The clinical results for 4 C stored and cryopreserved

Allograft Heart Valve and Valve- Processing s Kenneth L. Gall, BApplSc, Susan E. Smith, BSc, Christene A. Willmette, EN, and Mark F. O Brien, FRACS Department of Cardiac Surgery, The Prince Charles Hospital, Chermside, Brisbane, Australia Background. The impact of allograft valve viability on valve durability remains controversial. Analyses of our clinical results have demonstrated the superiority of the cryopreserved valve viable at the time of implantation over the 4 C stored valve nonviable at the time of implantation. In this study, we quantitatively assessed the effects on viability of current and past valve-processing protocols at The Prince Charles Hospital. Methods. The viability of pulmonary valves was quantitatively analyzed by thin-layer autoradiography to assess the effects of donor type, antibiotics, and valve storage. Results. Control valve segments obtained from beating-heart donor valves had a higher initial viability (0.92 0.02) than nonbeating-heart donor valves (0.66 0.03). Cryopreservation after low-dose antibiotic sterilization significantly reduced viability to 50% to 60% of the control, and in the presence of amphotericin B, viability dropped further to 10% to 36% of the control. After 7 days storage at 4 C, viability was reduced to 2% of control and to 0% viability after 21 days. Conclusions. For maximal preimplantation viability, valves should be procured as soon as possible after cessation of heart beat and should be cryopreserved if they are not to be clinically implanted within 1 to 2 days. Amphotericin B should not be used in conjunction with cryopreservation if viability is to be maximized. (Ann Thorac Surg 1998;65:1032 8) 1998 by The Society of Thoracic Surgeons The clinical results for 4 C stored and cryopreserved allograft aortic valves implanted at The Prince Charles Hospital (TPCH) continue to show after many analyses over time, the superiority of allografts viable at the time of implantation [1 3]. Many factors, including viability, can influence the durability of an allograft as assessed by freedom from structural deterioration. The importance of viability is disputed by many. There is no strong substantial evidence that viable cells persist and produce collagen and repair damaged matrix. Their presence (before implantation) indicates optimal tissue preservation overall. This at least may be the strongest argument for aiming for and maintaining allograft valve viability in valves for clinical implantation. The clinical outcome at TPCH for viable allograft aortic valves with respect to structural deterioration has been clearly defined [1] and compares more than favorably with other allograft valve series [4]. The results obtained at TPCH are of particular importance because of the level of control this institution has had over all aspects of collection, processing, and storage. The protocol for valve collection from nonbeating-heart donors has remained virtually the same for more than 27 years, and its most important aspect is an insistence on early collection, ie, within 24 hours after death (mean time, 15 hours). The antibiotic sterilization protocol has been altered only Accepted for publication Nov 3, 1997. Address reprint requests to Dr O Brien, Department of Cardiac Surgery, The Prince Charles Hospital, Rode Rd, Chermside, Brisbane 4032, Queensland, Australia. twice: to respond to the potential deleterious effects of amphotericin B and to reduce antibiotic incubation time from 24 hours to 6 hours (Table 1). A recent retrospective analysis of the microbiology of tissue samples obtained at collection, at processing, after antibiotic sterilization, and at implantation has shown the effectiveness of the current TPCH protocol of only 6 hours antibiotic sterilization [5]. Routine allograft valve viability assessment has been determined at TPCH by qualitative techniques including glucose utilization and tissue culture on co-processed pulmonary and tricuspid valve leaflet segments [1, 6]. However, for a valid analysis of factors affecting viability, a quantitative method of viability assessment was used. The method chosen reflected the ability of a selected cell type to perform a normal activity that is absent after disruption of normal cellular function. Fibroblasts, important in the maintenance of the valve matrix [7], are an ideal cell for such an analysis. They are known to be present in allograft valves at the time of implantation and, in the TPCH analyses, have been demonstrated to be present, although markedly variable and most often in reduced numbers, at valve explantation [8, 9]. A normal cellular function of fibroblasts is the incorporation of amino acids to produce collagen precursors. Thus, the viability of fibroblasts was assessed by the uptake of isotope-labeled proline as demonstrated by a thin-layer autoradiographic method based on the work of Van Der Kamp and associates [10]. This method of assessing viability, because it is based on cellular function, is 1998 by The Society of Thoracic Surgeons 0003-4975/98/$19.00 Published by Elsevier Science Inc PII S0003-4975(98)00085-X

Ann Thorac Surg GALL ET AL 1998;65:1032 8 PROCESSING AND ALLOGRAFT VALVE VIABILITY 1033 Table 1. Protocols for Allograft Valve Antibiotic Sterilization and Storage at The Prince Charles Hospital From 1969 to Date Interval (y) Antibiotics Incubation Time at 37 C (h) Storage Method 1969 1975 P, S, A 24 Refrigerated 4 C 1975 1988 P, S, A 24 Cryopreserved 1988 to date P, S 6 Cryopreserved A amphotericin B (10 g/ml); P penicillin (30 g/ml); S streptomycin (50 g/ml). considered superior to other methods including thiamine uptake and electron and light microscopy. The aim of this study was to assess quantitatively the effects on valve viability of the current TPCH protocol of valve procurement, antibiotic exposure, cryopreservation, and cryoprotectant removal at valve thaw. Previous TPCH protocols (see Table 1), including the use of the antifungal agent amphotericin B, 24-hour incubation, and refrigerated storage, were investigated to link the historical results of nonviable 4 C stored valves at TPCH. Gentamicin sulfate, although never used clinically at TPCH, was studied as a possible alternative to streptomycin sulfate. Material and Methods Valves Pulmonary allograft valves were used in this investigation in accordance with TPCH Ethics Committee (approval No. EC9104). Valves were obtained from two sources: beating-heart donors (multiorgan donors and heart transplant recipients) and nonbeating-heart donors where valves were procured within 24 hours after cessation of heart beat (autopsy source). The leaflets of the pulmonary valves were divided into two pieces. Thus each pulmonary valve used in this investigation provided one control segment and five segments for experimental use. Experimental Series Four series of experiments were performed to assess both the current and past methods of valve processing and storage. The experiments were performed on a minimum of 5 valves for each series. Series 1 used beating-heart donor valves (n 5) and investigated the effects of short exposure to antibiotics at room temperature, cryopreservation, and method of cryoprotectant removal (Table 2). Gentamicin, a possible substitute for streptomycin, was also studied. Series 2 used nonbeating-heart donor valves (n 5) (Table 3). treatment was similar to that in Series 1. However, an additional 6-hour 37 C incubation in antibiotics before cryopreservation, as used clinically, was performed. Series 3 used nonbeating-heart donor valves (n 5) and examined the effect of the presence of amphotericin B in the antibiotic sterilization medium (Table 4). Refrigerated storage was also studied in this series. Series 4 used nonbeating-heart donor valves (n 7) and investigated a reduced time exposure to amphotericin B and also non controlled-rate cryopreservation (CRC) (Table 5). Initial Immediately on receipt of the heart or heart valve block, a control half-leaflet segment was removed from the pulmonary valve and stored in 20 ml of Nutrient Medium 199 with Hanks salts (M199) (CSL Limited, Parkville, Victoria, Australia) at 4 C while the aortic valve was processed for clinical use. The remaining pulmonary valve segments were removed and placed in 20 ml of M199 with antibiotics and stored according to the protocol for the particular experimental series (see Tables 2 through 5). Cryopreservation For CRC, valve segments were packaged in 100 ml of M199 containing 10% dimethyl sulfoxide (DMSO) (Merck, Darmstadt, Germany) and frozen at a rate of Table 2. Series 1: Beating-Heart Donor Pulmonary Valves (n 5) Medium M199 PS PS PS PG PG Temperature 4 C RT RT RT RT RT Time (min) 60 60 60 60 60 60 CRC...... Y Y... Y DMSO washout...... Nonseq Seq... Seq Proportion 0.92 0.02 0.92 0.03 0.60 0.04 0.58 0.05 0.87 0.04 0.52 0.04 % Control... 99 65 63 95 56 CRC controlled-rate cryopreservation; DMSO dimethyl sulfoxide; M199 medium 199; Nonseq nonsequential; % Control mean percentage of initial viability (segment 1); PG medium 199 with penicillin (30 g/ml) and gentamicin (10 g/ml); PS medium 199 with penicillin (30 g/ml) and streptomycin (50 g/ml); Proportion viable fibroblasts as a proportion of total fibroblasts; RT room temperature; Seq sequential; Y yes.

1034 GALL ET AL Ann Thorac Surg PROCESSING AND ALLOGRAFT VALVE VIABILITY 1998;65:1032 8 Table 3. Series 2: Nonbeating-Heart Donor Pulmonary Valves (n 5) Medium M199 PS PS PS PG PG Temperature 1 4 C RT RT RT RT RT Time 1 (min) 60 60 60 60 60 60 Temperature 2 ( C)... 37 37 37 37 37 Time 2 (h)... 6 6 6 6 6 CRC...... Y Y... Y DMSO washout...... Nonseq Seq... Seq Proportion 0.58 0.06 0.61 0.05 0.29 0.08 0.32 0.05 0.61 0.03 0.32 0.08 % Control... 105 51 54 104 56 Abbreviations are the same as in Table 2. approximately 1 C per minute in a computer-controlled freezer (model 1010A; Cryomed, Mount Clemens, MI). 6 in series 4 was similarly packaged and underwent a non-crc process. The package was wrapped in cotton wool, placed inside a Styrofoam container, and suspended in the liquid nitrogen vapor immediately above the liquid nitrogen. The cryopreserved segments were stored in the vapor phase of a liquid nitrogen freezer at lower than 135 C for a minimum of 7 days before thawing. The non-crc process was used to mimic the clinical protocol before the introduction of CRC when valves were cryopreserved by suspension in the liquid nitrogen vapor phase. Thawing and DMSO Removal s were thawed by the current TPCH protocol of package immersion into a warm (42 C) normal saline solution bath. The temperature of the saline solution was maintained at 37 C during thawing. As soon as each segment was thawed, it was taken from the freezing solution. The DMSO was removed from the thawed segments by a single-step dilution or stepwise dilution method. The single-step dilution method involved four rinses of 2 minutes each in M199. The stepwise dilution method involved four rinses of 2 minutes each in M199 containing 5% DMSO, 2.5% DMSO, 0% DMSO, and 0% DMSO, respectively. All rinse solutions were maintained at lower than 10 C. Autoradiography proline uptake. s were rinsed in 20 ml of HAMS F-10 medium (ICN Pharmaceuticals Australasia Pty Ltd, Sydney, Australia) before being placed in vials containing 7.5 ml of HAMS F-10 medium, 0.75 ml of fetal bovine serum (CSL Limited), and 0.1 ml of l-[5-3 H]proline (37.0 MBq/mL) (Amersham Australia, Sydney, Australia). The open vials were placed in a 5.0% carbon dioxide incubator at 37 C for 6 hours. histologic processing and sectioning. After proline uptake, segments were placed in vials of cold (4 C) phosphate-buffered saline solution for two rinses of 5 Table 4. Series 3: Nonbeating-Heart Donor Pulmonary Valves (n 5) Medium M199 PSA PSA PSA PSA PSA Temperature 1 4 C RT RT RT RT RT Time 1 (min) 60 60 60 60 60 60 Temperature 2 ( C)... 37 37 37 37 37 Time 2 (h)... 24 24 24 24 24 CRC...... Y Y...... DMSO washout...... Nonseq Seq...... 4 C (days)............ 7 21 Proportion 0.62 0.05 0.53 0.05 0.10 0.02 0.06 0.01 0.01 0.01 0 % Control... 85 17 10 2 0 PSA medium 199 with penicillin (30 g/ml), streptomycin (50 g/ml), and amphotericin B (10 g/ml); Table 2. other abbreviations are the same as in

Ann Thorac Surg GALL ET AL 1998;65:1032 8 PROCESSING AND ALLOGRAFT VALVE VIABILITY 1035 Table 5. Series 4: Nonbeating-Heart Donor Pulmonary Valves (n 7) Medium M199 PSA PSA PSA PSA PSA Temperature 1 4 C RT RT RT RT RT Time 1 (min) 60 60 60 60 60 60 Temperature 2 ( C)......... 37 37... Time2(h)......... 6 6... CRC...... Y... Y... Non-CRC............... Y DMSO washout...... Seq... Seq Seq Proportion 0.74 0.04 0.64 0.07 0.27 0.04 0.71 0.05 0.26 0.06 0.20 0.04 % Control... 85 36 100 34 26 Non-CRC non-controlled-rate cryopreservation; PSA medium 199 with penicillin (30 g/ml), streptomycin (50 g/ml), and amphotericin B (10 g/ml); other abbreviations are the same as in Table 2. minutes duration to remove unincorporated proline. The leaflets were then put in vials containing 20 ml of 4% phosphate-buffered formaldehyde for 30 minutes before being transferred to an automatic tissue processor for routine histologic fixation, dehydration, wax embedding, and sectioning. nuclear emulsion. Slides were placed in slide racks, hydrated, and dipped into K2 nuclear emulsion (Ilford Imaging Australia, Mount Waverley, Victoria, Australia) to provide a thin film over the sections under darkroom conditions. The slides were dried in a 37 C oven and placed in lightproof boxes with silica gel for 21 days at 4 C. developing, staining, and mounting. Slides were developed (D-10; Kodak, Australia) and fixed (Na 2 S 2 O 3, 15.6 g/l, and K 2 S 2 O 5, 1.6 g/l). They were then stained with a routine hematoxylin and eosin method and mounted with glass coverslips. Counting The slides were randomly labeled with a sequential number to blind the observer performing the count. A single observer was used to ensure that the same criteria for viability were applied to all sections and that other cell types, eg, endothelial cells and white blood cells, were not included in the count. Fibroblast viability was indicated by a greater concentration of silver grains around the cell nucleus than in the background (Fig 1). Nonviable fibroblasts were easily identified by the absence of the silver grains. For each slide, one section was selected and examined under low- and then high-power light microscopy. Five random areas were observed in that section, and a total of between 700 and 1,550 fibroblasts were counted for each section examined. The code for the slide number was then broken, and the data for each segment were collated. The viability was calculated as the proportion of viable fibroblasts to total fibroblasts. A percentage of the initial viability (control) for similarly treated segments was also calculated and used for statistical comparisons between treatments. Statistical Analysis The mean percentage of the initial viability (control) was used for statistical comparisons between treatments within the same series and between series. A two-way analysis of variance without replication was used for this analysis with a p value of less than 0.05 considered significant. A regression analysis was performed on the initial viability (control) at processing over time from cessation of heart beat to collection with a p value of less than 0.05 being significant. Results A graphic representation of the initial viability (controls) of all valves examined plotted against time from asystole to processing for beating-heart donors and nonbeating- Fig 1. of pulmonary leaflet after autoradiography. Silver grains clustered around a cell nucleus identify a viable fibroblast. Nonviable fibroblasts are without silver grains. (Hematoxylin and eosin; 400 before 50% reduction.)

1036 GALL ET AL Ann Thorac Surg PROCESSING AND ALLOGRAFT VALVE VIABILITY 1998;65:1032 8 Fig 2. of pulmonary valves at time of processing: a comparison between beating-heart and nonbeating-heart donors. This represents the control viability for each donor. heart donors is depicted in Figure 2. A regression analysis of valve viability from all sources produced a correlation coefficient (r) of 0.692 and a test for correlation (p) of 0.0004. The beating-heart group had an r of 0.994 and a p of 0.0006 and the autopsy group, an r of 0.268 and a p of 0.298. The results for series 1 through 4 are shown at the end of Tables 2 through 5, respectively. The viability for each segment exposed to a particular treatment is expressed as the mean viability the standard error and as the mean percentage of the initial (control) viability. Pulmonary valve leaflet segments obtained from beating-heart donors, series 1 (see Table 2), showed a high initial (control) proportion of viable cells, 0.92 0.02. Exposure to antibiotics for 60 minutes at room temperature at the concentrations used clinically resulted in an insignificant decrease for penicillin and streptomycin to 99% of control and for penicillin and gentamicin to 95% of control. The current clinical protocol of antibiotic sterilization, CRC, and a stepwise dilution method for DMSO removal significantly reduced viability to 63% of control (p 0.05). Neither the single-step dilution method of DMSO removal (viability, 65% of control) nor the substitution of gentamicin for streptomycin in the antibiotic cocktail (viability, 56% of control) was significantly different. Pulmonary valve leaflet segments from the series 2 group of nonbeating-heart donors (see Table 3) showed the lower initial (control) proportion of viable cells of 0.58 0.06 encountered with nonbeating-heart donors. This did not change significantly after 60 minutes exposure to antibiotics (penicillin and streptomycin) at room temperature. The current clinical protocol for valves from this source resulted in a viability level of 54% of control after cryopreservation and DMSO washout. There was no significant difference between the two antibiotic cocktails used or the method of DMSO washout. The initial (control) proportion of viable cells from leaflet segments obtained from the group of nonbeatingheart donor pulmonary valves used in series 3 (see Table 4) was 0.62 0.05. decreased to 85% of control after 24 hours incubation with penicillin, streptomycin, and amphotericin B, but this was not significant. Subsequent cryopreservation followed by nonsequential washout of DMSO as in the processing protocol used at TPCH from 1975 to 1988 resulted in a significant reduction in viability to 17% (p 0.05) of control. No significant difference between the two methods of DMSO removal was noted. Storage of valves at 4 C for 7 days significantly reduced viability to 2% of control, and after 21 days, no viable cells were observed in any section. Pulmonary valve leaflet segments from the series 4 group of nonbeating-heart donors (see Table 5) demonstrated an initial (control) proportion of viable cells of 0.74 0.04. decreased to 85% of control after 60 minutes exposure to antibiotics (penicillin, streptomycin, and amphotericin B) at room temperature. However, this was not significant. A further 6-hour incubation in these antibiotics at 37 C did not reduce viability. However, after cryopreservation, viability was significantly reduced to 36% of control (p 0.05) after 60 minutes exposure at room temperature prior to cryopreservation and to 34% of control (p 0.05) after 6 hours incubation at 37 C prior to cryopreservation. The use of a non-crc technique produced a lower level of viability, 26% of control but this was not significant. Comment A number of methods for determining valve viability have been proposed. Methods using the uptake of radiolabeled thymidine, which relies on the division of cells as a marker of viability, are unsatisfactory for fibroblasts, which by nature are slow dividing [11]. A variety of other methods including assessment of collagen synthesis, protein synthesis, dye uptake, and dye exclusion are available. The method selected must apply to the cell type investigated [12]. Autoradiography has the advantage of assessment of cell function over a period of 6 hours and also examination of the cells in situ in their natural environment. Repeat counting and determination of overall valve cellularity can also be performed with this technique. For these reasons, autoradiography after tissue uptake of radioactive proline was selected as the quantitative method of determining viability. Because of the paucity of aortic valves for research use, the viability of pulmonary valves was examined in this study. Although the correlation between pulmonary and aortic valve viability has not been assessed, there is no obvious reason to suspect that there would be a significant difference. The initial viability of beating-heart valve donors, 0.92 0.02, compared with that of nonbeating heart donors, 0.66 0.03, indicates the superior viability of this valve source. Figure 2 demonstrates the reduction in viability with time from cessation of heart beat. In the beating-heart donor group, there is a significant correlation between time for processing and viability but not in the autopsy group. A longer time interval was not investigated, as all retrievals occur within the strict 24-hour period. Other factors such as donor weight, cooling, and initial leaflet cellularity may be important in determining

Ann Thorac Surg GALL ET AL 1998;65:1032 8 PROCESSING AND ALLOGRAFT VALVE VIABILITY 1037 the viability at the time of collection. For instance, a small donor with high cellularity may cool rapidly and provide a high level of viability at the time of valve procurement. The shortage of pediatric donors in our experience may warrant extending the 24 hours from death to collection interval in this particular group. However, the impact of extending this interval on the risk of increasing microbial contamination would require assessment before clinical protocol modification. The antibacterial effectiveness of the current TPCH antibiotic sterilization protocol of incubation in low-dose penicillin and streptomycin for the short time of only 6 hours at 37 C has been reported [5]. Now the gentleness of this protocol has been further demonstrated by the evidence of the maintenance of viability after incubation. The use of amphotericin B was discontinued at TPCH in 1988 in response to its potential deleterious effects on cell membranes. The results of this investigation confirm the reported toxicity of amphotericin B [13, 14]. Whereas the use of amphotericin B at 37 C for 6 and 24 hours had no significant effect on viability before cryopreservation, the combination of exposure to amphotericin B and cryopreservation was very deleterious to viability. An exposure time of 24 hours followed by cryopreservation produced a significantly lower level of viability (10%) compared with the level of viability obtained when the exposure was limited to 60 minutes (36%) and 6 hours (34%) followed by cryopreservation. The presence of the cryoprotectant DMSO increases the osmolarity of the freezing medium. Reports suggest that the DMSO should be removed gradually from the tissue to reduce the risk of rapid water movement into the cell, as this may lead to cellular disruption [15, 16]. In this study, a simple single-step method of DMSO removal produced a slightly lower level of viability than the stepwise dilution method of cryoprotectant removal, but the difference was not significant. The use of refrigerated storage of leaflet segments for 7 days resulted in a significantly lower viability than that obtained for all incubation and cryopreservation methods investigated. We do not recommend the practice of refrigerated storage for a few days or weeks followed by cryopreservation. Even if viability is not a consideration, the possibility of deterioration of the noncellular matrix requires attention. Controversy has surrounded and continues to surround the importance of the presence of viable fibroblasts within the allograft leaflet matrix at the time of implantation. At TPCH, there is evidence that cryopreserved valves, viable at the time of cryopreservation, have a much lower level of structural deterioration than nonviable valves, and this is supported by other published valve series [4]. In addition, from the work by Yacoub and colleagues [17] on homovital valves (from beating-heart donors, unprocessed, and implanted within 3 days), the inference can be made that valves with some or major fibroblast viability produced improved clinical long-term durability. The clinical experience at TPCH [3] and the pathologic analyses strongly support this theory. At explantation, TPCH cryopreserved valves are more likely to have leaflets that are thicker and lack perforation compared with refrigerated storage valves, nonviable at implantation. Tissue culture of explanted cryopreserved valves viable at the time of implantation has demonstrated both the presence of viable fibroblasts, although in markedly reduced numbers, and acellular leaflet segments. On these occasions when sufficient cells for chromosomal analysis have been cultured, the cells have always been shown to be of donor origin [8, 9]. Only on one occasion have fibroblasts been cultured from an explanted 4 C stored valve. This valve had been implanted within 7 days of procurement, and this result is consistent with a finding of this study that valves stored at 4 C contain viable cells up to around 7 days. Mochtar and associates [11] have also demonstrated this decrease in leaflet viability with time when valves are stored at 4 C. In summary, the results from this investigation into preimplantation viability support the current protocol for antibiotic sterilization and cryopreservation. Beatingheart donors show a much higher level of initial viability than nonbeating-heart donors and are therefore the preferred source of valves. Retrieval within 24 hours after death provides valves with an initial viability level of around 50% of the total fibroblast population. Although it is not within the TPCH protocol, if valves are to be stored longer than 3 days, we recommend that they be cryopreserved immediately after antibiotic sterilization. Amphotericin B should not be used in conjunction with cryopreservation. The simplicity of the rigid TPCH protocol has stood the test of time over 27 years. Valves retrieved under semisterile conditions within 24 hours require only a lowdose, short-term antibiotic exposure for sterilization. The clinical success of the allograft for aortic valve replacement gives credence to the importance of such a protocol of valve preparation, sterilization, and storage [5]. This coupled with the improved implantation technique of root replacement [18, 19], which minimizes valvular distortion, and a better understanding of the immune response in the young and elderly [20, 21] augurs well for further improvements in the durability of the allograft valve. The clinical results with valves that have been collected and cryopreserved soon after death are similar to those obtained with homovital valves implanted within 3 days [17]. This fact probably highlights the importance of minimal cellular and matrix disturbance before implantation. This may also help explain why the long-term durability of implanted allograft valves, including those stored under refrigerated conditions, cryopreserved and homovital, still cannot mimic that of the aortic valve within the immunosuppressed transplanted heart. This research project was supported by a grant from The Prince Charles Hospital Foundation. We thank Ms Anne Brosnan for her assistance during the development of this research project and Mr Ian Smith, MApplSc (Med Phys), for his analysis of the data.

1038 GALL ET AL Ann Thorac Surg PROCESSING AND ALLOGRAFT VALVE VIABILITY 1998;65:1032 8 References 1. O Brien MF, Stafford EG, Gardner M, et al. The viable cryopreserved allograft aortic valve. J Cardiac Surg 1987; 1(Suppl):153 67. 2. O Brien MF, McGiffin DC, Stafford EG, et al. Allograft aortic valve replacement: long-term comparative clinical analysis of the viable cryopreserved and antibiotic 4 C stored valves. J Cardiac Surg 1991;6(Suppl):534 43. 3. O Brien MF, Stafford EG, Gardner MAH, et al. Allograft aortic valve replacement: long-term follow-up. Ann Thorac Surg 1995;60:S65 70. 4. Grunkemeier GL, Bodnar E. Comparison of structural valve failure among different models of homograft valves. J Heart Valve Dis 1994;3:556 60. 5. Gall K, Smith S, Willmette C, Wong M, O Brien M. Allograft heart valve sterilization: a six-year in-depth analysis of a twenty-five-year experience with low-dose antibiotics. J Thorac Cardiovasc Surg 1995;110:680 7. 6. Watts LK, Duffy P, Field RB, Stafford EG, O Brien MF. Establishment of a viable homograft cardiac valve bank: a rapid method of determining homograft viability. Ann Thorac Surg 1976;21:230 6. 7. Van Der Kamp AWM, Nauta J. Fibroblast function and the maintenance of the aortic valve matrix. Cardiovasc Res 1979; 13:167 72. 8. O Brien MF, Stafford EG, Gardner MAH, Pohlner PG, Mc- Giffin DC. A comparison of aortic valve replacement with viable cryopreserved and fresh allograft valves, with a note on chromosome studies. J Thorac Cardiovasc Surg 1987;94: 812 23. 9. O Brien MF, Johnson N, Stafford EG, et al. A study of the cells in the explanted viable cryopreserved allograft valve. J Cardiac Surg 1988;3(Suppl):279 87. 10. Van Der Kamp AWM, Visser WJ, Van Dongen JM, Nauta J, Galjaard H. Preservation of aortic heart valves with maintenance of cell viability. J Surg Res 1981;30:47 56. 11. Mochtar B, Van Der Kamp AWM, Roza-De Jongh EJM, Nauta J. Cell survival in canine aortic heart valves stored in nutrient medium. Cardiovasc Res 1984;18:497 501. 12. Brockbank KGM, Bank HL. Measurement of postcryopreservation viability. J Cardiac Surg 1987;2(Suppl):145 51. 13. Feng XJ, Van Hove CEJ, Mohan R, Walter PJ, Herman AG. Effects of different antibiotics on the endothelium of the porcine aortic valve. J Heart Valve Dis 1993;2:694 704. 14. McNally RT, Brockbank KGM. The correlation between improved cellular viability and clinical performance in 5,000 cryopreserved human heart valves. ASAIO Trans 1991;37: M355 6. 15. Heacox AE, McNally RT, Brockbank KGM. Factors affecting the viability of cryopreserved allograft heart valves. In: Yankah AC, Hetzer R, Miller DC, Ross DN, Sommerville J, Yacoub MH, eds. Cardiac valve allografts 1962 1987. Darmstadt: Steinkopff Verlag, 1988:37 42. 16. Brockbank KGM. Basic principles of viable tissue preservation. In: Clarke DR, ed. Transplantation techniques and use of cryopreserved allograft cardiac valves and vascular tissues. Boston: Adams Publishing Group, 1989:9 23. 17. Yacoub M, Rasmi NRH, Sundt TM, et al. Fourteen-year experience with homovital homografts for aortic valve replacement. J Thorac Cardiovasc Surg 1995;110:186 94. 18. O Brien MF, Finney RS, Stafford EG, et al. Root replacement for all allograft aortic valves: preferred technique or too radical? Ann Thorac Surg 1995;60:S87 91. 19. Jones EL, Shah VB, Shanewise JS, et al. Should the freehand allograft be abandoned as a reliable alternative for aortic valve replacement? Ann Thorac Surg 1995;59:1397 404. 20. Zhao X-M, Green M, Frazer IH, Hogan P, O Brien MF. Donor-specific immune response after aortic valve allografting in the rat. Ann Thorac Surg 1994;57:1158 63. 21. O Brien MF. Homografts and autografts. In: Baue AE, ed. Glenn s textbook on thoracic and cardiovascular surgery; vol 2. 6th ed. Stamford, CT: Appleton and Lange, 1996:2005 42.