- Email: [email protected]
Clinical and Hemodynamic Results with the Carpentier-Edwards Porcine Bioprosthesis
Clinical and Hernodvnamic Results with the Carpentier=kdwardsPorcine Biopmsthesis Conrad Pelletier, M.D., Bernard R. Chaitman, M.D., Richard Baillot, M.D., Pere Guiteras Val, M.D., Raoul Bonan, M.D., and Ihor Dyrda, M.D. primary tissue failure is known to occur over time, particularly in younger patients [6,7]. The Hancock bioprosthesis has been the subject of numerous studies, but relatively few data have been published for the Carpentier-Edwards flexible, support-mounted heterograft. Since tissue preservation methods, stent design, and mounting procedures differ from manufacturer to manufacturer, it is important to evaluate results for the various bioprostheses separately. The Carpentier-Edwards bioprosthesis, which has been used at our hospital since May, 1976, accounts for more than 75'Yo of our valve implants. In a previous report, we showed that this bioprosthesis has good hemodynamic function in medium and large stent sizes but may be mildly stenotic in smaller stent sizes [8]. Since then, the frame of the sewing ring has been modified by further thinning of the cloth and frame structure to increase the ratio of the inner diameter to the outer diameter; this improved annulus model has been available for clinical use since mid-1980. The present report reviews our total experience with the The glutaraldehyde-preserved heterograft is a Carpentier-Edwards porcine valve. Clinical repopular valve device because of its silent func- sults are given, and the hemodynamic perfortion, low rate of hemolysis, and central flow mance of the standard and improved annulus pattern, and because it reduces the risk of sys- models is evaluated. temic embolism without requiring anticoagulation [l-31. Although the hemodynamic per- Material and Methods formance of porcine bioprostheses has been From May, 1976, to January, 1981, 699 found satisfactory, these valves have recently Carpentier-Edwards porcine valves were imundergone several modifications to increase planted in 605 patients. There were 325 women their effective orifice area [4, 51. The principal and 280 men, with a mean age of 52 years concern of their design, however, is durability: (range, 13 to 76 years). Thirty-one patients (5%) were younger than 30 years, and 26 patients (4%) were 70 years of age or older at the time of From the Departments of Surgery and Medicine, Montreal operation. Congestive heart failure, the most Heart Institute and Universite de Montreal Medical School, common indication for operation, was found in Montreal, Que, Canada. Presented at the Eighteenth Annual Meeting of The Society 503 patients (83%). Less frequently encounof Thoracic Surgeons, Jan 11-13, 1982, New Orleans, LA. tered indications included angina in 52 patients Address reprint requests to Dr. Pelletier, Department of (9%), embolic episodes in 15 patients (2.5%), Surgery, Montreal Heart Institute, 5000 E Belanger St, syncope in 8 patients (1%),arid other condiMontreal, Que, Canada HIT 1C8.
ABSTRACT Carpentier-Edwards bioprostheses were implanted in 605 patients, 509 of whom had a single valve replacement, and 96 of whom had a multiple valve replacement. There were 54 early deaths (8.9%) and 26 late deaths (4.3%).The five-year actuarial survival was 87% for aortic valve replacement, 83% for mitral valve replacement, and 81% for multiple valve replacement. Of the 525 survivors, all but 3 were followed for a total of 964 patient-years; 354 patients (68%)remained asymptomatic, and 95 patients (18%) were improved. The incidences of thromboembolism, endocarditis, and reoperation due to primary tissue failure of the bioprosthesis were 2.0, 1.3, and 0.1% per patient-year, respectively. The actuarial probability of being free of all valve-related complications was 93% after five years. Satisfactory hemodynamic performance of the bioprosthesis was demonstrated by postoperative studies done in 70 patients. Thus, the CarpentierEdwards porcine valve provides good clinical improvement, with a low incidence of valve-related complications and tissue failure at five years postoperatively.
612 0003-4975/82/120612-13$01.25 @ 1982 by The Society of Thoracic Surgeons
613
Pelletier et al: Carpentier-Edwards Porcine Bioprosthesis
PREOP
76
POSTOP 354
70
60
304 I
50
40 195
30
20
10
0
nn I
56
I1
FUNCTIONAL CLASS (NYHA)
hundred and nine patients (84%) underwent a single valve replacement; 96 had a multiple valve replacement. Coronary artery bypass grafting (CABG) was the procedure most often associated with valve replacement, and was done in 81 patients (13.4%). A tricuspid antions, such as prosthetic dysfunction, infection, nuloplasty was performed in 38 patients, and or hemolysis, in 23 patients (4%). Four patients the ascending aorta was replaced in 9 patients (0.5%) were asymptomatic and underwent oper- because of aneurysmal dilatation. Of the 704 ation for severe aortic stenosis. The preopera- implants, the standard Carpentier-Edwards tive functional incapacity of the patients bioprosthesis was used in 612 procedures and studied is shown in Figure 1. The majority of the improved annulus model in 87. In 5 patients the patients were in New York Heart Associa- having a multiple valve replacement, other tion Functional Class I11 or IV; patients operated prostheses were implanted (2 Hancock valves, on tor reasons other than heart failure or angina 2 Ionescu-Shiley valves, and 1 Omniscience are included in Class I. One hundred and valve*).The procedures were performed using a twenty patients (20%) had had previous valve bubbler oxygenator, with moderate hemodiluoperations. Predominant mitral stenosis was tion and systemic hypothermia between 22" the most common valve lesion, appearing in and 25°C. During aortic cross-clamping, cold 208 patients (34%), followed by aortic stenosis potassium cardioplegia was used for myocarin 182 patients (30%), and multiple valve dis- dial protection in 539 patients (89%); topical ease in 96 others (16%).Sixty-six patients (11%) cooling of the heart was associated with carwere operated on for pure mitral regurgitation, dioplegia in 283 patients (47%). Oral anticoagulants were given for 2 to 3 months after and 50 (8%), for pure aortic regurgitation. The types of valve replacement and asso- operation unless there was a major contraindiciated surgical procedures done in the patients in this series are shown in Table 1. Five- *Omniscience, Medical Incorporated, Minneapolis, MN. Fig I . N e w York Heart Association (NYHA) Functional Class preoperatively in 605 patients and postoperatively in 522 late survivors. Preoperative Class I includes patients operated on for reasons other than heart failure and angina.
614 The Annals of Thoracic Surgery Vol 34 No 6 December 1982
Table 1. Primary and Associated Surgical Procedures in 605 Patients ~~
Operation
No. with CABG"
Total No.=
Valve replacement MVR AVR MVR MVR MVR
+ AVR
+ tricuspid
+ AVR + tricuspid
Tricuspid Total
Associated procedures Tricuspid annuloplasty Replacement of ascending aorta Closure of ASD Closure of VSD Resection of left ventricular aneurysm Total
33 41 6 0 1 0
(5.5) (6.8) (1.0) (0.4)
81 (13.4)
... ...
... ...
... ...
274 (45.3) 232 (38.3) 86 (14.2) 7 (1.2) 3 (0.5) 3 (0.5) 605 (100.0)
38 (6.3) 9 (1.5) 8 (1.3) 2 (0.3) l(O.2) 58 (9.6)
"Numbers in parentheses are percentages of total (N = 605). CABG = coronary artery bypass grafting; MVR = mitral valve replacement; AVR = aortic valve replacement; ASD = atrial septal defect; VSD = ventricular septal defect.
cation. A permanent regimen of anticoagulation was maintained in 126 patients (23%) because of chronic atrial fibrillation or moderate-to-severe left atrial dilatation. Patients were seen 2, 6, and 12 months after operation for clinical evaluation and were followed yearly thereafter. Data were obtained in 433 patients (83%) through visits to the valve follow-up clinic and (17%) either by telephone interview or by their physicians in 89 others. The data were coded on a special valve clinic coding form and entered into the hospital computer. Recall of patients was facilitated by a computer-generated patient list updated on a monthly basis. A postoperative hemodynamic study was performed in 70 asymptomatic patients 6 to 15 months after operation using previously described techniques [B]. The rest and exercise gradients and effective orifice areas were measured by simultaneous pressure recordings. Mean and peak-to-peak systolic gradients across the aortic bioprosthesis were determined from five cardiac cycles, and the mean diastolic gradient across the mitral bioprosthesis was determined from five diastolic cycles. The effective valve area was calculated using the Gorlin formula, with the
mean gradient determined by manual planimetry [9]. Statistical analysis of the data was performed using Student's t test and the chi-square test. Overall survival and valve-related complications were determined by the lifie-table method [lo] and expressed as mean f standard error of the mean (SEM).
Results Early Mortality The operative and early postoperative (30-day) mortality was 8.9% (54 patients). The early mortality for each surgical procedure is shown in Table 2. The mortality for single valve replacement was less than 5%. Mortality increased to 15.5% in patients having a multiple valve procedure (mitral replacement with tricuspid annuloplasty, or double or triple valve replacement). In patients who had CABG operation associated with valve replacement, the operative mortality was 22.2%; this figure includes 3 patients in whom the ascending aorta was also replaced. Nine of the 54 patients who died early (16.7%) had had a previous valve replacement and were undergoing reoperation for valve dysfunction or additional valve disease.
615
Pelletier et al: Carpentier-Edwards Porcine Bioprosthesis
Table 2 . Early (30-Day) Mortality for Types of Valve Replacement Procedure
No. of Patients
No. of Deaths
Early Mortality (%)
MVR AVR Tricuspid valve replacement Multiple valve replacement Valve replacement + CABG Total
214 191 3 116 81 605
9 9 0 18 18 54
4.2 4.7 0 15.5 22.2 8.9
MVR = mitral valve replacement; AVR
=
aortic valve replacement; CABG = coronary artery bypass grafting.
Table 3 . Causes of Early and Late Death Etiology
No. of Early Deathsa
No. of Late Deathsa
Cardiac causes Heart failure Infarction Sudden death (arrhythmia) Valve-related causes Embolism Reoperation Endocarditis Noncardiac causes Hemorrhage Cerebral ischemia Others Unknown causes Total
43 (80)
10 (38.5)
27 11 5 0
... ... ...
10 (18)
2 1 7 5 (19) 2 2 1 10 (38.5)
8 1 1 1 (2) 54
1 0 9
1 (4) 26
"Numbers in parentheses are percentages of total for each column.
Cardiac problems were responsible for 80% of the early deaths (Table 3 ) . The cause of death is unknown in 1 patient, who died shortly after leaving the hospital.
diac causes in 10 others (see Table 3). In 5 patients, death was caused by valve-related complications. Late survival determined by the actuarial method showed that the five-year probability of survival was 87 k 4% (mean k SEM) for AVR, 83 k 4% for MVR, and 81 f 5% for multiple valve replacement (Fig 2). The overall mortality was directly correlated ( p < 0.001) with the preoperative functional incapacity (Table 4).
Late Mortality There were 26 late deaths (4.3%), which occurred from 2 to 44 months after operation (average, 17 months). Twelve deaths occurred in patients who had mitral valve replacement (MVR), 12 in patients who had aortic valve replacement (AVR), and 2 in patients who had Clinical Improvement multiple valve replacement. There were no Of the 525 survivors, 522 patients (99.4%) were deaths in the 3 patients undergoing isolated re- followed for 6 to 61 months after operation placement of the tricuspid valve. Late mortality (average follow-up, 22 months), for a total averaged 2.7% per patient-year. Death was due of 964 patient-years of follow-up. When last to cardiac causes in 10 patients and to noncar- seen, 354 patients (67.8%) were completely
616 The Annals of Thoracic Surgery Vol 34 No 6 December 1982
aortic mitral multiple
602
464
306
163
32
66
0.5
1
2
3
PTS
I
I
4
5
YEARS AFTER OPERATION
Fig 2. Actuarial survival for aortic, mitral, and multiple valve replacement, including early mortality. Patients at risk at the beginning of each interval are shown along the abscissa. Dashed lines indicate intervals with fewer than 50 patients at risk in each group.
episodes were associated with a residual neurological deficit, and two were fatal. The incidence of embolism for AVR, MVR, and multiple valve replacement was 1.7, 1.5, and 2.8% per patient-year, respectively. After 5 years, the probability, determined by actuarial analysis, asymptomatic, 95 patients (18.2%) had im- of being free of thromboembolic complications proved by at least one functional class, and 73 was 95.2 _+ 1.4% (Fig 3). There were twelve epiothers (14%) showed no improvement or con- sodes of endocarditis (1.3%per patient-year) in tinued to deteriorate postoperatively. Before 10 patients, including one operative infection, operation, 60% of the 605 patients were in and one death. All of these episodes occurred Functional Class I11 or IV,while 93% of the 522 within 2 years of operation, aftser which time survivors remained in Class I or 11 after opera- the probability of remaining free of infection tion (see Fig 1). The rate of clinical improve- was 97.7 f 0.7% (Fig 4). In 3 patients, blood ment was similar for each type of valve re- cultures were negative. placement (p > 0.05; Table 5). Reoperation was required in 5 patients during the follow-up period (0.5% per patientyear). At reoperation, the Carpentier-Edwards Valve-related Complications There were no cases of valve thrombosis. bioprosthesis was intact in 4 patients. In 1 paNineteen embolic episodes occurred in 17 pa- tient, a different heart valve was replaced betients (2% per patient-year); three of these cause of disease progression, and a Hancock
Table 4. Correlation between Preoperative Functional lncapacity and Overall Mortality (Early and Late)= Preoperative NYHA Functional Class
No. of Patients (N = 605)
No. of Deaths (N = 80)
I and I1
245 304 56
48 15
111
Iv
"Significance:p < 0.001. NYHA = New York Heart Association.
17
Mortality (%) 6.9 15.8 26.8
617 Pelletier et al: Carpentier-Edwards Porcine Bioprosthesis
Table 5. Degree of Postoperative Clinical Improvement for Valve Replacement Procedures (N= 522) Valve
Replacement Procedure Mitral
Aortic Tricuspid Multiple Totalb
No. Asymptomatica
No. Improveda
Total NO.^
145 (61.2) 154 (75.1) 3 (100) 52 (67.5) 354 (67.8)
53 (22.3) 25 (12.2) 0 17 (22.1) 95 (18.2)
198 (83.5) 179 (87.3) 3 (100) 69 (89.6) 449 (86.0)
i
NS
aNumbers in parentheses are percentages of total for each procedure. bNumbers in parentheses are percentages of total number of survivors followed. NS = not significant ( p > 0.05).
bioprosthesis was removed because of tissue failure in another patient. A major paravalvular leak was repaired in an aortic CarpentierEdwards bioprosthesis in 2 patients. In only 1 patient (0.1% per patient-year) was reoperation performed because of primary tissue failure of a Carpentier-Edwards bioprosthesis; this failure occurred 5 years after initial implantation in the aortic position in a 16-year-old girl. The
Fig 3 . Freedom from thromboembolism in patients leaving hospital after valve replacement with CarpentierEdwards bioprosthesis ( S E M < 1.5%). Patients at risk at the beginning of each interval are shown along the abscissa.
bioprosthesis showed dense calcification of the leaflets, one of which was ruptured. After 5 years, the actuarial probability of the Carpentier-Edwards valve remaining free of all valve-related complications was 92.7 f 1.4% (Fig 5). There were five hemorrhagic complications (0.5% per patient-year) due to anticoagulation in the 126 patients maintained on long-term treatment. One death occurred.
Postoperative Hemodynarnic Studies AORTIC VALVE BIOPROSTHESIS. Of the 35 aortic bioprostheses studied, 29 were the standard model and 6 were the improved annulus model. The peak systolic gradient averaged 11.4 f 1.1
57
e
2H 2I I-
LL
0 W W
E !-
Z W
:I;; 75
0 70 [r
W
a
551
0.5
,
I
;
;
457
295
149
50
15
1
2
3
4
5
YEARS AFTER OPERATION
PTS
618 The Annals of Thoracic Surgery Vol 34 No 6 December 1982
100
2 k
g
95
a
0
gZ
90
&
85
W
W
w
a
LL
80
I-
z
g
75
LI W
a 70 605
460
298
154
I
I I
0.5
1
2
3
55
,
21 PTS
I
4
5
38
4 PTS
YEARS AFTER OPERATION
Fig 4. Freedom f r o m endocarditis after valve replacement w i t h Carpentier-Edwards bioprosthesis ( S E M < 1 YO). (PTS = patients.)
I - ) W p
=
5
450
284
0
a W
a
0.5
1
2
138 1
I
4
3
4
5
YEARS AFTER OPERATION
Fig 5 . Actuarial probability of remaining free of all valve-related complications after replacement w i t h Carpentier-Edwards bioprosthesis ( S E M < 1.5 Yo). Patients at risk at t h e beginning of each interval are s h o w n along the abscissa. Dashed line indicates intervals with f e w e r t h a n 50 patients at risk. (PTS = patients.)
619 Pelletier et al: Carpentier-Edwards Porcine Bioprosthesis
N=12
25
N=17
c o : 4.9--.a
h
?! 5 E
20.
I-
15.
z
NS
T
5.2
- 9.6
P=O.Ol
N=6 6.7-12.4 NS
w 0 a
8
10.
Y
a w a
5.
0. 21-23
25-27
29-31
VALVE SIZE (mm)
Fig 6 . Residual peak systolic gradient across the Carpentier-Edwards bioprosthesis in the aortic position at rest (R) in 35 patients and during exercise (EX) in 30 patients, according to valve stent size. The cardiac output ( C O )is shown along the top of the figure. The difference between the rest and exercise gradients was statistically significant only for the 25 to 27 m m group. (NS = not significant).
mm Hg at rest (mean k SEM; range, 2 to 28 mm Hg). The gradient decreased with increasing stent size, averaging 13.4 k 2.2 mm Hg for the 21 to 23 mm valves, 11.3 f 1.5 mm Hg for the 25 to 27 mm valves, and 7.5 f 2.2 mm Hg for the 29 to 31 mm valves (Fig 6). Ergometric exercise
studies performed in 30 patients showed that cardiac output increased by an average of 82%, and that the gradient across the bioprosthesis also increased in each valve size group, although it was significant only in the 25 to 27 mm valves ( p = 0.01; see Fig 6). The effective orifice area averaged 1.4 k 0.1 cm2(range, 0.5 to 2.8 cm2)at rest and increased significantly with the size of the bioprosthesis (Table 6). During exercise, the effective orifice area increased in each stent size. Average increases were 12% for 21 to 23 mm valves, 15% for 25 to 27 mm valves, and 26% for 29 to 31 mm valves; however, the difference was significant ( p C 0.05) only for the 25 to 27 mm valves (see Table 6). Since
Table 6. Effective Orifice A rea of Aortic Bioprostheses at Rest (35 Patients) and during Exercise (30 Patients) Aortic Stent Size
Variable Effective area (cmzy Rest Exercise Cardiac output Increase with exercise
21-23 mm (N = 12)
25-27 mm (N = 17)
29-31 mm (N = 6)
1.14 f 0.09b 1.28 f 0.09
1.37 f 0.07b 1.57 f 0.1lc
1.93 f 0.25b 2.43 f 0.34
80%
85%
85Yo
aMean f standard error. bDifference between sizes: p < 0.05. cDifferencebetween rest and exercise: p < 0.05.
620 The Annals of Thoracic Surgery Vol 34 No 6 December 1982
N=22
co
2ol 15-
N=13 4.8
-
4.2 -7.0
7.4
* T
I ...... .;.;.;. :................ ........ ....... .............. .............. ....... .............. ....... ....... .............. .............. .............. .............. .............. .;.;.;.:.:.:.: ....... .............. .;.;.:.;.;.;.; ....... .............. .......
Table 7 . Effective Orifice Area of Mitral Bioprostheses at Rest (35 Patients) and during Exercise (30 Patients)
*
Mitral Stent Size
-.;.a
10-
5-
Variable
31-33 mm (N = 13)
Effective area (cm2)a
Rest Exercise Cardiac output Increase with
n27-29
27-29 mm (N = 22)
31-33 VALVE SIZE (mm)
Fig 7 . Residual mean diastolic gradient across the Carpentier-Edwards bioprosthesis in the mitral position at rest (R) in 35 patients and during exercise (EX) in 30 patients, according to valve stent size. The cardiac output (CO)is shown along the top of the figure. The difference between the rest and exercise gradients was statistically significant for both groups.
the hemodynamic performance of the improved annulus model was similar to that of the standard model (1.13 cm2versus 1.29 cm2),they were grouped together for the analysis. Grade 1/4 prosthetic regurgitation was found in 7 patients, and grade 2/4, in l patient. MITRAL VALVE BIOPROSTHESIS. Hemodynamic studies were performed in 35 patients with mitral bioprosthesis. There were 25 standard and 10 improved annulus valves. The mean mitral gradient at rest was 6.4 f 0.4 mm Hg (range, 2 to 12 mm Hg) and was correlated with stent size, averaging 7.1 f 0.6 mm Hg for the 27 to 29 mm group, and 5.1 f 0.3 mm Hg for the 31 to 33 mm group (Fig 7). Measurements obtained during exercise in 30 patients showed that cardiac output increased by an average of 58% and that the mean diastolic gradient increased significantly in both groups ( p < 0.001). The resting effective orifice area of the mitral valve bioprosthesis averaged 2.5 f 0.1 cm2(range, 1.4 to 4.0 cmz) and increased with stent size (Table 7). With exercise, the average area increased by 20% in the 27 to 29 mm group and by 17% in the 31 to 33 mm group ( p < 0.01; see Table 7). The improved annulus bioprosthesis tended to have a slightly larger orifice area than the
2.33 k 0.14 2.68 rt 0.16 2.80 f 0.Zb 3.14 k 0.24b 54 yo
67%
exercise aMean f standard error. bDifference between rest and exercise: p < 0.01.
standard valve (mean, 2.7 and 2.3 cm2, respectively), but since the difference was not significant, both models were grouped together for the analysis. Grade 114 mitral regurgitation was seen in 2 patients, and grade 2/4, in 1patient. Comment After a decade of extensive use, the value of biological valve devices is well recognized. However, few data have been reported for the Carpentier-Edwards valve, because it has only been available for clinical use since 1976. In the present study, a cohort of 605 patients in whom this porcine valve was implanted was followed prospectively using a computer-based follow-up program. The early mortality in this series is similar to that of other studies [ll-131. The operative risk in patients undergoing associated coronary and valve surgery has been decreasing lately, most likely because of better myocardial protection techniques during operation [14]. The five-year survival, which includes operative deaths, is also comparable to that reported for other mechanical or biological valves [12, 13, 15,161. Little difference in late survival with different valve devices is to be expected, since survival is mainly determined by the preoperative cardiac status, and most late deaths are due to cardiac conditions rather than to valve-related complications, as shown by this study and others [13, 15, 161.
621 Pelletier et al: Carpentier-Edwards Porcine Bioprosthesis
The principal advantage of biological valves is freedom from anticoagulation on a permanent basis in most patients. In more than 75% of our patients, anticoagulants were discontinued 2 to 3 months after surgery. Of the 126 patients maintained on a permanent anticoagulation regimen for cardiac reasons, 5 patients (4%) suffered hemorrhagic complications, which led to death in 1patient. Thus, the overall incidence of hemorrhage was 0.5% per patient-year in this series. This rate compares favorably to the incidence of 4.3% per patientyear with the Starr-Edwards prosthesis [131 and to the aggregate rate of 8% for major hemorrhage after 4 years in a Bjork-Shiley valve series [15]. In the latter study, more than 50% of the patients had suffered a major or a minor hemorrhagic episode after 4 years. Thromboembolic complications, infrequent in most series of porcine heterograft implants [12, 16, 171, were also uncommon in our patients. In contrast with the findings of Oyer and colleagues [MI, the rate of embolic episodes did not increase during the early postoperative period, probably as a result of a short-term anticoagulation regimen administered after operation in our patients. Instead, the actuarial curve shows a low but constant tendency toward embolism throughout the follow-up period, as was also the case with the Bjork-Shiley prosthesis [15]. While the overall incidence of embolism in the present study is similar to that reported for the Angell-Shiley valve (1.3% per patient-year) [17] and the Hancock bioprosthesis (2.9% per patient-year) [16], it compares favorably with the 4.4% per patient-year embolic rate for the Starr-Edwards prosthesis [131, and to the 85 to 88% chance of being free of that complication after 4 years with the Bjork-Shiley valve U51. In a recent review of their experience with the Ionescu-Shiley pericardial valve, Becker and associates [191 reported a cumulative thromboembolism-free rate of 94% at 3 to 4 years for AVR and of 71% for MVR. Since 80% of the episodes in the latter group occurred in patients who had atrial fibrillation, anticoagulation was strongly recommended for this condition. We did not observe any significant difference in embolic rate between aortic and mitral bioprotheses, and this may be attributed
to our policy of giving anticoagulants to patients with atrial fibrillation. Prosthetic valve thrombosis is not a rare complication with mechanical valve devices, and it is frequently fatal. A cumulative incidence of 3% after 4 years was reported for the Bjork-Shiley prosthesis in the aortic position; in the mitral position, a 13% incidence for the Bjork-Shiley and a 10% incidence for the Lillehei-Kaster valves were found despite anticoagulation [15, 201. Isolated instances of this complication have also occurred with Hancock bioprostheses [21]. Valve thrombosis, however, has not been observed in the present study. The hemodynamic performance of the Carpentier-Edwards valve appears to be satisfactory. In the mitral position, residual gradients with most mechanical and biological devices are similar [4, 22-24]. However, the effective orifice area of the Carpentier-Edwards value appears larger than most, and approaches that reported recently for the St. Jude prosthesis [24]. In the aortic position, the CarpentierEdwards valve is less obstructive than most mechanical devices [22, 23, 251 except the Hall-Kaster valve [26], and the gradient across the device is similar to that of the Hancock modified-orifice valve [27]. The effective orifice area is adequate, except in the smaller stent sizes where it is slightly stenotic. In addition, the increase in effective valve area with flow suggests some degree of valve leaflet inertia as a possible cause for smaller calculated surface areas of the bioprosthesis at low flow rates [B]. The Carpentier-Edwards bioprosthesis recently underwent modifications in its stent structure in order to improve the ratio of the inner diameter to the outer diameter, but this new design does not seem to have resulted in any appreciable improvement in the hemodynamic function in vivo. However, smallsized aortic valves (19 to 21 mm) were used in only 53 of our valve replacements (7.5%), despite the fact that aortoplastic enlargement of the aortic root was seldom performed. The narrow sewing ring of this valve may allow the insertion of larger sizes than would be possible with other bioprostheses. In contrast, Sanders and colleagues [7] have recommended that Hancock valves smaller than 25 mm should not
622 The Annals of Thoracic Surgery Vol 34 No 6 December 1982
be used in the aortic position and that surgical enlargement should be performed in the small aortic annulus [4, 71. Similarly, suboptimal hemodynamic performance has been reported with the Angell-Shiley porcine valve in both the aortic and mitral positions [28]. Substantial clinical improvement was obtained with the Carpentier-Edwards valve, with 93% of the survivors remaining in Functional Class I or I1 after valve replacement; these results are comparable to those of others using either mechanical valves or bioprostheses [12, 15, 16, 19, 201. However, the long-term durability of bioprostheses is still a major concern. The currently reported incidence of tissue failure of the valve is increasing with the availability of longer follow-up data. Calcification and leaflet tear are known to occur, particularly among younger patients, and the collagen framework of porcine valves has been found to deteriorate progressively over time [7, 29, 301. According to Magilligan and co-workers [6], the incidence of valve failure increases sharply after 5 years and reaches almost 50% in patients 35 years of age or younger, compared with a rate of about 10% in older patients. However, primary tissue failure of a Carpentier-Edwards valve occurred in only 1 patient in the present series, which includes 31 patients who were younger than 30 years of age at the time of operation. Degeneration of the bioprosthesis became manifest 5 years postoperatively in this patient, requiring reoperation. Although the rest of our younger patients are still free of clinical evidence of valve dysfunction, other instances of valve failure may be observed as these individuals move past the 5-year turning-point.
Conclusion The Carpentier-Edwards bioprosthesis is a safe and reliable heart valve device that provides excellent clinical improvement. The incidence of valve-related complications remains low up to 5 years, and freedom from long-term anticoagulation in most patients appears to be the major advantage over mechanical valves. Because of the high omurrence of primary tissue failure of bioprostheses in young patients, the Carpentier-Edwards valve should probably not be the primary choice for them; however, further
studies are needed to determine the cut-off age at which the incidence increases. Also, the long-term durability of this bioprosthesis still needs to be assessed by longer follow-up observation. The Carpentier-Edwards valve has shown satisfactory hemodynamic function, although it is mildly stenotic in smaller stent sizes. Future improvements to the valve should aim at increasing the inner diameter and modifying the sewing ring to permit insertion of larger size valves.
References 1. Rhodes GR, McIntosh CL: Evaluation of hemolysis following replacement of atrioventricular valves with porcine xenograft (Hancock) valves. J Thorac Cardiovasc Surg 73:312, 1977 2. Rainer EG, Sadler TR Jr, Clhristopher RA: Dynamic performance, strain forces and technical considerations of prosthetic aortic valves. In Davila JC (ed): Second Henry Ford Hospital International Symposium on Cardiac Surgery. New York, Appleton Century Crofts, 1976, p 402 3. Pipkin RD, Buch WS, Fogarty TJ: Evaluation of aortic valve replacement with a porcine xenograft without long-term anticoagulation. J Thorac Cardiovasc Surg 71:179, 1976 4. Lurie AJ, Miller RR, Maxwell KS, et al: Hemodynamic assessment of the glutaraldehyde preserved porcine heterograft in the aortic and mitral positions. Circulation 56:Suppl2:104,1977 5. Bonchek LI: Current status of cardiac valve replacement: selection of a prosthesis and indications for operation. Am Heart J 101:96, 1981 6. Magilligan DJ Jr, Lewis JW Jr, Jara FM, et al: Spontaneous degeneration of porcine bioprosthetic valves. Ann Thorac Surg 30:259, 1980 7. Sanders SP, Levy RJ, Freed MD, et al: Use of Hancock porcine xenografts in children and adolescents. Am J Cardiol46:429, 1980 8. Chaitman BR, Bonan R, Lepage G, et al: Hemodynamic evaluation of the CarpentierEdwards porcine xenograft. Circulation 60:1170, 1979
9. Gorlin R, Gorlin SG: Hydraulic formula for calculation of the area of the stenotic mitral valve, other cardiac valves, and central circulatory shunts. Am Heart J 41:1, 1951 10. Cutler SJ, Ederer F: Maximum utilization of the life table method in analyzing survival. J Chronic Dis 8:699, 1958 11. Bjork VO, Heme A: Ten years’ experience with the Bjork-Shiley tilting disc valve. J Thorac Cardiovasc Surg 78:331, 1979 12. Jamieson WRE, JanuszMT, Monro AI, et al: Early clinical exuerience with the Canoentier-Edwards
623
Pelletier et al: Carpentier-Edwards Porcine Bioprosthesis
porcine heterograft cardiac valve. Can J Surg 23:132, 1980 13. Teply JF, Grunkemeier GL, Sutherland HD, et al: The ultimate prognosis after valve replacement: an assessment at twenty years. Ann Thorac Surg 32:111, 1981 14. Paiement B, Maille JG, Boulanger M, et al: La visite pre-operatoire en chirurgie cardiovasculaire. Can Anaesth SOCJ 27:584, 1980 15. Karp RB, Cyrus RJ, Blackstone EH, et al: The Bjork-Shiley valve: intermediate term follow-up. J Thorac Cardiovasc Surg 81:602, 1981 16. Davila JC, Magilligan DJ, Lewis JW Jr: Is the Hancock porcine valve the best cardiac valve substitute today? Ann Thorac Surg 26:303, 1978 17. Angell WW, Angell JD, Sywak A: The AngellShiley porcine xenograft. Ann Thorac Surg 28:537, 1979 18. Oyer PE, Stinson EB, Reitz BA, et al: Long-term evaluation of the porcine xenograft bioprosthesis. J Thorac Cardiovasc Surg 78:343, 1979 19. Becker RM, Sandor L, Tindel M, Frater RWM: Medium- term follow-up of the Ionescu-Shiley heterograft valve. Ann Thorac Surg 32:120, 1981 20. Starek PJK, McLaurin LP, Wilcox BR, Murray GF: Clinical evaluation of the Lillehei-Kaster pivoting-disc valve. Ann Thorac Surg 22:362, 1976 21. Hetzer R, Hill JD, Kerth WJ, et al: Thrombosis and degeneration of Hancock valves: clinical and pathological findings. Ann Thorac Surg 26:317, 1978 22. Pyle RB, Mayer JE Jr, Lindsay WG, et al: Hemodynamic evaluation of Lillehei-Kaster and Starr-Edwards prostheses. Ann Thorac Surg 26:336, 1978 23. McAnulty JH, Morton M, Rahimtoola SH, et al: Hemodynamic characteristics of the composite strut ball valve prostheses (Starr-Edwards track valves) in patients on anticoagulants. Circulation 58:Suppl 1:159, 1978 24. Horstkotte D, Haerten K, Herzer JA, et al: Preliminary results in mitral valve replacement with the St. Jude medical prosthesis: comparison with the Bjork-Shiley valve. Circulation 64:Suppl 2:203, 1981 25. Bjork VO, Holmgren A, Olin C, et al: Clinical and haemodynamic results of aortic valve replacement with the Bjork-Shiley tilting disk valve prosthesis. Scand J Thorac Cardiovasc Surg 5:177, 1971 26. Nitter-Hauge S, Enge I, Sembe BKH, et al: Primary clinical experience with the Hall-Kaster valve in the aortic position: results at 3 months including hemodynamic studies. Circulation 6O:Suppl 1:55, 1979 27. Craver JM, King SB, Douglas JS, et al: Late hemodynamic evaluation of Hancock modified orifice aortic bioprosthesis. Circulation 6O:Suppl 1:93, 1979
28. Rivera R, Infantes C, Delcan JL, Rico M: Clinical and hemodynamic assessment of the AngellShiley porcine xenograft. Ann Thorac Surg 30:455, 1980 29. Dunn JM: Porcine valve durability in children. Ann Thorac Surg 32:357, 1981 30. Ferrans VJ, Spray TL, Billinghan ME, et al: U1trastructure of Hancock porcine valvular heterografts: pre- and post-implantation changes. Circulation 58:Suppl l : l O , 1978
Discussion DR. DONALD J . MAGILLIGAN JR (Detroit, MI): In October, 1981, we at Henry Ford Hospital in Detroit marked the tenth year in which we have used porcine bioprosthetic valves. With each year, we have seen a greater number of valves run the risk of degeneration; in fact, two years ago, in the eighth year, we reported the removal of 10 valves because of spontaneous degeneration. At this point we expected that the reoperation rate would increase linearly, but this did not occur; in 1981, we took out only half of the 20 valves that we predicted would need to be removed. The valves that were removed during 1981 were from the same group as those removed in the eighth year-valves that were implanted in the second and third years of our experience. We suspect that the bioprostheses manufactured subsequent to this time provide greater durability. We are optimistic enough about this to have given 22 patients a second bioprosthesis when their first valve degenerated . These patients are older now, and they have been followed for a shorter period. However, 1 patient whose bioprosthesis degenerated at 40 months has now gone 56 months without failure. We can conclude from this that degeneration does not appear to be an immunological event. In their paper, Dr. Pelletier and his colleagues state that they used the Carpentier-Edwards valve in only 75% of their implants. I would like them to be more specific about their selection criteria for a bioprosthesis. I would also like them to comment on the use of short-term anticoagulation. Finally, I would like to commend the authors on their precise and exhaustive reporting of results. This paper adds substantially to our knowledge of bioprosthetic valve function.
DR. ERIC JAMIESON (Vancouver, BC, Canada): By March, 1980, we had placed Carpentier-Edwards valves in 700 patients at the University of British Columbia, the St. Paul's Hospital and the Vancouver General Hospital. The duration of follow-up for these patients is similar to that for the 605 patients described by Dr. Pelletier and his co-workers. Early and late mortality rates are also similar for both groups.
624 The Annals of Thoracic Surgery Vol 34 No 6 December 1982
The incidence of thromboembolism in our series was slightly lower, at 1.4% per patient-year (15 episodes), compared with 2% per patient-year (19 episodes) in the Montreal series. However, embolism-free rates in both groups were similar: 96.4% in our patients, compared with 95.2% in the Montreal series. We had four fatal emboli. The major emboli occurred in patients who had mitral valve replacement, or in patients receiving multiple valves who were in atrial fibrillation. Ten percent of our patients who had aortic valve replacements and 45% of those undergoing mitral and multiple valve replacements were given long-term anticoagulation. I would like to ask Dr. Pelletier to comment on the role of anticoagulation in the patients who were at high risk of thromboembolism. Infective endocarditis was also a problem in our series. The incidence was 0.76% per patient-year, or 8 episodes, and there were five deaths. Primary valve dysfunction was seen in 1patient, a 15-year-old girl with a calcified prosthesis. Our survival and functional improvements were similar to the results of Dr. Pelletier and his colleagues. We did not evaluate the hemodynamic status of the patients in the postoperative period. We have continued to use the Carpentier-Edwards valve almost exclusively. To date we have placed 1,200 valves in 1,100 patients. Primary valve dysfunction has occurred in 4 patients, resulting from calcification in 2 children and from torn cusps in 2 adult patients.
In reply to Dr. Magilligan, we do not routinely implant a bioprosthesis in all patients. We currently use mechanical valve devices in approximately 25% of our patients. In selecting a valve substitute, we try to match the valve device to the patient’s needs. Young patients appear to be more susceptible to early degeneration of tissue valves, as Dr. Magilligan and his group reported a year ago; therefore, we generally recommend a mechanical prosthesis in patients younger than 30 to 35 years of age. We also prefer a mechanical prosthesis to a biological valve in those patients at higher risk for thromboembolic events after operation, including patients with chronic atrial fibrillation, a markedly enlarged left atrium, or atrial thrombosis. Long-term anticoagulation will have to be maintained in this group of patients, negating the main advantage of the bioprosthesis. We give anticoagulants to all patients with porcine valves for two to three months following operation; after this time, the anticoagulants are progressively stopped over a period of 7 to 10 days. Only those patients presenting major risk factors for thromboembolic complications are left on permanent anticoagulation. In regard to Dr. Jamieson’s comments, it is striking to see the great similarities between our findings and the results obtained by his group in Vancouver. Anticoagulation was maintained in 23% of the patients in our series, most of whom had MVR or double valve replacement. Very few patients were kept on permanent anticoagulation following AVR. DR. PELLETIER:
ABSTRACT Carpentier-Edwards bioprostheses were implanted in 605 patients, 509 of whom had a single valve replacement, and 96 of whom had a multiple valve replacement. There were 54 early deaths (8.9%) and 26 late deaths (4.3%).The five-year actuarial survival was 87% for aortic valve replacement, 83% for mitral valve replacement, and 81% for multiple valve replacement. Of the 525 survivors, all but 3 were followed for a total of 964 patient-years; 354 patients (68%)remained asymptomatic, and 95 patients (18%) were improved. The incidences of thromboembolism, endocarditis, and reoperation due to primary tissue failure of the bioprosthesis were 2.0, 1.3, and 0.1% per patient-year, respectively. The actuarial probability of being free of all valve-related complications was 93% after five years. Satisfactory hemodynamic performance of the bioprosthesis was demonstrated by postoperative studies done in 70 patients. Thus, the CarpentierEdwards porcine valve provides good clinical improvement, with a low incidence of valve-related complications and tissue failure at five years postoperatively.
612 0003-4975/82/120612-13$01.25 @ 1982 by The Society of Thoracic Surgeons
613
Pelletier et al: Carpentier-Edwards Porcine Bioprosthesis
PREOP
76
POSTOP 354
70
60
304 I
50
40 195
30
20
10
0
nn I
56
I1
FUNCTIONAL CLASS (NYHA)
hundred and nine patients (84%) underwent a single valve replacement; 96 had a multiple valve replacement. Coronary artery bypass grafting (CABG) was the procedure most often associated with valve replacement, and was done in 81 patients (13.4%). A tricuspid antions, such as prosthetic dysfunction, infection, nuloplasty was performed in 38 patients, and or hemolysis, in 23 patients (4%). Four patients the ascending aorta was replaced in 9 patients (0.5%) were asymptomatic and underwent oper- because of aneurysmal dilatation. Of the 704 ation for severe aortic stenosis. The preopera- implants, the standard Carpentier-Edwards tive functional incapacity of the patients bioprosthesis was used in 612 procedures and studied is shown in Figure 1. The majority of the improved annulus model in 87. In 5 patients the patients were in New York Heart Associa- having a multiple valve replacement, other tion Functional Class I11 or IV; patients operated prostheses were implanted (2 Hancock valves, on tor reasons other than heart failure or angina 2 Ionescu-Shiley valves, and 1 Omniscience are included in Class I. One hundred and valve*).The procedures were performed using a twenty patients (20%) had had previous valve bubbler oxygenator, with moderate hemodiluoperations. Predominant mitral stenosis was tion and systemic hypothermia between 22" the most common valve lesion, appearing in and 25°C. During aortic cross-clamping, cold 208 patients (34%), followed by aortic stenosis potassium cardioplegia was used for myocarin 182 patients (30%), and multiple valve dis- dial protection in 539 patients (89%); topical ease in 96 others (16%).Sixty-six patients (11%) cooling of the heart was associated with carwere operated on for pure mitral regurgitation, dioplegia in 283 patients (47%). Oral anticoagulants were given for 2 to 3 months after and 50 (8%), for pure aortic regurgitation. The types of valve replacement and asso- operation unless there was a major contraindiciated surgical procedures done in the patients in this series are shown in Table 1. Five- *Omniscience, Medical Incorporated, Minneapolis, MN. Fig I . N e w York Heart Association (NYHA) Functional Class preoperatively in 605 patients and postoperatively in 522 late survivors. Preoperative Class I includes patients operated on for reasons other than heart failure and angina.
614 The Annals of Thoracic Surgery Vol 34 No 6 December 1982
Table 1. Primary and Associated Surgical Procedures in 605 Patients ~~
Operation
No. with CABG"
Total No.=
Valve replacement MVR AVR MVR MVR MVR
+ AVR
+ tricuspid
+ AVR + tricuspid
Tricuspid Total
Associated procedures Tricuspid annuloplasty Replacement of ascending aorta Closure of ASD Closure of VSD Resection of left ventricular aneurysm Total
33 41 6 0 1 0
(5.5) (6.8) (1.0) (0.4)
81 (13.4)
... ...
... ...
... ...
274 (45.3) 232 (38.3) 86 (14.2) 7 (1.2) 3 (0.5) 3 (0.5) 605 (100.0)
38 (6.3) 9 (1.5) 8 (1.3) 2 (0.3) l(O.2) 58 (9.6)
"Numbers in parentheses are percentages of total (N = 605). CABG = coronary artery bypass grafting; MVR = mitral valve replacement; AVR = aortic valve replacement; ASD = atrial septal defect; VSD = ventricular septal defect.
cation. A permanent regimen of anticoagulation was maintained in 126 patients (23%) because of chronic atrial fibrillation or moderate-to-severe left atrial dilatation. Patients were seen 2, 6, and 12 months after operation for clinical evaluation and were followed yearly thereafter. Data were obtained in 433 patients (83%) through visits to the valve follow-up clinic and (17%) either by telephone interview or by their physicians in 89 others. The data were coded on a special valve clinic coding form and entered into the hospital computer. Recall of patients was facilitated by a computer-generated patient list updated on a monthly basis. A postoperative hemodynamic study was performed in 70 asymptomatic patients 6 to 15 months after operation using previously described techniques [B]. The rest and exercise gradients and effective orifice areas were measured by simultaneous pressure recordings. Mean and peak-to-peak systolic gradients across the aortic bioprosthesis were determined from five cardiac cycles, and the mean diastolic gradient across the mitral bioprosthesis was determined from five diastolic cycles. The effective valve area was calculated using the Gorlin formula, with the
mean gradient determined by manual planimetry [9]. Statistical analysis of the data was performed using Student's t test and the chi-square test. Overall survival and valve-related complications were determined by the lifie-table method [lo] and expressed as mean f standard error of the mean (SEM).
Results Early Mortality The operative and early postoperative (30-day) mortality was 8.9% (54 patients). The early mortality for each surgical procedure is shown in Table 2. The mortality for single valve replacement was less than 5%. Mortality increased to 15.5% in patients having a multiple valve procedure (mitral replacement with tricuspid annuloplasty, or double or triple valve replacement). In patients who had CABG operation associated with valve replacement, the operative mortality was 22.2%; this figure includes 3 patients in whom the ascending aorta was also replaced. Nine of the 54 patients who died early (16.7%) had had a previous valve replacement and were undergoing reoperation for valve dysfunction or additional valve disease.
615
Pelletier et al: Carpentier-Edwards Porcine Bioprosthesis
Table 2 . Early (30-Day) Mortality for Types of Valve Replacement Procedure
No. of Patients
No. of Deaths
Early Mortality (%)
MVR AVR Tricuspid valve replacement Multiple valve replacement Valve replacement + CABG Total
214 191 3 116 81 605
9 9 0 18 18 54
4.2 4.7 0 15.5 22.2 8.9
MVR = mitral valve replacement; AVR
=
aortic valve replacement; CABG = coronary artery bypass grafting.
Table 3 . Causes of Early and Late Death Etiology
No. of Early Deathsa
No. of Late Deathsa
Cardiac causes Heart failure Infarction Sudden death (arrhythmia) Valve-related causes Embolism Reoperation Endocarditis Noncardiac causes Hemorrhage Cerebral ischemia Others Unknown causes Total
43 (80)
10 (38.5)
27 11 5 0
... ... ...
10 (18)
2 1 7 5 (19) 2 2 1 10 (38.5)
8 1 1 1 (2) 54
1 0 9
1 (4) 26
"Numbers in parentheses are percentages of total for each column.
Cardiac problems were responsible for 80% of the early deaths (Table 3 ) . The cause of death is unknown in 1 patient, who died shortly after leaving the hospital.
diac causes in 10 others (see Table 3). In 5 patients, death was caused by valve-related complications. Late survival determined by the actuarial method showed that the five-year probability of survival was 87 k 4% (mean k SEM) for AVR, 83 k 4% for MVR, and 81 f 5% for multiple valve replacement (Fig 2). The overall mortality was directly correlated ( p < 0.001) with the preoperative functional incapacity (Table 4).
Late Mortality There were 26 late deaths (4.3%), which occurred from 2 to 44 months after operation (average, 17 months). Twelve deaths occurred in patients who had mitral valve replacement (MVR), 12 in patients who had aortic valve replacement (AVR), and 2 in patients who had Clinical Improvement multiple valve replacement. There were no Of the 525 survivors, 522 patients (99.4%) were deaths in the 3 patients undergoing isolated re- followed for 6 to 61 months after operation placement of the tricuspid valve. Late mortality (average follow-up, 22 months), for a total averaged 2.7% per patient-year. Death was due of 964 patient-years of follow-up. When last to cardiac causes in 10 patients and to noncar- seen, 354 patients (67.8%) were completely
616 The Annals of Thoracic Surgery Vol 34 No 6 December 1982
aortic mitral multiple
602
464
306
163
32
66
0.5
1
2
3
PTS
I
I
4
5
YEARS AFTER OPERATION
Fig 2. Actuarial survival for aortic, mitral, and multiple valve replacement, including early mortality. Patients at risk at the beginning of each interval are shown along the abscissa. Dashed lines indicate intervals with fewer than 50 patients at risk in each group.
episodes were associated with a residual neurological deficit, and two were fatal. The incidence of embolism for AVR, MVR, and multiple valve replacement was 1.7, 1.5, and 2.8% per patient-year, respectively. After 5 years, the probability, determined by actuarial analysis, asymptomatic, 95 patients (18.2%) had im- of being free of thromboembolic complications proved by at least one functional class, and 73 was 95.2 _+ 1.4% (Fig 3). There were twelve epiothers (14%) showed no improvement or con- sodes of endocarditis (1.3%per patient-year) in tinued to deteriorate postoperatively. Before 10 patients, including one operative infection, operation, 60% of the 605 patients were in and one death. All of these episodes occurred Functional Class I11 or IV,while 93% of the 522 within 2 years of operation, aftser which time survivors remained in Class I or 11 after opera- the probability of remaining free of infection tion (see Fig 1). The rate of clinical improve- was 97.7 f 0.7% (Fig 4). In 3 patients, blood ment was similar for each type of valve re- cultures were negative. placement (p > 0.05; Table 5). Reoperation was required in 5 patients during the follow-up period (0.5% per patientyear). At reoperation, the Carpentier-Edwards Valve-related Complications There were no cases of valve thrombosis. bioprosthesis was intact in 4 patients. In 1 paNineteen embolic episodes occurred in 17 pa- tient, a different heart valve was replaced betients (2% per patient-year); three of these cause of disease progression, and a Hancock
Table 4. Correlation between Preoperative Functional lncapacity and Overall Mortality (Early and Late)= Preoperative NYHA Functional Class
No. of Patients (N = 605)
No. of Deaths (N = 80)
I and I1
245 304 56
48 15
111
Iv
"Significance:p < 0.001. NYHA = New York Heart Association.
17
Mortality (%) 6.9 15.8 26.8
617 Pelletier et al: Carpentier-Edwards Porcine Bioprosthesis
Table 5. Degree of Postoperative Clinical Improvement for Valve Replacement Procedures (N= 522) Valve
Replacement Procedure Mitral
Aortic Tricuspid Multiple Totalb
No. Asymptomatica
No. Improveda
Total NO.^
145 (61.2) 154 (75.1) 3 (100) 52 (67.5) 354 (67.8)
53 (22.3) 25 (12.2) 0 17 (22.1) 95 (18.2)
198 (83.5) 179 (87.3) 3 (100) 69 (89.6) 449 (86.0)
i
NS
aNumbers in parentheses are percentages of total for each procedure. bNumbers in parentheses are percentages of total number of survivors followed. NS = not significant ( p > 0.05).
bioprosthesis was removed because of tissue failure in another patient. A major paravalvular leak was repaired in an aortic CarpentierEdwards bioprosthesis in 2 patients. In only 1 patient (0.1% per patient-year) was reoperation performed because of primary tissue failure of a Carpentier-Edwards bioprosthesis; this failure occurred 5 years after initial implantation in the aortic position in a 16-year-old girl. The
Fig 3 . Freedom from thromboembolism in patients leaving hospital after valve replacement with CarpentierEdwards bioprosthesis ( S E M < 1.5%). Patients at risk at the beginning of each interval are shown along the abscissa.
bioprosthesis showed dense calcification of the leaflets, one of which was ruptured. After 5 years, the actuarial probability of the Carpentier-Edwards valve remaining free of all valve-related complications was 92.7 f 1.4% (Fig 5). There were five hemorrhagic complications (0.5% per patient-year) due to anticoagulation in the 126 patients maintained on long-term treatment. One death occurred.
Postoperative Hemodynarnic Studies AORTIC VALVE BIOPROSTHESIS. Of the 35 aortic bioprostheses studied, 29 were the standard model and 6 were the improved annulus model. The peak systolic gradient averaged 11.4 f 1.1
57
e
2H 2I I-
LL
0 W W
E !-
Z W
:I;; 75
0 70 [r
W
a
551
0.5
,
I
;
;
457
295
149
50
15
1
2
3
4
5
YEARS AFTER OPERATION
PTS
618 The Annals of Thoracic Surgery Vol 34 No 6 December 1982
100
2 k
g
95
a
0
gZ
90
&
85
W
W
w
a
LL
80
I-
z
g
75
LI W
a 70 605
460
298
154
I
I I
0.5
1
2
3
55
,
21 PTS
I
4
5
38
4 PTS
YEARS AFTER OPERATION
Fig 4. Freedom f r o m endocarditis after valve replacement w i t h Carpentier-Edwards bioprosthesis ( S E M < 1 YO). (PTS = patients.)
I - ) W p
=
5
450
284
0
a W
a
0.5
1
2
138 1
I
4
3
4
5
YEARS AFTER OPERATION
Fig 5 . Actuarial probability of remaining free of all valve-related complications after replacement w i t h Carpentier-Edwards bioprosthesis ( S E M < 1.5 Yo). Patients at risk at t h e beginning of each interval are s h o w n along the abscissa. Dashed line indicates intervals with f e w e r t h a n 50 patients at risk. (PTS = patients.)
619 Pelletier et al: Carpentier-Edwards Porcine Bioprosthesis
N=12
25
N=17
c o : 4.9--.a
h
?! 5 E
20.
I-
15.
z
NS
T
5.2
- 9.6
P=O.Ol
N=6 6.7-12.4 NS
w 0 a
8
10.
Y
a w a
5.
0. 21-23
25-27
29-31
VALVE SIZE (mm)
Fig 6 . Residual peak systolic gradient across the Carpentier-Edwards bioprosthesis in the aortic position at rest (R) in 35 patients and during exercise (EX) in 30 patients, according to valve stent size. The cardiac output ( C O )is shown along the top of the figure. The difference between the rest and exercise gradients was statistically significant only for the 25 to 27 m m group. (NS = not significant).
mm Hg at rest (mean k SEM; range, 2 to 28 mm Hg). The gradient decreased with increasing stent size, averaging 13.4 k 2.2 mm Hg for the 21 to 23 mm valves, 11.3 f 1.5 mm Hg for the 25 to 27 mm valves, and 7.5 f 2.2 mm Hg for the 29 to 31 mm valves (Fig 6). Ergometric exercise
studies performed in 30 patients showed that cardiac output increased by an average of 82%, and that the gradient across the bioprosthesis also increased in each valve size group, although it was significant only in the 25 to 27 mm valves ( p = 0.01; see Fig 6). The effective orifice area averaged 1.4 k 0.1 cm2(range, 0.5 to 2.8 cm2)at rest and increased significantly with the size of the bioprosthesis (Table 6). During exercise, the effective orifice area increased in each stent size. Average increases were 12% for 21 to 23 mm valves, 15% for 25 to 27 mm valves, and 26% for 29 to 31 mm valves; however, the difference was significant ( p C 0.05) only for the 25 to 27 mm valves (see Table 6). Since
Table 6. Effective Orifice A rea of Aortic Bioprostheses at Rest (35 Patients) and during Exercise (30 Patients) Aortic Stent Size
Variable Effective area (cmzy Rest Exercise Cardiac output Increase with exercise
21-23 mm (N = 12)
25-27 mm (N = 17)
29-31 mm (N = 6)
1.14 f 0.09b 1.28 f 0.09
1.37 f 0.07b 1.57 f 0.1lc
1.93 f 0.25b 2.43 f 0.34
80%
85%
85Yo
aMean f standard error. bDifference between sizes: p < 0.05. cDifferencebetween rest and exercise: p < 0.05.
620 The Annals of Thoracic Surgery Vol 34 No 6 December 1982
N=22
co
2ol 15-
N=13 4.8
-
4.2 -7.0
7.4
* T
I ...... .;.;.;. :................ ........ ....... .............. .............. ....... .............. ....... ....... .............. .............. .............. .............. .............. .;.;.;.:.:.:.: ....... .............. .;.;.:.;.;.;.; ....... .............. .......
Table 7 . Effective Orifice Area of Mitral Bioprostheses at Rest (35 Patients) and during Exercise (30 Patients)
*
Mitral Stent Size
-.;.a
10-
5-
Variable
31-33 mm (N = 13)
Effective area (cm2)a
Rest Exercise Cardiac output Increase with
n27-29
27-29 mm (N = 22)
31-33 VALVE SIZE (mm)
Fig 7 . Residual mean diastolic gradient across the Carpentier-Edwards bioprosthesis in the mitral position at rest (R) in 35 patients and during exercise (EX) in 30 patients, according to valve stent size. The cardiac output (CO)is shown along the top of the figure. The difference between the rest and exercise gradients was statistically significant for both groups.
the hemodynamic performance of the improved annulus model was similar to that of the standard model (1.13 cm2versus 1.29 cm2),they were grouped together for the analysis. Grade 1/4 prosthetic regurgitation was found in 7 patients, and grade 2/4, in l patient. MITRAL VALVE BIOPROSTHESIS. Hemodynamic studies were performed in 35 patients with mitral bioprosthesis. There were 25 standard and 10 improved annulus valves. The mean mitral gradient at rest was 6.4 f 0.4 mm Hg (range, 2 to 12 mm Hg) and was correlated with stent size, averaging 7.1 f 0.6 mm Hg for the 27 to 29 mm group, and 5.1 f 0.3 mm Hg for the 31 to 33 mm group (Fig 7). Measurements obtained during exercise in 30 patients showed that cardiac output increased by an average of 58% and that the mean diastolic gradient increased significantly in both groups ( p < 0.001). The resting effective orifice area of the mitral valve bioprosthesis averaged 2.5 f 0.1 cm2(range, 1.4 to 4.0 cmz) and increased with stent size (Table 7). With exercise, the average area increased by 20% in the 27 to 29 mm group and by 17% in the 31 to 33 mm group ( p < 0.01; see Table 7). The improved annulus bioprosthesis tended to have a slightly larger orifice area than the
2.33 k 0.14 2.68 rt 0.16 2.80 f 0.Zb 3.14 k 0.24b 54 yo
67%
exercise aMean f standard error. bDifference between rest and exercise: p < 0.01.
standard valve (mean, 2.7 and 2.3 cm2, respectively), but since the difference was not significant, both models were grouped together for the analysis. Grade 114 mitral regurgitation was seen in 2 patients, and grade 2/4, in 1patient. Comment After a decade of extensive use, the value of biological valve devices is well recognized. However, few data have been reported for the Carpentier-Edwards valve, because it has only been available for clinical use since 1976. In the present study, a cohort of 605 patients in whom this porcine valve was implanted was followed prospectively using a computer-based follow-up program. The early mortality in this series is similar to that of other studies [ll-131. The operative risk in patients undergoing associated coronary and valve surgery has been decreasing lately, most likely because of better myocardial protection techniques during operation [14]. The five-year survival, which includes operative deaths, is also comparable to that reported for other mechanical or biological valves [12, 13, 15,161. Little difference in late survival with different valve devices is to be expected, since survival is mainly determined by the preoperative cardiac status, and most late deaths are due to cardiac conditions rather than to valve-related complications, as shown by this study and others [13, 15, 161.
621 Pelletier et al: Carpentier-Edwards Porcine Bioprosthesis
The principal advantage of biological valves is freedom from anticoagulation on a permanent basis in most patients. In more than 75% of our patients, anticoagulants were discontinued 2 to 3 months after surgery. Of the 126 patients maintained on a permanent anticoagulation regimen for cardiac reasons, 5 patients (4%) suffered hemorrhagic complications, which led to death in 1patient. Thus, the overall incidence of hemorrhage was 0.5% per patient-year in this series. This rate compares favorably to the incidence of 4.3% per patientyear with the Starr-Edwards prosthesis [131 and to the aggregate rate of 8% for major hemorrhage after 4 years in a Bjork-Shiley valve series [15]. In the latter study, more than 50% of the patients had suffered a major or a minor hemorrhagic episode after 4 years. Thromboembolic complications, infrequent in most series of porcine heterograft implants [12, 16, 171, were also uncommon in our patients. In contrast with the findings of Oyer and colleagues [MI, the rate of embolic episodes did not increase during the early postoperative period, probably as a result of a short-term anticoagulation regimen administered after operation in our patients. Instead, the actuarial curve shows a low but constant tendency toward embolism throughout the follow-up period, as was also the case with the Bjork-Shiley prosthesis [15]. While the overall incidence of embolism in the present study is similar to that reported for the Angell-Shiley valve (1.3% per patient-year) [17] and the Hancock bioprosthesis (2.9% per patient-year) [16], it compares favorably with the 4.4% per patient-year embolic rate for the Starr-Edwards prosthesis [131, and to the 85 to 88% chance of being free of that complication after 4 years with the Bjork-Shiley valve U51. In a recent review of their experience with the Ionescu-Shiley pericardial valve, Becker and associates [191 reported a cumulative thromboembolism-free rate of 94% at 3 to 4 years for AVR and of 71% for MVR. Since 80% of the episodes in the latter group occurred in patients who had atrial fibrillation, anticoagulation was strongly recommended for this condition. We did not observe any significant difference in embolic rate between aortic and mitral bioprotheses, and this may be attributed
to our policy of giving anticoagulants to patients with atrial fibrillation. Prosthetic valve thrombosis is not a rare complication with mechanical valve devices, and it is frequently fatal. A cumulative incidence of 3% after 4 years was reported for the Bjork-Shiley prosthesis in the aortic position; in the mitral position, a 13% incidence for the Bjork-Shiley and a 10% incidence for the Lillehei-Kaster valves were found despite anticoagulation [15, 201. Isolated instances of this complication have also occurred with Hancock bioprostheses [21]. Valve thrombosis, however, has not been observed in the present study. The hemodynamic performance of the Carpentier-Edwards valve appears to be satisfactory. In the mitral position, residual gradients with most mechanical and biological devices are similar [4, 22-24]. However, the effective orifice area of the Carpentier-Edwards value appears larger than most, and approaches that reported recently for the St. Jude prosthesis [24]. In the aortic position, the CarpentierEdwards valve is less obstructive than most mechanical devices [22, 23, 251 except the Hall-Kaster valve [26], and the gradient across the device is similar to that of the Hancock modified-orifice valve [27]. The effective orifice area is adequate, except in the smaller stent sizes where it is slightly stenotic. In addition, the increase in effective valve area with flow suggests some degree of valve leaflet inertia as a possible cause for smaller calculated surface areas of the bioprosthesis at low flow rates [B]. The Carpentier-Edwards bioprosthesis recently underwent modifications in its stent structure in order to improve the ratio of the inner diameter to the outer diameter, but this new design does not seem to have resulted in any appreciable improvement in the hemodynamic function in vivo. However, smallsized aortic valves (19 to 21 mm) were used in only 53 of our valve replacements (7.5%), despite the fact that aortoplastic enlargement of the aortic root was seldom performed. The narrow sewing ring of this valve may allow the insertion of larger sizes than would be possible with other bioprostheses. In contrast, Sanders and colleagues [7] have recommended that Hancock valves smaller than 25 mm should not
622 The Annals of Thoracic Surgery Vol 34 No 6 December 1982
be used in the aortic position and that surgical enlargement should be performed in the small aortic annulus [4, 71. Similarly, suboptimal hemodynamic performance has been reported with the Angell-Shiley porcine valve in both the aortic and mitral positions [28]. Substantial clinical improvement was obtained with the Carpentier-Edwards valve, with 93% of the survivors remaining in Functional Class I or I1 after valve replacement; these results are comparable to those of others using either mechanical valves or bioprostheses [12, 15, 16, 19, 201. However, the long-term durability of bioprostheses is still a major concern. The currently reported incidence of tissue failure of the valve is increasing with the availability of longer follow-up data. Calcification and leaflet tear are known to occur, particularly among younger patients, and the collagen framework of porcine valves has been found to deteriorate progressively over time [7, 29, 301. According to Magilligan and co-workers [6], the incidence of valve failure increases sharply after 5 years and reaches almost 50% in patients 35 years of age or younger, compared with a rate of about 10% in older patients. However, primary tissue failure of a Carpentier-Edwards valve occurred in only 1 patient in the present series, which includes 31 patients who were younger than 30 years of age at the time of operation. Degeneration of the bioprosthesis became manifest 5 years postoperatively in this patient, requiring reoperation. Although the rest of our younger patients are still free of clinical evidence of valve dysfunction, other instances of valve failure may be observed as these individuals move past the 5-year turning-point.
Conclusion The Carpentier-Edwards bioprosthesis is a safe and reliable heart valve device that provides excellent clinical improvement. The incidence of valve-related complications remains low up to 5 years, and freedom from long-term anticoagulation in most patients appears to be the major advantage over mechanical valves. Because of the high omurrence of primary tissue failure of bioprostheses in young patients, the Carpentier-Edwards valve should probably not be the primary choice for them; however, further
studies are needed to determine the cut-off age at which the incidence increases. Also, the long-term durability of this bioprosthesis still needs to be assessed by longer follow-up observation. The Carpentier-Edwards valve has shown satisfactory hemodynamic function, although it is mildly stenotic in smaller stent sizes. Future improvements to the valve should aim at increasing the inner diameter and modifying the sewing ring to permit insertion of larger size valves.
References 1. Rhodes GR, McIntosh CL: Evaluation of hemolysis following replacement of atrioventricular valves with porcine xenograft (Hancock) valves. J Thorac Cardiovasc Surg 73:312, 1977 2. Rainer EG, Sadler TR Jr, Clhristopher RA: Dynamic performance, strain forces and technical considerations of prosthetic aortic valves. In Davila JC (ed): Second Henry Ford Hospital International Symposium on Cardiac Surgery. New York, Appleton Century Crofts, 1976, p 402 3. Pipkin RD, Buch WS, Fogarty TJ: Evaluation of aortic valve replacement with a porcine xenograft without long-term anticoagulation. J Thorac Cardiovasc Surg 71:179, 1976 4. Lurie AJ, Miller RR, Maxwell KS, et al: Hemodynamic assessment of the glutaraldehyde preserved porcine heterograft in the aortic and mitral positions. Circulation 56:Suppl2:104,1977 5. Bonchek LI: Current status of cardiac valve replacement: selection of a prosthesis and indications for operation. Am Heart J 101:96, 1981 6. Magilligan DJ Jr, Lewis JW Jr, Jara FM, et al: Spontaneous degeneration of porcine bioprosthetic valves. Ann Thorac Surg 30:259, 1980 7. Sanders SP, Levy RJ, Freed MD, et al: Use of Hancock porcine xenografts in children and adolescents. Am J Cardiol46:429, 1980 8. Chaitman BR, Bonan R, Lepage G, et al: Hemodynamic evaluation of the CarpentierEdwards porcine xenograft. Circulation 60:1170, 1979
9. Gorlin R, Gorlin SG: Hydraulic formula for calculation of the area of the stenotic mitral valve, other cardiac valves, and central circulatory shunts. Am Heart J 41:1, 1951 10. Cutler SJ, Ederer F: Maximum utilization of the life table method in analyzing survival. J Chronic Dis 8:699, 1958 11. Bjork VO, Heme A: Ten years’ experience with the Bjork-Shiley tilting disc valve. J Thorac Cardiovasc Surg 78:331, 1979 12. Jamieson WRE, JanuszMT, Monro AI, et al: Early clinical exuerience with the Canoentier-Edwards
623
Pelletier et al: Carpentier-Edwards Porcine Bioprosthesis
porcine heterograft cardiac valve. Can J Surg 23:132, 1980 13. Teply JF, Grunkemeier GL, Sutherland HD, et al: The ultimate prognosis after valve replacement: an assessment at twenty years. Ann Thorac Surg 32:111, 1981 14. Paiement B, Maille JG, Boulanger M, et al: La visite pre-operatoire en chirurgie cardiovasculaire. Can Anaesth SOCJ 27:584, 1980 15. Karp RB, Cyrus RJ, Blackstone EH, et al: The Bjork-Shiley valve: intermediate term follow-up. J Thorac Cardiovasc Surg 81:602, 1981 16. Davila JC, Magilligan DJ, Lewis JW Jr: Is the Hancock porcine valve the best cardiac valve substitute today? Ann Thorac Surg 26:303, 1978 17. Angell WW, Angell JD, Sywak A: The AngellShiley porcine xenograft. Ann Thorac Surg 28:537, 1979 18. Oyer PE, Stinson EB, Reitz BA, et al: Long-term evaluation of the porcine xenograft bioprosthesis. J Thorac Cardiovasc Surg 78:343, 1979 19. Becker RM, Sandor L, Tindel M, Frater RWM: Medium- term follow-up of the Ionescu-Shiley heterograft valve. Ann Thorac Surg 32:120, 1981 20. Starek PJK, McLaurin LP, Wilcox BR, Murray GF: Clinical evaluation of the Lillehei-Kaster pivoting-disc valve. Ann Thorac Surg 22:362, 1976 21. Hetzer R, Hill JD, Kerth WJ, et al: Thrombosis and degeneration of Hancock valves: clinical and pathological findings. Ann Thorac Surg 26:317, 1978 22. Pyle RB, Mayer JE Jr, Lindsay WG, et al: Hemodynamic evaluation of Lillehei-Kaster and Starr-Edwards prostheses. Ann Thorac Surg 26:336, 1978 23. McAnulty JH, Morton M, Rahimtoola SH, et al: Hemodynamic characteristics of the composite strut ball valve prostheses (Starr-Edwards track valves) in patients on anticoagulants. Circulation 58:Suppl 1:159, 1978 24. Horstkotte D, Haerten K, Herzer JA, et al: Preliminary results in mitral valve replacement with the St. Jude medical prosthesis: comparison with the Bjork-Shiley valve. Circulation 64:Suppl 2:203, 1981 25. Bjork VO, Holmgren A, Olin C, et al: Clinical and haemodynamic results of aortic valve replacement with the Bjork-Shiley tilting disk valve prosthesis. Scand J Thorac Cardiovasc Surg 5:177, 1971 26. Nitter-Hauge S, Enge I, Sembe BKH, et al: Primary clinical experience with the Hall-Kaster valve in the aortic position: results at 3 months including hemodynamic studies. Circulation 6O:Suppl 1:55, 1979 27. Craver JM, King SB, Douglas JS, et al: Late hemodynamic evaluation of Hancock modified orifice aortic bioprosthesis. Circulation 6O:Suppl 1:93, 1979
28. Rivera R, Infantes C, Delcan JL, Rico M: Clinical and hemodynamic assessment of the AngellShiley porcine xenograft. Ann Thorac Surg 30:455, 1980 29. Dunn JM: Porcine valve durability in children. Ann Thorac Surg 32:357, 1981 30. Ferrans VJ, Spray TL, Billinghan ME, et al: U1trastructure of Hancock porcine valvular heterografts: pre- and post-implantation changes. Circulation 58:Suppl l : l O , 1978
Discussion DR. DONALD J . MAGILLIGAN JR (Detroit, MI): In October, 1981, we at Henry Ford Hospital in Detroit marked the tenth year in which we have used porcine bioprosthetic valves. With each year, we have seen a greater number of valves run the risk of degeneration; in fact, two years ago, in the eighth year, we reported the removal of 10 valves because of spontaneous degeneration. At this point we expected that the reoperation rate would increase linearly, but this did not occur; in 1981, we took out only half of the 20 valves that we predicted would need to be removed. The valves that were removed during 1981 were from the same group as those removed in the eighth year-valves that were implanted in the second and third years of our experience. We suspect that the bioprostheses manufactured subsequent to this time provide greater durability. We are optimistic enough about this to have given 22 patients a second bioprosthesis when their first valve degenerated . These patients are older now, and they have been followed for a shorter period. However, 1 patient whose bioprosthesis degenerated at 40 months has now gone 56 months without failure. We can conclude from this that degeneration does not appear to be an immunological event. In their paper, Dr. Pelletier and his colleagues state that they used the Carpentier-Edwards valve in only 75% of their implants. I would like them to be more specific about their selection criteria for a bioprosthesis. I would also like them to comment on the use of short-term anticoagulation. Finally, I would like to commend the authors on their precise and exhaustive reporting of results. This paper adds substantially to our knowledge of bioprosthetic valve function.
DR. ERIC JAMIESON (Vancouver, BC, Canada): By March, 1980, we had placed Carpentier-Edwards valves in 700 patients at the University of British Columbia, the St. Paul's Hospital and the Vancouver General Hospital. The duration of follow-up for these patients is similar to that for the 605 patients described by Dr. Pelletier and his co-workers. Early and late mortality rates are also similar for both groups.
624 The Annals of Thoracic Surgery Vol 34 No 6 December 1982
The incidence of thromboembolism in our series was slightly lower, at 1.4% per patient-year (15 episodes), compared with 2% per patient-year (19 episodes) in the Montreal series. However, embolism-free rates in both groups were similar: 96.4% in our patients, compared with 95.2% in the Montreal series. We had four fatal emboli. The major emboli occurred in patients who had mitral valve replacement, or in patients receiving multiple valves who were in atrial fibrillation. Ten percent of our patients who had aortic valve replacements and 45% of those undergoing mitral and multiple valve replacements were given long-term anticoagulation. I would like to ask Dr. Pelletier to comment on the role of anticoagulation in the patients who were at high risk of thromboembolism. Infective endocarditis was also a problem in our series. The incidence was 0.76% per patient-year, or 8 episodes, and there were five deaths. Primary valve dysfunction was seen in 1patient, a 15-year-old girl with a calcified prosthesis. Our survival and functional improvements were similar to the results of Dr. Pelletier and his colleagues. We did not evaluate the hemodynamic status of the patients in the postoperative period. We have continued to use the Carpentier-Edwards valve almost exclusively. To date we have placed 1,200 valves in 1,100 patients. Primary valve dysfunction has occurred in 4 patients, resulting from calcification in 2 children and from torn cusps in 2 adult patients.
In reply to Dr. Magilligan, we do not routinely implant a bioprosthesis in all patients. We currently use mechanical valve devices in approximately 25% of our patients. In selecting a valve substitute, we try to match the valve device to the patient’s needs. Young patients appear to be more susceptible to early degeneration of tissue valves, as Dr. Magilligan and his group reported a year ago; therefore, we generally recommend a mechanical prosthesis in patients younger than 30 to 35 years of age. We also prefer a mechanical prosthesis to a biological valve in those patients at higher risk for thromboembolic events after operation, including patients with chronic atrial fibrillation, a markedly enlarged left atrium, or atrial thrombosis. Long-term anticoagulation will have to be maintained in this group of patients, negating the main advantage of the bioprosthesis. We give anticoagulants to all patients with porcine valves for two to three months following operation; after this time, the anticoagulants are progressively stopped over a period of 7 to 10 days. Only those patients presenting major risk factors for thromboembolic complications are left on permanent anticoagulation. In regard to Dr. Jamieson’s comments, it is striking to see the great similarities between our findings and the results obtained by his group in Vancouver. Anticoagulation was maintained in 23% of the patients in our series, most of whom had MVR or double valve replacement. Very few patients were kept on permanent anticoagulation following AVR. DR. PELLETIER: