Early and Late Phase Events Following Bioprosthetic Tricuspid Valve Replacement LYNN 6. McGRATH, M.D., CHAO CHEN, PH.D., BRIDGET M. BAILEY, B.S.N., JAVIER FERNANDEZ, M.D., GLENN W. LAUB, M.D., and MARK S. ADKINS, M.D. Department of Surgery, Deborah Heart and Lung Center; Browns Mills, and Department of Surgety University of Medicine and Dentistry of New Jersey, Robert Wood Johnson Medical School, New Brunswick, New Jersey ABSTRACT From 1961 through 1987,9,247 patients underwentan intracardiac repair for valvular heart disease. Five hundredthirty patients had a procedurethat included a tricuspid valve operation (6%),with tricuspid valve replacement performed in 175 patients ( ~ V Oof ) , whom 154 had a bioprosthetic valve implanted (1.7%). These 154 patients with a bioprosthetic valve in the tricuspid position are the subject of this review. There were 27 males and 127 females. Ages ranged from 10 to 75 years. There was tricuspid valve insufficiency in 139 patients (90%),and stenosis plus insufficiency in 15 (10%). Carpentier-Edwards prostheses were implanted in 83 (54%), lonescu-Shiley in 55 (35%), Hancock in 12 (8%), and Mitroflow in 4 (3%). Concomitant procedureswere performed in 146 patients (95%).At least one previous operation had been performed in 86 patients (56%). Preoperatively, 139 patients were in functional Class 111 or IV (90%). Hospital death occurred in 20 patients (1 3%). Logistic regression analysis revealed that incremental risk factors for hospital death included increasing peripheral edema preoperatively (p = 0.04), and use of a Hancock prosthesis in the tricuspid position (p = 0.03). All 134 hospital survivors were followed at a mean of 66.01 months, range 1 to 162 months. There were 70 late deaths (52%). Log-rank test indicated that incremental risk factors for late death were: longer cross-clamp time at repair (p = 0.0007); higher pulmonary artery systolic pressure preoperatively (p = 0.01); earlier date of surgery (p = 0.03); and larger tricuspid prosthesis size (p = 0.06). The incidence of valve-related events for bioprostheses implanted in the tricuspid position was: paravalvular leak (O%), thrombosis (1 %), and endocarditis (2%).Seventeen patients (1 2.7%)had structural failure of a bioprosthesis in the tricuspid position, with 15 having a reoperation performed at a mean of 74 months post repair. Fourteen of the patients undergoing reoperation required tricuspid valve re-replacement, and one a tricuspid valve thrombectomy. Actuarial freedom from tricuspid valve re-replacement was 7'0% at 10 years. Different bioprosthesis type was the sole incremental risk factor for the requirementfor late tricuspid valve re-replacement(p = 0.0001). We conclude that bioprostheses in the tricuspid position are at relatively low risk of valve-related events. However, there is an important hazard for decreased overall survival in these patients. Risk factors for premature late death are related to earlier date of surgery, more complex repairs requiring prolonged aortic occlusion to accomplish, and signs of increasing right heart failure. bioprosthetic valves, structural failure, tricuspid insufficiency, tricuspid stenosis
Presented at the Vth International Symposium on Cardiac Bioprostheses, May 24-27, 1991, Avignon, France.
Vol. 7,No. 3, 1992
Address for correspondence: Lynn B. McGrath, M.D., Chairman* Department Of Surgeryi Deborah Heart and Lung Center, 200 Trenton Road, Browns Mills, NJ 08015. Fax: (609) 893-1212.
Journal of Cardiac Surgery
245
McGRATH. CHEN, BAILEY, ET AL.
The surgical literature is replete with reports regarding the fate of mechanical prostheses implanted in the tricuspid position.l-1° However, there is little information available regarding early and late phase events following bioprosthetic valve implantation in the tricuspid position.11-15 The purpose of this review was to retrospectively analyze events in a consecutive series of patients having tricuspid valve replacement utilizing a bioprosthetic valve. METHODS Patients
A retrospective analysis was made of 9,247 consecutive patients, each of whom underwent a cardiac valvular operation at the Deborah Heart and Lung Center between January 1, 1961 and December 31,1987. Five hundred thirty of these patients had a repair that required a tricuspid valve operation (5.7%). Three hundred fifty-one of those patients underwent a tricuspid valve repair (3.8%). The remaining 175 patients had a tricuspid valve replacement procedure performed (1.9%). Twenty-one patients had a mechanical prosthesis placed in the tricuspid position, and the remaining 154 had a bioprosthetic valve implanted (1.7%). These latter 154 patients were analyzed for this report. FoI10 W-u p
All patients were followed in 1990 by telephone call to the patient, the family, and/or the referring physician. Follow-up was performed at a mean of 66 months, range 1 to 162 months postoperatively. Statistical analysis
The methods used to determine differences between groups included the calculation of p values by Chi-square or Fisher methods for categorical variables, and t-tests for continuous variables. Time-related events were analyzed by product-limit estimator and log-rank tests. Hazard function analysis was used to assess instantaneous risk of events.
246
RESULTS
Patients
There were 27 males (17.5%) and 127females (82.5%). The mean age at repair was 55.8 years with a range of 10 to 75 years. Four patients were in preoperative New York Heart Association Functional Class I (2.6%), 11 patients were in Class If (7.2%), 90 were in Class Ill (58.4%), and 49 were in Class IV (31.8%). One hundred twenty-two of the 154 patients had undergone preoperative right ventricular angiocardiographic assessment of the severity of tricuspid valve incompetence, by methods previously described! Grade 1 incompetence was found in 2 patients (1.6%), grade 2 in 15 patients (12.3%), grade 3 in 32 patients (26.2%), and grade 4 in 73 patients (59.9%). The predominant pathology found at the time of operation was tricuspid valve insufficiency in 139 (90.3%), and a combination of tricuspid valve stenosis plus insufficiency in 15 patients (9.7%). The prostheses placed in the tricuspid position included Carpentier-Edwards valves (Edwards CVS Div., Baxter Healthcare Corp., Irvine, CA, USA) in 83 patients (53.9%), lonescu-Shiley in 55 (35.7%), Hancock (Medtronic, Inc., Minneapolis, MN, USA) in 12 (7.80/0), and Mitroflow (Symbion Inc., Salt Lake City, UT, USA) in 4 (2.6%). Four of the patients had failed primary repair with secondary tricuspid valve replacement. At least one previous operation had been performed in 86 patients (55.8%), two previous operations had been performed in 20 patients (12.9%), and three previous operations had been performed in 3 patients (1.95%). Associated valvular disease was common. Mitral insufficiency was present in 116 patients (75.3%), mitral stenosis in 100 (64.9%), aortic stenosis in 15 (9.7%), and aortic insufficiency in 85 (55.2%). Coronary artery disease requiring coronary artery bypass grafting was present in 19 patients (12.3%). Overall, concomitant procedures were performed in 146 patients (94.8%) (Table 1).
Journal of Cardiac Surgery
Vol. 7, No. 3,1992
TISSUE VALVES IN T H E TRlClUSPlD POSITION
TABLE 1 Bioprosthetic Tricuspid Valve Replacement Deborah Heart and Lung Center 1961-1 987 N = 154 ~~
~
Concomitant Procetdures Procedures No. Mitral Valve Replacement 96 Mitral Valve Replacement + Aortic Valve Replacement 30 Mitral Valve Replacement + CABG 8 Aortic Valve Replacement 3 Mitral Valve Repair 3 Aortic Valve Replacement + Mitral Repair 2 2 Mitral Repair + Aortic Valve Repair Aortic Valve Replacement + Mitral Valve Replacement + CABG 1 Aortic Repair + Mitral Valve Replacement 1 Total 146' Associated procedures performed in conjunction with bioprosthetic tricuspid valve replacement. 'Eight patients underwent isolated tricusDid valve replacement. CABG .= coronary artery bypass grafting.
Survival Of the 154 patients, 20 (12.9%) suffered hospital death (70% confidence limits: 10.1%16.4%). The modes of hospital death included acute cardiac failure in 8 patients (5.2%), subacute cardiac failure in 7 (4.5%), chronic cardiac failure in 1 (0.6%), 1 died of bleeding (0.6%), 1 of endocarditis (O.6%), 1 of sepsis (0.6%), and 1 died of a neurological event (0.6%). An extensive univariate analysis was conducted to determine those incremental risk factors for the event of hospital death. Factors ultimately noted to be significant by logistic regression analysis included increasing peripheral edema preoperatively (p = 0.04) and the use of a Hancock prosthesis in the tricuspid position (p = 0.03). All of the 134 hospital survivors were followed late postoperatively at a mean of 66 months, range 1 to 162 months. Late death occurred in 70 (52.2%) patients (70% confidence limits: 47.4%-57%). Modes of late death included chronic cardiac failure in 17 (24.3%), acute cardiacfailure in 9(12.9%), subacutecardiacfailure in 8 (11.4%), infective endocarditis in 6 (8.6%), sepsis in 6 (8.60/0), myocardial infarction in 5 (5.7%), sudden death in 2 (2.9%), neurological in 2 (2.9%), bleeding in 1 (1A%), anticoagulantrelated hemorrhage in 1 (1.4%), valve strut failure in 1 (1.4%), valve-related thrombosis in 1
Vol. 7,No. 3,1992
YO 62.3 19.5 5.2 2.0 2.0 1.3 1.3 0.6 0.6 94.8
(1.4%), arrhythmia in 1 (1.4%), and noncardiac in 11 (15.7%). Log-rank test demonstrated that incremental risk factors for late death included: longer total aortic cross-clamp time (p = 0.0007); higher preoperative pulmonary artery systolic pressure (p = 0.01); earlier date of suirgery (p = 0.03); and larger tricuspid bioprosthesis size (p = 0.06). Actuarial survival, including hospital death, was 35% at 10 years (Fig. 1). Hazard function arialysis of death revealed that there was a high early instantaneous risk, an intermediate phase, arid an increasing late phase risk (Fig. 2). At late follow-up, 44 patients were noted to be in New York Heart Association Functional Class I (68.7%), 12 were in Class II (18.8%), 3 were in Cllass Ill (4.7%), and 1 was in Class IV (1.6%).
Valve-related events Seventeen of the hospital survivors developed late structural failure of a bioprosthetic valve in the tricuspid position (12.7%). Two of these had documented stenotic and incompetent bioprosthetic valves, but declined reoperation. Fifteen of the patients underwent reoperation, performed at a mean interval of 74.8 months, range 3 to 136 months postoperatively (Table 2). Of those 15 piatients, 1 patient with a Hancock valve had a successful tricuspid valve thrombectomy at 8 months following the original operation, and the
J o u r n a l of Cardiac Surgery
247
MCGRATH, CHEN, BAILEY, ET AL. OVERALL ACTUARIAL FREEDOM FROM BOTH HOSPITAL AND LATE DEATH AND 70% CONFIDENCE LIMIT N=l54
YO
F R E
E D 0
M
-----
0.70 0.65 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.150.10
0
*....
a*.
..
=*-..\
.. I
I
I
I
12
24
36
48
I
I
I
I
I
60
72
84
96
108
I
120
I
132
I
i
144
156
TIME (MONTHS) # of patients at 135 months: 10
Figure 1. Actuarial survival, including hospital death.
HAZARD FUNCTION OF HOSPITAL AND LATE DEATH N * 154
10
30
50
70 90 TIME (months)
110
130
150
Figure 2. Hazard function of hospital and late death.
other 14 required tricuspid valve re-replacement. Figure 3 indicates that actuarial freedom from reoperation for bioprosthetic valves in the tricuspid position at 10 years was about 65%. A hazard function analysis of the event reoperation for the tricuspid valve position is depicted in Fig248
ure 4. This nomogram indicates that most of the reoperations occurred after 8 years following surgery. A log-rank test was used to compare actuarial freedom from reoperation of different prostheses simultaneously. The results showed a significant
Journal of Cardiac Surgery
Vol. 7 , No. 3,1992
TISSUE VALVES IN THE TRlClJSPlD POSITION
TABLE 2 Bioprosthetic Tricuspid Valve Replacement Deborah Heart and Lung Center 1961-1 987 N = 134* Late Reoperation Involving Tricuspid Valve Tricuspid Findings at BioDrosthesis N No. ReODeration Carpentier 83 10 Calcification 1 Endocarditis 1 Pannus lngrowth 55 1 Aspergilloma Obstructing lonescu-Shiley the Valve Hancock 12 1 Thrombus Mitroflow 4 1 Calcification Operative findings at reoperation for bioprosthetic valve failure in-the tricuspid-position. 'Mean interval to reoperation = 74.8 months, range 3-136 months.
difference among the four types of prostheses (p = 0.001). A further analysis involving pair-wise comparison revealed that these four types of valves can be divided into two homogeneous groups: lonescu-Shiley plus Carpentier group; and Hancock plus Mitroflow group. Within-group differences are not significant; however, the difference between the two groups reaches a highly significant level (p = O.OOOl), as indicated in Figure 5. f-ive patients developed tricuspid paravalvular leak postoperatively; however, none of them required reoperation. The actuarial freedom from paravalvular leak in the tricuspid position is shown in Figure 6. Overall, 14 of the hospital survivors developed infective endocarditis (10.5%), 3 occurring in the mitral position, 8 in an unknown position, and 3 documented on a tricuspid bioprosthesis (2%). Figure 7 indicates actuarial freedom from bioprosthetic valve endocarditis at the tricuspid position, with no events occurring after 9 months postoperatively.
ACTUARIAL FREEDOM FROM REOPERATION FOR TRlCUSPlD VALVE POSITION AND 70% CONFIDENCE LIMIT # OF PATIENTS WITH SUCH EVENTS = 15 N=l54 1 .oo
0.95 0.90 0.85 0.80
F R E E
0
M
0.75
0.70 0.65 0.60
0.55
0.40 0
12
24
36
48
60
72
I
I
1
I
I
I
1
84
96
108
120
132
144
156
TIME (MONTHS) # of patients at risk after the second episode: 118, 11, 88, 69, 41, 37, 36, 30, 18, 16, 14, 7.
Figure 3 . Actuarial freedom from reoperation for bioprosthetic valve failure in the tricuspid position, including 70% confidence limits.
Vol. 7, No. 3,1992
Journal of Cardiac Surgery
249
MCGRATH, CHEN, BAILEY, ET AL. HAZARD FUNCTON OF iRRICUSPIDVALVE FAILURE N I154 0.007
-
0.006
..
0.005
..
0.004
..
.. 0.002 .' 0.003
O.0OI 0.000
I
10
30
50
70
90
110
130
150
TIME (months)
Figure 4. Hazard function of bioprosthetic valve failure in the tricuspidposition. ACTUARl4LFREEDOMFRWREOPERATKNATTHE TRlCUSPlDVALVE POsrrlONFOR VARJCUS BlOPRosMESES AND 7% CONFIDEWE LIMITS
N-154
I
lonescu8hiley and Carpentier.
0 #ofreoDerationi13. N-1%
0.4
! 0
200
100
TIME (MONTH) log-nnk tort of equality Over types of prO8the8k p=O.OoOl
after the second epl8ode: lonorcu-Shlley and Carpontler group: 110, 69, 42, 38. 37, 30, 18, 16. 14, 7 limncock and Mltroflow group: 6, 3
# of patlents at rl8k
Figure 5 . Actuarial freedom from reoperation for bioprosthetic valve failure in the tricuspid position, according to prosthesis type.
DISCUSSION
Patients with valvular heart disease uncommonly require tricuspid valve surgery.' Earlier reports indicated that nonsurgical treatment of those patients with tricuspid valve regurgitation
250
undergoing surgery for severe mitral valve disease would lead to a satisfactory However, this conservative approach occasionally led to a poor result, and the requirement for late reoperation for residual or
Journal of Cardiac Surgery
Vol. 7, No. 3,1992
TISSUE VALVES IN THE TRlCLlSPlD POSITION
0.90-
0.92
0.M
*%.
I
I
I
Figure 6. Actuarial freedom from paravalvular leak involving the tricuspid position, including 70% confidence limits.
progressive tricuspid valve incompetence. Therefore, presently, patients with important tricuspid valve disease, undergoing surgical intervention for lefi-sided valvular disease, will usually have tricuspid valve surgery as well. Tricuspid valve replacement in the modern era has been utilized with decreasing frequency. The reasons for this are m u l t i f a c t ~ r i a l .Earlier ~~-~~ surgical treatment for patients with mitral and aortic valve disease has perhaps led to a relative decrease in the number of patients presenting with end-stage evidence of right ventricular failure and severe degrees of tricuspid valve incompetence. Furthermore, there have been important advances in the use of tricuspid valve reparative technique^?^ Although implantation of mechanical prostheses in the tricuspid position has been shown to have an important risk of associated complications,l-l0in particular a high incidence of thromboembolic events, little information is available regarding morphological changes occurring in bioprosthetic valves utilized in the tricuspid position. Roberts et al.25 reported degenerative changes in mitral and tricuspid bioprostheses explanted simultaneously, and found that in
Vol. 7 , No. 3,1992
cornparison, those in the tricuspid position apparently were at reduced risk for biodegeneration. Nonetheless, in an accompanying editorial, Horowitz maintained that bioprosthetic valves may not be appropriate in the tricuspid position, as the low pressure found in the right heart may induce stasis, with the attendant risk for incomplete valve opening. Particularly in patients with reduced cardiac output and a dilated, fibrillating right atrium, it was postulated that for a bioprosthesis in the tricuspid position, there may not be a sufficient rate of pressure rise to open the bioprosthetic valve adequately, and this could potentially lead to valve failureJ6 Although the use of bioprosthetic and homograft tissue valves for tricuspid valve replacement has been advocated,27there is little long-term follow-up data reported, and therefore the performance of bioprostheses in the tricuspid position is not secure. However, Guerra et aI.l4 reported good results with the use of bioprosthetic valves in the tricuspid position, and favored these even when mechanical prostheses were implanted in the left side of the heart. In this review, an important effort was made to
Journal of Cardiac Surgery
251
McGRATH, CHEN, BAILEY, ET AL.
a
ACTUARIAL FREEDOM FROM TRICUSPID BIOPROSTHESIS ENDOCARDITIS AND 70% CONFIDENCE LIMIT # OF PATIENTS WITH SUCH EVENTS = 3 N =l54
0.90
%
F R E E
0.75 0.70 0.65 0.60
0.55 0.50 0.45 0.40
I I
I
I
I
I
I
I
I
I
I
I
I
I
0
12
24
36
48
60
72
84
96
108
120
132
144
1 156
TIME (MONTHS) # of patients at risk after the second episode: 132, 108, 30. 'The upper limit at month 105 Is equal to the estimated survival at month nine, which is 0.9764. The lower limit a1 month 105 was computed as i f we had one more event at month 105, the lower limit is 0.94455.
Figure 7. Actuarial freedom from prosthetic valve endocarditis involving the tricuspid position, including 70% confidence limits.
patients with severe forms of right ventricular focus the study at the tricuspid valve level; howfailure, with the expectation of a low risk of valever, in any study of patients with multivalvular vular-related events; however, there is an imporcardiac disease, it is problematic to track events tant risk of hospital and premature late death. specifically related to the tricuspid position. The results reported herein with the use of tricuspid REFERENCES valve replacement utilizing bioprostheses indicate that the risk of valve thrombosis, prosthetic 1. Bach RJ, From AHL, Castaneda AR, et al: Late valve endocarditis, paravalvular leak, and structhrombotic obstruction of Starr-Edwardstricuspid tural valve failure requiring reoperation were at a valve prosthesis. Chest 61:613,1972. 2. Suwansirikul S, Glassman E, Raia F, et al: Late low risk; however, there is an important risk for thrombosis of Starr-Edwards tricuspid ball valve hospital death in those patients undergoing leftprosthesis. Am J Cardiol34:737,1974. sided valve surgery who also require replace3.Sanfelippo PM, Colonel L, Giuliani ER, et at: ment of the tricuspid valve. Although the Tricuspid valve prosthetic replacement: Early and mortality rate reported in this article is less than late results with the Starr-Edwards prosthesis. J Thorac Cardiovasc Surg 71:441,1976. in some other reported s e r i e ~ , ~ ~clinical *~*-~~ 4. Bourdillon PDV, Sharratt GP: Malfunction of Bjorkevidence for more severe degrees of right Shiley valve prosthesis in tricuspid position. Br ventricular failure was clearly an important risk Heart J 38: 1149,1976. factor for hospital and late death. 5. Raj MVJ, Srinivas V, Evans DW: Thrombotic jamWe conclude that bioprosthetic valves can be ming of a tricuspid prosthesis. Br Heart J 38:1355, 1976. implanted in the tricuspid position in those
252
Journal of Cardiac Surgery
Vol. 7, No. 3, 1992
TISSUE VALVES IN THE TRlCUSPlD POSITION
6. lnberg MV, Havia I, Arstila M: Thrombolytic treatment for thrombotic complication of valve prosthesis after tricuspid valve replacement. Scand J Thorac Cardiovasc Surg 11:195,1977. 7. Jugdutt BI, Fraser RS, Lee SJK, et al: Long-term survival after tricuspid valve replacement: Results with seven different prostheses. J Thorac Cardiovasc Surg 74:20,1977. 8. Peterffy A, Henze A, Jonasson R, et al: Clinical evaluation of the Bjork-Shiley tilting disc valve in the tricuspid position. Scand J Thorac Cardiovasc Surg 12:179,1978. 9. Thorburn CW, Morgan JJ, Shanahan MX, et al: Long-term results of tricuspid valve replacement and the problem of prosthetic valve thrombosis. AmJCardiol51:1128,1983. 10. Boskovic D, Elezovic 1, Boskovic D, et al: Late thrombosis of the Bjork-Shiley tilting disc valve in the tricuspid position: Thrombolytic treatment with streptokinase. J Thorac Cardiovasc Surg 91:1,1986. 11. Rhodes GR, Mclntosh CL, Redwood DR, et al: Clinical and hemodynamic results following triple valve replacement: Mechanical vs. porcine xenograft prostheses. Circulation 56(Suppl 2):11123, 1977. 12. Del Campo C, Akalin H, McKenzie FN: Evaluation of the Carpentier-Edwards porcine bioprosthesis in the tricuspid position. Tex Heart lnst J 9:299, 1982. 13. Pasque M, Williams WG, Coles JG, et al: Tricuspid valve replacement in children. Ann Thorac Surg 44: 164, 1987. 14. Guerra F, Bortolotti U, Thiene G, et al: Long-term performance of the Hancock porcine bioprosthesis in the tricuspid position: A review of fortyfive patients with fourteen-year follow-up. J Thorac Cardiovasc Surg 99:838,1990. 15. Eng J, Ravichandran PS, Kay PH, et al: Long-term results of lonescu-Shiley valve in the tricuspid position. AnnThorac Surg 51:200,1990. 16. McGrath LB, Chen C, Bailey BM, et al: Determination of the need for tricuspid valve replacement: Value of preoperative right ventricular angiocardiography. J lnvas Cardiol3:35,1991. 17. McGrath LB, Gonzalez-Lavin L, Bailey BM, et al: Tricuspid valve operations in 530 patients: Twentyfive year assessment of early and late phase events. J Thorac Cardiovasc Surg 99:124,1990. 18. Braunwald NS, Ross J Jr, Morrow AG: Conservative management of tricuspid regurgitation in patients undergoing mitral valve replacement. Circulation 35,36(Pt 2):163,1967.
Vol. 7, No. 3, 1992
19. Carpentier A, Deloche A, Hanania G, et al: Surgical management of acquired tricuspid valve dislease. J Thorac Cardiovasc Surg 67:53,1974. 20. Baxter RH, Bain WH, Rankin RJ, et al: Tricuspid valve replacement: A five-year appraisal. Thorax 30:158,1975. 21. Breyer RH, McClenathan JH, Michaelis LL, et al: ‘Tricuspid regurgitation: A comparison of non‘operative management, tricuspid annuloplasty, and tricuspid valve replacement. J Thorac Cardiovasc Surg 72:867,1976. 22. Kratz JM, Crawford FA Jr, Stroud MR, et al: Trends and results in tricuspid valve surgery. Chest 88:837,1985. 23. Cohen SR, Sell JE, Mclntosh CL, et al: Tricuspid regurgitation in patients with acquired, chronic, pure mitral regurgitation: Nonoperative management, tricuspid valve annuloplasty, and tricuspid valve replacement. J Thorac Cardiovasc Surg 94:488, 1987. 24. Grondin P, Meere C, Limet R, et al: Carpentier’s annulus and DeVega’s annuloplasty: The end of the tricuspid challenge. J Thorac Cardiovasc Surg 70:852,1975. 25. Cohen SR, Silver MA, Mclntosh CL, Roberts WC: Comparison of late (62-1 40 months) degenerative changes in simultaneously implanted and explanted porcine (Hancock) bioprostheses in the tricuspid and mitral valve positions in six patients. Am J Cardiol53:1599, 1984. 26. Spray TI, Roberts WC: Structural changes in porcine xenografts used as substitute cardiac valves. Am J Cardiol40:319, 1977. 27. Barratt-Boyes BG, Rutherford JD, Whitlock RML, et al: A review of surgery for acquired tricuspid valve disease, including an assessment of the stented semilunar homograft valve, and the results of operation for multivalvular heart disease. Aust N Z J Surg 58:23,1988. 28. Stephenson LW, Kouchoukos NT, Kirklin JW: Triple-valve replacement: An analysis of eight years’ experience. Ann Thorac Surg 23:327, 1977. 29. Macmanus Q, Grunkemeier G, Starr A: Late results of triple valve replacement: A 14-year review. Ann Thorac Surg 25:402, 1978. 30. Kara M, Langlet MF, Blin D, et al: Triple valve procedures: An analysis of early and late results. Thorac Cardiovasc Surg 34: 17,1986. 31. Baughman KL, Kallman CH, Yurchak PM, et al: Predictors of survival after tricuspid valve surgery. Am J Cardiol54:137, 1984.
Journal of Cardiac !Surgery
253
Presented at the Vth International Symposium on Cardiac Bioprostheses, May 24-27, 1991, Avignon, France.
Vol. 7,No. 3, 1992
Address for correspondence: Lynn B. McGrath, M.D., Chairman* Department Of Surgeryi Deborah Heart and Lung Center, 200 Trenton Road, Browns Mills, NJ 08015. Fax: (609) 893-1212.
Journal of Cardiac Surgery
245
McGRATH. CHEN, BAILEY, ET AL.
The surgical literature is replete with reports regarding the fate of mechanical prostheses implanted in the tricuspid position.l-1° However, there is little information available regarding early and late phase events following bioprosthetic valve implantation in the tricuspid position.11-15 The purpose of this review was to retrospectively analyze events in a consecutive series of patients having tricuspid valve replacement utilizing a bioprosthetic valve. METHODS Patients
A retrospective analysis was made of 9,247 consecutive patients, each of whom underwent a cardiac valvular operation at the Deborah Heart and Lung Center between January 1, 1961 and December 31,1987. Five hundred thirty of these patients had a repair that required a tricuspid valve operation (5.7%). Three hundred fifty-one of those patients underwent a tricuspid valve repair (3.8%). The remaining 175 patients had a tricuspid valve replacement procedure performed (1.9%). Twenty-one patients had a mechanical prosthesis placed in the tricuspid position, and the remaining 154 had a bioprosthetic valve implanted (1.7%). These latter 154 patients were analyzed for this report. FoI10 W-u p
All patients were followed in 1990 by telephone call to the patient, the family, and/or the referring physician. Follow-up was performed at a mean of 66 months, range 1 to 162 months postoperatively. Statistical analysis
The methods used to determine differences between groups included the calculation of p values by Chi-square or Fisher methods for categorical variables, and t-tests for continuous variables. Time-related events were analyzed by product-limit estimator and log-rank tests. Hazard function analysis was used to assess instantaneous risk of events.
246
RESULTS
Patients
There were 27 males (17.5%) and 127females (82.5%). The mean age at repair was 55.8 years with a range of 10 to 75 years. Four patients were in preoperative New York Heart Association Functional Class I (2.6%), 11 patients were in Class If (7.2%), 90 were in Class Ill (58.4%), and 49 were in Class IV (31.8%). One hundred twenty-two of the 154 patients had undergone preoperative right ventricular angiocardiographic assessment of the severity of tricuspid valve incompetence, by methods previously described! Grade 1 incompetence was found in 2 patients (1.6%), grade 2 in 15 patients (12.3%), grade 3 in 32 patients (26.2%), and grade 4 in 73 patients (59.9%). The predominant pathology found at the time of operation was tricuspid valve insufficiency in 139 (90.3%), and a combination of tricuspid valve stenosis plus insufficiency in 15 patients (9.7%). The prostheses placed in the tricuspid position included Carpentier-Edwards valves (Edwards CVS Div., Baxter Healthcare Corp., Irvine, CA, USA) in 83 patients (53.9%), lonescu-Shiley in 55 (35.7%), Hancock (Medtronic, Inc., Minneapolis, MN, USA) in 12 (7.80/0), and Mitroflow (Symbion Inc., Salt Lake City, UT, USA) in 4 (2.6%). Four of the patients had failed primary repair with secondary tricuspid valve replacement. At least one previous operation had been performed in 86 patients (55.8%), two previous operations had been performed in 20 patients (12.9%), and three previous operations had been performed in 3 patients (1.95%). Associated valvular disease was common. Mitral insufficiency was present in 116 patients (75.3%), mitral stenosis in 100 (64.9%), aortic stenosis in 15 (9.7%), and aortic insufficiency in 85 (55.2%). Coronary artery disease requiring coronary artery bypass grafting was present in 19 patients (12.3%). Overall, concomitant procedures were performed in 146 patients (94.8%) (Table 1).
Journal of Cardiac Surgery
Vol. 7, No. 3,1992
TISSUE VALVES IN T H E TRlClUSPlD POSITION
TABLE 1 Bioprosthetic Tricuspid Valve Replacement Deborah Heart and Lung Center 1961-1 987 N = 154 ~~
~
Concomitant Procetdures Procedures No. Mitral Valve Replacement 96 Mitral Valve Replacement + Aortic Valve Replacement 30 Mitral Valve Replacement + CABG 8 Aortic Valve Replacement 3 Mitral Valve Repair 3 Aortic Valve Replacement + Mitral Repair 2 2 Mitral Repair + Aortic Valve Repair Aortic Valve Replacement + Mitral Valve Replacement + CABG 1 Aortic Repair + Mitral Valve Replacement 1 Total 146' Associated procedures performed in conjunction with bioprosthetic tricuspid valve replacement. 'Eight patients underwent isolated tricusDid valve replacement. CABG .= coronary artery bypass grafting.
Survival Of the 154 patients, 20 (12.9%) suffered hospital death (70% confidence limits: 10.1%16.4%). The modes of hospital death included acute cardiac failure in 8 patients (5.2%), subacute cardiac failure in 7 (4.5%), chronic cardiac failure in 1 (0.6%), 1 died of bleeding (0.6%), 1 of endocarditis (O.6%), 1 of sepsis (0.6%), and 1 died of a neurological event (0.6%). An extensive univariate analysis was conducted to determine those incremental risk factors for the event of hospital death. Factors ultimately noted to be significant by logistic regression analysis included increasing peripheral edema preoperatively (p = 0.04) and the use of a Hancock prosthesis in the tricuspid position (p = 0.03). All of the 134 hospital survivors were followed late postoperatively at a mean of 66 months, range 1 to 162 months. Late death occurred in 70 (52.2%) patients (70% confidence limits: 47.4%-57%). Modes of late death included chronic cardiac failure in 17 (24.3%), acute cardiacfailure in 9(12.9%), subacutecardiacfailure in 8 (11.4%), infective endocarditis in 6 (8.6%), sepsis in 6 (8.60/0), myocardial infarction in 5 (5.7%), sudden death in 2 (2.9%), neurological in 2 (2.9%), bleeding in 1 (1A%), anticoagulantrelated hemorrhage in 1 (1.4%), valve strut failure in 1 (1.4%), valve-related thrombosis in 1
Vol. 7,No. 3,1992
YO 62.3 19.5 5.2 2.0 2.0 1.3 1.3 0.6 0.6 94.8
(1.4%), arrhythmia in 1 (1.4%), and noncardiac in 11 (15.7%). Log-rank test demonstrated that incremental risk factors for late death included: longer total aortic cross-clamp time (p = 0.0007); higher preoperative pulmonary artery systolic pressure (p = 0.01); earlier date of suirgery (p = 0.03); and larger tricuspid bioprosthesis size (p = 0.06). Actuarial survival, including hospital death, was 35% at 10 years (Fig. 1). Hazard function arialysis of death revealed that there was a high early instantaneous risk, an intermediate phase, arid an increasing late phase risk (Fig. 2). At late follow-up, 44 patients were noted to be in New York Heart Association Functional Class I (68.7%), 12 were in Class II (18.8%), 3 were in Cllass Ill (4.7%), and 1 was in Class IV (1.6%).
Valve-related events Seventeen of the hospital survivors developed late structural failure of a bioprosthetic valve in the tricuspid position (12.7%). Two of these had documented stenotic and incompetent bioprosthetic valves, but declined reoperation. Fifteen of the patients underwent reoperation, performed at a mean interval of 74.8 months, range 3 to 136 months postoperatively (Table 2). Of those 15 piatients, 1 patient with a Hancock valve had a successful tricuspid valve thrombectomy at 8 months following the original operation, and the
J o u r n a l of Cardiac Surgery
247
MCGRATH, CHEN, BAILEY, ET AL. OVERALL ACTUARIAL FREEDOM FROM BOTH HOSPITAL AND LATE DEATH AND 70% CONFIDENCE LIMIT N=l54
YO
F R E
E D 0
M
-----
0.70 0.65 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.150.10
0
*....
a*.
..
=*-..\
.. I
I
I
I
12
24
36
48
I
I
I
I
I
60
72
84
96
108
I
120
I
132
I
i
144
156
TIME (MONTHS) # of patients at 135 months: 10
Figure 1. Actuarial survival, including hospital death.
HAZARD FUNCTION OF HOSPITAL AND LATE DEATH N * 154
10
30
50
70 90 TIME (months)
110
130
150
Figure 2. Hazard function of hospital and late death.
other 14 required tricuspid valve re-replacement. Figure 3 indicates that actuarial freedom from reoperation for bioprosthetic valves in the tricuspid position at 10 years was about 65%. A hazard function analysis of the event reoperation for the tricuspid valve position is depicted in Fig248
ure 4. This nomogram indicates that most of the reoperations occurred after 8 years following surgery. A log-rank test was used to compare actuarial freedom from reoperation of different prostheses simultaneously. The results showed a significant
Journal of Cardiac Surgery
Vol. 7 , No. 3,1992
TISSUE VALVES IN THE TRlClJSPlD POSITION
TABLE 2 Bioprosthetic Tricuspid Valve Replacement Deborah Heart and Lung Center 1961-1 987 N = 134* Late Reoperation Involving Tricuspid Valve Tricuspid Findings at BioDrosthesis N No. ReODeration Carpentier 83 10 Calcification 1 Endocarditis 1 Pannus lngrowth 55 1 Aspergilloma Obstructing lonescu-Shiley the Valve Hancock 12 1 Thrombus Mitroflow 4 1 Calcification Operative findings at reoperation for bioprosthetic valve failure in-the tricuspid-position. 'Mean interval to reoperation = 74.8 months, range 3-136 months.
difference among the four types of prostheses (p = 0.001). A further analysis involving pair-wise comparison revealed that these four types of valves can be divided into two homogeneous groups: lonescu-Shiley plus Carpentier group; and Hancock plus Mitroflow group. Within-group differences are not significant; however, the difference between the two groups reaches a highly significant level (p = O.OOOl), as indicated in Figure 5. f-ive patients developed tricuspid paravalvular leak postoperatively; however, none of them required reoperation. The actuarial freedom from paravalvular leak in the tricuspid position is shown in Figure 6. Overall, 14 of the hospital survivors developed infective endocarditis (10.5%), 3 occurring in the mitral position, 8 in an unknown position, and 3 documented on a tricuspid bioprosthesis (2%). Figure 7 indicates actuarial freedom from bioprosthetic valve endocarditis at the tricuspid position, with no events occurring after 9 months postoperatively.
ACTUARIAL FREEDOM FROM REOPERATION FOR TRlCUSPlD VALVE POSITION AND 70% CONFIDENCE LIMIT # OF PATIENTS WITH SUCH EVENTS = 15 N=l54 1 .oo
0.95 0.90 0.85 0.80
F R E E
0
M
0.75
0.70 0.65 0.60
0.55
0.40 0
12
24
36
48
60
72
I
I
1
I
I
I
1
84
96
108
120
132
144
156
TIME (MONTHS) # of patients at risk after the second episode: 118, 11, 88, 69, 41, 37, 36, 30, 18, 16, 14, 7.
Figure 3 . Actuarial freedom from reoperation for bioprosthetic valve failure in the tricuspid position, including 70% confidence limits.
Vol. 7, No. 3,1992
Journal of Cardiac Surgery
249
MCGRATH, CHEN, BAILEY, ET AL. HAZARD FUNCTON OF iRRICUSPIDVALVE FAILURE N I154 0.007
-
0.006
..
0.005
..
0.004
..
.. 0.002 .' 0.003
O.0OI 0.000
I
10
30
50
70
90
110
130
150
TIME (months)
Figure 4. Hazard function of bioprosthetic valve failure in the tricuspidposition. ACTUARl4LFREEDOMFRWREOPERATKNATTHE TRlCUSPlDVALVE POsrrlONFOR VARJCUS BlOPRosMESES AND 7% CONFIDEWE LIMITS
N-154
I
lonescu8hiley and Carpentier.
0 #ofreoDerationi13. N-1%
0.4
! 0
200
100
TIME (MONTH) log-nnk tort of equality Over types of prO8the8k p=O.OoOl
after the second epl8ode: lonorcu-Shlley and Carpontler group: 110, 69, 42, 38. 37, 30, 18, 16. 14, 7 limncock and Mltroflow group: 6, 3
# of patlents at rl8k
Figure 5 . Actuarial freedom from reoperation for bioprosthetic valve failure in the tricuspid position, according to prosthesis type.
DISCUSSION
Patients with valvular heart disease uncommonly require tricuspid valve surgery.' Earlier reports indicated that nonsurgical treatment of those patients with tricuspid valve regurgitation
250
undergoing surgery for severe mitral valve disease would lead to a satisfactory However, this conservative approach occasionally led to a poor result, and the requirement for late reoperation for residual or
Journal of Cardiac Surgery
Vol. 7, No. 3,1992
TISSUE VALVES IN THE TRlCLlSPlD POSITION
0.90-
0.92
0.M
*%.
I
I
I
Figure 6. Actuarial freedom from paravalvular leak involving the tricuspid position, including 70% confidence limits.
progressive tricuspid valve incompetence. Therefore, presently, patients with important tricuspid valve disease, undergoing surgical intervention for lefi-sided valvular disease, will usually have tricuspid valve surgery as well. Tricuspid valve replacement in the modern era has been utilized with decreasing frequency. The reasons for this are m u l t i f a c t ~ r i a l .Earlier ~~-~~ surgical treatment for patients with mitral and aortic valve disease has perhaps led to a relative decrease in the number of patients presenting with end-stage evidence of right ventricular failure and severe degrees of tricuspid valve incompetence. Furthermore, there have been important advances in the use of tricuspid valve reparative technique^?^ Although implantation of mechanical prostheses in the tricuspid position has been shown to have an important risk of associated complications,l-l0in particular a high incidence of thromboembolic events, little information is available regarding morphological changes occurring in bioprosthetic valves utilized in the tricuspid position. Roberts et al.25 reported degenerative changes in mitral and tricuspid bioprostheses explanted simultaneously, and found that in
Vol. 7 , No. 3,1992
cornparison, those in the tricuspid position apparently were at reduced risk for biodegeneration. Nonetheless, in an accompanying editorial, Horowitz maintained that bioprosthetic valves may not be appropriate in the tricuspid position, as the low pressure found in the right heart may induce stasis, with the attendant risk for incomplete valve opening. Particularly in patients with reduced cardiac output and a dilated, fibrillating right atrium, it was postulated that for a bioprosthesis in the tricuspid position, there may not be a sufficient rate of pressure rise to open the bioprosthetic valve adequately, and this could potentially lead to valve failureJ6 Although the use of bioprosthetic and homograft tissue valves for tricuspid valve replacement has been advocated,27there is little long-term follow-up data reported, and therefore the performance of bioprostheses in the tricuspid position is not secure. However, Guerra et aI.l4 reported good results with the use of bioprosthetic valves in the tricuspid position, and favored these even when mechanical prostheses were implanted in the left side of the heart. In this review, an important effort was made to
Journal of Cardiac Surgery
251
McGRATH, CHEN, BAILEY, ET AL.
a
ACTUARIAL FREEDOM FROM TRICUSPID BIOPROSTHESIS ENDOCARDITIS AND 70% CONFIDENCE LIMIT # OF PATIENTS WITH SUCH EVENTS = 3 N =l54
0.90
%
F R E E
0.75 0.70 0.65 0.60
0.55 0.50 0.45 0.40
I I
I
I
I
I
I
I
I
I
I
I
I
I
0
12
24
36
48
60
72
84
96
108
120
132
144
1 156
TIME (MONTHS) # of patients at risk after the second episode: 132, 108, 30. 'The upper limit at month 105 Is equal to the estimated survival at month nine, which is 0.9764. The lower limit a1 month 105 was computed as i f we had one more event at month 105, the lower limit is 0.94455.
Figure 7. Actuarial freedom from prosthetic valve endocarditis involving the tricuspid position, including 70% confidence limits.
patients with severe forms of right ventricular focus the study at the tricuspid valve level; howfailure, with the expectation of a low risk of valever, in any study of patients with multivalvular vular-related events; however, there is an imporcardiac disease, it is problematic to track events tant risk of hospital and premature late death. specifically related to the tricuspid position. The results reported herein with the use of tricuspid REFERENCES valve replacement utilizing bioprostheses indicate that the risk of valve thrombosis, prosthetic 1. Bach RJ, From AHL, Castaneda AR, et al: Late valve endocarditis, paravalvular leak, and structhrombotic obstruction of Starr-Edwardstricuspid tural valve failure requiring reoperation were at a valve prosthesis. Chest 61:613,1972. 2. Suwansirikul S, Glassman E, Raia F, et al: Late low risk; however, there is an important risk for thrombosis of Starr-Edwards tricuspid ball valve hospital death in those patients undergoing leftprosthesis. Am J Cardiol34:737,1974. sided valve surgery who also require replace3.Sanfelippo PM, Colonel L, Giuliani ER, et at: ment of the tricuspid valve. Although the Tricuspid valve prosthetic replacement: Early and mortality rate reported in this article is less than late results with the Starr-Edwards prosthesis. J Thorac Cardiovasc Surg 71:441,1976. in some other reported s e r i e ~ , ~ ~clinical *~*-~~ 4. Bourdillon PDV, Sharratt GP: Malfunction of Bjorkevidence for more severe degrees of right Shiley valve prosthesis in tricuspid position. Br ventricular failure was clearly an important risk Heart J 38: 1149,1976. factor for hospital and late death. 5. Raj MVJ, Srinivas V, Evans DW: Thrombotic jamWe conclude that bioprosthetic valves can be ming of a tricuspid prosthesis. Br Heart J 38:1355, 1976. implanted in the tricuspid position in those
252
Journal of Cardiac Surgery
Vol. 7, No. 3, 1992
TISSUE VALVES IN THE TRlCUSPlD POSITION
6. lnberg MV, Havia I, Arstila M: Thrombolytic treatment for thrombotic complication of valve prosthesis after tricuspid valve replacement. Scand J Thorac Cardiovasc Surg 11:195,1977. 7. Jugdutt BI, Fraser RS, Lee SJK, et al: Long-term survival after tricuspid valve replacement: Results with seven different prostheses. J Thorac Cardiovasc Surg 74:20,1977. 8. Peterffy A, Henze A, Jonasson R, et al: Clinical evaluation of the Bjork-Shiley tilting disc valve in the tricuspid position. Scand J Thorac Cardiovasc Surg 12:179,1978. 9. Thorburn CW, Morgan JJ, Shanahan MX, et al: Long-term results of tricuspid valve replacement and the problem of prosthetic valve thrombosis. AmJCardiol51:1128,1983. 10. Boskovic D, Elezovic 1, Boskovic D, et al: Late thrombosis of the Bjork-Shiley tilting disc valve in the tricuspid position: Thrombolytic treatment with streptokinase. J Thorac Cardiovasc Surg 91:1,1986. 11. Rhodes GR, Mclntosh CL, Redwood DR, et al: Clinical and hemodynamic results following triple valve replacement: Mechanical vs. porcine xenograft prostheses. Circulation 56(Suppl 2):11123, 1977. 12. Del Campo C, Akalin H, McKenzie FN: Evaluation of the Carpentier-Edwards porcine bioprosthesis in the tricuspid position. Tex Heart lnst J 9:299, 1982. 13. Pasque M, Williams WG, Coles JG, et al: Tricuspid valve replacement in children. Ann Thorac Surg 44: 164, 1987. 14. Guerra F, Bortolotti U, Thiene G, et al: Long-term performance of the Hancock porcine bioprosthesis in the tricuspid position: A review of fortyfive patients with fourteen-year follow-up. J Thorac Cardiovasc Surg 99:838,1990. 15. Eng J, Ravichandran PS, Kay PH, et al: Long-term results of lonescu-Shiley valve in the tricuspid position. AnnThorac Surg 51:200,1990. 16. McGrath LB, Chen C, Bailey BM, et al: Determination of the need for tricuspid valve replacement: Value of preoperative right ventricular angiocardiography. J lnvas Cardiol3:35,1991. 17. McGrath LB, Gonzalez-Lavin L, Bailey BM, et al: Tricuspid valve operations in 530 patients: Twentyfive year assessment of early and late phase events. J Thorac Cardiovasc Surg 99:124,1990. 18. Braunwald NS, Ross J Jr, Morrow AG: Conservative management of tricuspid regurgitation in patients undergoing mitral valve replacement. Circulation 35,36(Pt 2):163,1967.
Vol. 7, No. 3, 1992
19. Carpentier A, Deloche A, Hanania G, et al: Surgical management of acquired tricuspid valve dislease. J Thorac Cardiovasc Surg 67:53,1974. 20. Baxter RH, Bain WH, Rankin RJ, et al: Tricuspid valve replacement: A five-year appraisal. Thorax 30:158,1975. 21. Breyer RH, McClenathan JH, Michaelis LL, et al: ‘Tricuspid regurgitation: A comparison of non‘operative management, tricuspid annuloplasty, and tricuspid valve replacement. J Thorac Cardiovasc Surg 72:867,1976. 22. Kratz JM, Crawford FA Jr, Stroud MR, et al: Trends and results in tricuspid valve surgery. Chest 88:837,1985. 23. Cohen SR, Sell JE, Mclntosh CL, et al: Tricuspid regurgitation in patients with acquired, chronic, pure mitral regurgitation: Nonoperative management, tricuspid valve annuloplasty, and tricuspid valve replacement. J Thorac Cardiovasc Surg 94:488, 1987. 24. Grondin P, Meere C, Limet R, et al: Carpentier’s annulus and DeVega’s annuloplasty: The end of the tricuspid challenge. J Thorac Cardiovasc Surg 70:852,1975. 25. Cohen SR, Silver MA, Mclntosh CL, Roberts WC: Comparison of late (62-1 40 months) degenerative changes in simultaneously implanted and explanted porcine (Hancock) bioprostheses in the tricuspid and mitral valve positions in six patients. Am J Cardiol53:1599, 1984. 26. Spray TI, Roberts WC: Structural changes in porcine xenografts used as substitute cardiac valves. Am J Cardiol40:319, 1977. 27. Barratt-Boyes BG, Rutherford JD, Whitlock RML, et al: A review of surgery for acquired tricuspid valve disease, including an assessment of the stented semilunar homograft valve, and the results of operation for multivalvular heart disease. Aust N Z J Surg 58:23,1988. 28. Stephenson LW, Kouchoukos NT, Kirklin JW: Triple-valve replacement: An analysis of eight years’ experience. Ann Thorac Surg 23:327, 1977. 29. Macmanus Q, Grunkemeier G, Starr A: Late results of triple valve replacement: A 14-year review. Ann Thorac Surg 25:402, 1978. 30. Kara M, Langlet MF, Blin D, et al: Triple valve procedures: An analysis of early and late results. Thorac Cardiovasc Surg 34: 17,1986. 31. Baughman KL, Kallman CH, Yurchak PM, et al: Predictors of survival after tricuspid valve surgery. Am J Cardiol54:137, 1984.
Journal of Cardiac !Surgery
253
