| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 1995;110:224-238
© 1995 Mosby, Inc.
SURGERY FOR ACQUIRED HEART DISEASE |
Bethesda, Md.
Received for publication April 28, 1994. Accepted for publication Sept. 29, 1994. Address for reprints: Victor J. Ferrans, MD, PhD, Chief, Ultrastructure Section, Pathology Branch, NHLBI, Bldg. 10/2N240, National Heart, Lung and Blood Institute, National Institutes of Health, 9000 Rockville Pike, Bethesda, MD 20892.
Abstract
Morphologic studies and calcium analyses were made on mitral valve allografts from 12 juvenile sheep surviving 12 to 24 weeks after mitral valve replacement. Before implantation, the allografts were treated with 0.625% glutaraldehyde (group I, n = 4) or with cold antibiotic solution (group II, n = 8). Three group I animals died 12 to 19 weeks after implantation because of dysfunction of calcified valves; the surviving animal also had extensive allograft calcification. One group II animal died of mitral regurgitation; the valves of the other seven (including five with regurgitation shown by Doppler and ventriculographic studies) were explanted at 19 to 24 weeks. Chordal rupture related to calcific deposits was found in all group I valves. Leaflet perforations (n = 4) and ruptured chordae (n = 4), each caused by connective tissue deterioration, were found in group II valves. Inflammatory reaction was absent or minimal in group I valves but moderate or severe in group II valves. Fibrous sheaths were thicker in group II than in group I valves. Calcium levels were much higher in group I than in group II valves. Calcification in group I valves was diffuse and involved collagen, elastic fibers, and connective tissue cells and matrix; in group II valves, it was localized in connective tissue cells. Thus glutaraldehyde-treated allografts failed because of extensive calcification, whereas antibiotic-preserved allografts underwent deterioration of connective tissue and infiltration by inflammatory cells. (J THORAC CARDIOVASC SURG 1995;110:224-38)
Two major types of tissue valves have been developed for use as replacement heart valves: valvular heterografts (bioprostheses) and allografts (homografts).
1,2 Because of reasons related to ease of procurement, standardization of methods of preimplantation processing, and availability from commercial sources, valvular heterografts have been used much more extensively than allografts. Although several different methods of preimplantation processing have been used for human heart valve allografts, 3-5 at the present time the majority of these valves are processed with solutions containing antibiotics and antifungal agents, often followed by cryopreservation. Most experimental and clinical studies of allografts have used either aortic or pulmonary valves, and these have been used for aortic and mitral valve replacement, 6-9 as well asfor reconstruction of the right ventricular outflow tract 10-12 and for "the pulmonary switch," 13 with excellent long-term results. Although early reports documented the successful implantation of mitral and tricuspid valve allografts in human patients, 14,15 long-term follow-up revealed late deterioration of allograft function in every case. 16 Other investigators subsequently reported implantation of mitralallografts in the mitral 17 and tricuspid 18 positions in patients. A recent study, based on the experience gained with porcine aortic valvular heterografts, which are processed by exposure to low concentrations of glutaraldehyde, has used porcine mitral valves treated with this agent before implantation in human patients. 19 Nevertheless, no data have been reported on the use of allografts or heterografts constructed of mitral valvular tissue and implanted in the atrioventricular position in experimental animal models. A considerable amount of information has been gathered on the results of implantation of variously processed porcine aortic and pericardial bioprostheses in the juvenile sheep model. 20,21 Recently, replacement of a portion of the mitral valve by an antibiotic-treated and cryopreserved mitral allograft tissue has been reported in sheep. 22 In view of these reports, it was considered of interest to use the sheep model to evaluate the results obtained by replacing the mitral valve with fresh, antibiotic-preserved or glutaraldehyde-treated mitral allografts.
METERIALS AND METHODS
Animals used.
Donor allografts were obtained from a mixed breed (Rambouillet, Dorset, Hampshire, and Suffolk) of male and female juvenile domestic sheep. Sheep of the same breed and from the same commercial source (Ovine Biotechnologies, Inc., New Hope, Pa.) served as recipients. The sheep were 25 to 35 weeks old and weighed 20 to 35 kg. All animals received care as prescribed by the Guide for the Care and Use of Laboratory Animals (NIH Publication No. 80-23, revised 1985). Procedures for preoperative, intraoperative, and postoperative management were approved by the Animal Care and Use Committee of the National Heart, Lung, and Blood Institute.
Preparation of the allografts.
Anesthesia was induced with diazepam (0.4 mg/kg) and sodium pentobarbital (5 mg/kg) and maintained with either 0.5% to 1.0% halothane or 1% to 2% isoflurane in oxygen. After administration of beef lung sodium heparin sulfate (300 units/kg), the animals were killed with potassium chloride (30 mEq) given intravenously. Valves were procured under clean but not surgically sterile conditions. The mitral valve was excised with a 2 to 3 mm rim of left atrial muscle above the anulus and 2 to 3 mm of the tips of the papillary muscles. Only valves with single discrete anteromedial and posterolateral papillary muscle heads were considered suitable for implantation. If the valvular anulus accommodated standard commercial valve sizers of 25 or 27 mm, they were considered to be of appropriate sizes for implantation. After being harvested, the valves were placed in either phosphate-buffered 0.625% glutaraldehyde (pH 7.4) or buffer and antibiotic solution containing cefoxitin 240 µg/ml, lincomycin 120 µg/ml, polymyxin 100 µg/ml, and vancomycin 50 µg/ml in 0.9% saline. They were maintained at 4º C until implantation (1 to 7 days for the antibiotic-treated valves and 2 to 16 weeks for those fixed with glutaraldehyde). Before implantation, the valves were passed through three washes of 0.9% saline containing beef lung sodium heparin 5000 units/500 ml.
Surgical implantation.
The recipients for the allograft valves were fasted overnight before the operation. Anesthesia was induced and maintained as described earlier. Mechanical ventilation was instituted via a 9 mm cuffed endotracheal tube with a volume-cycled mechanical ventilator at a rate of 12 to 14 cycles per minute, a tidal volume of 10 to 15 ml/kg, and 100% oxygen. The mechanics of ventilation were adjusted according to frequently determined arterial blood gas analyses. Electrocardiographic signals were recorded from four limb leads. Procaine penicillin, 900,000 IU administered intramuscularly, and gentamicin, 75 mg administered intravenously, were given as prophylactic antibiotics. Succinylcholine chloride was given to provide muscle relaxation before thoracotomy and was not repeated.
Systemic blood pressure was monitored via a right femoral artery catheter. A Swan-Ganz catheter (Baxter Healthcare Corp., Edwards Division, Santa Ana, Calif.) was passed into the pulmonary artery via the right femoral vein, and the pulmonary artery pressure was similarly monitored and recorded. Cardiac output was determined by computer analysis of the thermodilution curves.
Under surgically sterile conditions, a thoracotomy incision, sparing the pectoralis major muscle, was made in the left fourth intercostal space. The left internal thoracic vessels were ligated, the left lung was retracted caudally and dorsally, and the pericardium was opened and retracted.
The circuit for cardiopulmonary bypass consisted of a Cobe roller pump system (model 043600-000, Cobe Laboratories, Inc, Lakewood, Colo.) attached to a Mediflex 3/8 x 2/32; -inch extracorporeal tube (Gish Biomedical, Inc., Santa Ana, Calif.) and a Bentley BCR 2500 cardiotomy reservoir with an attached bubble oxygenator (Bentley Laboratories Division, Irvine, Calif.). The prime was prepared by delivering 500 ml of 0.9% sodium chloride and 2 units of autologous blood into the reservoir, to which 54.4 mg of calcium chloride and 6000 units of beef lung sodium heparin sulfate were then added. The bypass circuit was activated, air was evacuated by means of a three-way stopcock attached to the filter, and oxygen was delivered to the oxygenator through an arterial filter at a rate of 1 L/min. The prime was warmed and maintained between 37° and 39° C.
The left femoral artery was cannulated for arterial perfusion, and the right ventricle was cannulated through the main pulmonary artery for the venous return. Normothermic cardiopulmonary bypass was instituted at a flow rate of 100 ml/min per kilogram and adjusted to maintain a mean blood pressure of 70 to 80 mm Hg. The aorta was not clamped; hypothermia was not attempted, and the heart was neither arrested nor fibrillated. The left ventricle was vented through an apical stab incision by means of an 18F catheter. The distal pulmonary artery was occluded with an umbilical tape and attached snare.
A left atriotomy was made between the left atrial appendage and the entrance of the left hemiazygous vein into the coronary sinus. The papillary muscles were transected at their bases and the mitral valve was excised, with a 2 to 3 mm remnant of valvular tissue left at the anulus. Nonabsorbable interrupted mattress sutures of 3-0 Tevdek polyester (Deknatel Div., Pfizer, Inc., Fall River, Mass.) buttressed with Teflon felt pledgets were passed through the atrial aspect of the mitral anulus. The sutures were passed through the anulus of the allograft valve so that the orientation of the valve being implanted was identical to that of the valve that had been excised. Any redundancy of allograft annular tissue was plicated at the commissures. The papillary muscles were reimplanted at the sites of the recipient's papillary muscle bases with a single Teflon felt-reinforced horizontal mattress suture of 0 Tevdek polyester placed transmurally for each papillary muscle. The atriotomy was repaired with 6-0 Prolene suture (Ethicon, Inc., Somerville, N.J.) and the pulmonary artery snare was released. Air was evacuated from the left atrium and the left ventricle, and the animal was separated from cardiopulmonary bypass. Intravenous infusions of isoproterenol or nitroglycerin, or both, were used as necessary. A single chest tube was placed in the anterior mediastinum through a stab incision made at the sixth intercostal space. The femoral artery was repaired with 6-0 Prolene suture, and the groin and chest were closed in layers.
After operation, the animals were maintained in intensive care oxygen compartments with oxygen concentrations set at 40% for the first 24 to 36 hours. Furosemide, 150 mg twice daily, and digoxin, 0.25 mg once daily, were given for 10 to 14 days. After recovery, the animals were returned to the National Institutes of Health Animal Center, Poolesville, Maryland, and maintained by conventional ovine husbandry techniques.
Hemodynamic studies at the time of explantation.
Anesthesia was induced and maintained as before. A Swan-Ganz catheter was positioned in the main pulmonary artery via the femoral vein. A catheter was positioned in the right common femoral artery for monitoring systemic arterial pressure and arterial blood gases. These catheters were interfaced with a physiologic recorder with the use of fluid-filled pressure transducers. Arterial blood gases and pH were maintained within physiologic ranges. The thorax was opened through a bilateral transverse thoracotomy in the fifth intercostal space and extended across the sternum. Left atrial and left ventricular pressures were obtained from intracavitary catheter-tipped transducers positioned transmurally. Four consecutive cardiac cycles were analyzed for each hemodynamic determination. A hydrostatic standard was used for calibration of all pressure recordings. Left ventricular and left atrial pressures were superimposed and recorded simultaneously. The areas between the left ventricular and left atrial pressure curves during the diastolic filling periods were measured by planimetry. The mean diastolic pressure differences were obtained as the averages of four beats. Effective mitral orifice areas (MOA) were calculated with the use of a modified Gorlin and Gorlin equation: MOA = CO/(37.9 x DFT x 
Pmean x HR) where CO = cardiac output, DFT = diastolic filling time, Pmean = mean transvalvular pressure difference, and HR = heart rate. The thermodilution method with determinations of an average of five injections of 5 ml of ice-cold (0º to 4º C) saline solution was used for measuring cardiac output and, in the absence of significant mitral regurgitation, it was assumed to be identical to the transmitral flow rate per minute.
A 7F pigtail catheter was positioned in the left ventricular apex, and left ventriculography was performed to assess mitral regurgitation. Ultrasound studies (imaging, standard spectral Doppler echocardiography, and two-dimensional, color-encoded Doppler echocardiography) were also made to evaluate mitral regurgitation. The animal was killed with intracardiac potassium chloride.
Morphologic evaluation of valves.
All major organs were weighed and examined grossly and microscopically for evidence of infarction and hemosiderin deposition (hematoxylin-eosin and Pert's stain for iron). For evaluation of valve-induced hemolysis, determinations were made of hematocrit, red blood cell count, hemoglobin, reticulocyte count, serum free hemoglobin, serum lactate dehydrogenase, and bilirubin.
After removal of the heart, the valve was examined grossly in situ for evidence of dysfunction. It was further evaluated for the presence of vegetations, thrombotic material, signs of structural degradation, tears or perforations, fibrous tissue ingrowth, perivalvular leaks, and calcific deposits. Any vegetations were cultured. In situ photographs of the inflow and outflow surfaces of the valve were taken. The prosthesis was removed from the heart and fixed in McDowell's solution. 23 After fixation, gross photographs of the inflow, outflow, and lateral aspects of the valve were taken with the valve immersed in saline solution. Macrophotographs were taken at magnifications from 2 x to 8 x with a dissecting photomicroscope. A roentgenogram of the valve was taken with soft tissue techniques. For quantitative evaluation of mineralization, one half of each of the anterior and posterior leaflets was submitted for analyses of calcium and phosphorus content by atomic absorption spectrophotometry. Before the calcium and phosphorus analyses, the halves of the leaflets were subdivided into leaflet proper, chordae tendineae, and papillary muscle portions. The other half of each leaflet was processed for light and electron microscopy. 1 The following groups of valves (each, n = 2) were used as controls: (1) fresh, unprocessed mitral valves that were excised immediately after the animal was killed and were placed directly in McDowell's solution 23; (2) unimplanted glutaraldehyde-treated valves, and (3) unimplanted antibiotic-preserved valves.
RESULTS
Glutaraldehyde-treated valves were implanted in seven sheep. Two of them died of pulmonary edema and mitral regurgitation and calcification within 12 weeks after operation; another died of bacterial endocarditis at 18 weeks. These three animals were excluded from the study. Of the remaining four animals, three died of valvular dysfunction and cardiopulmonary failure between 12 and 19 weeks, and another was put to death at 19 weeks. The latter animal showed severe mitral stenosis with a mean left atrial pressure of 12 mm Hg, mean transmitral valvular gradient of 6 mm Hg, a forward cardiac output of 1.68 L/min, and a calculated valve area of only 0.5 cm2 . Doppler and cineventriculography confirmed the presence of severe mitral regurgitation caused by fracture of calcified chordae tendineae.
Antibiotic-preserved valves were implanted in 14 sheep. Six of them died in less than 12 weeks after operation and were excluded from the study. Three of these animals had bacterial endocarditis, and the other three had severe pulmonary edema, mitral regurgitation, and rupture of the chordae tendineae. The remaining eight sheep survived 12 to 24 weeks. One sheep died of mitral regurgitation at 12 weeks. The other seven were put to death after hemodynamic study. Five of the seven animals showed moderate to severe mitral regurgitation by Doppler echocardiographic and/or cineventriculographic study. For all seven animals, hemodynamic findings (expressed as mean ± standard error of the mean) were as follows: mean left atrial pressure, 19.5 ± 1.2 mm Hg; mean diastolic pressure gradient, 8.5 ± 1.1 mm Hg; Gorlin and Gorlin calculated mitral valve area, 1.1 ± 0.1 cm2 , with forward cardiac output of 2.18 ± 0.21 L/min. Increased pressure gradients and decreased calculated valve areas were thought to be influenced by the increased transmitral flow because of the regurgitation.
Cardiac gross anatomical findings
Explanted group I (glutaraldehyde-treated) valves.
On gross anatomic inspection, three valves appeared stenotic and one regurgitant. Fibrous sheaths extended for variable distances from the valve ring toward the free edges of the leaflets and from the papillary muscles onto the chordae tendineae. All valves contained calcium deposits, which formed either flat, yellowish plaques or raised nodules. The chordae tendineae had lost their elasticity and contained extensive yellowish deposits of calcium (Fig. 1, A). Several calcific nodules were also present on these chordae (Fig. 1, B). Some of the secondary chordae were ruptured in each of the four valves.
|
24 or abrasion byexcessively long ends of sutures. 25 Other chordae were thinner than normal, and rupture of the primary chordae was found in four valves (three of which also had perforations of the leaflets). The tips of these ruptured chordae were tapered and noncalcified (Fig. 1, G).
Radiographic findings
Explanted group I valves.
The extent of calcification was much greater in group I than in group II valves. The great majority of the chordae tendineae showed diffuse calcification. The leaflets also had diffuse calcification, especially at the valve anulus. The calcific deposits extended from this area to the points of insertion of the chordae into the leaflets (Fig. 2, A).
|
Quantitative measurements of calcium contents
The calcium content of explanted valves was much higher in group I than in group II animals. In fact, the values in group II valves were only slightly higher than those in normal (untreated) control valves. Calcium analyses showed values (milligrams of calcium per gram dry tissue weight ± standard error of the mean) of 249 ± 83, 88 ± 56 and 100 ± 32 in the anterior leaflet, posterior leaflet, and chordae, respectively, in group I allografts, and 1.4 ± 0.3, 3.0 ± 1.1, and 11.7 ± 6.4 in the anterior leaflet, posterior leaflet and chordae, respectively, in group II allografts. Corresponding values for control valves were 0.38 ± 0.05, 0.38 ± 0.07, and 0.26 ± 0.03.
Histologic Observations
Unimplanted, untreated valves.
In each leaflet, four distinct tissue layers were evident (Fig. 3, A) from the atrial to the ventricular aspects: the atrialis, the spongiosa, the fibrosa, and the ventricularis. The leaflet surfaces were covered by endothelial cells. The atrialis contained layered elastic fibers and a few collagen fibers. Fibroblasts and myofibroblasts were present in all four layers of the leaflet and the chordae tendineae. The spongiosa contained elastic fibers, collagen fibers, and abundant proteoglycan material. Cardiac myocytes, adipose tissue cells, and capillaries were found in the spongiosa in the basal third of the leaflet. The fibrosa consisted of thick bundles of collagen fibers. The ventricularis contained small amounts of elastic fibers and collagen fibers. This layer usually extended only along the proximal two thirds of the leaflet; in the distal third, the fibrosa was directly subjacent to the endothelial lining. The connective tissue at the free edge of the leaflet was loosely arranged, with a few bundles of collagen and abundant proteoglycan materials.
|
Unimplanted group I valves.
The structure of the unimplanted group I valves was normal. Tissues from glutaraldehyde-treated valves tended to be more eosinophilic than those of the untreated valves.
Unimplanted group II valves.
The changes observed in the unimplanted group II valves included focal loss of endothelial cells and edema of the spongiosa.
Explanted group I valves.
Fibrous sheathing was evident in the leaflets and chordae tendineae of all valves (Fig. 4, A) and involved the inflow surfaces to a greater extent than the outflow surface. The fibrous sheaths were partially lined by endothelial cells of host origin (Fig. 4, B). The main cellular components of the fibrous sheaths were fibroblasts and myofibroblasts, as well as a few macrophages and lymphocytes. The collagen fibers in the leaflets were more eosinophilic than those in the fibrous sheaths (because of the preimplantation treatment of the valves with glutaraldehyde). The fibroblasts and myofibroblasts in the leaflets (Fig. 4, C) were considered to be of donor origin. However, a few fibroblasts and macrophages of host origin appeared to have infiltrated from the fibrous sheaths into the leaflet tissue (Fig. 4, D). All leaflets showed calcification, mainly in the fibrosa at the base of the valve (Fig. 4, A). Calcific deposits were diffusely present on the collagen fibers. These deposits were homogeneous in their central portions and granular in their peripheral portions.
|
Explanted group II valves.
Fibrous sheathing was evident in the leaflets and chordae tendineae of all eight valves. At least the proximal one third of both surfaces of the leaflets was covered by fibrous sheathing (Fig. 5, A). Both the extent and the thickness of the fibrous sheathing were greater on the outflow surfaces than on the inflow surfaces. In five valves, fibrous sheathing that originated from ventricular endocardium lined the chordae tendineae and extended into the areas of the free edge and distal one third of the leaflets (Fig. 5, B). All fibrous sheaths were lined by endothelial cells; however, the remaining portions of the leaflets were completely denuded. In contrast to the findings in group I valves, the staining characteristics of the collagenous fibers were similar in the fibrous sheaths and in the leaflets. Calcific deposits were not found in any of the fibrous sheaths.
|
Fibrous sheaths lined by endothelial cells covered most of the chordae tendineae (Fig 5, D). Fibroblasts, lymphocytes, macrophages, and plasma cells were found in these fibrous sheaths and in subjacent areas of the chordae (Fig. 5, E). The collagen bundles in the fibrous cores were pale and frayed, particularly in areas of cellular infiltration (Fig. 5, F). Calcific deposits were found in the fibrous cores of the chordae in three valves.
Electron microscopic observations
Unimplanted, untreated valves.
Most of the interstitial connective tissue cells were myofibroblasts, which had prominent rough endoplasmic reticulum, Golgi apparatus, relatively large bundles of actinlike filaments, and discontinuous, poorly developed basement membranes. Some of the cytoplasmic processes of adjacent myofibroblasts were in contact, thus providing intercellular connections. Collagen fibrils formed thin bundles in the atrialis and the ventricularis. In the spongiosa these bundles were small and scattered. The fibrosa was characterized by thick bundles of collagen fibrils. Elastic fibers were composed of amorphous components (central elastin cores, which appeared lucent in sections stained with uranyl acetate and lead citrate and electron-dense in sections stained by the Kajikawa method 26) and peripherally located, longitudinally aligned microfibrils. Elastic fibers were most numerous in the atrialis but were also frequent in the ventricularis. Small elastic fibers were scattered in the spongiosa. At the free edge of the leaflet, elastic fibers were small and irregularly oriented. Proteoglycan granules were star-shaped and were most numerous in the spongiosa. These granules were also found diffusely at the free edge of the leaflet.
Myofibroblasts were the main cellular components of the chordae. The outer zones of the chordae contained elastic fibers and small amounts of proteoglycan granules and collagen fibrils. The inner cores consisted of thick bundles of collagen fibrils, among which myofibroblasts and a few small elastic fibers were scattered.
Unimplanted group I valves.
The surfaces were partially denuded of endothelial cells. Some of the myofibroblasts showed swelling of the rough endoplasmic reticulum and mitochondria and margination of nuclear chromatin; others showed damaged plasma membranes. The structure of the collagen fibrils and the elastic fibers was normal. Proteoglycan granules were markedly reduced in number. Some of the remaining granules were smaller than normal and lacked elongated projections. A mild to moderate degree of interstitial edema was present.
Unimplanted group II valves.
Many endothelial cells were lost and remaining cells showed pyknotic nuclei and lytic changes in their cytoplasm. Myofibroblasts showed similar changes. Collagen bundles were separated and loosely arranged, but the periodicity of the collagen fibrils was still visible. Sparse proteoglycan granules were present in the interstitium. The microfibrils of the elastic fibers sometimes had lost their parallel arrangement. Interstitial edema was diffusely present. Cell degeneration was more severe in this group than in group I valves.
Explanted group I valves.
Endothelial cells of donor origin were lost from the valvular surfaces and the endothelial cells of host origin covered the fibrous sheaths (Fig. 6, A). The fibroblasts and myofibroblasts in the fibrous sheaths showed parallel arrangements, and cytoplasmic processes of adjacent cells were in contact. The amorphous components of elastic fibers in the fibrous sheaths often had a finely granular appearance. Masses of microfibrils with parallel arrangements also were found among the fibroblasts and myofibroblasts in the fibrous sheaths. These findings were considered to be indicative of the formation of new elastic fibers. Proteoglycan granules also were found in the fibrous sheaths, but not in donor tissue.
|
27Fig. 6, B and C) and others showed a spiraling appearance in longitudinal sections. Collagen fibrils in the donor tissue showed normal periodicity. Amorphous and very finely granular and fibrillar materials were present between the collagen fibrils in these valves and often were aggregated into masses (Fig. 6, D). Some of the mesenchymal cells were relatively well preserved. Others were damaged. Calcification involved both the mesenchymal cells and the extracellular components of connective tissue. Intramitochondrial, intralysosomal, and intranuclear calcific deposits, consisting of needle-shaped, electron-dense crystals, were frequently found in these mesenchymal cells. Necrotic cell fragments also were calcified. Two patterns of calcification were observed in the collagen fibrils (Fig. 6, F and G): calcification of entire fibrils and calcification localized to the surfaces of the fibrils. The latter deposits were continuous with deposits involving adjacent areas of extracellular matrix. The amorphous components of elastic fibers also had calcific deposits composed of needle-shaped crystals (Fig. 6, F). Microfibrils were not clearly identified in these elastic fibers. In peripheral areas of large calcific masses, very electron-dense, round particles were present in close association with the microfibrillar components of the elastic fibers (Fig. 6, G).
Explanted group II valves.
Some of the denuded areas of the surfaces were covered by fibrin deposits, whereas in others the collagen fibrils were exposed. A few of the collagen fibrils in the fibrous sheaths showed a spiraling appearance (Fig. 7, A). Elastic fibers in the fibrous sheaths were similar to those in group I valves. In donor tissues, mesenchymal cells showed more severe changes than in group I valves. All cells showed disruption of the plasma membranes and the organelles (Fig. 7, B). Calcific deposits were rarely found in these cells. Newly formed elastic fibers, characterized by finely granular appearance with numerous microfibrils, were present near mesenchymal cells of host origin that had infiltrated into leaflet tissue (Fig. 7, C). Proteoglycan material was found only around these cells. The collagen fibrils throughout the donor tissue showed a loose arrangement and were more widely separated from each other than was the case in group I valves (Fig. 7, D). "Collagen flowers" were not found. Numerous microfibrils, which were morphologically similar to those associated with elastin, were scattered through all layers of donor tissue (Fig. 7, E and F). Similar accumulations of such fibrils were not found in group I valves.
|
The present study demonstrates that the results obtained with the sheep mitral valve allograft model depend on the type of preimplantation processing of the allograft. These results clearly show that the glutaraldehyde-treated allografts undergo rapid, severe calcification resulting in marked stenosis and functional failure. These changes also can result in mitral regurgitation from fracture of calcified chordae tendineae. In contrast, the allografts preserved by treatment with antibiotics undergo progressive disruption of the connective tissue in the chordae tendineae and leaflets and infiltration by inflammatory cells. These alterations lead to progressive mitral regurgitation. Thus the findings of the present study are in agreement with those of Maxwell, Gavin, and Barratt-Boyes, 8,9 who found that glutaraldehyde-treated porcine aortic valve xenografts implanted in the mitral position in patients calcified very frequently but detached from the commissures very infrequently, whereas antibiotic-preserved aortic valve allografts implanted in the mitral position showed much less calcification but had a much higher incidence of cuspal damage and detachment.
Glutaraldehyde-treated valves.
This study provides the first evaluation of the postimplantation changes that develop in glutaraldehyde-treated mitral valves implanted orthotopically. The calcification observed in the present study is histologically and ultrastructurally similar to that previously reported with glutaraldehyde-fixed porcine aortic and bovine parietal pericardial bioprostheses implanted in the sheep model. 20,21 As described previously, this is a model for accelerated calcification of biologic tissue valves. Gross anatomic examination disclosed that in all four group I valves the calcific deposits involved the leaflets and the chordae tendineae, thus resulting in loss of valvular mobility. It became obvious that the mechanical forces exerted under these conditions on calcified chordae tendineae were sufficient to cause their fracture. The microscopic patterns of calcification observed in these valves were similar to those described in porcine aortic valves and bovine pericardial valves implanted after glutaraldehyde fixation in the sheep model 20,21 and inpatients. 28,29 The results of the present study emphasize the importance of calcification of the chordae tendineae in causing valvular stenosis and regurgitation in mitral valvular allografts. However, other changes in connective tissue components were much less severe in group I than in group II valves.
Three types of ultrastructural alterations of collagen were observed in the present study: (1) "collagen flowers," (2) spiraling collagen fibrils, and (3) wide separation and loss of collagen fibrils. The "collagen flowers" were observed only in group I valves and were similar to those described in several inheritable disorders of connective tissue. 27 Such fibrils have been thought to result from either abnormal synthesis or partial degradation of collagen and are not specific for any disease entity. In the valves in the present study, such fibrils probably occurred as consequences of partial degradation of glutaraldehyde-crosslinked collagen fibrils. Nevertheless, we have not observed "collagen flowers" in explanted glutaraldehyde-fixed porcine aortic valvular bioprostheses or bovine pericardial bioprostheses. The spiraling substructure also has been interpreted as indicative of partial unraveling of the collagen fibrils. In contrast, the loss of collagen fibrils observed in group II valves is clearly indicative of complete disruption.
Antibiotic-preserved valves.
Chordal rupture and cuspal perforations were the main anatomic changes leading to failure of antibiotic-preserved mitral valve allografts. Damage to the connective tissue elements, particularly collagen, led to these alterations. Fibrous sheathing was present on the chordae tendineae and on the leaflets; however, the extent to which this connective tissue reinforcement developed was not adequate to prevent chordal rupture and leaflet perforation. It seems likely that the connective tissue breakdown in these valves resulted from several factors including autolytic changes before implantation, inflammatory reaction, and wear and tear associated with the mechanical forces related to opening and closure of the mitral allograft.
In explanted group II valves, we found extensive accumulations of connective tissue microfibrils that were morphologically similar to the microfibrils associated with elastic fibers. However, such accumulations of microfibrils were not observed in unimplanted valves, and their mechanism of formation remains unknown.
We observed a cell-mediated immunologic reaction, characterized by infiltration of lymphocytes, plasma cells, and macrophages, even in areas covered by fibrous sheaths. This reaction was associated with disruption of the connective tissue components. Macrophage-derived collagenases probably played an important role in this disruption of the connective tissue. A similar reaction has been observed in explanted antibiotic-preserved human aortic valve allografts. 30 The fact that such a reaction was minimal in glutaraldehyde-treated allografts is in agreement with the concept that glutaraldehyde fixation produces a marked decrease in the antigenicity of transplanted valvular tissues. 31 This treatment also increases the crosslinking and the stability of collagen, 29 accounting for the paucity of collagen disruption in the explanted group I valves.
The only connective tissue cells and endothelial cells that appeared to be viable in explanted group II valves were localized in the fibrous sheaths and their immediate vicinity. The connective tissue cells of donor origin present in these valves were disrupted. Similar observations have been made in antibiotic-preserved human aortic valve allografts. 30 Thus such cells do not maintain a functional role in the turnover and renewal of donor valvular connective tissue after implantation of the allograft. It remains to be determined whether or not cryopreservation of mitral valve allografts will result in improved viability of donor connective tissue cells and endothelial cells for long periods after implantation. However, no growth of cells of donor origin was found in tissue cultures of cryopreserved aortic valve allografts explanted 8 to 15 months after implantation in the descending aorta of sheep. 32 In contrast, growth of cells of donor origin was found in 43% of similarly implanted pulmonic valves. The reason for this difference is not known. In human beings, aortic valve allografts have lasted for very long periods. 3-13 Structural changes develop in such allografts, including loss of nuclei of connective tissue cells and a homogeneous appearance of the connective tissue. Other studies have demonstrated severe loss of endothelial cells in cryopreserved human 33 and rat 34 aortic valve allografts. These changes are in many ways comparable with those observed in the present study. These results suggest that under certain circumstances the viability of connective tissue cells and endothelial cells of donor origin is not necessary for good long-term function of aortic valve allografts. However, we consider that loss of connective tissue cell viability, leading to inability to synthesize new collagen fibrils to replace those that are degraded during normal tissue turnover, is the ultimate cause of chordal and cuspal disruption in the mitral valve allografts in the present study.
The antibiotic-preserved mitral valve allografts in the sheep model failed more quickly and more severely than would have been expected on the basis of previous experience with aortic and pulmonary valve allografts implanted in human beings 3-13 and in sheep. 32 Clinical and experimental studies of mitral valve allografts have been reported previously. Van Vliet and associates 35 found that fresh mitral valve allografts implanted in dogs maintained satisfactory function for up to 1 year and reported that these valves were gradually remodeled by fibrous connective tissue of host origin without apparent functional deterioration. Similar results were reported by Bernhard and colleagues. 36 However, other studies have shown that mitral valve allografts implanted in human beings 10,16,37 and in dogs 38,39 either had a low rate of survival or exhibited late deterioration with a high incidence of rupture of the chordae tendineae. We believe that the arrangement of the connective tissue components in aortic valve leaflets makes them more resistant to damage by mechanical forces during valvular opening and closing. Specifically, a large majority of the collagen fibrils in aortic valves are arranged in fibrous cords that are oriented parallel to the edge of the cusp and attach at the commissures. This arrangement is clearly different from that in mitral leaflets and chordae tendineae. The role of the chordae tendineae in mitral valve closure is without functional parallel in aortic valve closure. This unique role places the mitral chordae at high risk of functional deterioration because of the substantial mechanical forces to which these structures are subjected during left ventricular systole. The lower pressures in the right side of the heart may account for the long-term durability of tricuspid valve allografts implanted in the tricuspid position in dogs. 40
In conclusion, the present study reports the results obtained with two types of orthotopic mitral valve allografts in sheep. Glutaraldehyde-fixed allografts failed because of extensive calcification involving the chordae and the leaflets, and antibiotic-preserved allografts failed because of extensive alterations in collagen. The development of fibrous sheathing of recipient origin was not sufficient to provide adequate structural reinforcement of antibiotic-treated allografts undergoing connective tissue deterioration.
Acknowledgments
We acknowledge the assistance of the veterinary professional and technical staff of the Laboratory of Animal Medicine and Surgery, National Heart, Lung, and Blood Institute.
Footnotes
From the Pathology Brancha and the Laboratory of Animal Medicine and Surgery,b National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Md. 
References
This article has been cited by other articles:
|
|
 |
|
  F. Bauer, M. Jones, T. Shiota, M. S. Firstenberg, J. X. Qin, H. Tsujino, Y. J. Kim, M. Sitges, L. A. Cardon, A. D. Zetts, et al. Left ventricular outflow tract mean systolic acceleration as a surrogatefor the slope of the left ventricular end-systolic pressure-volume relationship J. Am. Coll. Cardiol., October 2, 2002; 40(7): 1320 - 1327. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y J Kim, M Jones, T Shiota, H Tsujino, J X Qin, F Bauer, M Sitges, J Kwan, L A Cardon, A D Zetts, et al. Effect of load alterations on the effective regurgitant orifice area in chronic aortic regurgitation Heart, October 1, 2002; 88(4): 397 - 400. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Mori, R. A. Rusk, M. Jones, X.-N. Li, T. Irvine, A. D. Zetts, and D. J. Sahn A new dynamic three-dimensional digital color doppler method for quantification of pulmonary regurgitation: validation study in an animal model J. Am. Coll. Cardiol., September 18, 2002; 40(6): 1179 - 1185. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Sitges, M. Jones, T. Shiota, D. L. Prior, J. X. Qin, H. Tsujino, F. Bauer, Y. J. Kim, D. Deserranno, N. L. Greenberg, et al. Interaliasing distance of the flow convergence surface for determining mitral regurgitant volume: a validation study in a chronic animal model J. Am. Coll. Cardiol., October 1, 2001; 38(4): 1195 - 1202. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. S. Kumar, S. K. Choudhary, A. Mathur, A. Saxena, R. Roy, and P. Chopra Homograft mitral valve replacement: Five years' results J. Thorac. Cardiovasc. Surg., September 1, 2000; 120(3): 450 - 458. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Shiota, M. Jones, S. Aida, M. Chikada, H. Tsujino, T. El-Kadi, and D. J. Sahn Validation of the accuracy of both right and left ventricular outflow volume determinations and semiautomated calculation of shunt volumes through atrial septal defects by digital color Doppler flow mapping in a chronic animal model J. Am. Coll. Cardiol., August 1, 1999; 34(2): 587 - 593. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Mori, T. Shiota, M. Jones, S. Wanitkun, T. Irvine, X. Li, A. Delabays, N. G. Pandian, and D. J. Sahn Three-Dimensional Reconstruction of the Color Doppler–Imaged Vena Contracta for Quantifying Aortic Regurgitation : Studies in a Chronic Animal Model Circulation, March 30, 1999; 99(12): 1611 - 1617. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Shiota, M. Jones, M. Chikada, C. E. Fleishman, J. B. Castellucci, B. Cotter, A. N. DeMaria, O. T. von Ramm, J. Kisslo, T. Ryan, et al. Real-time Three-dimensional Echocardiography for Determining Right Ventricular Stroke Volume in an Animal Model of Chronic Right Ventricular Volume Overload Circulation, May 19, 1998; 97(19): 1897 - 1900. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Bernal, J. M. Rabasa, J. C. Cagigas, F. Val, and J. M. Revuelta Behavior of Mitral Allografts in the Tricuspid Position in the Growing Sheep Model Ann. Thorac. Surg., May 1, 1998; 65(5): 1326 - 1330. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Shiota, M. Jones, A. Delabays, X. Li, I. Yamada, M. Ishii, P. Acar, S. Holcomb, N. G. Pandian, and D. J. Sahn Direct Measurement of Three-dimensionally Reconstructed Flow Convergence Surface Area and Regurgitant Flow in Aortic Regurgitation : In Vitro and Chronic Animal Model Studies Circulation, November 18, 1997; 96(10): 3687 - 3695. [Abstract] [Full Text]
|
