Primary Tissue Valve Degeneration in Glutaraldehvde-Preserved Porcine Biomostheses: Hancock I Vekus Carpentier-Edwards at 4- to 7-Years Follow-up

Transcription

1 Primary Tissue Valve Degeneration in Glutaraldehvde-Preserved Porcine Biomostheses: A Hancock I Vekus Edwards at 4- to 7-Years Follow-up Francisco Nistal, M.D., Edurne Artifiano, M.D., and Ignacio Gallo, M.D. ABSTRACT In the 32-month period between April, 978, and December, 980, 292 patients, divided into two equal groups, received a glutaraldehyde porcine bioprosthesiseither Hancock or Edwards (CE)-as mitral valve substitute. Every patient receiving a mitral porcine xenograft during that time was included in the study. The type of bioprosthesis was always selected by the surgeon and not randomly. Preoperative clinical characteristics, associated surgical procedures, valve implantation sizes, and follow-up data showed no relevant differences between the two groups. There were three instances of primary tissue failure in the Hancock group and six in the CE (linearized rates of 0.49 and 0.97 events percentage of patienvyears, respectively). Mean duration of explanted valves and microscopic findings were similar in both groups. Primary tissue failure was more frequent in patients under 40 years of age in both groups, although differences were not statistically significant. A marginally significant trend toward greater incidence of tissue failure in patients of 40 years of age and older was seen in the CE group when compared with the Hancock group. Actuarial survival of the bioprostheses free from primary tissue failure was 6.5 years of 95? 3% (mean -C standard error) for Hancock and % for CE (p = ). No significant difference in terms of durability has been found between the two most popular glutaraldehyde porcine bioprostheses, although the behavior of the CE in patients older than 40 years should be reassessed in a study with a larger number of patients and a longer follow-up period. Tissue valve prostheses have become widely accepted in the last decade because they are less thrombogenic than mechanical valves and do not necessitate anticoagulant therapy in a large proportion of patients [l-5. We use this type of valve in the majority of our patients because it is difficult to maintain them in reliable anticoagulant treatment. From the Servicios de Cirugia Cardiovascular and Anatomia Patol6gica, Hospital Nacional Marques de Valdecilla, Universidad de Santander, Spain. Accepted for publication Jan 24, 986, Address reprint requests to Dr. Gallo, Servicio de Cirugia Cardiovascular, Hospital Nacional Valdecilla, Santander, Spain. The durability of these prostheses remains a problem, especially in young adults and children [6-9. Several reports on primary valve tissue failure with large numbers of patients have been published, mainly on the Hancock porcine bioprosthesis [l, 2, 4, 5, 0-3. A few papers also have dealt with tissue degeneration in the Edwards porcine heterograft [4-7. This study compares the incidence of primary failure in two parallel series of patients who received as mitral valve substitute either a Hancock or a Edwards (CE) porcine bioprosthesis. Material and Methods Between April, 978, and December, 980, a total of 447 patients received in our institution a glutaraldehydepreserved porcine bioprosthesis as valve substitute in mitral or aortic position (or both). All patients with a mitral xenograft, isolated or not, were reviewed. Two groups of 46 patients received either a Hancock or a CE valve in the mitral position. Patients not discharged from hospital were excluded from the study (9 in the Hancock and 4 in the CE groups). No special selection or randomization was used to select the type of bioprosthesis for each individual patient; the surgeon s preference was the determinant. Each surgeon tended to implant the same type of valve, with few cases of crossover. Preoperative clinical data for both groups are listed in Table according to the type of prosthesis. Standardized operative techniques included cold ischemic arrest with topical hypothermia for myocardial protection, moderate systemic hypothermia, and uninterrupted aortic cross-clamping. Cold potassium cardioplegia was used in all patients. Handling and rinsing of the bioprosthesis during implantation were also standardized. Associated surgical procedures were performed in 84 patients of the CE group and in 8 of the Hancock group (Table 2). Table 3 shows the distribution of implantation sizes of the valves and mean values. Primary tissue valve failure was established at hemodynamic study, reoperation, or autopsy by detection of valve incompetence or stenosis due to cuspal tears, perforations, calcific deposits, or stiffening of the leaflets, in the absence of an active infective process. The definition of primary failure includes not only reoperation but also death with either autopsy or hemodynamic confirmation of valve failure. Living patients with hemodynamic or clinical diagnosis of valve dysfunction were not included 568 Ann Thorac Surg 42: , Nov 986

2 ~~ 569 Nistal, ArtiAano, and Gallo: Degeneration in Porcine Mitral Valve Bioprostheses Table. Preoperative Clinical Characteristics Characteristic (n = 46) (n = 46) p Value Age Mean 2 SD Patients < 40 Patients > 40 Range Sex Male Female Functional class (NYHA) I I I IV Mean 2 SD Type of lesion Stenosis Insufficiency Mixed lesion Prosthetic dysfunction Etiology Congenital SBE Rheumatic Ischemic Others 49 f 2 29 (20%) 7 (80%) (5%) 7 (49%) 94 (64%) 29 (20%) (3%) 35 (24%) 9 (62%) (%) 6 (4%) 22 (83%) 3 (2%) 4 (0%) (6%) 23 (84%) (40%) 88 (60%) 88 (60%) 35 (24%) (6%) 95 (65%) 5 (3%) 4 (3%) 2 (83%) 0 20 (4%) p <.05 p <.05 SD = standard deviation; NYHA = New York Heart Association; SBE = subacute bacterial endocarditis. Table 3. Implantation Valve Sizes Hancock (n = 46) Edwards (n = 46) % Failures % Failures Size % Valves per Size % Valves per Size Mean size 2 SD f.8 () Values are not significantly different from their homologues. SD = standard deviation; = not significant. Table 4. allow-up Data Variable (n = 26) (n = 32) p Value Mean follow-up f 7 f SD (months) Range (months) Patients lost 0 Total follow-up period (patient/years) SD = standard deviation; = not significant Table 2. Associated Surgical Procedures Procedure (n = 46) (n = 46) p Value Tricuspid valvuloplasty 4 (28%) 50 (34%) Tricuspid valve replace- (%) (%) ment Aortic valve replacement 55 (38%) 42 (29%) Double valve replace- 0 ment CABG 3(2%) No associated surgery 65 (44%) 62 (42%) CABG = coronary artery bypass graft; = not significant. because they are waiting for confirmation. All surviving patients were routinely evaluated at this center. Those who failed to report to the hospital were contacted by letter. Follow-up was conducted until December, 984. Explanted valves were examined, described, and photographed macroscopically, processed by standard techniques for light microscopy, and stained by hematoxilid eosin. Statistical analysis was performed by the Student s t test, the chi square, and the 2 x 2 contingency test for unhomogeneity evaluation of the preoperative data. The incidence of primary tissue valve failure was determined by actuarial analysis and linearized occurrence rates. Actuarial curves of tissue failure were calculated by the &plan and Meier method. A standard error of the mean is noted for each point in the curve. Comparison of the actuarial curves was performed using normal distribution. Results There were 27 hospital survivors (of whom was lost to follow-up) in the Hancock group and 32 in the CE group. These 258 patients were followed for to 79 months (Table 4). The mean follow-up periods were 59 % 20 (range -79) for the Hancock group and (range 3-79) months for the CE group. Cumulative follow-up periods were 69 and 65 patient/years for Hancock and CE groups, respectively. A total of 30 patients (7 in the Hancock and 3 in the CE groups) died late postoperatively. Among the late deaths in the Hancock group, 3 patients (57, 48, and 55 years of age) died of unknown cause and without autopsy at 24, 6, and 43 months after implantation. Two of these patients had postoperative symptoms: had respiratory insufficiency, the other exhibited a partially recovered cerebral embolism. In the CE group a woman

3 570 The Annals of Thoracic Surgery Vol 42 No 5 November 986 Table 5. Primary Tissue Failure Variable (n = 3) (n = 6) Age Sex M M F M F M M M M FOIIOW-UP (months) Valve size Tear + + Calcification + + Tear and + + +a + + calcification "Carpentier aortic unaltered + = present. who was 7 years old at operation died in the fortyseventh postoperative month; the cause of death was unknown, and no necropsic information was available. Mean implantation valve size was similar in the two series. There were no differences in the distribution of valve sizes between the groups, and we did not observe significant clustering of failing valves in either small or large sizes in either group (Table 3). Three Hancock valves had to be explanted because of primary tissue failure between 52 and 62 months. In the CE group, 6 valves were explanted for the same reason between 3 and 74 months. Mean duration of the valves before removal was not significantly different between the groups (58 * 5 months for Hancock and months for CE). Linearized rates of primary valve failure were 0.49 failing valves per 00 patient/years and 0.97 per 00 patienuyears for the Hancock and CE groups, respectively. In patients whose valve later degenerated the mean age at the time of implantation was 30 years (range 25-38) for the Hancock group and 49 (range 36-57) for the CE group. Actuarial rate of freedom from primary tissue valve failure was 95.3 * 2.7% for the Hancock patients and 84.4 & 8.6% for CE patients at 79 months of follow-up (Table 5 and Fig ). ; ; y a00 z I: y 0 ICE. I*.O,.".l I,'.0*.., 0! I I Fig 2. Actuarial comparison of age cohorts from the Hancock and Edwards (CE) groups. Comparison of patients younger and older than 40 years in CE recipients yields no significant diference, while in the Hancock group a marginally significant difference is observed (p =.06). Comparison of younger patientsfrom both groups shows no significant difference but between patients older than 40 years there is a marginally significant difference (p =.06). 2 I A - I $ 0 P (dce n 9 CE) 'q d PI:., :,", 2, ::, :, :I, 5% I Fig 3. Actuarial analysis of the influence of sex on the appearance of primary tissue failure. Among Hancock recipients no difference is observed between mule and female patients, but male patients in the Edwards (CE) group display lower actuarial rates at 72 and 79 months than found in females (differences marginally significant: p =,078 at 72 months, p =.056 at 79 months). t I - HancOcU 2... carpentter. Edwards 80.- ' 7h 87!I4 08 ( I 'I ' 4.4?.8% I I Fig I. Actuarial analysis of the incidence of primary tissue failure in Hancock and Edwards porcine bioprostheses at 7-years' follow-up. Values are noted plus or minus standard error. Comparison of age cohorts of both groups using actuarial analysis (Fig 2) showed no difference in the failure rate among CE recipients younger or older than 40 years (p =.8). A marginally significant difference (p =.06) was found in the Hancock group between both age cohorts. Primary tissue failure rate was not significantly different between the two groups in patients less than 40 years of age ( p = 3). However, a marginally significant difference was observed between both groups of patients older than 40 years at 72 and 79 months of followup (p =.06). Sex (Fig 3) had no apparent influence on the failure rate in the Hancock group, but among CE recipients a striking difference was observed between male and female patients; however, the difference was only mar-

4 ~ 57 Nistal, Artifiano, and Gallo: Degeneration in Porcine Mitral Valve Bioprostheses g 90 J 9s? 3 S m c t 4 s,i 33 or < 33 a.i f I7.\ :.O <, a., : u I3 I<,,, 8. 3, Sll I I Pas- Months Fig 4. Actuarial analysis of the influence of implantation valve size on the appearance of primary tissue failure. Valves size 33 or smaller behave similarly in the Hancock group, but size 33 valves seem to degenerate at a greater rate than smaller valves in the Edwards group; however, the difference is not statistically significant (p =.3). ginally significant (p =.078 at 72 months; p =.056 at 79 months). A similar phenomenon was also observed when valve size cohorts from both groups were compared (Fig 4). Size 33 CE valves seemed to degenerate at a greater rate than the rest of valves from this group or any of the valves in the Hancock group, but the difference was not statistically significant ( p =.3). Macroscopic examination showed tear and calcification in 2 cases and isolated tear in in the Hancock group. Three CE valves had tear and calcification; 2 displayed isolated calcification and showed tear alone. Microscopically no substantial differences were observed between the groups. All explanted valves had calcific deposits to some extent, and collagen degeneration was usually associated. Comment The characteristics of our patient population have strongly conditioned our policy in valvular surgery. Because our patients live far from the hospital and cannot be maintained under anticoagulation, we need to take a conservative attitude, and when valve replacement is necessary we use bioprostheses. Since 974 we have routinely implanted glutaraldehyde porcine bioprostheses; durability was, even then, an important matter to be considered by the surgeon. Although many patients would have to be reoperated with time, we thought most of them would probably benefit from a period practically free of concern. Initially we only implanted Hancock porcine bioprostheses but since April, 978, we have also used the CE glutaraldehyde-preserved porcine xenograft. In the period of time between April, 978, and December, 980, a total of 5 porcine valves were implanted in mitral or aortic position in 447 patients in our hospital. A mitral replacement, isolated or not, was performed in 292 patients, half of them receiving a Hancock xenograft and half a CE. When hospital deaths and patients lost to follow-up have been eliminated, 26 survivors in the Hancock group and 32 in the CE remained for comparison. Preoperative data show no statistically significant differences in age (mean values, patients younger than 40, and range), functional class, type of lesion (stenosis, insufficiency, mixed lesions or prosthetic dysfunction), or etiology (Table ). Comparison of sex distribution showed a significantly larger proportion of women in the CE than in the Hancock group. Distribution of associated surgical procedures and implantation valve sizes (Tables 2 and 3) is also similar, and there is no difference between the groups in mean and cumulative follow-up times. Hospital mortality was not significantly different between the groups; on the other hand, we have seen no published evidence that this variable could influence the primary tissue failure rate. Late death of unknown cause without a necropsic study accounted for 3 cases in the Hancock group and in the CE group. These patients could represent occult primary tissue failure events, but all of them were older than 40 years; considering the actuarial rates for these cohorts (Fig 2), this hypothesis seems unlikely. Furthermore, because 2 of the Hancock patients dying late postoperatively had had permanent postoperative noncardiac symptoms, there may be a different explanation for their decease. The strict comparison of mortality, as well as other phenomena like thromboembolism or endocarditis, goes beyond the scope of this paper. Incidence of primary tissue degeneration requiring reoperation or proved at autopsy seems to be higher in the CE than in the Hancock group, as judged by the linearized rates. However, examination of actuarial curves shows no statistical difference, although the CE series has a consistently lower rate of freedom from primary failure starting in the second postoperative year. One should be cautious when interpreting these results, however, because the differences are rather small; longer follow-up times would be required to magnify them. Age of the patients at operation, in our own experience [5] and that of others [l, 6-0], acts as a factor that predisposes to primary tissue failure. We have not been able to demonstrate this tendency (Fig 2), but we did observe a trend in that direction, especially in the Hancock group. In both groups, actuarial rates of failure among patients younger than 40 years were not significantly different. Interestingly, a marginally significant difference among patients older than 40 years was observed. This difference, however, is not very large and could disappear with longer follow-up periods. Sex has been reported to have no influence on the prevalence of primary tissue failure [lo]. Our observations (Fig 3) agree in general with this statement; however, a higher incidence (but not statistically significant) of primary failure was encountered among males in the CE group. This finding should be reassessed in the future. A similar phenomenon seems to occur with size 33 valves in the CE group (Fig 4), but the difference is in this case clearly nonsignificant ( p =.3).

5 572 The Annals of Thoracic Surgery Vol 42 No 5 November 986 Comparison of the long-term performance of two heart valve prostheses is difficult because many factors involved (for example, patient population, different surgical techniques) vary among surgical teams. That is the reason why the most accurate results would be obtained through a single hospital study, ideally randomized, with large and similar patient groups, followed for sufficient time (within fixed time limits). We believe that our study fulfills most of these criteria and therefore should constitute an adequate approach to this question. Durability still remains the major problem of tissue valves and represents a serious concern for surgeons. Primary tissue valve failure of the Hancock porcine bioprosthesis has been extensively reported [I, 2, 4-3]. However, there is not much information available about the intermediate and long-term behavior of the CE valve; although many reports on porcine bioprostheses describe the use of both types of valves, no specific comparison between them has been accomplished so far. Jamieson and associates [4] in a recent report show a linearized rate of primary failure for isolated mitral replacement patients of event per 00 patient/years and of.6 per 00 patient/years for multiple valve replacement, provided that all failing valves in this group were in the mitral position. They reported an overall linearized incidence of mitral primary tissue failure of. events per 00 patient/years, but this figure is probably slightly underestimated because the authors do not state whether every patient in the multiple valve replacement group had the mitral valve replaced. Our rates are thus comparable with those of Jamieson and associates. Deloche and co-workers (7, reviewing the experience of the Broussais Hospital in Paris, found no differences in primary tissue valve degeneration between two groups of 88 and 04 patients who had mitral valves replaced by Hancock or CE xenografts between 974 and 978; there was a maximum follow-up time of 6 years. Gross histologic appearance was also similar in both groups. These authors, however, do not see any primary failure before the fourth postoperative year, while in our experience the phenomenon appears after the second postoperative year in the CE series. In summary, it can be stated that durability of the CE porcine bioprosthesis is very acceptable in the intermediate term, and we see no clear difference in durability between it and the Hancock porcine xenograft. We gratefully acknowledge Dr. Roman Valiente for statistical advice and Ana Ruiz for secretarial assistance. References. Oyer PE, Miller DC, Stinson EB, et al: Clinical durability of the Hancock porcine bioprosthetic valve. J Thorac Cardiovasc Surg 80:824, Cohn LH, Mudge GH, Pratter F, et al: Five- to eight-year follow-up of patients undergoing porcine heart-valve replacement. N Engl J Med 304:258, Ionescu MI, Smith DR, Hassan SS, et al: Clinical durability of the pericardial xenograft valve: ten years experience with mitral valve replacement. Ann Thorac Surg 34:265, Gallo I, Ruiz 8, Durdn CMG. Five- to eight-year follow-up of patients with the Hancock cardiac bioprosthesis. J Thorac Cardiovasc Surg 86:897, Gallo I, Ruiz 8, Nistal F, et al: Degeneration in porcine bioprosthetic cardiac valves: incidence of primary tissue failure among 938 bioprostheses at risk. Am J Cardiol 53:06, Curcio CA, Commerford PJ, Rose AG, et al: Calcification of glutaraldehyde-preserved porcine xenografts in young patients. J Thorac Cardiovasc Surg 8:62, Geha AS, Laks H, Stansel HC, et al: Late failure of porcine valve heterografts in children. J Thorac Cardiovasc Surg 78:35, Kutsche LM, Oyer P, Shumway N, et al: An important complication of Hancock mitral valve replacement in children. Circulation 6O:Suppl :98, Sanders SP, Levy RJ, Freed MD, et al: Use of Hancock porcine xenografts in children and adolescents. Am J Cardiol 46:429, Magilligan DJ Jr, Lewis JW Jr, Jara FM, et al: Spontaneous degeneration of the porcine bioprosthetic valve. Ann Thorac Surg 30:259, 980. Magdligan DJ, Lewis JW, Heinzerling RH, et al: Fate of a second porcine bioprosthetic valve. J Thorac Cardiovasc Surg 85362, Schoen FJ, Collins JJ, Cohn LH: Long-term failure rate and morphologic correlations in porcine bioprosthetic valves. Am J Cardiol 5:957, Bolooki H, Mallon S, Kaiser G, et al: Failure of Hancock xenograft valve: importance of valve position (4- to 9-year follow-up). Ann Thorac Surg 36:246, Jamieson WRE, Pelletier LC, Janusz MT, et al: Five-year evaluation of the Edwards porcine bioprosthesis. J Thorac Cardiovasc Surg 88:324, Janusz MT, Jamieson WRE, Gerein AN, et al: Long-term follow-up of patients with porcine cardiac valve prostheses. Can J Surg 26:60, Gallo, Ruiz 8, Durdn CMG. Clinical experience with the Edwards porcine bioprosthesis: short-term results (from 2 to 4.5 years). Thorac Cardiovasc Surgeon 3:277, Deloche A, Perier P, Bourezak H, et al: A 4-year experience with valvular bioprostheses: valve survival and patient survival. In Cohn LH and Galluci V (eds): Cardiac Bioprostheses. New York, Yorke, 982, pp 25-34