table 1) was encountered in four of the six hearts that had been linked to a LVAD during life. The extent of the commissural fusion (which was purely fibrous in 3 patients and was a mixture of fibrous and organizing thrombus in 1 patient) is indicated in Table 1. Mean length of commissural fusion was 5.4 mm (range, 2 to 14 mm; SD 5.0 mm). No correlation was found between the duration of LVAD implantation and the generation of aortic valve thrombosis and commissural fusion. Mean duration of implantation of the six LVADs was days (range, 26 to 689 days; SD days). The LVADs in the 3 patients with firm, fibrous union of the affected commissures had been implanted for days (range, 26 to 689 days; SD days). Patient 1, whose device had been implanted for the second longest period (152 days) showed a greater (14 mm) extent of cuspal apposition by an organizing intercuspal thrombus that had been deposited on the ventricular aspect of the left coronary and noncoronary cusps (Fig 1). Patient 3 was a woman with dilated cardiomyopathy whose LVAD had been implanted for 689 days. Her myocardial function had improved in the interim, and the patient had been discharged from hospital, but the LVAD had to be removed because of drive-line sepsis. A month later the patient suddenly collapsed at home, was partially resuscitated, and died shortly thereafter. Autopsy revealed fusion of two com- Fig 2. Heart 3. (A) Fibrous fusion of commissure between left coronary cusp and noncoronary cusp (next to forceps). The fusion between the right coronary and left coronary cusps was disrupted before photography. (B) Histologic section of fused portion of left coronary and noncoronary cusps. The original valve cusps appear dark red, and the intervening area of fusion appears paler and represents less dense, more recently formed collagen, probably due to organization of an intercuspal fibrin deposit. (Elastic van Gieson stain; 20.) Fig 3. Heart 5. Macroscopic appearance of healed, fibrous commissural fusion between left coronary cusp (right, held by forceps) and noncoronary (left) cusp of aortic valve.

4 Ann Thorac Surg ROSE ET AL 2000;70: LVAD-INDUCED AORTIC STENOSIS 1273 Fig 4. (A) Healed, fibrous commissural fusion of two aortic valve cusps in an additional patient not discussed in the article, whose LVAD had been implanted for 165 days (see Addendum). (B) Transverse section of fused commissure. (Elastic van Gieson stain; 20.) missures of her aortic valve (Fig 2). Because the valvular complication had been unsuspected, the echocardiogram performed just before removal of the LVAD had not specifically visualized the aortic valvular commissures, nor had it measured aortic valve gradients. Patient 5, whose device had been implanted for the shortest period (26 days), showed healed, fibrous fusion of a single commissure (Fig 3). Comment A recent comprehensive review of implantable LVADs [11] lists only the following as possible complications of LVADs: (1) the most common early complications are bleeding, right-sided heart failure, air embolism, and progressive multisystem organ failure; and (2) the most common late complications are infection, thromboembolism, and device failure. Although the review discusses hemodynamic features of the blood flow in LVAD recipients with various valvular lesions, no mention is made of any valvular complications in the native heart resulting from the LVAD. It was recommended that patients with preexisting mitral stenosis or aortic regurgitation may require correction of the valvulopathy before implantation of a LVAD [11]. As far as we are aware, there has been no previous report describing aortic valve pathology induced by a LVAD. The development of aortic stenosis will have a less deleterious effect in patients scheduled for biological cardiac transplantation, but it holds potentially serious implications for the following populations: (1) patients in whom the LVAD is aimed to serve as a bridge to recovery [6], and (2) patients whose device fails. Muller and colleagues [6] have described 5 patients with idiopathic dilated cardiomyopathy treated with long-term LVAD therapy who were successfully weaned from mechanical support. The wearable electrical devices currently available have external backup mechanisms to continue temporary support [11, 12]. If the backup mechanism should not be available or fail, the native heart must provide systemic support until the device can be repaired. Under such circumstances, the presence of significant aortic stenosis could have highly unfavorable effects on the function of a diseased left ventricle that has not been called upon to support the full circulation for many months or years.

5 1274 ROSE ET AL Ann Thorac Surg LVAD-INDUCED AORTIC STENOSIS 2000;70: When set in the automatic mode, the HeartMate LVAD automatically ejects when it reaches 90% of filling capacity. This keeps the left ventricle unloaded, and the aortic valve remains permanently closed [13]. Stasis of blood on the ventricular aspect of the permanently immobilized aortic valve cusps (more particularly in the commissural region), combined with low anticoagulation, favors development of thrombi in this region. Increased thrombin generation in patients with LVAD acts as an additional factor favoring thrombosis. Thrombi are unlikely to form on the aortic aspect of the aortic valve or in the sinuses of Valsalva, which are washed by blood pumped in from the LVAD. The native aortic valve will only open when pressure in the left ventricle exceeds that in the ascending aorta, as happens, for example, during the regular cycle of venting the pneumatically powered drive line or during exercise. With exercise (a rare event in patients with LVADs), increased preload results in a more forceful left ventricular contraction and in Doppler evidence of significant flow across the native aortic valve [14, 15]. Organization of such thrombus leads to commissural fusion and acquired aortic stenosis. The severity of the latter will depend on the number and extent of the aortic valvular commissural fusions. Fusion of a single commissure leads to an acquired bicuspid aortic valve (as seen in heart 1 and, to a lesser extent, in heart 5). Fusion of two commissures was observed in heart 3. Fusion of all three commissures, which was not observed in this study, may also be predicted to occur. The aortic valvular stenosis resulting from the LVAD is easily distinguishable from senile calcific aortic stenosis, which does not produce commissural fusion and shows cuspal calcification [16]. Chronic rheumatic aortic stenosis shows commissural fusion, but the valve cusps are diffusely thickened, fibrosed, and vascularized and may show inflammation. The mitral valve is also usually affected in patients with rheumatic aortic stenosis. In aortic stenosis associated with LVAD-induced commissural fusion, the aortic leaflets appear normal apart from the commissural fusion. Infection is common in patients with implanted LVADs. In the series reported by McCarthy and colleagues [17], 55% of patients had bloodstream infection during LVAD support. The presence of thrombus on the aortic valve may predispose to infective endocarditis, the pathogenesis of which involves infection of a preexisting bland valvular thrombus. In future clinical examination of patients with LVADs, we shall be paying particular attention to the native aortic valve. In theory, thrombi on the aortic valve cusps and partial commissural fusion may be prevented by allowing intermittent opening of the aortic valve by left ventricular systole, induced by periodic venting of the LVAD. A built-in pump-venting cycle would be of value in long-term implants. Addendum Subsequent to the submission of this manuscript, we have encountered 3 additional patients with partial aortic valve commissural fusion associated with the presence of a LVAD. The heart of 1 of these patients, who had a LVAD for 165 days, is illustrated in Figure 4. References 1. De Vries WC, Anderson JL, Joyce LD, et al. Clinical use of the total artificial heart. N Engl J Med 1984;310: Griffith BP, Hardesty RL, Kormos RL, et al. Temporary use of the Jarvik-7 total artificial heart before transplantation. N Engl J Med 1987;316: Portner PM, Oyer PE, Pennington DG, et al. Implantable left ventricular assist system: bridge to transplant and the future. Ann Thorac Surg 1989;47: Kormos RL, Murali S, Dew MA, et al. Chronic mechanical circulatory support, rehabilitation, low morbidity, and superior survival. Ann Thorac Surg 1994;57: Hunt SA. Current status of cardiac transplantation. JAMA 1998;280: Muller J, Wallukat G, Weng Y-G, et al. Weaning from mechanical cardiac support in patients with idiopathic dilated cardiomyopathy. Circulation 1997;96: Rose EA, Frazier OH. Resurrection after mechanical circulatory support. Circulation 1997;96: Scott-Burden T, Tock CL, Bosley JP, Schwarz JJ, Engler DA, Casscells SW. Genetically engineered cellular linings for ventricular assist devices [Abstract]. Cardiovasc Pathol 1998;7: Slater JP, Rose EA, Levin HR, et al. Low thromboembolic risk without anticoagulation using advanced-design left ventricular assist devices. Ann Thorac Surg 1996;62: Spanier T, Oz M, Levin H, et al. Activation of coagulation and fibrinolytic pathways in patients with left ventricular assist devices. J Thorac Cardiovasc Surg 1996;112: Goldstein DJ, Oz MC, Rose EA. Implantable left ventricular assist devices. N Engl J Med 1998;339: Rose EA, Goldstein DJ. Wearable long-term mechanical support for patients with end-stage heart disease: a tenable goal. Ann Thorac Surg 1996;61: McCarthy PM, Nakatani S, Vargo R, et al. Structural and ventricular histologic changes after implantable LVAD insertion. Ann Thorac Surg 1995;59: Branch KR, Dembitsky WP, Peterson KL, et al. Physiology of the native heart and Thermo Cardiosystems left ventricular assist device complex at rest and during exercise: implications for chronic support. J Heart Lung Transplant 1994;13: Jaski BE, Branch KR, Adamson R, et al. Exercise hemodynamics during long-term implantation of a left ventricular assist device in patients awaiting heart transplantation. J Am Coll Cardiol 1993;22: Rose AG. Etiology of valvular heart disease. Curr Opinion Cardiol 1996;11: McCarthy PM, Schmitt SK, Vargo RL, Gordon S, Keys TF, Hobbs RE. Implantable LVAD infections: implications for permanent use of the device. Ann Thorac Surg 1996;61: