Perioperative events in patients with failed mechanical and bioprosthetic valves

Perioperative Events in Patients With Failed Mechanical and Bioprosthetic Valves Lynn B. McGrath, MD, Javier Fernandez, MD, Glenn W. Laub, MD, William A. Anderson, MD, Bridget M. Bailey, BSN, and Chao Chen, PhD Department

of Thoracic and Cardiovascular

Surgery, Deborah Heart and Lung Center, Browns Mills, New Jersey

From 1963 through 1991, 1037 patients underwent reoperative valvular procedures. The 478 patients having reoperations for either failed bioprosthetic (n = 212) or mechanical (n = 266) valves were evaluated. There were 210 male (44%) and 268 female (56%;) patients. The mean age at reoperation of the patients in the bioprosthesis group was 59.7 years and and that in the mechanical valve group was 56.1 years (p = 0.0006). The mean interval to the time of reoperation was 84.7 months in the mechanical valve group and 74 months in the bioprosthesis group. There was no difference between the two groups in the functional class at reoperation. More severe mitral valve stenosis and incompetence, more severe aortic valve stenosis, and higher right ventricular and pulmonary arterial pressures were noted in the biopros-

thesis group than in the mechanical valve group. Hemolysis (p = 0.05) was more prevalent in the patients with mechanical valves than in the ones with bioprostheses. A longer aortic occlusion time @ = 0.0001) and longer cardiopulmonary bypass time (p = 0.0001) were required for the reoperations in the bioprosthesis group. The operative mortality was 13.2% for the bioprosthesis patients and 12.4% for the mechanical valve patients. The risk factors for hospital death included the cross-clamp time (p = O.OOOl), the functional class (p = O.OOOOl), the presence of ascites (y = 0.02), hepatomegaly (p = 0.002), and decreasing ejection fraction (p = 0.05). We conclude that mechanical valve failures do not produce catastrophic events resulting in poor reoperative results. (Ann Thorac Surg 1995;6O:S475-8)

A

was at another valve position. This left 478 patients for the evaluation. The prior valve replacement in these 478 patients had consisted of either a bioprosthetic (n = 212) or a mechanical (n = 266) prosthesis, and they were undergoing reoperation at the same valve site. Of these 478 patients, 447 (94%) had had one previous operation, 29 had had two previous operations (6X), 1 had had three previous operations (0.2%), and 1 had had four prior operations (0.2%). The mean interval from the time of the most recent operation to that of the current reoperation at the Deborah Heart and Lung Center was 78 months (range, 0 to 406 months). There were 117 male (44%) and 149 female (56%) patients in the mechanical valve group. There were 93 male (44%) and 719 female patients in the bioprosthesis group (56%). The mean age of the 266 patients with mechanical valves was 56.1 years (range, 6.9 to 81 years), and the mean age of the 212 patients with bioprosthetic valves was 59.7 years (range, 6.9 to 87 years) (p [Wilcoxon] = 0.0006).

lthough bioprosthetic valves often succumb to various modes of intrinsic valve dysfunction, among the rationales for continuing to recommend the use of these prostheses has been their anticipated subtle manner of valve failure, permitting greater ease of performance of elective reoperative procedures [l, 21. In contrast, it has been inferred that patients undergoing reoperative procedures for failed mechanical prostheses present with catastrophic valvular dysfunction and unsatisfactory outcomes are expected at fhe time of reoperation [3, 41. The purpose of this study was to examine the nature of events occurring in patients undergoing reoperation for failure of bioprosthetic and mechanical valves to determine differences in the clinical status, modes of valve failure, and operative results. Material

and Methods

Patients From January 1, 1963, through December 31, 1991, 1037 consecutive patients who had had at least one previous valvular operation underwent a reoperative cardiac procedure at the Deborah Heart and Lung Center. Each of these 1037 patients’ records were reviewed, and 559 were excluded from analysis in this study: the 66 who had had a prior native valve repair and the 493 who had had a prior valve replacement but whose current reoperation Presented in part theses, Vancouver,

at the VI International Svmposium BC, July 29-31, lYY1.

Address reprmt requests to LIr M&rath, rah Heart and Lung Center, 200 lrenton

Lkywtment I
0 1995 by The Society of Thoracic Surgeons

for Cardiac

Bioproh-

of Surgei->, D&w Mills, NJ 0X015.

Statistical Analysis The two groups were initially tested for multiple demographic and preoperative variables to assess their comparability. Postoperative complication rates and operative mortality were compared for the two groups using standard statistical procedures: Wilcoxon’s test, Fisher’s exact test, and the ,$ test. Two-way analysis of variance was adopted to evaluate the effect of the prosthesis type 0003-4975/95/$9.50 0003-4975(95)00444-P

S476

CARDIAC PATIENT

BIOPROSTHESES STATUS AFTER

\‘ALVE

MC GRATl FAILURE

I ET AL

and valve position on the intervals from the time of the previous operation to that of the current reoperation. The Society of Thoracic Surgeons guidelines for data reporting were followed [5]. Results

Patients Multiple variables were examined to determine differences in patient status at the time of presentation for reoperation. There was no statistically significant difference in the functional class (p [,$ test] = 0.72) or in the incidence of renal dysfunction, peripheral emboli, or prosthetic valve endocarditis between the two groups at the time of presentation for reoperation. However, there were important differences in the incidence of certain events. There was a higher incidence of intravascular hemolysis in the mechanical valve group (21 patients [SO/,]) than in the bioprosthesis group (7 patients [3.3%]) (p [Fisher’s exact] = 0.05). A history of cerebrovascular accident was more prevalent in the mechanical group (38 patients [14.2%]) than it was in the bioprosthesis group (18 patients [8.5%]) (~7 [Fisher’s exact] = 0.06). Coronary artery disease was more common in the bioprosthesis group (32 patients [15.1%]) than it was in the mechanical valve group (26 patients [lo%]) (r? [Fisher’s exact] = 0.07). Of the 204 patients with mechanical prostheses who were taking Coumadin (crystalline warfarin sodium; DuPont Pharmaceuticals, Wilmington, DE), no patient had had an episode of major anticoagulant-related hemorrhage. The preoperative catheterization data indicated that patients with failed bioprostheses had evidence of higher right ventricular and pulmonary arterial pressures (eg, mean pulmonary arterial pressure: 39 mm Hg in bioprosthesis group versus 32 mm Hg in mechanical valve group; ~7 [Wilcoxon] = 0.0001). In addition, the bioprosthesis group had more severe aortic valve stenosis (p [Wilcoxon] = 0.0002) and mitral valve stenosis and incompetence (p = 0.0001). However, there was no hemodynamic difference between the bioprosthesis and mechanical valve groups in terms of the degree of tricuspid stenosis, the angiographically shown severity of tricuspid valve incompetence, or the angiographically shown severity of aortic valve incompetence. Procedures Procedures performed at the current reoperation consisted of valvular replacements in 411 patients (85.9%), repair of a paravalvular leak in 61 patients (12.8%), valvular thrombectomy in 11 patients (2.3”~,,), removal of a valvular pannus in 3 patients with Bjork-Shiley valves (0.62”/0), and rotation of a Bjiirk-Shiley aortic prosthesis in 1 patient (0.21%). Overall, 487 procedures were performed in the 478 patients. The mean time interval from the previous valve replacement to the current valve replacement (n = 430 prostheses) was significantly less in the bioprosthesis group at (74 months; range, 0.2 to 307 months) than it was in the mechanical valve group (84.7 months; range, 0 to 242 months) (p [Fisher’s exact] = 0.02).

Ann

Thorac Surg 1995;60:5475-8

At reoperation, of the 147 prostheses implanted in the aortic position, there were 52 St. Jude Medical valves (40%), 52 BjGrk-Shiley valves (35%), 20 Ionescu-Shiley valves (13%), 5 Starr-Edwards valves (4%), 4 MedtronicHall valves (3%), 4 Carpentier-Edwards valves (3%), 2 Smeloff-Cutter valves (l%), and 1 Mitroflow valve (0.5%), for a total of 122 mechanical valves (83%) and 25 bioprosthetic valves (17%) implanted in the aortic position at reoperation. At reoperation, of the 273 prostheses implanted in the mitral position, there were 94 St. Jude Medical valves (34%), 70 BjGrk-Shiley valves (25%), 35 CarpentierEdwards valves (73%), 32 Ionescu-Shiley valves (12%), 28 Beall valves (lo%), 6 Medtronic-Hall valves (2%), 4 Hancock valves (2%), 2 Tascon valves (l%), and 2 SmeloffCutter valves (l%), for a total of 200 mechanical valves (73%) and 73 bioprosthetic valves (27%) implanted in the mitral position at reoperation. At reoperation, of the 10 prostheses implanted in the tricuspid position, there were 4 Carpentier-Edwards valves (40%), 3 Ionescu-Shiley valves (30%), 2 St. Jude Medical valves (20”/“), and 1 Beall valve (lo%), for a total of 7 bioprosthetic valves (70%) and 3 mechanical valves (30%) implanted in the tricuspid position at reoperation. At reoperation 20 patients (9.4%) in the bioprosthesis group and 27 patients (10.2%) in the mechanical valve group underwent associated coronary artery bypass grafting (p [Fisher’s exact] = 0.88).

Mechanical Valve Patients There were 145 mitral valves, 62 aortic valves, and 2 tricuspid valves replaced in the mechanical valve group, for a total of 209 valves replaced in the 266 patients. The intraoperative findings in the patients with the failed mechanical valves (289 findings in 266 patients) consisted of paravalvular leaks in 105 (36%), evidence of disc, poppet, or ball wear in 73 (25”/0), thrombosis in 52 (18%), pannus formation in 41 (14”%), prosthetic endocarditis in 11 (4%), patient-prosthesis mismatch in 4 (1.4%), and poppet escape in 3 (1%). Procedures performed at reoperation included 209 valvular re-replacements (740/u), 59 paravalvular leak repairs (20.8%), 11 thrombectomies (4%), 3 pannus removal procedures (lo&), and 1 rotation of an aortic prosthesis (0.35%), for a total of 283 procedures in the 266 patients. The prostheses explanted included 112 Beall valves (54%), 49 BjGrk-Shiley valves (24%), 13 Starr-Edwards valves (6%), 13 Cross-Jones valves (6%), 7 St. Jude Medical valves (4%), 3 Smeloff-Cutter valves (l%), 3 LilleheiKaster valves (l%), 3 Medtronic-Hall valves (l%), 1 McGovern valve (0.5%), 1 Braunwald-Cutter valve (0.5%), 1 Omniscience valve (0.5%), 1 Sorin valve (0.5%), 1 Cooley valve (0.5%), and 1 Wada-Cutter valve (0.5%), for a total of 209 valves explanted from the 266 patients. Of the 160 (77%) mechanical prostheses implanted at reoperation, there were 92 Bjiirk-Shiley valves (44%), 39 St. Jude Medical valves (19”/0), 27 Beall valves (13%), and 2 Medtronic-Hall valves (1%). Of the 49 (23%) bioprosthetic valves implanted, there were 28 CarpentierEdwards valves (13%), 16 Ionescu-Shiley valves (S%), 3

Ann Thorac 1995;60:5475-8

CARDlAC

Surg

Hancock valves (l%), and 2 Tascon valves there was a total of 209 valves implanted patients undergoing valvular replacement.

(1%). Thus in the 203

Bioprosthetic Valve Patients There were 128 mitral valves, 85 aortic valves, and 8 tricuspid valves replaced in the bioprosthesis group, for a total of 221 valves replaced in the 212 patients. The intraoperative findings in the patients with the failed bioprosthetic valves (263 findings in 212 patients) consisted of calcification in 101 (48%), leaflet tear in 83 (39%), paravalvular leak in 29 (14%), prosthetic valve endocarditis in 13 (6%), leaflet wear in 12 (6%), leaflet disruption in 11 (5%), pannus ingrowth in 10 (5”&), and fibrosis with stenosis in 4 (2%). Procedures performed at reoperation included 221 valvular replacements (98%), 3 paravalvular leak repairs (1.37%), and 1 valvular thrombectomy (0.4%), for a total of 225 procedures in the 212 patients. The prostheses explanted included 83 Ionescu-Shiley valves (37%), 79 Carpentier-Edwards valves (360/n), 31 Hancock valves (14%), 16 heterografts (7”/0), 10 Mitrotlow valves (5%), and 2 homograft valves (l%), for a total of 221 valves explanted from the 212 patients. Of the 163 (74%) mechanical valves implanted at reoperation, there were 116 St. Jude Medical valves (52’!h), 30 Bjork-Shiley valves (14%), 8 Medtronic-Hall valves (3’%), 5 Starr-Edwards valves (2%), 2 Beall valves (lo/o), and 2 Smeloff-Cutter valves (1%). Of the 58 (26%) bioprosthetic valves implanted, there were 39 Ionescu-Shiley valves (17%), 15 Carpentier-Edwards valves (7%), 3 Mitroflow valves (l%), and 1 Hancock valve (0.4%). Thus there was a total of 221 valves implanted in the 212 patients.

Paravalvular Leak There were 62 paravalvular leaks repaired in 61 patients, 32 at the mitral position and 30 at the aortic position. The mean interval from the time of the previous valvular replacement to the repair of the paravalvular leak was 32.2 months (range, 0.1 to 133 months). Of the valves involved, 59 were mechanical (95%) and 3 were bioprostheses (5%). The mechanical prostheses involved consisted of 32 Bjork-Shiley valves (54%), 14 St. Jude Medical valves (24%), 7 Beall valves (12%), 3 Starr-Edwards valves (5%), and 3 Medtronic-Hall valves (5%). The mean interval from the previous mechanical valvular replacement to the repair of the paravalvular leak was 32.5 months (range, 0.1 to 133 months). A bioprosthetic paravalvular leak was repaired in 3 patients, each of whom had an Ionescu-Shiley valve. The mean interval from the previous valvular replacement to the repair of the bioprosthetic paravalvular leak was 27.8 months (range, 1 to 65 months).

BIOPROSTHESES PATIENT STATUS

AFTER

McCRATH VALVE

ET AL FAILURE

S477

months) after implantation. Five Beall valves and one Medtronic-Hall valve in the mitral position were thrombectomized, and these were done a mean of 30 months (range, 5 to 80 months) postoperatively. Only one bioprosthesis was thrombectomized, a Hancock valve, and this was done 8 months after tricuspid valve replacement.

Postoperative Events There was no difference between the two groups in the incidence of multiple events, which included use of an intraaortic balloon or the development of neurologic events, hepatic failure, pulmonary failure, renal failure, wound infection, atria1 arrhythmias, or ventricular arrhythmias. The sole postoperative event that was significantly different between the two groups in terms of incidence, for those patients undergoing valvular replacement, was surgically induced complete atrioventricular dissociation requiring pacemaker implantation, which occurred in 26 patients in the bioprosthesis group (12.3%) and in 10 patients in the mechanical valve group (3.8%) (p [Fisher’s exact] = 0.0007).

Survival There were 28 hospital deaths (13.2%) in the bioprosthesis group and 33 hospital deaths (12.4%) in the mechanical valve group (p [Fisher’s exact] = 0.89). The modes of hospital death in the mechanical valve group consisted of chronic cardiac failure in 12 (36%), acute cardiac failure in 7 (21%), sepsis in 3 (9%), bleeding in 3 (9%), endocarditis in 2 (6%), arrhythmia in 1 (3%), a neurologic event in 1 (3”/0), renal failure in 1 (3%), pulmonary failure in 1 (3%), multisystem organ failure in 1 (30/b), and mesenteric thrombosis in 1 (3%). The modes of hospital death in the bioprosthesis group consisted of acute cardiac failure in 9 (32%), chronic cardiac failure in 6 (21’%), sepsis in 6 (21%), endocarditis in 3 (lo%), a neurologic event in 1 (4%), bleeding in 1 (4?L), anticoagulant-related bleeding in 1 (4%), and renal failure in 1 (4%). Multiple variables were analyzed to determine the incremental risk factors for hospital death, and this analysis identified higher preoperative functional class (p [J test] = O.OOOOl), longer aortic cross-clamp time (p [Wilcoxon] = O.OOOl), longer total cardiopulmonary bypass time (~7 [Wilcoxon] = O.OOOl), hepatomegaly (~7 [Fisher’s exact] = 0.002), ascites (p [Fisher’s exact] = 0.02), and a decreasing preoperative ejection fraction (p [Wilcoxon] = 0.05) as risk factors. Actuarial survival was similar for both the bioprosthesis and mechanical valve groups (p [log-rank] = 0.13) (Fig 1).

Thrombectomy

Comment

There were 12 valvular thrombectomies performed as the primary procedure, and this was necessitated by prosthesis failure, Five Bjork-Shiley valves in the aortic position were thrombectomized at a mean of 27 months (range, 3 to 60

The issue of the intrinsic failure of bioprostheses versus the disadvantages of the anticoagulant therapy needed by patients with mechanical prostheses is an important consideration when it comes to choosing a prosthetic valve [6,7]. There have been reports seeking to minimize

s478

CARDLAC PATIENT

BIOPROSTHESES ST4TUS AFTER

VAI.VE

McGRATH FAILURE

Ann

ET Al

tence, tricular

%SURVIVAL -

Biologx.al

t

Mechanical

pulmonary the patients

in

arterial with

and right venbioprostheses.

Therefore, although there has been the perception that patients with failed mechanical valves are less well at the time of reoperation, indeed the patients with bioprostheses had evidence of more severe valvular dysfunction, together with pulmonary hypertension and right ventricular failure, at least according to the hemodynamic as-

40 3,~.

sessment

P(LOG-RANK)=0.13

done

immediately

before

reoperation.

We conclude

20 10 0

and increasing pressures

Thorac Surg 1995;60:5475-8

5

0

10

15

I 20

25

YEARS

the drawbacks of reoperations required to manage intrinsic dysfunction of bioprosthetic valves IS]. The durability of bioprosthetic devices may or may not be enhanced by recent modifications [9]. In addition, often patients who receive bioprosthetic valves so that they do not have to take anticoagulants subsequently require anticoagulation because of the development of emboli or tachydysrhythmias. The problems related to anticoagulation have been ameliorated in part by the use of low-dose protocols [lo]. In fact, in the series reported on here no patient suffered a major anticoagulant-related event. Indeed, in an earlier study of patients with multivalvular heart disease, the use of anticoagulation was found to confer an overall survival advantage [ll]. Therefore the advantages of bioprosthetic valves seem to have become less clear and the criteria for their use more narrow 1721. Bioprosthetic valves are useful for those patients who have a contraindication to anticoagulant therapy and in young women who desire pregnancy [13]. The risk of thromboembolic events is also reported to be less in patients with biologic valves [14] and resistance to the development of infective endocarditis is reported to be increased [15]. Another purported advantage of bioprosthetic valves is their anticipated subtle manner of failure, in contrast to the catastrophic mechanisms of failure anticipated for mechanical valves. The purpose of this report was to examine the perioperative clinical status, modes of valve dysfunction, and surgical outcomes in those patients undergoing reoperation for failed mechanical and bioprosthetic valves. The goal was to determine whether

there

were

important

differences

in the

status

of

patients in the two groups. From the standpoint of perioperative clinical status, there was no difference in the functional class between the two groups. Also of note was the fact that the patients with worse

erative

failed bioprosthetic hemodynamic status

catheterization

valves at the

showed evidence time of their prereop-

than did concurrent

patients

of

with

failed mechanical valves. There was also an important increase in the severity of obstruction at the mitral and aortic valve levels, more severe mitral valve incompe-

that patients with mechanical prostheses are not in worse condition at the time of reoperation for valve failure. The relatively low risk of valve-related events associated with current mechanical valve prostheses and the low incidence of anticoagulant-related hemorrhage found in this study should be considered when devising treatment strategies that involve the choice of valvular prostheses.

References 1. Teoh

kowski tion

KH, Ivanov H. Clinical

of bioprosthetic

J, Weisel RD, and Doppler valve

failure

Daniel LB, Darcel echocardiographic after

10 years.

IC, Raevalua-

Supplement

IV Circulation 1990;82(Suppl 4):110-6. 2. Cohn LH. Valve re-replacement in the asymptomatic tient. Ann Thorac Surg 1991;51:357-8. 3. Kontos GJ, Schaff HV, Orszulak TA, Puga FJ, Pluth Danielson

4.

5.

6.

7.

8. 9. 10.

11.

12.

13.

14.

15.

GK.

Thrombotic

obstruction

of disc

valves:

paJR, clini-

cal recognition and surgical management. Ann Thorac Surg 1989;48:60-5. Lund 0, Pilegaard HK, Magnussen K, Knudsen MA, Nielsen TT, Albrechtsen OK. Long-term prosthesis-related and sudden cardiac-related complications after valve replacement for aortic stenosis. Ann fhorac Surg 1990;50:396-406. Edmunds H. Clark RE. Cohn LH. Miller DC, Weisel RD. Guidelines for reporting morbid& and mortality after cardiac valvular operations. Ann Thorac Surg 1988;46:257-9. MacArthur KJD, Bain WH, Turner MA, et al. Mechanical versus biological heart valves: a ten-year comparison in a single centre. Eur J Cardiothorac Surg 1988;2:143-50. Hammermeister KE, Sethi GK, Henderson WG, Oprian C, Kim T, Rahimtoola S. A comparison of outcomes in men 11 years after heart-valve replacement with a mechanical valve or bioprosthesis. N Engl J Med 1993;328:1289-96. Cohn LH, Aranki SF, Rizzo RJ, et al. Decrease in operative risk of reoperative valve surgery. Ann Tborac Surg 1993;56:15-21. Williams MA. The intact bioprosthesis-early results. J Cardiac Surg 1988;3:347-58. Saour JN, Sieck JO, Mamo LAR, Callus AS. Trial of different intensities of anticoagulation in patients with prosthetic heart valve. N Engl J Med 1990;322:428-32. McGrath LB, Gonzalez-Lavin L, Bailey BM, Grunkemeier GL, Fernandez J, Laub GW. Tricuspid valve operations in 530 patients: twenty-five-year assessment of early and late phase events. J Thorac Cardiovasc Surg 1990;99:124-33. Magilligan DJ, Lewis JW Jr, Stein I’, Alam M. The porcine bioprosthetic heart valve: experience at 15 years. Ann Thorac Surg 1989;48:324-30. Salazar E, Zajarias A, Gutierrez N, Iturbe I. The problem of cardiac valve prostheses, anticoagulants, and pregnancy. Circulation 1984;70:1-169-77. Borkon AM, Soule LM, Baughman KL, et al. Comparative analysis of mechanical and bioprosthetic valves after aortic valv; replacement. J Thorac Cardiovasc Surg 1987;94:20-33. Haydock D, Barratt-Boyes B, Macedo T, Kirklin JW, Blackstone E. Aortic valve replacement for active infectious endocarditis in 108 patients: a comparison of freehand allograft valves with mechanical prostheses and bioprostheses. J Thorac

Cardiovasc

Surg

1992;103:130-9.