A tions in the United States annually [l]. Despite. Calcification of Porcine Valves: A Successful New Method of Antimineralization

Transcription

1 Calcification of Porcine Valves: A Successful New Method of Antimineralization John Parker Gott, MD, Pan-Chih, MD, Lynne M. A. Dorsey, MBA, John L. Jay, MD, G. Kimble Jett, MD, Frederick J. Schoen, MD, PhD, Jean-Marie Girardot, PhD, and Robert A. Guyton, MD Carlyle Fraser Heart Center, Crawford Long Hospital of Emory University, Atlanta, Georgia; Biomedical Design, Incorporated, Atlanta, Georgia; and Department of Pathology, Brigham and Women's Hospital and Harvard Medical School, Boston, Massachusetts Despite distinct advantages over mechanical cardiac valve prostheses, the use of bioprosthetic valves remains limited due to poor long-term durability, primarily as a result of tissue calcification. A novel anticalcification process, based on treatment of porcine bioprostheses with a derivative of oleic acid, has been developed by one of us (J.M.G.) (US Patent Number 4,976,733). This process employing 2-aminooleic acid (AOA) was tested in a juvenile sheep model. Terminal studies after a 20-week interval included hemodynamic, radiographic, morphologic, and quantitative tissue calcium analyses. All control valves (n = 4) had thickened, immobile, heavily calcified leaflets, whereas all AOA-treated valves (n = 8) were pliable and free of calcium deposits. Calculated valve orifice areas for controls (0.9 f 0.2 cm') (mean f standard error of the mean) was less than for AOA-treated valves (2.0 f 0.3 cm2) (p < 0.05). Radiographic calcification scores were greatly elevated in the control (25.5 f 5.6) versus AOA-treated valves (0.5 f 0.5) (p < 0.002). In quantitative mineralization studies, the mean calcium content of the control leaflets was 129 f 21 milligrams per gram dry weight cusp tissue versus 7.7 f 5.8 mg/g for AOA-treated valves (p < 0.001). Pathologic examination confirmed heavy calcification in the control leaflets, which was essentially absent in the AOA-treated leaflets. However, cuspal hematomas in areas of shuctural loosening and surface roughening were noted in AOA-treated valves. This anticalcification process dramatically reduced mineralization of porcine valve prostheses in this model. ( 1992;53:207-16) bout 40,000 patients undergo cardiac valve opera- A tions in the United States annually [l]. Despite advances in valvular reconstructive techniques, the majority of patients undergoing valve procedures receive prosthetic replacement [2]. Development of prosthetic devices has evolved mainly along two lines: mechanical and bioprosthetic. The advantage of superior durability for mechanical valves is offset by the requirement for patient anticoagulation to suppress thromboembolic problems. Conversely, bioprosthetic valve patients do not require indefinite anticoagulation, but the valves are prone to calcification and degenerative failure prompting reoperation. The ability to suppress this tissue mineralization could allow safer expansion of the benefits of bioprostheses to a wider patient population. Accordingly, a recently developed method of tissue anticalcification based on treatment with 2-aminooleic acid (AOA)[3] was tested using a standard porcine bioprosthetic valve in a juvenile sheep model previously demonstrated to mineralize tissue valves rapidly [&]. Presented at the Twenty-seventh Annual Meeting of The Society of Thoracic Surgeons, San Francisco, CA, Feb 16-20, Address reprint requests to Dr Guyton, Division of Cardiothoracic Surgery, Emory University School of Medicine, 25 Prescott St, Atlanta, GA Material and Methods Twelve juvenile sheep underwent successful mitral valve implantation with modification of a previously described technique [4-61. These animals, 2 males and 10 females, were 14 to 24 weeks of age with a body mass range of 25 to 31.5 kg at the time of implantation. After an overnight fast the animals were given atropine (0.2 mg/kg) and acetylpromazine (0.55 mg/kg) intramuscularly followed by shearing of the operative and access sites. Parenteral prophylactic antimicrobial injections (cephapirin and penicillin G benzathine and procaine suspensions) were begun preoperatively and continued for 72 hours. The animals were then given general anesthesia with ketamine (22 mgkg) followed by halothane inhalation and neuromuscular blockade with pancuronium (0.1 mg/kg). A large-bore rumen tube was passed orally for gastrointestinal decompression. The animals were placed in the right lateral decubitus position, the right femoral artery was cannulated for arterial inflow from the cardiopulmonary bypass circuit (Cobe VPCML, Lakewood, CO), and the left femoral artery was used for arterial pressure and blood gas monitoring. A left thoracotomy incision was used for cardiac exposure. Venous return was from retrograde cannulation of the right ventricle through the main pulmonary artery. Myocardial protection was afforded by antegrade cold, oxygenated, crystalloid cardioplegia and by The Society of Thoracic Surgeons /92/$5.00

2 208 GOTTETAL 1992;53: perfusion was carried out with moderate systemic hypothermia. The prosthetic mitral valves were placed in the supraannular position through a left atriotomy using interrupted atrially based pledgeted sutures. The mural leaflet and its chordal attachments were preserved. The animals were extubated within a few hours of operation and were ambulatory and taking food and water the evening of the operation. Parenteral analgesics were given for 48 hours postoperatively. The single left chest tube was intermittently aspirated and then removed 24 to 48 hours postoperatively. Therapeutic anticoagulation was not employed in any animal. Although thought to affect tissue calcification, there was no attempt to limit oral or intravenous calcium administration [7]. In addition to the 12 surviving animals there were two operative deaths. It was observed that these animals had coronary air emboli with postoperative ventricular dysfunction and arrhythmias leading to their deaths. In the right lateral decubitus position the left main coronary artery is superiorly oriented, and it readily and passively fills with air when the left heart is opened for valve insertion. A modification of the operative technique whereby the left main coronary artery was controlled with a Rumel tourniquet on a 3-0 polypropylene suture was instituted. This suture was passed around the left main coronary artery and the vessel was occluded during the aortic cross-clamp period. Occlusion began after cardioplegia delivery and continued until complete deairing of the cardiac chambers after removal of the cross-clamp. This maneuver appeared to have prevented coronary air emboli in the remainder of the experiments. These two operative deaths are not considered further. Of the 12 implants, four were control valves, which were clinical quality 25-mm porcine bioprostheses (Medtronic Heart Valve Division, Irvine, CA). The experimental valves were prepared by treating the same type of valve with AOA in a newly developed process [3]. The animals were housed in a controlled indoor environment, were exercised daily, had access to feed and water ad libitum, and were under close veterinary medical supervision during the preoperative, convalescent, and chronic phases of the study. The sheep were given levamisole nematocidal treatment preoperatively and vaccinated against tetanus, Clostridium perfringens C and D, vibrio, pasteurella, and ovine ecthyma. All animals received humane care in compliance with the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH publication No , revised 1985). After the 20-week implantation period the animals were anesthetized and instrumented through a median sternotomy for the hemodynamic studies. High-fidelity micromanometers (Millar Instruments, Houston, TX) were placed in the left atrium and left ventricle for continuous pressure measurements. Thermodilution cardiac outputs were obtained with balloon-tipped flow-directed pulmonary artery catheters (Edwards Swan-Ganz catheter; Baxter, Santa Ana, CA). The data were accumulated through an analog to digital conversion board (Data Translation, Inc, Marlborough, MA), sampled every 4 milliseconds, stored, and analyzed through a 386 microprocessor using an interactive program developed in this laboratory (James M. Bradford, PhD, copyright Emory University). Valve orifice areas were calculated using the modified Gorlin formula for the mitral valve [8]: MVA = CO/[(HR) (DFp ) (37.9) (a)] where MVA = mitral valve area, CO = cardiac output, HR = heart rate, DFp = diastolic filling period, 37.9 = empiric constant, and grad = mean mitral valve diastolic gradient. After the hemodynamic studies the animals were killed with an injection of a commercial euthanasia solution (Beuthanasia-D; Schering Corporation, Kenilworth, NJ). The valves were excised, gross photographs were taken, and x-ray exposures of the valves were then made using a uniform mammographic technique. The sheep underwent a formal postmortem examination by a boardcertified veterinary pathologist (Dirck Dillehy, DVM, PhD). A simple scoring system was employed to grade the radiographic appearance of the calcium deposits. The leaflet deposits were assigned scores based on the deposit size and number. A sum was then determined for each valve (see Appendix). The valves were placed in 4% formaldehyde. After fixation, the aortic wall and cusps were removed from the stent. A radial section of each valve cusp and its adjacent aortic wall was processed for histologic examination. The remainder of the leaflet segments were rinsed five times in distilled water, lyophilized overnight, weighed, and acid hydrolyzed under vacuum at 90 C with 2 ml of ultrapure redistilled 6 N HCl (GFS Chemicals). After 24 hours the samples were dried under vacuum and the residue resuspended in 1 ml of 0.3 N HCl then diluted with 0.01 N HCl. Quantitative calcium levels were then obtained using inductively coupled plasma analysis (Georgia Tech Research Institute, Atlanta, GA). Specimens were embedded in JB-4 glycol methacrylate medium (Polysciences, Inc, Warrington, PA). Sections 2 to 3 pm thick were cut with glass knives and stained with hematoxylin and eosin, von Kossa s stain (for calcium phosphates), and the Movat pentachrome stain (for collagen, elastin, and mucopolysaccharides). Statistical analyses were carried out on a microprocessor (Intel 386/24 megahertz, Santa Clara, CA) using proprietary software (RS1 release 4.2, BBN Software Products Corporation, Arlington, VA) with the advice of a statistician (Elmer Hall, PhD, Emory University, Atlanta, GA). Data are displayed as mean * standard error of the mean. Comparison between AOA and control groups for valve orifice areas was made with a one-tailed unpaired t test. The radiographic scores followed a non-gaussian distribution and Fisher s exact test was employed. Quantitative calcification scores were compared by two-tailed unpaired t tests. Significant differences were said to exist at the p < 0.05 level.

3 1992; GOTTETAL 209 centimeters squared Control T AOA Fig 1. Mean valve orifice areas for control and 2-aminooleic acidtreated (AOA) valve implants (mean * standard error of the mean). Results The sheep generally did well through the planned 20- week implant period. Four animals had an apparent perioperative peripheral nerve injury leading to a functional loss comparable with foot drop. Two injuries promptly resolved, and the two permanent losses were successfully treated by splinting of the extremity allowing satisfactory ambulation. One animal with a control valve implant had sudden development of signs of pulmonary congestion at 10 weeks after implantation and died within hours of onset. Hemodynamic studies were not done. Postmortem examination of that animal showed a tightly stenotic, heavily calcified valve with cardiac chamber and pulmonary findings compatible with severe mitral stenosis. Similar symptoms developed in 2 additional control valve animals, which underwent early elective terminal hemodynamic studies, one 13 and one 19 weeks after implantation. Only 1 of the AOA-treated valve animals did not reach the planned 20-week implant period. This animal had development of pneumonia and for humane reasons was put to death after hemodynamic study at 10 weeks. The mean increase in mass for the control valve sheep was 4.7 kg as compared with 16.4 kg for the AOA animals. The animals with valves treated with AOA had larger calculated valve orifices than the untreated, control-valve sheep. The calculated valve areas are displayed in graphic form in Figure 1. Typical digitized waveforms collected from a control and from a treated valve implantation are shown in Figure 2. Photographs representative of the gross appearance of the typical extensive calcification of the untreated valves and the freedom from calcium deposits of the AOAtreated valves are shown in Figure 3. As illustrated by the photographs, on gross examination the control valves were extensively mineralized with immobile leaflets and multiple calcific excrescences as seen with bioprosthetic valves explanted from patients after degenerative failure. The AOA-treated valves had pliable leaflets free of gross calcific deposits. However, focal cuspal hematomas and areas of cuspal thinning were noted in virtually all valves. I Light microscopic studies confirmed the difference in mineralization between AOA and untreated valves. However, in virtually all cases, AOA-treated valves had morphologic features suggesting generalized tissue degradation, including structural loosening, surface roughening, and deep cuspal collections of erythrocytes. Control, untreated valves had these features only rarely. Focal surface roughening was noted to be present on AOAtreated valves before implantation. Histologic features of treated and control valves are illustrated in Figure 4. The mammographic exposures (Fig 5) demonstrate the marked contrast between the radiolucent AOA-protected valves and the heavily opacified calcium deposits on the control valves. The highly significantly different grouped radiographic scores are given in the bar graphs of Figure 6. The leaflet quantitative calcification levels are highly significantly elevated in control valves compared with AOA-treated valves (Fig 7). By way of comparison, the quantitative calcium level for untreated, nonimplanted control valves (n = 6) was 0.6 k 0.08 mg calciudg dry tissue. Examination of individual AOA-treated valve quantitative calcium levels showed that six of the eight valves had levels so low that they were no different from the untreated, nonimplanted control valve levels. The salient feature found at necropsy was universal pulmonary or systemic vascular congestion compatible with mitral stenosis in the control valve sheep. Atelectasis and pneumonia were found in some animals in both groups. No macroscopic intracardiac thrombi were noted in any animals. None of the valves were thought to have evidence of endocarditis, and the sewing rings were well endothelialized. Comment The optimal surgical treatment of cardiac valvular disease may be restoration of the native valve to a satisfactory functional state through conservative reparative procedures [9]. Unfortunately, the valvular pathology most often encountered at operation is irreparable due to advanced rheumatic, degenerative, ischemic, or calcific destruction necessitating prosthetic valve replacement. For more than two decades reliable mechanical and bioprosthetic alternatives for valve replacement have been available. Homograft valves remain an option, but use continues to be limited by supply, increased technical demands for implantation, and yet unproven advantage. For most patients the prosthetic valve choice is based on an individualized judgment weighing the durability and increased risk of thromboembolism for the mechanical valve versus freedom from chronic anticoagulation but increased risk for tissue degeneration and reoperation for the bioprosthetic valve. The choice is influenced by patient age and life expectancy, access to medical care, childbearing plans, contraindication to anticoagulation, lifestyle, assessment of medical compliance, cardiac rhythm, alterations of calcium metabolism, sizing and gradient considerations, position, and institutional or surgeon biases [lo-151.

4 210 GOTTETAL 1992; pp ~~RV~1~%~95-88:W :15:18:47> Fig 2. (A) Representative digitized W E <# left atrial and left ventricular wave- R-l cap iro forms used for determination of transtxe-2#23$% #B FI f.rwp P2JUt valvular gradients and valve orifice areas through customized software. ;/;&;i2t rma 1.1 Note the high mean gradient of 43.8 mm Hg for the control valve I 48 after volum; loading to give a cardiac output of 8.6 Llmin. (B) Revre- 30 seniative waveforms for a treated valve after volume loading to give a cardiac output of 10.5 Llmin las8 18 I 2 Hit m y key to continue B 1 OW I wrr k k $1; k Caw ifrawt iaoou Pat. Put 1 hi2 1 : k hi Hit m y key to continue The clinical problem of porcine tissue valve failure is illustrated by the valve survival data (freedom from reoperation or death) from Bolooki and associates [16] which are representative of other series. After pooling their data from Carpentier-Edwards and Hancock implants (no significant difference), 5-year valve survival approximates 84% to 91% of implants and falls to 70% to 75% at 8 years. Follow-up from other series at 14 to 15 years revealed a freedom from valve failure as low as 31% to 41% [17, 181. It is well established that the leading cause of clinical bioprosthetic valve failure is primary tissue degeneration, which occurs in 60% to 75% of cases [13, 191. The most common clinical cause (approximately 85% [19]) of this degeneration is related to valve tissue calcification. As bioprostheses otherwise function well as valve replacement devices, the importance of the search for methods of attenuating the mineralization of these valves is evident. The pathogenesis of bioprosthetic calcification is incompletely understood but does not appear to be completely explained by a single mechanism. There are certainly factors that are host related, as well as those specific to the bioprosthesis itself. The physical forces to which the valve tissue is exposed [20, 211, recipient mineral metabolism [22], and host environment at the valve locus [l] may all play a role in facilitating calcification. It is clear that mechanical stress plays a role in promotion of calcification and is thus an active area of research and development efforts. However, it does not appear that mechanical deformation is necessary for mineralization to proceed [l, 13, 23, 241. Exposure to the extracellular ions associated with the host mineral metabolism is a prerequisite for bioprosthetic calcification. Systemic efforts to manipulate this variable have demonstrably reduced bioprosthetic calcification in animal models but have potentially serious ramifications through deleterious effects on bone metabolism [25]. Strategies for controlled release of antimineralization compounds in the vicinity of the valve might serve to reduce the systemic toxicity of this approach [26]. Certainly factors related to the bioprosthesis are involved in the initiation of calcification. A main focus for mitiga-

5 1992;53: GOTT ETAL 211 Fig 3. (A) Gross photograph of an explanted control valve demonstrating heavy calcification (arrows) of all leaflets. ( B ) An 2-aminooleic acidtreated valve that is well endothelialized and free of any mineralization. (C) Closer view of a control valve leaflet with a large cuspal calcium deposit (arrow). (D)Close-up of pliable, deposit-free, 2-aminooleic acid-treated leaflet.

6 212 GOTTETAL 1992:53: Fig 4. Light microscopic findings: ( A ) Control valve mineralization with large intrinsic calcific deposit. W o n Kossa stain [calcium phosphates black], x.50.) (B,C,D) Features of 2-aminooleic acid-treated valves: (B) Cuspul hematoma. (Movat stain, X.50.) (0Generalized structural loosening, with superficial (open arrow) and deep (large filled arrow) hematomas, and loss of surface integrity (small filled arrow). (Hemafoxylin and eosin, x.50.) ( D ) Surface roughening and disruption (arrow). (Hematoxylin and eosin, x 100.)'

7 1992; GOITETAL Control 213 AOA Fig 6. The mean radiographic scores for the two groups exhibiting the highly significant difference in calcitcm deposition (mean f standard error). (AOA = 2-aminwleic acid.) A N-lauryl sarcosine) [22], dyes (toluidine blue), impregnation techniques (polyacrylamide) [22], calcium binding or apatitic inhibitors (diphosphonates) [27], ionic antagonists (aluminum) [28], and other treatments have exhibited reduction of calcification in various animal models [22, 29, 301. Thus, many different approaches to anticalcification by modification of the biological material have been devised and have yielded variable reduction of mineral deposition. Unfortunately, no anticalcification technique has yet been demonstrated to deter mineralization in the clinical setting [31]. As research efforts have focused on the problem of bioprosthetic calcification it has become apparent that the glutaraldehyde preservation process is a two-edged sword. Glutaraldehyde is used for tanning all currently available clinical bioprostheses. Carpentier and associates [32] demonstrated that glutaraldehyde pretreatment in the manufacturing process established a tissue valve sufficiently durable for clinical use. Glutaraldehyde was B Fig 5. (A)Mammographic exposure of a control valve showing extensive areas of leaflet opacification corresponding to calcium deposition. (B)2-Aminooleic acid-treated valve mammograph which is radiolucent, indicating freedom from calcification. tion of mineralization has been an attempt to modify chemically the preserved biological material in a preimplantation process. A broad range of compounds and treatment processes have demonstrated antimineralization efficacyin protecting implanted Porcine aortic valves or bovine pericardial tissue. Various surfactants (sodium dodecyl sulfate [ie, T6], polysorbate-80, Triton X-100, Control AOA Fig 7. The mean quantitative leaflet calcium content of control and 2-aminooleic acid-treated (AOA) valve groups is very significantly different (mean k standard error).

8 214 GO'ITETAL 19!32; superior in this regard to formaldehyde. The glutaraldehyde-generated covalent cross-links between the bioprosthetic tissue collagen molecules explain the tensile advantage. Unfortunately, this same glutaraldehyde preservation method is a significant promoter of the bioprosthetic calcification process [23]. A new technique for bioprosthetic antimineralization was devised that takes advantage of the universal glutaraldehyde preservation technique for these devices. Using the free aldehyde groups associated with glutaraldehyde preservation, 2-aminooleic acid was covalently bound to porcine aortic leaflets and tested in the rat subcutaneous implant model [3]. These trials demonstrated essentially complete protection from calcification of the leaflets. The subsequent step in assessing anticalcification efficacy and possible clinical relevance and utility was large mammal testing of intact, treated valves. This anticalcification agent, 2-aminooleic acid, was applied in a proprietary process [3] after the manufacture of standard clinical quality and otherwise unmodified porcine bioprostheses. These valves were then tested in the mitral position in juvenile sheep. Jones and associates [P6] standardized and have extensive experience with a juvenile sheep model for chronic bioprosthetic valve implantation studies. The important feature of the model is the rapid calcification of standard porcine bioprostheses over weeks or months in a way that mimics the process that might take years in mature animals or humans. This allows timely acquisition of data on anticalcification treatments in clinically useful valve sizes and designs in an orthotopic position exposed to typical cyclical loading conditions and flows. Furthermore, Jones and colleagues [33] have demonstrated in these comparative studies that antimineralization efficacy in the small mammal subcutaneous implant model does not always correlate with efficacy in the chronic sheep intracardiac implants, thus emphasizing the need for this advanced screening methodology. There was absolute concordance in all tests designed to measure the valvular calcification in this juvenile sheep study. There was a clear reduction and near elimination of mineralization documented for the treated valves by every test employed. The hemodynamic studies demonstrated the functional correlation of the degree of calcification with stenosis in the control valves and lower gradients and larger areas with the AOA valves. The simple examination and palpation of the explanted valves was sufficiently convincing regarding the anticalcification effect, but this was clearly nonquantitative. To quantitate the difference in calcification between the two valve groups, two independent methods of determination were used. The x-ray examination of the valves was quantifiable and confirmed the success of AOA in preventing bioprosthetic mineralization. Previously described radiographic scoring techniques for assessment of bioprosthetic valve calcification were reviewed [33, 341. A simple method was used that measured the size and number of calcium deposits without particular emphasis on topographical distribution, as in the detailed scoring system of Gallo and associates [%I. For this screening study the complexity of scoring systems directed partly toward determining patterns and possible etiology of calcification was unnecessary. These mammographic examinations confirmed the protection from calcification by AOA. In addition to the x-ray score, a quantitative biochemical determination of leaflet calcium was made. This was the most rigorous and objective of the measures of success used in this study and was uniformly indicative of AOA protection compared with control valve calcium levels. Although no gross valve failures were noted in the AOA valves, subtle signs of structural deterioration were apparent. These features appear to be related to the technical aspects of AOA treatment, as modification of AOA valve processing methodology has eliminated the valve surface roughening while preserving the anticalcification properties of AOA in the rat subcutaneous model (351. The potential for toxicologic problems with AOA appears to be remote from the biocompatibility study data (on file, Biomedical Design, Inc). Furthermore, the compound is chemically similar to oleic acid, a normal component of human lipid metabolism. Although not yet tested in large mammal circulatory studies, this anticalcification process appears to hold promise for similar efficacy in pericardial bioprostheses [3]. Aminooleic acid treatment of porcine bioprostheses has demonstrated exceptional reduction of valvular tissue mineralization in a widely accepted animal model of accelerated bioprosthetic calcification. This process opens a novel avenue of study in the search for improved clinical bioprosthetic antimineralization. This project was supported in part by a grant from Biomedical Design, Incorporated, Atlanta, GA. Biomedical Design, Incorporated, is the assignee of the patent for this antimineralization process. Dr Girardot is the co-owner of Biomedical Design, Inc. No other author has any financial interest in this method. References 1. Schoen FJ, Harasaki H, Kim KM, Anderson HC, Levy RJ. Biomaterial-associated Calcification: pathology, mechanisms, and strategies for prevention. J Biomed Mater Res Appl Biomater 1988;22 A1:ll Cosgrove DM. Valve reconstruction versus valve replacement. In: Current heart valve prostheses. In: Crawford FA, ed. Cardiac surgery: state of the art reviews, vol 1. Philadelphia: Hanley and Belfus, Girardot JM. A method for retarding or preventing the calcification of a prosthesis implanted in a mammal. US patent no. 4,976, Jones M, Bamhart GR, Chavez AM, Jett GK, Rose DM, Ishihara T, Ferrans VJ. Experimental evaluation of bioprosthetic valves implanted in sheep. In: Cohn LH, Gallua V, eds. Cardiac bioprostheses. New York Yorke Medical, Bamhart GR, Jones M, Ishihara T, Rose DM, Chavez AM, Ferrans VJ. Degeneration and calcification of bioprosthetic cardiac valves. Am J Pathol 1982; Bamhart GR, Jones M, Ishihara T, Chavez AM, Rose DM,

9 1992; GOTTETAL Ferrans VJ. Bioprosthetic valvular failure. J Thorac Cardiovasc Surg 1982;83: Carpentier A, Carpentier S, Goussef N, et al. Calcium intake as a factor affecting calcification in glutaraldehyde treated biological material. Artif Organs 1982;5(Suppl):l. Cohen MV, Gorlin R. Modified orifice equation for the calculation of mitral valve area. Am Heart J 1972;84: Deloche A, Jebara VA, Relland JYM, et al. Valve repair with Carpentier techniques. The second decade. J Thorac Cardiovasc Surg 1990;99: Ferrans VJ, Boyce SW, Billingham ME, Jones M, Ishihara T, Roberts WC. Calcific deposits in porcine bioprostheses. Structure and pathogenesis. Am J Cardiol 1980;46: Gallo I, Nistal F, Blasquez R, Arbe E, Artiiiano E. Incidence of primary valve failure in porcine bioprosthetic heart valves. 1988;45: Geha AS, Laks H, Stansel HC, et al. Late failure of porcine valve heterografts in children. J Thorac Cardiovasc Surg 1979;78: Schoen FJ, Kujovich JL, Levy RJ, St. John Sutton M. Bioprosthetic heart valve pathology, clinical pathologic features of valve failure and pathobiology of calcification. Cardiovasc Clin 1988;18: Kirklin JW, Barrett-Boyes BG. Mitral valve disease with or without tricuspid valve disease. In: Cardiac surgery. New York Wiley, 1986:364. Spencer FC. Acquired disease of the mitral valve. In: Sabiston DC, Spencer FC, eds. Gibbon s surgery of the chest. Phdadelphia: W. B. Saunders, 1989:1235. Bolooki H, Kaiser GS, Mallon SM, et al. Comparison of long-term results of Carpentier-Edwards and Hancock bioprosthetic valves. 1986;42: Foster AH, Greenberg GJ, Underhill DJ, McIntosh CL, Clark RE. Intrinsic failure of Hancock mitral prostheses: 10 to 15-year experience. 1987;44: Gallucci V, Mazzuco A, Bortolotti U, Milano A, Guerra F, Thiene G. The standard Hancock porcine bioprosthesis: overall experience at the University of Padova. J Cardiac Surg 1988;3(Suppl): Schoen FJ, Hobson CE. Anatomic analysis of removed prosthetic heart valves: causes of failure of 33 mechanical valves and 58 bioprostheses, 1980 to Hum Pathol 1985;16: Wright JTM, Eberhardt CE, Gibbs ML, Saul T, Gilpin CB. Hancock 11: an improved bioprosthesis. In: Cohn LH, Gallucci V, eds. Cardiac bioprostheses. New York Yorke Medical Books, Thubrikar MJ, Deck JD, Aouad J, et al. Role of mechanical stress in calcification of aortic bioprosthetic valves. J Thorac Cardiovasc Surg 1983;86: Carpentier A, Nashef A, Carpentier S, Ahmed A, Goussef N. Techniques for prevention of calcification of valvular prostheses. Circulation 1984;7O(Suppl 1): Levy RJ, Schoen FJ, Levy JT, et al. Biological determinants of dystrophic calcification and osteocalcin deposition in glutaraldehyde-preserved porcine aortic valve leaflets implanted subcutaneously in rats. Am J Pathol 1983;113: Schoen FJ, Levy RJ, Nelson AC, Bemhard WF, Nashef A, Hawley M. Onset and progression of experimental bioprosthetic heart valve calcification. Lab Invest 1985;52: Levy RJ, Schoen FJ, Lund SA, Smith MS. Prevention of leaflet calcification of bioprosthetic heart valves with diphospho- nate injection therapy. Experimental studies of optimal doses and therapeutic durations. J Thorac Cardiovasc Surg 1987; Golumb G, Langer R, Schoen FJ, et al. Controlled release of diphosphonate to inhibit bioprosthetic heart valve calcification: dose-response and mechanistic studies. J Control Re1 1986;4: Levy RJ, Hawley MA, Schoen FJ, et al. Inhibition by diphosponate compounds of calcification of porcine bioprosthetic heart valve cusps implanted subcutaneously in rats. Circulation 1985;71: Webb CL, Schoen FJ, Levy RJ. A]+++ preincubation inhibits calcification of bioprosthetic heart valve tissue in the rat subdermal model. Trans Am SOC Artif Organs 1988;34: Arbustini E, Jones M, Moses RD, Eidbo EE, Carroll RJ, Ferrans VJ. Modification by the Hancock T6 process of calcification of bioprosthetic cardiac valves implanted in sheep. Am J Cardiol 1984;53: Krammsch DM, Chan CT. The effect of agents interfering with soft tissue calcification and cell proliferation on calcific fibrous-fatty plaques in rabbits. Circ Res 1978;42: Oury JH, Angel1 WW, Koziol JA. Comparison of Hancock I and Hancock I1 bioprostheses. J Cardiac Surg 1988;3: Carpentier A, Lemaigre G, Robert L, et al. Biological factors affecting long term results of valvular heterografts. J Thorac Cardiovasc Surg 1969;58: Jones M, Eidbo EE, Hilbert SL, Ferrans VJ, Clark RE. The effects of anticalcification treatments on bioprosthetic heart valves implanted in sheep. Trans Am SOC Artif Organs 1988;X Gallo I, Nistal F, Artiiiano E, et al. The behavior of pericardial versus porcine valve xenografts in the growing sheep model. J Thorac Cardiovasc Surg 1987;93: Girardot MN, Girardot JM, Schoen FJ. Alpha amino oleic acid, a new compound, prevents calcification of bioprosthetic heart valves. Trans SOC Biomater 1991;14:114. Appendix 1. Table used for determining radiographic calcification scores Calcium >O-lmm >1-2mm >2-3mm >Mmm deposit size Weighted score for each deposit by size Leaflet 1 x,x... X... 2 X,XJ X J... X 3 X J x,x Subtotals Sum of all leaflet scores 22 equals valve radiographic score x = one calcium deposit corresponding to size range of column heading.

10 216 GOlTETAL 1992; DISCUSSION DR RICHARD E. CLARK (Pittsburgh, PA): Dr Michael Jones at the National Heart, Lung, and Blood Institute is the person who developed the juvenile sheep model for testing various valves, particularly those that use biologic tissues. Over the past 11 years he has implanted more than 1,500 heart valves in sheep. He was unable to be here and kindly consented to my request to show some of his data. Data from more than 240 animals demonstrated the effects of various antimineralization treatments for porcine aortic valves and the lack of any salutory effects for bovine pericardial valves. Calcium content was determined by atomic spectroscopy, and was measured in milligrams per gram. There were a few treatments with very low calcium content after 20 weeks of implantation. Polysorbate-treated porcine valves had 7.6 mg of calcium per gram of tissue, which was the lowest value found. Dr Gott s findings when compared with the data of Dr Jones show that this new process results in calcium content as low as Dr Jones has been able to find in 10 years with a host of other treatments. I would like to make a second comment on Dr Gott s paper and ask him a couple of questions. In data from a large series comprising 18 standard porcine valves and 15 treated by the PV2 process implanted into the mitral position in juvenile sheep for 20 weeks, there was a very wide distribution of calcium content for each valve treatment as a function of individual animals. Dr Gott showed similar data. This emphasizes the point that one needs a large number of complete experiments to determine if there is a high degree of efficacy. Dr Gott had only eight treated valves. If one takes the eight lowest values for standard valves from Dr Jones s experience, one might have made a different interpretation of the degree of efficacy of Dr Girardot s process. The first question is, are you going to increase the number of animals in the study to determine the distribution of calcium content? The second question is, did the treatment alter the ultrastructure of collagen? Dr Victor Ferrans in the Pathology Branch at the National Heart, Lung, and Blood Institute and Dr Jones have been able to find that some anticalcification treatments affect the major collagen bundles, suggesting a destructive effect that may become manifest as decreased fatigue resistance to flexion. Have your explanted valves had ultrastructural analyses, and have nonimplanted valves been tested in a durability tester? DR GOlT: In answer to your first question, indeed, we are excited enough in Dr Guyton s laboratory with this compound and this technique to pursue this further. We do plan another series of experiments to increase the number of animals with the treated valve that have been studied with the new modification of the technique. This modification appears to eliminate these structural problems. With regard to the second question, no, these valves have not been studied at the ultrastructural level. However, they have been studied extensively at the light microscopic level. This was the only source of concern in the whole study, that is, the structural loosening and surface roughening that I mentioned in the treated valves. Again, this appears to have been remedied by a change in the technique for application of the AOA to the valve itself. With regard to collagen chemistry or deformation or degrada- tion or change related to the AOA treatment, I cannot comment specifically on that. DR JACK W. LOVE (Santa Barbara, CA): Dr Gott and his colleagues have presented really impressive results with Dr Girardot s anticalcification treatment of porcine valves in the juvenile sheep model. This represents one more approach at attempting to retard calcification and improve durability of the xenograft valve. The issue of antigenicity of xenograft tissue as a determinant of calcification and durability has not been settled. Xenograft tissue after glutaraldehyde tanning retains some antigenicity. This was clearly shown by Carpentier and subsequent investigators. We have continued to pursue the concept of a nonantigenic, noncalcifying, more durable bioprosthesis made of autogenous tissue in the operating room at the time of operation for valve replacement. Our valves are made in the operating room from a rectangle of pericardium treated with a 5-min immersion in 0.6% glutaraldehyde. The valve can be fabricated in about 5 minutes. We have implanted these valves in the juvenile sheep model for 5 months, 20 weeks, with no other treatment of the tissue. These valves are competent by ventricular injection at cardiac catheterization, the hemodynamics are normal, the tissue is soft and pliable, and the valves are not calcified. Dr Gott, have you looked at the antigenicity of your treated and your untreated valves? DR ROBERT W. M. FRATER (Bronx, NY): I do not think you answered Dr Clark s question, which is, have you tested these devices in the wear tester after treating them with the AOA, and do you have a notion of what is happening to the valve and to the glutaraldehyde portion of it when you treat it with AOA? FROM THE AUDIENCE: Is the AOA chemically bound, or could it be that the 20-week test, accelerated test of calcification, is also an accelerated test of washout? It may be in the clinical situation the AOA will not be there in 2 years when we really need it. DR GOTT: I apologize for not answering Dr Clark s question about the mechanical wear testing as Dr Frater pointed out. The new treatment, the modification of the AOA treatment, has been tested in a mechanical cycle device in more than 50 million cycles, which corresponds to years of human cardiac cycling, and it does appear that there is no mechanical problem with the valve. With regard to the antigenicity question, in the literature there appears to be some ambiguity about whether the T or B cell function plays a role in the calcification of these valves. The answer is not quite clear. That is a fascinating valve that you made there in the operating room. In answer to the specific question, no, we did not look for any elements of antigenicity in our model. The only thought we had about antigenicity was that we used bovine source heparin rather than porcine heparin when we placed the valves. With regard to the chemical properties of AOA and the interaction with the valve, again, this is speculation, but the compound is designed to have a covalent link with the valve, and it appears in rat subcutaneous washout studies that the compound does remain bound significantly to the valve tissue over weeks and months.