Simultaneous implantation of St. Jude Medical aortic and mitral prostheses

J

THORAC CARDIOVASC SURG

1987;94:733-9

Simultaneous implantation of S1. Jude Medical aortic and mitral prostheses Since January 1980, 92 consecutive patients received St. Jude Medical aortic and mitral prostheses simultaneously. Mean age was 57.6 years (standard deviation 12.4~ 14 were 70 years or older. Twenty-threehad a previous cardiac operation and 22 had additional proceduresperformed at the time of double valve replacement.Before the operation 62 % of the patients were in New York Heart Association functional class ill and 29 % were in class IV or required emergency operation. There were six (6.5%) deaths within 30 days. None of the hospital deaths were valve related; all occurred in patients who had additional risk concerns. FoUow-up is 100% complete and ranges from 2 to 80 months, totaling 242 patient-years(mean 33.8 months). AU except four hospital survivors reached class I or n and 40 patients (47%) remain asymptomatic. The actuarial survival rates are 82% at 1 year,70% at 3 years, and 60% at 5 years. Causes of late death include heart failure (10),sudden, unexplained death (five), reoperation for coronary artery disease (one), noncardiac (four), and valve related (five). The linearized rate of fatal valve-related events is 2.1 % pt-yr, A total of 22 valve-related complications(including five fatal) occurred is 18 patients, for a linearized rate or incidence of 9.1 %jpt-yr. Eleven thromboembolic episodes (rate 4.6% jpt-yr) occurred in nine patients; three of these (1.2% jpt-yr) were fatal. Thromboembolic and bleeding complications represented 64 % of all valve-related complications. Four patients had six episodes of prosthetic valve endocarditis (incidence 2.5% jpt-yr), of which one (incidence 0.4% jpt-yr) was fatal. Paravalvular leak contributed to the fifth valve-related death. At 5 years, 83 % of patients were free of thromboembolic complications; 94 % were free of anticoagulant-related hemorrhage; and 71 % were free of all valve-related complications. There are few comparable data for patients who have had simultaneous replacement of aortic and mitral valves with other mechanical prostheses. The total incidence of valve-related complicationsfor patients with bioprosthesesranges between3.9% jpt-yr and 10.4% jpt-yr and is similar to the 9.1 %jpt-yr observed in the present series. The type of valve-related complication (thromboemboli and bleeding versus valve deterioration) is the principal difference between St. Jude Medical and bioprosthetic valves in patients who require simultaneous replacement of aortic and mitral valves.

Frederick Armenti, MD, Larry W. Stephenson, MD, and L. Henry Edmunds, Jr., MD,

Philadelphia. Pa.

h e bileaflet St. Jude Medical prosthesis was introduced clinically in 1977 with the hope that the wide, 85-degree opening angle of the leaflets and pyrolytic carbon construction would reduce the incidence of valve-related complications." Since 1977, nearly 100,000 St. Jude Medical prostheses have been

From the Division of Cardiothoracic Surgery, Department of Surgery, Hospital of the University of Pennsylvania, Philadelphia, Pa.

implanted" and several investigators have reported their experiences.?" Although some reports include experience with simultaneous aortic and mitral valve replacement.?" none has focused critically on this subset of patients. This study reports our experience with simultaneous replacement of the aortic and mitral valves with the St. Jude Medical prosthesis and compares valverelated complications with those reported for other cardiac valve prostheses.

Received for publication Oct. 30, 1986.

Methods

Accepted for publication Dec. 29, 1986.

From Jan. 1, 1980, through June 1, 1986, 92 consecutive patients had replacement of the aortic and mitral valves with the St. Jude Medical bileaflet valve at the Hospital of the University of Pennsylvania. During this period, eight other

Address for reprints: L. Henry Edmunds, Jr., MD, Hospital of the University of Pennsylvania, 3400 Spruce St., 4th Floor, Silverstein Pavilion, Philadelphia, PA 19104.

733

The Journal of

734

Thoracic and Cardiovascular Surgery

Armenti, Stephenson, Edmunds

Table m. Additional procedures

Table I. Preoperative data No. of patients

Description

NYHA functional class lor II

8 57 23 4

III IV

Emergency Previous cardiac operation Mitral commissurotomy Valve prosthesis Rhythm Sinus Atrial fibrillation

12

Coronary arterial bypass grafts Tricuspid annuloplasty Tricuspid valve replacement Coronary arterial bypass grafts and tricuspid annuloplasty Closure of atrial septal defect Subtotal Simultaneous insertion of permanent pacemaker Operative insertion of intra-aorta balloon Subtotal Total

II

(n = 56) (ml/dl) Cardiac index (n = 63) (L/m 2/min)

6 2

2 I

22

2 10

12 34

37 55

Table IV. Size of prosthesis Valve size (mm)

Table n. Catheterization data L VEDP (n = 66) (torr) pCWP (n = 66) (torr) Ejection fraction (n = 51) (%) Arteriovenous O2 difference

II

No. Aortic

Mean ± SD

Range

6

19

13.7 ± 6.2

0-30

23

25 25

25.2 ± 8.5 60.4 ± 19.5 6.1 ± 1.3

10-48 16-81

25

24

10

3.5-9.0

27 29

2.1 ± 0.7

1.2-4.9

21

Mitral

27 29 31

Legend: SO, Standard deviation. LVEOP, Left ventricular end-diastolic pressure. PCWP, Pulmonary capillary wedge pressure.

patients had replacement of both aortic and mitral valves with other mechanical valves or bioprostheses. Preoperatively, eight patients (9%) were in New York Heart Association (NYHA) class II, 57 (62%) class III, and 27 (29%) class IV or required emergency operation. Operations were performed by six surgeons. The status of all patients was known in August 1986; total follow-up of the group is 242 patient-years (mean 33.8 months). Warfarin sodium (Coumadin) anticoagulation was started 2 days after the operation in all patients. Heparin was used if warfarin intake was delayed. Occasional patients « 10) also received platelet inhibitors (aspirin or dipyridamole). The ages of the 92 patients ranged from 14 to 82 years, mean 57.6 years (standard deviation 12.4). Forty-three were male. Fourteen were 70 years or older at operation; 16 were less than 50. Preoperative data are presented in Tables I and II. One patient had a cardiac arrest 15 minutes before transfer to the operating room and a second arrest after transfer to the operating table. This patient and three others had emergency or urgent operation without prior cardiac catheterization. Three others had· emergency cardiac catheterization and immediate operation for left atrial thrombosis, acute fulminant pulmonary edema, or cardiogenic shock. One additional patient had endotracheal intubation before operation for pulmonary edema. Two had acute endocarditis of natural valves and one acute prosthetic endocarditis. Nearly all patients had either a history of rheumatic fever or pathologic confirmation of rheumatic heart disease. Operations were performed with a bubble or membrane oxygenator and cold crystalloid cardioplegic solution. Bypass time ranged from 101 to 308 minutes, mean 181 ± 86. The

2

8 48

36

aorta was clamped for 82 to 245 mm minutes,mean 132 ± 32. All valves were inserted with interrupted polyester mattress sutures with Teflon felt pledgets in the left atrium (mitral) or downstream to the valve (aortic). Twenty-two patients (25%) had additional procedures, excluding simultaneous insertion of permanent pacemakers and the intra-aortic balloon (Table III). The sizes of the prostheses are listed in Table IV; no patient required enlargement of the aortic anulus. Three patients had insertion of the intra-aortic balloon in the intensive care unit postoperatively (one early death). Seven patients (7.5%) required reexploration for persistent bleeding and four others for late pericardial serosanguineous effusion (three) or a trapped Swan-Ganz catheter (one). None of these patients died early. The mean stay in the hospital after the operation was 16.4 days (standard deviation 11.0, median 13 days). Twenty-one patients (23%) were hospitalized for more than 2 weeks after the operation. The incidence or linearized rate of valve-related morbidity and mortality is calculated as events per 100 patient-years of follow-up (i.e., 2.42) and presented as percent per patient-year (%/pt-yr). The terms incidence and linearized rate are used interchangeably. Morbidity and mortality data are also calculated and plotted actuarially as patients (not events) alive or free of a particular complication with respect to time after operation. 14. 15

Results Six patients died within 30 days of operation. The patient, aged 57 years, who had two episodes of cardiac

Volume 94 Number 5 November 1987

St. Jude Medical prostheses

735

100 80 ...J


->>

60

a: ~

en

40

~

20 0

0 92

12

24

67

51

36

40

48 30

60 16

72 2

MONTHS NO. of PATI ENTS

Fig. 1.. Actuarial survival rate of 92 patients undergoing simultaneous replacement of aortic and mitral valves with St. Jude Medical prostheses. No. of Patients refers to patients at risk. Error bars indicate standard error of the mean.

arrest immediately before the operation, had had previous aortic valve replacement and bypass grafts and could not be weaned from cardiopulmonary bypass. Another patient, aged 59 years, who previously had had mitral valve replacement and two strokes, had emergency operation for left atrial thrombosis and failed to regain consciousness. In one patient with suprasystemic pulmonary arterial pressures, the suture line of the thin, friable right atrium ruptured 11 hours after operation. In a fourth patient, who previously had undergone mitral commissurotomy, uncontrollable supraventricular atrial tachycardia developed 3 days after operation. Two patients, both aged 78 years, who had bar calcification of the posterior mitral anulus, died of ventricular rupture at the site of the bar 14 and 18 hours after operation. 16 Early operative mortality was 6.5%. Because five of the six patients who died early had conditions known to increase operative risk,16,17 no assessment of the influence of preoperative functional class, associated procedures, or age on early death is possible.All patients who died were in NYHA functional class IV before operation. Fig. 1 presents the actuarial survival curve of these 92 patients and Table V lists the causes of death and intervals between operation and death. For 57 patients in NYHA functional class III before operation, the actuarial survival rate at 3 years is 76% and at 5 years, 62%. The actuarial survival rate for 27 patients in class IV is 70% at 1 year and 52% at 3 years. Thirteen of 31 (42%) early and late deaths occurred in patients who were in functional class IV before the operation.

Table V. Late deaths Causes of late death Valve related Prosthetic endocarditis Stroke Paravalvular leak Sudden, unexplained Heart failure Reoperation, elsewhere, for CAD Noncardiac

Postop. time (mo)

3 2,12,64 6 1.5, 2, 3, 20, 36 2, 2, 2, 6, 18, 21, 25, 29,37,48 49 12, 17, 23, 56

Legend: CAD, Coronary artery disease.

Five valve-related late deaths occurred. One patient died of septicemia and postoperative endocarditis resulting from Staphylococcus epidermidis. He was readmitted and died at emergency reoperation 3 months after the initial procedure. Another patient had a paravalvular leak of the mitral valve within 1 month of operation, but multiple blood cultures were negative and there was no clinical evidence of septicemia or endocarditis at any time postoperatively. Complete heart block and tricuspid insufficiency developed 3 months later, and she died of heart failure 6 months postoperatively. Reoperation was ruled out since the amount of the leak was small and ventricular performance was severely reduced. Although arguable, her death is listed as valve related. Three patients died of strokes. One of these had a nonfatal stroke with residual deficit 33 months after operation before the fatal stroke at 64 months. Another patient

7 36

The Journal of Thoracic and Cardiovascular

Armenti, Stephenson, Edmunds

Surgery

L,------,-+-+--........lf

80

TE

L.;:-__-.-;_..,L---'----...J T E

z o

+

ACH

'----r---..., ALL

~

« u

...J

c,

::t:

o

60

u

, , , , o 1.----.-----..------..------..------..-----------"

o

92

12 59

i

24 45

36 33

48 25

60

12

72 2

MONTHS NO. of PATIENTS

Fig. 2. Actuarial calculations of patients free of major valve-related complications. PVE, Prosthetic valve endocarditis. ACH, Anticoagulant-related hemorrhage. TE, Thromboembolism. TE + ACH, Thromboembolism and anticoagulant-related hemorrhage together. All, All valve-related complications. No. of Patients refers to patients at risk for all valve-related complications. Error bars indicate standard error of the mean.

had a stroke before operation and the third had had two previous mitral commissurotomies. The linearized rate or incidence of fatal valve-related events was 2.1%/ pt-yr. Five patients died suddenly and unexpectedly while doing well at home. Three of these patients were in chronic atrial fibrillation, one in sinus rhythm, and one had a permanent pacemaker. Ten patients died of progressive heart failure at various intervals postoperatively. Noncardiac causes of death were renal failure, pneumonia, cirrhosis, and metastatic carcinoma. Ten patients (10.9%) died between 2 and 6 months after the operation. Fourteen patients (one of whom died later of a stroke) had 17 nonfatal valve-related complications for an incidence of 7.0%/pt-yr. Eight strokes in seven patients were not fatal. Three nonfatal strokes caused residual deficits and five, including two events in the same patient, did not. No patients had recognized emboli to other organs and no patients had prosthetic valve thrombosis. Seven of the nine patients with thromboemboli were in atrial fibrillation. The incidence of thromboembolism was higher in the first postoperative year than in subsequent years (Fig. 2). The linearized rate of fatal thromboembolism is 1.2%/pt-yr; of embolic events with residual defects, 1.2%/pt-yr; of embolic events without residual defects, 2.1%/pt-yr; and of all (11) thromboembolic events, 4.6%/pt-yr. The actuarial curve of patients who are free of all thromboembolic events is presented in Fig. 2.

No patient died of complications of anticoagulation. Three patients have required transfusion for gastrointestinal bleeding (two) and nose bleed (one). The incidence of all anticoagulant-related complications is 1.2%/ptyr. A total of four patients had six episodes of postoperative prosthetic valve endocarditis, and one of these died during emergency reoperation. The second patient, who was addicted to heroin and had active Pseudomonas endocarditis of his native valves at the original cardiac operation, had recurrent Pseudomonas endocarditis of his prosthetic valves. This patient had successful reoperation 13 months after his first valve replacement procedure. His is currently free of infection 30 months later. The third patient had positive blood cultures for Staphylococcus epidermidis 2 weeks after operation, but was treated with antibiotics for 8 weeks and remains well 32 months later. The fourth patient, aged 69 years, had had three separate episodes of prosthetic endocarditis and all have been successfully controlled by antibiotics. Staphylococcus epidermidis endocarditis developed 3 weeks after operation and subsequently two episodes of Streptococcus bovis endocarditis developed 24 and 36 months later. She remains apparently well without antibiotics 40 months after the valve operation. The incidence of fatal prosthetic endocarditis is O.4%/pt-yr and that of all prosthetic endocarditis, 2.5%/pt-yr. No other patients have had reoperation. The incidence of reoperation therefore is 0.8%/pt-yr. One patient has a murmur of paravalvular leak and hemoly-

Volume 94 Number 5 November 1987

sis and may require operation. No other patients have any valve-related complications. Eighteen of 92 patients (19.6%) have had valve-related fatal or nonfatal complications in 242 patient-years of follow-up (including the patient with paravalvular leak and hemolysis). Five (5.4%) complications were fatal. The linearized rate of all fatal and nonfatal prosthesis-related complications is 9.1%/pt-yr (22 events in 242 patient-years). Sixty-one patients remain alive 2 to 80 months after operation. Forty of these patients are currently in NYHA functional class I and 21 are in class II. The change in functional classification of the 96 hospital survivors 2 to 3 months after discharge is illustrated in Fig. 3. All except three survivors improved by at least one functional classification and 56 (61%) improved by two classifications or reached class I. Discussion The early hospital mortality of 6.5% is greater than the 1.2% mortality reported earlier" but is less than the hospital mortality of mitral valve replacement alone." This apparent paradox probably reflects the fact that concomitant aortic valve disease augments the hemodynamic burden caused by mitral valve disease and produces disabling symptoms before left ventricular contractility is severely compromised. Moreover, rheumatic fever remains the primary cause of combined valve disease, whereas nearly half of our patients who require mitral valve replacement alone had mitral valve dysfunction resulting from ischemia." Ischemic mitral valve disease is associated with a higher operative mortality than is rheumatic mitral disease." Operative mortality is obviously important, and hospital death represents a failure of either the prescription or execution of an operation. However, failure to prescribe high-risk operations for incapacitated, unhappy patients who are bereft of other therapeutic options is a lost opportunity to help those who will survive. Surgeons and cardiologists understandably have focused on operative mortality of groups ofpatients as an index of the safety of an operation. However, to a great extent, hospital mortality reflects the preoperative condition of the patient. Many factors such as advanced age, functional classification, emergency operation, acute endocarditis, preoperative myocardial infarction, stroke, arrhythmias, renal disease, pulmonary disase, and malnutrition, are associated with increased hospital deaths." 29 Although the possibility is always present, few patients now die of poorly executed operations or errors in postoperative care. Within wide limits, surgeons can adjust operative mortality by preoperative selection; this limits the utility of the operative mortality statistic as an index of the safety of an operation.

St. Jude Medical prostheses

737

I

n

n

m

m 6

PREOPERATIVE ------II~_

POSTOPERATIVE

Fig. 3. Change in NYHA functional classification for 92 patients. Postoperative evaluation was made 2 to 3 months after operation.

Patients who die within 30 days after the operation do not benefit from operation. Since full recovery of older patients often requires 3 to 6 months, one can argue that only those patients who survive more than 6 months and whose functional classification improves to class I or II status benefit from operation in terms of less disability and improved longevity or both. By this analysis, only 76 (83%) of our patients benefited from the operation. The others all died within 6 months of operation. Although far from ideal, this percentage is sufficiently high to justify operation in an increasingly older population and in patients who have risk factors that significantly increase the possibility of operative death. In this series, four had emergency (immediate) operation, three others had urgent operation without prior cardiac catheterization, three had acute endocarditis, 27 (including the 10 just mentioned) were in class IV, 14 were more than 70 years of age, and many others had serious associated diseases or other known risk factors. Patients who were in NYHA functional class IV had the highest operative risk (22%) and the least benefit in terms of improvement to class I or II (67%) or increment in longevity (52% 3-year actuarial survival rate). Similarly, age more than 70 years was associated with a 14% operative mortality, a 79% improvement to class I or II, and a 61% 3-year actuarial survival rate. Fourteen of the 22 valve-related late complications (64%) were due to thromboembolism and anticoagulant-related bleeding, and only one was due, somewhat arbitrarily, to a mechanical complication of the valve.

738

The Journal of Thoracic and Cardiovascular Surgery

Armenti, Stephenson, Edmunds

The mechanical durability of the St. Jude Medical valve and central flow characteristics with tolerable backflow' are distinct advantages of this valve over other tissue and mechanical valves. Reoperation is usually required only for prosthetic infection. Nevertheless, the incidence of fatal and disability-causing strokes and the life-long need for anticoagulants remain major disadvantages of the prosthesis. With time, the incidence of thromboembolic complications can be expected to decrease," but the incidence of bleeding complications will likely remain constant. Kinsley, Antunes, and Colsen?have recently reported their experience with simultaneous replacement of aortic and mitral valves with St. Jude Medical prostheses. They observed a 2.0%/pt-yr late valve-related mortality (versus 2.1%/pt-yr) and a 5-year survival rate of 76% (versus 60%). Although Kinsley's group did not report the incidence of nonfatal hemorrhage, freedom from all valve-related complications at 5 years was 76%, which is similar to our 71%. In our series, the total incidence of all valve-related complications is 9.1%./pt-yr. Although there are many excellent series of simultaneous replacement of aortic and mitral valves with other prostheses, only a few reports provide incidence and actuarial data of valve-related morbidity and mortality. No comparable current data are available for bare-strut Starr-Edwards, Omniscience, and Medtronic Hall prostheses. Sethia and associates" report a 3.3%/pt-yr incidence of valve-related complications after replacement with standard Bjork-Shiley prostheses. Mean age of patients was 47 years and 5-year survival rate, excluding an operative mortality of 12.6%,was 81%. For bioprostheses, the incidence of valve-related complications is 4.9/° 6.0,21 and 6.9%/pt-yr22 with Hancock porcine heterografts, 1O.4%/pt-yr in a small series of Carpentier-Edwards prostheses," and 3.9%/pt-yr with the Ionescu-Shiley pericardial valve." Variations in selection criteria, heterogeneity of patient populations, and paucity of data preclude valid comparisons between different valve prostheses used for simultaneous aortic and mitral valve replacement. Available data suggest that the incidence of valverelated complications for St. Jude Medical and various bioprostheses is similar. The principal difference is in the type of complication. Neither mechanical failure nor thrombosis occurred in this series; thromboembolism, anticoagulant-related bleeding, and prosthetic endocarditis were the major causes of valve-related morbidity and mortality. In contrast, patients with two bioprostheses have fewer thromboembolic and bleeding complications, have a similar incidence of infection, but more often require reoperation for deterioration of the

valve.":" Since the incidence of bioprosthetic deterioration increases with time," 28 the superior durability of the St. Jude Medical prosthesis may prove decisive. Thus we continue to recommend this prosthesis for simultaneous aortic and mitral valve repalcement. We thank Stuart Bogom, Microcomputer Coordinator, Population Studies Center, University of Pennsylvania, for the actuarial calculations and plots. REFERENCES 1. Emery RW, Mettler E, Nicoloff DM. A new cardiac prosthesis: the St. Jude mechanical cardiac valve. Circulation 1979;60(Pt 2):148-54. 2. Nicoloff DM, Emory RW, Arom KV, et al. Clinical and hemodynamic results with the St. Jude Medical cardiac valve prosthesis: A three-year experience. J THoRAc CARDIOVASC SURG 1981;82:674-83. 3. Yoganathan AP, Chaux A, Gray RJ, DeRobertis M, Matloff JM. Flow characteristics of the St. Jude prosthetic valve: an in vitro and in vivo study. Artif Intern Organs 1982;6:288-93. 4. Lillehei CWo The St. Jude Medical prosthetic heart valve: results from a 5-year multicenter experience. In: Horstkotte D, Loogen F, eds. Update in heart valve replacement. Dusseldorf, Federal Republic of Germany: Steinkopff Verlag Darmstadt, 1986:3-18. 5. Crawford FA Jr, Kratz JM, Sade RM, Stroud MR, Bartles DM. Aortic and mitral valve replacement with the St. Jude Medical prosthesis. Ann Surg 1984;199:75361. 6. Baudet EM, Oca CC, Roques XF, et al. A 51/2 year experience with the St. Jude Medical cardiac valve prosthesis. J THORAC CARDIOVASC SURG 1985;90:13744. 7. Czer LSC, Matloff J, Chaux A, DeRobertis M, Yoganathan A, Gray RJ. A 6-year experience with the St. Jude Medical valve: hemodynamic performance, surgical results, biocompatibility and follow-up. J Am ColI Cardiol 1985;6:904-12. 8. Arom KV, Nicoloff DM, Kersten TE, Northrup WF III, Lindsay WG~ Six years of experience with the St. Jude Medical valvular prosthesis. Circulation 1985;72(Pt 2):11153-8. 9. Kinsley RH, Antunes MJ, Colsen PRo St. Jude Medical valve replacement: an evaluation of valve performance. J THORAC CARDIOVASC SURG 1986;92:349-60. 10. Duveau D, Michaud JL, Despins P, et al. Mitral valve replacement with St. Jude Medical prosthesis: incidence of thromboembolic events in 349 patients. Eur Heart J 1984;5(Suppl D):49-52. 11. Burckhardt D, Hoffmann A, Vogt S, Pfisterer M, Hasse J, Gradel E. Clinical evaluation of the St. Jude Medical heart valve prosthesis. J THORAC CARDIOVASC SURG 1984;88:432-8. 12. Duncan JM, Cooley DA, Livesay 11, et al. The St. Jude

Volume 94 Number 5 November 1987

13.

14.

15. 16.

17.

18.

19.

20.

Medical valve: early clinical results in 253 patients. Texas Heart Inst J 1983;10:11-6. Horstkotte 0, Korfer R, Seipel L, Bircks W, Loogen F. Late complications in patients with Bjork-Shiley and St. Jude Medical heart valve replacement. Circulation 1983; 68(Pt 2):I1I75-84. Kalbfleisch JD, Prentice RL. The statistical analysis of failure time data. New York: John Wiley & Sons, 1980. SAS User's Guide. Statistics, 5th ed. SAS Institute Inc., Cary, North Carolina: 1985:529-57. Cammack PL, Edie RN, Edmunds LH Jr. Bar calcification of the mitral anulus: a risk factor in mitral valve surgery. J THORAC CARDIOVASC SURO 1987;94:399-404. Edmunds LH Jr, Addonizio VP Jr, Tepe NA. Valvular heart disease: prosthetic valve replacement. In: Parmely WW, Chatterjee K, Cheitlin M, Rapaport E, Karliner J, Scheinman M, eds. Cardiology. Philadelphia, JP Lippincott Co (in press). Stephenson LW, Edie RN, Harkin AH, Edmunds LH Jr. Combined aortic and mitral valve replacement: changes in practice and prognosis. Circulation 1984;69:640-4. Chaux A, Czer LS, Matloff JM, et al. The St. Jude Medical bileaflet valve prosthesis: a five year experience. J THORAC CARDIOVASC SURG 1984;88:706-17. Kirklin JW, Barratt-Boyes BG. Cardiac surgery. New York: John Wiley & Sons, 1986:431-45.

St. Jude Medical prostheses 739

21. Sethia B, Turner MA, Lewis S, Rodger RA, Bain WH: Fourteen years' experience with Bjork-Shiley tilting disc prosthesis. J THORAC CARDIOVASC SURG 1986;91:35061. 22. Zussa C, Ottino G, DiSumma M, et al. Porcine cardiac bioprostheses: evaluation of long-term results in 990 patients. Ann Thorac Surg 1985;39:243-50. 23. Janusz MT, Jamieson WRE, Allen P, et al. Experience with the Carpentier-Edwards porcine valve prosthesis in 700 patients. Ann Thorac Surg 1982;34:625-33. 24. Reul GJ Jr, Cooley DA, Duncan JM, et al. Valve failure with the Ionescu-Shiley bovine periocardial bioprosthesis: analysis of 2080 patients. J Vase Surg 1985;192-204. 25. Oyer PE, Stinson EB, Reitz BA, Miller DC, Rossiter SJ, Shumway NE. Long-term evaluation of the porcine xenograft bioprosthesis. J THORAC CARDIOVASC SURG 1979;78:343-50. 26. Gallo JI, Ruiz B, Carrion MF, Gutierrez JA, Vega JL, Duran CMG. Heart valve replacement with the Hancock bioprosthesis: a 6-year review. Ann Thorac Surg 1981; 31:444-9. 27. Magilligan OJ Jr, Lewis JW Jr, Tilley B, Peterson E. The porcine bioprosthetic valve. J THORAC CARDIOVASC SURG 1985;89:499-507. 28. Gallo I, Nistal F, Artifiano E: Six to ten year follow-up of patients with the Hancock cardiac bioprosthesis. J THORAC CARDIOVASC SURG 1986;92:14-20.