- Email: [email protected]
Isolated Mitral Valve Replacement with the Hancock Bioprosthesis: A 13-Year Appraisal
Isolated Mitral Valve Replacement with the Hancock Bioprosthesis: A 13-Year Appraisal Vincenzo Gallucci, M.D., Uberto Bortolotti, M.D., Aldo Milano, M.D., Carlo Valfrk, M.D., Alessandro Mazzucco, M.D., and Gaetano Thiene, M.D. ABSTRACT Four hundred seventy-six patients underwent isolated mitral valve replacement (MVR) with the glutaraldehyde-preserved porcine Hancock bioprosthesis from March, 1970, through December, 1981. There were 312 female and 164 male patients ranging in age at operation from 9 to 68 years (average, 53 years). Associated surgical procedures were performed in 35 patients. Hospital mortality was 13%, the main cause of death being lowoutput syndrome. The survivors were followed from 1.6 to 13.2 years (mean, 5.2 years). Cumulative duration of follow-up is 2,180 patient-years and is 97% complete. Overall late mortality is 3.1 f 0.3% per patient-year, and actuarial survival is 73.8 f 3.4% at 13 years. Embolic accidents occurred in 45 patients and were fatal in 13; the linearized incidence of postoperative systemic thromboemboli is 2.1 f 0.3% per patient-year. Reoperation was necessary in 49 patients: in 4 because of valve endocarditis, with 3 deaths; in 6 because of perivalvular leak, with no deaths; in 2 because of left atrial thrombosis; and in 37 because of valve dysfunction due to primary tissue failure, caused mainly by calcific degeneration of the tissue, with 5 operative deaths. Actuarial freedom from primary tissue failure is 58 & 6.6% at 13 years. Extended follow-up after MVR with the Hancock bioprosthesis confirms the satisfactory performance and low thrombogenicity of this device up to 13 years after operation. Calcific degeneration of the cusps remains the chief cause of valve failure leading to reoperation and may represent a major problem. Therefore, until further data on the prevention of the calcific process are available from ongoing experimental studies, we will use the Hancock bioprosthesis for MVR in patients selected according to age and specific situations.
In our institution, the first glutaraldehyde-processed Hancock porcine bioprosthesis was implanted in March, 1970. Our early and medium-term results indicated this device had a low thrombogenicity and an acceptable hemodynamic performance, which led us to share the iniFrom the Departments of Cardiovascular Surgery and Pathology, University of Padova Medical School, Padova, Italy. Presented at the Twentieth Annual Meeting of The Society of Thoracic Surgeons, San Antonio, TX, Jan 23-25, 1984. Address reprint requests to Dr. Gallucci, Centro di Chirurgia Cardiovascolare, Universite di Padova, Via Giustiniani, 2, 35100 Padova, Italy.
571
tial enthusiasm of others and justified its extensive clinical use [l, 21. A limited durability, however, soon appeared to be the major drawback of the prosthesis, rendering it unsuitable for particular patients, such as children and young people [3-51. Data on patients followed for 10 years or longer are currently still scarce; hence, the actual long-term durability of the porcine bioprosthesis remains to be clearly defined. We review here our 13-year experience with the Hancock bioprosthesis in patients undergoing isolated mitral valve replacement (MVR).
Material and Methods From March, 1970, through December, 1981, 476 patients underwent isolated MVR with the Hancock bioprosthesis. There were 312 female and 164 male patients ranging in age from 9 to 68 years (average, 53 years). Associated surgical procedures, predominantly tricuspid annuloplasty and coronary artery bypass grafting, were performed in 35 patients (7.4%).The cause of valvular disease was rheumatic in 87.6% of patients, degenerative in 6.3%, congenital in 1.7%, ischemic in 1.3%, and infective (endocarditis) in 0.6%; 2.5% of the patients required replacement of a failing prosthesis previously implanted. Preoperatively, 26 patients were classified in New York Heart Association (NYHA) Functional Class 11, 349 in Class 111, and 101 in Class IV.
Surgical Technique All patients underwent MVR through a median sternotomy incision under standard cardiopulmonary bypass with moderate to deep systemic hypothermia. Initially, interrupted aortic cross-clamping with coronary perfusion and topical cooling was employed, but this method was replaced subsequently by a single aortic cross-clamp time. Cold potassium cardioplegia has been used routinely since 1977. The surgical technique has remained quite uniform throughout the period of study. The Hancock bioprosthesis was inserted using multiple interrupted sutures of nonabsorbable material, which were buttressed by Teflon pledgets when a calcified annulus was found. Anticoagulant Treatment Oral anticoagulation was started on removal of the chest tubes, usually on the second postoperative day. Subsequently, all patients remained on a regimen of coumarin derivatives (warfarin sodium) for 3 months. Then oral administration of anticoagulants was discontinued, regardless of the heart size and rhythm. In recent years, a more restrictive policy has been adopted. Long-term
572 The Annals of Thoracic Surgery Vol 38 No 6 December 1984
anticoagulation is currently maintained in patients with a giant left atrium, chronic atrial fibrillation, or left atrial thrombi at operation and in those with a combination of these findings.
Valve Failure and Thromboembolism In defining valve failure, we adopted the criteria used by the Stanford group [6, 71. Primary tissue failure is that due to cusp tears, perforation, 01' calcification in the absence of infection, as established at reoperation or at postmortem examination. Thromboembolic events were classified as being major or minor. Those resulting in peripheral acute arterial obstruction or causing permanenl neurological damage constituted major events, and those associated with temporary neurological deficits without long-term sequelae constituted minor events. Follow-Up Reevaluation of operative survivors was carried out during the 6-month interval betwcsen January and June, 1983. Information on current clinical status and postoperative complications was gather1.d in most instances by direct evaluation and interview or through questionnaires answered by the patient or the referring cardiologist. All explanted bioprostheses, whether removed at reoperation or at postmortem exe mination, underwent a detailed morphological and radiological investigation, as described elsewhere [8-lo]. Statistical Analysis Analysis of the rate of patient survival, the incidence of postoperative thromboembolic events, and the incidence of valve failure has been expressed in an actuarial fashion [ll], reporting the 70% confidence limits and excluding operative deaths. The linearized rate of thromboembolic events, overall valve dysfunction, and primary tissue failure was calculated on individual yearly postoperative intervals rather than on the entire follow-up period [12].
Results Operative Mortality and Follow-up There were 63 operative death!; (13%). Most of these patients were in NYHA Class N and died of low-output syndrome. Of the 413 patients discharged from the hospital, 69 died in the late postoperative period. This corresponds to a linearized rate of 3.1 ? 0.3% per patientyear. Those patients surviving clperation were followed from a minimum of 1.6 to a maximum of 13.2 years; 292 patients were followed for 5 years and 15, for 13 years. The mean follow-up was 5.2 years with a cumulative duration of follow-up of 2,180 patient-years, which is 97% complete (0.5 lost to follow-up per patient-year). Actuarial survival is 87 f 1.5%at 5 years after operation and 73.8 f 3.4% at 13 years (Fig 1). The functional classification of the 285 current survivors is as follows: 80
1001
-
411 392 364 330 292 246
1
2
3
4
5
6
YEARS
199 148 100 69
7
8
9
53
36
I5
1011 1 2 1 3
POSTOP
Fig 1. Actuarial survival following isolated mitral valve replacement with the Hancock xenograft. Operative deaths are not included. Numbers on the abscissa indicate the patients at risk.
patients in NYHA Functional Class I, 176 in Class 11, and 29 in Class 111.
Overall Valve Dysfunction and Reoperations Bioprosthetic valve dysfunction was observed in 51 patients (linearized rate of 2.3 f 0.3% per patient-year). Forty-nine underwent reoperation, and 2 died of valve failure. Reoperation was required because of primary tissue failure in 37 patients, prosthetic valve endocarditis in 4, paravalvular leak in 6, and left atrial thrombosis in 2. The linearized rate of reoperation for each of these causes in the entire follow-up period was as follows: primary tissue failure, 1.7 f 0.3%per patient-year; prosthetic valve endocarditis, 0.2 0.9% per patient-year; perivalvular leak, 0.3 0.1% per patient-year; and left atrial thrombosis, 0.1 +- 0.1%per patient-year. Primary tissue failure was caused predominantly by tissue calcification, leading either to stenosis due to cusp stiffening or to incompetence because of commissural detachment and tears, or a combination of both (Figs 24). Calcific degeneration, therefore, represents the leading cause of primary tissue failure in our series [lo]. Reoperation for primary tissue failure was undertaken an average of 8 years after MVR (range, 2.8 to 11.6 years). There were 5 operative deaths (13.5%). Prosthetic valve endocarditis occurred in 7 patients (linearized rate of 0.3 f 0.1%per patient-year). Reoperation was done in 4 patients at an average interval of 1.8 years after MVR (range, 0.2 to 3.6 years). There were 3 deaths. Of the remaining 3 patients, 1 died of valve incompetence and 2 were successfully treated with antibiotics. Reoperation for perivalvular leak was performed in 6 patients at an average of 2.4 years after MVR (range, 0.1 to 7.0 years). There were no deaths. Two patients underwent reoperation because of left atrial thrombosis, and both died. Another patient died of a cerebral embolism; at postmortem examination, the valve appeared stenotic due to thrombosis of two sinuses. The actuarial probability of being free from all causes of bioprosthetic valve dysfunction (Fig 5) and from primary tissue failure (Fig 6) are 56.8 6.5% and 58 f
*
*
*
573
Gallucci et al: Isolated MVR with Hancock Bioprosthesis
B
A
Fig 2. (A)Atrial v i m of a mitral porcine xenograft explanted after 133 months. Note the fibrous tissue overgrowth causing severe reduction of the effective orifice area, and the calcificfragmentation of one cusp. (B) Histological section of one cusp from the same device is covered by a thick, fibrous pannus on the atrial aspect. (H&E; x 12 before 5% reduction.)
6.6%,respectively, at 13 years. Figures 5 and 6 also show the linearized rate of overall bioprosthetic valve dysfunction and primary tissue failure calculated yearly after operation. An attempt was made to determine whether there was a difference in durability depending on the year of valve implantation. At 8 years, the behavior of valves implanted in 1971, 1972, 1974, and 1975 did not differ significantly; those implanted in 1970 and 1973 showed a lower durability, but the number of patients entering the study in these 2 years is too small for statistical comparison with the others (Table).
Thromboembolic Complications and Hemorrhage Forty-five thromboembolic events were observed during the follow-up period. This corresponds to a linearized rate of 2.1 0.3% per patient-year. Thirteen of the events were fatal (linearized rate of 0.6 & 0.1% per patient-year). The actuarial probability of being free from a thromboembolic event at 13 years is 79.2 ? 4.0% (Fig 7). Figure 7 also shows the linearized rate of thromboembolism, as calculated on yearly intervals. Thirty-six patients
*
sustaining a thromboembolic event were in atrial fibrillation, and 9 were in sinus rhythm. Fourteen were on a long-term regimen of anticoagulants, and 31 were not. Among the current survivors, 58% are in atrial fibrillation and 42%, in sinus rhythm. As a result of our current policy, 53% of the survivors are on a regimen of anticoagulant treatment. Anticoagulant-related hemorrhage occurred in 9 patients, a linearized rate of 0.4 f 0.1% per patient-year. Four of them died, a linear rate of 0.2 ~fr0.1% per patient-year.
Comment The glutaraldehyde-preserved Hancock bioprosthesis has been the valve of choice in many centers during the last decade [l,2,6, 7, 13-15]. Reports from these institutions have stressed the advantages of this device, such as its low thrombogenicity and acceptable hemodynamics in the large sizes, but they also have discussed its limited long-term durability. Because of these results, a considerable amount of experimental work has been undertaken and has led to the manufacture of new bioprostheses [16, 171. Improvements in stent design and modification of tissue processing techniques may give these new bioprosthetic valves an increased resistance to fatigue and wear. We believe that data that can help in the assessment of the long-term performance of the old generation of bioprostheses, such as the standard Hancock porcine valve, must be considered of paramount
574 The Annals of Thoracic Surgery Vol 38 No 6 December 1984
A
Fig 3. (A)Atrial aspect ofa mitral bioprosthesis explanted after 111 months because of severe stenosis. Calcific stiffening of the cusps is evident. ( B ) Histological section of one cusl7 shows massive calcific deposits within the valvular tissue. (H&E; x 19 before 5% reduction.)
importance. It is with these valves that the new ones will be compared in the future. The present study, based on an extended follow-up, confirms previous observations from our institution and from others [I, 2, 6, 7, 13-15] regarding the satisfactory clinical performance of the Hancock bioprosthesis. In fact, 13 years after MVR, the actuarial survival approximates 74%, a figure that compare:, favorably with those reported by others using the same or other biological valves [6, 7, 14, 181. Since the questionable long-term durability of the porcine xenograft represents its ma lor disadvantage, we have focused particularly on the rate and causes of valve failure. In our experience, dysfurction of the Hancock valve in the mitral position is predominantly due to primary tissue failure, which is a consequence of cusp calcification; in fact, primary tissue failure without evidence of calcific deposits is exceedingly rare [19]. Recently, we [lo] evaluated calcification of porcine bioprostheses on a pathological basis and found that it appears to be a time-related phenomenon affecting not only bioprosthetic valves implanted in children [3-51, as reported in the past, but also thos? inserted in adults [9,
B 10, 20, 211. Since predisposing factors are seldom detected, calcification of the bioprosthesis can be considered an unavoidable stage in the natural history of the standard porcine valve. Therefore, modifications of the stent design, with the aim of reducing stress on the cusps, and the use of anticalcific treatments in tissue processing seem justifiable, although their effectiveness will be demonstrated only after a clinical experience comparable to those achieved with the old models. Valve durability may be affected by the condition of the tissue at the time of valve fixation. In fact, many reports have demonstrated that processed but unimplanted porcine bioprosthetic valves may already show structural changes, such as loss of endothelial lining, that might predispose to plasma protein insudation and eventual calcification [8, 9, 221. Furthermore, Magilligan and colleagues [23] found that bioprostheses implanted early in their experience behave differently from those implanted later. Those used in 1974 had a lower incidence of degeneration compared with those employed in 1972 and 1973, which suggests a more careful manufacture of the former. We were unable to confirm these findings, at least in our MVR population, since valves implanted from 1970 through 1975 had approximately the same rate of failure at 8 years. This excludes those implanted in 1970 and 1973, because the number of patients entering the study in these years was too limited for a meaningful statistical comparison. In our experience, porcine bioprosthetic valves used in the mitral position show a negligible incidence of primary tissue fail-
575 Gallucci et al: Isolated MVR with Hancock Bioprosthesis
B
A
Fig 4. ( A ) Ventricular aspect of a niitral porcine valve recovered at reoperation after 84 months. A calcific vegetation at one commissure with detachment of two cusps caused severe incompetence. ( B ) Roentgenogram of the same device confirms the presence of gross calcific deposits.
-s w I-
100 -
-s1 0 0--
80-
?80-
U
a
a
a
.A
60-
460-
a
U
3 c
$
0-0
3 U
1
1
58 : 6 6
U c
405 3 8 8 361 326 290 245
198 150 100
71
55
39
15
4
403 387 360 326 290 245
A
Fig 5 . (A)Actuarial probability of being free from overall bioprosthetic valve (BP) dysfunction. ( B ) Linearized rate of overall bioprosthetic valve dysfunction during each yearly postoperative interval. Numbers in parentheses indicate the valve failures. (PT-YR = patient-year.)
B
1
2
3
4
5 6
YEARS
PTF
198 150 100 71
7
53
39
15
8 9 1011 1213
POSTOP
EVENTS
Fig 6 . (A)Actuarial rate of primary tissue failure (PTF). (B) Linearized rate of primary tissue failure. The numbers in parentheses indicate the events occurring during each yearly postoperative interval. (IT-YR = patient-year.)
576 The Annals of Thoracic Surgery- Vol 38 No 6 December 1984
Actuarial Rate of Bioprosthesis Dysfzrnction According to Year of Mitral Valve Replacement Year
No. of Patients"
Percent at 8 'Years
No. of Patientsb
1970 1971 1972 1973 1974 1975
17 40 35 10 24 51
75.7 f 12 93.8 f 4.2 100 62.5 2 17 95.3 f 4.5 94.5 2 3.7
11 31 24 8 18 37
"Patients entering the study. bPatients reaching the eighth postoperative year.
=
i
79254 0
60 U
U
A
B
404 382 351
i
2
316 275 228
178 I25
80
5 4 Y E5A R S6 P OiS T O8P 9
TE
54
40
101'1
28
12
1'213
I
EVENTS
Fig 7. (A)Actuarial and ( B ) linearized rates of postoperative thromboembolic (TE) events after mitral valve replacement with a porcine bioprosthesis. The numbers in parentheses indicate the events during each year. (PT-YR = patient-year.)
ure up to the sixth postoperative year, the highest rate of failure at 7 years, and no further acceleration thereafter. These data are in agreement with the observations of others [12]. Reoperation for primary tissue failure is usually performed on an elective basis, the dysfunction of porcine bioprostheses being progressive. IUevertheless, the mortality observed in this series indicates that reoperation in such patients must not be regarded as a trivial problem. Especially in young patients, the surgeon will be faced with the likelihood of repeated reoperations and a progressively increasing operative risk. Therefore, our current trend is to employ the porcine valve in patients selected by age and specific situations [24, 251. Endocarditis can still be considered a rare cause of
valve dysfunction. It occurred in 7 patients in this series, 4 of whom underwent reoperation with a high mortality, as previously reported [26]. In 2 patients infection was controlled and cured by antibiotic treatment. This is a very recent experience for our group and was not reported in our latest survey of patients with a Hancock bioprosthesis [24]. We now believe that possibly the infected porcine xenograft can be sterilized, as previously stressed by others [13, 15, 271. Previous authors [l, 2, 6, 14, 15, 281 maintained that the Hancock bioprosthesis has a low thrombogenicity, and our findings support this claim. In fact, the probability of being free from thromboembolic events 13 years after MVR is quite high, approximating 80%. The majority of these events occurred in the first postoperative years, many in the first postoperative months [l]. This suggests that temporary anticoagulation after MVR with a porcine valve appears reasonable and advisable. However, the need for long-term anticoagulation in this cohort of patients remains a debated issue. Chronic atrial fibrillation is a predisposing factor to left atrial thrombi, an obvious source of systemic emboli [29, 301. Therefore, our current policy, like that of others [14, 151, is to keep patients in atrial fibrillation on a regimen of anticoagulation indefinitely. We do believe that it is difficult to evaluate the role of atrial fibrillation in the incidence of thromboembolism, since more than 50% of our current survivors are on a regimen of anticoagulants. Some of the thromboembolic events in the late postoperative period might be related more to bioprosthesis failure, with embolization of cusp fragments, than to true thromboemboli [31]. In these patients, assessment of prosthetic function seems more important than the institution of anticoagulant treatment. We conclude that the Hancock bioprosthesis in the mitral position continues to perform satisfactorily after 13 years. Calcific degeneration is the main cause of dysfunction requiring reoperation which may represent a major consideration. However, until further data on the prevention of calcification and the extension of bioprosthetic durability are available, we will continue to implant this device in patients selected according to age and specific situations. This work was supported in part by Grant No. 82.02738.04 of the Consiglio Nazionale delle Ricerche, Rome, Italy. The authors thank Agostino Leorin for his technical assistance.
References 1. C6vese PG, Gallucci V, Morea M, et al: Heart valve replacement with the Hancock bioprosthesis:analysis of long-term results. Circulation 56Suppl2111, 1977 2. Casarotto D, Bortolotti U,Thiene G, et al: Long-term results (from 5 to 7 years) with the Hancock-SGP bioprosthesis. J Cardiovasc Surg (Torino) 20:399, 1979 3. Kutsche LM, Oyer PE, Shumway NE, Baum D: An important complication of Hancock mitral valve replacement in children. Circulation 6O:Suppl 1:98, 1979
577 Gallucci et al: Isolated MVR with Hancock Bioprosthesis
4. Saunders SP, Levy RJ, Freed MD, et al: Use of Hancock porcine xenografts in children and adolescents. Am J Cardiol46429, 1980 5. Bortolotti U, Thiene G: Calcification of porcine heterografts implanted in children. Chest 80:117, 1981 6. Oyer PE, Stinson EB, Reitz BA, et al: Long-term evaluation of the porcine xenograft bioprosthesis. J Thorac Cardiovasc Surg 78:343, 1979 7. Oyer PE, Miller DC, Stinson EB, et a1 Clinical durability of the Hancock porcine bioprosthesis. J Thorac Cardiovasc Surg 80824, 1980 8. Thiene G, Arbustini E, Bortolotti U, et al: Pathological substrates of porcine valve dysfunction. In Cohn LH, Gallucci V (eds): Cardiac Bioprostheses. New York, Yorke Medical, 1982, p 378 9. Valente M, Bortolotti U, Arbustini E, et al: Glutaraldehydepreserved porcine bioprosthesis: factors affecting performance as determined by pathologic studies. Chest 83:607, 1983 10. Milano A, Bortolotti U, Talenti E, et al: Calcific degeneration as the main cause of porcine bioprosthetic valve failure. Am J Cardiol 53:1066, 1984 11. Anderson RP, Bonchek LI, Grunkemeier GE, et al: The analysis and presentation of surgical results by actuarial methods. J Surg Res 16:224, 1974 12. Oyer PE, Stinson EB, Miller DC, et al: Clinical analysis of the Hancock porcine bioprosthesis. In Cohn LH, Gallucci V (eds): Cardiac Bioprostheses. New York, Yorke Medical, 1982, p 539 13. Davila JC, Magilligan DJ Jr, Lewis JW Jr: Is the Hancock porcine valve the best cardiac valve substitute today? Ann Thorac Surg 26:303, 1978 14. Cohn LH, Mudge GH, Pratter F, Collins JJ Jr: Five- to eightyear follow-up of patients undergoing porcine heart valve replacement. N Engl J Med 304258, 1981 15. Gallo JI, Ruiz B, Carrion MF, et al: Heart valve replacement with the Hancock bioprosthesis: a 6-year review. Ann Thorac Surg 31:444, 1981 16. Gabbay S, Frater RWM: The unileaflet heart valve bioprosthesis: new concept. In Cohn LH, Gallucci V (eds): Cardiac Bioprostheses. New York, Yorke Medical, 1982, p 411 17. Wright JTM, Eberhardt CE, Gibbs ML, et al: Hancock 11: an improved bioprosthesis. In Cohn LH, Gallucci V (eds): Cardiac Bioprostheses. New York, Yorke Medical, 1982, p 425 18. Ionescu MI, Tandon AP, Saunders NR, et al: Clinical durability of the pericardial xenograft valve: 11 years’ experience. In Cohn LH, Gallucci V (eds): Cardiac Bioprostheses. New York, Yorke Medical, 1982, p 42 19. Arbustini E, Bortolotti U, Valente M, et al: Cusp disruption by massive lipid infiltration: a rare cause of porcine valve dysfunction. J Thorac Cardiovasc Surg 84:738, 1982 20. Cipriano PR, Billingham ME, Oyer PE, et al: Calcification of porcine prosthetic heart valves: a radiographic and light microscopic study. Circulation 66:1100, 1982 21. Schoen FJ, Collins JJ Jr, Cohn LH: Long-term failure and morphologic correlation in porcine bioprosthetic heart valves. Am J Cardiol 51:957, 1983 22. Ferrans VJ, Spray TL, Billingham ME, Roberts WC: Structural changes in glutaraldehyde-treated porcine heterografts used as substitute cardiac valves: transmission and scanning electron microscopic observations in 12 patients. Am J Cardiol41:1161, 1978 23. Magilligan DJ Jr, Lewis JW Jr, Stein I’D, et al: Decreasing incidence of porcine bioprosthetic degeneration. In Cohn
24.
25. 26. 27. 28. 29.
30. 31.
LH, Gallucci V (eds): Cardiac Bioprostheses. New York, Yorke Medical, 1982, p 559 Gallucci V, Valfre C, Mazzucco A, et a1 Heart valve replacement with the Hancock bioprosthesis: a 5-11 year followup. In Cohn LH, Gallucci V (eds): Cardiac Bioprostheses. New York, Yorke Medical, 1982, p 9 Bortolotti U, Milano A, Mazzucco A, et al: Pregnancy in patients with a porcine bioprosthesis. Am J Cardiol501051, 1982 Bortolotti U, Thiene G, Milano A, et al: Pathological study of infective endocarditis on Hancock porcine bioprostheses. J Thorac Cardiovasc Surg 81:934, 1981 Magilligan DJ Jr, Quinn EL, Davila JC: Bacteremia, endocarditis, and the Hancock valve. Ann Thorac Surg 24:508,1977 Borkon AM, McIntosh CL, Von Rueden TJ, Morrow AG: Mitral valve replacement with the Hancock bioprosthesis: five- to ten-year follow-up. Ann Thorac Surg 32:127, 1981 Thiene G, Bortolotti U, Panizzon G, et al: Pathological substrates of thrombus formation after heart valve replacement with the Hancock bioprosthesis. J Thorac Cardiovasc Surg 80:414, 1980 Edmiston WA, Harrison EC, Duick GF, et al: Thromboembolism in mitral porcine valve recipients. Am J Cardiol 41:508, 1978 Bortolotti U, Milano A, Thiene G, et al: Evidence of impending embolization of a calcific cusp fragment from a mitral porcine xenograft. Thorac Cardiovasc Surg 30:405, 1982
Discussion DR. LAWRENCE H. COHN (Boston, MA): As far as I know, Dr. Gallucci and co-workers were the first to implant the present glutaraldehyde-preserved porcine valve. Their experience is obviously the longest in the world, and Dr. Gallucci has made a very excellent presentation. I think, based on really very meaningful statistical data, we now have some conclusions at the 10year level. I am not sure that on review of the material we can claim to have similarly meaningful 13-year data, because the last points on Dr. Gallucci’s curves show small numbers of patients at risk and the standard error of the mean is higher than we would like to be statistically correct about conclusions at that point. The first point of discussion concerns operative mortality. This has declined, as it has in everybody’s hands, over the past several years. Our series, which extends from 1972, is fairly comparable in terms of numbers to that of Dr. Gallucci. A major difference is that we have employed coronary bypass grafting in about 22% of our patients, roughly about three times the incidence in Dr. Gallucci’s series. I wonder if he would comment on his group’s present approach to coronary angiography in patients with mitral valve disease and whether or not he thinks it may have something to do with the recent improvements in operative mortality. The rate of thromboembolism with this valve is remarkably consistent throughout the world. There are a number of series-those from Stanford and Paris as well as our own-that document about 2 embolic events per year of follow-up. We do have a marked increase in emboli in patients who are in atrial fibrillation. Our data on valve dysfunction with the porcine valve include primary valve dysfunction, perivalvular leak, and bacterial endocarditis. Our final figure of 1.5 events per patient-year does not reflect the “total valve failure” suggested by the Stanford
578 The Annals of Thoracic Surgery Vol 38 No 6 December 1984
group, because we have not included the thromboembolic rate in this calculation. I think that is a good concept, and I congratulate Dr. Gallucci for bringing it .o our attention. There are a few differencesin primary dysfunction rates between our series and that of Dr. Gallucci. Primary tissue failure is not strictly linear, as he has suggested. For instance, 9 patients whom we operated on in 1972 are at risk for valve dysfunction, but it has not occurred in m y of them to date. Yet, from the next year, 3 or 4 patients out of a group at risk of about 25 have had tissue failure. I think Don Magilligan first pointed out that this is not a strictly time-related phenomenon and that it may reflect variability in fabrication techniques in the early days of the manufacture of this valve. Perhaps Dr. Gallucci would comment on this. Finally, Dr. Gallucci, could you tell us how many patients in whom dysfunction developed were either children or less than 35 years of age? Our overall age rang: is considerably older, as it is with most series from the United States compared with those from Europe. Perhaps this is one of the reasons why dysfunction rates vary. This is an excellent analysis, and I was very pleased to have the opportunity to discuss it. w. R. ERIC JAMIESON (Vancouver, BC, Canada): I appreciate the opportunity to discuss Professor Gallucci’s paper, an excellent presentation documenting the longest series in the world with the Hancock porcine bioprosthesis. At the University of British Ccilumbia, the CarpentierEdwards porcine bioprosthesis has been utilized since 1975, and there have been approximately 1,700 implants. I will briefly compare a series of 423 MVRs done between 1975 to 1981 with Dr. Gallucci’s series of 476 done between 1970 and 1981. The 5year survival is higher in the Carpentier-Edwards series (80% versus 71%) but probably reflects the lower operative mortality in this series of implants, which were performed in the latter half of the decade with advanced myocardial preservation. The valve-related complications are relatively similar in both groups, especially with regard to per prosthetic leaks and prosthetic valve endocarditis. The incider ce of thromboembolism is lower in the Carpentier-Edwards series (1.6% versus 2.1% per patient-y ear). There have been 6 instances of prmary tissue failure in our series, with no deaths at reoperation. The degeneration was primarily due to tears and not to calc fication. Dr. Gallucci indicated that his group’s failures were due to calcification. Did tears and perforations accompany the calcification? I also wonder whether Dr. Gallucci considered utilizing the multiple decrement analysis method advocated by Bodnar and colleagues for evaluation of valve performance. The first assessment is the freedom from all valve-related complications (thromboembolism, anticoagulant hemorrhage, prosthetic DR.
valve endocarditis, periprosthetic leak, and primary valve failure). This denotes the actuarial proportion of patients who remain free from any valve-related complications, including those causing death or reoperation. The other two assessments are freedom from mortality related to valve-related complications and freedom from mortality or reoperation related to valverelated complications. In evaluation of patients with the Carpentier-Edwards prosthesis with follow-up intervals over 5 years, the freedom from all valve-related complications was 84%, the freedom from valve-related mortality and reoperation was 92%, and the freedom from valve-related mortality was 97%. Using this methodology, Dr. Craig Miller of Stanford University presented data at a meeting of the American College of Surgeons showing that at the 10-year period, the Hancock porcine bioprosthesis performed better than a standard mechanical prosthesis. DR. GALLUCCI: Thank you very much, Dr. Cohn and Dr. Jamieson, for your remarks. With reference to Dr. Cohn’s observations, the current operative mortality for MVR is around 6 or 7%. This improvement is probably explained by changes in surgical techniques and also by the introduction of cardioplegia. As far as coronary bypass grafting is concerned, in our center we depend mainly on what the cardiologists report to us. We currently require a more careful examination, which includes coronary angiography. In general, we now ask that coronary angiography be done in every patient with valvular disease who is older than 55 years. In the past, this was not routinely performed, and I agree with Dr. Cohn that probably some patients might have died because of episodes of acute coronary insufficiency. The number of thromboembolic episodes are just about the same in Dr. Cohn’s series and our series. Definitely such episodes are more frequent in patients in atrial fibrillation. We took into consideration the difference in behavior of the Hancock valve according to the year of implantation and tried to trace a curve of it. We are aware that the group at the Henry Ford Hospital in Detroit found a difference in durability in bioprostheses depending on the year of implantation. However, in examining the valves implanted from 1970 to 1976, we could not find such a difference, as I mentioned in the presentation. As for the age of the patients, a few of those operated on at the beginning of our experience were children, usually with complex congenital malformations, and all of them had to undergo reoperation because the valve did not last more than 2 or 3 years. Dr. Jamieson, all of the explanted valves undergo roentgenographic, gross, histological, and electron microscopic examination. We have found that almost all valves with some kind of dysfunction had cusp calcification.
In our institution, the first glutaraldehyde-processed Hancock porcine bioprosthesis was implanted in March, 1970. Our early and medium-term results indicated this device had a low thrombogenicity and an acceptable hemodynamic performance, which led us to share the iniFrom the Departments of Cardiovascular Surgery and Pathology, University of Padova Medical School, Padova, Italy. Presented at the Twentieth Annual Meeting of The Society of Thoracic Surgeons, San Antonio, TX, Jan 23-25, 1984. Address reprint requests to Dr. Gallucci, Centro di Chirurgia Cardiovascolare, Universite di Padova, Via Giustiniani, 2, 35100 Padova, Italy.
571
tial enthusiasm of others and justified its extensive clinical use [l, 21. A limited durability, however, soon appeared to be the major drawback of the prosthesis, rendering it unsuitable for particular patients, such as children and young people [3-51. Data on patients followed for 10 years or longer are currently still scarce; hence, the actual long-term durability of the porcine bioprosthesis remains to be clearly defined. We review here our 13-year experience with the Hancock bioprosthesis in patients undergoing isolated mitral valve replacement (MVR).
Material and Methods From March, 1970, through December, 1981, 476 patients underwent isolated MVR with the Hancock bioprosthesis. There were 312 female and 164 male patients ranging in age from 9 to 68 years (average, 53 years). Associated surgical procedures, predominantly tricuspid annuloplasty and coronary artery bypass grafting, were performed in 35 patients (7.4%).The cause of valvular disease was rheumatic in 87.6% of patients, degenerative in 6.3%, congenital in 1.7%, ischemic in 1.3%, and infective (endocarditis) in 0.6%; 2.5% of the patients required replacement of a failing prosthesis previously implanted. Preoperatively, 26 patients were classified in New York Heart Association (NYHA) Functional Class 11, 349 in Class 111, and 101 in Class IV.
Surgical Technique All patients underwent MVR through a median sternotomy incision under standard cardiopulmonary bypass with moderate to deep systemic hypothermia. Initially, interrupted aortic cross-clamping with coronary perfusion and topical cooling was employed, but this method was replaced subsequently by a single aortic cross-clamp time. Cold potassium cardioplegia has been used routinely since 1977. The surgical technique has remained quite uniform throughout the period of study. The Hancock bioprosthesis was inserted using multiple interrupted sutures of nonabsorbable material, which were buttressed by Teflon pledgets when a calcified annulus was found. Anticoagulant Treatment Oral anticoagulation was started on removal of the chest tubes, usually on the second postoperative day. Subsequently, all patients remained on a regimen of coumarin derivatives (warfarin sodium) for 3 months. Then oral administration of anticoagulants was discontinued, regardless of the heart size and rhythm. In recent years, a more restrictive policy has been adopted. Long-term
572 The Annals of Thoracic Surgery Vol 38 No 6 December 1984
anticoagulation is currently maintained in patients with a giant left atrium, chronic atrial fibrillation, or left atrial thrombi at operation and in those with a combination of these findings.
Valve Failure and Thromboembolism In defining valve failure, we adopted the criteria used by the Stanford group [6, 71. Primary tissue failure is that due to cusp tears, perforation, 01' calcification in the absence of infection, as established at reoperation or at postmortem examination. Thromboembolic events were classified as being major or minor. Those resulting in peripheral acute arterial obstruction or causing permanenl neurological damage constituted major events, and those associated with temporary neurological deficits without long-term sequelae constituted minor events. Follow-Up Reevaluation of operative survivors was carried out during the 6-month interval betwcsen January and June, 1983. Information on current clinical status and postoperative complications was gather1.d in most instances by direct evaluation and interview or through questionnaires answered by the patient or the referring cardiologist. All explanted bioprostheses, whether removed at reoperation or at postmortem exe mination, underwent a detailed morphological and radiological investigation, as described elsewhere [8-lo]. Statistical Analysis Analysis of the rate of patient survival, the incidence of postoperative thromboembolic events, and the incidence of valve failure has been expressed in an actuarial fashion [ll], reporting the 70% confidence limits and excluding operative deaths. The linearized rate of thromboembolic events, overall valve dysfunction, and primary tissue failure was calculated on individual yearly postoperative intervals rather than on the entire follow-up period [12].
Results Operative Mortality and Follow-up There were 63 operative death!; (13%). Most of these patients were in NYHA Class N and died of low-output syndrome. Of the 413 patients discharged from the hospital, 69 died in the late postoperative period. This corresponds to a linearized rate of 3.1 ? 0.3% per patientyear. Those patients surviving clperation were followed from a minimum of 1.6 to a maximum of 13.2 years; 292 patients were followed for 5 years and 15, for 13 years. The mean follow-up was 5.2 years with a cumulative duration of follow-up of 2,180 patient-years, which is 97% complete (0.5 lost to follow-up per patient-year). Actuarial survival is 87 f 1.5%at 5 years after operation and 73.8 f 3.4% at 13 years (Fig 1). The functional classification of the 285 current survivors is as follows: 80
1001
-
411 392 364 330 292 246
1
2
3
4
5
6
YEARS
199 148 100 69
7
8
9
53
36
I5
1011 1 2 1 3
POSTOP
Fig 1. Actuarial survival following isolated mitral valve replacement with the Hancock xenograft. Operative deaths are not included. Numbers on the abscissa indicate the patients at risk.
patients in NYHA Functional Class I, 176 in Class 11, and 29 in Class 111.
Overall Valve Dysfunction and Reoperations Bioprosthetic valve dysfunction was observed in 51 patients (linearized rate of 2.3 f 0.3% per patient-year). Forty-nine underwent reoperation, and 2 died of valve failure. Reoperation was required because of primary tissue failure in 37 patients, prosthetic valve endocarditis in 4, paravalvular leak in 6, and left atrial thrombosis in 2. The linearized rate of reoperation for each of these causes in the entire follow-up period was as follows: primary tissue failure, 1.7 f 0.3%per patient-year; prosthetic valve endocarditis, 0.2 0.9% per patient-year; perivalvular leak, 0.3 0.1% per patient-year; and left atrial thrombosis, 0.1 +- 0.1%per patient-year. Primary tissue failure was caused predominantly by tissue calcification, leading either to stenosis due to cusp stiffening or to incompetence because of commissural detachment and tears, or a combination of both (Figs 24). Calcific degeneration, therefore, represents the leading cause of primary tissue failure in our series [lo]. Reoperation for primary tissue failure was undertaken an average of 8 years after MVR (range, 2.8 to 11.6 years). There were 5 operative deaths (13.5%). Prosthetic valve endocarditis occurred in 7 patients (linearized rate of 0.3 f 0.1%per patient-year). Reoperation was done in 4 patients at an average interval of 1.8 years after MVR (range, 0.2 to 3.6 years). There were 3 deaths. Of the remaining 3 patients, 1 died of valve incompetence and 2 were successfully treated with antibiotics. Reoperation for perivalvular leak was performed in 6 patients at an average of 2.4 years after MVR (range, 0.1 to 7.0 years). There were no deaths. Two patients underwent reoperation because of left atrial thrombosis, and both died. Another patient died of a cerebral embolism; at postmortem examination, the valve appeared stenotic due to thrombosis of two sinuses. The actuarial probability of being free from all causes of bioprosthetic valve dysfunction (Fig 5) and from primary tissue failure (Fig 6) are 56.8 6.5% and 58 f
*
*
*
573
Gallucci et al: Isolated MVR with Hancock Bioprosthesis
B
A
Fig 2. (A)Atrial v i m of a mitral porcine xenograft explanted after 133 months. Note the fibrous tissue overgrowth causing severe reduction of the effective orifice area, and the calcificfragmentation of one cusp. (B) Histological section of one cusp from the same device is covered by a thick, fibrous pannus on the atrial aspect. (H&E; x 12 before 5% reduction.)
6.6%,respectively, at 13 years. Figures 5 and 6 also show the linearized rate of overall bioprosthetic valve dysfunction and primary tissue failure calculated yearly after operation. An attempt was made to determine whether there was a difference in durability depending on the year of valve implantation. At 8 years, the behavior of valves implanted in 1971, 1972, 1974, and 1975 did not differ significantly; those implanted in 1970 and 1973 showed a lower durability, but the number of patients entering the study in these 2 years is too small for statistical comparison with the others (Table).
Thromboembolic Complications and Hemorrhage Forty-five thromboembolic events were observed during the follow-up period. This corresponds to a linearized rate of 2.1 0.3% per patient-year. Thirteen of the events were fatal (linearized rate of 0.6 & 0.1% per patient-year). The actuarial probability of being free from a thromboembolic event at 13 years is 79.2 ? 4.0% (Fig 7). Figure 7 also shows the linearized rate of thromboembolism, as calculated on yearly intervals. Thirty-six patients
*
sustaining a thromboembolic event were in atrial fibrillation, and 9 were in sinus rhythm. Fourteen were on a long-term regimen of anticoagulants, and 31 were not. Among the current survivors, 58% are in atrial fibrillation and 42%, in sinus rhythm. As a result of our current policy, 53% of the survivors are on a regimen of anticoagulant treatment. Anticoagulant-related hemorrhage occurred in 9 patients, a linearized rate of 0.4 f 0.1% per patient-year. Four of them died, a linear rate of 0.2 ~fr0.1% per patient-year.
Comment The glutaraldehyde-preserved Hancock bioprosthesis has been the valve of choice in many centers during the last decade [l,2,6, 7, 13-15]. Reports from these institutions have stressed the advantages of this device, such as its low thrombogenicity and acceptable hemodynamics in the large sizes, but they also have discussed its limited long-term durability. Because of these results, a considerable amount of experimental work has been undertaken and has led to the manufacture of new bioprostheses [16, 171. Improvements in stent design and modification of tissue processing techniques may give these new bioprosthetic valves an increased resistance to fatigue and wear. We believe that data that can help in the assessment of the long-term performance of the old generation of bioprostheses, such as the standard Hancock porcine valve, must be considered of paramount
574 The Annals of Thoracic Surgery Vol 38 No 6 December 1984
A
Fig 3. (A)Atrial aspect ofa mitral bioprosthesis explanted after 111 months because of severe stenosis. Calcific stiffening of the cusps is evident. ( B ) Histological section of one cusl7 shows massive calcific deposits within the valvular tissue. (H&E; x 19 before 5% reduction.)
importance. It is with these valves that the new ones will be compared in the future. The present study, based on an extended follow-up, confirms previous observations from our institution and from others [I, 2, 6, 7, 13-15] regarding the satisfactory clinical performance of the Hancock bioprosthesis. In fact, 13 years after MVR, the actuarial survival approximates 74%, a figure that compare:, favorably with those reported by others using the same or other biological valves [6, 7, 14, 181. Since the questionable long-term durability of the porcine xenograft represents its ma lor disadvantage, we have focused particularly on the rate and causes of valve failure. In our experience, dysfurction of the Hancock valve in the mitral position is predominantly due to primary tissue failure, which is a consequence of cusp calcification; in fact, primary tissue failure without evidence of calcific deposits is exceedingly rare [19]. Recently, we [lo] evaluated calcification of porcine bioprostheses on a pathological basis and found that it appears to be a time-related phenomenon affecting not only bioprosthetic valves implanted in children [3-51, as reported in the past, but also thos? inserted in adults [9,
B 10, 20, 211. Since predisposing factors are seldom detected, calcification of the bioprosthesis can be considered an unavoidable stage in the natural history of the standard porcine valve. Therefore, modifications of the stent design, with the aim of reducing stress on the cusps, and the use of anticalcific treatments in tissue processing seem justifiable, although their effectiveness will be demonstrated only after a clinical experience comparable to those achieved with the old models. Valve durability may be affected by the condition of the tissue at the time of valve fixation. In fact, many reports have demonstrated that processed but unimplanted porcine bioprosthetic valves may already show structural changes, such as loss of endothelial lining, that might predispose to plasma protein insudation and eventual calcification [8, 9, 221. Furthermore, Magilligan and colleagues [23] found that bioprostheses implanted early in their experience behave differently from those implanted later. Those used in 1974 had a lower incidence of degeneration compared with those employed in 1972 and 1973, which suggests a more careful manufacture of the former. We were unable to confirm these findings, at least in our MVR population, since valves implanted from 1970 through 1975 had approximately the same rate of failure at 8 years. This excludes those implanted in 1970 and 1973, because the number of patients entering the study in these years was too limited for a meaningful statistical comparison. In our experience, porcine bioprosthetic valves used in the mitral position show a negligible incidence of primary tissue fail-
575 Gallucci et al: Isolated MVR with Hancock Bioprosthesis
B
A
Fig 4. ( A ) Ventricular aspect of a niitral porcine valve recovered at reoperation after 84 months. A calcific vegetation at one commissure with detachment of two cusps caused severe incompetence. ( B ) Roentgenogram of the same device confirms the presence of gross calcific deposits.
-s w I-
100 -
-s1 0 0--
80-
?80-
U
a
a
a
.A
60-
460-
a
U
3 c
$
0-0
3 U
1
1
58 : 6 6
U c
405 3 8 8 361 326 290 245
198 150 100
71
55
39
15
4
403 387 360 326 290 245
A
Fig 5 . (A)Actuarial probability of being free from overall bioprosthetic valve (BP) dysfunction. ( B ) Linearized rate of overall bioprosthetic valve dysfunction during each yearly postoperative interval. Numbers in parentheses indicate the valve failures. (PT-YR = patient-year.)
B
1
2
3
4
5 6
YEARS
PTF
198 150 100 71
7
53
39
15
8 9 1011 1213
POSTOP
EVENTS
Fig 6 . (A)Actuarial rate of primary tissue failure (PTF). (B) Linearized rate of primary tissue failure. The numbers in parentheses indicate the events occurring during each yearly postoperative interval. (IT-YR = patient-year.)
576 The Annals of Thoracic Surgery- Vol 38 No 6 December 1984
Actuarial Rate of Bioprosthesis Dysfzrnction According to Year of Mitral Valve Replacement Year
No. of Patients"
Percent at 8 'Years
No. of Patientsb
1970 1971 1972 1973 1974 1975
17 40 35 10 24 51
75.7 f 12 93.8 f 4.2 100 62.5 2 17 95.3 f 4.5 94.5 2 3.7
11 31 24 8 18 37
"Patients entering the study. bPatients reaching the eighth postoperative year.
=
i
79254 0
60 U
U
A
B
404 382 351
i
2
316 275 228
178 I25
80
5 4 Y E5A R S6 P OiS T O8P 9
TE
54
40
101'1
28
12
1'213
I
EVENTS
Fig 7. (A)Actuarial and ( B ) linearized rates of postoperative thromboembolic (TE) events after mitral valve replacement with a porcine bioprosthesis. The numbers in parentheses indicate the events during each year. (PT-YR = patient-year.)
ure up to the sixth postoperative year, the highest rate of failure at 7 years, and no further acceleration thereafter. These data are in agreement with the observations of others [12]. Reoperation for primary tissue failure is usually performed on an elective basis, the dysfunction of porcine bioprostheses being progressive. IUevertheless, the mortality observed in this series indicates that reoperation in such patients must not be regarded as a trivial problem. Especially in young patients, the surgeon will be faced with the likelihood of repeated reoperations and a progressively increasing operative risk. Therefore, our current trend is to employ the porcine valve in patients selected by age and specific situations [24, 251. Endocarditis can still be considered a rare cause of
valve dysfunction. It occurred in 7 patients in this series, 4 of whom underwent reoperation with a high mortality, as previously reported [26]. In 2 patients infection was controlled and cured by antibiotic treatment. This is a very recent experience for our group and was not reported in our latest survey of patients with a Hancock bioprosthesis [24]. We now believe that possibly the infected porcine xenograft can be sterilized, as previously stressed by others [13, 15, 271. Previous authors [l, 2, 6, 14, 15, 281 maintained that the Hancock bioprosthesis has a low thrombogenicity, and our findings support this claim. In fact, the probability of being free from thromboembolic events 13 years after MVR is quite high, approximating 80%. The majority of these events occurred in the first postoperative years, many in the first postoperative months [l]. This suggests that temporary anticoagulation after MVR with a porcine valve appears reasonable and advisable. However, the need for long-term anticoagulation in this cohort of patients remains a debated issue. Chronic atrial fibrillation is a predisposing factor to left atrial thrombi, an obvious source of systemic emboli [29, 301. Therefore, our current policy, like that of others [14, 151, is to keep patients in atrial fibrillation on a regimen of anticoagulation indefinitely. We do believe that it is difficult to evaluate the role of atrial fibrillation in the incidence of thromboembolism, since more than 50% of our current survivors are on a regimen of anticoagulants. Some of the thromboembolic events in the late postoperative period might be related more to bioprosthesis failure, with embolization of cusp fragments, than to true thromboemboli [31]. In these patients, assessment of prosthetic function seems more important than the institution of anticoagulant treatment. We conclude that the Hancock bioprosthesis in the mitral position continues to perform satisfactorily after 13 years. Calcific degeneration is the main cause of dysfunction requiring reoperation which may represent a major consideration. However, until further data on the prevention of calcification and the extension of bioprosthetic durability are available, we will continue to implant this device in patients selected according to age and specific situations. This work was supported in part by Grant No. 82.02738.04 of the Consiglio Nazionale delle Ricerche, Rome, Italy. The authors thank Agostino Leorin for his technical assistance.
References 1. C6vese PG, Gallucci V, Morea M, et al: Heart valve replacement with the Hancock bioprosthesis:analysis of long-term results. Circulation 56Suppl2111, 1977 2. Casarotto D, Bortolotti U,Thiene G, et al: Long-term results (from 5 to 7 years) with the Hancock-SGP bioprosthesis. J Cardiovasc Surg (Torino) 20:399, 1979 3. Kutsche LM, Oyer PE, Shumway NE, Baum D: An important complication of Hancock mitral valve replacement in children. Circulation 6O:Suppl 1:98, 1979
577 Gallucci et al: Isolated MVR with Hancock Bioprosthesis
4. Saunders SP, Levy RJ, Freed MD, et al: Use of Hancock porcine xenografts in children and adolescents. Am J Cardiol46429, 1980 5. Bortolotti U, Thiene G: Calcification of porcine heterografts implanted in children. Chest 80:117, 1981 6. Oyer PE, Stinson EB, Reitz BA, et al: Long-term evaluation of the porcine xenograft bioprosthesis. J Thorac Cardiovasc Surg 78:343, 1979 7. Oyer PE, Miller DC, Stinson EB, et a1 Clinical durability of the Hancock porcine bioprosthesis. J Thorac Cardiovasc Surg 80824, 1980 8. Thiene G, Arbustini E, Bortolotti U, et al: Pathological substrates of porcine valve dysfunction. In Cohn LH, Gallucci V (eds): Cardiac Bioprostheses. New York, Yorke Medical, 1982, p 378 9. Valente M, Bortolotti U, Arbustini E, et al: Glutaraldehydepreserved porcine bioprosthesis: factors affecting performance as determined by pathologic studies. Chest 83:607, 1983 10. Milano A, Bortolotti U, Talenti E, et al: Calcific degeneration as the main cause of porcine bioprosthetic valve failure. Am J Cardiol 53:1066, 1984 11. Anderson RP, Bonchek LI, Grunkemeier GE, et al: The analysis and presentation of surgical results by actuarial methods. J Surg Res 16:224, 1974 12. Oyer PE, Stinson EB, Miller DC, et al: Clinical analysis of the Hancock porcine bioprosthesis. In Cohn LH, Gallucci V (eds): Cardiac Bioprostheses. New York, Yorke Medical, 1982, p 539 13. Davila JC, Magilligan DJ Jr, Lewis JW Jr: Is the Hancock porcine valve the best cardiac valve substitute today? Ann Thorac Surg 26:303, 1978 14. Cohn LH, Mudge GH, Pratter F, Collins JJ Jr: Five- to eightyear follow-up of patients undergoing porcine heart valve replacement. N Engl J Med 304258, 1981 15. Gallo JI, Ruiz B, Carrion MF, et al: Heart valve replacement with the Hancock bioprosthesis: a 6-year review. Ann Thorac Surg 31:444, 1981 16. Gabbay S, Frater RWM: The unileaflet heart valve bioprosthesis: new concept. In Cohn LH, Gallucci V (eds): Cardiac Bioprostheses. New York, Yorke Medical, 1982, p 411 17. Wright JTM, Eberhardt CE, Gibbs ML, et al: Hancock 11: an improved bioprosthesis. In Cohn LH, Gallucci V (eds): Cardiac Bioprostheses. New York, Yorke Medical, 1982, p 425 18. Ionescu MI, Tandon AP, Saunders NR, et al: Clinical durability of the pericardial xenograft valve: 11 years’ experience. In Cohn LH, Gallucci V (eds): Cardiac Bioprostheses. New York, Yorke Medical, 1982, p 42 19. Arbustini E, Bortolotti U, Valente M, et al: Cusp disruption by massive lipid infiltration: a rare cause of porcine valve dysfunction. J Thorac Cardiovasc Surg 84:738, 1982 20. Cipriano PR, Billingham ME, Oyer PE, et al: Calcification of porcine prosthetic heart valves: a radiographic and light microscopic study. Circulation 66:1100, 1982 21. Schoen FJ, Collins JJ Jr, Cohn LH: Long-term failure and morphologic correlation in porcine bioprosthetic heart valves. Am J Cardiol 51:957, 1983 22. Ferrans VJ, Spray TL, Billingham ME, Roberts WC: Structural changes in glutaraldehyde-treated porcine heterografts used as substitute cardiac valves: transmission and scanning electron microscopic observations in 12 patients. Am J Cardiol41:1161, 1978 23. Magilligan DJ Jr, Lewis JW Jr, Stein I’D, et al: Decreasing incidence of porcine bioprosthetic degeneration. In Cohn
24.
25. 26. 27. 28. 29.
30. 31.
LH, Gallucci V (eds): Cardiac Bioprostheses. New York, Yorke Medical, 1982, p 559 Gallucci V, Valfre C, Mazzucco A, et a1 Heart valve replacement with the Hancock bioprosthesis: a 5-11 year followup. In Cohn LH, Gallucci V (eds): Cardiac Bioprostheses. New York, Yorke Medical, 1982, p 9 Bortolotti U, Milano A, Mazzucco A, et al: Pregnancy in patients with a porcine bioprosthesis. Am J Cardiol501051, 1982 Bortolotti U, Thiene G, Milano A, et al: Pathological study of infective endocarditis on Hancock porcine bioprostheses. J Thorac Cardiovasc Surg 81:934, 1981 Magilligan DJ Jr, Quinn EL, Davila JC: Bacteremia, endocarditis, and the Hancock valve. Ann Thorac Surg 24:508,1977 Borkon AM, McIntosh CL, Von Rueden TJ, Morrow AG: Mitral valve replacement with the Hancock bioprosthesis: five- to ten-year follow-up. Ann Thorac Surg 32:127, 1981 Thiene G, Bortolotti U, Panizzon G, et al: Pathological substrates of thrombus formation after heart valve replacement with the Hancock bioprosthesis. J Thorac Cardiovasc Surg 80:414, 1980 Edmiston WA, Harrison EC, Duick GF, et al: Thromboembolism in mitral porcine valve recipients. Am J Cardiol 41:508, 1978 Bortolotti U, Milano A, Thiene G, et al: Evidence of impending embolization of a calcific cusp fragment from a mitral porcine xenograft. Thorac Cardiovasc Surg 30:405, 1982
Discussion DR. LAWRENCE H. COHN (Boston, MA): As far as I know, Dr. Gallucci and co-workers were the first to implant the present glutaraldehyde-preserved porcine valve. Their experience is obviously the longest in the world, and Dr. Gallucci has made a very excellent presentation. I think, based on really very meaningful statistical data, we now have some conclusions at the 10year level. I am not sure that on review of the material we can claim to have similarly meaningful 13-year data, because the last points on Dr. Gallucci’s curves show small numbers of patients at risk and the standard error of the mean is higher than we would like to be statistically correct about conclusions at that point. The first point of discussion concerns operative mortality. This has declined, as it has in everybody’s hands, over the past several years. Our series, which extends from 1972, is fairly comparable in terms of numbers to that of Dr. Gallucci. A major difference is that we have employed coronary bypass grafting in about 22% of our patients, roughly about three times the incidence in Dr. Gallucci’s series. I wonder if he would comment on his group’s present approach to coronary angiography in patients with mitral valve disease and whether or not he thinks it may have something to do with the recent improvements in operative mortality. The rate of thromboembolism with this valve is remarkably consistent throughout the world. There are a number of series-those from Stanford and Paris as well as our own-that document about 2 embolic events per year of follow-up. We do have a marked increase in emboli in patients who are in atrial fibrillation. Our data on valve dysfunction with the porcine valve include primary valve dysfunction, perivalvular leak, and bacterial endocarditis. Our final figure of 1.5 events per patient-year does not reflect the “total valve failure” suggested by the Stanford
578 The Annals of Thoracic Surgery Vol 38 No 6 December 1984
group, because we have not included the thromboembolic rate in this calculation. I think that is a good concept, and I congratulate Dr. Gallucci for bringing it .o our attention. There are a few differencesin primary dysfunction rates between our series and that of Dr. Gallucci. Primary tissue failure is not strictly linear, as he has suggested. For instance, 9 patients whom we operated on in 1972 are at risk for valve dysfunction, but it has not occurred in m y of them to date. Yet, from the next year, 3 or 4 patients out of a group at risk of about 25 have had tissue failure. I think Don Magilligan first pointed out that this is not a strictly time-related phenomenon and that it may reflect variability in fabrication techniques in the early days of the manufacture of this valve. Perhaps Dr. Gallucci would comment on this. Finally, Dr. Gallucci, could you tell us how many patients in whom dysfunction developed were either children or less than 35 years of age? Our overall age rang: is considerably older, as it is with most series from the United States compared with those from Europe. Perhaps this is one of the reasons why dysfunction rates vary. This is an excellent analysis, and I was very pleased to have the opportunity to discuss it. w. R. ERIC JAMIESON (Vancouver, BC, Canada): I appreciate the opportunity to discuss Professor Gallucci’s paper, an excellent presentation documenting the longest series in the world with the Hancock porcine bioprosthesis. At the University of British Ccilumbia, the CarpentierEdwards porcine bioprosthesis has been utilized since 1975, and there have been approximately 1,700 implants. I will briefly compare a series of 423 MVRs done between 1975 to 1981 with Dr. Gallucci’s series of 476 done between 1970 and 1981. The 5year survival is higher in the Carpentier-Edwards series (80% versus 71%) but probably reflects the lower operative mortality in this series of implants, which were performed in the latter half of the decade with advanced myocardial preservation. The valve-related complications are relatively similar in both groups, especially with regard to per prosthetic leaks and prosthetic valve endocarditis. The incider ce of thromboembolism is lower in the Carpentier-Edwards series (1.6% versus 2.1% per patient-y ear). There have been 6 instances of prmary tissue failure in our series, with no deaths at reoperation. The degeneration was primarily due to tears and not to calc fication. Dr. Gallucci indicated that his group’s failures were due to calcification. Did tears and perforations accompany the calcification? I also wonder whether Dr. Gallucci considered utilizing the multiple decrement analysis method advocated by Bodnar and colleagues for evaluation of valve performance. The first assessment is the freedom from all valve-related complications (thromboembolism, anticoagulant hemorrhage, prosthetic DR.
valve endocarditis, periprosthetic leak, and primary valve failure). This denotes the actuarial proportion of patients who remain free from any valve-related complications, including those causing death or reoperation. The other two assessments are freedom from mortality related to valve-related complications and freedom from mortality or reoperation related to valverelated complications. In evaluation of patients with the Carpentier-Edwards prosthesis with follow-up intervals over 5 years, the freedom from all valve-related complications was 84%, the freedom from valve-related mortality and reoperation was 92%, and the freedom from valve-related mortality was 97%. Using this methodology, Dr. Craig Miller of Stanford University presented data at a meeting of the American College of Surgeons showing that at the 10-year period, the Hancock porcine bioprosthesis performed better than a standard mechanical prosthesis. DR. GALLUCCI: Thank you very much, Dr. Cohn and Dr. Jamieson, for your remarks. With reference to Dr. Cohn’s observations, the current operative mortality for MVR is around 6 or 7%. This improvement is probably explained by changes in surgical techniques and also by the introduction of cardioplegia. As far as coronary bypass grafting is concerned, in our center we depend mainly on what the cardiologists report to us. We currently require a more careful examination, which includes coronary angiography. In general, we now ask that coronary angiography be done in every patient with valvular disease who is older than 55 years. In the past, this was not routinely performed, and I agree with Dr. Cohn that probably some patients might have died because of episodes of acute coronary insufficiency. The number of thromboembolic episodes are just about the same in Dr. Cohn’s series and our series. Definitely such episodes are more frequent in patients in atrial fibrillation. We took into consideration the difference in behavior of the Hancock valve according to the year of implantation and tried to trace a curve of it. We are aware that the group at the Henry Ford Hospital in Detroit found a difference in durability in bioprostheses depending on the year of implantation. However, in examining the valves implanted from 1970 to 1976, we could not find such a difference, as I mentioned in the presentation. As for the age of the patients, a few of those operated on at the beginning of our experience were children, usually with complex congenital malformations, and all of them had to undergo reoperation because the valve did not last more than 2 or 3 years. Dr. Jamieson, all of the explanted valves undergo roentgenographic, gross, histological, and electron microscopic examination. We have found that almost all valves with some kind of dysfunction had cusp calcification.