Appraisal therapy
Surgical
and reappraisal Edited
by Arthur
treatment
C. DeGraff,
of valvutar
Part V. Prosthetic replacement
of cardiac
Alan
heart
F. Lyon, and Julian
Frieden
disease
of the mitral valve
CURRENTSTATUS Frank C. Spencer, M.D. New York, N. Y.
P
rosthetic replacement of the mitral valve first became possible with the introduction of the ball-valve prosthesis by Starr and Edwards’ in 1961. In the seven years since this historic achievement, over 15,000 mitral prostheses have been sold by the Edwards Laboratories, well indicating the widespread need for prosthetic replacement. With experience, the risk of operation has progressively decreased. In 1967, Starr and associates2 reported personal experience with 278 patients undergoing mitral valve replacement since 1960. The operative mortality rate for isolated valve replacement decreased from 50 per cent to between 7 and 22 per cent, an average of 18 per cent. Of particular significance was the fact that the operative mortality rate for multiple valve replacement was almost identical, decreasing from 28 per cent in 1962 to 12 per cent in 1967. During this time, several other mitral prostheses have been developed and recent studies have investigated homografts and heterografts. This article will consider these different types of mitral prostheses, present indications and operative risks, and long-term prognosis and management. From
576
the
Department
of Surgery.
American Heart Journal
New
York
University
Types of mitral prostheses Ball-valve prostheses. Most experience has been with the ball-valve prosthesis originally developed by Starr, and previously employed by Hufnagel in the descending aorta before extracorporeal circulation was possible. The cardiovascular service at New York University has performed over 200 mitral replacements, evaluating both ball valves and disc valves, with the author preferring the ball-valve prosthesis. A most important consideration with a ball-valve prosthesis is to select the correct size, using one in which the cage of the prosthesis can easily be accommodated in the cavity of the left ventricle. This selection is particularly crucial in patients with mitral stenosis without enlargement of the left ventricle. Most difficulties with the ball-valve prosthesis have resulted from an unduly large prosthesis in which the ventricular septum or the wall of the ventricle impinged on the cage of the prosthesis, resulting in erosion of the ventricular wall or restriction of free motion of the ball with subsequent thrombosis. The disc valve, a “low-profile” prosthesis, has a distinct advantage with a small ventricular cavity and is preferred for such paMedical
Center,
October,
550 First
1968
Ave.,
Vol.
New
York,
76, No.
N. Y. 10016.
4, pp. 576-580
Volume Number
76 1
Surgical treatment of valvular heart disease
tients. However, even with disc valves, protrusion of muscle into the borders of the prosthesis with restriction of motion of the disc has occurred; recent designs have “muscle guards” on the prosthesis to minimize the risk of this complication. For tricuspid replacement, because of the size and contour of the right ventricular cavity, this author has preferred a disc prosthesis, although Starr has reported excellent results with the ball valve. Except for the consideration of choosing a prosthesis of appropriate size, reported experiences with disc and ball-valve prostheses have not demonstrated any superiority of one type over the other. The main problem with prosthetic valves of all types has been thromboembolism.3r4 A significant development in the past year has been the widespread introduction of prostheses which are completely covered with a porous Teflon cloth, including both the metal rim of the prosthesis as well as the metallic struts. In April, 1968, Braunwald and Morrow5 reported that only one embolus had occurred following 55 valve replacements with cloth-covered valves, and Beall noted two emboli after 176 replacements. In a series of 110 patients, Starr found 5 per cent with embolism soon after operation, but only 1 per cent subsequently. If these initial experiences are confirmed, the incidence of embolism with the cloth-covered prostheses will be in the range of 1 to 3 per cent, about one tenth of that with earlier prostheses. At the same time the cloth-covered prosthesis was introduced, the Silastic ball was replaced with a hollow steel ball, primarily because of experience with erosion or fracture of the Silastic ball with aortic prostheses. These factors are considered in more detail in the companion article in this series on Prosthetic cardiac valves. A different type of ball-valve prosthesis was developed by Smeloff, termed the Cutter prosthesis. In this type of prosthesis, the cage mechanism is partly in the atrium and partly in the ventricle, thus limiting the extent to which the cage extends into the ventricular cavity. A “controlled leak” is designed in the prosthesis to permit a small amount of leakage of blood when the ball is seated, hoping by this washing effect during each systolic
577
contraction to minimize the deposition of thrombi. Although this prosthesis has had extensive clinical trial, no consistent advantage has been found over the conventional Starr-Edwards prosthesis. Disc prostheses. In these prostheses a disc, rather than a ball, is used for the mobile, occluding mechanism. With a disc, the cage enclosing the disc is much shorter than the cage in ball valves. Hence, such prostheses are termed “low-profile” prostheses. Their main advantage is that they can be inserted into a ventricular cavity of small size with little risk of impingement on the surrounding ventricular wall or septum. Although the hemodynamic flow patterns are theoretically more desirable than those with a ball valve, the clinical significance of these differences has not been apparent. The incidence of thromboembolism has been similar to that with a ball valve. Theoretically, less force is required to hold a disc valve in position, because the force of itnpact of the disc against the seating ring of the prosthesis is less than that with a ball-valve prosthesis. A disadvantage of the disc prosthesis is that free rotation of the disc may not occur and result in unequal wear, whereas the randomly spinning ball of a ball-valve prosthesis is more immune to wear. Several different types of disc prostheses have been developed, including the Hufnagel prosthesis, the Kay-Shiley prosthesis, the Earle Kay prosthesis, the Cross-Jones prosthesis, and more recently the Beall prosthesis. No significant clinical differences have been reported among these different types. Most of these have recently been modified so that the metallic surfaces are completely covered with cloth. Homograft and heterograft prostheses. With the increasing use of aortic homograft prostheses, considerable effort has been expended to develop mitral homograft prostheses. With the chordae tendineae, however, replacement of the mitral valve with a homograft is much more difficult. In April, 1968, Angel1 and associates6 reported experiences with 70 patients in whom the mitral valve was replaced with a fresh aortic homograft valve, which was attached to a prosthetic ring and then sutured to the mitral amlulus in an in-
578
Spencer
verted position. Similarly, Ionescu and colleagues’ have described 50 patients in whom a heterograft aortic valve from a pig was used in a similar manner. Although initial experiences have been satisfactory, long-term results are not available. Operative indications and hazards Indications. At present, operation should be recommended only for seriously disabled patients, particularly individuals who are unable to work and despite intensive medical therapy are developing progressive signs of congestive heart failure. The reason for restricting prosthetic replacement to advanced forms of heart disease is that the combined operative and late mortality rate following prosthetic replacement of the mitral valve has been about 25 per cent. The 28 per cent average mortality rate in 155 patients reported by Starr has already been mentioned. In 100 patients reported by Morrow and associates,8 the operative death rate was 17 per cent and late deaths were 7 per cent, a total of 24 per cent. With improvements in operative technique, operative risk has decreased to about 10 per cent. Hopefully, the cloth-covered prostheses will decrease the late risk from thromboembolism, but significant longterm data are not yet available. Only when prostheses are available with a negligible risk of thromboembolism should operation be recommended in earlier stages of the disease. Operative hazards. At operation the main hazards are air embolism and hemorrhage from friable, calcified tissues. With increasing experience, deaths from either of these causes have progressively decreased. Following operation, the low cardiac output syndrome has similarly become less frequent, most likely because of careful transfusion following operation to keep left atria1 pressure elevated to 15 to 20 mm. Hg. Arrhythmias are one of the more frequent postoperative complications and require intensive observation and therapy with digitalis, potassium, procaine amide, and selective electrical pacing of the heart with a pacemaker wire implanted at the time of operation. Pulmonary, renal, and hepatic complications of minor degree are frequently seen but
Am. Heart .I. Ortobcr, 1968
are seldom fatal. Thromboembolism may occur in the early postoperative course, but again is infrequent if anticoagulation is begun four to five days after operation. Infection has fortunately become rare with the intensive use of large amounts of bactericidal antibiotics during and following operation. Most of the patients receive Staphcillin in large amounts during operation and for one month afterward. In the past six years, infection of the prosthetic valve (organism : Candida ulbicans) occurred in only one patient. long-term prognosis and management It is important to realize that the mortality risk in the two years following discharge from the hospital is about equal to that associated with operation. This well emphasizes the importance of careful long-term medical management following prosthetic replacement, considering surgical replacement of a mitral valve as only one phase in the treatment of the patient with valvular disease. Many causes of late deaths can probably be avoided by early detection and treatment of complications. In the report by Starr, about one half of the late deaths were due to the prosthesis, either from thromboembolism, detachment of the prosthesis, or other complications. In about one half of the patients, however, death was apparently due to an arrhythmia or to myocardial disease, for only myocardial fibrosis was found at postmortem examination. In seven late deaths reported by Morrow and associates, two were due to infection of the prosthetic valve and all others were related to malfunction of the prosthesis. In long-term management of the patient with a prosthetic mitral valve, three areas of therapv should be considered: 1. Ant&oapZation. This author’s preference is to maintain anticoagulation permanently, administering Coumadin (sodium warfarin) daily to keep the prothrombin time about twice the normal rate. Anticoagulant therapy is not without its hazards, however, because of the risk of hemorrhage. One death has occurred in our patients in the last two years, directly related to sodium warfarin therapy, primarily from negligence on the part of the
Surgical treatment
patient. Major hemorrhage occurred in two other patients, however, despite careful medical observation. Fortunately, these subsided without serious sequelae in each instance. With the new cloth-covered prostheses, it may be found possible to stop anticoagulation after several months, at which time the prostheses should presumably be covered with a new endothelium. When a surgical operation is performed, such as a cholecystectomy, anticoagulant therapy must be stopped for a short period of time. Our preference has been to give heparin until a few hours before operation and then to resume heparin one to two days afterward. 2. Antcbiotk therapy. Antibiotic therapy, usually sodium methicillin (Staphcillin), is given for four to five weeks following operation. With this program bacterial infections have not occurred in the past several years. Subsequently, antibiotics are recommended if surgical procedures are performed which might induce a transient bacteremia, such as extraction of a tooth or a cystoscopic examination. Fatal endocarditis developing after such procedures has been reported by several investigators. 3. Cardiac function. The degree of recovery of cardiac function following prosthetic mitral replacement must be carefully evaluated by the physician over a period of several months. Such evaluation should include the presence of symptoms, regression of heart size and signs of ventricular hypertrophy on the electrocardiogram, and tolerance for the intake of sodium. An attempt should be made to correct chronic atria1 fibrillation by electrical cardioversion a few weeks following operation. In patients who have been fibrillating for long periods of time, a permanent conversion to sinus rhythm can be obtained in about 30 to 35 per cent. In the group of patients reported by Starr in 1967, 65 per cent were free of all cardiac symptoms and taking a regular diet, and roentgenograms had showed a dramatic reduction in cardiac size. Twenty-five per cent had mild restriction of exercise ability and did not require salt restriction. Ten per cent had persistent cardiomegaly and required permanent intensive medical ther-
qf vnlv~dar
heart disease
579
apy, obtaining little benefit from operation. In the 76 patients surviving operation reported by Morrow and associates,8 47 were classified as Class I, 26 as Class II, and three as Class III. The reason for the variation in results is not known. It may be that some patients have a superimposed myocardial disease (a rheumatic myocarditis?), but in any event, the response following operation is not a uniform one and must be determined by the physician over a long period of time. If cardiac symptoms persist following operation, several possibilities should be evaluated. The most important is whether mitral insufficiency has developed from a perivalvular leak around the rim of the prosthesis, usually due to one of the sutures anchoring the prosthesis cutting through the fibrous tissue of the residual annulus. Such a diagnosis can be confirmed or excluded only by angiography of the left ventricle. It is surprising that in some patients a significant leak can be present without an audible murmur. Hence, repeat cardiac catheterization and angiography should be performed in any patient with persistent cardiac symptoms. Other possibilities include the presence of additional valvular disease, such as aortic or tricuspid valvular disease. Rarely, a serious arrhythmia can precipitate cardiac failure in the absence of other findings. An increased pulmonary vascular resistance almost always subsides following mitral replacement and is rarely a cause of persistent difficulty. Finally, when other possibilities have been excluded, the possibility of intrinsic disease of the left ventricular muscle must be considered. This is usually characterized by persistent cardiomegaly, with a left ventricular enddiastolic pressure above 15 mm. Hg, and absence of any signs of mitral regurgitation. Because of the varying response of different patients to operation, the salt intake and the physical exertion of the patient should be carefully evaluated by the physician for several months following operation. This author personally has seen more than one patient in whom a sudden change in physical activity, or an excessive salt intake several months fol-
580
.4,x.
Spencer
lowing operation, cardiac failure.
Heart
October,
precipitated
edema and
Summary
Extensive experience has been gained with prosthetic replacement of the mitral valve since it was first introduced by Starr in 1961. Operative mortality following prosthetic replacement has decreased to about 10 per cent, but 10 to 15 per cent of patients subsequently die in the first two years following operation, either from complications related to the prosthesis or from their underlying cardiac disease. Hence, at present, operation should be recommended only for patients with progressive disability from valvular disease. Following mitral replacement, about 65 per cent of patients become asymptomatic, 2.5 per cent have mild symptoms, and for unknown reasons about 10 per cent remain with significant disability. A potentially very significant advance has been made in the past year with the introduction of cloth-covered prostheses, which from presently available data indicate that the incidence of thromboembolism may be decreased from the range of 20 to 25 per cent to as low as 2 to 4 per cent. This advance, if confirmed, ma> greatly liberalize future indications for
1.
1968
mitral valve replacement. Only preliminary experiences are yet available with homograft and heterograft prostheses. REFERENCES M. L.: Mitral replace1. Starr, A., and Edwards, ment: Clinical experience with a ball-valve prosthesis, Ann. Surg. 154:726,1961. 2. Starr, A., Herr, R. H., and Wood, J. A.: hlitral replacement. Review of six years’ experience, J. Thoracic & Cardiovas. Surg. 54:333, 1967. 3. Spencer, F. C., Cortes, L., Marcarenhas, G., Ifuku, M., and Koepke, J.: Mechanism of thrombus formation upon Starr-Edwards prosthetic mitral valves, Ann. Surg. 165:814, 1967. N. S.. and Bonchek. I,. I.: Preven4. Braunwald. tion of thrombus formation on rigid prosthetic heart valves by ingrowth of autogenous tissue. J. Thoracic &- Cacdiovas. Surg. 54:630, 1967: 3. Braunwald. N. S.. and Morrow, ii. G.: Tissue ._ ingrowth and the’rigid heart valve: Review of clinical and experimental experience during the past year, J. Thoracic & Card&as. Surg. 56:307, 1968. bV. W., Stinson, E. B., Iben, A. B.. and 6. Angel], Shumway, N. E.: Multiple valve replacement with the fresh aortic homograft, J. Thoracic & Cardiovas. Surg. 56:323, 1968. lonescu, hl. I., Wooler, G. H., Whitaker, \V., Smith, D. Ii., Taylor, S. H., and Hargreaves, M. D.: Heart valve replacement with reinforced aortic heterografts: Technique and results J. Thoracic & Cardiovas. Surg. 56:333, 1968. Morrow, A. G., Oldham, H. N., Elkins, Ii. C., and Braunwald, E.: Prosthetic replacement of the mitral valve. Preoperative and postoperative clinical and hemodynamic assessments in 100 patients, Circulation 35:962, 1967.