Survival and Valve Failure After Aortic Valve Replacement Kevin H. Teoh, MD, Richard D. Weisel, MD, Joan Ivanov, RN, and Susan A. Slattery Division of Cardiovascular Surgery, the Toronto General Hospital and the University of Toronto, Toronto, Ontario, Canada
A prospective evaluation of 412 consecutive patients undergoing isolated aortic valve replacement between January 1982 and December 1985 was performed in an attempt to identify the determinants of survival and valve failure. A variety of valves were inserted to permit a prospective evaluation of alternative valves including: Bjork-Shiley mechanical (n = 37), Ionescu-Shiley pericardial (n = 261), Hancock pericardial (n = 78), and Carpentier-Edwards porcine (n = 36). Thirteen patients died in the hospital (3.2%) and 47 patients died in the follow-up period producing an actuarial survival of 81% f 3% at 48 months. Survival was independently predicted by advancing age, preoperative New York Heart Association functional class, and the presence of endocarditis ( p < 0.05 by Cox regression analysis). The majority of patients were symptomatically improved (New York Heart Association class I or 11: 21% preoperative, 88% postoperative). Freedom from structural valve dysfunction, prosthetic valve endocarditis, and reopera-
tion for valve-related complications were 95% f 2%, 95% f 2%, and 92% f 2% at 48 months, respectively. These valve-related complications occurred more frequently in younger patients and in those with a Hancock pericardial valve (freedom from structural valve dysfunction, 89% f 5%; prosthetic valve endocarditis, 84% 9%; reoperation, 78% 2 10%; p < 0.05 by Cox regression). Freedom from thromboembolism was 88% f 2% at 48 months; it was significantly lower in patients with a preoperative thromboembolic event and was not influenced by the type of prosthesis inserted. Freedom from anticoagulantrelated hemorrhage was 85% 8% at 48 months and was not influenced by any preoperative factors. Aortic valve replacement resulted in excellent symptomatic improvement. Postoperatively, older patients died whereas younger patients suffered valve-related complications. The Hancock pericardial valve had an unacceptably high incidence of premature valve failure. (Ann Tliorac Surg 1991;52:270-5)
T
Clinical Preseizta tioii
he recent experience of isolated aortic valve replacement at the Toronto General Hospital was reviewed to determine the survival and incidence of valve-related complications. An analysis of preoperative patient characteristics was performed in a n attempt to identify those factors that influenced postoperative survival and complications. In addition, a variety of prostheses were used, which permitted an evaluation of their performance and influence on patient mortality and morbidity.
Material and Methods Patients Between January 1982 and December 1985, 412 consecutive patients underwent isolated aortic valve replacement at the Toronto General Hospital. One hundred twentyfive patients required concomitant coronary bypass grafting, and they were included in the study. During this interval, 46 patients had combined aortic and mitral valve operations and were excluded from the analysis, as were 46 patients who had concomitant ascending aortic procedures. Preoperative, operative, and postoperative data were collected. Accepted for publication April 16, 1991 Address reprint requests to Dr Weisel, Cardiovascular Surgery, Toronto General Hospital, EN 14-215, 200 Elizabeth St, Toronto, Ont, M5V 1L7, Canada.
0 1991 by The Society of Thoracic Surgeons
*
Sex, age, New York Heart Association (NYHA) functional class, presence of endocarditis, urgency of operation, and history of previous sternotomy were recorded prospectively (Table 1). Most of the patients were men (71%)and more than 60 years of age (53%).The majority of patients were in NYHA class 111 or IV (79%). Twenty-seven patients (7%) required aortic valve replacement for native valve endocarditis. Fifty-two patients (13%)required urgent operations (within 72 hours of cardiac catheterization). Forty-five patients (11%)had undergone a previous sternotomy.
Cardiac Catheterization Cardiac catheterization data were reviewed to determine the valvular lesion, ventricular function, and extent of coronary disease (Table 2). Cardiac pressures were recorded. Cardiac output was calculated by the thermodilution or Fick method and was used to calculate the aortic valve area. Left ventricular ejection fraction was determined by planimetry from a single-plane right anterior oblique contrast ventriculogram. One hundred seventy-six patients (43%) had predominant aortic stenosis defined as a peak systolic gradient greater than 40 mm Hg, a n aortic valve area less than 1 cm2, and aortic regurgitation of 1+ or less. The peak systolic aortic valve gradient was calculated either by simultaneous pressure measurements or by the pullback 0003-4975/91/’$3,50
TEO H ETA L AORTIC VALVE REPLACEMENT
Ann Thorac Surg 1991:52:270-5
271
Table 1 . Preoperative Patient Profile Total
Variable
No. of patients Sex (males)
Age (Y)
No. of Patients
Bjork-Shiley
Ionescu-Shiley
Hancock
Mechanical
Pericardial
Pericardial
Carpentier-Edwards Porcine
412 293 59 2 14
37 31 (84%) 48.7 2 14“
261 176 (67%) 60.4 ? 13
78 59 (76%) 58.7 ? 14
36 27 (75%) 60.2 2 14
3 (8%) 6 (16%) 22 (59%) 6 (16%) 6 (16%) 16 (43%) 9 (24%)
10 (4%) 41 (16%) 162 (62%) 48 (18%) 14 (5%) 17 (6%) 31 (12%)
4 (5%) 20 (26%) 40 (51%) 14 (18%) 3 (4%) 10 (13%) 7 (9%)
Preoperative NYHA class I I1 111
IV 27 45 52
Endocarditis Previous sternotomy Nonelective
0 (0%)
3 (8%) 26 (72%) 7 (19%) 4 (11%) 2 (6%)” 5 (14%)‘
* p < 0.05. NYHA = New York Heart Association.
technique. The severity of aortic regurgitation was estimated by the extent of left ventricular opacification after aortic root injection of contrast medium (minimal = 1+, mild = 2+, moderate = 3+, severe = 4+). Ninety-three patients (22%)had predominant aortic regurgitation characterized by 3+ or greater opacification of the left ventricle after aortic root injection and a pulse pressure greater than 60% of the systolic aortic pressure. Mixed aortic valve disease with features of both aortic stenosis and regurgitation was found in 143 patients (35%).Mitral and tricuspid valve stenosis or regurgitation were infrequent (mitral, 6%; tricuspid, 0.2%) and were hemodynamically insignificant. Coronary angiography showed a stenosis of greater than 50% (diameter) narrowing of the major coronary artery supplying one of three cardiac regions (left anterior descending, circumflex, and right coronary arteries) in 158 patients (38%).Sixty-nine patients (17%)had single-vessel disease, 51 patients (12%)had double-vessel disease, and
38 patients (9%) had triple-vessel disease. Left main coronary stenosis was present in 14 patients (3%).
Operative Techniques The anesthetic management and operative techniques were similar among the anesthetists and the five surgeons participating in the study. Fentanyl citrate was used for induction and maintenance of anesthesia. Moderate hemodilution (hematocrit, 0.20 to 0.25) and systemic hypothermia (25°C)were maintained during cardiopulmonary bypass. Multidose cold crystalloid cardioplegia was used in the majority of patients (n = 387), and some recent patients received blood cardioplegia (n = 25). Cardioplegic solution was delivered by hand-held cannulas every 20 minutes to maintain myocardial temperatures less than 15°C. Additional topical hypothermic protection was afforded by a cooling pad. Concomitant coronary artery bypass grafting was performed with saphenous veins. Distal anastomoses were performed be-
Table 2. Cardiac Catheterization Data ~~
Variable Aortic valve lesion Stenosis
Regurgitation Mixed
Coronary artery disease Left ventricular ejection fraction 0.60
Total No. of
Patients 176 193 143 158
Bjork- Shiley
Pericardial
Hancock Pericardial
Carpentier-Edwards Porcine
116 (44%) 53 (20%) 92 (25%) 100 (38%)
33 (42%) 17 (22%) 28 (36%) 31 (40%)
13 (36%) 10 (28%) 13 (36%) 15 (42%)
3 (8%)
15 (6%)
8 (22%) 9 (24%) 17 (46%)
47 (18%) 52 (20%) 147 (56%)
2 (3%) 8 (10%) 19 (24%) 48 (63%)
2 (6%) 7 (18%) 8 (22%) 19 (53%)
Mechanical 14 (38%) 13 (35%) 10 (27%) 12 (32%)
Ionescu-Shiley
272
TEOHETAL AORTIC VALVE REPLACEMENT
Ann Thorac Surg 1991;52:27&5
fore valve replacement to permit vein graft infusions with cardioplegic solutions. The cardiopulmonary bypass time was 118 k 31 minutes and the aortic cross-clamp time was 78 19 minutes.
*
standard error of the mean. In the text the p values refer to the Cox multivariant regression analysis, whereas the p values depicted in the figures refer to Wilcoxon or MantelCox test for the univariate analyses.
Prosthesis Selection of the prosthesis was based on the patient‘s age and history of previous thromboembolism or bleeding disorder. The Bjork-Shiley mechanical valve was inserted in 37 patients who were younger (49 14 years) or had a history of thromboembolism requiring anticoagulation. Bioprotheses tended to be inserted in older patients (60 14 years) and those who had a contraindication to anticoagulation. However, the surgeon’s choice of prosthesis was based on a discussion with the patient and his or her cardiologist. A variety of bioprosthetic valves were inserted. The Ionescu-Shiley pericardial valve was inserted in 261 patients, the Hancock pericardial valve in 78 patients, and the Carpentier-Edwards porcine valve in 36 patients. All valves were inserted with interrupted pledgeted sutures.
*
*
Results
Mortality Sixty patients died postoperatively. Thirteen patients (3.2%)died during their hospital stay or within 30 days of operation. The causes of death were myocardial infarction and cardiogenic shock in 10 patients, hemorrhage in 2 patients, and stroke in 1 patient. There were 47 late deaths resulting in an actuarial survival of 81% 3% at 48 months. Ten patients with preoperative coronary artery disease died because of myocardial infarction, and 14 patients died of a malignancy. Patients with coronary artery disease who did not undergo concomitant coronary artery bypass grafting had a lower 48-month survival (67% t 9%) than patients who had coronary artery disease and underwent coronary bypass grafting with aortic valve replacement (80% 5%) or patients without coronary artery disease (84% *3%). These differences approached statistical significance ( p = 0.06). Fifteen patients died because of valve-related complications (endocarditis, 5 patients; anticoagulant-related hemorrhage, 4; thromboembolism, 5; structural valve failure, 1).There were eight sudden unexplained deaths. Therefore, 23 patients were believed to have had a valve-related death: 14 confirmed, 8 unexplained, and 1 lost to follow-up. Cox regression analysis (Fig 1) identified three independent predictors of survival: advancing age, NYHA functional class, and endocarditis. Twenty-seven patients had endocarditis. These patients were younger (48 16 versus 60 L 13 years; p < 0.01) and had a greater incidence of urgent operation (63% versus 9%; p < 0.01) than the patients who did not have endocarditis. Valve-related deaths occurred more frequently in older patients (survival at 48 months: 70
0
12
24
36
48
VALVE TYPE d MECHANICAL 4 PORCINE --APERICARDIAL HANCOCK
--
-.-+
valves. Only 19% of patients with mechanical valves received antiplatelet agents. Fifteen patients suffered an anticoagulant-related hemorrhage, of which 4 cases were fatal. Freedom from anticoagulant-related hemorrhage was 94% 5 1%at 48 months. Anticoagulant-related liemorrhage was not influenced by antiplatelet drug administration. Nineteen patients underwent reoperation for valverelated complications. Ten patients required reoperation for structural valve dysfunction. Prosthetic valve etidocarditis was the indication in 5 patients. Four patients underwent reoperation because of a periprosthetic leak. Actuarial freedom from reoperation was 92% 2 2% at 48 months. Young age and the Hancock prosthesis were independent risk factors for reoperation. (Figure 4 depicts the univariate analysis.)
Postoperative New York Heart Associatiori Class The majority of patients had symptomatic improvement after aortic valve replacement. Eighty-eight percent of patients were in NYHA class I or I1 postoperatively, compared with 21 70 preoperatively. Eighty-thrce percent of patients had an improvement in one or more NYHA classes or no change if they were already in NYHA class I or I1 preoperatively. A worsening of NYHA functional class or a failure to improve from NYHA class 111 or 1V was found in 17% of patients.
Comment The patients studied underwent operation between 1982 and 1985. This permitted an evaluation of the results of aortic valve replacement using current methods of anesthetic management, surgical techniques, and postoperative care.
Follow-up was 99.8% complete, and most postoperative events were confirmed such that undetected events were unlikely. Although the duration of follow-up was limited to the intermediate term (48 months), clinically relevant findings with regard to survival and valve-related complications were apparent. However, the number of patients and the duration of the follow-up were limited, and therefore significant findings may not have been detected due to the low power of this study. The in-hospital mortality and long-term survival for this study were 3.2% and 81% 5 3% at 48 months, respectively. Our experience at the Toronto General Hospital was similar to those of other centers who report hospital mortality ranging from 3.7% to 13% and 4-year survival ranging from 75% to 90% [ P 7 ] . Advancing age and poor preoperative functional class adversely affected survival. Other authors have also found these two factors to influence survival after aortic valve replacement [&8]. Survival was lower in older patients, reflecting the predominance of elderly patients undergoing aortic valve replacement. Fifty-three percent of patients were more than 60 years of age. Although operative mortality was higher in older patients [9], they were symptomatically improved and valve-related complications were few. Patients with preoperative functional impairment, particularly those in class IV, had significantly lower survival. Functional impairment may reflect marked left ventricular dysfunction. Other studies have found indices of left ventricular dysfunction to adversely influence postoperative survival [4, 8, 91. Survival after aortic valve replacement may be improved by intervening surgically before patients become severely symptomatic. Survival was also lower in patients with endocarditis before operation. Most of these patients (63%) required urgent operation because of severe hemodynamic decompensation and florid sepsis, which also influenced survival. Freedom from thromboembolism and anticoagulantrelated hemorrhage was 88% ? 2% and 94% 1% at 48 months, respectively. Similar results have been obtained by other centers [5, lo]. Cohn and associates [ll]found a greater incidence of thromboembolism with mechanical valves compared with bioprosthetic valves. We found that patients with a preoperative history of thromboembolism were at risk of postoperative thromboembolism regardless of the type of prosthesis inserted. Perhaps the patients in the series of Cohn and associates [ l l ] received mechanical valves because of preoperative indications for anticoagulation such as thromboembolism. Thromboembolic events occurred more frequently in patients receiving anticoagulants. We could not identify the clues that led the cardiologists to order anticoagulants for patients at risk of subsequent thromboemboli. Younger patients were more prone to valve-related complications such as structural valve dysfunction, prosthetic valve endocarditis, and reoperation. The studies of Blackstone [12], Mitchell [ 8 ] ,Borken [ 5 ] , and their associates also found young age to increase the likelihood of these complications. The explanation for the increased risk in younger patients is not known.
*
Ann Thorac Surg 1991;52:27&5
The Hancock pericardial prosthesis displayed an unacceptably high incidence of structural valve dysfunction, prosthetic valve endocarditis, and reoperation. Prosthetic valve endocarditis likely developed in prostheses with intrinsic abnormalities. Goldman and co-workers [13] and Bortolotti and colleagues [ 141 have also recently reported early failure of the Hancock pericardial valve. In view of their findings, the Hancock pericardial valve is no longer used at our institution. Patients with the Hancock valve are being followed up closely in anticipation of eventual valve dysfunction. In summary, contemporary aortic valve replacement was accomplished with very acceptable operative mortality and resulted in marked symptomatic improvement. Late survival was influenced by preoperative patient factors: age, functional class, and endocarditis. Valverelated complications were similar between mechanical and bioprosthetic valves. However, premature failure of the Hancock pericardial valve resulted in an unacceptable rate of valve-related complications. Supported by the Canadian Heart Foundation and the Heart and Stroke Foundation of Ontario. We thank Ms Jeri Azuma for assistance with data collection. We appreciate the assistance in the preparation of the illustrations and manuscript from Ms Mindy Madonik, Ms Eslyn Kesper, and Ms Carol Davison. We acknowledge the continuing support of the cardiac surgeons at the Toronto General Hospital, Drs Tirone E. David, Ronald J. Baird, Hugh E. Scully, and Lynda L. Mickleborough.
References 1. Edmunds LH, Clark RE, Cohn LH, et al. Guidelines for reporting morbidity and mortality after cardiac valvular operations. Ann Thorac Surg 1988;46:257-9. 2. Kaplan EL, Meier P. Non-parametric estimations from incomplete observations. J Am Stat Assoc 1958;53:457-81.
TEOHETAL AORTIC VALVE REPLACEMENT
275
3. Cox DR. Regression models and life tables (with discussion). J R Stat Soc (Series 8) 1972;34:187. 4. Lytle BW, Cosgrove DM, Gill CC, Taylor PC, et al. Aortic valve replacement combined with myocardial revascularization: late results and determinants of risk for 471 in-hospital survivors. J Thorac Cardiovasc Surg 1988;95:402-14. 5. Borkon AM, Soule LM, Baughman KL, et al. Comparative analysis of mechanical and bioprosthetic valves after aortic valve replacement. J Thorac Cardiovasc Surg 1987;94:2&33. 6. Stewart S, Cianciotta C, Hicks GL, DeWeese JA. The LilleheiKaster aortic valve prosthesis. J Thorac Cardiovasc Surg 1988;95:102>30. 7. Lindblom D. Long-term clinical results after aortic valve replacement with Bjiirk-Shiley prosthesis. J Thorac Cardiovasc Surg 1988;95:658-67. 8. Mitchell RS, Miller DC, Stinson EB, et al. Significant patientrelated determinants of prosthetic valve performance. J Thorac Cardiovasc Surg 1986;91:807-17. 9. Christakis GT, Weisel RD, David TE, Salerno TA, Ivanov J, and the Cardiac Surgeons at the University of Toronto. Predictors of operative survival after valve replacement. Circulation 1988;78:(Suppl 1):25-34. 10. Cobanoglu A, Fessler CL, Guvendik L, Grunkemeier G, Starr A. Aortic valve replacement with the Starr-Edwards prosthesis: a comparison of the first and second decades of followup. Ann Thorac Surg 1988;45:248-52. 11. Cohn LH, Allred EN, DiSesa VJ, Sawtelle K, Shemin RJ, Collins JJ Jr. Early and late risk of aortic valve replacement: a 12-year concomitant comparison of the porcine bioprosthetic and tilting disk prosthetic aortic valves. J Thorac Cardiovasc Surg 1984;88:695-705. 12. Blackstone EH, Kirklin JW. Death and other time-related events after valve replacement. Circulation 1985;72:753-67. 13. Goldman B, Scully H, Tong C, et al. Clinical results of pericardial xenograft valves: the lonescu-Shiley and Hancock valves. Can J Cardiol 1988;4:328-32. 14. Bortolotti U, Milano A, Thiene G, et al. Early mechanical failures of the Hancock pericardial xenograft. J Thorac Cardiovasc Surg 1987;94:200-7.
A prospective evaluation of 412 consecutive patients undergoing isolated aortic valve replacement between January 1982 and December 1985 was performed in an attempt to identify the determinants of survival and valve failure. A variety of valves were inserted to permit a prospective evaluation of alternative valves including: Bjork-Shiley mechanical (n = 37), Ionescu-Shiley pericardial (n = 261), Hancock pericardial (n = 78), and Carpentier-Edwards porcine (n = 36). Thirteen patients died in the hospital (3.2%) and 47 patients died in the follow-up period producing an actuarial survival of 81% f 3% at 48 months. Survival was independently predicted by advancing age, preoperative New York Heart Association functional class, and the presence of endocarditis ( p < 0.05 by Cox regression analysis). The majority of patients were symptomatically improved (New York Heart Association class I or 11: 21% preoperative, 88% postoperative). Freedom from structural valve dysfunction, prosthetic valve endocarditis, and reopera-
tion for valve-related complications were 95% f 2%, 95% f 2%, and 92% f 2% at 48 months, respectively. These valve-related complications occurred more frequently in younger patients and in those with a Hancock pericardial valve (freedom from structural valve dysfunction, 89% f 5%; prosthetic valve endocarditis, 84% 9%; reoperation, 78% 2 10%; p < 0.05 by Cox regression). Freedom from thromboembolism was 88% f 2% at 48 months; it was significantly lower in patients with a preoperative thromboembolic event and was not influenced by the type of prosthesis inserted. Freedom from anticoagulantrelated hemorrhage was 85% 8% at 48 months and was not influenced by any preoperative factors. Aortic valve replacement resulted in excellent symptomatic improvement. Postoperatively, older patients died whereas younger patients suffered valve-related complications. The Hancock pericardial valve had an unacceptably high incidence of premature valve failure. (Ann Tliorac Surg 1991;52:270-5)
T
Clinical Preseizta tioii
he recent experience of isolated aortic valve replacement at the Toronto General Hospital was reviewed to determine the survival and incidence of valve-related complications. An analysis of preoperative patient characteristics was performed in a n attempt to identify those factors that influenced postoperative survival and complications. In addition, a variety of prostheses were used, which permitted an evaluation of their performance and influence on patient mortality and morbidity.
Material and Methods Patients Between January 1982 and December 1985, 412 consecutive patients underwent isolated aortic valve replacement at the Toronto General Hospital. One hundred twentyfive patients required concomitant coronary bypass grafting, and they were included in the study. During this interval, 46 patients had combined aortic and mitral valve operations and were excluded from the analysis, as were 46 patients who had concomitant ascending aortic procedures. Preoperative, operative, and postoperative data were collected. Accepted for publication April 16, 1991 Address reprint requests to Dr Weisel, Cardiovascular Surgery, Toronto General Hospital, EN 14-215, 200 Elizabeth St, Toronto, Ont, M5V 1L7, Canada.
0 1991 by The Society of Thoracic Surgeons
*
Sex, age, New York Heart Association (NYHA) functional class, presence of endocarditis, urgency of operation, and history of previous sternotomy were recorded prospectively (Table 1). Most of the patients were men (71%)and more than 60 years of age (53%).The majority of patients were in NYHA class 111 or IV (79%). Twenty-seven patients (7%) required aortic valve replacement for native valve endocarditis. Fifty-two patients (13%)required urgent operations (within 72 hours of cardiac catheterization). Forty-five patients (11%)had undergone a previous sternotomy.
Cardiac Catheterization Cardiac catheterization data were reviewed to determine the valvular lesion, ventricular function, and extent of coronary disease (Table 2). Cardiac pressures were recorded. Cardiac output was calculated by the thermodilution or Fick method and was used to calculate the aortic valve area. Left ventricular ejection fraction was determined by planimetry from a single-plane right anterior oblique contrast ventriculogram. One hundred seventy-six patients (43%) had predominant aortic stenosis defined as a peak systolic gradient greater than 40 mm Hg, a n aortic valve area less than 1 cm2, and aortic regurgitation of 1+ or less. The peak systolic aortic valve gradient was calculated either by simultaneous pressure measurements or by the pullback 0003-4975/91/’$3,50
TEO H ETA L AORTIC VALVE REPLACEMENT
Ann Thorac Surg 1991:52:270-5
271
Table 1 . Preoperative Patient Profile Total
Variable
No. of patients Sex (males)
Age (Y)
No. of Patients
Bjork-Shiley
Ionescu-Shiley
Hancock
Mechanical
Pericardial
Pericardial
Carpentier-Edwards Porcine
412 293 59 2 14
37 31 (84%) 48.7 2 14“
261 176 (67%) 60.4 ? 13
78 59 (76%) 58.7 ? 14
36 27 (75%) 60.2 2 14
3 (8%) 6 (16%) 22 (59%) 6 (16%) 6 (16%) 16 (43%) 9 (24%)
10 (4%) 41 (16%) 162 (62%) 48 (18%) 14 (5%) 17 (6%) 31 (12%)
4 (5%) 20 (26%) 40 (51%) 14 (18%) 3 (4%) 10 (13%) 7 (9%)
Preoperative NYHA class I I1 111
IV 27 45 52
Endocarditis Previous sternotomy Nonelective
0 (0%)
3 (8%) 26 (72%) 7 (19%) 4 (11%) 2 (6%)” 5 (14%)‘
* p < 0.05. NYHA = New York Heart Association.
technique. The severity of aortic regurgitation was estimated by the extent of left ventricular opacification after aortic root injection of contrast medium (minimal = 1+, mild = 2+, moderate = 3+, severe = 4+). Ninety-three patients (22%)had predominant aortic regurgitation characterized by 3+ or greater opacification of the left ventricle after aortic root injection and a pulse pressure greater than 60% of the systolic aortic pressure. Mixed aortic valve disease with features of both aortic stenosis and regurgitation was found in 143 patients (35%).Mitral and tricuspid valve stenosis or regurgitation were infrequent (mitral, 6%; tricuspid, 0.2%) and were hemodynamically insignificant. Coronary angiography showed a stenosis of greater than 50% (diameter) narrowing of the major coronary artery supplying one of three cardiac regions (left anterior descending, circumflex, and right coronary arteries) in 158 patients (38%).Sixty-nine patients (17%)had single-vessel disease, 51 patients (12%)had double-vessel disease, and
38 patients (9%) had triple-vessel disease. Left main coronary stenosis was present in 14 patients (3%).
Operative Techniques The anesthetic management and operative techniques were similar among the anesthetists and the five surgeons participating in the study. Fentanyl citrate was used for induction and maintenance of anesthesia. Moderate hemodilution (hematocrit, 0.20 to 0.25) and systemic hypothermia (25°C)were maintained during cardiopulmonary bypass. Multidose cold crystalloid cardioplegia was used in the majority of patients (n = 387), and some recent patients received blood cardioplegia (n = 25). Cardioplegic solution was delivered by hand-held cannulas every 20 minutes to maintain myocardial temperatures less than 15°C. Additional topical hypothermic protection was afforded by a cooling pad. Concomitant coronary artery bypass grafting was performed with saphenous veins. Distal anastomoses were performed be-
Table 2. Cardiac Catheterization Data ~~
Variable Aortic valve lesion Stenosis
Regurgitation Mixed
Coronary artery disease Left ventricular ejection fraction 0.60
Total No. of
Patients 176 193 143 158
Bjork- Shiley
Pericardial
Hancock Pericardial
Carpentier-Edwards Porcine
116 (44%) 53 (20%) 92 (25%) 100 (38%)
33 (42%) 17 (22%) 28 (36%) 31 (40%)
13 (36%) 10 (28%) 13 (36%) 15 (42%)
3 (8%)
15 (6%)
8 (22%) 9 (24%) 17 (46%)
47 (18%) 52 (20%) 147 (56%)
2 (3%) 8 (10%) 19 (24%) 48 (63%)
2 (6%) 7 (18%) 8 (22%) 19 (53%)
Mechanical 14 (38%) 13 (35%) 10 (27%) 12 (32%)
Ionescu-Shiley
272
TEOHETAL AORTIC VALVE REPLACEMENT
Ann Thorac Surg 1991;52:27&5
fore valve replacement to permit vein graft infusions with cardioplegic solutions. The cardiopulmonary bypass time was 118 k 31 minutes and the aortic cross-clamp time was 78 19 minutes.
*
standard error of the mean. In the text the p values refer to the Cox multivariant regression analysis, whereas the p values depicted in the figures refer to Wilcoxon or MantelCox test for the univariate analyses.
Prosthesis Selection of the prosthesis was based on the patient‘s age and history of previous thromboembolism or bleeding disorder. The Bjork-Shiley mechanical valve was inserted in 37 patients who were younger (49 14 years) or had a history of thromboembolism requiring anticoagulation. Bioprotheses tended to be inserted in older patients (60 14 years) and those who had a contraindication to anticoagulation. However, the surgeon’s choice of prosthesis was based on a discussion with the patient and his or her cardiologist. A variety of bioprosthetic valves were inserted. The Ionescu-Shiley pericardial valve was inserted in 261 patients, the Hancock pericardial valve in 78 patients, and the Carpentier-Edwards porcine valve in 36 patients. All valves were inserted with interrupted pledgeted sutures.
*
*
Results
Mortality Sixty patients died postoperatively. Thirteen patients (3.2%)died during their hospital stay or within 30 days of operation. The causes of death were myocardial infarction and cardiogenic shock in 10 patients, hemorrhage in 2 patients, and stroke in 1 patient. There were 47 late deaths resulting in an actuarial survival of 81% 3% at 48 months. Ten patients with preoperative coronary artery disease died because of myocardial infarction, and 14 patients died of a malignancy. Patients with coronary artery disease who did not undergo concomitant coronary artery bypass grafting had a lower 48-month survival (67% t 9%) than patients who had coronary artery disease and underwent coronary bypass grafting with aortic valve replacement (80% 5%) or patients without coronary artery disease (84% *3%). These differences approached statistical significance ( p = 0.06). Fifteen patients died because of valve-related complications (endocarditis, 5 patients; anticoagulant-related hemorrhage, 4; thromboembolism, 5; structural valve failure, 1).There were eight sudden unexplained deaths. Therefore, 23 patients were believed to have had a valve-related death: 14 confirmed, 8 unexplained, and 1 lost to follow-up. Cox regression analysis (Fig 1) identified three independent predictors of survival: advancing age, NYHA functional class, and endocarditis. Twenty-seven patients had endocarditis. These patients were younger (48 16 versus 60 L 13 years; p < 0.01) and had a greater incidence of urgent operation (63% versus 9%; p < 0.01) than the patients who did not have endocarditis. Valve-related deaths occurred more frequently in older patients (survival at 48 months: 70
0
12
24
36
48
VALVE TYPE d MECHANICAL 4 PORCINE --APERICARDIAL HANCOCK
--
-.-+
valves. Only 19% of patients with mechanical valves received antiplatelet agents. Fifteen patients suffered an anticoagulant-related hemorrhage, of which 4 cases were fatal. Freedom from anticoagulant-related hemorrhage was 94% 5 1%at 48 months. Anticoagulant-related liemorrhage was not influenced by antiplatelet drug administration. Nineteen patients underwent reoperation for valverelated complications. Ten patients required reoperation for structural valve dysfunction. Prosthetic valve etidocarditis was the indication in 5 patients. Four patients underwent reoperation because of a periprosthetic leak. Actuarial freedom from reoperation was 92% 2 2% at 48 months. Young age and the Hancock prosthesis were independent risk factors for reoperation. (Figure 4 depicts the univariate analysis.)
Postoperative New York Heart Associatiori Class The majority of patients had symptomatic improvement after aortic valve replacement. Eighty-eight percent of patients were in NYHA class I or I1 postoperatively, compared with 21 70 preoperatively. Eighty-thrce percent of patients had an improvement in one or more NYHA classes or no change if they were already in NYHA class I or I1 preoperatively. A worsening of NYHA functional class or a failure to improve from NYHA class 111 or 1V was found in 17% of patients.
Comment The patients studied underwent operation between 1982 and 1985. This permitted an evaluation of the results of aortic valve replacement using current methods of anesthetic management, surgical techniques, and postoperative care.
Follow-up was 99.8% complete, and most postoperative events were confirmed such that undetected events were unlikely. Although the duration of follow-up was limited to the intermediate term (48 months), clinically relevant findings with regard to survival and valve-related complications were apparent. However, the number of patients and the duration of the follow-up were limited, and therefore significant findings may not have been detected due to the low power of this study. The in-hospital mortality and long-term survival for this study were 3.2% and 81% 5 3% at 48 months, respectively. Our experience at the Toronto General Hospital was similar to those of other centers who report hospital mortality ranging from 3.7% to 13% and 4-year survival ranging from 75% to 90% [ P 7 ] . Advancing age and poor preoperative functional class adversely affected survival. Other authors have also found these two factors to influence survival after aortic valve replacement [&8]. Survival was lower in older patients, reflecting the predominance of elderly patients undergoing aortic valve replacement. Fifty-three percent of patients were more than 60 years of age. Although operative mortality was higher in older patients [9], they were symptomatically improved and valve-related complications were few. Patients with preoperative functional impairment, particularly those in class IV, had significantly lower survival. Functional impairment may reflect marked left ventricular dysfunction. Other studies have found indices of left ventricular dysfunction to adversely influence postoperative survival [4, 8, 91. Survival after aortic valve replacement may be improved by intervening surgically before patients become severely symptomatic. Survival was also lower in patients with endocarditis before operation. Most of these patients (63%) required urgent operation because of severe hemodynamic decompensation and florid sepsis, which also influenced survival. Freedom from thromboembolism and anticoagulantrelated hemorrhage was 88% ? 2% and 94% 1% at 48 months, respectively. Similar results have been obtained by other centers [5, lo]. Cohn and associates [ll]found a greater incidence of thromboembolism with mechanical valves compared with bioprosthetic valves. We found that patients with a preoperative history of thromboembolism were at risk of postoperative thromboembolism regardless of the type of prosthesis inserted. Perhaps the patients in the series of Cohn and associates [ l l ] received mechanical valves because of preoperative indications for anticoagulation such as thromboembolism. Thromboembolic events occurred more frequently in patients receiving anticoagulants. We could not identify the clues that led the cardiologists to order anticoagulants for patients at risk of subsequent thromboemboli. Younger patients were more prone to valve-related complications such as structural valve dysfunction, prosthetic valve endocarditis, and reoperation. The studies of Blackstone [12], Mitchell [ 8 ] ,Borken [ 5 ] , and their associates also found young age to increase the likelihood of these complications. The explanation for the increased risk in younger patients is not known.
*
Ann Thorac Surg 1991;52:27&5
The Hancock pericardial prosthesis displayed an unacceptably high incidence of structural valve dysfunction, prosthetic valve endocarditis, and reoperation. Prosthetic valve endocarditis likely developed in prostheses with intrinsic abnormalities. Goldman and co-workers [13] and Bortolotti and colleagues [ 141 have also recently reported early failure of the Hancock pericardial valve. In view of their findings, the Hancock pericardial valve is no longer used at our institution. Patients with the Hancock valve are being followed up closely in anticipation of eventual valve dysfunction. In summary, contemporary aortic valve replacement was accomplished with very acceptable operative mortality and resulted in marked symptomatic improvement. Late survival was influenced by preoperative patient factors: age, functional class, and endocarditis. Valverelated complications were similar between mechanical and bioprosthetic valves. However, premature failure of the Hancock pericardial valve resulted in an unacceptable rate of valve-related complications. Supported by the Canadian Heart Foundation and the Heart and Stroke Foundation of Ontario. We thank Ms Jeri Azuma for assistance with data collection. We appreciate the assistance in the preparation of the illustrations and manuscript from Ms Mindy Madonik, Ms Eslyn Kesper, and Ms Carol Davison. We acknowledge the continuing support of the cardiac surgeons at the Toronto General Hospital, Drs Tirone E. David, Ronald J. Baird, Hugh E. Scully, and Lynda L. Mickleborough.
References 1. Edmunds LH, Clark RE, Cohn LH, et al. Guidelines for reporting morbidity and mortality after cardiac valvular operations. Ann Thorac Surg 1988;46:257-9. 2. Kaplan EL, Meier P. Non-parametric estimations from incomplete observations. J Am Stat Assoc 1958;53:457-81.
TEOHETAL AORTIC VALVE REPLACEMENT
275
3. Cox DR. Regression models and life tables (with discussion). J R Stat Soc (Series 8) 1972;34:187. 4. Lytle BW, Cosgrove DM, Gill CC, Taylor PC, et al. Aortic valve replacement combined with myocardial revascularization: late results and determinants of risk for 471 in-hospital survivors. J Thorac Cardiovasc Surg 1988;95:402-14. 5. Borkon AM, Soule LM, Baughman KL, et al. Comparative analysis of mechanical and bioprosthetic valves after aortic valve replacement. J Thorac Cardiovasc Surg 1987;94:2&33. 6. Stewart S, Cianciotta C, Hicks GL, DeWeese JA. The LilleheiKaster aortic valve prosthesis. J Thorac Cardiovasc Surg 1988;95:102>30. 7. Lindblom D. Long-term clinical results after aortic valve replacement with Bjiirk-Shiley prosthesis. J Thorac Cardiovasc Surg 1988;95:658-67. 8. Mitchell RS, Miller DC, Stinson EB, et al. Significant patientrelated determinants of prosthetic valve performance. J Thorac Cardiovasc Surg 1986;91:807-17. 9. Christakis GT, Weisel RD, David TE, Salerno TA, Ivanov J, and the Cardiac Surgeons at the University of Toronto. Predictors of operative survival after valve replacement. Circulation 1988;78:(Suppl 1):25-34. 10. Cobanoglu A, Fessler CL, Guvendik L, Grunkemeier G, Starr A. Aortic valve replacement with the Starr-Edwards prosthesis: a comparison of the first and second decades of followup. Ann Thorac Surg 1988;45:248-52. 11. Cohn LH, Allred EN, DiSesa VJ, Sawtelle K, Shemin RJ, Collins JJ Jr. Early and late risk of aortic valve replacement: a 12-year concomitant comparison of the porcine bioprosthetic and tilting disk prosthetic aortic valves. J Thorac Cardiovasc Surg 1984;88:695-705. 12. Blackstone EH, Kirklin JW. Death and other time-related events after valve replacement. Circulation 1985;72:753-67. 13. Goldman B, Scully H, Tong C, et al. Clinical results of pericardial xenograft valves: the lonescu-Shiley and Hancock valves. Can J Cardiol 1988;4:328-32. 14. Bortolotti U, Milano A, Thiene G, et al. Early mechanical failures of the Hancock pericardial xenograft. J Thorac Cardiovasc Surg 1987;94:200-7.
