Onset and progression of calcification in porcine aortic bioprosthetic valves implanted as orthotopic mitral valve replacements in juvenile sheep

Transcription

1 Onset and progression of calcification in porcine aortic bioprosthetic valves implanted as orthotopic mitral valve replacements in juvenile sheep The purpose of this study was to characterize the onset and progression of mineralization in porcine bioprosthetic valves implanted in sheep and to test the hypothesis that such valves simulate calcification that is observed clinically and in other experimental models. Hancock I porcine aortic bioprosthetic valves (Medtronic Heart Valve Division, Irvine, Calif.) were implanted as orthotopic mitral valve replacements in juvenile sheep, retrieved after 1 to 124 days, and analyzed as follows: gross inspection, radiography, light, transmission, and surface scanning electron microscopy, and calcium analysis by absorption spectroscopy. Mineralization increased with increasing time after implantation in both valve cusps and adjacent aortic wall. Mean cuspal calcification was 80 ~g/mg in valves removed after 3 to 4 months. Nevertheless, considerable variability among valves was apparent in the level of calcification noted at specific time intervals. Virtually all aspects of the morphologic characteristics were identical to those previously noted for clinical explants and experimental specimens, both subcutaneous and circulatory. In particular, ultrastructural examination revealed that the earliest calcific deposits were associated with devitalized cuspal connective tissue cells and their fragments. Collagen calcification was sparse. Both surface scanning and transmission electron microscopy indicated a lack of endothelial or blood-derived cells on the valves at all sampling times. We conclude that porcine bioprosthetic valves implanted as mitral valves in sheep provide a useful calcification model, simulating morphologic and pathobiologic events that occur clinically and in noncirculatory models. However, sufficient specimen replicates must be done to overcome variability in calcification among valves and sampling sites. (J THORAe CARDIOVASC SURG 1994;108:880-7) Frederick J. Schoen, MD, PhD,a Danielle Hirsch, PhD,b Richard W. Bianco,c and Robert 1. Levy, MD,b Boston, Mass, Ann Arbor, Mich., and Minneapolis, Minn. Structural dysfunction caused by primary tissue failure, especially calcification, remains the major cause of failure of porcine bioprosthetic valves and contributes to From the Department of Pathology, Brigham and Women's Hospital and Harvard Medical School, Boston, Mass."; the Department of Pediatrics and Communicable Diseases, University of Michigan Medical Center, Ann Arbor, Mich. b ; and the Department of Surgery (Division of ), University of Minnesota, Minneapolis, Minn.c Supported in part by grant HL from the National Institutes of Health. Received for publication Feb. 10, Accepted for publication May 31, Address for reprints: Frederick J. Schoen, MD, PhD, Department of Pathology, Brigham and Women's Hospital, 75 Francis St., Boston, MA Copyright ri) 1994 by Mosby-Year Book, Inc /94 $ /1/58198 the failure of many pericardial valves. 1-5 A well-characterized large animal model for bioprosthetic heart valve replacement that reproduces clinically observed primary tissue failure, especially cuspal calcification, would be extremely useful in the investigation of mechanisms of valve deterioration and in potentially developing and testing therapeutic strategies. Large animal implants of tricuspid and mitral orthotopic valve replacements or heterotopic conduit or assist device-mounted implants in sheep or calves can be an important source of pathophysiologic data to elucidate determinants and mechanisms of calcification in the circulatory environment. 6-9 Moreover, such studies presently serve as a critical component of the preclinical testing phase of a new tissue valve configuration or tissue pretreatment, especially for evaluating the safety and efficacy of antimineralization therapies Despite the wide experimental use of orthotopic mitral valve replacements in sheep, the initiating mechanisms 880

2 The Journal of Thoracic and Volume 108. Number 5 Schoen et al and evolution of calcification have not been characterized in detail and the clinical relevance of this model has not previously been confirmed. Therefore, this study characterized the onset and progression of calcification of porcine aortic bioprosthetic valves implanted as orthotopic mitral valve replacements in sheep and tested the hypothesis that porcine valves implanted as orthotopic mitral valve replacements in young sheep progressively develop intrinsic cuspal and aortic wall calcification which simulates in extent, mechanisms, and morphologic features that observed clinically and in other models used for investigating valve mineralization. Materials and methods Valves and animal model. Twenty purchased clinical-grade Hancock I (Standard) porcine bioprosthetic valves (size 25 mm, Medtronic Heart Valve Division, Irvine, Calif.) were implanted into the mitral position of Dorset or Columbian cross-bred sheep, 3 to 4 months of age, weighing 28 to 38 kg, by means of techniques previously described in detail. l? In brief, valves were implanted into the mitral position under halothane/isoflurane inhalation anesthesia via a standard left thoracotomy through the fourth intercostal space with the aid of cardiopulmonary bypass. Preoperatively, the aniinals were vaccinated against Clostridium C and D and received ticarcillin disodium and gentamicin sulfate. Approximately 2 weeks after the operation, the animals were transferred to a farm, where they were monitored. Subcutaneous heparin sulfate (1000 units) was given twice a day for 2 days; piroxicam (10 mg by mouth) was given each day beginning on postoperative day 1, but no anticoagulation was administered long term. At monthly intervals, blood samples were obtained for complete chemistries, complete blood count with differential, reticulocyte count, and coagulation profile. The sheep used for this study were fed diets containing the recommended daily calcium intake for young developing sheep. Full blood screens, including aerobic/anaerobic blood cultures, were obtained. The sheep were killed electively at monthly intervals for 1 to 4 months. Whenever possible, immediately before the animals were killed, with the animals anesthetized, intubated, and their lungs ventilated, right- and left-sided catheterization studies were performed via the jugular vein and carotid artery, respectively, to do angiographic imaging and estimate left atrial pressure (via pulmonary wedge) and left ventricular pressures (to obtain mitral valvular gradients). Animals were then heparinized (3000 units) and put to death under full anesthesia with an overdose of potassium chloride. This was followed by complete necropsy. All animals received humane care in compliance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the "Guide for the Care and Use of Laboratory Animals" prepared by the Institute of Laboratory Animal Resources and published by the National Institutes of Health (NIH Publication No , revised 1985). Explant analysis. Valves were retrieved at times up to 124 days. Early explants were obtained from seven animals that either died unexpectedly (n = 6) or were put to death because of early postoperative difficulties (n = 1) after 1 to 40 days. In all such cases, necropsy was done; the valves were taken at Fig. 1. Radiographs and gross photograph of explanted valves. A, Radiograph at 23 days demonstrating focal calcification of the aortic wall. Radiograph (B)and photograph (C) of valve explanted at 106 days showing prominent calcification of both cusps (arrow in C) and aortic wall.

3 8 8 2 Schoen et al. The Journal of Thoracic and November 1994 R ~ A Implant duration (months) "",so j ~ ci e at a ; u li B "'" 'so I. 0 '" e ~ ; u c 0 o~ WAll I 0 CUSP Implant duration (day.) 15. c 15. Implant durauon (days) Fig. 2. Calcium concentration versus implantation time, separately measured in valve cusps and aortic walls. A, Data for valve cohorts segregated among implant durations. Mean calcium concentrations of cusp and wall are similar at a similar duration offunction. Calcification to levels above those in unimplanted valves was noted even in valves functioning less than 1 month. B, Data for cusp plus wall in individual valves, derived by summation of cuspal and wall calcium concentrations. Although the trend toward increased mineralization with increasing time after implantation remains evident, there is considerable variability among valves. C, Data plotted as individual cusp/wall data points. Both cuspal and aortic wall calcification increase with time. necropsy and placed in 10% neutral buffered formalin. The organs were examined for emboli. In no case was an unexpected death considered to be valve related. For each of the 13 valves taken electively, the left coronary cusp was immediately dissected from the valve, with portions frozen for mineral analysis and potential histochemical/immunohistochemical analysis that might be suggested by the morphologic results (e.g., alkaline phosphatase enzyme activity or identification of adherent blood celis); however, none of the latter was believed to be indicated. An additional portion was placed in Karnovsky's fixative for transmission and scanning electron microscopy. The remainder of the prosthesis was removed with a thin rim of annular tissue and immersed in Karnovsky's fixative. After fixation, all retrieved valves were photographed, radiographed, and analyzed grossly and by light microscopy, and calcium was quantitated. Yalve dissection involved careful removal of the entire tissue complex from the stent, followed by amputation of the cusps from the adjacent aortic wall at their junctions. After dissection, a sample of each remaining fixed cusp was subjected to calcium and phosphorus analysis. Quantitative determinations for calcium were done individually on the separate cusps and the aortic wall portion of each cusp by atomic absorption spectroscopy, using previously described methods. 9, 18, 19 A section of each cusp and the adjacent aortic wall was selected for light microscopy. Routine processing for light microscopy included dehydration through graded ethanois and embedding of specimens in glycolmethacrylate (GMA, Polysciences Inc., Warrington, Pa.). Sections 3 Jlm thick were stained with hematoxylin and eosin for overall morphologic features and the von Kossa stain for calcium phosphates. Cuspal tissue processed for transmission electron microscopy was fixed in 2.5% glutaraldehyde and 2.0% paraformaldehyde in cacodylate buffer at ph 7.4, postfixed in 2.0% osmium tetroxide, dehydrated in ethanol in propylene oxide, and embedded in Poly/Bed 812 (Polysciences). Sections were cut at 60 nm, stained with lead citrate and uranyl acetate, and examined with a JEOL-lOOCX transmission electron microscope (leol, Cranford, N.J.), at an accelerating voltage of 80 KY. Tissue for scanning electron microscopy was fixed as described earlier, dehydrated in graded ethanols, critical point dried, coated with gold-palladium, and examined in an lsi -DS-130 scanning electron microscope (lsi, Santa Clara, Calif.) at 20 KY.

4 The Journal of Thoracic and Volume 108, Number 5 Schoen et al Results At the time of terminal elective study, technically satisfactory hemodynamic studies were performed in nine animals. Four animals put to death at 35 to 63 days had valvular gradients of 5 mm Hg or less and five animals kiued at 97 to 124 days had mitral valve gradients of 6 to 16 mm Hg. Mild-to-moderate mitral stenosis (left atrialleft ventricular gradient> 10 mm Hg) was present in only two of the six animals surviving more than 63 days; no animal had mitral regurgitation or severe mitral stenosis. All valves were grossly functional at the time of explantation. No instances of bioprosthesis-associated valve infection were found, and no evidence of prominent tissue pannus overgrowth, thrombotic deposits or vegetations, cuspal immobilization or retraction, commissural fusion, cuspal perforations, or intracuspal hematomas was observed on any valve. Radiographs and a typical gross photograph of removed valves are illustrated in Fig. 1. The earliest calcific deposits were noted in the cuspal commissures and basal attachment sites. Bioprostheses removed after long-term function (> 3 months) typically had focal nodular intrinsic cuspal calcification and prominent aortic wall mineralization. In all, 13 of the 20 porcine valvular bioprostheses had radiographic evidence of calcific deposits within cusps and/or the aortic wall, including all valves functioning more than 23 days. The relationship between calcium content measured by atomic absorption spectroscopy and the duration of function is shown quantitatively for the cusps and wall portions measured separately in Fig. 2. A consistent trend was noted toward increased mineralization with increasing time after implantation (Fig. 2, A). The maximal level of cuspal calcification was 109 to 118 /-lg/mg at 3 to 4 months (mean 80 /-lg/mg). However, considerable variability was observed among valves (Fig. 2, B) and among individual anatomic components of valves (Fig. 2, C). Moreover, no consistent relationship existed between cuspal and aortic wall calcification within the same valve. Light microscopy demonstrated that the overall structure of the valve cusps was well preserved in areas not calcified, without apparent endothelium, inflammatory cells, or thrombus. Calcification of the aortic valve cusps was noted as early as 31 days and calcification of the aortic wall was noted as early as 23 days. The morphologic features that characterized valve calcification are summarized in Fig. 3. Calcification initially appeared as smau, clearly intrinsic nodules that were scattered throughout the cusps and aortic wall portions, noted earliest in the right coronary muscle shelf (Fig. 3, A). Otherwise, the earliest cuspal calcific deposits were noted in the fibrosa (Fig. 3, B), but later deposits expanded the spongiosa. The earliest deposits and the edges of expanded nodules suggested an initial calcification mechanism that primarily involved cell remnants (Fig. 3, C). No extrinsic mineralization was noted. Mineral deposits were also noted diffusely throughout the aortic media, with initial deposits forming between elastic lamellae, apparently associated with devitalized smooth muscle ceus of the aortic wall media (Fig. 3, D). Later calcific deposits appeared to also decorate elastic lamellae. Subsequently, calcific deposits tended to form diffusely in the aorta, eventuating in confluent nodules (Fig. 3, E), The earliest specimen suitable for electron microscopy was derived from an animal killed electively 31 days after the operation. Initial mineral deposits were clearly localized predominantly to devitalized connective tissue cells of the underlying cuspal matrix and their fragments (Fig. 4). The earliest deposits were noted in the nuclei, in the cytoplasm appearing to be at residual organelles, or associated with the plasma membrane. Deposits associated with collagen and elastin were difficult to locate, even in longer term implants. Surface structure, revealed by surface scanning and transmission electron microscopy, was similar throughou t all valves examined, irrespective of time. The predominant surface was a relatively rough fibrillar structure (presumably derived from both subendothelial valvular connective tissue and minimal adherent plasma proteins or thrombus) (Fig. 5). No endothelial cells were noted. Blood cells and platelets were only rarely identified. Discussion The most important results of this study were as follows: (1) Mitral bioprosthetic valves implanted in sheep calcified progressively in a m anner which simulated that noted clinically and also subcutaneously in rodents; (2) calcific deposits were found in all porcine valvular bioprostheses implanted in sheep for periods longer than 23 days and were later prominent in both cusps and aortic walls; (3) tissue overgrowth was not excessive; (4) although calcification was progressive with increasing implant duration, considerable variability in calcification was noted among cusps and aortic walls at a given interval; (5) as in other models, cuspal calcification was intrinsic initially and largely localized to devitalized cells; and (6) cellular blood-surface interaction was minimal at all time periods examined. In contrast to the relatively common finding of cuspal tears in bioprostheses failing after human implantation, no perforations or tears were found in the valves implanted in sheep. Pathobiology and progression of calcification. All aspects of the structure of the calcific deposits in the present study were similar to those widely reported in both

5 884 Schoen et al. The Journal of Thoracic and November 1994 Fig. 3. Light microscopic features of mineralization. A, Early calcification, noted most prominently in the right coronary muscle shelf. B, Intrinsic calcific nodule in cuspal fibrosa. C, High-power photomicrograph of expanding edge of cuspal calcific deposit that demonstrates cell-oriented initial deposits (arrow). D, Early calcific deposits on aortic wall, apparently localized to aortic wall cells. E, Aortic wall media diffusely and heavily involved by calcification. All stained with von Kossa reagent; magnifications: A, X 90; B, X 70; C, X 140; D and E, X 80. clinical and subcutaneous implant specimens. The most frequent locations for calcific deposits in these valves were the aortic wails, the muscle shelf cif the right coronary cusp, and elsewhere in the cusps. Moreover, in virtually all previously reported clinical, experimental circulatory, and subcutaneous implant models, calcification has been shown to heavily involve the residual connective tissue cells and their fragments and, to a variable and usually lesser extent, collagen and elastin.9, We have previously hypothesized that the critical mechanism in calcification of bioprosthetic tissue involves the inability of the devitalized, mechanically fragmented cells (presumably having dysfunctional mechanisms for calcium exclusion) to maintain a low intracellular content of free calcium in the presence of high extracellular calcium. The phosphorus necessary to form the earliest calcium phosphate mineral nuclei appears to derive from the organic phosphorus present in

6 The Journal of Thoracic and Volume 108, Number 5 Schoen et al Fig. 4. Ultrastructure of connective tissue cell-associated calcific deposits (arrows) in connective cell nucleus (A) and cytoplasm (B) of sheep implants. Ultrathin sections stained with uranyl acetate and lead citrate. Bar = 3,urn. Fig. 5. Morphologic characteristics of cuspal surface. A, Scanning electron microscopic appearance of surface of a valve explanted after 4 months, indicating the lack of endothelial cell coating. The predominant surface is a rough fibrillar structure derived from the subendothelial valvular connective tissue as well as adherent plasma proteins and perhaps small fragments of thrombus. Only rare blood-derived cells were identified. This appearance was identical at all examination times. B, Transmission electron microscopic appearance of surface of valve cusp, demonstrating a lack of cellular coating and the presence of superficially imbibed plasma proteins immediately beneath the surface. Ultrathin section stained with uranyl acetate and lead citrate. Bar = 3,urn. the cellular membranes themselves. Moreover, although the pathophysiology of both cusp and aortic wall calcification predominantly involves mineral deposits in the devitalized underlying cells and their fragments, some evidence exists that calcification of the aortic wall also involves elastin in both clinical and experimental specimens. 23,24 Thus the results of this study emphasize the pathophysiologic continuum of calcification mechanisms in the spectrum of in vivo small animal noncirculatory models through the clinical situation. Taken with the data on aortic wall calcification, they also emphasize that, although mechanical action clearly accelerates mineralization,25 the essential calcification mechanism in bioprosthetic valves does not require mechanical activity in the form of valve flexing. Cellular blood-surface interaction noted by scanning and transmission electron microscopy was minimal in the present study. The valves were not endothelialized at any time period noted, and only rare leukocytes and/or platelets were noted on the surfaces. This suggests that cellular and/or proteinaceous interactions of blood with the surfaces of porcine valves do not strongly affect the pathophysiology of calcification and that the blood merely provides a chemical environment permissive to calcification. Moreover, aortic wall calcification is expected to be

7 8 8 6 Schoen et al. The Journal of Thoracic and November 1994 inconsequential in stent-mounted porcine valves in which the aortic wall bulk is small, largely restrained against the stent, and hidden under a cloth covering. However, with the recent consideration of stentless porcine valves,26,27 calcification of the aortic wall portion could become a limiting problem, as it has been in many valvular allografts that also contain an untethered portion of the aortic (or pulmonary arterial) wall. 28 Calcium measurements. The quantitative cuspal calcification results of this study are consistent with those of previous reports. The mean cuspal calcification concentration was 80 /-lg/mg and the maximum was 109 to 118 /-lg/mg at 3 to 4 months. In Carpentier-Edwards porcine valves implanted in five calves as mitral replacements for 69 to 142 days, the mean cuspal calcium concentration was 86 /-lg/mg. 9 Hancock porcine valves implanted in 25 to 32 kg juvenile sheep as mitral replacements for 5 months as controls for comparison with anticalcification treatments had a mean cuspal calcium concentration of 129 /-lg/mg.1 5 In another antimineralization study, cusps of standard Carpentier-Edwards and Hancock porcine valves implanted as.mitral valve replacements in sheep had calcium contents of 65 to 100 /-lg/mg after approximately 5 months. lo, II The extent of calcification that occurs in 3 to 4 months in the sheep is slightly less than that previously measured in failed clinical explants (mean 113 /-lg/mg for valves implanted a mean of7 years); valves with tears, the most frequent failure mode, had mean 105 /-lg/mg. 29 Therefore, these porcine valve implants in sheep attained a level of calcification approximating that which exists shortly before the onset of the secondary tears that induce clinical failures. These results should be independent of porcine valve type, because Hancock and Carpentier-Edwards valves have similar durability and modes of failure. 3D Moreover, it is likely that similar results would be obtained for pericardial valves implanted in sheep, because pericardium and porcine valves have a similar capacity and mechanism for mineralization. 19 Nevertheless, we documented marked variability in the quantitated calcification among valves and in cuspal and aortic wall calcification at each time period. Marked variability in calcification among valves, and indeed among portions of the same valve, is also seen in clinical specimens 28 and, to a lesser extent, in specimens derived from the rat subcutaneous model. 9, 18 The source of this variation is likely multifactorial, including host, implant, and technicaldeterminants. 31 Irrespective of cause, the wide variability among valves necessitates that an ample number of replicative experiments be done at each time period in preclinical studies so that an adequate representation is obtained. Implications. The major clinical importance of this study lies in its detailed characterization of an animal model for an important human condition, bioprosthetic valve calcification. On the one hand, these results suggest that subcutaneous models can be adequately used to rapidly and economically screen antimineralization therapies and other new technologies for efficacy and safety. However, another phase of preclinical testing necessitates that the effects of blood and blood-surface interaction, mechanical activity, and an actual valve configuration be evaluated. We have confirmed that orthotopic mitral valve implants in sheep satisfy this need. Innovations such as systemic or local drug administration, alternative fixation schemes, tissue pretreatments, and new bioprosthetic valve designs and biomaterials, including pericardial bioprostheses and trileaf'let polymer valves, must be tested in a large animal model before the clinical use of new products, as required by the Food and Drug Administration. However, the present study emphasizes the importance of cautious interpretation of bulk calcium data on new valve designs and therapeutic strategies. Conclusions We conclude that (1) porcine bioprosthetic valves implanted as mitral valves in sheep provide a useful calcification model, simulating events that occur clinically; (2) the use ofthis model as a preclinical evaluation of new valve treatments, designs, and biomaterials requires sufficient replicates to overcome specimen variability in calcification; and (3) the morphologic features observed in porcine valves suggest that calcification pathophysiology in the circulatory environment is identical to that which occurs outside the circulation. We are grateful to Sally Brinkman, Rose Marie Clack, Joelle Drader, John Mrachek, Sara Murray, Elena Rabkin, Helen Shing, and Joseph Trachy for technical assistance and to Claudia Davis for typing this manuscript. REFERENCES 1. Milano A, Bortolotti U, Talenti E, et al. Calcific degeneration as the main cause of porcine bioprosthetic valve failure. Am J Cardiol 1984;53: Schoen FJ, Hobson CEo Anatomic analysis of removed prosthetic heart valves: cause of failure of 33 mechanical valves and 58 bioprostheses, Hum Pathol 1985; 16: Reul GJ, Cooley DA, Duncan JM, et al. Valve failure with the Ionescu-Shiley bovine pericardial bioprosthesis: analysis of 2680 patients. J Vase Surg 1985;2: Schoen FJ, Fernandez J, Gonzalez-Lavin L, Cernaianu A. Cause of failure and pathologic findings in surgicallyremoved Ionescu-Shiley standard bovine pericardial heart valve bioprostheses: emphasis on progressive structural deterioration. Circulation 1987;76:

8 The Journal of Thoracic and Volume 108, Number 5 Schoen et al Turina J, Hess OM, Turina M, Krayenbuehl HP. Cardiac bioprostheses in the 1990s. Circulation 1993;88: Lian JB, Levy RJ, Bernhard WF, Szycher M. LVAD mineralization and ),-carboxyglutamic acid containing proteins in normal and pathologically mineralized tissues. Trans Am Soc ArtifIntern Organs 1981 ;46: Barnhart GR, Jones M, Ishihara T, Rose DM, Chavez AM, Ferrans VJ. Degeneration and calcification of bioprosthetic cardiac valves: bioprosthetic tricuspid valve implantation in sheep. Am J Pathol 1982; I 06: Barnhart, GR, Jones M, Ishihara T, Chavez AM, Rose D M, F errans V J. Failure of porcine aortic and bovine pericardial prosthetic valves: an experimental investigation in young sheep. Circulation I 982;66(Suppl 1): Schoen FJ, Levy RJ, Nelson AC, Bernhard WF, Nashef A, Hawley M. Onset and progression of experimental bioprosthetic heart valve calcification. Lab Invest I 985;).i.: Jones M, Eidbo EE, Hilbert SL, et al. Anticalcification treatments of bioprosthetic heart valves: in-vivo studies in sheep. J Cardiac Surg 1989;4: II. Arbustini E, Jones M, Moses RD, et al. Modification ofthe Hancock T6 process of calcification of bioprosthetic cardiac valves implanted in sheep. Am J CardioI1984;53: Thiene G, Laborde F, Valente M, et al. Experimental evaluation of porcine-valved conduits processed with a calcium retarding agent (T6). J THORAC CARDIOV ASC SURG 1986; 91: Valente M, Minarini M, Ius P, et al. Durability of glutaraldehyde-fixed pericardia I v alve prostheses: clinical and animal experimental studies. J Heart Valve Dis 1992;1: Schoen FJ, Levy RJ, Hilbert SL, Bianco RW. Antimineralization treatments for bioprosthetic heart valves: assessment of efficacy and safety. J THORAC CARDTOV ASC SURG 1992;104: Gott GP, Chih P, Dorsey LMA, et al. Calcification ofporcine valves: a successful new method of antimineralization. Ann Thorac Surg 1992;53: Shemin RJ, Schoen FJ, Hein R, Austin J, Cohn LH. Hemodynamic and pathologic evaluation of a unileaflet pericardial bioprosthetic valve. J THORAC CARDIOV ASC SURG 1988;95: Irwin E, Lang G, Clack R, et al. Long-term evaluation of prosthetic mitral valves in sheep. J Invest Surg 1993;6: Levy RJ, Schoen FJ, Levy JT, et al. Biologic determinants of dystrophic calcification and osteocalcin deposition in glutaraldehyde-preserved porcine aortic valve leaflets implanted subcutaneously in rats. Am J PathoI1983;113: Schoen FJ, Tsao JW, Levy RJ. Calcification of bovine pericardium used in cardiac v alve bioprostheses: implications for the mechanisms of bioprosthetic tissue mineralization. Am J PathoI1986;123: Ferrans V J, Boyce SW, Billingham ME, Jones J, Ishihara T, Roberts We. Calcific deposits in porcine bioprostheses: structure and pathogenesis. Am J Cardiol 1980;46: Valente M, Bortolotti U, Thiene G. Ultrastructural substrates of dystrophic calcification in porcine bioprosthetic valve failure. Am J Pathol 1985;119: Webb CL, Schoen FJ, Flowers WE, Alfrey AC, Horton C, Levy RJ. Inhibition of mineralization of glutaraldehydepretreated bovine pericardium by AICl 3. Am J Pathol 1991 ;138: Schoen FJ. Interventional and surgical cardiovascular pathology-dinical correlations and basic principles. Philadelphia: WB Saunders, Webb CL, Nguyen NM, Schoen FJ, Levy RJ. Calcification of allograft aortic wall in a rat subdermal model. Am J PathoI1992;141: Thubrikar MJ, Deck JD, Aouad J, Nolan SP. Role of mechanical stress in calcification of aortic bioprosthetic valves. J THORAC CARDIOVASC SURG 1993;86: David TE, Bos J, Rakowski H. Aortic valve replacement with the Toronto SPV bioprosthesis. J Heart Valve Dis 1992;1 : Konertz W, Hamann P, Schwammenthal E, Breithardt G, Scheid HH. Aortic valve replacement with stentless xenografts. J Heart Valve Dis 1992;1: Cleveland DC, Williams WG, Razzouk AJ, et al. Failure of cryopreserved homograft valved conduits in the pulmonary circulation. Circulation 1992;86(Suppl):IIl Schoen FJ, Kujovich JL, Webb CL, Levy RJ. Chemically determined mineral content of explanted porcine aortic valve bioprostheses: correlation with radiographic assessment of calcification and clinical data. Circulation 1987;76: Sarris GE, Robbins RC, Miller DC, et al. Randomized, prospective assessment of bioprosthetic valve durability: Hancock versus Carpentier-Edwards valves. Circulation 1993;88: Schoen FJ, Kujovich JL, Levy RJ, Sutton MSJ. Bioprosthetic valve failure. Cardiovasc Clin 1988;18: