- Email: [email protected]
Five-Year Clinical Experience with the Omniscience Cardiac Valve
Five-Year Clinical Experience with the Omniscience Cardiac Valve Richard DeWall, M.D., L. Conrad Pelletier, M.D., Antonio Panebianco, M.D., George Hicks, M.D., Benjamin Schuster, M.D., Raoul Bonan, M.D., Jean-Paul Martineau, M.D., and Luke Yip, M.D. ABSTRACT Clinical data on the Omniscience cardiac valve prosthesis (sizes 19 to 31 mm) were obtained from 326 patients (155,aortic valve replacement [AVR]; 125, mitral valve replacement [MVRI; and 46, double-valve replacement) during a five-year period (650 patient-years) with 96% accountability. Mean age was 56 f 12 years, and 40% (130)of the patients were 61 years old or older. Thirtyone percent (101)had prior or concomitant coronary artery bypass grafting procedures, 9% (28) had a previous malfunctioning prosthesis, and 17% (57)had other cardiac surgical procedures. Eighty percent were in New York Heart Association (NYHA) Functional Class I11 or IV preoperatively. Early mortality was 10% (34). Five-year actuarial thrombus-free rate is 96% for patients who underwent AVR and 95% for patients having MVR. Postoperatively, 89% (172/193)were in NYHA Class I or 11, and 84% (163/ 193) improved by at least one Functional Class. A significantly high proportion of the postoperative hematological values fall within normal ranges: red blood cell count, 94% (176/187);hematocrit, 88% (166/188);and hemoglobin, 93% (176/190).Ninety-eight percent (287/292)are free from clinical anemia. Four of the 5 patients with clinical anemia had a preoperative history of this condition. Levels of lactic dehydrogenase in 3 patients suggested the probable presence of clinically significant hemolysis, although all 3 have normal hematological values. Actuarial five-year survival for patients who underwent AVR or MVR are similar (82% and 80%, respectively), indicating satisfactory and comparable levels of safety and performance. Based on the duration of the study and the absence of restrictive selection criteria, which would eliminate high-risk patients, these clinical data demonstrate a reliable degree of safety and performance for this prosthesis. Since the successful implantation of prosthetic aortic and mitral valves by Harken and associates [l]and Starr and Edwards [2], respectively, more than twenty years ago, many variations of valve prosthesis design have been From the Kettering Medical Center, Dayton, OH, Institut de Cardiologie de Montreal, Montreal, PQ, Canada, and Lehigh Valley Hospital, Allentown, PA. Presented at the Twentieth Annual Meeting of The Society of Thoracic Surgeons, San Antonio, TX, Jan 23-25, 1984. Address reprint requests to Dr. DeWaU, 421 ThomhiU Rd, Dayton, OH 45419.
275
developed. The first Omniscience cardiac valve prostheses were implanted at Kettering Medical Center in Dayton, OH, in August, 1978. The patient underwent both aortic valve replacement (AVR) and mitral valve replacement (MVR) and is still doing well. This paper summarizes the first five years of experience of three North American medical centers with the Omniscience cardiac valve prosthesis implanted in patients undergoing AVR, MVR, or double-valve replacement (AVR MVR).
+
Material and Methods Between August, 1978, and June, 1983,326 consecutive patients underwent cardiac valve replacement with an Omniscience prosthesis at Kettering Medical Center, Lehigh Valley Hospital in Allentown, PA, and Institut de Cardiologie de Montreal in Quebec. There were 155 male patients (48%)and 171 female patients (52%).The mean age was 56 ? 12 years (range, 17 to 79 years). One hundred fifty-five (48%)underwent AVR, 125 (38%)had MVR, and 46 (14%) had double-valve replacement. Valve replacement was carried out using standard techniques. Special care was given to selection of valve size as well as valve clearance and orientation. Thirtyone percent (101) of the patients underwent prior or concomitant coronary artery bypass grafting procedures, 9% (28) had replacement of a previously implanted malfunctioning prosthesis, and 17% (57) had other cardiac surgical procedures. Only 43% of the patients (140) were free from any prior or concomitant cardiac surgical procedures (Fig 1). Warfarin anticoagulation alone was prescribed for 94% of the patients, and warfarin plus antiplatelet therapy was prescribed for 4%. Neither aspirin alone nor antiplatelet therapy alone was prescribed for any patient. Current patient follow-up was carried out through office visits and telephone contact with patients and referring physicians. Only 12 patients were lost to followup, resulting in a patient accountability of 96%. Mean follow-up was 28 2 14 months (range, 2 to 60 months). Actuarial rates of patient survival were calculated for the overall patient population as well as for patients having isolated AVR and MVR [3]. The probability of remaining thromboembolism free was also determined by actuarial method for the patients who had isolated AVR and MVR [3]. Patient functional status was determined according to New York Heart Association (NYHA) Classification criteria. Preoperatively, 27% of patients were in NYHA Class IV,53% in Class 111,17%in Class 11,
276 The Annals of Thoracic Surgery Vol 38 No 3 September 1984
Causes of 42 Late Postoperative Deaths Cause Cardiac related /CARDIACRELATED PROCEDURES
/
ADDITIONAL CARDIAC SURQERY 57%
\
NO ADDITIONAL CARDIAC SURQERY
Arrhythmia Congestive heart failure Myocardial infarction Cardiac arrestlsudden death Embolus Valve thrombosis Periprosthetic leak Pannus overgrowth Not cardiac related
No. of
Patients 34 10 8
1 8
1
2 3 1 8
Fig 1 . Associated surgical procedures. ( 0 s = Omniscience; CABG = coronary artery bypass grafting.)
and 3% in Class I. NYHA Classification was not available in hospital records of 4 surviving patients.
Results Thirty-four of the 326 patients (10.4%)did not survive the early postoperative period. Mortality ranged from 8.4%(13 patients) and 8.7%(4) for patients having AVR and double-yalve replacement, respectively, to 13.6% (17) for those having MVR. Sixty-five percent of these operative deaths occurred within forty-eight hours of operation. Cardiac causes of death included congestive heart failure (9 patients), low cardiac output syndrome (8), hemorrhage or cardiac tamponade or both (7), myocardial infarction (3), cardiac arrest (l),and arrhythmia (1). Noncardiac causes of death included renal failure (2 patients), respiratory arrest (l), pneumonia (l), and unknown (1). No death was determined to be valve related. Sixty-nine percent of the operative deaths occurred in patients in NYHA Class IV and 22% in patients in Class III.
Survival Two hundred ninety-two patients at risk (operative survivors) were followed for up to five years. Causes of the 42 late deaths are presented in the Table. Preoperatively, 4 of the patients who died late were in NYHA Class II,19 were in Class 111, 15 were in Class IV, and 4 were of unknown class. The overall actuarial survival probability was 80% at five years, with 82% and 80% for patients who underwent AVR and MVR, respectively (Fig 2). As demonstrated in Figure 3, late overall survival (AVR, MVR, and double-valve replacement) is influenced by preoperative NYHA Class. Actuarial survival at four years for patients known to be in NYHA Class I1 before operation was 93%, 85% for Class 111, and 60%for Class IV. The number of patients originally in NYHA Class I was too small to make a valid comparison. Patients aged 61 years or older and in NYHA Functional Class IV preoperatively were considered to be at high risk. This group comprised 11% (32/292) of the atrisk population analyzed for long-term survival, yet 26% (11142) of the late deaths occurred in this group.
Functional Class Figure 4 demonstrates the functional improvement that occurred after valve replacement, as reflected by changes in NYHA Classification. Of 193 patients whose preoperative and postoperative NYHA Class is known, 89% are now in Class I or 11. Eighty-four percent of all patients have improved by at least one Functional Class, 14% remain in the same Class, and only 2% have declined by one Class. Complications None of the complications that occurred was determined to be related to intrinsic valve failure or malfunction. THROMBOEMBOLIC EVENTS. Six transient ischemic episodes occurred during the late postoperative period (0.92%per patient-year), four in patients who had AVR and two in patients who had double-valve replacement. Eleven patients experienced serious thromboembolic events (valve thrombosis or thromboembolism with residual effects) during this period. Among patients who underwent AVR there were three instances of valve thrombosis, two of which were fatal, resulting in a 96% actuarial thrombus-free rate at five years (Fig 5). The five-year actuarial thrombus-free rate of 95% for patients who had MVR is computed on the basis of five nonfatal cases of valve thrombosis (see Fig 5). Among patients who had double-valve replacement, there were three thromboembolic events (two instances of valve thrombosis and one instance of thromboembolism with residual effects), and all were fatal. The patient who experienced thromboembolism was in atrial fibrillation and died of a probable cerebral embolus within hours of the placement of bilateral antral windows as treatment for severe sinusitis. In 2 of these patients, pannus overgrowth caused interference with the valve mechanism. Valve thrombosis occurred secondarily. In 6 of the 17 patients experiencing either a transient ischemic episode or a serious thromboembolic event, the anticoagulation therapy was compromised shortly before the occurrence of the complication. One of the 6 patients had a subtherapeutic prothrombin time and 2
277 DeWall et al: Five-Year Experience with Omniscience Cardiac Valve
j: <
o AVR
50
A
MVR
T
S.E.
0
patlonto a t rlah ontorlna Intorrol
I
-.--fowor thon 1 0 potlont.
2
>
5 cn
40
4
I
I I
0
1
I
I
I
2
I I
3
I
1
4
5
I
IMPLANT PERIOD (YEARS)
Fig 2 . Actuarial postoperative suruival for patients at risk undergoing aortic valve replacement (AVR) and mitral valve replacement (MVR). (S.E. = standard error.)
(1600) (610)
'1
3 40
cn
PREOP N V W A
1
T
I
111
a
(
p o t l o n t n o t r1.k
._ .. f0w.r
AIV
ontorlna Intorval
than ton pntlOnt.
/r
I
I
I
I
0
1
2
3
4
IMPLANT PERIOD (YEARS)
Fig 3 . Overall survival by preoperative New York Heart Association (NYHA) Functional Class. (S.E.= standard error.)
had been removed from anticoagulant therapy, 1 because of gastrointestinal bleeding and the other because of hypermenorrhagia. For the other 3 patients, a temporary interruption of anticoagulation was necessary to perform dental work (1 patient), a right carotid endartarectomy (l),and leg amputation (1). ANTICOAGULATION. One nonfatal and three fatal hemorrhages due to anticoagulant therapy occurred. Hence, the overall incidence was 0.2% per patient-year.
PERIPROSTHETIC LEAK. Fourteen episodes of periprosthetic le,ak occurred in 12 patients in this series, 6 who had AVR, 4 who had MVR, and 2 who had double-valve replacement. In both of the last patients, the leak occurred in the mitral position. Three patients had bacterial endocarditis, and 1patient had Marfan's syndrome. In 2 patients, stitches were placed in tissue of very poor quality; there was little or no subsequent healing. Eight of these 12 patients were successfully reoperated on, 2
278 The Annals of Thoracic Surgery Vol 38 No 3 September 1984
70
died after operation, and 2 died with no operation. For 1 of these patients, although periprosthetic leak was found at autopsy, death was due to congestive heart failure. The overall incidence of this event is 2.2% per patient-year.
n=193 potionta
60
s
50
Blood Analyses An assessment of available current postoperative hematological values demonstrated that a high proportion exceeded the lower normal levels: 88% (166/188) of the hematocrit values, 93% (176/190) of the hemoglobin values, and 94% (176/187) of the red blood cell counts. Ninety-eight percent of all at-risk patients were free from clinical anemia. Of the 5 patients treated postoperatively for this condition, 4 had a preoperative history of the anemia. The presence and severity of intravascular hemolysis is often gauged by serum levels of lactic dehydrogenase (LDH). A level of only two to three times the normal LDH level rules out the presence of high-grade intravascular hemolysis [4]. Utilizing this criterion, only 3 patients experienced LDH levels above this range. However, their hematocrit, hemoglobin, and red blood cell values were higher than the lower normal limits. These results indicate that although mild intravascular hemolysis may be occurring in some patients, severe anemia does not result.
I-
w z
u
40 V
z W
a
0
w
if
30
20
10
0
It
Ill I V
OPERATIVE
Fig 4 . Contrast between preoperative and postoperative New York Heart Association Functional Classifications.
Comment The Omniscience cardiac valve has provided satisfactory clinical results during this five-year study. Operative mortality more accurately reflects the patient's preoperative condition and other surgical factors than valverelated variables. None of the early deaths was valve related. Late death, although influenced by the design and the resultant safety and performance factors of a
Fig 5. Actuarial thrombus-free rate for patients at risk undergoing aortic valve replacement (AVR) and mitral valve replacement (MVR). 7'hromboembolic events included thromboembolism with residual effect and valve thrombosis. (S.E. = standard error.)
* soL: d a 4
m
z
-
0 AVR
A
MVR T S.E.
n w 80W
a
1L
m
....(owor
-
than 10 patlonta Mk mtoring Int.rvd
( 1 patlmta at
IMPLANT PERIOD (YEARS)
279 DeWall et al: Five-Year Experience with Omniscience Cardiac Valve
prosthetic valve, is still a function of patient-related factors. None of the late deaths was due to mechanical or structural failure of the valve. The five-year survival (82% and 80% for AVR and MVR patients, respectively) of the Omniscience valve compares favorably with that reported for all types of prosthetic valves, including bioprostheses. In these other studies, five-year actuarial survival for patients who underwent AVR ranged from 63 to 84%and from 66 to 84% for patients who had MVR [5-91. Preoperative NYHA Functional Class represents a patient-related factor that carries a strong influence on late postoperative survival (see Fig 3). As expected, an increased risk for late mortality was apparent in older patients (61 years or older) who were in NYHA Class IV preoperatively. Williams and associates [101 reported 36month survival for both Hancock and CarpentierEdwards bioprostheses implanted in patients in NYHA Class I1 and Class I11 (94% and 87%, respectively). This survival is comparable to that of the Omniscience valve. For patients in a more advanced disease stage (NYHA Class IV), the three-year survival for the Omniscience valve is distinctly higher (70%)than that reported for the Hancock and Carpentier-Edwards valves (59%). The low rate of transient ischemic episodes (0.92%per patient-year) and the observed five-year actuarial thrombus-free rates of 96% and 95% for AVR and MVR patients, respectively, compare favorably with rates for other prostheses, including both mechanical valves and bioprostheses [4, 10-14al. It is important to note that the excellent five-year thrombus-free rate of 95%for patients who had MVR contrasts sharply with the high rate of mitral valve thrombosis reported for the Omniscience prosthesis by Fananapazir and colleagues [W] in 1983. This difference may be due in large part to the combination of posterior orientation and large valve sizes employed by these investigators, which may increase the probability of disc interference by tissue remnants or the ventricular wall itself. This combination of large mitral valve size and posterior orientation has not been employed at any of the three centers in the present study. Other patient-related factors such as compromise of anticoagulant therapy [16-19], history of atrial fibrillation [18, 191, and history of rheumatic heart disease 120-231 may have been involved also. In addition, Fananapazir and associates [W] reported a high incidence of perivalvular leak. The use of 2-0 sutures in a continuous suture technique was identified. As Dhasmana and colleagues (241 pointed out, a statistically significant increase in the development of perivalvular leaks occurs when this combination is utilized. In addition, Omniscience valves were implanted in 8% of their patients after infective endocarditis or in replacement of other malfunctioning prostheses. Both categories of patients are prone to perivalvular Ieaks. The quality of life for the patients still alive has been improved dramatically: 89% are postoperatively in NYHA Class I or I1 (see Fig 4). The data reported here, coupled with low incidences for other valve-related com-
plications, indicate a high degree of performance and safety of the Omniscience cardiac valve prosthesis.
References 1. Harken DE, Soroff HS, Taylor WJ, et al: Partial and complete prosthesis in aortic insufficiency. J Thorac Cardiovasc Surg 40:744, 1960 2. Starr A, Edwards ML: Mitral replacement: clinical experience with a ball-valve prosthesis. Ann Surg 1W726, 1961 3. Grunkemeier GL, Starr A Actuarial analysis of surgical results: rationale and method. Ann Thorac Surg 24:404,1977 4. Lefrak EA, Starr A Cardiac Valve Prostheses. New York, Appleton-Century-Crofts, 1979 5. McHenry MM, Smeluff EA, Mathof HJ, et al: Long-term results after single aortic and mitral valve replacement with the present model of Smeloff-Cutter valves. J Thorac Cardiovasc Surg 75:704, 1978 6. BlackstoneEH, Kirklin JW, Pluth JR, et al: The performance of the Braunwald-Cutter aortic prosthetic valve. Ann Thorac Surg 23:302, 1977 7. Lepley D Jr, FIemma RJ, Mullen DC, et al: Long-term follow-up of the Bjork-Shiley prosthetic valve used in the mitral position. Ann Thorac Surg 30164, 1980 8. Bjork VO, Henze A: Ten years’ experience with the BiorkShiley tilting disc valve. J Thorac Cardiovasc Surg 7&331, 1979 9. Angel1 NW: A nine-year experience with the Angell-Shiley xenograft and a comparative literature review of the porcine bioprosthesis versus the mechanical prosthesis. In Proceedings of the Symposium on Bioprosthetic Cardiac Valves, Munich, April 5-7, 1979, pp 81-104 10. Williams JB, Karp RB, Kirklin JW, et al: Considerations in selection and management of patients undergoing valve replacement with glutaraldehyde-fixed porcine bioprostheses. Ann Thorac Surg 30247, 1980 11. Macmanus Q, Grunkemeier GL, Lambert LE, et a1 Year of operation as a risk factor in the late results of valve replacement. J Thorac Cardiovasc Surg 80834, 1980 12. Mattingly WT Jr, OConnor W, Zeok JV, Todd EP: Thrombotic catastrophe in the patient with multiple Bjork-Shiley prostheses. Ann Thorac Surg 35:253, 1983 13. Horstkotte D, Korfer R, Seipel L, et al: Late complications in patients with Bjork-Shiley and St. Jude Medical heart valve replacement. Circulation 68Suppl2175, 1983 14. Rossiter SJ, Miller DG, Stinson EB, et al: Hemodynamicand clinical comparison of the Hancock modified orifice and standard orifice bioprostheses in the aortic position. J Thorac Cardiovasc Surp. 80% 1980 14a. Starr DS, Lawrie GM: Howell JF, M o m s GC Jr: Clinical experience with the Smeloff-Cutterprosthesis: 1- to 12-year follow-up. Ann Thorac Surg 30:448, 1980 15. Fananapazir L, Clarke DB, Dark JF, et al: Results of valve replacement with the Omniscience prusthesis. J Thorac Cardiovasc Surg 86:621, 1983 16. Edmunds LH Jr: Thromboembolic complications of current cardiac valvular prostheses (collective review). Ann Thorac Surg 34:96, 1982 17. Katholi RE, Nolan SP, McGuire LB: Living with prosthetic heart valves: subsequent non-cardiac operations and the risk of thromboembolism or hemorrhage. Am Heart J 92:162, 1976 18. Farah E, Cadhilac M, Boubaker A, et al: Thromboembolic risk in prosthetic valve carriers: prevention trial. In Throm-
280 The Annals of Thoracic Surgery Vol 38 No 3 September 1984
19. 20. 21. 22. 23.
24.
boembolism and Prosthetic Valves. Scand J Haematol [Suppl] 38:97, 1981 Pumphrey CW, Fuster V, Chesebro JH: Systemic thromboembolism in valvular heart disease and prosthetic heart valves. Mod Concepts Cardiovasc Dis 51:131, 1982 Steele PP, Weily HS, Davies H, Genton E: Platelet survival in patients with rheumatic heart disease. N Engl J Med 290:537, 1974 Steele PP, Weily HS, Rainwater J, Vogel R Platelet survival time and thromboembolism in patients with mitral valve prolapse. Circulation 60:42, 1979 Steele PP, Rainwater J: Favorable effect of sulfinpyrazone on thromboembolism in patients with rheumatic heart disease. Circulation 62462, 1980 Baggenstoss AH, Titus JL: Rheumatic and collagen disorders of the heart. In Gould SE (ed): Pathology of the Heart and Blood Vessels. Third edition. Springfield, IL, Thomas, 1968 Dhasmana JP, Blackstone EH, Kirklin JW, Kouchoukos NT: Factors associated with periprosthetic leakage following primary mitral valve replacement: with special consideration of the suture technique. Ann Thorac Surg 35:170, 1983
Discussion DR. DONALD J . MAGILLIGAN (Detroit, MI): I appreciate the oppor-
tunity to review the experience of Dr. DeWall, one of the true innovators in thoracic surgery, regarding his five-year experience with the Omniscience cardiac valve. With each new valve, careful follow-up is necessary to determine whether the prosthesis represents a technological breakthrough or just another valve. Dr. DeWall has shown that patient survival and improvement in NYHA Functional Class are as good with this valve as with other prostheses. But survival and improvement in symptoms are gross variables and have been too imprecise to distinguish one valve from another. In the absence of hemodynamic data, we must focus on durability and thromboembolism to determine the effectiveness of this valve. With respect to durability, the Omniscience valve has not broken or disrupted in this group of 326 patients with a mean follow-up of 28 months. The results with respect to thromboembolismare important. I compared the results that Dr. DeWall presented today with
those obtained by Marshall and associateswith the Bjork-Shiley valve (J Thorac Cardiovasc Surg 85:902, 1983). The Bjork-Shiley group comprised 357 patients followed for a mean of 46 months. The Omniscience group had 326 patients followed for a mean of 28 months. The incidence of thromboembolism was 23%with the Bjork-Shiley valve and only 6%with the Omniscience valve. However, it must be remembered that these patients were followed an average of 2.3 years only. The incidence of valve thrombosis was 3% for the BjorkShiley valve and 4% for the Omniscience valve. The ratio of thrombosis to total thromboembolic events was 13% for the Bjork-Shiley valve and an alarming 65% for the Omniscience valve. Thromboembolism is a linear-related event, and as the follow-up on the Omniscience patients approaches the 46 months of the Bjork-Shiley study, it can be anticipated that more thromboembolic events will occur. Dr. DeWall, can you give us any assurance that as these patients are followed longer the ratio of valve thrombosis to total thromboembolic events will decline from this clearly unacceptable level? I commend Dr. DeWall for a very honest study and thank the Society for inviting me to discuss it. In response to Dr. Magilligan, hemodynamic data were not a subject of this paper but are part of another communication. I can only testify that the hemodynamic data of the Omniscience valve are competitive with other types of valves. Relative to thrombotic problems in prosthetic valves, it is increasingly apparent from the literature on valves that orientation is of importance with all types of prosthetic valves. I have gathered data from seventeen centers that use the Omniscience valve and correlated the incidence of problems with valve size and orientation in the mitral position. Anterior orientation of the larger orifice is virtually free from complications at all centers, even with the use of larger sized prostheses. Posterior orientation has a small incidence of problems that is comparable with other valve types, except in one center that uses an average valve size of 30 mm. Researchers there experienced an unusually high incidence of one problem, that is, disc hang-up postoperatively subsequent to cardiac shrinkage after implantation relieves the cause of cardiac hypertrophy or dilatation. This practice is most unusual considering the admonitions in the valve surgical literature for more than twenty years against the use of oversized valves. DR. DE WALL:
275
developed. The first Omniscience cardiac valve prostheses were implanted at Kettering Medical Center in Dayton, OH, in August, 1978. The patient underwent both aortic valve replacement (AVR) and mitral valve replacement (MVR) and is still doing well. This paper summarizes the first five years of experience of three North American medical centers with the Omniscience cardiac valve prosthesis implanted in patients undergoing AVR, MVR, or double-valve replacement (AVR MVR).
+
Material and Methods Between August, 1978, and June, 1983,326 consecutive patients underwent cardiac valve replacement with an Omniscience prosthesis at Kettering Medical Center, Lehigh Valley Hospital in Allentown, PA, and Institut de Cardiologie de Montreal in Quebec. There were 155 male patients (48%)and 171 female patients (52%).The mean age was 56 ? 12 years (range, 17 to 79 years). One hundred fifty-five (48%)underwent AVR, 125 (38%)had MVR, and 46 (14%) had double-valve replacement. Valve replacement was carried out using standard techniques. Special care was given to selection of valve size as well as valve clearance and orientation. Thirtyone percent (101) of the patients underwent prior or concomitant coronary artery bypass grafting procedures, 9% (28) had replacement of a previously implanted malfunctioning prosthesis, and 17% (57) had other cardiac surgical procedures. Only 43% of the patients (140) were free from any prior or concomitant cardiac surgical procedures (Fig 1). Warfarin anticoagulation alone was prescribed for 94% of the patients, and warfarin plus antiplatelet therapy was prescribed for 4%. Neither aspirin alone nor antiplatelet therapy alone was prescribed for any patient. Current patient follow-up was carried out through office visits and telephone contact with patients and referring physicians. Only 12 patients were lost to followup, resulting in a patient accountability of 96%. Mean follow-up was 28 2 14 months (range, 2 to 60 months). Actuarial rates of patient survival were calculated for the overall patient population as well as for patients having isolated AVR and MVR [3]. The probability of remaining thromboembolism free was also determined by actuarial method for the patients who had isolated AVR and MVR [3]. Patient functional status was determined according to New York Heart Association (NYHA) Classification criteria. Preoperatively, 27% of patients were in NYHA Class IV,53% in Class 111,17%in Class 11,
276 The Annals of Thoracic Surgery Vol 38 No 3 September 1984
Causes of 42 Late Postoperative Deaths Cause Cardiac related /CARDIACRELATED PROCEDURES
/
ADDITIONAL CARDIAC SURQERY 57%
\
NO ADDITIONAL CARDIAC SURQERY
Arrhythmia Congestive heart failure Myocardial infarction Cardiac arrestlsudden death Embolus Valve thrombosis Periprosthetic leak Pannus overgrowth Not cardiac related
No. of
Patients 34 10 8
1 8
1
2 3 1 8
Fig 1 . Associated surgical procedures. ( 0 s = Omniscience; CABG = coronary artery bypass grafting.)
and 3% in Class I. NYHA Classification was not available in hospital records of 4 surviving patients.
Results Thirty-four of the 326 patients (10.4%)did not survive the early postoperative period. Mortality ranged from 8.4%(13 patients) and 8.7%(4) for patients having AVR and double-yalve replacement, respectively, to 13.6% (17) for those having MVR. Sixty-five percent of these operative deaths occurred within forty-eight hours of operation. Cardiac causes of death included congestive heart failure (9 patients), low cardiac output syndrome (8), hemorrhage or cardiac tamponade or both (7), myocardial infarction (3), cardiac arrest (l),and arrhythmia (1). Noncardiac causes of death included renal failure (2 patients), respiratory arrest (l), pneumonia (l), and unknown (1). No death was determined to be valve related. Sixty-nine percent of the operative deaths occurred in patients in NYHA Class IV and 22% in patients in Class III.
Survival Two hundred ninety-two patients at risk (operative survivors) were followed for up to five years. Causes of the 42 late deaths are presented in the Table. Preoperatively, 4 of the patients who died late were in NYHA Class II,19 were in Class 111, 15 were in Class IV, and 4 were of unknown class. The overall actuarial survival probability was 80% at five years, with 82% and 80% for patients who underwent AVR and MVR, respectively (Fig 2). As demonstrated in Figure 3, late overall survival (AVR, MVR, and double-valve replacement) is influenced by preoperative NYHA Class. Actuarial survival at four years for patients known to be in NYHA Class I1 before operation was 93%, 85% for Class 111, and 60%for Class IV. The number of patients originally in NYHA Class I was too small to make a valid comparison. Patients aged 61 years or older and in NYHA Functional Class IV preoperatively were considered to be at high risk. This group comprised 11% (32/292) of the atrisk population analyzed for long-term survival, yet 26% (11142) of the late deaths occurred in this group.
Functional Class Figure 4 demonstrates the functional improvement that occurred after valve replacement, as reflected by changes in NYHA Classification. Of 193 patients whose preoperative and postoperative NYHA Class is known, 89% are now in Class I or 11. Eighty-four percent of all patients have improved by at least one Functional Class, 14% remain in the same Class, and only 2% have declined by one Class. Complications None of the complications that occurred was determined to be related to intrinsic valve failure or malfunction. THROMBOEMBOLIC EVENTS. Six transient ischemic episodes occurred during the late postoperative period (0.92%per patient-year), four in patients who had AVR and two in patients who had double-valve replacement. Eleven patients experienced serious thromboembolic events (valve thrombosis or thromboembolism with residual effects) during this period. Among patients who underwent AVR there were three instances of valve thrombosis, two of which were fatal, resulting in a 96% actuarial thrombus-free rate at five years (Fig 5). The five-year actuarial thrombus-free rate of 95% for patients who had MVR is computed on the basis of five nonfatal cases of valve thrombosis (see Fig 5). Among patients who had double-valve replacement, there were three thromboembolic events (two instances of valve thrombosis and one instance of thromboembolism with residual effects), and all were fatal. The patient who experienced thromboembolism was in atrial fibrillation and died of a probable cerebral embolus within hours of the placement of bilateral antral windows as treatment for severe sinusitis. In 2 of these patients, pannus overgrowth caused interference with the valve mechanism. Valve thrombosis occurred secondarily. In 6 of the 17 patients experiencing either a transient ischemic episode or a serious thromboembolic event, the anticoagulation therapy was compromised shortly before the occurrence of the complication. One of the 6 patients had a subtherapeutic prothrombin time and 2
277 DeWall et al: Five-Year Experience with Omniscience Cardiac Valve
j: <
o AVR
50
A
MVR
T
S.E.
0
patlonto a t rlah ontorlna Intorrol
I
-.--fowor thon 1 0 potlont.
2
>
5 cn
40
4
I
I I
0
1
I
I
I
2
I I
3
I
1
4
5
I
IMPLANT PERIOD (YEARS)
Fig 2 . Actuarial postoperative suruival for patients at risk undergoing aortic valve replacement (AVR) and mitral valve replacement (MVR). (S.E. = standard error.)
(1600) (610)
'1
3 40
cn
PREOP N V W A
1
T
I
111
a
(
p o t l o n t n o t r1.k
._ .. f0w.r
AIV
ontorlna Intorval
than ton pntlOnt.
/r
I
I
I
I
0
1
2
3
4
IMPLANT PERIOD (YEARS)
Fig 3 . Overall survival by preoperative New York Heart Association (NYHA) Functional Class. (S.E.= standard error.)
had been removed from anticoagulant therapy, 1 because of gastrointestinal bleeding and the other because of hypermenorrhagia. For the other 3 patients, a temporary interruption of anticoagulation was necessary to perform dental work (1 patient), a right carotid endartarectomy (l),and leg amputation (1). ANTICOAGULATION. One nonfatal and three fatal hemorrhages due to anticoagulant therapy occurred. Hence, the overall incidence was 0.2% per patient-year.
PERIPROSTHETIC LEAK. Fourteen episodes of periprosthetic le,ak occurred in 12 patients in this series, 6 who had AVR, 4 who had MVR, and 2 who had double-valve replacement. In both of the last patients, the leak occurred in the mitral position. Three patients had bacterial endocarditis, and 1patient had Marfan's syndrome. In 2 patients, stitches were placed in tissue of very poor quality; there was little or no subsequent healing. Eight of these 12 patients were successfully reoperated on, 2
278 The Annals of Thoracic Surgery Vol 38 No 3 September 1984
70
died after operation, and 2 died with no operation. For 1 of these patients, although periprosthetic leak was found at autopsy, death was due to congestive heart failure. The overall incidence of this event is 2.2% per patient-year.
n=193 potionta
60
s
50
Blood Analyses An assessment of available current postoperative hematological values demonstrated that a high proportion exceeded the lower normal levels: 88% (166/188) of the hematocrit values, 93% (176/190) of the hemoglobin values, and 94% (176/187) of the red blood cell counts. Ninety-eight percent of all at-risk patients were free from clinical anemia. Of the 5 patients treated postoperatively for this condition, 4 had a preoperative history of the anemia. The presence and severity of intravascular hemolysis is often gauged by serum levels of lactic dehydrogenase (LDH). A level of only two to three times the normal LDH level rules out the presence of high-grade intravascular hemolysis [4]. Utilizing this criterion, only 3 patients experienced LDH levels above this range. However, their hematocrit, hemoglobin, and red blood cell values were higher than the lower normal limits. These results indicate that although mild intravascular hemolysis may be occurring in some patients, severe anemia does not result.
I-
w z
u
40 V
z W
a
0
w
if
30
20
10
0
It
Ill I V
OPERATIVE
Fig 4 . Contrast between preoperative and postoperative New York Heart Association Functional Classifications.
Comment The Omniscience cardiac valve has provided satisfactory clinical results during this five-year study. Operative mortality more accurately reflects the patient's preoperative condition and other surgical factors than valverelated variables. None of the early deaths was valve related. Late death, although influenced by the design and the resultant safety and performance factors of a
Fig 5. Actuarial thrombus-free rate for patients at risk undergoing aortic valve replacement (AVR) and mitral valve replacement (MVR). 7'hromboembolic events included thromboembolism with residual effect and valve thrombosis. (S.E. = standard error.)
* soL: d a 4
m
z
-
0 AVR
A
MVR T S.E.
n w 80W
a
1L
m
....(owor
-
than 10 patlonta Mk mtoring Int.rvd
( 1 patlmta at
IMPLANT PERIOD (YEARS)
279 DeWall et al: Five-Year Experience with Omniscience Cardiac Valve
prosthetic valve, is still a function of patient-related factors. None of the late deaths was due to mechanical or structural failure of the valve. The five-year survival (82% and 80% for AVR and MVR patients, respectively) of the Omniscience valve compares favorably with that reported for all types of prosthetic valves, including bioprostheses. In these other studies, five-year actuarial survival for patients who underwent AVR ranged from 63 to 84%and from 66 to 84% for patients who had MVR [5-91. Preoperative NYHA Functional Class represents a patient-related factor that carries a strong influence on late postoperative survival (see Fig 3). As expected, an increased risk for late mortality was apparent in older patients (61 years or older) who were in NYHA Class IV preoperatively. Williams and associates [101 reported 36month survival for both Hancock and CarpentierEdwards bioprostheses implanted in patients in NYHA Class I1 and Class I11 (94% and 87%, respectively). This survival is comparable to that of the Omniscience valve. For patients in a more advanced disease stage (NYHA Class IV), the three-year survival for the Omniscience valve is distinctly higher (70%)than that reported for the Hancock and Carpentier-Edwards valves (59%). The low rate of transient ischemic episodes (0.92%per patient-year) and the observed five-year actuarial thrombus-free rates of 96% and 95% for AVR and MVR patients, respectively, compare favorably with rates for other prostheses, including both mechanical valves and bioprostheses [4, 10-14al. It is important to note that the excellent five-year thrombus-free rate of 95%for patients who had MVR contrasts sharply with the high rate of mitral valve thrombosis reported for the Omniscience prosthesis by Fananapazir and colleagues [W] in 1983. This difference may be due in large part to the combination of posterior orientation and large valve sizes employed by these investigators, which may increase the probability of disc interference by tissue remnants or the ventricular wall itself. This combination of large mitral valve size and posterior orientation has not been employed at any of the three centers in the present study. Other patient-related factors such as compromise of anticoagulant therapy [16-19], history of atrial fibrillation [18, 191, and history of rheumatic heart disease 120-231 may have been involved also. In addition, Fananapazir and associates [W] reported a high incidence of perivalvular leak. The use of 2-0 sutures in a continuous suture technique was identified. As Dhasmana and colleagues (241 pointed out, a statistically significant increase in the development of perivalvular leaks occurs when this combination is utilized. In addition, Omniscience valves were implanted in 8% of their patients after infective endocarditis or in replacement of other malfunctioning prostheses. Both categories of patients are prone to perivalvular Ieaks. The quality of life for the patients still alive has been improved dramatically: 89% are postoperatively in NYHA Class I or I1 (see Fig 4). The data reported here, coupled with low incidences for other valve-related com-
plications, indicate a high degree of performance and safety of the Omniscience cardiac valve prosthesis.
References 1. Harken DE, Soroff HS, Taylor WJ, et al: Partial and complete prosthesis in aortic insufficiency. J Thorac Cardiovasc Surg 40:744, 1960 2. Starr A, Edwards ML: Mitral replacement: clinical experience with a ball-valve prosthesis. Ann Surg 1W726, 1961 3. Grunkemeier GL, Starr A Actuarial analysis of surgical results: rationale and method. Ann Thorac Surg 24:404,1977 4. Lefrak EA, Starr A Cardiac Valve Prostheses. New York, Appleton-Century-Crofts, 1979 5. McHenry MM, Smeluff EA, Mathof HJ, et al: Long-term results after single aortic and mitral valve replacement with the present model of Smeloff-Cutter valves. J Thorac Cardiovasc Surg 75:704, 1978 6. BlackstoneEH, Kirklin JW, Pluth JR, et al: The performance of the Braunwald-Cutter aortic prosthetic valve. Ann Thorac Surg 23:302, 1977 7. Lepley D Jr, FIemma RJ, Mullen DC, et al: Long-term follow-up of the Bjork-Shiley prosthetic valve used in the mitral position. Ann Thorac Surg 30164, 1980 8. Bjork VO, Henze A: Ten years’ experience with the BiorkShiley tilting disc valve. J Thorac Cardiovasc Surg 7&331, 1979 9. Angel1 NW: A nine-year experience with the Angell-Shiley xenograft and a comparative literature review of the porcine bioprosthesis versus the mechanical prosthesis. In Proceedings of the Symposium on Bioprosthetic Cardiac Valves, Munich, April 5-7, 1979, pp 81-104 10. Williams JB, Karp RB, Kirklin JW, et al: Considerations in selection and management of patients undergoing valve replacement with glutaraldehyde-fixed porcine bioprostheses. Ann Thorac Surg 30247, 1980 11. Macmanus Q, Grunkemeier GL, Lambert LE, et a1 Year of operation as a risk factor in the late results of valve replacement. J Thorac Cardiovasc Surg 80834, 1980 12. Mattingly WT Jr, OConnor W, Zeok JV, Todd EP: Thrombotic catastrophe in the patient with multiple Bjork-Shiley prostheses. Ann Thorac Surg 35:253, 1983 13. Horstkotte D, Korfer R, Seipel L, et al: Late complications in patients with Bjork-Shiley and St. Jude Medical heart valve replacement. Circulation 68Suppl2175, 1983 14. Rossiter SJ, Miller DG, Stinson EB, et al: Hemodynamicand clinical comparison of the Hancock modified orifice and standard orifice bioprostheses in the aortic position. J Thorac Cardiovasc Surp. 80% 1980 14a. Starr DS, Lawrie GM: Howell JF, M o m s GC Jr: Clinical experience with the Smeloff-Cutterprosthesis: 1- to 12-year follow-up. Ann Thorac Surg 30:448, 1980 15. Fananapazir L, Clarke DB, Dark JF, et al: Results of valve replacement with the Omniscience prusthesis. J Thorac Cardiovasc Surg 86:621, 1983 16. Edmunds LH Jr: Thromboembolic complications of current cardiac valvular prostheses (collective review). Ann Thorac Surg 34:96, 1982 17. Katholi RE, Nolan SP, McGuire LB: Living with prosthetic heart valves: subsequent non-cardiac operations and the risk of thromboembolism or hemorrhage. Am Heart J 92:162, 1976 18. Farah E, Cadhilac M, Boubaker A, et al: Thromboembolic risk in prosthetic valve carriers: prevention trial. In Throm-
280 The Annals of Thoracic Surgery Vol 38 No 3 September 1984
19. 20. 21. 22. 23.
24.
boembolism and Prosthetic Valves. Scand J Haematol [Suppl] 38:97, 1981 Pumphrey CW, Fuster V, Chesebro JH: Systemic thromboembolism in valvular heart disease and prosthetic heart valves. Mod Concepts Cardiovasc Dis 51:131, 1982 Steele PP, Weily HS, Davies H, Genton E: Platelet survival in patients with rheumatic heart disease. N Engl J Med 290:537, 1974 Steele PP, Weily HS, Rainwater J, Vogel R Platelet survival time and thromboembolism in patients with mitral valve prolapse. Circulation 60:42, 1979 Steele PP, Rainwater J: Favorable effect of sulfinpyrazone on thromboembolism in patients with rheumatic heart disease. Circulation 62462, 1980 Baggenstoss AH, Titus JL: Rheumatic and collagen disorders of the heart. In Gould SE (ed): Pathology of the Heart and Blood Vessels. Third edition. Springfield, IL, Thomas, 1968 Dhasmana JP, Blackstone EH, Kirklin JW, Kouchoukos NT: Factors associated with periprosthetic leakage following primary mitral valve replacement: with special consideration of the suture technique. Ann Thorac Surg 35:170, 1983
Discussion DR. DONALD J . MAGILLIGAN (Detroit, MI): I appreciate the oppor-
tunity to review the experience of Dr. DeWall, one of the true innovators in thoracic surgery, regarding his five-year experience with the Omniscience cardiac valve. With each new valve, careful follow-up is necessary to determine whether the prosthesis represents a technological breakthrough or just another valve. Dr. DeWall has shown that patient survival and improvement in NYHA Functional Class are as good with this valve as with other prostheses. But survival and improvement in symptoms are gross variables and have been too imprecise to distinguish one valve from another. In the absence of hemodynamic data, we must focus on durability and thromboembolism to determine the effectiveness of this valve. With respect to durability, the Omniscience valve has not broken or disrupted in this group of 326 patients with a mean follow-up of 28 months. The results with respect to thromboembolismare important. I compared the results that Dr. DeWall presented today with
those obtained by Marshall and associateswith the Bjork-Shiley valve (J Thorac Cardiovasc Surg 85:902, 1983). The Bjork-Shiley group comprised 357 patients followed for a mean of 46 months. The Omniscience group had 326 patients followed for a mean of 28 months. The incidence of thromboembolism was 23%with the Bjork-Shiley valve and only 6%with the Omniscience valve. However, it must be remembered that these patients were followed an average of 2.3 years only. The incidence of valve thrombosis was 3% for the BjorkShiley valve and 4% for the Omniscience valve. The ratio of thrombosis to total thromboembolic events was 13% for the Bjork-Shiley valve and an alarming 65% for the Omniscience valve. Thromboembolism is a linear-related event, and as the follow-up on the Omniscience patients approaches the 46 months of the Bjork-Shiley study, it can be anticipated that more thromboembolic events will occur. Dr. DeWall, can you give us any assurance that as these patients are followed longer the ratio of valve thrombosis to total thromboembolic events will decline from this clearly unacceptable level? I commend Dr. DeWall for a very honest study and thank the Society for inviting me to discuss it. In response to Dr. Magilligan, hemodynamic data were not a subject of this paper but are part of another communication. I can only testify that the hemodynamic data of the Omniscience valve are competitive with other types of valves. Relative to thrombotic problems in prosthetic valves, it is increasingly apparent from the literature on valves that orientation is of importance with all types of prosthetic valves. I have gathered data from seventeen centers that use the Omniscience valve and correlated the incidence of problems with valve size and orientation in the mitral position. Anterior orientation of the larger orifice is virtually free from complications at all centers, even with the use of larger sized prostheses. Posterior orientation has a small incidence of problems that is comparable with other valve types, except in one center that uses an average valve size of 30 mm. Researchers there experienced an unusually high incidence of one problem, that is, disc hang-up postoperatively subsequent to cardiac shrinkage after implantation relieves the cause of cardiac hypertrophy or dilatation. This practice is most unusual considering the admonitions in the valve surgical literature for more than twenty years against the use of oversized valves. DR. DE WALL: