D derly population has consisted chiefly of evaluation

Transcription

1 Cardiac Valve Replacement in the Elderly: Clinical Perf ormance of Biological Prostheses W. R. Eric Jamieson, MD, Lawrence H. Burr, MD, A. Ian Munro, MD, Robert T. Miyagishima, MD, and Alfred N. Gerein, MD Vancouver General Hospital and St. Paul's Hospital, The University of British Columbia, Vancouver, British Columbia, Canada From 975 to 987,,27 elderly patients underwent,223 valve replacements with the Carpentier-Edwards standard or supraannular porcine bioprostheses in,47 operations. Of the total patient population seen during these years, 33.5% receiving a standard porcine bioprosthesis and 48.6% receiving a supraannular bioprosthesis were 65 years of age or older. Of this elderly patient population, 465 patients were between 65 and 69 years old; 68 patients, 7 and 79 years old; and 52 patients, 8 years old and older. Aortic valve replacement was performed in 635 patients, mitral valve replacement in 47 patients, tricuspid valve replacement in 2 patients, and multiple-valve replacement in 8 patients. The cumulative follow-up was 3,957 patient-years. Early mortality was 9.5%: 7.3% for the 65- to 69-year-old group,.7% for the 7- to 79-year-old group, and 5.4% for the group 8 years old and older. Late mortality was 5.5% per patientyear: 4.2% per patient-year for the 65- to 69-year-old group, 6.3% per patient-year for the 7- to 79-year-old group, and 4.% per patient-year for the group 8 years old and older. Valve-related causes contributed to 7 early deaths and 33 late deaths. The overall patient survival, including operative deaths, was 7.7% f.6% at 5 years and 47.8% f 3.7% at and 2 years. The freedom from all valve-related complications was 52.%? 6.% at and 2 years. The overall rate of valve-related complications was 5.% per patient-year (fatal complications,.3% per patient-year). The overall rate of thromboembolism was 2.3% per patient-year and the freedom from thromboembolism, 69.6% f 5.2% at and 2 years. The freedom from structural valve deterioration was 8.8% f 8.% at and 2 years: 7.7%?.% at and 2 years for the 65- to 69-year-old group, 97.9% &.2% at years for the 7- to 79-year-old group, and % at 2 years for the group 8 years old and older. At and 2 years, the freedom from valve-related death was 83.7% f 4.3% and the freedom from reoperation, 73.3% f 8.6%. The freedom from valve-related death, residual morbidity from thromboembolism and anticoagulant-related hemorrhage, and reoperation was 6.7% f 7.% at and 2 years. The clinical performance of porcine bioprostheses in the elderly patient population has been excellent. The early mortality increases in patients 7 years old or older. Structural valve deterioration is essentially nonexistent at and 2 years in patients 7 years of age or older. (Ann Thorac Surg 989;48:73-85) ocumentation of valvular cardiac surgery in the el- D derly population has consisted chiefly of evaluation of early and late mortality in justification of the operation [-28].The quality of life of this patient population has also been given considerable attention [3, 4, 6-8, 2, 3, 7-9]. Detailed evaluation of valve-related complications with either biological or mechanical prostheses has not been performed. The experience of The University of British Columbia with cardiac valve replacement was originally documented in 98 [4] and 98 [5]. The role of porcine bioprostheses is presented here in an evaluation of,27 consecutive patients aged 65 years and older. Material and Methods The teaching hospitals of The University of British Columbia, Vancouver General Hospital and St. Paul's Hospital, have used porcine bioprostheses since 975. From 975 to Presented at the Twenty-fourth Annual Meeting of the Society of Thoracic Surgeons, New Orleans, LA, Sep 2628, 988. Address reprint requests to Dr Jamieson, Division of Cardiovascular and Thoracic Surgery, Department of Surgery, University of British Columbia, 9 W th Ave, No. 3, Vancouver, BC, Canada V5Z 4E3. 987, the Carpentier-Edwards standard porcine bioprosthesis was implanted in,83 patients and the Carpentier- Edwards supraannular porcine bioprosthesis, in,536 patients. Of the total patient population, 4.4% were 65 years of age and older, 33.5% of the patients receiving a standard bioprosthesis between 975 and 982 and 48.6% of the population given a supraannular bioprosthesis from 982 to 987. The elderly patient population is detailed in Tables through 4. There were,27 patients who received,223 Table. Patient Population by Age Total Total Total Age (yr) Patients Operations Valves G Total,27",47,223 a Eight patients are in both the 65- to 69-year-old group and the 7- to 79-year-old group. 989 by The Society of Thoracic Surgeons /89/$3.5

2 ~ ~~ 74 JAMIESON ET AL Ann Thorac Surg 989;48: 7M35 Table 2. Patient Population by Valve Replaced Type of Valve Total Total Total Replacement Patients Operations Valves Aortic Mitral Tricuspid Multiple = Total,27b,47,223 a This represents 69 aortic valves, 79 mitral valves, and 8 triscupid valves Five patients are in both the aortic and mitral valve replacement groups, patient is in both the aortic valve and multiple-valve replacement groups, and is in both the mitral and multiple-valve replacement groups valves in,47 operations. Four hundred sixty-five patients were between 65 and 69 years old, 68 were between 7 and 79 years old, and 52 were 8 years old and older (see Table ). Aortic valve replacement (AVR) was performed in 635 patients, mitral valve replacement (MVR) in 47 patients, tricuspid valve replacement in 2 patients, and multiple-valve replacement in 8 patients (see Table 2). The distribution of patients in the Carpentier-Edwards porcine standard and supraannular bioprostheses subsets is shown in Tables 3 and 4. The average follow-up for the three age groups was 3.9 years for patients 65 to 69 years old, 3.2 years for patients 7 to 79 years old, and 2.5 years for the group 8 years old or more. The cumulative follow-up was 3,956.7 patientyears:,845.7 patient-years for the 65- to 69-year-old group,,983.2 patient-years for the 7- to 79-year-old group, and 27.8 patient-years for the group 8 years old or older (Table 5). Of the cumulative follow-up, AVR accounted for 2,329. patient-years; MVR,,397.2 patientyears; and multiple-valve replacement, patientyears (Table 6). Results Early mortality was 9.5% (9 patients): 7.3% for the 65- to 69-year-old group,.7% for the 7- to 79-year-old group, and 5.4% for the group 8 years old or more (Table 7). Forty patients undergoing AVR died, 58 having MVR died, and undergoing multiple-valve replacement died Table 3. Patient Population for Carpentier-Edwards Standard Biopros thesis Age Group Total Total Total (Yd Patients Operations Valves Total 396a 4 46 a Three patients are in the 65- to 69-year-old group twice, and patient is in the 7- to 79-year-old group twice. Table 4. Patient Population for Carpentier-Edwards Supraannular Bioprosthesis Age Group Total Total Total (Yr) Patients Operations Valves Total 746a a One patient is in the 65- to 69-year-old group twice, and 3 patients are in the 7- to 79-year-old group twice. (Table 8). The causes of early death are detailed in Table 9. The valve-related causes included thromboembolism (3 patients), prosthetic valve endocarditis ( patient), paraprosthetic leak (2), and hemorrhage as a result of antithromboembolic therapy (). Early mortality was influenced by concomitant procedures and previous cardiac operations. Early mortality was 2.8% among patients with concomitant procedures and 6.8% among those without additional procedures (Table lo), and 3.2% among patients with previous cardiac operations and 9.2% among those without such operations (Table ). The causes of early death by age group are documented in Table 2. Late mortality expressed as a linearized occurrence rate was 5.5% per patient-year overall; it was 4.2% per patientyear for the 65- to 69-year-old group, 6.3% per patientyear for the 7- to 79-year-old group, and 4.% per patient-year for patients 8 years old or more (Table 3). Late mortality for AVR was 5.2% per patient-year; for MVR, 6.2% per patient-year; and for multiple-valve replacement, 5.4% per patient-year (Table 4). The causes of late death are shown in Table 5. The valve-related causes included thromboembolism (9 patients), prosthetic valve endocarditis (5), paraprosthetic leak ( patient), hemorrhage due to antithromboembolic therapy (4), clinical valve dysfunction (l), and structural valve deterioration (3). patient survival, including operative deaths, was 7.7% &.6% at 5 years and 47.8% f 3.7% at and 2 years (Fig ). For the group 65 to 69 years old, survival was 77.5% k 2.% at 5 years and 6.9% f 4.4% at and 2 years; for the group 7 to 79 years old, survival was 67.6% f 2.3% at 5 years; and for the group 8 years old or more, survival was 37.7% f 6.2% at 3 years. The freedom from all valve-related complications, mor- Table 5. Cumulative Follow-up by Age Age Total Total (Yr) Operations Patient Total ,47,845.7, ,956.7

3 Ann Thorac Surg 989;48:73-85 JAMIESON ET AL 75 Table 6. Cumulative Follow-up by Valve Replaced Type of Valve Total Total Replacement Operations Patient- Aortic Mitral Tricuspid Mu tiple Total ,47 2,239., ,956.7 Table 8. Early Deaths By Value Replaced" Type of Valve Total Total Replacement Operations Deaths % Aortic Mitral Tricuspid 2 Multiple Total, a Early mortality is 3-day mortality. bidity and mortality, was 52.% * 6.% at and 2 years (Fig 2). At years, the freedom from all valve-related complications was 5.9% * 8.3% for the 65- to 69-year-old group, 53.8% & 8.5% for 7- to 79-year-old group, and 56.4% & 23.6% for the group 8 years old and older. The hazard intervals by year indicating yearly rates and mortality rates are demonstrated in Figure 2. The linearized occurrence rate for valve-related complications was 5.% per patient-year (63 events) and the mortality rate,.3% per patient-year (4 events). The overall rate for valve-related complications was 4.% per patient-year for the 65- to 69-year-old group (.7% per patient-year, fatal), 4.4% per patient-year for the 7- to 79-year-old group (.2% per patient-year, fatal), and 3.9% per patient-year for the group 8 years old or older (3.2% per patient-year, fatal). The linearized occurrence rate and number of events for all valve-related complications are shown in Table 6 by age group and in Table 7 by valve replaced. The overall rates were as follows: thromboembolism, 2.3% per patient-year (92 events); antithromboembolic therapyrelated hemorrhage,.6% per patient-year (23 events); prosthetic valve endocarditis,.3% per patient-year ( events); paraprosthetic leak,.3% per patient-year (2 events); structural valve deterioration,.4% per patientyear (5 events); and clinical valve dysfunction,.3% per patient-year (ten events). The mortality rate for thromboembolism was.6% per patient-year (22 deaths); antithromboembolic therapy-related hemorrhage,.% per patient-year (5 deaths); prosthetic valve endocartidis,.2% per patient-year (6 deaths); paraprosthetic leak,.8% per patient-year (3 deaths); structural valve deterioration,.8% per patient-year (3 deaths); and clinical valve dysfunction,.3% per patient-year ( death). Reoperation for valve-related complications was per- Table 7. Early Deaths by Age" Age Total Total (Yr) Operations Deaths % Total, a Early mortality is 3-day mortality. formed in 33 patients (.8% per patient-year) (see Table 6). The linearized occurrence rate for AVR was.3% per patient-year (eight events); for MVR,.6% per patientyear (22 events); and for multiple-valve replacement,.4% per patient-year (three events) (see Table 7). The mortality rate for reoperation was.2% per patient-year (7 deaths). Of the 7 patients who died, reoperation was necessary for structural valve deterioration (3), prosthetic valve endocarditis (2), paraprosthetic leak (l), and clinical valve dysfunction (). The morbidity and mortality associated with thromboembolism by age group are detailed in Table 8. The overall linearized occurrence rate for major events was.6% per patient-year (64 events). The overall linearized rate for minor events was.7% per patient-year (28 events). The mortality rate for the group 65 to 69 years old was.3% per patient-year (6 deaths); for the group 7 to 79 years old,.7% per patient-year (4 deaths); and for the group 8 years old or more,.6% per patient-year (2 deaths). The morbidity and mortality associated with thromboembolism are also detailed by valve replaced (Table 9). The freedom from individual valve-related complications is shown in Figures 3 through. The linearized occurrence rate of each complication, depicting total and fatal events, by yearly hazard intervals is also illustrated Table 9. Causes of Early Death Cause No. of Patients Myocardial infarctiodarrhythmia 28 Low output syndrome/congestive heart failure 35 Thromboembolism Prosthetic valve endocarditis Paraprosthetic leak Antithromboembolic therapy-related hemorrhage Hemorrhage 6 Aortic dissection Other 22 Total 9= a This represents 4 deaths among patients having aortic valve replacement, 58 deaths among patients having mitral valve replacement, and deaths among patients having multiple-valve replacement.

4 76 JAMIESON ET AL Ann Thorac Surg 989;48: 7>85 Table. Early Deaths With and Without Concomitant Procedure Type of Valve With Without Replacement Operations % Deaths Operations % Deaths Aortic Mitral Tricuspid Multiple Total Table. Early Deaths With and Without Previous Cardiac Procedure Type of Valve With Without Replacement Operations % Deaths Operations % Deaths Aortic Mitral Tricuspid Multiple Total , for each complication. The freedom from thromboembolism was 69.6% rt: 5.2% overall at and 2 years (see Fig 3). The freedom for the 65- to 69-year-old group was 78.2% % at and 2 years; for the 7 to 79-year-old group, 9.8%?.5% at 5 years; and for the group 8 years and older, 88.8% * 7.6% at 5 years. The freedom from thromboembolism by valve replaced is shown in Figure 4. The freedom for AVR was 65.5% f 6.5% at and 2 years; for MVR, 79.3% & 5.5% at years; and for multiple-valve replacement, 97.% 2 2.% at years. Thromboembolism and related deaths occurred throughout the observation period with a trend toward increased occurrence in the latter years of the period. The freedom from hemorrhage related to antithrom- Table 2. Causes of Early Death by Age Age (yr) Cause Myocardial infarctiodarrhythmia Low output syndrome/ congestive heart failure Thromboembolism Prosthetic valve endocarditis Paraprosthetic leak Antithromboembolic therapyrelated hemorrhage Hemorrhage Aortic dissection Other Total boembolic therapy was 97.6% f.6% at 5 years and 95.3% rt:.3% at and 2 years (see Fig 5). There was no appreciable difference between the three age groups. This complication occurred throughout the observation period. The freedom from prosthetic valve endocarditis was 99.%?.3% at 5 years and 97.7% +.9% at and 2 years (see Fig 6). Again, there was no difference between the age groups. This complication occurred early and midterm in the observation period. The freedom from paraprosthetic leak was 98.8% rt:.4% at 5 years and 98.2% f.7% at and 2 years (see Fig 7). Paraprosthetic leak occurred only in the 65- to 69-year-old and 7- to 79-year-old groups. The freedom from clinical valve dysfunction was 98.8% &.4% at 5 years and 95.9% rt: 2.6% at and 2 years (see Fig 8). This complication occurred only in the 65- to 69-year-old and 7- to 79-year-old groups. The freedom from structural valve deterioration was 99.% &.4% at 5 years and 8.8% +- 8.% at and 2 years (see Fig 9). The freedom for the 65- to 69-year-old group was 7.7% f.% at and 2 years; for the 7- to 79-year-old group, 97.9%?.2% at years; and for the group 8 years old and older, % at 2 years. The freedom from structural Table 3. Late Deaths by Age Age Total Total (Yr) Patient- Deaths %/Patient-Year 65-69, , Total 3,

5 Ann Thorac Surg 989:48: 7M5 JAMIESON ET AL 77 Table 4. Late Deaths by Valve Replaced Type of Valve Total Total Replacement Patient- Deaths %/Patient-Year Aortic 2, Mitral, Tricuspid Multiple Total 3, ap 7 - eo- 6 - valve deterioration for AVR was 99.4% k.6% at and 2 years; for MVR, 98.3%?.% at 5 years; and for multiple-valve replacement, 95.2%? 4.6% at 5 years (see Fig ). Structural valve deterioration commenced at 4 years and increased yearly, with a large increase during the tenth year of observation. The multiple decrement analysis of valve-related complications is shown in Figures through 6. The freedom from reoperation was 97.8%?.6% at 5 years and 73.3% k 8.6% at and 2 years (see Fig ). The freedom from reoperation at and 2 years was 63.3% 2.2% for the 65- to 69-year-old group, 94.5% *.8% for the 7- to 79-year-old group, and % for the group 8 years old and older. Reoperation occurred throughout the observation period, but the late increase parallels structural valve deterioration. The freedom from valve-related death was 96.3% 2.7% at 5 years and 83.7%? 4.3% at and 2 years (see Figs 2, 3). The freedom from death was 9.2% * 4.5% at and 2 years for the 65- to 69-year-old group, 77.5% * 8.2% at years for the 7- to 79-year-old group, Table 5. Causes of Late Death Cause Myocardial infarctionlarrhythmia Congestive heart failure T h r o m b o e m b o ism Prosthetic valve endocarditis Paraprosthetic leak Structural valve deterioration Antithromboembolic therapy-related hemorrhage Clinical valve dysfunction Gastrointestinal hemorrhage Hemorrhage Sepsis Sudden unexpected death (arrhythmia) Noncardiac-related cause Other Total No. of Patients a This represents 2 deaths among patients having aortic valve replacement, 86 deaths among patients having mitral valve replacement, death among patients having tricuspid valve replacement, and 2 deaths among patients having multiple-valve replacement overall lmo lyr 3yr 6yr 9yr 2yr Follow-up Time Fig. Patient suraival. and 59.5% -t 24.7% at years for the group 8 years old and older (see Fig 2). Death from valve-related causes occurred throughout the observation period but increased late with reoperation and the higher rate of thromboem- Table 6. Linearized Occurrence Rates of Valve-Related Complications by Age Morbidity, Mortality Thromboembolism Antithromboembolic therapy-rela ted hemorrhage Clinical valve dysfunction Prosthetic valve endocarditis Paraprosthetic leak Structural valve deterioration Reoperation 34 (.8) 55 (2.8) 3 (2.4) 92 (2.3) 6 (.3) 4 (.7) 2 (.6) 22 (.6) 2 (.7) (.5) (.8) 23 (.6) 3 (.2) (.5) (.8) 5 (.) 5 (.3) 5 (.3) (.3) l(.5) (.3) 4 (.2) 6 (.3) (.8) (.3) (.5) 4 (.2) (.8) 6 (.2) 6 (.3) 6 (.3) 2 (.3) l(.5) 2 (.) 3 (.8) 2 (.7) 3 (.2) 5 (.4) 2 (.) (.5) 3 (.8) 2 (.) 3 (.7) 33 (.8) 3 (.2) 4 (.2) 7 (.2) a Data in parentheses are percent per patient-year or episodes per patient-years.

6 78 JAMIESON ET AL Ann Thorac Surg 989; loo] g 7- U Q p! ao- Y overall ( a- z 3- nnnnnp n,, II l I I I I I I /, alive Fig 2. Freedom from all valve-related complications. bolism. The freedom from valve-related death was 84.2% f 5.2% for AVR at and 2 years, 8.6%? 9.4% for MVR at years, and 96.5% f 2.5% for multiple-valve replacement at years (see Fig 3). The freedom from valve-related death and reoperation was 94.6% f.9% at 5 years and 64.3% f 8.% at and 2 years (see Fig 4). The freedom for the 65- to 69-year-old group was 6.2%?.8% at and 2 years; for the 7- to 79-year-old group, 74.4% & 8.% at years; and for the group 8 years old and older, 59.5% f 24.7% at 8 and years. The hazard rates parallel the rates of structural valve deterioration and deaths related to reoperation. The freedom from valve-related death and residual morbidity (treatment failure) was 92.8%?.% at 5 years and 75.4% f 4.5% at and 2 years (see Fig 5). The freedom for the 65- to 69-year-old group was 8.8%? 5.4% at and 2 years; for the 7- to 79-year-old group, 69.6% f 7.8% at years; and for the group 8 years old and older, 56.4% f 23.6% at 8 and years. This assessment considers patients who died of valve-related complications or had residual neurological deficits from thromboembolism, antithromboembolic therapy-related hemorrhage, or prosthetic valve endocarditis. The hazard rates indicate the mortality and residual morbidity occurrence rates. The freedom from valve-related death, residual morbidity, and reoperation was 6.7% f 7.% at and 2 years (see Fig 6). The freedom for the 65- to 69-year-old group was 6.5% f 9.7% at and 2 years; for the 7- to 79-year-old group, 67.4%? 7.6% at years; and for the group 8 years old and older, 56.4%? 23.6% at 8 and years. The hazard linearized occurrence rates indicate an increasing incidence throughout the observation period. Table 7. Linearized Occurrence Rates of Value-Related Complications by Value Replaced" Type of Valve Replacement Morbidity, Mortality Aortic Mitral Multiple Thromboembolism Antithromboembolic therapy-related hemorrhage Clinical valve dysfunction Prosthetic valve endocarditis Paraprosthetic leak Structural valve deterioration 54 (2.3) 3 (.6) 6 (.3) 2 (.9) 4 (.2) l(.4) 5 (.2) 4 (.2) 4 (.2) l(.4) 36 (2.6) 8 (.6) 4 (.) 3 (.2) 6 (.4) 4 (.3) l(.7) 7 (.5) 2 (.) (.4) 3 (.9) 3 (.2) 2 (.9) (.5) 3 (.4) 2 (.9) (.5) l(o.5) l(o.5) 92 (2.3) 22 (.6) 23 (.6 5 (.) (.3) l(.3) (.3) 6 (.2) 2 (.3) 3 (.8) 5 (.4) 3 (.8) Reoperation 8 (.3) 22 (.6) 3 (.4) 33 (.8) 3 (.) 3 (.2) (.5) 7 (.2) a Data in parentheses are percent per patient-year, or episodes per patient-years.

7 Ann Thorac Surg 989;48: JAMIESON ET AL 79 Data in parentheses are percent per patient-year or episodes per patient-years. The preoperative and postoperative New York Heart Association functional classification is shown in Figure 7. Preoperatively, 92.3% of the patients were in either class I or IV, whereas postoperatively, 93.% were in class I or. The overall experience is summarized in Tables 2 through 23. The causes of the 33 reoperations are shown in Table 2 and the causes of the valve-related deaths (7 early and 33 late), in Table 2. The freedom from each valve-related complication and composites of complications are summarized in Tables 22 and 23. eo overall i b n Fig 3. Freedom from thromboembolism by age. alive Comment The majority of reports on valve replacement have dealt with the safety of the procedure and the general quality of life of patients following valve procedures [l-29. There has been limited documentation of valve-related complications in the elderly population [7]. In previous reports, we [4, 5 recommended that indications for valve replacement in the elderly be similar to those for the general population and that advanced age alone not be a contraindication to surgical management. The operative mortality in various series of elderly patients was summarized in 98. During the 96s, the operative mortality for all reported valve replacements ranged between 5.3% and 4% [l, 2, &3]. During the decade of the 97s, the mortality fell to a range of 3.7% to.7% [%5, 4, 5. In the 96s, the operative mortality for AVR ranged from 8.7% to 33%; for MVR, 2% to 37%; and for multiple-valve replacement, 5% to 75%. In the 97s, the mortality range fell appreciably; for AVR, % to 6%; for MVR, 5.6% to 9%; and for multiple-valve replacement. % to 33%. Table 9. Thromboembolism by Valve Replaced" Type of Valve Replacement Morbidity, Mortality Aortic Mitral Multiple The operative mortality we [5] reported for the 97s was 8.8% overall and 4.7% for AVR,.4% for MVR, and 37.5% for multiple-valve replacement. The present series has some overlap with the previous series, for all operations with porcine bioprostheses from 975 to 987 are included. We know of no report that considers exclusively 6 6; Minor 2 (.9) 8 (.6) 28 (.7)..._ J _..._ I 4;! MVR MR 32 5 overall n I Major 34 (.5) 28 (2.) 2 (.9) 64 (.6) alive 3 (.6) (o.6) (.5) 22 (o'6) Fig 4. Freedom from thromboembolism by type of valve replacement. a Data in parentheses are percent per patient-year or episodes per (AVR = aortic valve replacement; MR = multiple-valve replacement; patient-years. MVR = mitral value replacement.)

8 8 JAMIESON ET AL Ann Thorac Surg 989;48:73-85 ' " :i 7 : :! ; overall 3 8 a 2 \ 4 a Rl2 i l _. n o i z 3 4 s a 7 8 a i o i i i z alive Fig 5. Freedom from hemorrhage related to antithromboembolic therapy. E 'p a3 t # - 5 " \ # 8-3 (I a U o z 3 4 s a 7 a a 2 Fig 7. Freedom from paraprosthetic leak. alive ao-+ overall the mortality in the 98s. Early mortality in the present overlapping series was 9.5% overall and 6.3% for AVR, 3.5% for MVR, and 3.8% for multiple-valve replacement. Only multiple-valve replacement showed improvement compared with our data for the 97s. The series presented here included a substantial number of high-risk patients, for 46.6% had concomitant procedures and.% had had previous cardiac operations. The mortality for valve replacements performed with concomitant procedures was 2.8% and for those without, 6.8%. The influence of previous procedures on mortality was also evident; 9.2% of patients without and 3.2% of those with such a procedure died early. Another extensive series with a group of 36 patients I # 7 " : :" :! 3 a a ao overall # 7o-l i ao overall \ a R 4 t $ i z o i s a 7 8 a (z \ 4 : ' j a$ Rl 2 o z 3 4 s a 7 a a 2 Fig 6. Freedom from prosthetic valve endocarditis. alive Fig 8. Freedom from clinical valve dysfunction. alive

9 Ann Thorac Surg 989:48: 7-5 JAMIESON ET AL 8 8 a, 8 t 74 " ii i! overall iy U \ 4 a R l S i O 7 B i (2 alive Fig 9. Freedom from structural valve deterioration by age. ar % overall 3 8 B 2 ld4-l \ a a R s o Fig. Freedom from reoperation. alive older than 7 years is that reported by Bessone and colleagues [7] in 985. This study revealed an early mortality of.5% for AVR, 5.7% for MVR, and 4.7% for multiple-valve replacement. In the present series, considering only patients 7 years of age and older, the overall early mortality was.% (75/676 patients). In 987, Blakeman and co-workers [8] reported only a 3% mortality among patients 75 years of age and older undergoing isolated AVR. Craver and colleagues [2], in 984, documented an early mortality of 5.% for isolated AVR and of 5.4% for AVR and concomitant coronary AVR MVR MR overall 3 6 i n 'il s o to il ;2 olive Fig. Freedom from structural valve deterioration by type of valve replacement. (AVR = aortic valve replacement; MR = multiple-valve replacement; MVR = mitral valve replacement.) i ar 4 "j \ \ \--r i i 2! l+--+l- ar B 2 Fig 2. Freedom from valve-related death by age. alive 65-69

10 82 JAMIESON ET AL Ann Thorac Surg 989;48:7=5 " :AVR MVR MR overall ; E B t" " overall i 3 i i i2 6 s 2!.a R 3 l+--d-l : alive Fig 3. Freedom from valve-related death by type of valve replacement. (AVR = aortic valve replacement; MR = multiple-valve replacement; MVR = mitral valve replacement.) 6 s 2 L 3 : acve Fig 5. Freedom from valve-related death and residual morbidity (treatment failure). artery bypass grafting in a series of 52 patients older than 7 years. The late mortality is of considerable importance and was summarized in a previous study [5]. Late mortality has been as great as 9.% per patient-year [3]. In the large series of Bessone and coauthors [7], late mortality was 7.6% per patient-year. In the present series, the rate was 5.5% per patient-year (AVR, 5.2% per patient-year; MVR, 6.2% per patient-year; and multiple-valve replacement, 5.4% per patient-year). By age group, late mortality 8 - E 7- U 7 U * % overd, I I 3 6 (2 n 6 lb+dd-l % 24 ar K) 2 I fatd afive Fig 4. Freedom from valve-related death and reoperation. 3-6 s 2 & 2 9- \ - ar overall 3 6 ti n nnn asve Fig 6. Freedom from valve-related mortality, residual morbidity, and reopera t ion.

11 Ann Thorac Surg 989;48:7M5 JAMIESON ET AL 83 Table 2. Causes of Valve-Related Death Complication Early Late Total 'O 8 eo 4 2 I II IV NYHA D PREOP I POSTOP Fig 7. New York Heart Association (NYHA) classification before and after operation. was 4.2% per patient-year for the group 65 to 69 years old, 6.3% per patient-year for the group 7 to 79 years old, and 4.% per patient-year for the group 8 years old and older. The overall patient survival was 7% at 5 years and 48% at and 2 years. Survival was 65% at 6 years in the series of Bessone and associates [7]. To our knowledge, this report provides the first overall assessment of valve-related complications of valve procedures in an elderly population. The freedom from all valve-related complications (including thromboembolism, hemorrhage related to antithromboembolic therapy, prosthetic valve endocarditis, paraprosthetic leak, and structural valve deterioration) at and 2 years was 52%. At and 2 years, the freedom from thromboembolism was 7% and from structural valve deterioration, 8%, while the freedom from all other complications was greater than 95%. Thromboembolic events increased with time, thus indicating other possible causes of thromboembolism attributed to the prostheses without detailed investigation. Structural valve deterioration requires special consideration. The overall freedom from this complication at Thromboembolism Antithromboembolic 4 5 therapy-related hemorrhage Clinical valve dysfunction Prosthetic valve 5 6 endocarditis Paraprosthetic leak 2 3 Structural valve 3 3 deterioration Total" 7 (9) 33 (29) 4 (328) Numbers in parentheses are the total number of deaths. and 2 years was 8.8%. There was a substantial difference in the age groups: 7.7% for the 65- to 69-year-old group at and 2 years, 97.9% for the 7- to 79-year-old group at years, and % for the group 8 years old or older at 2 years. This information illustrates the considerable safety of porcine bioprostheses for patients older than 7 years. The performance of the aortic prostheses was superior to that of the mitral prostheses both for MVR and multiple-valve replacement. Moront and Katz [3] reported calcification of a mitral prosthesis in a 72- year-old woman on a regimen of calcium carbonate therapy. The overall performance of the porcine bioprostheses is best illustrated by considering the multiple decrement analysis of the complications. The freedom from reoperation, for all complications, was 73.3% at and 2 years and was greater in patients 7 years old or older compared with those less than 7 years of age, thus reflecting the influence of structural valve deterioration. The overall freedom from valve-related death was 83.7% at and 2 Table 22. Freedom From Valve-Related Complications" Table 2. Causes of Valve-Related Reoperation Complication Thromboembolism Antithromboembolic therapy-related hemorrhage Clinical valve dysfunction Prosthetic valve endocarditis Paraprosthetic leak Structural valve deterioration Total a This total accounts for all the reoperations Total " Complication 5 2 Thromboembolism 9.8 t t t 5.2 Antithromboembolic 97.6 t r.3 therapy-related hemorrhage Clinical valve dysfunction 98.8 t f t 2.6 Prosthetic valve t endocarditis Paraprosthetic leak 98.8 t t f.7 Structural valve 99. t t t 8. deterioration Reoperation 97.8 t t t 8.6 a Data are shown as the percent freedom -C the standard error.

12 84 JAMIESON ET AL Ann Thorac Surg 989;48: 7H5 Table 23. Freedom From Valve-Related Composite Complications" Complication 5 2 Valve-related death t and reoperation Valve-related death 96.3 f rl f 4.3 Treatment failure 92.8 f f f 4.5 Valve-related death, 9.3 f. 6.7 f f 7. residual morbidity, and reoperation All valve-related 86.7 f f f 6. complications a Data are shown as the percent freedom L the standard error. years; the age influence was reversed with the greater freedom in patients younger than 7 years of age, thereby reflecting the late mortality from thromboembolism in the older age groups (7 years old and more). The freedom from valve-related death and reoperation was 64.3% at and 2 years, while the freedom from valve-related death and residual morbidity was 75.4%. Residual morbidity is defined as permanent neurological deficit from thromboembolism, hemorrhage due to antithromboembolicrelated therapy, and embolic complications of prosthetic valve endocarditis. The most important index of performance is the 6.7% freedom from the combination of valve-related death, residual morbidity, and reoperation at and 2 years. The porcine bioprostheses provide excellent clinical performance in the elderly. The risk of structural valve deterioration (primary tissue failure) is low and essentially negligible at and 2 years in patients older than 7 years. We are indebted to our colleagues Dr G. F.. Tyers, Dr M. T. Janusz, Dr H. Ling, Dr P. Allen, Dr H. Tutassaura, and Dr R. I. Hayden for allowing the inclusion of their patients in this study. We extend appreciation to Eva Germann, computer programmer, and Joan MacNab and Florence Chan for their extensive efforts in the preparation of this work. References. Austen WG, De Sanctis RW, Buckley MJ, Mundth ED, Scannell JG. Surgical management of aortic valve disease in elderly. JAMA 97;2: Amad A, Starr A. Valve replacement of geriatric patients. Br Heart J 969;3: Bessone LN, Pupello DF, Blank RH, Harrison EE, Sbar S. Valve replacement in patients over 7 years. Ann Thorac Surg 977;24: Henze A. Aortic valve replacement in patients over the age of 6. Scand J Thorac Cardiovasc Surg 974;8:. 5. Quinlan R, Cohn LH, Collins JJ. Determinants of survival following cardiac operations in elderly patients. Chest 975; 68: Oh W, Hickman R, Emanuel R, et al. Heart valve surgery in 4 patients over the age of 6. Br Heart J 973;35: Shanahan MX, Windsor HM, Golding L. Open heart surgery in the elderly. Aust N Z J Surg 97;32: Barnhost DA, Giuliani ER, Pluth JR, Danielson GK, Wallace RB, McGoon DC. Open-heart surgery in patients more than 65 years old. Ann Thorac Surg 974;8: Guthrie RB, Sepllberg RD, Benedict JS, Buhl TL. Open heart valve surgery in patients 65 and older. Arch Surg 972;5:43.. Hildner FJ, Linhart JW, Samet P, Piccinini J, Masten JL, Greenberg JJ. Clinical and hemodynamic comparisons of valve replacement in patients over and under age 6. Ann Thorac Surg 969;743W5.. Bowles LT, Hallman GL, Cooley DA. Open heart surgery on the elderly: results of 54 patients sixty years of age or older. Ann Thorac Surg 969; Finegan RE, Gianelly RE, Harrison DC. Aortic stenosis in the elderly: relevance of age to diagnosis and treatment. N Engl J Med 969;28: Copeland JG, Griepp RB, Stinson EB, Shumway NE. Isolated aortic valve replacement in patients older than 65 years. JAMA 977; Jamieson WRE, Thompson DM, Munro AI. Cardiac valve replacement in elderly patients. Can Med Assoc J 98; 23: Jamieson WRE, Dooner J, Munro AI, et al. Cardiac valve replacement in the elderly: a review of 32 consecutive cases. Circulation 98;64(Suppl 2): Beall AC. Prosthetic heart valves for the elderly. Med Instrum 982;6(4): Bessone LN, Pupello DF, Blank RH, Lopez-Cuenea E, Hiro SP, Ebra G. Valve replacement in the elderly: a long term appraisal. J Cardiovasc Surg (Torino) 985;26: Blakeman BM, Pifarre R, Sullivan HJ, et al. Aortic valve replacement in patients 75 years old and older. Ann Thorac Surg 987;44: Commerford PJ, Curcio A, Albanese M, Beck W. Aortic valve replacement in the elderly. S Afr Med J 98;59(27): Craver JM, Goldstein J, Jones EL, Knapp WA, Hatcher CR Jr. Clinical, hemodynamic and operative descriptors affecting outcome of aortic valve replacement in elderly versus young patients. Ann Surg 984;99: DeBono HB, English TAH, Milstein BB. Heart valve replacement in the elderly. Br Med J 978;2: Hochberg MS, Derhac WM, Conkle DM, McIntosh CL, Epstein SE, Morrow AG. Mitral valve replacement in elderly patients: encouraging postoperative clinical and hemodynamic results. J Thorac Cardiovasc Surg 979;77: Jones TW, Thomas GI, Stavney LS, Monhas DR. Aortic valve replacement and the senior citizen. Am Surg 979;45: Nicolaou N, Kinsley RH. Mitral valve replacement in the elderly. S Afr Med J 984;65(5):59MOO. 25. Santinga JT, Flora J, Kirsh M, Baublis J. Aortic valve replacement in the elderly. J Am Geriatr SOC 983;3(4): Schulte HD, Bircks W, Drian A, Matejic 8. Coronary and valvular surgery in elderly patients (>7 years). Thorac Cardiovasc Surg 987;35: Storstein, Efskind L. Aortic valve replacement in elderly patients. Acta Med Scand 979;26: Stephenson LW, MacVaugh H, Edmunds LH Jr. Surgery using cardiopulmonary bypass in the elderly. Circulation 978;58: Hines GL, Boa BH, Reed GE. Safety of aortic valve replacement in septuagenarians. NY State J Med 977;77: Moront MG, Katz NM. Early degeneration of a porcine aortic valve bioprosthesis in the mitral position in an elderly woman and its association with long-term calcium carbonate therapy. Am J Cardiol 987;59:6-7.

13 Ann Thorac Surg 989;48:73-85 JAMIESON ET AL 85 DISCUSSION DR CONRAD L. PELLETIER (Montreal, Que, Canada): Dr Jamieson is to be complimented for this very timely and wellpresented study. These data are important in two respects. First, they are important in regard to the target population. Elderly patients represent a growing proportion of patients undergoing valve replacement. It is therefore important to assess the risk/ benefit ratio of this operation in this group of patients. Dr Jamieson has shown that early mortality increases with age, noncardiac-related causes being responsible for 25% of those deaths. We reported similar findings in a study published a couple of years ago, and stressed the need for careful patient selection to decrease noncardiac-related mortality and morbidity. In addition, the percentage of operative deaths is doubled when associated procedures are performed. I ask Dr Jamieson to comment on his approach to associated coronary lesions. Are all stenoses bypassed or only life-threatening ones? Also, what proportion of patients do associated coronary procedures represent in this series? Second and mainly, these data are important in terms of the choice of the proper valve prosthesis for this age group. This study comes at a time when the use of bioprostheses is becoming somewhat more selective. We, like others, have changed our attitude regarding the choice of prosthesis. Bioprostheses represented more than 9% of our total valve implants in 979 and slightly less than 5% in 987. The results reported by Dr Jamieson indicate that the elderly patient is well suited for placement of a bioprosthesis, not only because of the limited life expectancy, but more so because valve durability appears excellent up to years in this age group, thus permitting us to avoid anticoagulants. However, I noticed that there was a higher incidence of thromboembolic complications in older patients. This may be at least in part related to the higher prevalence of atherosclerotic cerebral vessel disease with advanced age. Or could it be due to the absence of anticoagulation, even in patients with risk factors? My second question therefore addresses the problem of anticoagulation, both during the early postoperative period and during long-term follow-up. Finally, what proportion of your patients was kept permanently on a regimen of anticoagulation therapy? Again, I congratulate Dr Jamieson and his colleagues for their excellent and well-documented study. DR CARY W. AKINS (Boston, MA): I congratulate Dr Jamieson on his usual excellent statistical assessment of a large series of patients. I do have one question, however. If you define structural valve deterioration as that determined only at reoperation or autopsy, did patients more than 8 years old actually undergo reoperation as their valves began to deteriorate or were they merely allowed to die? The survival rate for the group 8 years old and more was obviously much lower. The question is, have you somewhat confused the issue, and is your conclusion that patients older than 8 years have little valve deterioration actually valid? DR J. EDWARD OKIES (Portland, OR): I have a question about the size limitation of the prosthesis. Some elderly patients have a very small annulus. What is the lower size limit of tissue valve that you would place, and do you perform any reconstruction of the annulus to allow use of a bigger valve? DR JAMIESON: I thank the discussants for their comments and questions. Dr Pelletier, in regard to your question on coronary artery bypass, we performed concomitant procedures (approximately 95% for coronary artery disease) in 525 of our,27 patients. Coronary artery bypass was performed for hemodynamically significant lesions, up to three per patient. The older patients, those 7 years of age or more, had a higher incidence of thromboembolism. The higher incidence may reflect an increased incidence of extracranial cerebrovascular disease. The patients were not generally investigated to determine the presence or absence of such disease. Thromboembolic events were considered valve-related if there were no known precipitating factors. Long-term anticoagulation is used after MVR in patients with chronic atrial fibrillation. Patients with AVR and MVR who are in sinus rhythm are generally not maintained with long-term anticoagulation. Dr Akins, your comments regarding the patient group 8 years old and older are certainly well taken. The freedom from structural valve degeneration and reoperation is exceptional. There were 52 patients in this group, and 8 early deaths and 8 late deaths. To our knowledge, patients have not been refused reoperation. The patient interviews did not reveal mention, proposal, or refusal of reoperation. It remains possible that family physicians or cardiologists did not give consideration to reoperation for valve-related complications. Dr Okies, the patients in this study were managed without aortic root enlargement. Aortic root enlargement is generally not performed in our institutions. In the majority of patients, the aortic valve size is not smaller than 2 mm. A 9-mm aortic valve is occasionally placed in low-weight women with limitation of activity level. We previously evaluated functional class by valve size, and found no correlation.