A prospective evaluation of the Bjbrk-Shiley, Hancock, and Carpentier-Edwards heart valve prostheses

Transcription

1 THERAPY AND PREVENTION VALVE REPLACEMENT A prospective evaluation of the Bjbrk-Shiley, Hancock, and Carpentier-Edwards heart valve prostheses PETER BLOOMFIELD, M.R.C.P., ARTHUR H. KITCHIN, F.R.C.P.. DAVID J. WHEATLEY, M.D., F.R.C.S.,* PHILIP R. WALBAUM, F.R.C.S., WALTER LUTZ, M.SC., AND HUGH C. MILLER, F.R.C.P. ABSTRACT From 1975 to 1979, 540 patients undergoing valve replacement were entered into a randomized trial and received either a Bjdrk-Shiley (273 patients) or a porcine heterograft prosthesis (initially a Hancock valve [107 patients] and later a Carpentier-Edwards prosthesis [ 160 patients]). Two hundred and sixty-two patients required mitral valve replacement, 210 required aortic valve replacement, 60 required mitral and aortic valve replacement, and eight also required associated tricuspid valve replacement (six mitral valve replacement; two mitral plus aortic valve replacement). Analysis of 34 preoperative and operative variables showed the treatment groups to be well randomized. In-hospital mortality was not significantly different among patients receiving the three prostheses for aortic valve replacement (7.6% overall) and mitral plus aortic valve replacement (10% overall), but there was a higher in-hospital mortality for patients undergoing mitral valve replacement with the Carpentier- Edwards prosthesis (15.5% compared with 8.8% overall; p =.03). This difference could not be explained on the basis of any preoperative or operative variable. Median follow-up was 5.6 (range 2.8 to 8.3) years. Actuarial survival after mitral valve replacement was % at 7 years, that after aortic valve replacement was 69.6 ± 9.6% at 7 years, and that after mitral plus aortic valve replacement was 62.5 ± 20.0% at 7 years. There was no significant difference in actuarial survival of patients receiving the three prostheses within the mitral, aortic, and mitral plus aortic valve replacement groups, nor was there a difference when these groups were amalgamated. Thirty-seven patients required reoperation for valve failure (15 with Bjbrk-Shiley, 12 with Hancock, and 10 with Carpentier-Edwards valves; p = NS) and 11 died at reoperation (four with Bjdrk-Shiley, four with Hancock, and three with Carpentier-Edwards valves; overall operative mortality 29.7%). Up to 7 years after surgery, there was no significant difference in the incidence of thromboembolism in patients with the different prostheses undergoing mitral or aortic valve replacement. There were too few patients undergoing mitral plus aortic valve replacement for meaningful comparison. There was no significant beneficial effect of anticoagulants in patients undergoing mitral or aortic valve replacement with porcine prostheses, but patients were not randomly allocated to anticoagulant treatment. All patients with Bjdrk-Shiley prostheses received anticoagulants. Multivariate analysis of factors associated with embolism identified atrial fibrillation with mitral valve replacement (p <.001) and age less than 65 years (p <.01) and a rheumatic cause of valvular disease (p <.01) with aortic valve replacement. The risks of anticoagulation were low, with an overall incidence of complications of approximately one per 100 years treatment. To date no significant advantage of any of the three prostheses has been observed, but further follow-up is necessary because important differences may yet emerge. Circulation 73, No. 6, , From the Departments of Cardiology and Cardiac Surgery, Royal IN THE EARLY 1970s glutaraldehyde-treated porcine Infirmary, the Department of Cardiology, Western General Hospital, valves mounted on a wire stent were introduced for and the Medical Computing and Statistics Unit, University of Edin- cardiac valve replacement and appeared in initial burgh, Edinburgh, Scotland. Supported by a grant from the Scottish Home and Health Department. short-term studies to be durable and to carry a reduced Address for correspondence: Dr. Hugh C. Miller, Department of risk of thromboembolism compared with mechanical Cardiology, Royal Infirmary, Edinburgh, Scotland. ls e Received Oct. 17, 1985; revision accepted Jan. 23, prostheses. Presented in part at the 58th Scientific Sessions of the American Heart However, in 1975 the long-term results with porcine Association, Washington, D.C., November heterograft prostheses had not been evaluated in a pro- *Current address: Department of Cardiac Surgery, Glasgow Royal Infirmary, Glasgow, Scotland. spective trial, so we commenced a prospective ran- Vol. 73, No. 6, June

2 BLOOMFIELD et al. domized trial comparing the porcine heterograft with the Bjork-Shiley tilting-disk prosthesis. Our objectives were to compare the two types of prostheses for operative risk, initial and late clinical results, durability, incidence of thromboembolism, and incidence of complications related to anticoagulation. The last of the study participants entered the trial in August We now report for the first time the results in the 540 patients who were entered into the trial and have now been followed for up to 8 years Patients and methods All patients with valvular heart disease treated in the two main cardiology referral hospitals in Edinburgh during the study period were seen by six consultant cardiologists, all of whom participated in the study. All patients requiring heart valve replacement were considered for the study. Patients were excluded if long-term anticoagulation was contraindicated (as in the case of young women anticipating pregnancy) or if they were reluctant to take anticoagulants over a long term. Patients already entered in the study who required valve rereplacement were not reentered. Three of the five consultant cardiac surgeons participated. A comprehensive clinical profile, including details of history, physical examination, the 12-lead electrocardiogram (ECG), the chest x-ray, and cardiac catheterization, was recorded for each patient. When coronary artery disease was suspected clinically coronary angiography was performed. Significant coronary artery disease was said to be present if there was a 70% or greater lesion in one of the three major coronary arteries. A total of 34 preoperative clinical variables was recorded for each patient. Operative details and method of randomization. The Bjork-Shiley valve used was the 60 degree spherical disk valve (model ABP in the aortic position and model MBRP with a supra-annular sewing ring in the mitral position) and the Hancock valve was the stabilized, glutaraldehyde-processed bioprosthesis (model 242 in the aortic and model 342 in the mitral position). The Carpentier-Edwards valve used was the glutaraldehyde-preserved, flexible, support-mounted porcine bioprosthesis (model 2625 in the aortic and model 6625 in the mitral position). At the time of surgery if there were no technical reasons, e.g., small aortic root, precluding randomization, an envelope was opened indicating the prosthesis allocated to the patient. Randomization was performed in groups of 10 to ensure even distribution of patients among surgeons and treatment groups. The operative technique was standardized for all patients; cold cardioplegia was used routinely. Data recorded included total aortic cross-clamp time, total bypass time, the amount of blood given, and the need for postoperative inotropic support. Postoperative anticoagulation policy. All patients received warfarin for 2 months after surgery, and patients with Bjork- Shiley prostheses received the drug throughout the study period. Patients with porcine heterograft prostheses subsequently received warfarin at the discretion of their own cardiologists. Collection of follow-up data. At each follow-up visit functional class, drug therapy, physical findings, chest x-ray, and the ECG were documented. The occurrence of any embolic event, defined as an episode of transient or sustained neurologic deficit, amaurosis fugax, limb ischemia of sudden onset, or any other clinical event that could possibly be attributed to an embolism, was recorded. The presence of any residual physical signs after embolism or any complication of anticoagulation was recorded on a standard data sheet for computer-based storage and analysis. Valve failure was defined as dysfunction of the prosthesis resulting in regurgitation or stenosis followed by replacement of the prosthesis or death, infective endocarditis, or multiple thromboembolic events necessitating replacement of prosthesis. Deaths. Operation-related deaths were defined as those occurring before hospital discharge. Later deaths were classified as "cardiac," "noncardiac," or "possibly cardiac" on the basis of stated clinical data or information gathered at autopsy. Death due to systemic embolism or a complication of anticoagulation was classified as "cardiac." If the cause of death was attributable to the prosthesis it was recorded as such. Deaths of those patients undergoing valve rereplacement who died at reoperation were recorded as prosthesis-related deaths, whereas patients undergoing successful reoperation were thereafter withdrawn from further analysis. Statistical analysis. Statistical analysis employed t test analysis for the continuous variables of age and length of follow-up and the Fischer exact test and Pearson chi-square analysis for categorical analysis. The rates of survival, valve failure, thromboembolism, and anticoagulant-related hemorrhage were expressed in the form of standard life tables, as previously described.4 The generalized Wilcoxon and Mantel-Haenzel tests were used to test for differences in survival. The chi-squared test for symmetry and the Pearson chi-squared test for contingency tables were used to analyze differences in functional class. Continuous data are presented as mean + SD, and actuarial probability estimates as mean ± 2SEM. Results Patients. From December 1975 to August patients were referred by the six participating cardiologists for valve replacement by the three participating TABLE 1 Baseline characteristics of patients undergoing mitral (MVR), aortic (AVR), and combined (MVR + AVR) valve replacement MVR MVR AVR AVR (n = 262) (n = 210) (n = 60) Age <50 years > Female sex NYHA class I and II NYHA class III and IV NYHA class IV at some time Cause of disease Rheumatic Nonrheumatic Associated CAD present Prior mitral valvotomy Previous valve replacement Stenosis Regurgitation Stenosis plus regurgitation All values are percentages. CAD = coronary artery disease. CIRCULATION

3 THERAPY AND PREVENTION-VALVE REPLACEMENT surgeons. Seventy-two patients were excluded and 540 patients were randomly assigned to a treatment group. During this period a total of 811 patients underwent valve replacement in Edinburgh Royal Infirmary. Those not included in the trial were referred directly for surgery from other centers or were operated on by surgeons not participating in the trial. Of the 540 randomized patients 262 required mitral valve replacement, 210 aortic valve replacement, and 60 combined mitral and aortic valve replacements. Eight patients underwent tricuspid valve replacement, six in addition to mitral valve replacement and two in addition to mitral plus aortic valve replacement. These patients were excluded from further analysis because of the small numbers. Fourteen patients underwent concomitant coronary artery bypass surgery; they were evenly distributed among the treatment groups and their data were included in the analysis. The baseline characteristics of patients undergoing aortic, mitral, and mitral plus aortic valve replacement are given in table 1. Hemodynamic data on all patients have been summarized and are available in table form on request.* Results of randomization. From January 1977, because of a significant cost advantage, patients randomized to treatment with a porcine heterograft prosthesis received a Carpentier-Edwards (160 patients) instead of a Hancock prosthesis (107 patients). All patients randomly assigned to receive a mechanical prosthesis received the Bjdrk-Shiley valve (273 patients). To ensure that time of operation did not result in a separate bias, results were compared for patients receiving the Bjdrk-Shiley valve in the first period of the trial and those undergoing surgery in the second period of the study. Results were similar for both periods and findings are therefore presented as a simple comparison between the three types of prostheses. No statistical difference with respect to values for any of 34 preoperative variables was found between patients receiving Bjdrk-Shiley, Hancock, and Carpentier-Edwards prostheses within the groups undergoing aortic, mitral, or mitral plus aortic valve replacement. Patients not undergoing catheterization were randomly distributed among the treatment groups. Also, the three surgeons performing the procedures placed equal numbers of each type of valve. *See National Auxiliary Publications Service document No for 2 pages of supplementary material. Order from NAPS c/o Microfiche Publications, P.O. Box 3513, Grand Central Station, New York, NY Remit in advance U.S. funds only, $7.75 for photocopies or $4.00 for microfiche. Outside the U.S. and Canada add postage of $4.50 for the first 20 pages and $1.00 for each additional page; $1.50 for microfiche postage. Vol. 73, No. 6, June 1986 TABLE 2 In-hospital mortality for patients undergoing aortic valve replacement by type of prosthesis within three age groups (<50, 50 to 65, and >65 years) Age group < >65 All ages Bjork-Shiley valve Death Total % Hancock valve Death Total So Carpentier-Edwards valve Death Total % A total of 16 of 210 patients (7.6%) died in-hospital. There was no significant difference between the groups. In-hospital mortality. In-hospital mortality for patients undergoing aortic and mitral valve replacement is given for each prosthesis within three age groups (less than 50 years, 50 to 65 years, and greater than 65 years old) in tables 2 and 3. In-hospital mortality was not significantly different for patients with the three prostheses undergoing aortic valve replacement (table 2) and mitral plus aortic valve replacement,* but the number of patients in the latter group was small. Among those undergoing mitral valve replacement, in-hospital mortality was greater for patients receiving the Carpentier-Edwards prosthesis (15.5%) than for all patients as a group (8.8%) (p =.03; table 3). This difference was apparent in all three age groups and could not be explained on the basis of any preoperative variable or after careful analysis of operative and postoperative details. No special problems with the prosthesis were identified by the surgeons. Deaths were due to a variety of causes, including low-output state, myocardial infarction, coagulation disorder, cerebral and peripheral embolism, ruptured abdominal aortic aneurysm, and in one patient rupture of the ventricular wall by the stent of the prosthesis. Follow-up after discharge from the hospital. Of the 487 patients leaving the hospital, 445 have been followed up by their referring cardiologist or local hospital- *See National Auxiliary Publications Service document No for 2 pages of supplementary material. Order from NAPS c/o Microfiche Publications, P.O. Box 3513, Grand Central Station, New York, NY Remit in advance U.S. funds only, $7.75 for photocopies or $4.00 for microfiche. Outside the U.S. and Canada add postage of $4.50 for the first 20 pages and $1.00 for each additional page; $1.50 for microfiche postage. 1215

4 BLOOMFIELD et al. TABLE 3 In-hospital mortality for patients undergoing mitral valve replacement by type of prosthesis within three age groups (<50, 50 to 65, and >65 years) Age group < >65 All ages Bjork-Shiley valve Death Total % Hancock valve Death Total % Carpentier-Edwards valve Death Total % A total of 23 of 262 patients (8.8%) died in-hospital. p =.03 for differences between Bjork-Shiley and Carpentier-Edwards and Hancock and Carpentier-Edwards valves. based physician, 36 responded to a questionnaire, and six have been lost to follow-up. The median duration of follow-up has been 5.6 years, with a range of 2.8 to 8.3 years. Clinical results Functional class. Comparison of New York Heart Association class of patients on optimal medical treatment before operation, at the first postoperative visit, and at the latest follow-up visit showed a highly significant improvement in functional class after surgery for all patients, and no significant change at later follow-up (table 4). There was no significant difference between the patients with the three prostheses in any of the groups. In addition, there was no significant difference in the pattern of improvement for individual patients in the different prosthesis groups. Late postoperative deaths. There have been 113 late postoperative deaths (20.9%) during the trial. The causes of death (cardiac, possibly cardiac, noncardiac, or not known) are categorized in table 5. The cardiac deaths were further divided into those that could and TABLE 4 Mean NYHA class (±+ SD) of patients on optimal medical treatment before surgery, at first postoperative visit, and at latest follow-up Preoperative Postoperative Class at latest class class follow-up AVR ± MVR 2.66± MVR + AVR 2.70± ±0.83 AVR = aortic valve replacement; MVR = mitral valve replacement. could not be directly attributed to the prosthesis, but since autopsy was performed on only a proportion of patients who died, this information is incomplete. Most cases of possible cardiac death were sudden and information from an autopsy was not usually available. Reoperation for valve failure. Details of the 37 patients who underwent valve rereplacement are listed in table 6. There was no significant difference in the need for valve rereplacement in patients in the three prosthesis groups. In eight patients in the Bjork-Shiley group thrombotic obstruction of the valve was found at reoperation; in two of these patients anticoagulant control was known to be poor at the time. In five patients in the Hancock group and six patients in the Carpentier-Edwards group failure was due to degeneration of the cusps of the prosthesis. This occurred at various times after operation. In some patients the cusps had calcified and in others the leaflets had ruptured, thus making the prosthesis regurgitant. In all 12 patients with periprosthetic leak the prosthesis was found to be intact at reoperation. Bacterial endocarditis. Bacterial endocarditis was the reason for reoperation in four patients, one with a Bjork-Shiley, two with a Hancock, and one with a Carpentier-Edwards valve (table 6). Three of these patients died at surgery. Another four patients (two with a Bjork-Shiley, one with a Hancock, and one with TABLE 5 Causes of late postoperative deaths Cardiac death Prosthesis related Embolism Complication of anticoagulation Thrombosed valve Valve dysfunction Endocarditis Reoperation Total prosthesis related Autopsy performed (including reoperation) Total cardiac deaths Possible cardiac death Noncardiac death Not known Total Prosthetic valve used a Carpentier-Edwards valve) died as a result of endocarditis. Two patients with Bjork-Shiley aortic pros- Carpentier- Bjork-Shiley Hancock Edwards (n = 266) (n = 107) (n = 159) CIRCULATION

5 THERAPY AND PREVENTION-VALVE REPLACEMENT TABLE 6 Data on reoperation for valve failure Reason for reoperation Throm- MVR botic Peri- + obstruc- Leaflet prosthetic Prosthesis AVR MVR AVR Total Deaths tion failure leak SBE Bjork-Shiley (n = 266) Hancock (n = 107) 4A Carpentier (n = 159) 1 6B Total (n 532) AVR aortic valve replacement; MVR = mitral valve replacement; SBE = subacute bacterial endocarditis. ANineteen millimeter AVR removed due to persisting gradient. BDetails of one MVR reoperation not available. theses were successfully treated with antibiotics for bacterial endocarditis. Survival. There was no difference between results of analysis of survival in the categories studied when the noncardiac deaths were included and when they were excluded and therefore only the results for total survival are presented. All patients. Analysis of actuarial survival for all patients in the study showed no significant difference between those with the three different prostheses up to 7.0 years of follow-up (figure 1). Actuarial survival for all patients was 63.2% ± 8.7% at 7.0 years. Aortic valve replacement. Analysis of actuarial survival for patients undergoing aortic valve replacement showed no significant difference between those with B0-70- OX 60- Survival '\ ACTUARIAL SURVIVAL ALL VALVES (mean 5.6, range 3 -Byears ) 1~~~~~~~~~~ Hosp 12 2 Years Bjork Shiley n = (266) Hancock n= (107) Carpentier Edwards (n = 159) 3 4 After Implantation a FIGURE 1. Actuarial survival curves, including in-hospital mortality (Hosp), for all patients undergoing mitral, aortic, and mitral plus aortic valve replacement. The numbers below the figure indicate the patients at risk during follow-up. Vol. 73, No. 6, June the three different prostheses (figure 2). Actuarial survival for all patients undergoing aortic valve replacement was % at 7.0 years. Mitral valve replacement. Analysis of actuarial survival for patients undergoing mitral valve replacement showed no significant difference between those with the three different prostheses (figure 3). Actuarial survival for all patients undergoing mitral valve replacement was 56.7 ± 7.0% at 7.0 years. In-hospital mortality for patients with the Carpentier-Edwards prosthesis was higher than that for patients with the other two prostheses, but there was no significant difference in long-term survival between those in the three prosthesis groups whether or not in-hospital mortality was included in the analysis. Aortic and mitral valve replacement. There was no significant difference in actuarial survival between pa o Survival Hosp 1/2 ACTUARIAL SURVIVAL AORTIC VALVE REPLACEMENT Bjark Shiley (n=108) Hancock (n=46) Carpentier Edwards (n=56) Years After Implantation , : FIGURE 2. Actuarial survival curves for patients undergoing aortic valve replacement with each type of prosthesis used

6 BLOOMFIELD et al. 100* 90 o60 Survival 50 ACTUARIAL SURVIVAL MITRAL VALVE R FIGURE 3. Actuarial survival curves for patients valve replacement with each type of prosthesis us REPLACEMENT patients had more than one embolism. Eight patients with Bjork-Shiley prostheses (six undergoing mitral, one undergoing aortic, and one undergoing mitral plus 80 aortic valve replacement) had a total of 20 embolic 70- events. Two patients with Hancock prostheses (both undergoing aortic valve replacement) had four embolic events and two patients with Carpentier-Edwards Bjork Shiley (n=129) Hancock (n=49) prostheses (one receiving mitral and one mitral plus Carpentier Edwards(n= 84) 40- aortic valve replacement) had four embolic events. The incidence of thromboembolism was analyzed for 30 each prosthesis within the group undergoing aortic and 20 that undergoing mitral valve replacement. Too few patients underwent combined aortic and mitral valve 10 replacement to allow a meaningful comparison. Hosp Aortic valve replacement. There were 18 embolic Years After Implantation events in patients undergoing aortic valve replacement ) with no fatalities. In five patients (four receiving a ' is Hancock, and one receiving a Carpentier-Edwards valve) the embolism was a major event and residual unde m signs were present at subsequent follow-up (table 7). ed.sinweepeetasusqetflo-p(al7) In figure 4 the survival curves for patients alive and tients in the three prosthesis groups, biut the numbers free from embolism are compared with curves of sim- patients was ple survival for patients undergoing aortic valve re- were small. Actuarial survival for the1se % at 7.0 years. placement receiving the three types of prosthesis. At Thromboembolism. Only the first tihromboembolic 7.0 years % of patients receiving Bjork- Twelve Shiley aortic valve replacements were alive and event was considered in the actuarial arnalysis Hancock '9 ' X X Carpentier Edwards._ -0 \S S S 'X Years THROMBOEMBOLISM: AORTIC VALVE REPLACEMENT 30- H f% -A. U Years Survival Survival Free from Embolism FIGURE 4. Actuarial survival and survival free from embolism for patients undergoing aortic valve replacement with the three prostheses CIRCULATION

7 TABLE 7 Incidence of emboli in patients treated and those not treated with anticoagulants THERAPY AND PREVENTION-VALVE REPLACEMENT Fatal emboli or Atrial No. of with Patient-years Emboli per 100 Prosthesis and anticoagulant treatmenta fibrillation emboli residua of treatment patient-years Aortic valve replacement Bjdrk-Shiley + anticoagulants (100) Hancock + anticoagulants (7) Carpentier-Edwards + anticoagulants (8) Hancock, no anticoagulants (34) p = NS Carpentier-Edwards, no anticoagulants (45) Mitral valve replacement Bjdrk-Shiley + anticoagulants (122) Hancock + anticoagulants (16) Carpentier-Edwards + anticoagulants (20) p = NS Hancock, no anticoagulants (30) Carpentier-Edwards, no anticoagulants (51) ANumbers in parentheses are initial numbers of patients in each group. ± 6.2% were alive and free from embolism. At 7.0 years the figures for Hancock aortic valve replacements were % and %, respectively, and those for Carpentier-Edwards aortic valve replacements were 71.9 ± 9.5% and %, respectively. There was no significant difference in the overall curves for the three prosthesis groups. Mitral valve replacement. Patients undergoing mitral valve replacement were significantly more likely to experience a thromboembolic event than patients undergoing aortic valve replacement (p <.005). There were 44 embolic events with three fatalities, two in patients with Bjdrk-Shiley prostheses and one in a patient with a Hancock prosthesis. There were 19 major embolic events that produced residual signs noted at subsequent follow-up (12 in patients with Bjdrk-Shiley, three in those with Hancock, and four in those with Carpentier-Edwards valves; table 7). Actuarial curves for survival and survival free from thromboembolism for patients undergoing mitral valve replacement with each of the prostheses is shown in figure 5. At 7.0 years 58.5 ± 5.2% of patients receiving Bjbrk- Shiley mitral valve replacements were alive and 48.8 ± 9.2% were alive and free from embolism. For Hancock mitral valve replacements the values were 65.3 ± 7.8% and 49.3 ± 8. 1%, respectively, and for Carpentier-Edwards mitral valve replacements, 46.4 ± 7.8% and 39.2 ± 8.8%, respectively. There was no significant difference in the overall curves for the three prosthesis groups. Effects ofanticoagulation on incidence ofthromboembolism. Initial policy with respect to anticoagulation was as described above. Any patient not receiving anticoagu- Vol. 73, No. 6, June 1986 lants who experienced an embolic event was started on anticoagulants. During the course of the trial the policy of some physicians toward anticoagulation in patients with porcine heterograft valves changed; some patients who had not been receiving these drugs and had not experienced an embolic event were nevertheless prescribed anticoagulants. Because of this variation in the time that each individual patient may have been treated with anticoagulants we calculated the total number of patient-years with and without treatment for each prosthesis group. We expressed the incidence of embolism as the number of emboli occurring in patients taking anticoagulants at the time of embolism divided by the total number of patient-years of treatment with anticoagulants, and made similar calculations for patients not treated with anticoagulants at the time of embolism (table 7). There was no significant difference in the ratio of number of emboli to years of treatment exposure for patients who did and those who did not receive anticoagulants. Those patients receiving a Hancock mitral valve replacement had a low rate of embolism, but the number of patients was small and the difference did not reach statistical significance. Furthermore, there was no significant difference in the occurrence of emboli with residua in patients in these treatment groups. Effect of other variables on incidence of thromboembolism. The effect of preoperative and postoperative variables on the incidence of thromboembolism was analyzed by multiple regression analysis. For those undergoing aortic valve replacement, age less than 65 years (p <.01 ) and rheumatic cause for valvular disease (p <.01 ) were found to be significantly related to incidence of 1219

8 BLOOMFIELD et al. 100Q Bjork 'X. Shiley /0 Survival S0 ----*~ o0 H Years L_ a THROMBOEMBOLISM: MITRAL VALVE REPLACEMENT Survival Survival Free from Embolism FIGURE 5. Actuarial survival and survival free from embolism for patients undergoing mitral valve replacement with the three prostheses. thromboembolism. Six of nine patients with rheumatic aortic valve disease had clinically significant mitral valve disease. For patients undergoing mitral valve replacement the presence of atrial fibrillation was found to be the only variable predictive of thromboembolism and was highly significant (p <.001). The presence of these variables was randomly distributed among the patients in the three prosthesis groups. Seventy-eight patients of 262 undergoing mitral valve replacement remained in sinus rhythm and of these, six of 34 with Bjork-Shiley, one of 17 with Hancock, and one of 27 with Carpentier-Edwards prostheses suffered a thromboembolic event. In this small subgroup of patients the incidence of thromboembolism in patients with Bjork-Shiley prostheses was higher than for patients with the porcine prostheses, but the difference did not reach statistical significance (p =.12). Complications ofanticoagulation. There were 18 events, 14 in patients with Bjork-Shiley prostheses and two each in patients with Hancock and Carpentier-Edwards prostheses. Two anticoagulant-related deaths occurred in patients with Bjork-Shiley prostheses. The number of events per 100 patient-years of treatment was similar for the three prostheses: 1.0, 1. 5, and 1.5 for Bjork- Shiley, Hancock, and Carpentier-Edwards prostheses, respectively Discussion There have been several nonrandomized studies reporting results of valve replacement with porcine heterograft and Bjork-Shiley prostheses,5-'0 but no randomized prospective trials comparing them. After a median follow-up period of 5.6 years in a large randomized prospective trial we have found no significant difference in survival of patients receiving Bjork-Shiley, Hancock, and Carpentier-Edwards prostheses for mitral, aortic, or combined mitral and aortic valve replacements. Multiple factors have been shown to influence early and late survival after valve replacement." 1-6 Randomization in this study resulted in the random distribution among the treatment groups of 34 preoperative clinical variables. The three participating surgeons treated equal numbers of patients in the three prosthesis groups. Of 612 patients considered for the trial, 88.3% fulfilled the entry criteria. We therefore believe that the treatment groups were well randomized and that any difference in outcome can be considered to represent an effect of the prosthesis used. The randomization process was potentially compromised by the change in the type of porcine prosthesis used (the Carpentier-Edwards prosthesis replaced the Hancock prosthesis) after the trial had commenced. CIRCULATION

9 THERAPY AND PREVENTION-VALVE REPLACEMENT This introduced the possibility that the time of operation might bias results. However, results of Bjork- Shiley valve replacement were not significantly different during the early and late periods of the trial and we therefore felt justified in considering the patients who received this valve as one group. The change from the Hancock to the Carpentier-Edwards prosthesis after 14 months resulted in a larger cohort of patients receiving the latter and the duration of follow-up for those patients was less. Deaths from noncardiac causes occurred with equal frequency among the different patient groups. Whenever possible clinical or autopsy information was sought to determine whether the cause of death could be attributed to the prosthetic valve, but this was not available in all cases. From the information available it appears that there was no difference in the incidence of prosthesis-related death among the patients with the three prostheses. In-hospital mortality for patients receiving the Carpentier-Edwards prosthesis for mitral valve replacement was significantly higher than that for patients receiving Hancock or Bjork-Shiley prostheses. Careful analysis of hospital records failed to provide an explanation and no specific problems were identified by the surgeons involved. The overall in-hospital mortality for mitral valve replacement was 8.8%, which is similar to the figures from the United Kingdom Cardiac Surgery Register of 7.0%, 9.6%, and 8.5% for the years 1977, 1978, and 1979.'7 An increased mortality for patients receiving Carpentier-Edwards prosthesis for mitral valve replacement has not been reported in other nonrandomized series of valve replacement surgery.'8 There was no increased in-hospital mortality associated with any one of the prostheses used for aortic valve replacement, with an overall hospital mortality of 7.6%. Results from the U.K. Cardiac Surgery Register for the years 1977, 1978, and 1979 were 8.6%, 7.9%, and 7.4% respectively. 17 There were too few patients undergoing combined aortic and mitral valve replacement to make meaningful comparisons and the overall operative mortality of 10% compares favorably with results from the U.K. Cardiac Surgery Register of 17.0%, 16.2% and 16.0% for the years 1977, 1978, and '7 Reoperation for valve failure was necessary in 37 patients and resulted in 11 deaths. The high operative mortality reflects the fact that many of these operations were carried out as emergency procedures. Thrombotic occlusion was the most common reason for reoperation in patients receiving Bjork-Shiley prostheses, as has been reported in other series.9 10 Degeneration of the valve cusps or rupture of the cusps was the most common cause of porcine heterograft failure. Severe obstruction also developed with heterograft valves and one 26-year-old patient with a Carpentier-Edwards mitral prosthesis developed thickening of the valve cusps with severe functional obstruction within 3 weeks of implantation. This patient has already been reported. 19 One further patient, 24 years old, developed thickening, calcification, and obstruction of a Hancock aortic prosthesis within 10 months of surgery and required reoperation. The fact that younger patients may more readily develop obstruction in heterograft prostheses8 20 was not recognized at the beginning of the study and accordingly younger patients were not excluded. There was no significant difference in the frequency with which reoperation was required for patients with the three prostheses and no difference in the mortality associated with such surgery. It has been stated that the risk of valve rereplacement is less with porcine heterograft prostheses because the time course of cusp degeneration is more prolonged than that of thrombolic occlusion that may occur with tilting-disk prostheses.2' There was no evidence of such a difference in this trial, although the number of patients was small. Indeed, several patients with heterograft failure presented as acute emergencies. The rate of porcine valve degeneration may accelerate 6 or more years after surgery,' 22, 23 and further follow-up of patients in this study is clearly important. There was no difference in the incidence of thromboembolism in patients with the three prostheses undergoing aortic, mitral, or aortic plus mitral valve replacement. Thromboembolism was associated with age less than 65 years and a rheumatic cause of heart disease in those with aortic valve replacement and with atrial fibrillation in those with mitral valve replacement. Only in the small group of patients undergoing mitral valve replacement who remained in sinus rhythm did the type of prosthesis exert an effect, with a higher rate of thromboembolism associated with the Bjork-Shiley prosthesis, although this did not reach statistical significance. The effect of anticoagulants in patients undergoing heterograft valve replacement was difficult to assess because patients were not randomly assigned to anticoagulant treatment and the results were likely to be biased by the physicians' choice of treatment. There appeared to be no beneficial effect of anticoagulants in patients undergoing aortic valve replacement, and there was only a small (nonsignificant) advantage of their use in patients receiving the Hancock valve, but not the Carpentier-Edwards prosthesis, for mitral valve replacement. It is therefore difficult to make recommendations on the basis of these results Vol. 73, No. 6, June

10 BLOOMFIELD et al. but it would appear to be a reasonable policy to prescribe anticoagulants for all patients undergoing aortic valve replacement who have coexisting mitral valve disease and all patients undergoing mitral valve replacement who are in atrial fibrillation. The risks of anticoagulation in this trial were low, with an overall incidence of complications of approximately one per 100 patient-years, which is in keeping with a previous report from one of our institutions.24 In summary, after a median period of follow-up of 5.6 years no significant advantage of the use of any of the three prostheses studied has yet appeared, but further follow-up is necessary since important differences may yet emerge. We thank Mrs. J. Chalmers and Mr. W. Adams of the Medical Computing and Statistics Unit, and Mrs. Jean Cunningham for secretarial support throughout the study. References 1. McLung JA, Stein JA, Ambrose JA, Herman MV, Reed GE: Prosthetic heart valves; a review. Prog Cardiovasc Dis 26: 237, Carpentier KA, Delouche A, Relland J, Fabiani JN, Forman J, Camilleri JP, Soyer R, Dubost C: Six year follow-up of glutaraldehyde-preserved heterografts. J Thorac Cardiovasc Surg 68: 771, Stinson EB, Griepp RB, Oyer PE, Shumway NE: Long-term experience with porcine aortic valve xenografts. J Thorac Cardiovasc Surg 73: 54, Grunkemeier GL, Lambert LE, Bonchek LI, Starr A: An improved statistical method for assessing the results of operation. Ann Thorac Surg 20: 289, Oyer PE, Stinson EB, Reitz BA, Miller DC, Rossiter SJ, Shumway NE: Long-term evaluation of the porcine xenograft prosthesis. J Thorac Cardiovasc Surg 78: 343, Cohn LH, Mudge GH, Pratter F, Collins JJ: Five to eight year follow-up of patients undergoing porcine heart valve replacement. N Engl J Med 304: 258, Angell WA, Angell JD, Kosek JC: Twelve year experience with glutaraldehyde-preserved porcine xenografts. J Thorac Cardiovasc Surg 83: 493, Magilligan DJ, Lewis JW, Tilley B, Peterson E: The porcine bioprosthetic valve: twelve years later. J Thorac Cardiovasc Surg 89: 499, Bjork VO, Henze A: Ten years experience with the Bjork-Shiley tilting disc valve. J Thorac Cardiovasc Surg 78: 331, Marshall WG, Kouchoukos NT, Karp RB, Williams JB: Late results after mitral valve replacement with the Bjork-Shiley and porcine prostheses. J Thorac Cardiovasc Surg 85: 902, Chaffin JS, Daggett WM: Mitral valve replacement: a nine year follow-up of risks and survivals. Ann Thorac Surg 27: 312, Phillips HR, Levine FH, Carter JE, Boucher CA, Osbakken MD, Okada RD, Akins CW, Daggett WM, Buckley MJ, Pohost GJ: Mitral valve replacement for isolated mitral regurgitation; analysis of clinical course and late post-operative left ventricular ejection fraction. Am J Cardiol 48: 647, Hammermeister KE, Fischer L, Kennedy JW, Samuels S, Dodge HT: Prediction of late survival in patients with mitral valve disease from clinical, haemodynamic and quantitative angiographic variables. Circulation 57: 341, Czer LS, Gray RJ, DeRobertis MA, Bateman TM, Stewart ME, Chaux A, Matloff JM: Mitral valve replacement: impact of coronary artery disease and determinants of prognosis after revascularisation. Circulation 70(suppl 1) 1-198, Hirshfield JW, Epstein SE, Roberts AJ, Glancy DL, Morrow AG: Indices predicting long-term survival after valve replacement in patients with aortic regurgitation and patients with aortic stenosis. Circulation 50: 1190, Mirsky I, Henschke C, Hess OM, Krayenbuehl HP: Predictors of post-operative performance in aortic valve disease. Am J Cardiol 48: 295, English TAJ, Bailey AR, Dark JF, Williams WG: The UK Cardiac Surgery Register Br Med J 289: 1205, Jamieson WRE, Pelletier LC, Janusz MT, Chairman BR, Tyres GFO, Miyagishima RT: Five year evaluation of the Carpentier- Edwards porcine bioprosthesis. J Thorac Cardiovasc Surg 88: 324, Forfar JC, Cotter L, Morritt GN: Severe and early stenosis of porcine heterograft mitral valve. Br Heart J 40: 1184, Curcio CA, Commerford PJ, Rose AG, Stevens JE, Barnard MS: Calcification of gluteraldehyde preserved porcine xenografts in young patients. J Thorac Cardiovasc Surg 81: 621, Craver JM, Jones EL, McKeown P, Bone DK, Hatcher CR, Kandrach M: Porcine cardiac xenograft valves. Analysis of survival valve failure and explantation. Ann Thorac Surg 34: 16, Angell WW, Angell JD: Porcine valves. Prog Cardiovasc Dis 23: 141, Oyer PE, Miller DC, Stinson EB, Reitz BA, Moreno-Cabral RJ, Shumway NE: Clinical durability of the Hancock porcine bioprosthetic valve. J Thorac Cardiovasc Surg 80: 824, Forfar JC: A 7 year analysis of haemorrhage in patients on longterm anticoagulant treatment. Br Heart J 42: 128, CIRCULATION