Five-year evaluation of the Carpentier-Edwards porcine bioprosthesis

J

THoRAc CARDIOVASC SURG

88:324-333, 1984

Original Communications

Five-year evaluation of the Carpentier-Edwards porcine bioprosthesis The Carpentier-Edwards porcine bioprosthesis compares favorably to other valve.substitutes, with a low incidence of valve-related complications after up to 6 years' follow-up. From April, 1976, to February, 1978, 397 prostheses were implanted in 355 patients at the University of British Columbia and the Montreal Heart Institute: aortic valve replacement (AVR), 155; mitral valve replacement (MVR), 154; tricuspid valve replacement (fVR), five; multiple replacement (MR), 41 patients. Previous cardiac operatiOllS had been performed in 58 patients (16;3%). Concomitant cardiac procedures, including myocardial revascularization, were performed in 66 patients (18.5% ~ The 30 day mortality was 8.7% (31 patients (AVR, 4.5%; MVR 9.0%; TVR 20%; MR 21.9%~ The cumulative follow-up was 1,367 The late mortality is 3.9% per patient-year (AVR patient-years (range 6 to 81 months, mean 51 months~ 3.4%, MVR 3.6%, MR 6.3%). The long-term valve-related complication rates (expressed as percent per patient-year and number of events) are as follows: thromboembolism (AVR 1.1 [seven], MVR 1.7 [101 MR 3.2 [four)); prosthetic valve endocarditis (AVR 0.6 [four], MVR 0.3 [two], MR 1.6 [two]~ primary tissue failure (AVR 0.16 [one], MVR 1.0 [six], MR 1.6 [two)); anticoagulant-related hemorrhage 0.07 (one~ and reoperation (AVR 0.48 [three], MVR 1.32 [eight1 MR 2.4 [three~ Actuarial survival rates at 6 years, including operative deaths, are as follows: AVR 81.4% ± 5.6%, MVR 70.0% ± 6.7%, and MR 52.1 % ± 12.5%. TIJe overall survival rate was 74.0% ± 4.5% at 6 years. The freedom from all valve-related complications at 72 months was 83.9 % ± 4.8 %. The freedom from valve-related death was 97.3% ± 1.8% at 72 months. The freedom from valve-related death and reoperation at 72 months was 91.5 % ± 3.8 %, and the freedom from reoperation alone was 93.4 % ± 3.5 %. There were no deaths due to reoperation for primary tissue failure. Performance to date with the Carpentier-Edwards porcine bioprosthesis has been excellent, with an evaluation interval extending from 60 to 81 months.

W. R. Eric Jamieson, M.D., L. Conrad Pelletier, M.D., Michael T. Janusz, M.D., Bernard R. Chaitman, M.D., G. Frank O. Tyers, ~.D., and Robert T. Miyagishima, M.D., Vancouver, B. C, and Montreal, Quebec, Canada h e glutaraldehyde-preserved porcine bioprosthesis was introduced in 1969 to reduce the risk of thrombo-

From the Division of Cardiovascular and Thoracic Surgery, Department of Surgery, Vancouver General Hospital and St. Paul's Hospital, University of British Columbia, Vancouver, B. C, Canada, and the Montreal Heart Institute, University of Montreal, Montreal, Quebec, Canada. Read at the Ninth Annual Meeting of The Western Thoracic Surgical Association, Colorado Springs, Colo., June 19-22, 1983. Address for reprints: W. R. E. Jamieson, M.D., Department of Surgery. 3rd Floor, 910 W. 10th Ave., Vancouver, B. C, Canada V5Z 4E3.

324

embolism associated with mechanical prostheses.I The Hancock prosthesis and the Carpentier-Edwards prosthesis are the two major porcine bioprostheses available today. The Hancock valve has been extensively studied, up to 10 years, for functional performance and valverelated complications>" The Carpentier-Edwards prosthesis was introduced in 1975 and has had less extensive evaluation of clinical performance.'>" The two porcine prostheses differ in technique of tissue ftxation and preservation, tissue selection and mounting, and prosthesis design and construction. The long-term durability of both these prostheses must be fully evaluated.

Volume 88

Carpentier-Edwards porcine bioprosthesis 3 2 5

Number 3 September, 1984

This study reports the experience of two centers with the Carpentier-Edwards porcine bioprosthesis implanted for periods of more than 5 years in 355 patients. Patients and methods The Carpentier-Edwards porcine bioprosthesis was first implanted at the University of British Columbia teaching hospitals (UBC) in April, 1976, and the Montreal Heart Institute (MHI) in May, 1976. By January, 1983,2,095 patients had received the Carpentier-Edwards valve: 1,315 patients at UBC and 780 patients at MHI. The two centers have implanted 397 prostheses in 355 patients, with follow-up of 5 years or longer (UBC 235, MHI 120). There were 183 males and 172 females ranging in age from 15 to 82 years, with a mean of 57.0 years. Valve replacements were performed singly in the aortic position in 155 patients, mitral position in 154 patients, and tricuspid position in five patients. Multiple valve replacements were performed in 41 patients. The operative techniques of valve replacement at both centers were at the discretion of the individual surgeons. Since late 1976, hypothermic potassium crystalloid cardioplegia has been used for myocardial protection. Concomitant procedures, especially myocardial revascularization, were performed as indicated by complete investigation. Long-term anticoagulation was utilized in the presence of chronic atrial fibrillation, intracardiac thrombus, or dilated left atrium. The vast majority of mitral valve replacement (MVR) and multiple replacement (MR) patients received early postoperative anticoagulation, whereas the majority of patients undergoing aortic valve replacement (AVR) did not receive anticoagulation. Definitions The valve-related complications evaluated were those recommended by Oyer and colleagues' for assessment of both mechanical and tissue valves. The valve-related complications were thromboembolism, anticoagulantrelated hemorrhage, prosthetic valve endocarditis, periprosthetic leak, and prosthesis failure. Reoperations or death resulting from any of these are considered to be valve-related reoperations or deaths. Thromboembolism is defined as all new, focal neurological deficits, either transient or permanent. Cerebrovascular accidents occurring intraoperatively are not considered to be valve-related thromboembolism. Anticoagulant-related hemorrhage included internal or external bleeding that necessitated hospital care, including transfusion, or extensive outpatient care. Prosthetic valve endocarditis was defined as any documented infection of the prosthesis. Operative management was dictated by the presence

% 100

80 60

40

20

o

III II Pre-op

IV

II III Post-op

IV

Fig. 1. Preoperative and postoperative New York Heart Association classifications.

of progressive and severe congestive heart failure, systemic emboli, and persistent sepsis. Periprosthetic leak includes all episodes of periprosthetic leak documented by cardiac catheterization, reoperation, or autopsy. This excludes those patients having a periprosthetic leak caused by infective endocarditis. In those cases the complication is listed as infective endocarditis. Prosthesis failure is defmed as any structural failure of the prosthesis causing stenosis or insufficiency, including leaflet disruption, calcification, stent failure, or thrombosis, documented by cardiac catheterization, reoperation, or autopsy. This excludes valve failures resulting from infective endocarditis or periprosthetic leak, which are listed separately. Reoperation was defined as re-replacement for any cause. Reoperation for repair or replacement of another native valve or other cardiac pathology unrelated to the prosthesis (e.g., coronary artery disease) is not included. The methodology of valve assessment reported by Bodnar, Wain, and Haberman" was used in our evaluation of valve performance. With this method, valve survival and complication-free valve survival were determined. This assessment allowed for separation of valve survival from patient survival. Freedom from valverelated death (or freedom from fatal valve-related complications) denotes the actuarial proportion of patients who are alive or have died of a non-valverelated cause. Freedom from valve-related mortality or reoperation denotes the actuarial proportion of patients who remain free from valve-related death or from reoperation for a valve-related cause. Freedom from all valve-related complications is the actuarial proportion of patients who remain free from any valve-related complications, including those causing death or reoperation.

3 26

The Journal of Thoracic and Cardiovascular Surgery

Jamieson et al.

Table ill. Causes of late postoperative deaths

Table I. Operative and early* deaths Valve replacement operation

No. of operations

No. of deaths

Aortic Mitral Tricuspid Multiple

155 154 5 41

7 14 I 9

9.0 20.0 21.9

Totals

355

31

8.7

%

4.5

• Less than 30 days.

4 2 I

Thromboembolism Prosthetic valve endocarditis Peri prosthetic leak Congestive heart failure Acute myocardial infarction/arrhythmia Reoperation Pulmonary embolism Aortic dissection Aortic aneurysm Other

I 16

Total

53*

9 16 I I

2

'Aortic valve replacement, 21; mitral valve replacement, 22; tricuspid valve replacement. two; multiple ~alve replacement, eight.

Table II. Causes of early postoperative death Low output syndrome/congestive heart failure Myocardial infarction/arrhythmia Technical error Sepsis Hemorrhage Mesenteric infarction Noncardiac cause

17

Total

31

7 I I I I

3

Statistical analysis Standard actuarial (life-table method) and linearized statistical techniques were used to describe survival and the incidence of valve-related complications. 16-18 The multiple decrement analysis of Bodnar, Wain, and Haberman" was used to evaluate valve performance. Results

FoUow-up evaluation The clinical evaluation was obtained by direct contact with the patients and/or attending physicians during a 4 month closing period prior to February, 1983. The average duration of follow-up of surviving patients was 51 months (range 6 to 81 months). The cumulative follow-up for all patients was 1,367 patient-years (AVR, 625 patient-years; MVR, 603 patient-years; tricuspid valve replacement [TVR], 12 patient-years; and MR, 126 patient-years). The patient evaluation was 99.3% complete (322 of the 324 patients who survived operation). Preoperative status Patients' symptoms were assessed according to the New York Heart Association functional classification. The majority of patients were in Class III or IV (57% in Class III and 18% in Class IV); 24% were in Class II and 1% in Class I (Fig. 1). Previous cardiac surgery had been performed in 58 patients (16.3%) (previous valve replacement or commissurotomy in 55; myocardial revascularization in two; and repair of coarctation in one). Concomitant surgical procedures were performed in 66 patients (18.5%). These included myocardial revascularization in 37 patients, pacemaker insertion in 13 patients, repair of another valve in 11 patients, resection of an ascending aortic aneurysm in three patients, and correction of a congenital defect in two patients.

Early mortality and complications. The overall hospital (30 day) mortality for the 355 patients was 8.7% or 31 patients (Table I). AVR was performed in 155 patients with seven deaths (4.5%), MVR was performed in 154 patients with 14 deaths (9.0%), and TVR was performed in five patients with one death (20.0%). MR was performed in 41 patients with nine deaths (21.9%). Low-output syndrome and/or congestive failure was the cause of 54.8% (17/31) of the early deaths (Table II). The other causes of death were myocardial infarction and/or arrhythmias (seven), technical error (one), sepsis (one), hemorrhage (one), mesenteric infarction (one), and noncardiac causes (three). Late mortality and complications. There have been 53 late deaths in the series (14.9%). The late mortality, expressed as a linearized occurrence rate, was 3.9% per patient-year, overall. The late mortality for AVR is 3.4% per patient-year (21); for MVR, 3.6% per patientyear (22); and for MR, 6.3% per patient-year (eight). The causes of late death are detailed in Table III. The valve-related causes were thromboembolism (four), prosthetic valve endocarditis (two), and periprosthetic leak (one). The cardiac-related causes were congestive heart failure (nine) and myocardial infarction and/or arrhythmia (16). The other causes were reoperation (one), pulmonary embolism (one), aortic dissection (two), aortic aneurysm (one), and miscellaneous causes (16). The overall patient survival, including operative

Volume 88

327

Carpentier-Edwards porcine bioprosthesis

Number 3 September, 1984

Table IV. Valve dysfunction-valve-related complications (linear mortality and morbidity rates*)

Patient Survival

Type of valve replacement Mortality-morbidity

Thromboembolism Overall Fatal Prosthetic valve endocarditis Overall Fatal

Aortic

Multiple

7 (1.1) 0

10 (1.7) 3 (0.5)

4 (3.2) 1 (0.8)

4 (0.6) 1 (0.2)

2 (0.3) 0

2 (1.6) 1 (0.8)

'Expressed as percent per patient-year or episodes per 100 patient-years.

deaths, expressed by the life-table method, is 74.0% ± 4.5% at 6 years (Fig. 2). AVR patients is 81.4% ± 6.7%; for MVR patients, 70.0% ± 6.7%; and for MR patients, 52.1% ± 12.5%. The postoperative functional classification showed considerable improvement in the majority (Fig. I). Ninety-two percent of the patients were in Class I or II postoperatively (Class I, 69.5%; Class II, 22.5%; Class III, 6%; and Class IV, 2%). Valve-related complications. The significant valverelated late complications were thromboembolism, prosthetic valve endocarditis, primary tissue failure, periprosthetic leak, and anticoagulant-related hemorrhage. The linearized occurrence rate for valve-related complications was 3.0% per patient-year (41 events) and the fatality rate was 0.6% per patient-year (eight events). The freedom from all valve-related complications was 83.9% ± 4.9% at 6 years. Thromboembolism. There were 21 postoperative thromboembolic events in the total patient population (Table IV). The distribution of major and minor thromboembolic events is delineated in Table V. Of the seven events in AVR patients, three were major events and four were minor events, with no deaths. Of the 10 episodesin MVR patients, one was minor and nine were major, with three deaths. Embolic events occurred in four MR patients, one minor and three major, with one death. The linearized occurrence rate of thromboembolism is shown in Tables IV and V. The rate of embolic events for the overall group of patients was 1.5% per patientyear. The thromboembolic rate following AVR was 1.1% per patient-year, 1.7% for MVR, and 3.2% for MR. The linearized occurrence rate for deaths related to thromboembolism was 0.50% per patient-year for MVR and 0.80% for MR, with no deaths in patients with AVR. Anticoagulation complications. There was only one anticoagulant-related hemorrhage documented in the

50 I

I

6 12

I

24

I

I

I

36

60

48

72

I

84

Months Fig. 2. Patient survival. AVR, Aortic valve replacement. MVR, Mitral valve replacement. MR, Multiple valve replacement.

Table V. Thromboembolism Events Valve

Total

AVR MVR MR

7 (1.1)* 10 (1.7) 4 (3.2)

Totals

21 (1.5)

Major 3

9 3

15 (1.1)

I

Minor 4 I 1 6 (0.4)

Mortality

o 3 1

4 (0.3)

Legend: AYR. Aortic valve replacement. MVR, Mitral valve replacement. MR.

Multiple valve replacement. 'Expressed as percent per patient-year or episodes per 100 patient-years.

total series, I V2 years after MVR (0.07% per patientyear). The patient survived. Approximately 10% of AVR patients and 45% of MVR patients received long-term anticoagulation. Prosthetic valve endocarditis. Prosthetic valve endocarditis occurred in eight patients. Four instances followed AVR, two followed MVR, and two followed MR (Table IV). The linearized occurrence rate of endocarditis following AVR was 0.60% per patient-year; following MVR, 0.30% per patient-year; and following MR, 1.6% per patient-year. The linearized occurrence rate for deaths due to endocarditis was 0.20% per year for AVR and 0.80% per year for MR. There were no deaths from endocarditis after MVR. Periprosthetic leaks. Significant periprosthetic leaks occurred in two patients, one after AVR and one after MVR. Both patients underwent reoperation, and one died (MVR). The linearized occurrence rate for periprosthetic leak was 0.16% for AVR and 0.20% for

The Journal of

3 2 8 Jamieson et al.

Thoracic and Cardiovascular Surgery

Table VI. Valve dysfunction-valve-related complications (linear mortality and morbidity rates*)

100

Type of valve replacement Mortality-morbidity ._- valve-related mortality ·······valve-related mortality or re-operation --- freedom from all valve-related complications

I

I

6 12

I

24

I

36

48

I

60

I

72

I

84

Months

Fig. 3. Freedom from valve-related death and complications.

MVR. There were no periprosthetic leaks following MR. The rate of death related to periprosthetic leak was zero for AVR and 0.20% for MVR (Table VI). Primary tissue failure. Primary tissue failure has occurred in nine patients (0.66% per patient-year) (AVR, one; MVR, six; MR, two), as shown in Table VI. The linearized occurrence rate for AVR patients is 0.16% per patient-year; for MVR patients, 1.0% per patient-year, and for MR, 0.16% per patient-year. Eight patients had reoperation, with no deaths. Primary tissue failure was identified at autopsy in one MVR patient, so that the linearized occurrence rate for death from MVR was 0.20% per patient-year. The patients with primary tissue failure were reoperated upon 32 to 67 months after the initial operative procedure (mean 48.2 months). Ofthe nine patients, the mitral prosthesis was involved in eight and the aortic prosthesis in only one. Two of the mitral prostheses had been implanted during multiple replacements. Of the nine patients, seven were female and only two were male. Three of the patients were adolescents when the valve was inserted. Regurgitation from prosthesis incompetence owing to leaflet tears was the predominant lesion in five patients (aortic prosthesis, one; mitral prosthesis, four). Stenosis was predominant in four patients, all having a mitral prosthesis; due to calcification in three patients and to tissue ingrowth in one. Of the three patients who had valve insertions as adolescents, two presented with a mitral prosthesis calcified at 30 and 36 months, and one with an aortic prosthesis presented with a tom leaflet and calcification of all three leaflets at 62 months. Reoperation for valve-related complications. Fourteen patients (AYR, three; MVR, eight; MR, three) had reoperation for valve-related complications. There were four reoperations related to prosthetic valve endocarditis (AVR, one; MVR, two; MR, one), with one death

Primary tissue failure Overall Fatal Periprosthetic leak Overall Fatal

Aortic

Multiple

1 (0.16) 0

6 (1.0) 1 (0.2)

2 (1.6) 0

1 (0.16) 0

1 (0.2) 1 (0.2)

0 0

'Expressed as percent per patient-year or episodes per 100 patient-years.

(AVR). There were tworeoperations for periprosthetic leak (AVR, one; MVR, one), with one death (MVR). Eight reoperations were due to primary tissue failure (AVR, one; MVR, five; MR, two). All patients who underwent reoperation for primary tissue failure survived. There were no reoperations for thromboembolism or anticoagulant-related hemorrhage. Freedom from valve-related complications. The freedom from valve-related death and complications is shown in Figs. 3 to 6. The freedom from all valve-related complications at 6 years was 83.9% ± 4.8%. The freedom from valve-related death was 97.3% ± 1.8% at 6 years. The freedom from valve-related death or reoperation at 6 years was 91.5% ± 3.8%, and the freedom from reoperation alone was 93.4% ± 3.5%. Discussion This study was undertaken to document the performance of the Carpentier-Edwards porcine bioprosthesis implanted for more than 5 years. The Hancock porcine bioprosthesis was introduced in 1970 and the Carpentier-Edwards valve in 1975. The Hancock prosthesis has received extensive evaluation, but concern continues over bioprosthesis durability. Since the CarpentierEdwards valve has distinct differences in tissue fixation and preservation and prosthesis construction, long-term assessment of this porcine prosthesis is also warranted. The thrust for standardization of assessment of prosthesis performance has come from Miller," Bodnar," and their colleagues. Emphasis has been placed on comprehensive comparison of all valve-related complications, so that the various mechanical and tissue valves can be analyzed adequately. Bodnar, Wain, and Haberman" recommend that the valve(s) must be made the subject of the investigation, and that valve survival and valve-related deaths and complications be recorded separately from overall patient survival. The standard valve-related complications, namely, thromboembolism, anticoagulant-related hemorrhage,

Volume 88 Number 3 September, 1984

Carpentier-Edwards porcine bioprosthesis 3 2 9

100 E o

-0 Q)

~

u..

~

90

E

- t __ ! __

r-!-!-1- y-

90

1-

r- r- 1_ r-r-r

0

-0

80

e Q)

u..

~

70 314 304 296

60 0'(

289

279

263

172

24

36 Months

48

60

72

70 311

60 0'(

I

I

1 6 12

25

80

84

Fig. 4. Freedom from valve-related death.

prosthetic valve endocarditis, periprosthetic leak, and valve failure, are generally recognized by most investigators, but seldom are all used for valve assessment. Minale," Schonbeck," and their associates considered all these valve-related complications in their comparisons of porcine bioprostheses and mechanical prostheses. On the other hand, Ionescu and co-workers" considered only thromboembolism and valve failure when assessing the Ionescu-Shiley pericardial valve. The Stanford University group does not consider all valve-related complications in the assessment of valve failure.'? This group considers the occurrence of any of the following events to constitute valve failure: (1) new postoperative regurgitant murmur unless proved to result from a periprosthetic leak, (2) thromboembolism necessitating reoperation or resulting in death, (3) thrombotic valvular occlusion resulting in reoperation or death, (4) prosthetic valve endocarditis necessitating reoperation or causing death, and (5) significant hemodynamic dysfunction resulting in reoperation or death.'? We believe that all valve-related complications must be considered even though they may not necessitate reoperation or result in death. The standardization' of valve-related complications, including strict definitions, will facilitate future comparison of all types of valvular prostheses. The Starr-Edwards valve performance is regarded as the standard against which newer mechanical and tissue valve bioprostheses require comparison. The performance characteristics of the Starr-Edwards Model 6120 mitral prosthesis recently have been documented extensively." Considering all valve-related complications, 22% of early hospital deaths and 27% of late deaths are valve related. Miller and colleagues" concluded that the long-term performance of the Starr-Edwards prosthesis Model 6120 was markedly suboptimal. Thromboembolism. Edmunds" has provided extensive documentation on thromboembolism with tissue and mechanical prostheses. The incidence of thrombo-

301 293

286

273

257

165

25

6 12

24

36

48

60

72

, 84

Months

Fig. 5. Freedom from valve-related death and reoperation.

100 E

90

o

-0

e Q)

u..

~

80 70 311 296 284

60 0'(

273

261

242

155

23

24

36

48

60

n

I

1 6 12

I

84

Months

Fig. 6. Freedom from all valve-related complications.

embolism in the aortic position is less than 2% per patient-year with biological valves without anticoagulation and with mechanical valves with anticoagulation. With MVR the incidence of thromboembolism is approximately 4% per patient-year with or without anticoagulation with biological valves and with anticoagulation with mechanical valves. The anticoagulantrelated mortality and morbidity are 0.17% and 2.2% per patient-year, respectively, with mechanical valves. Without anticoagulation of mechanical prostheses, the incidence of thromboembolism rises threefold to sixfold. Most series of biological valves have documented anticoagulant use on the long term in 10% of AVR and 40% to 60% of MVR and MR patients." This has also been our own long-term experience.v":" The incidence of thromboembolism is 1.5% per patient-year with the Carpentier-Edwards valve compared to 5.7% per patient-year with the Starr-Edwards prosthesis." The incidence of anticoagulant-related hemorrhage is also much lower, the incidence being 0.07% and 1.5% per patient-year, respectively. Oyer and colleagues 2• 3 have reported extensively on the Hancock porcine prosthesis, and their data on thromboembolism are similar to ours. Our fmdings

330

The Journal of Thoracic and Cardiovascular Surgery

Jamieson et al.

and these reports by others have shown the porcine valve to be relatively free from embolic complications and anticoagulant-related morbidity and mortality. Hill and associates" have reported that long-term anticoagulation increases the risk of bleeding with the porcine valve but does not influence the risk of thromboembolism; yet, they recommend anticoagulation for states of restrictive flow related to large atrium and atrial fibrillation. Hetzer, Topalidis, and Borst" have also reported that the risk of hemorrhage from anticoagulants is greater than the risk of embolism, even though all emboli occur in patients with chronic atrial fibrillation. These authors believe that multiple risk factors-i-chronic atrial fibrillation, atrial thrombi, and history of embolism-must be present to recommend anticoagulants beyond 3 months after MVR. The Ionescu-Shiley pericardial valve also has been reported to have a low incidence of thromboembolismY,25 Prosthetic valve endocarditis. There appears to be no difference between mechanical and tissue valves with regard to prosthetic valve endocarditis." Our incidence of 0.6% per patient-year is similar to 0.7% for the Hancock prosthesis? and 0.5% per patient-year for the Starr-Edwards valve. 19 The operative mortality for prosthetic valve endocarditis remains high at approximately 25%. In our series the medical management of prosthetic valve endocarditis was successful in three of four patients conservatively managed. Valve. failure. The question of durability is of paramount concern to cardiac surgeons. The valve failure rate in our series is 0.66% per patient-year, which compares favorably to the experience of Oyer's group2,3 with the Hancock prosthesis and Craver's group' with a combined Hancock and Carpentier-Edwards experience. Magilligan and colleagues," 28 in their series of Hancock valves, have shown that valve failure rate rises sharply after 5 years. This has not been our experience to date with the Carpentier-Edwards prosthesis. Magilligan and colleaguesv-" found 20% of Hancock prostheses falling at 9 years. Lakier and associates,' evaluating the same series of Hancock implants, found that degeneration was a function of duration of implantation and age of the patient at implantation. The higher incidence of early failures in children and young adults has been well documented by several investigators.":" The predicated replacement-free rate in children at 5 years is 58.5%.32 Mechanical valves have potentiality of improved durability. Miller and co-workers" found that durability was the most favorable characteristic of the Starr-Edwards prosthesis, but they did document that 16% of failures were due to thrombotic occlusion and tissue ingrowth. The valve occlusion rate was 0.6% per

patient-year and the fatility rate was 0.5% per patientyear. The valve occlusion rates with the Bjork-Shiley, Smeloff-Cutter, and Lillehei-Kaster mechanical prostheses have been much higher at 2.5% per patientyear.":" Karp and associates" have documented 16 thrombotic occlusions in 15 patients with 13 deaths among 643 patients with the Bjork-Shiley prosthesis. Reoperation. Carver and colleagues' have shown that catastrophic failure is less likely with biological prostheses than mechanical prostheses. The reoperation rate'in that series was 2.3% with an 8.2% mortality. In our series there has been no fatality to date for reoperation for valve failure. If the mortality from reoperation can be maintained at an acceptably low level, the potentially reduced durability of biological valves can be tolerated by surgeons and patients. Durability of the biological prosthesis may be improved with nonpressurized tissue fixation, refined glutaraldehyde solution, flexible stents, and improved hemodynamics.v" Angell, Angell, and Kosek" believe that these factors will extend durability to an ideal of 15 to 20 years. Carpentier and associates" have used these principles in formulating the supra-annular prosthesis that presently is in investigational use. Conclusion In this and other studies we have found the freedom from valve-related complications with the CarpentierEdwards valve to be an acceptable 87% to 93% at 5 to 6 years. Biological valves have a performance similar to that of mechanical valves with regard to susceptibility to endocarditis, and, to date, their durability has been comparable to that of mechanical valves. Their performance with regard to thromboembolism, patient survival, and overall morbidity has been superior to that of mechanical prostheses. We thank Joan MacNab and Manon Bellemare, research assistants, for their assistance in the follow-up assessment. We also thank the surgical staff of the Montreal Heart Institute, Vancouver General Hospital, and St. Paul's Hospital for inclusion of their patients in this assessment. REFERENCES Carpentier A, Lemaigre G, Ladislas R, Carpentier S, Dubost C: Biological factors affecting long-term results of valvular heterografts. J THORAC CARDIOVASC SURG 58:467-483, 1969 2 Oyer PE, Stinson EB, Reitz BA, Miller DC, Rossiter SJ, Shumway NE: Long-term evaluation of the porcine xenograft bioprosthesis. J THORAC CARDIOVASC SURG 78:343350, 1979 3 Oyer PE, Stinson EB, Griepp RB, Shumway NE: Valve replacement with the Starr-Edwards and Hancock pros-

Volume 88 Number 3 September, 1984

thesis. Comparative analysis of late morbidity and mortality. Ann Thorac Surg 186:301-307, 1977 4 Hill JD, La Follette L, Szarnicki RJ, Avery GJ II, Wilson RM, Gerbode F, Kerth WJ, Rodvien R: Risk-benefit analysis of warfarin therapy in Hancock mitral valve replacement. J THORAC CARDIOVASC SURG 83:718-723, 1982 5 Craver JM, Jones EL, McKeown P, Bone OK, Hatcher CR Jr, Kandrach M: Porcine cardiac xenograft valves. Analysis of survival, valve failure and explantation. Ann Thorac Surg 34:16-21,1982 6 Geha AS, Hammond GL, Laks H, Stansel HC, Glenn WWL: Factors affecting performance and thromboembolism after porcine xenograft cardiac valve replacement. J THORAC CARDIOVASC SURG 83:377-384, 1982 7 Lakier JB, Khaja F, Magilligan OJ, Goldstein S: Porcine xenograft valves-long-term (60-89 months) follow-up. Circulation 62:313-318, 1980 8 Jamieson WRE, Janusz MT, Miyagishima RT, Munro AI, Tutassaura H, Gerein AN, Burr LH, Allen P: Embolic complications of porcine heterograft cardiac valves. J THORAC CARDIOVASC SURG 81:626-631, 1981 9 Marshall WG Jr, Kouchoukos NT, Karp RB, Williams JB: Late results after mitral valve replacement with the Bjork-Shiley and porcine prostheses. J THORAC CARDIOVASC SURG 85:902-910, 1983 10 Oyer PE, Stinson EB, Miller DC, Jamieson SW, Reitz BA, Baumgartner W, Shumway NE: Clinical analysis of the Hancock porcine bioprosthesis, Cardiac Bioprostheses. Proceedings of the Second International Symposium, V Gallucci, LH Cohn, eds., New York, 1982, Yorke Medical Books, pp 539-551 II Gallucci V, Valfre C, Mazzucco A, Bortolotti V, Milano A, Chioin R, Dalla Volta S, Cevese PG: Heart valve replacement with the Hancock bioprostheses. A 5-11 year follow-up, Cardiac Bioprostheses. Proceedings of the Second International Symposium, LH Cohn, V Gallucci, eds., New York, 1982, Yorke Medical Books, pp 9-24 12 Pelletier C, Chaitman BR, Baillot R, Val PG, Bonan R, Dynda I: Clinical and hemodynamic results with the Carpentier-Edwards porcine bioprosthesis. Ann Thorac Surg 34:612-624, 1982 13 Janusz MT, Jamieson WRE, Allen P, Gerein AN, Miyagishima RT, Burr LH, Tutassallra H, Munro AI, Tyers GFO: Long-term follow-up of patients with porcine cardiac valve prostheses. Can J Surg 26:160-162, 1983 14 Janusz MT, Jamieson WRE, Allen P, Munro AI, Miyagishima RT, Tutassaura H, Burr LH, Gerein AN, Tyers GFO: Experience with the Carpentier-Edwards porcine valve prosthesis in 700 patients. Ann Thorac Surg 34:625633, 1982 15 Jamieson WRE, Janusz MT, Tyers GFO, Allen P, Munro AI, Burr LH, Gerein AN, Ling H, Miyagishima RT, Tutassaura H: Early durability of the Carpentier-Edwards porcine bioprosthesis, Concepts and Controversies in Cardiovascular Surgery, MJ Kaplitt, JB Borman, eds., New York, 1983, Appleton-Century-Crofts, pp 111-133

Carpentier-Edwards porcine bioprosthesis 3 3 1

16 Bodnar E, Wain WH, Haberman S: Assessment and comparison of the performance of cardiac valves. Ann Thorac Surg 34: 146-156, 1982 17 Kaplan EL, Meier P: Non-parametric estimations from incomplete observations. J Am Stat Assoc 53:457-481, 1958 18 Gross AJ, Clark VA: Survival distributions. Reliability applications in the Biomedical Sciences, New York, 1975, John Wiley & Sons, Inc., pp 240-243 19 Miller DC, Oyer PE, Stinson EB, Reitz BA, Jamieson SW, Baumgartner WA, Mitchell RS, Shumway NE: Ten to fifteen year assessment of the performance characteristics of the Starr-Edwards Model 6120 mitral valve prosthesis. J THORAC CARDIOVASC SURG 85:1-20, 1983 20 Minale C, Bardos P, Bourg NP, Messmer BJ: Early and late results of porcine bioprostheses versus mechanical prostheses in aortic and mitral positions, Cardiac Bioprostheses. Proceedings of the Second International Symposium, LH Cohn, V Gallucci, eds., New York, 1982, Yorke Medical Books, pp 143-153 21 Schonbeck M, Egloff L, Kugelmeier J, Rothlin M, Senning A, Turina M: Porcine bioprostheses versus mechanical valvular substitutes. A retrospective comparative analysis, Cardiac Bioprostheses. Proceedings of the Second International Symposium, LH Cohn, V Gallucci, eds., New York, 1982, Yorke Medical Books, pp 192204 . 22 Ionescu MI, Tandon AP, Saunders NR, Chidambaram M, Smith DR: Clinical durability of the pericardial xenograft valve: 11 years' experience, Cardiac Bioprosthesis. Proceedings of the Second International Symposium, LH Cohn, V Gallucci, eds., New York, 1982, Yorke Medical Books, pp 42-60 23 Edmunds LH Jr: Thromboembolic complications of current cardiac valvular prostheses. Ann Thorac Surg 34:96105, 1982 24 Hetzer R, Topalidis T, Borst HG: Thromboembolism and anticoagulation after isolated mitral valve replacement with porcine heterografts, Cardiac Bioprostheses. Proceedings of the Second International Symposium. LH Cohn, V Gallucci, eds., New York, 1982, Yorke Medical Books, pp 172-183 25 Ionescu MI, Smith DR, Hasan SS, Chidambaram M, Tandon AP: Clinical durability of the pericardial xenograft valve. Ten years' experience with mitral replacement. Ann Thorac Surg 34:265-277, 1982 26 Baumgartner WA, Miller DC, Reitz BA, Oyer PE, Jamieson SW, Stinson EB, Shumway NE: Surgical treatment of prosthetic valve endocarditis. Ann Thorac Surg 35:87-104, 1983 27 Magilligan OJ, Lewis JW, Heinzerling RH, Smith 0: Fate of a second porcine bioprosthetic valve. J THORAC CARDIOVASC SURG 85:326-370, 1983 28 Magilligan OJ Jr, Lewis JW Jr, Stein PO, Lakier J, Smith DR: Decreasing incidence of porcine bioprosthetic degeneration, Cardiac Bioprostheses. Proceedings of the Second International Symposium, LH Cohn, V Gallucci, eds.,

The Journal of Thoracic and Cardiovascular Surgery

332 Jamieson et al.

29

30

31

32

33

34

35

36

37

38

39

40

41

42

New York, 1982, Yorke Medical Books, pp 559-570 Fiddler GI, Gerlis LM, Walker DR, Scott 0, Williams GJ: Calcification of glutaraldehyde-preserved porcine and bovine xenograft valves in young children. Ann Thorac Surg 35:257-261, 1983 Human DG, Joffe HS, Fraser CB, Barnard CN: Mitral valve replacement in children. J THORAC CARDIOVASC SURG 83:873-877, 1982 Silver MM, Pollock J, Silver MD, Williams WG, Trusler GA: Calcification in porcine xenograft valves in children. Am J Cardiol 45:685-689, 1980 Williams DB, Danielson GK, McGoon DC, Puga FJ, Mair DO, Edwards WD: Porcine heterograft valve replacement in children. J THORAC CARDIOVASC SURG 84:446-450, 1982 Gardner TJ, Roland JMA, Neill CA, Donahoo JS: Valve replacement in children. A fifteen year perspective. J THORAC CARDIOVASC SURG 83: 178-185, 1982 Odell JA: Calcification of porcine bioprostheses in children, Cardiac Bioprostheses. Proceedings of the Second International Symposium, LH Cohn, V Gallucci, eds., New York, 1982, Yorke Medical Books, pp 231-237 Villani M, Bianchi T, Vanini V, Tiraboschi R, Crupi GC, Pezzica E, Parenzan L: Bioprosthetic valve replacement in children, Cardiac Bioprostheses. Proceedings of the Second International Symposium, LH Cohn, V Gallucci, eds., New York, 1982, Yorke Medical Books, pp 248-255 Karp RB, Cyrus RJ, Blackstone EH, Kirklin JW, Kouchoukos NT, Pacifico AD: The Bjork-Shiley valve. Intermediate-term follow-up. J THORAC CARDIOVASC SURG 81:602-614, 1981 Dale J: Arterial thromboembolic complications in patients with Bjork-Shiley and Lillehei-Kaster aortic disc valve prostheses. Am Heart J 93:715-722, 1977 Zwart HHJ, Hicks G, Schuster B, Nathan M, Labrah F, Wenzke F, Ahmed T, De Wall RA: Clinical experience with the Lillehei-Kaster valve prostheses. Ann Thorac Surg 28:158-165, 1979 Carpentier A, Nashref A, Carpentier S, Goussef N, Reiland J, Levy RJ, Fishbein MC, Asmar BEl, Benomar M, Sayed S El, Gouge-Donzeau PG: Prevention of tissue valve calcification by chemical techniques, Cardiac Bioprostheses. Proceedings of the Second International Symposium, LH Cohn, V Gallucci, New York, 1982, Yorke Medical Books, pp 320-330 Wright JTM, Eberhardt CE, Gibbs ML, Saul T, Gilpin CB: Hancock II: An improved bioprosthesis, Cardiac Bioprostheses. Proceedings of the Second International Symposium, LH Cohn, V Gallucci, eds., New York, 1982, Yorke Medical Books, pp 425-444 Angell WW, Angell JD, Kosek JC: Twelve-year experience with glutaraldehyde-preserved porcine xenografts. J THORAC CARDIOVASC SURG 83:493-502, 1982 Carpentier A, Lane E, Carpentier S, Deloche A, Chauvaud S, Maxwell S: Continuing improvements in valvular bioprostheses. J THORAC CARDIOVASC SURG 83:27-42, 1982

Discussion DR. D. CRAIG MILLER Stanford. Calif.

I would like to expand on this eloquent presentation. Dr. Jamieson and his collaborators have presented a value paper of truly exemplary standards. This paper should stand as a model in the assessment of valves and valve-related complications. The authors have used comprehensive definitions and have employed multiple decrement analysis when considering multiple events. This paper is truly the 5 year benchmark for this particular brand of xenograft bioprosthesis. It contains n6W information but confirms what almost everyone has previously assumed-that this valve is durable for upwards of 5 years. The primary tissue failure rate is low during this first 5 year interval; furthermore.eight patients with primary tissue failure underwent reoperation without mortality. This low risk should now be the rule in most hospitals for elective reoperation, and it is an important factor when weighing the tradeoff in durability with a tissue valve (i.e., the risk of reoperation) against the cumulative incidence of valve-related complications with a mechanical valve. The percentage of patients free from fatal valve-related complications is also extremely low at 6 years; however, the real question lies in the 10 to 15 year follow-up interval. Comparison of patients with mechanical and tissue valves at only 5 years artifactually biases the results in favor of the tissue valve. To offer a glimpse of the 10 year expectations with another brand of porcine xenograft, the valve failure rates appear to be identical for the aortic and the mitral valves. At 10 years we expect approximately 80% of patients to be free from primary tissue failure and 70% to be free from valve failure due to all causes (including endocarditis). We interpret these fmdings from Stanford to continue to constitute compelling arguments for the use of porcine bioprostheses in essentially all adult patients. DR. GEORGE MILLER Sacramento, Calif.

Our experience with this valve is limited to 2 years. During this time, we have implanted 190 Carpentier-Edwards valves, most of them in the mitral position. We know of no primary tissue failure, and we have removed only one, because of endocarditis. We began using this valve because of the rather high incidence of subvalvular fibrosis developing at 6 to 10 years in the Smeloff mitral valve. The Smeloff aortic valve, however, has remained durable and relatively trouble-free. We have had some experience in removing previously placed valves, and we have come upon a method we believe worth reporting. It is superior from the standpoint of preservation of all the annular tissue as well as prevention of shredding and particulate embolization of remaining sewing ring, particularly the filler material in the Hancock sewing ring. The key is the fact that the tissue growth from the anulus to the valve sewing ring has little tensile strength or actual ingrowth into the sewing ring. The valve continues to be held primarily by the sutures, even at 8 to 10 years. Thus, each suture knot is pulled up with a clamp and cut separately.

Volume 88 Number 3 September, 1984

Carpentier-Edwards porcine bioprosthesis 3 3 3

The neoendocardial peel, which is not particularly adherent to the sewing ring, is rather easily defined. This peel is begun near the edge of the sewing ring, and the valve then is simply peeled out of the anulus without damage either to the remaining anulus or to the sewing ring. We make no claim as to originality, but we have found this technique to be quite effective.

I believe that we are at the crossroads in the assessment of the durability and function of these two bioprostheses. Perhaps we are observing a real clinical race between these two tissue valves, which have been commercially available for a similar length of time. Dr. Miller brought up the most important point-we have to see how they behave in the years to come.

DR. LORENZO GONZALEZ-LAVIN

DR. JAMIESON (Closing) I thank the discussants for their favorable comments. I am especially appreciative of Dr. Craig Miller's remarks regarding our assessment of valve performance. This multiple decrement analysis of valve performance, as stated by Dr. Miller, should be utilized in the evaluation of all types of prosthetic valves. The Stanford University experience with the Hancock porcine prosthesis at 5 years is similar to ours with the CarpentierEdwards porcine valve. The low incidence of valve-related complications gives tissue valves superior performance characteristics at 5 to 7 years. Durability of tissue valves will determine the superiority of performance of tissue and mechanical prostheses at 10 to 15 years. I extend my appreciation to Dr. George Miller for relating his experience. I agree with Dr. Gonzales-Lavin that the lonescu-Shiley pericardial valve is an acceptable prosthesis. Mr. lonescu has reported a very low incidence of thromboembolism. The incidence of tissue failure is essentially the same with porcine and pericardial valves. The improvements of low-pressure fixation and flexible stents in all tissue valves should improve long-term durability. We will continue to evaluate our valve performance and bring forward future assessments.

Palo Alto, Calif

The combined clinical series of Vancouver and Montreal with the Carpentier-Edwards bioprosthesis is impressive. This is an important report, mainly for those of us interested in the long-term function of tissue valves. In addition to using porcine valves, we have routinely utilized the lonescu-Shiley valve in the aortic position. Between February, 1977, and March, 1983,216 patients have been operated upon at the Ingham Medical Center in Michigan and the Palo Alto Medical Foundation in California. The follow-up is now between 3 months and 6 years, with a mean of 30 months. We have found the durability of the bovine pericardial valve to be outstanding. In the entire series, there have been only two instances of intrinsic tissue failure, both due to ectopic calcification, for 0.4 episodes per 100 patient-year. Of interest, in our series calcification has not occurred in the younger age group. Actuarially, 96.5% of patients are free from intrinsic tissue failure. There have not been any valve-related deaths in the absence of infective endocarditis. In addition to the good durability, the incidence of thromboembolism (these are aortic valves) without the routine use of anticoagulants has been very low at 0.7 episodes per 100 patient-years. Clinically, all surviving patients are in Class I and II, even those with valve sizes less than 21 mm in diameter.

Bomtd volumes available to subscribers Bound volumes of THE JOURNAL Of THORACIC AND CARDIOVASCULAR SURGERY are available to subscribers (only) for the 1984 issues from the Publisher, at a cost of $58.50 ($73.00 international) for Vol. 87 (January-June) and Vol. 88 (July-December). Shipping charges are included. Each bound volume contains a subject and author index and all advertising is removed. Copies are shipped within 30 days after publication of the last issue of the volume. The binding is durable buckram with the JOURNAL name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact The C. V. Mosby Company, Circulation Department, 11830 Westline Industrial Drive, St. Louis, Missouri 63146, USA; phone (800) 325-4177, ext. 351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular JOURNAL subscription.