Twelve-Y ear Experience with Mitral Valve Replacement Joseph W. Rubin, M.D., Robert G. Ellison, M.D., H. Victor Moore, M.D., Rollie J. Harp, M.S., a n d William S. Hitch, M.D. ABSTRACT Between 1962 and 1974,203mitral prostheses were implanted in 201 patients. Of the 102 survivors, 29 have Beall, 25 Kay-Shiley, 22 StarrEdwards (SE) 6000,and 27 SE 6320 valves. Full rehabilitation was achieved in 25 patients with Beall and 23 with SE 6320 valves. Sixteen with SE 6000 valves remain normally active. Only 8 with Kay-Shiley prostheses have resumed normal activities. Systemic embolization occurred with the following frequencies per 1,000 patient-months: 13.7for those receiving the Kay-Shileyvalve; 7.2 in the SE 6000 group; 4.3 after SE 6320 implantation; and 3.1 for the Beall group. Other prosthesis-related complications that were much less frequent included detachment (lo),bacterial endocarditis (5), and hemolysis (10).Three Kay-Shiley valves malfunctioned. Life table analyses reveal the following survival rates: 33% after 11 years in the SE 6000patients, 50% after 7.5 years in the Kay-Shiley group, 69% 2.5 years after SE 6320 implantation, and 65% 3.5 years after replacement with a Beall valve. Evidence is presented to support the extension of operative treatment to patients with less advanced valvular heart disease. Postoperative anticoagulation remains an unresolved issue despite lower rates of thromboembolism. More cumulative analyses of survival and morbidity and follow-up hemodynamic data are needed to assess the Beall and SE 6320 prostheses now employed in our valve replacement program.

A

t the beginning of the last decade, valve replacement was reserved for

those patients who were critically ill and benefiting less and less from medical treatment. With improvement in the design of prostheses and increasing operative experience, morbidity and mortality as a result of valve replacement have declined. Nevertheless, postoperative systemic embolization has persisted, and that together with the resulting need for anticoagulation have remained the principal drawbacks of the available prosthetic valves. Although it may be possible to extend surgical therapy to patients with less advanced disease, the disadvantages of implantable devices indicate the ongoing necessity of exercising cautious patient selection, especially among individuals who are only mildly disabled. With these considerations in mind we have reviewed our own experience with mitral valve replacement. The Medical College of Georgia began its open-heart program in 1956. However, the first mitral valve replacement was not performed until 1962. During From the Division of Thoracic and Cardiac Surgery, Medical College of Georgia, Augusta, Ga. Presented at the Twenty-first Annual Meeting of the Southern Thoracic Surgical Association, Williamsburg, Va., Nov. 7-9, 1974. Address reprint requests to Dr. Rubin, Division of Thoracic and Cardiac Surgery, Medical College of Georgia, Augusta, Ga. 30902.

VOL. 19, NO. 6, JUNE, 1975

659

RUBIN ET AL.

the twelve-year period ending June 30, 1974, isolated mitral valve replacement accounted for 20% of all open-heart operations. Since April, 1962, the following four types of mitral prosthesis have been employed: the Starr-Edwards Models 6000 and 6320 caged-ball valves (SE 6000 and SE 6320) and the Kay-Shiley and Beall low-profile disc valves. With evolution in prosthesis design during this study period, substitution of one device for another in clinical use was prompted by the promise of increased survival and a solution to recurrent complications such as thromboembolism, relative obstruction of blood flow in the prosthesis, detachment, destruction of formed elements of blood, poppet damage, and tissue ingrowth interfering with the traverse of the occluder or orifice size. This report presents our experience with these prostheses. Analysis of patient status during follow-up has permitted us to assess comparative performance of the four prostheses selected for mitral valve replacement during the study period.

Clinical Material Two hundred three isolated mitral valve replacements were performed in 201 patients between April, 1962, and July, 1974. The ages ranged from 4 to 73 years with a median age of 40 years (Fig. 1). There were 127 female and 74 male patients. Hemodynamic evaluation by cardiac catheterization was performed preoperatively in all but 16 patients. Exposure of the heart was gained initially through a left posterolateral thoracotomy and later in the series through a right anterolateral thoracotomy or median sternotomy. Open cardiotomy, usually by direct left atrial incision, was performed under various conditions of extracorporeal circulation. Early in the clinical program normothermic perfusion was employed. At that time the extracorporeal circuit with a disc oxygenator was primed with whole blood collected in ACD solution. Later in the series the standard technique for cardiopulmonary bypass was hypothermic perfusion with hemodilution priming and a bubble

70

1

FIG. 1 . Age and sex distribution of 201 patients undergoingmitral valve replacement.

61

60 74 MALES

127 FEMALES

50 -

20 I

44

40-

30-

20-

IC

15

10-

L

0-10

660

10-20

20-30

30-40 40-50 YEARS

50-60

60-70

THE ANNALS OF THORACIC SURGERY

70-80

Mitral Valve Replacement oxygenator. Intermittent elective ischemic cardioplegia was induced in most instances by cross-clampingthe ascending aorta. Alternatively,the heart was electrically fibrillated. All patients were anticoagulated with warfarin prior to discharge from the hospital. Although continuous anticoagulation was recommended, considerable difficulty in maintaining a satisfactory therapeutic level was encountered in many patients. Follow-up examinations of long-term survivors were usually performed at half-yearly or yearly intervals. In some instances the family physician provided postoperative data by either letter or telephone interview. Thirteen patients were lost to follow-upor moved to remote areas. In the remainder, recovery wasjudged according to level of rehabilitation achieved.

Results Table 1 outlines preliminary clinical data, classifying patients with respect to etiology of the valve disease, hemodynamic lesion, associated conditions, and functional status. No statistically significant relationship was demonstrated between these commonly selected preoperative factors and survival. Twenty-six TABLE 1 . PREOPERATIVE DATA A N D RELATED MORTALITY IN 201 PATIENTS UNDERGOING MITRAL VALVE REPLACEMENT

Data

No. of Patients

No. of Deaths

195 4 2 20 1

86 1 0 87

75 74 34 18 201

31 26 19 11 87

28 56

0 19

10 6

4 3 1

2 33 83

0 16 30 15 26 87

Etiology Rheumatic Congenital Ruptured chordae Total Hemodynamic lesion Pure MS Pure MR Mixed MS & MR Multiple-valve disease Total Associated conditions Previous cardiac procedure Pulmonary hypertension (PA > 40 mm Hg) Tricuspid regurgitation Systemic thromboembolism Bacterial endocarditis Functional Class (N.Y.H.A.) I I1 I11 IV Not listed Total MS

=

mitral stenosis; MR

=

7

27

56 20 1

mitral regurgitation.

VOL. 19, NO. 6 , JUNE, 1975

66 1

RUBIN ET AL. TABLE 2. EXPERIENCE WITH MITRAL VALVE REPLACEMENT IN 201 PATIENTS

Valve

No. of Procedures

Operative Deaths

Late Deaths

Long-term Survivors

Lost to Follow-up

Beall Kay-Shiley SE 6000 SE 6320 Total

43 55 67 38 203

10

4 14 24 3 45

29 25* 22 27 102*

0 8 4 1 13

8

17 7 42

*Two patients underwent replacement of Kay-Shiley prostheses. One died of myocardial infarction in the early postoperative period. The other is fully rehabilitated with a SE 6320 valve.

patients had undergone previous repair of the mitral valve by closed o r open methods, and 2 had had previous aortic valve replacement. All these patients have survived. More than half of the patients have persistent atrial fibrillation. Experience with each of the four prostheses is summarized in Table 2. Throughout the period of study there were 87 deaths following mitral valve replacement, 42 occurring in the early postoperative period. T h e most frequent causes of early death were myocardial failure and respiratory complications. Forty-five deaths occurred 1 to 107 months after operation. Table 3 lists prosthesis-related complications in 147 patients who survived operation. Twenty-six of the 45 late deaths were from these complications. Thirty-six patients suffered recurrent thromboembolism, the most frequent cause of morbidity. Of the 36, the Kay-Shiley and SE 6000 groups each accounted for 16 patients. In contrast, systemic emboli developed in only 2 of 33 individuals who received the Beall prosthesis and in 2 of 30 in the SE 6320 group. T h e TABLE 3. PROSTHESIS-RELATED COMPLICATIONS AND MORTALITY IN 147 PATIENTS WHO SURVIVED OPERATION

Complication Thromboembolism Detachment Bacterial endocarditis Further cardiac operation Repair of leak Prosthesis malfunction Aortic or tricuspid replacement Other Hemorrhage due to anticoagulation Hemolysis Tot a1

Beall Kay-Shiley SE 6000 SE 6320 No. of (N = 33) (N = 39*) ( N = 46) ( N = 30*) Deaths

2 3 1 1

4$

16 3 2 3 1 2 4$

16 4 2

2

3 1

3 2$

1

13

2t 2 2 1 3 26

*One patient is included who had a Kay-Shiley prosthesis replaced with a SE 6320 valve. tDied 2 and 8 years following reoperation, respectively, with progressive myocardial disease and recurrent congestive heart failure. $One patient in each group has peribasilar leak.

662

THE ANNALS OF THORACIC SURGERY

Mitral Valve Replacement mortality from this complication for the entire group of 147 patients was 8.8%(13 persons) following 55 recorded episodes of thromboembolism. Other, less frequent valve-related complications outlined in detail in Table 3 included periprosthetic leak, infection of the prosthesis, hemorrhage due to anticoagulation, hemolysis, and reoperation. Noteworthy is the paucity of complications among patients in the SE 6320 group. Hemolysis occurred with considerable frequency early after implantation of low-profiledisc valves. This problem was self-limiting in all but 3 patients in the Beall group, who nevertheless are fully rehabilitated. Peribasilar detachments have also accelerated the destruction of formed blood elements. Closure of the leaks has satisfactorily reduced the hemolysis rate. The cause of death in 10 patients was either unknown or noncardiac. Late, progressive left ventricular dysfunction of undetermined etiology was identified clinically in 9 patients. Four in the SE 6000 group who had been well died suddenly, presumably from ventricular arrhythmias. Since thromboembolism was the greatest single cause of late morbidity, the incidence of embolic phenomena was related to aggregate follow-up months for each prosthesis. Patients with Kay-Shiley and SE 6000 prostheses have had aggregate follow-up of 1,895 and 3,490 months, respectively. The Kay-Shiley valve had the worst embolism record (13.7 per 1,000 patient-months), followed by the SE 6000 (7.2 per 1,000 patient-months). Occurrence of embolization following implantation of the SE 6320 prostheses has been 4.3 per 1,000 patient-months after 470 months of follow-up. For Beall valves we have observed 3.1 emboli per 1,000 patient-months after 644 months of follow-up. For each prosthesis group a series of actuarial curves has been developed using the life table method [5](see Figs. 2-5). The computations make use of all survival information and the times of the first embolic episode from the date of mitral valve replacement, collected up to the closing time of the study. A point on the survival curve for a particular year represents the cumulative proportion surviving from the time of mitral valve replacement until that year. The margin of standard error for the survival rate is also indicated for each year of follow-up since valve replacement. The actuarial curves for the study of embolism represent the cumulative proportion of patients surviving at the end of the first postoperative month and remaining free from embolism. The occurrences of the other prosthesis-related complications listed in Table 3 were too infrequent to permit derivation of useful actuarial curves. Sixty-five percent of the patients who received Beall valves are surviving and 92% are free from embolism (Fig. 2). Sixty-nine percent of the SE 6320 patients are alive; 86% are without embolism (Fig. 3). In the Beall group, no late deaths have occurred in the past 24 months of follow-up. Maximum follow-up in the SE 6320 and Beall groups has been 26 and 40 months, respectively. Only 33% of the SE 6000 group are still alive, the longest survival being 107 months; 45%of the survivors are free from embolism (Fig. 4). Fifty percent of the Kay-Shiley group have lived 90 months, 56% of them without incidence of systemic embolization (Fig. 5). In contrast to patients who received the Beall VOL. 19, NO. 6, J U N E , 1975

663

RUBIN ET AL.

86% Embolism free

69% Survival

FIG. 2 . Mortality a d thromboembolismfollowing isolated mitral valve replacement with the Beallposthesis: 43 procedures with 33 operative survivors.

YEARS

prosthesis (see Fig. 2), survival continues to decline parallel to the rate of thromboembolism in the SE 6000 and Kay-Shiley groups (see Figs. 4,5). Our experience with the SE 6320 valve has been too brief to indicate the long-term risk of embolism and death by the life table method. Examination of the functional status of 102 survivors at the end of the study

9 2 % Embolism tree

FIG. 3. Mortality and thromboembolismfollowing replacement with the Stan-Edwards 6320 pasthesis: 38 procedures with 31 operative survivors.

6 5 % Survival

YEARS

664

THE ANNALS OF THORACIC SURGERY

Mitral Valve Replacement

80-

5045.hEmbdism ,Embolism free

33% Survival

YEARS

FIG. 4 . Mortality and thromboembolism following replacement with the Starr-Edwards 6000 prosthesis: 67 procedures with 50 operative survivors.

period disclosed that 72 were fully rehabilitated (Table 4).Twenty-five of 29 in the Beall group and 23 of 27 who received the SE 6320 prosthesis have resumed normal activities. Among 22 still alive in the SE 6000 group, 16 have regained full activity. These results are in contrast to the outcome in 25 patients who are late FIG. 5. Mortality and thromboembolism following replacement with the Kay-Shilq prosthesis: 55 procedures with 47 operative survivors.

100

90

80

n

E

..

70

0

a

6

z

60 56 X Embolism free

401;

50

I M o I

50% Survival

rI:

:

;

:

;

:

;

2

3

4

5

6

7

8

YEARS

VOI.. 19. NO. 6. JUNE. 1975

665

RUBIN ET AL. TABLE 4. FUNCTIONAL RECOVERY IN 102 LONG-TERM SURVIVORS

Status Poor Asymptomatic but not working Normal activities Lost to follow-up Total

Beall

No. of Patients Kay-Shiley SE 6000

2

6*

2 25

5 8 8 25*

29

2 16 4 22

SE 6320

3 23* 1

27

*Included are 3 patients who required prosthesis replacement. One died in the early postoperative period. A second died before reoperation. The third is alive and well with a SE 6320 valve.

survivors of Kay-Shiley valve implantation. Although 8 have been lost to followup, only 8 of the 25 in this group are known to have been rehabilitated completely. Twelve patients in all groups were asymptomaticbut unable to work. In 8 patients recovery was adversely affected by recurrent emboli, prosthesis malfunction, associated valve disease, progressive left ventricular dysfunction, or peribasilar leak. Six of the 8 patients had Kay-Shiley prostheses. Two submitted to reoperation when follow-up hemodynamic evaluation revealed signs of obstructed left atrial outflow thought to be caused by prosthesis malfunction. One patient is now well with a SE 6320 valve. The other died in the early postoperative period of a myocardial infarction. Although one valve was incriminated as a source for recurrent thromboembolism, neither of the two replaced prostheses was found at reoperation to be physically obstructed by thrombus or fibrosis. A third patient with a Kay-Shileyvalve developed signs of progressive obstruction of the left atrial outflow tract and died before the prosthesis could be replaced. At postmortem examination the valve orifice was seen to be narrowed by ingrowth of fibrous tissue. Two patients with Beall valves have not improved. Detachment has caused excessive hemolysis in 1, who will undergo reoperation in the near future. In the other patient poor left ventricular function discovered at repeat cardiac catheterization was attributed to moderately severe aortic stenosis and regurgitation. Comment

The period encompassed by this study represents the era of exponential development of the science and expertise of valve replacement. The progressive decrease in morbidity and mortality following mitral valve replacement is due to improving surgical management and better prosthetic material. These advances and the published data of several investigators suggest that long-term results may be improved by extending operative intervention to patients who are less disabled and continue to respond to medical therapy [2,3,11,17]. For example, as outlined in Table 1, this study illustrates the unpredictable outcome (61 deaths in 143 patients) in the presence of congestive heart failure and varying disabilities categorized as N.Y.H.A. Functional Class I1 to IV. Similarly, the fate of individu-

666

THE ANNALS OF THORACIC SURGERY

Mitral Valve Replacement als with multiple diseased valves could not be predicted if only the abnormal mitral valve was replaced (11 deaths in 18 patients). In contrast, all the patients who had received palliation (and protection) from a previous repair of the mitral valve survived replacement. Furthermore, survival was improved in instances of ruptured chordae tendineae or recurrent bacterial endocarditis when the indication for operation was more urgent than that of chronic congestive failure still responsive to medical therapy. Severe hemodynamic abnormalities may be masked by the capacity of the heart to compensate without causing clinical disability, particularly with the aid of digitalis and diuretics. Traditionally, operation had been suggested when there was chronic cardiac decompensation and continued medical management was of no avail. In our early experience with the SE 6000 and Kay-Shiley valves (seeTable 3) this delay may have been justified because of the significant incidence of prosthesis-related complications [6, 10, 141. Although Starr and his associates previously emphasized cautious selection of only the most severely incapacitated for operation, their latest results with the SE 6320 have led them to advocate earlier intervention [3, 15, 161. There is considerable assurance that the present version of the Beall valve and the SE 6320 both carry a greatly reduced incidence of systemic embolization. Furthermore, improved surgical management and better prosthetic material may diminish further the low rates of detachment (lo/ 147), infection (5/147), and hemolysis (10/147) that we have observed. Nevertheless, our current experience reveals persistent incidence of thromboembolism, 4.3 episodes per 1,000 patient-months with the SE 6320 prosthesis and 3.1 with the Beall valve. In order to establish a case for early operative intervention, the risks of prosthesis-related late mortality and morbidity should be reliably less than the risks of death and morbidity, particularly from systemic embolization, in patients with chronic valvular heart disease. According to actuarial data, the Kay-Shiley and SE 6000 valves have not satisfied these criteria (see Figs. 4,5). After examining the life table analyses of survival and thromboembolism in the Beall group, early operation is advisable provided that operative deaths, which have occurred in our experience, can be ascribed to an advanced state of disability due to heart disease. N o deaths have occurred beyond 1 year of observation in this group, and 92%of the survivors have remained free from embolism (see Fig. 2). At present this is the only evidence in our experience that supports valve replacement at early stages of functional disability. Our patients with SE 6320 valves have not been followed long enough to establish late risks of valve-related morbidity and death (see Fig. 3). Additional postoperative hemodynamic data are needed to assess our results with the Beall and SE 6320 prostheses. Historically, the low-profile disc valve design represented a departure from the caged-ball SE 6000. Although the caged-ball prosthesis was hemodynamically reliable, the bulky cage structure was considered a disadvantage and a source of complications. These included impingement of the cage on the interventricular septum, leading to arrhythmias, erosion, thrombus formation, and subsequent embolization and interference with ball motion [4, 161. The Kay-Shiley valve was

VOI.. 19. NO. 6 , JUNE. 1975

667

RUBIN E T AL. introduced with the expectation of improving operative results [4, 8, 91. Our findings of alarming rates of thromboembolism and hemodynamic results indicative of stenosis of the valve are similar to the experiences reported by others [7, 121. After we discontinued use of the Kay-Shiley prosthesis, superior results have been obtained with the SE 6320 and Beall valves. Despite reduced embolism rates, the use of postoperative anticoagulation, particularly following implantation of a Beall or SE 6320 valve, is an unresolved question. Patients with advanced rheumatic heart disease and a mitral prosthesis are continually exposed to the risk of embolism, due not only to the prosthesis but also to a dilated left atrium or atrial fibrillation. For these reasons we continue to employ postoperative anticoagulation. No hemorrhagic complications have occurred in these two patient groups. Several investigators did not administer anticoagulants after mitral replacement with the Beall prosthesis and have reported embolism rates similar to those found in our series [ 1,2, 131. T h e decision for or against anticoagulation after use of the newer prostheses should be based upon individual clinical judgment after comparing the risks of anticoagulation and thromboembolism in a given patient population. Rigorous statistical analysis to aid in resolving this issue seems impossible. T h e indication for continuing anticoagulation in survivors with the Kay-Shiley o r SE 6000 valve appears clear. Actuarial analyses of these patient groups (see Figs. 4, 5 ) reveal a high risk of embolism.

References 1. Beall, A. C., Bloodwell, R. D., Arbegast, N. R., Liotta, D., Cooley, D. A., and DeBakey, M. E. Mitral Valve Replacement with Dacron-covered Disk Prosthesis to Prevent Thromboembolism: Clinical Experience in 202 Cases. In L. A. Brewer, 111 (Ed), Prosthetic Heart Values. Springfield, Ill.: Thomas, 1970. Chap 2 1. 2. Beall, A. C., Jr., Morris, G. C., Jr., Noon, G. P., Guinn, G. A., Reul, G. J., Jr., Lefrak, E. A., and Greenberg, S. D. An improved mitral valve prosthesis. Ann Thorac Surg 15:25, 1973. 3. Bonchek, L. I., Anderson, R. P., and Starr, A. Mitral valve replacement with clothcovered composite-seat prostheses. J Thorac Cardiouasc Surg 67:93, 1974. 4. Cooley, D. A., Bloodwell, R. D., and Hallman, G. L. Mitral valve replacement with a discoid prosthesis. Ann Thorac Surg 3:487, 1956. 5. Cutler, S.J., and Ederer, F. Maximum utilization of the life table method in analyzing survival. J Chron Dis 8:699, 1958. 6. Herr, R., Starr, A., McCord, C. W., and Wood, J. A. Special problems following valve replacement. Ann Thoruc Surg 1:403, 1965. 7. Hughes, R. K., and Carey, J. S. Experience with the Kay-Shiley Mitral Valve. In L. A. Brewer, 111(Ed),ProstheticHeart Values. Springfield, 111.: Thomas, 1970. Chap 40. 8. Kay, J. H . , Kawashima, Y., Kagawa, Y., Tsuji, H . K., a n d Redington, J. V. Experimental mitral valve replacement with a new disc valve. Ann Thoruc Surg 2:485, 1966. 9. Kay, J. H., Tsuji, H. K., Redington, J. V., Mendez, A., Saji, K., Kamata, K., Yokoyama, T., Magidson, O., and Krohn, B. Experiences with the Kay-Shiley Disk Valve. In L. A. Brewer, 111(Ed),ProstheticHeart Values. Springfield, 111.: Thomas, 1970. Chap42. 10. Levine, F. H., Copeland, J. G., and Morrow, A. G. Prosthetic replacement of the mitral valve. Circulation 47:518, 1973. 1 1 . Nichols, H. T., Fernandez, J., Morse, D., and Gooch, A. S. Improved results in 336 patients with the isolated mitral Beall valve replacement. Chest 62:266, 1972. 12. Payton, B. C., Vogel, J. H. K., Ovary, H., and Blount, S. G., Jr. Follow-up Results on

668

THE ANNALS OF THORACIC SURGERY

Mitral Valve Replacement

13. 14. 15. 16.

17.

the Kay-Shiley Valve. In L. A. Brewer, 111(Ed),ProstheticHeart Values. Springfield, Ill.: Thomas, 1970. Chap 41. Stanford, W., Lindberg, E. F., and Armstrong, R. G. Implantation of heart valve prostheses without anticoagulants. J Thorac Cardiovasc Surg 63:648, 1972. Starr, A. Mitral valve replacement with ball valve prosthesis. Br Heart J 33:47, 1971. Starr, A., and Edwards, M. L. Mitral replacement: Clinical experience with a ballvalve prosthesis. Ann Surg 154:726, 1961. Starr, A., Herr, R. H., and Wood, J. A. Mitral replacement: Review of six years experience. J Thorac Cardioumc Surg 54:333, 1967. Winter, T. Q., Reis, R. L., Glancy, D. L., Roberts, W. C., Epstein, S. E., and Morrow, A. G. Current status of the Starr-Edwards cloth covered prosthetic cardiac valves. Circulation 45: 1, 1972.

NOTICE FROM THE SOUTHERN THORACIC SURGICAL ASSOCIATION T h e Twenty-second Annual Meeting of the Southern Thoracic Surgical Association will be held at the Fairmont-Roosevelt Hotel, New Orleans, La., on November 6-8, 1975. Reservations may be made by writing to the Reservations Manager, Fairmont-Roosevelt Hotel, N e w Orleans, La. Application for membership in the Southern Thoracic Surgical Association, on forms provided by the Association, should be sent directly to J. Alex Haller, Jr., M.D. [Chairman of the Membership Cominittee], The Johns Hopkins Hospital, Baltimore, Md.2 1205. The deadline for application t o membership is September 1, 1975. Papers that are accepted for the program will be considered for publication in The Annals and must be submitted to the Editor by October 15, 1975. J A M E S W. BROOKS,M.D. Secretar~-Ti-e~~surer

VOL. 19, N O . 6, JUNE, 1975

669