The growing sheep model has been used extensively

Transcription

1 Behavior of Mitral Allografts in the Tricuspid Position in the Growing Sheep Model José M. Bernal, MD, José M. Rabasa, MD, Juan C. Cagigas, MD, Fernando Val, MD, and José M. Revuelta, MD Departments of Cardiovascular Surgery and Anatomic Pathology, Hospital Universitario Valdecilla, Universidad de Cantabria, Santander, Spain Background. On the basis of a previous experience in a chronic sheep model in which partial mitral allografts remained viable and properly functioning 12 months after operation, we assessed the results obtained by replacing the tricuspid valve with fresh antibioticpreserved mitral allografts. Methods. Twenty 3-month-old sheep with a mean weight of kg underwent cardiopulmonary bypass and had a fresh antibiotic-preserved mitral allograft implanted in the tricuspid position with the heart beating under normothermic conditions. The tricuspid valve apparatus was not excised. After a mean follow-up of 13.2 months, the allograft was evaluated by gross inspection and light and electron microscopy. Results. Nine sheep died of technical causes within the first week after operation and 2 at 4 and 6 months of infective endocarditis of the allograft. The hemodynamic study before heart explantation revealed residual tricuspid incompetence in 3 of the 9 survivors. Macroscopic examination showed flexible valves with no signs of structural deterioration, calcification, or thrombosis. Under light and scanning electron microscopic examination, allografts were almost completely denuded of endothelial cells and showed loosely arranged connective tissue with scarce signs of inflammatory reaction. Despite these findings, allografts were free from major structural damage. Conclusions. The mitral homograft could be an alternative to replacement of the tricuspid valve with a bioprosthesis or a mechanical prosthesis. (Ann Thorac Surg 1998;65: ) 1998 by The Society of Thoracic Surgeons The growing sheep model has been used extensively for the investigation of both valve repair techniques and durability of bioprosthetic materials. Since 1980, this model has been used in our surgical research laboratory for different research purposes [1 3]. Although early experiments with mitral valve allografts were carried out more than 40 years ago [4], experience with a sheep model in which partial mitral allografts remained viable and properly functioning 12 months after operation [3] renewed interest in the use of mitral allografts. Different groups [5 9] have subsequently reported the use of partial or complete cryopreserved mitral homografts for repairing or replacing the cardiac valves. On the other hand, other authors [10, 11] have obtained satisfactory initial results with the use of the stentless porcine mitral bioprosthesis. Despite this recent clinical information, the viability of mitral allografts has been evaluated in few experimental studies, and the results are sometimes conflicting [3, 12, 13]. In view of these reports, it was considered of interest to use the growing sheep model to assess the results obtained by replacing the tricuspid valve with fresh, antibioticpreserved mitral allografts. Accepted for publication Dec 22, Address reprint requests to Dr Bernal, Department of Cardiovascular Surgery, Hospital Universitario Valdecilla, E Santander, Spain. Material and Methods All animals used in this study received humane care in compliance with the Guide for the Care and Use of Laboratory Animals (NIH publication 85-23, revised 1985) and the Principles of Laboratory Animal Care formulated by the National Society for Medical Research and approved by the Institutional Animal Care and Use Committee of the University of Cantabria. Preparation of Allografts Mitral allografts were procured through the municipal abattoir of Camargo, Santander, from industrially sacrificed young sheep. The heart was explanted under nonsterile conditions and transferred to the laboratory of experimental cardiac surgery where under surgically sterile conditions, mitral allografts were immediately prepared with a rim of left atrial muscle and aortic valve left around the mitral annulus. Only valves with two single papillary muscles were considered suitable for implantation. The size of the valve annulus was measured to choose the most appropriate allografts for implantation. In all cases, the time between heart explantation and dissection of the mitral allograft was less than 2 hours. Homografts were immersed for 24 hours in a modified Hanks solution with a ph of 6.8 to 7.0 and the following antibiotic composition: cefoxitin sodium (240 mg/ml), 1998 by The Society of Thoracic Surgeons /98/$19.00 Published by Elsevier Science Inc PII S (98)

2 Ann Thorac Surg BERNAL ET AL 1998;65: EXPERIMENTAL MITRAL ALLOGRAFTS 1327 lincomycin hydrochloride monohydrate (120 mg/ml), vancomycin (50 mg/ml), polymyxin B sulfate (100 mg/ml), and amphotericin B (25 mg/ml). Samples for bacterial culture were obtained, and the fresh mitral allografts were then moved to a TC 199 culture medium. They were maintained at 4 C until implantation (range, 3 to 9 days; mean duration, 4.4 days). Only sterile allografts were considered adequate for implantation. Surgical Implantation Twenty sheep 3 months old with a mean weight of kg (range, 21 to 26 kg) were obtained from the Faculty of Veterinary Medicine in León, Spain. The recipients of the allograft valves were given no food for 24 hours before the operation. Anesthesia was induced and maintained with sodium thiopental, 5 mg/kg. Cefamandole rafate, 1 g, was given intravenously at the time of induction of anesthesia and 6 hours after the end of operation as a prophylactic antibiotic. The heart was exposed through a right anterolateral thoracotomy in the fourth intercostal space. The right femoral artery was cannulated for arterial perfusion, and the cava veins were cannulated for venous return. During cardiopulmonary bypass, the heart was maintained beating under normothermic conditions, as the aorta was not cross-clamped. Mitral allografts were prepared for implantation before cardiopulmonary bypass was instituted. All tissue surrounding the mitral annulus was removed, and the papillary muscles were transected 8 to 10 mm from their cusps. A 3-0 pledget-supported polypropylene suture was passed through each papillary muscle so that pledgets were placed at the site of insertion of the chordae tendineae. The tricuspid valve was exposed through a right atriotomy parallel to the atrioventricular sulcus, and the coronary sinus was vented. The mitral allograft was implanted with the normal anatomic orientation preserved, ie, the septal leaflet of the allograft was sutured to the septal segment of the tricuspid annulus. As previously described [14], the tricuspid valve apparatus was not excised (transvalvular implantation). The diameter of the mitral allografts ranged between 27 and 31 mm (31-mm allografts were the most commonly implanted). The implantation technique varied during the study period according to the results obtained. In the first 10 animals (group 1), each papillary muscle was fixed to the corresponding recipient point (endocardial implantation) with sutures. Implantation of the allograft around the tricuspid annulus was completed with a 4-0 polypropylene running suture. In the following 10 animals (group 2), the surgical technique was modified, and the papillary muscles were anchored to the free edge of the right ventricular wall using a Teflon felt reinforced suture. The most appropriate sites for preserving the topographic anatomy of the graft were selected in the right ventricle, and the needles of the 3-0 pledget-supported polypropylene suture were passed through the ventricular wall thickness to the epicardial surface. The graft annulus was sutured directly to the native tricuspid septal leaflet, with care taken to avoid the atrioventricular node. The atriotomy was repaired, and the animal was separated from cardiopulmonary bypass. The femoral artery was repaired, and the sternotomy was closed. After operation, the animals were maintained in a heated postoperative recovery room for 24 hours. Once they had recovered, they were returned to the animal care facility. Postoperative analgesia consisted of butorphanol tartrate, 0.5 mg intramuscularly every 8 hours for the first 24 hours. The condition of the animals was evaluated daily by us and by a veterinary technician for 7 days after the operation. Animals were then transferred to the farm and maintained by conventional ovine husbandry techniques. Explantation of Allografts After a mean follow-up of 13.2 months (range, 12 to 15 months), animals were transferred to the research unit. Anesthesia was induced and maintained as before. A catheter was positioned in the external jugular vein, and right atrial and right ventricular pressures were recorded simultaneously. A dose of heparin sodium, 3 mg/kg, was administered, and the animal was killed with a single dose of sodium thiopental, 30 mg/kg. After the heart was removed, the inflow surface of the valve was examined. The free wall of the right ventricle was opened to examine the outflow surface, the chordae tendineae, and the papillary muscles. Each valve was divided into two halves with their respective leaflets, chordae tendineae, papillary muscles, and right ventricular wall portions. One half of each graft was submitted for histologic study, and the other was examined under scanning electron microscopy. The tissue samples were fixed with 2.5% glutaraldehyde in a solution of cacodylate buffer (ph 7.3), critical-point dried, coated with gold/palladium by the ion sputtering method, and examined in an SEM 501 scanning electron microscope. Results Seven sheep from group 1 (n 10) and 2 from group 2 (n 10) died within the first week after operation of complete atrioventricular block (n 3), tears in the interventricular septum or the free wall of the right ventricle (n 3), heart failure (n 2), and disruption of the papillary muscles of the graft (n 1). These 9 animals were excluded from the study. Eleven sheep were transferred to the farm, and 2 of them died at 4 and 6 months after operation of infective endocarditis of the allograft, which was demonstrated by microscopic examination. Nine animals survived and were electively killed. The hemodynamic study performed immediately before each animal was killed demonstrated a mean right atrial pressure of mm Hg and mean systolic and telediastolic right ventricular pressures of mm Hg and mm Hg, respectively. Simultaneous recording of the right atrial and ventricular pressures showed residual tricuspid incompetence in 3 of the 9 sheep. Mean heart rate was beats per minute. Macroscopic examination of the heart showed no intracardiac thrombi or signs of infective endocarditis (Figs

3 1328 BERNAL ET AL Ann Thorac Surg EXPERIMENTAL MITRAL ALLOGRAFTS 1998;65: Fig 3. Homograft leaflet almost denuded of endothelial cells. Fibroblasts and collagen fibers have replaced the connective tissue of the native valve. There are few inflammatory mononuclear cells. (Hematoxylin and eosin; 40 before 52% reduction.) Fig 1. Mitral homograft implanted in the tricuspid position in a sheep for 13 months. The graft has a whitish appearance and is covered by a fibrous sheath without structural damage to the chordae tendineae. The papillary muscle of the posterior leaflet (P) is implanted on the papillary muscle of the right ventricle. 1, 2). Leaflets of the allografts had a whitish appearance, except for the free margins, although they retained their original flexibility. Signs of structural deterioration, such as calcification, rupture, or shrinkage of the leaflets, were not observed. The chordae tendineae were apparently normal and had not lost their elasticity. The heads of the papillary muscles were covered by fibrous sheaths. In two functionally incompetent valves, prolapse of one or both leaflets without elongated or ruptured chordae was probably caused by inappropriate implantation of the papillary muscles (too close to the tricuspid ring). Rupture of the primary chordae was found in the other regurgitant valve. Incorporation of the allograft into the valve annulus was visible only because of the polypropylene suture. In all instances, the graft diameter was 1 to 4 mm smaller than it was before operation. Examination with a light microscope revealed some important features (Fig 3). The leaflet surfaces of the allograft were covered by fibrous connective tissue with collagen-producing fibroblasts that had replaced the connective tissue of the native valve. It showed a poorly structured collagen network. The thickness of the fibrous sheathing was greater on the valve ring and the atrial aspect than on the free edge and the ventricular aspect. Most portions of the allograft were completely denuded of endothelial cells. Discrete signs of infiltration of mononuclear cells in the native valve leaflets were observed. Two allografts showed chondroid and bone metaplasia with bone marrow production in the leaflet of the implanted valve. Calcifications were not seen except for a graft leaflet with an area of dystrophic calcification. The chordal stroma was composed of connective tissue and collagen fibers without infiltration by mononuclear cells or calcification (Fig 4). The chordal surface was com- Fig 2. Mitral homograft implanted in the tricuspid position in a sheep for 15 months. Chordae tendineae are those of the native tricuspid valve. Fig 4. Chorda tendinae of mitral homograft explanted after 12 months. Stroma are formed by connective tissue with fibroblasts and collagen fibers. There are no signs of infiltration of mononuclear cells. (Hematoxylin and eosin; 40 before 40% reduction.)

4 Ann Thorac Surg BERNAL ET AL 1998;65: EXPERIMENTAL MITRAL ALLOGRAFTS 1329 Fig 5. Ultrastructure of the leaflet of a mitral homograft explanted at 12 months shows an acellular surface. Collagen fibers are loosely arranged and show a tendency to form circles. (Uranyl acetate and lead citrate; 640 before 52% reduction.) pletely acellular. The stroma of the ruptured chorda was acellular with fibrin and polymorphonuclear cells and inflammatory mononuclear cells in the rupture zone. Scanning microscopic examination revealed coverage of leaflets and chordae by collagen fibers. They were loosely arranged with a tendency toward circles in the leaflets (Fig 5), and they were oriented chiefly along the main axis in the chordae tendineae (Fig 6). Comment Although a number of clinical and experimental studies of the feasibility of mitral valve allografts were undertaken before clinical research on aortic valve allografts [4, 15], homograft replacement of the mitral valve was soon abandoned because of disappointing surgical results stemming from technical difficulties and tissue viability related problems [16]. In contrast, more than three decades later, aortic homografts are considered the best substitutes for aortic valve replacement. Interest in mitral valve homografts has revived over this decade. On the one hand, laboratory investigations Fig 6. Ultrastructure of a chordae tendinae in a mitral allograft explanted at 14 months shows a completely denuded surface. (Uranyl acetate and lead citrate; 80 before 52% reduction.) [3] with a chronic sheep model showed that fresh or cryopreserved mitral allografts remained viable and properly functioning 12 months after operation. This study opened the way to a novel concept of mitral valve repair, ie, the use of a partial mitral homograft to replace the corresponding segment in a diseased valve. The first successful partial mitral homograft implantation was performed in 1992 [7]. On the other hand, Vrandecic [10], Morea [11], and their colleagues have obtained satisfactory initial results with the stentless porcine mitral bioprosthesis. In contrast to the clinical experience [5, 6, 8, 9] recently reported, the scarcity of experimental data to support scientifically the use of mitral homografts in humans is surprising. Since the first attempts in the experimental laboratories, reported in 1954 by Robicsek [4], in 1964 by Cachera and coworkers [17], and in 1965 by Rastelli and associates [15] and Van Vliet and colleagues [18] to the present time, mitral allograft implantation has been poorly investigated in animal models, sometimes with controversial results. The experimental study [3] in which a partial mitral allograft was used to replace a segment of the mitral valve showed excellent morphologic and functional results. In the sheep model, Vetter and coauthors [12] documented by conventional echocardiography unimpaired valvular motion of a mitral allograft with expanded polytetrafluoroethylene chordal support after 5 months of implantation. These results are in contrast with those in a recent report by Tamura and associates [13]. In an elegant experimental study using electron microscopy, they compared failure modes of glutaraldehyde-treated versus antibiotic-preserved mitral allografts implanted in sheep. Valves were explanted at 19 to 24 weeks. Glutaraldehydetreated allografts failed because of extensive calcification involving the chordae and the leaflets, thus resulting in loss of valvular mobility. Antibiotic-preserved grafts failed because of chordal rupture and cuspal perforations secondary to deterioration of connective tissue elements, particularly collagen. According to Tamura and associates, the connective tissue breakdown in the valves of fresh mitral allografts can result from several factors including autolytic changes before implantation, host inflammatory reaction, and wear and tear associated with the mechanical forces related to the opening and closing of the allograft. These results alerted surgeons to the weak points at which rupture can occur, ie, the papillary muscle anastomosis and the chordal insertions. In the case of a patient reported by Yankah and coworkers [19], late graft failure (44 months after operation) was due to rupture of the scarred graft papillary muscle. Our study was conducted to assess the behavior of mitral allografts implanted in the tricuspid position in the chronic sheep model. For reasons involving the surgical technique and because of anatomic variations in the papillary muscles in the right ventricle, the graft papillary apparatus was implanted mainly in the free ventricular wall. Transventricular papillary muscle fixation has already been described even for allograft mitral valve replacement [19]. Implantation of mitral homografts is

5 1330 BERNAL ET AL Ann Thorac Surg EXPERIMENTAL MITRAL ALLOGRAFTS 1998;65: more difficult than implantation of aortic homografts, because the papillary muscle anastomosis is the most important technical aspect for surgical success. Preservation of the native valve and transventricular fixation allows the surgeon not only to establish proper anatomic orientation and alignment of the subvalvular apparatus, but also to maintain its shock-absorbing capacity. Early mortality was high as a result of technical conditions in the sheep model, ie, the need to perform the procedure on a beating heart. Nevertheless, none of the surviving animals showed major tricuspid insufficiency at 12 to 15 months after operation. Two animals died of infective endocarditis, demonstrated by microscopic examination. The explantation of the hearts was not performed under sterile conditions, and the preparation of the surgical field in sheep is difficult because of their thick wool. These factors might have contributed to contamination of the allografts. This incidence of infective endocarditis in the sheep model has been noted by other investigators [12, 13]. Results of macroscopic examination and light and scanning microscopy agree in part with the experience of Tamura and associates [13]. Fresh mitral allografts treated with antibiotics were almost completely denuded of endothelial cells. The connective tissue was highly disorganized, and collagen fibrils in the fibrous sheath tended to exhibit a circular arrangement. The absence of chordal rupture and leaflet perforation, in our experience, may be explained by the scarcity of mononuclear cell infiltration of the leaflets and the absence of signs of inflammatory reaction in the chordae tendineae. Implantation of a mitral allograft in the tricuspid position where atrial and ventricular tensions are much lower than in the left-sided position may partially explain discrepancies in the results of different studies. In our opinion, the experimental data available are not sufficient to liberalize the use of complete mitral homografts for the replacement of mitral valves. In surgical candidates with tricuspid endocarditis, implantation of a mitral homograft may be an alternative to the classic valvectomy without valve replacement. The same may apply to patients with severe tricuspid valve disease if repair is not feasible. Neither bioprostheses nor mechanical prostheses constitute an adequate valve substitute. Although mitral homografts may be a valid possibility, further experience is needed to assess the performance of mitral allografts in the tricuspid position. Supported by research grant 93/0495 from Fondo de Investigaciones Sanitarias, Madrid, Spain. We are indebted to Marta Pulido, MD, for editing the manuscript and to Professor Jose L. Ojeda for examining the scanning electron micrographs. References 1. Gallo I, Frater RWM. Experimental atrioventricular bioprosthetic valve insertion: a simple and successful technique. Thorac Cardiovasc Surg 1983;31: Revuelta JM, Garcia-Rinaldi R, Gaite L, Val F, Garijo F. Generation of chordae tendineae with polytetrafluoroethylene stents. J Thorac Cardiovasc Surg 1989;97: Revuelta JM, Cagigas JC, Bernal JM, Val F, Rabasa JM, Lequerica MA. Partial replacement of mitral valve by homograft. An experimental study. J Thorac Cardiovasc Surg 1992;104: Robicsek F. Cardiac valve transplantation. Acta Med Acad Sci Hung 1954;5: Pomar JL, Mestres C-A, Pare JC, Miro JM. Management of persistent tricuspid endocarditis with transplantation of cryopreserved mitral homografts. J Thorac Cardiovasc Surg 1994;107: Di Summa M, Donegani E, Zattera GF, Pansini S, Morea M. Successful orthotopic transplantation of a fresh tricuspid valve homograft in a human. Ann Thorac Surg 1993;56: Revuelta JM, Bernal JM, Rabasa JM. Partial homograft replacement of mitral valve. Lancet 1994;344: Acar C, Farge A, Ramsheyi A, et al. Mitral valve replacement using a cryopreserved mitral homograft. Ann Thorac Surg 1994;57: Acar C, Tolan M, Berrebi A, et al. Homograft replacement of the mitral valve. Graft selection, technique of implantation, and results in forty-three patients. J Thorac Cardiovasc Surg 1996;111: Vrandecic MP, Gontijo BF, Fantini FA, et al. Heterologous mitral valve transplant: the first 50 patients clinical analysis. Eur J Cardiothorac Surg 1995;9: Morea M, DePaulis R, Galloni M, Gastaldi L, di Summa M. Mitral valve replacement with the Biocor stentless mitral valve: early results. J Heart Valve Dis 1994;3: Vetter HO, Dagge A, Liao K, et al. Mitral allograft with chordal support: echocardiographic evaluation in sheep. J Heart Valve Dis 1995;4: Tamura K, Jones M, Yamada I, Ferrans VJ. A comparison of failure modes of glutaraldehyde-treated versus antibioticpreserved mitral valve allografts implanted in sheep. J Thorac Cardiovasc Surg 1995;110: Revuelta JM, Bernal JM, Rabasa JM. Transvalvular technique for implantation of a mitral valve homograft. J Thorac Cardiovasc Surg 1996;111: Rastelli GC, Berguis J, Swan HJC. Evaluation of the function of the mitral valve after homotransplantation in the dog. J Thorac Cardiovasc Surg 1965;49: Sievers HH, Lange PE, Yankah AC, Wessel A, Bernhard A. Allogeneous transplantation of the mitral valve: an open question. Thorac Cardiovasc Surg 1985;33: Cachera JP, Salvatore L, Hermant J, Herbinet B. Reconstructions plastiques de l appareil mitral chez le chien au moyen de valves mitrales homologues conservés. Raport préliminaire. Ann Chir Thorac Cardiovasc 1964;3: Van Vliet PD, Titus JK, Berghis J, Ellis FH Jr. Morphologic features of homotransplanted canine mitral valves. J Thorac Cardiovasc Surg 1965;49: Yankah AC, Sievers HH, Lange PE, Bernhard A. Clinical report on stentless mitral allografts. J Heart Valve Dis 1995; 4:40 4.