- Email: [email protected]
Clinical experience with the Lillehei-Kaster cardiac valve prosthesis
Clinical experience with the Lillehei-Kaster cardiac valve prosthesis This paper reviews our experience with the Lillehei-Kaster pivoting disc prosthesis in 155 patients with aortic and mitral valve disease. We employed 189 valves during the period 1971 to 1974. The early surgical mortality rates for isolated mitral, isolated aortic, and combined mitral and aortic valve replacements were 9 per cent, 14 per cent, and 3 per cent, respectively. Postoperatively, there was no evidence of significant hemolysis, and the gradients across the prostheses were satisfactory. Clinical evaluation of prosthetic function was made difficult by the infrequency of an opening click and the common occurrence of mid-diastolic murmurs even with minimal gradients. The most disturbing complication, which has led to our abandoning use of this valve, was thrombosis, which occurred in at least 10 per cent of the mitral and 5 per cent of the aortic valves. The cause is thought to be late prosthetic disproportion as the heart shrinks in size.
Abdul S. Mitha, M.R.C.P. (Lond.)., Rodney E. Matisonn, M . B . , Ch.B., F.C.P. (S.A.), Bernard T. le Roux, Ch.M. (Cape Town), F.R.C.S. (Edin.), and Elliot Chesler, M.D., F.R.C.P., F . A . C . C , Durban, South Africa
r rom the paucity of available literature, experience with the Lillehei-Kaster pivoting disc appears to be less extensive than that with the various other prosthetic cardiac valves. 1 ' 2 The valve has been in clinical use since 1970 and was designed to eliminate thromboembolism, diminish obstruction to flow, diminish hemolysis, and avoid premature degradation. Both the aortic and mitral Lillehei-Kaster valves comprise a Pyrolite* disc suspended in a weldless titanium housing encircled by a Teflon fabric sewing ring. In the closed position the disc is inclined to 18 degrees with the annular valve housing, and in the open position to 80 degrees. The disc-retaining struts arise diagonally from the outflow area of the valve housing. Pulse duplicator studies have indicated favorable hemodynamic characteristics for the valve, and autopsy studies in animals have shown secure healing and absence of thrombus and valve wear. 1 From the Cardiac and Cardio-Thoracic Units, Wentworth Hospital, and the University of Natal, Durban, South Africa. Supported by a grant from the Medical Research Council of South Africa. Received for publication April 9, 1976. Accepted for publication June 11. 1976. Address for reprints: Professor E. Chesler, Cardiac Unit, Wentworth Hospital. P.B. Jacobs, Natal. 4026. *Pyrolite is the registered trademark of Gulf Energy and Environmental Systems Co.
This paper reports our clinical experience with 189 Lillehei-Kaster valves in 155 patients with aortic and mitral valve disease during the period 1971 to 1974. Isolated mitral valve replacement The mitral valve was replaced in 75 patients. Their mean age was 33 years ( ± 15.9) and 66 per cent of the patients were younger than 40 years old. The majority were black. Fifty were women and 25 men. According to the New York Heart Association classification, all the patients were placed preoperatively into Functional Classes III and IV. Cardiac catherizations were performed in 28 patients before operation and the average mean pulmonary artery pressure was 49 mm. Hg. Thirteen patients had associated, hemodynamically insignificant aortic valve disease and 43 (57 per cent) had established atrial fibrillation. Most of the patients had rheumatic valvular disease. Of 7 with mitral incompetence, 4 had spontaneous rupture of the chordae tendineae, one had infective endocarditis, and 2 had myxomatous change (Table I). The sizes of the 77 prostheses used are given in Table II. Surgical results Early deaths. There were seven deaths within the first 30 days of operation resulting in a mortality rate of 9 per cent (Table III). One patient died within 24 hours 401
The Journal of Thoracic and Cardiovascular Surgery
4 0 2 Mitha et al.
Table I. Isolated mitral valve replacement: Clinico-pathological material
Table III. Isolated mitral valve replacement: Early postoperative deaths in 75 patients No.
Cause
Mitral stenosis and incompetence (previous valvotomy 19) Calcific mitral stenosis Nonrheumatic mitral incompetence Atrial septal defect and mitral incompetence Periprosthetic mitral incompetence (Starr-Edwards valve) Thrombosed Cutter valve
60
Thrombosed prostheses Intraoperative cerebral embolism Postoperative myocardial infarction Uncontrolled arrhythmias Low output Mediastinitis from persistent wound infection
1 1 1 1 2 1
Total
7 (9%)
Total
75
5 7 1 1 1
Table IV. Isolated mitral valve replacement: Late deaths
Table II. Isolated mitral valve replacement: Sizes of the prosthetic valves Valve size
No. of patients
16 18 20 22 25
4 13 20 19 21
Total
No. of patients
77
of operation and autopsy demonstrated a thrombosed prosthesis. Late deaths. There were ten late deaths among the 68 patients who survived the operation, and the causes are given in Table IV. Two patients died suddenly and 2 died of pulmonary edema; no autopsy was performed and the cause is undetermined. Frequency of thrombosis of the prosthesis. This occurred in 7 patients and involved eight of 77 valves (10 per cent). Three of these valves were size 20, three were size 22, and two were size 25. Four patients were admitted with acute pulmonary edema including a man whose prostheses thrombosed twice despite permanent anticoagulation therapy; all but one survived emergency operations in which inverted homograft aortic valves were inserted. One patient had systemic embolism and was also found to have thrombus on the valve at a second operation. Two patients died suddenly at home, and medicolegal autopsy demonstrated thrombosis of the prostheses (Fig. 1). Of the 77 valves there were thus nine thrombosed prostheses, one early and eight late (12 per cent), and at least four deaths, one early and three late (5 per cent), attributable to this cause. Thromboembolism. A minor cerebral embolus occurred in one patient who made a satisfactory recovery. Periprosthetic mitral insufficiency. This occurred
in 68
survivors Cause
No.
Thrombosed prosthesis Sudden (undetermined cause) Pulmonary edema (undetermined cause) Myocardial infarction Renal failure
2 2 3 2 1
Total
10(15r/r)
in 3 patients and was sufficiently severe to require a second operation in 2 of them. Postoperative myopathy. Three patients had impaired left ventricular function with ejection fractions of less than 35 per cent. Two have since died of intractable cardiac failure. Hemolysis. Clinical hemolysis has been insignificant and only one patient required iron and folic acid to correct anemia. Postoperative auscultatory findings. In only 3 of the 68 patients could an opening click be heard; the closing click was invariably single. A mid-diastolic murmur was detected in 20, and these findings were confirmed by phonocardiograms (Fig. 2). There was no correlation between the occurrence of a middiastolic murmur and the magnitude of the gradient across the prosthesis. Postoperative hemodynamic evaluation. Cardiac catheterization and cineangiography was performed in 25 patients 3 to 9 months after operation, and the findings are detailed in Table V. On the basis that the Gorlin formula usually underestimates the area of a prosthetic valve by 20 per cent, the values obtained are satisfactory. Isolated aortic valve replacement Aortic valve replacement with the Lillehei-Kaster valve was performed in 49 patients during the same period; 34 were men, 15 were women, and their
Volume 72 Number 3 September, 1976
Lillehei-Kaster
prosthesis
403
Fig. 1. /I, Interior of left atrium (Z*4) viewed from above showing organized thrombus on the Lillehei-Kaster prosthesis (LK). B, Section through heart showing organized thrombus occluding the disc of the prosthesis. Ao, aorta. LV and RV, left and right ventricles, respectively. average age was 44 years. Ninety per cent of the patients were placed in Group IV of the New York Heart Association classification, and there was an approximately equal distribution of aortic stenosis and aortic incompetence. The clinicopathological features are detailed in Table VI. Surgical results Early deaths. In the uncomplicated group, there were two deaths; one patient died of intractable ventricular fibrillation and the other could not be weaned from cardiopulmonary bypass. In the group of 10 complicated cases there were five deaths, resulting from persistent infective endocarditis, ventricular fibrillation, respiratory failure, progressive aortic dissec-
tion, and intractable hemmorrhage, respectively. There were no deaths in those who had a previous operation, and the over-all mortality rate was 14 per cent (Table VI). Late deaths and complications. Among the 42 survivors there were two deaths (5 per cent), both the result of thrombosis of the prostheses. The mean follow-up period was 14 months. One patient had a minor cerebral embolus with resultant hemiparesis. Hemolysis was not a complication in any of the survivors. Postoperative hemodynamic evaluation. Six of the 42 patients underwent cardiac catherization, and the gradients as assessed by transeptal puncture are given in Table VII.
404
The Journal of Thoracic and Cardiovascular Surgery
Mitha et al.
Fig. 2. Simultaneously recorded electrocardiogram (ECG) and phonocardiogram (Phono) from an asymptomatic patient with a Lillehei-Kaster valve, showing initial low frequency sound "M" followed by a single closing sound (C.C). The opening click (O.C.) is followed by a mid-diastolic rumble (MDM). A2 and P2, Aortic and pulmonic components of the second sound, respectively. Table V. Isolated mitral valve replacement: Valve size
No. of patients
25 22 20 16
7 7 9 2
Postoperative
Mean gradient (mm. Hg) 7 6.57 8.33 9
± ± ± ±
hemodynamic findings in 25 patients Mean valve area (sq. cm.) 2.33 1.87 1.81 1.45
3.2 2.7 5.2 4.24
Combined mitral and aortic valve replacement There were 31 patients (15 men, and 16 women) and their average age was 31 years. All the patients were in Grade IV of the New York Heart Association classification. Early deaths. There was one early death from intraoperative cerebral damage, resulting in an operative mortality rate of 3 per cent. Late deaths and complications. Among the 30 survivors there were eight deaths of which four were sudden, the causes were undetermined because autopsy was not performed (Table VIII). The causes of death in the remaining four instances were bacterial endocarditis, pulmonary embolism, postoperative paraplegia, and a thrombosed mitral prosthesis, respectively. The latter occurred in a patient in whom the mitral prosthesis thrombosed twice; attempted replacement of the second prosthesis with a homograft was unsuccessful. Significant hemolysis occurred in one patient and another sustained major cerebral embolism. Anticoagulant therapy Of the entire series of 155 patients, there were 140 survivors after operation. The majority of these patients were black people living in rural areas under poor socioeconomic circumstances, so that anticoagulant control was impossible. Sixteen patients living in urban areas were maintained on permanent anticoagulation therapy, and thrombosis of the prosthesis occurred in 4
± ± ± ±
.74 .42 .81 .07
In vitro area (sq. cm.) 4.91 3.8 3.14 2.01
of them (25 per cent). Among the remaining 124 patients who did not receive anticoagulants, there were seven instances of thrombosed prostheses (6 per cent). Of the 11 thrombosed prostheses, the mitral was involved in 9 and the aortic in 2 patients, respectively. Comment Lillehei and associates 1, 2 recently reported their results in 139 patients who underwent heart valve replacement with the Lillehei-Kaster prosthesis. Seventy patients had isolated aortic valve replacement, 9 of whom (14.3 per cent) died within the first month. During the next 24 months 5 patients died, and no deaths were attributed to malfunction of the valve or thromboembolism. Forty-nine patients had their mitral valve replaced, and 3 (6.1 per cent) died in the first postoperative month. An additional 7 patients (14 per cent) died during a follow-up period of 2 to 24 months. Except for one patient who died after a second operation for replacement of a malfunctioning prosthesis, improperly positioned at the first operation, no late postoperative deaths were related to valve malfunction or thromboembolism. In both the aortic and mitral positions, intravascular hemolysis was minimal and measured gradients and calculated valve areas at catherterization were satisfactory. In terms of absence of significant intravascular hemolysis and satisfactory hemodynamics assessed at cardiac catherization, our results are similar (Tables V
Volume 72 Number 3 September, 1976
Lillehei-Kaster prosthesis
405
Table VI. Isolated aortic valve replacement: Clinicopathological material No.
Operative deaths
Uncomplicated AVD Complicated cases AS + VSD AI (+ aneurysm of aorta, ostial stenosis) Impaired function of LV Chronic lung disease Aneurysm of sinus of Valsalva Dissecting aneurysm of aorta Valsalva fistula into RV Renal failure Previous surgery Starr-Edwards valve with periprosthetic leak Starr-Edwards valve with hemolysis Leaking homograft Starr-Edwards with SBE
34
2
Totals
49
1 1 1 1 1 1 3 1 1 1 2 1
■itiS-.
5
0
7 (14%)
Legend: AVD, Aortic valve disease. AS. Aortic stenosis. VSD, Ventricular septal defect. AI, Aortic insufficiency. LV, Left ventricle. RV, Right ventricle. SBE, Subacute bacterial endocarditis.
Table VII. Isolated aortic valve replacement: Hemodynamic data of 6 patients
Fig. 3. Left ventricular cineangiogram, right anterior oblique projection, showing afillingdefect produced by thrombus (T) below the Lillehei-Kaster prosthesis. Ao, Aorta. LV, Left ventricle. Table VIII. Combined aortic and mitral valve replacement: Late deaths in 30 survivors
Valve size
PSG (mm. Hg)
Valve area (sq. cm.)
Cause
22 20 20 18 16 16
4 15 19 12 23 40
2.4 1.2 1.9 1.1 1.03 0.93
Thombosis of mitral prosthesis Sudden undetermined Infective endocarditis Pulmonary embolism Paraplegia
Total
No.
: (27%)
Legend: PSG, Pulmonary-systemic gradient.
and VII). The short follow-up period of 28 months precludes any comment on the frequency of systemic thromboembolism. In our series, the incidence of definite prosthetic thrombosis was 10 per cent in the mitral position and 5 per cent in the aortic position. However, if those patients who died suddenly and in who autopsies were not performed are included, the incidence could be much higher. Because of these disappointing results the use of the prosthesis was abandoned in 1974. Clinical evaluation of the Lillehei-Kaster prosthesis postoperatively is difficult, and valve dysfunction is not readily detected by auscultation. In contrast to cagedball or non-tilting disc mitral valve prostheses, which have clearly defined opening and closing sounds on auscultation, we could detect and document opening clicks in only 4 per cent of normally functioning mitral
prostheses. Additionally, mid-diastolic murmurs were audible in 30 per cent of cases, irrespective of the gradient across the valve; persumably this murmur is related to turbulence around the disc. Similar findings have been reported by other observers. In our experience the closing click is invariably single (Fig. 2). Gibson and associates3 reported that the closing sound is always twofold, with a small initial high-frequency component and a second high-frequency component separated by less than 0.03 second. Previous work has clearly demonstrated, however, that this initial component of the first heart sound ("M") is present in normal subjects and also in patients with Starr-Edwards prosthetic valves whether they are in sinus rhythm or atrial filbrillation.4 The nature of origin of " M , " which occurs at the time of crossover of the left atrial and left ventricular pressures,5 is uncertain, but it is probably
4 0 6 Mitha et al.
The Journal of Thoracic and Cardiovascular Surgery
Fig. 4. Chest roentgenograms in an asymptomatic patient with a Lillehei-Kaster prosthesis showing dramatic change in heart size. A, Preoperative. B and C, Six weeks and 28 months after operation, respectively. produced by resonance of the myocardium and atrioventricular ring and is unrelated to prosthetic events. The presentation of cases of abrupt valve dysfunction was typically one of acute pulmonary edema or congestive cardiac failure with marked tachycardia or uncontrolled atrial fibrillation. Because of the difficulty in assessing valve dysfunction by auscultation, we found chest roentgenograms particularly helpful in diagnosis by demonstrating increasing cardiomegaly and, in particular, sudden left atrial enlargement and pulmonary venous hypertension. Fluoroscopy is of no value because the disc is not radiopaque. Cardiac catherization and cineangiography demonstrated significant gradients across the prosthesis, with filling defects produced by thromboses on the atrial and ventricular aspects of the valve (Fig. 3). The problem of dysfunction of disc prostheses has been reviewed recently by Roberts and colleagues.6 Important factors are (1) intrinsic stenosis of the prosthesis, (2) inadequate space in the ventricles or
ascending aorta to freely accomodate the prosthesis when disproportion exists or develops, (3) thrombus formation on the prostheses, and (4) prosthetic variance. It would appear from our findings that prosthetic dysfunction was in no way related to anticoagulant therapy. Prosthetic thrombosis occurred more frequently in those patients who received anticoagulants (25 per cent) than in those who did not (4 per cent). Prosthetic disproportion, that is a prosthesis which is too large for the aortic or ventricular cavity, is probably the most important mechanism pertinent to our series. Most of the mitral prostheses used were large—60 of the 77 prostheses were sizes 20 to 25, and these were the valves involved by thrombosis. The smaller prostheses (sizes 16 and 18) were not affected in this manner. It is likely that disproportion was present, with minimal impingement and cocking of the disc on the ventricular endocardium resulting in thrombosis of the prosthesis and the left atrium. Apart from one case in which thrombosis occurred within 24 hours of operation, it is extremely improbable that dispropor-
Volume 72 Number 3 September, 1976
tion and cocking of the disc was present at the conclusion of the operation. In the majority, some other factor must have been responsible for the disproportion. Most of our patients are young, having had virulent, acute rheumatic fever at an early age. Their hearts are usually markedly dilated, and we have frequently observed dramatic shrinkage in heart size following correction of the hemodynamics 7 (Fig. 4). It is probable that disproportion is related to this change, with resultant impingement of the prosthetic disc on the endocardium of the left ventricle as the chamber decreases in size. Other factors, such as unsuspected severe aortic incompetence with the regurgitant jet interfering with disc opening, were not present. 6 It is now well appreciated that the caged-ball type of prosthesis may obstruct left ventricular outflow or inflow after valve replacement, and thrombus formation on these valves is not uncommonly found at autopsy. 8 Usually the amount of thrombus on the struts or primary orifice is insufficient to interfere with ball or poppet movement. Prostheses of the disc type, however, are much more vulnerable to abrupt valve dysfunction with an immobile disc than is the caged-ball variety. Our disturbing experience cannot be attributed to faulty valve design but more appropriately to failure to recognize the potential for subsequent disproportion. Deliberate insertion of the smaller prostheses may avoid this complication, particularly in those cases in which the left ventricle is dilated and volume overloaded from mitral or aortic incompetence. REFERENCES I Lillehei, C. W., Kaster, R. L., Coleman, M., and Bloch, J. H.: Heart Valve Replacement with the Lillehei-Kaster
Lillehei-Kaster
2
3
4 5
6
7
8
prosthesis
407
Pivoting Disc Prosthesis, N.Y. State J. Med 74: 1426, 1974. Lillehei, C. W., Carlson, R. G., and Kaster, R. L., et al.: Heart Valve Replacement with the Lillehei-Kaster Pivoting Disc Prosthesis, Presented before the Society for Vascular Surgery, Carmel, Calif., June, 1972. Cited by Brawley, R. K., Donahoo, J. D., andGott, V. L.: Current Status of the Beall, Bjork-Shiley, Braunwald-Cutter Cardiac Valve Prostheses, Am. J. Cardiol. 35: 855, 1975. Gibson, T. C , Starek, P. J. K., Roos, S., and Craige, E.: Echocardiographic and Phonocardiographic Characteristics of the Lillehei-Kaster Mitral Valve Prostheses, Circulation 49: 434, 1974. Armstrong, T. G., and Gotsman, M. S.: Initial Low Frequency Vibrations of the First Heart Sound, Br. Heart J. 35: 691, 1973. Lakier, J. B., Pocock, W. A., Gale, G. E., and Barlow, J. B.: Haemodynamic and Sound Events Preceding First Heart Sound in Mitral Stenosis, Br. Heart J. 34: 1 152, 1972. Roberts, W. C , Fishbein, M. C , and Golden, A.: Cardiac Pathology after Valve Replacement by Disc Prosthesis. A Study of 61 Necropsy Patients, Am. J. Cardiol. 35: 740, 1975. Van der Horst, R. L., Joshi, M. A., le Roux, B. T., and Rogers, N. M. A., and Gotsman, M. S.: The Chest X-ray After Mitral Valve Replacement in Childhood, S. Afr. Med. J. 46: 1933, 1972. Kalke, B., Korns, M. E., Goott, B., Lillehei, C. W., and Edwards, J. E.: Engagement of Ventricular Myocardium by Open-Cage Atrioventricular Valvular Prosthesis, J. THORAC. CARDIOVASC. SURG. 58: 92,
1969.
9 Roberts, W. C , Buckely, B. A., and Morrow, A. C : Pathologic Anatomy of Cardiac Valve Replacement: A Study of 24 Necropsy Patients, Progr. Cardiovasc. Dis. 15: 539, 1973.
Abdul S. Mitha, M.R.C.P. (Lond.)., Rodney E. Matisonn, M . B . , Ch.B., F.C.P. (S.A.), Bernard T. le Roux, Ch.M. (Cape Town), F.R.C.S. (Edin.), and Elliot Chesler, M.D., F.R.C.P., F . A . C . C , Durban, South Africa
r rom the paucity of available literature, experience with the Lillehei-Kaster pivoting disc appears to be less extensive than that with the various other prosthetic cardiac valves. 1 ' 2 The valve has been in clinical use since 1970 and was designed to eliminate thromboembolism, diminish obstruction to flow, diminish hemolysis, and avoid premature degradation. Both the aortic and mitral Lillehei-Kaster valves comprise a Pyrolite* disc suspended in a weldless titanium housing encircled by a Teflon fabric sewing ring. In the closed position the disc is inclined to 18 degrees with the annular valve housing, and in the open position to 80 degrees. The disc-retaining struts arise diagonally from the outflow area of the valve housing. Pulse duplicator studies have indicated favorable hemodynamic characteristics for the valve, and autopsy studies in animals have shown secure healing and absence of thrombus and valve wear. 1 From the Cardiac and Cardio-Thoracic Units, Wentworth Hospital, and the University of Natal, Durban, South Africa. Supported by a grant from the Medical Research Council of South Africa. Received for publication April 9, 1976. Accepted for publication June 11. 1976. Address for reprints: Professor E. Chesler, Cardiac Unit, Wentworth Hospital. P.B. Jacobs, Natal. 4026. *Pyrolite is the registered trademark of Gulf Energy and Environmental Systems Co.
This paper reports our clinical experience with 189 Lillehei-Kaster valves in 155 patients with aortic and mitral valve disease during the period 1971 to 1974. Isolated mitral valve replacement The mitral valve was replaced in 75 patients. Their mean age was 33 years ( ± 15.9) and 66 per cent of the patients were younger than 40 years old. The majority were black. Fifty were women and 25 men. According to the New York Heart Association classification, all the patients were placed preoperatively into Functional Classes III and IV. Cardiac catherizations were performed in 28 patients before operation and the average mean pulmonary artery pressure was 49 mm. Hg. Thirteen patients had associated, hemodynamically insignificant aortic valve disease and 43 (57 per cent) had established atrial fibrillation. Most of the patients had rheumatic valvular disease. Of 7 with mitral incompetence, 4 had spontaneous rupture of the chordae tendineae, one had infective endocarditis, and 2 had myxomatous change (Table I). The sizes of the 77 prostheses used are given in Table II. Surgical results Early deaths. There were seven deaths within the first 30 days of operation resulting in a mortality rate of 9 per cent (Table III). One patient died within 24 hours 401
The Journal of Thoracic and Cardiovascular Surgery
4 0 2 Mitha et al.
Table I. Isolated mitral valve replacement: Clinico-pathological material
Table III. Isolated mitral valve replacement: Early postoperative deaths in 75 patients No.
Cause
Mitral stenosis and incompetence (previous valvotomy 19) Calcific mitral stenosis Nonrheumatic mitral incompetence Atrial septal defect and mitral incompetence Periprosthetic mitral incompetence (Starr-Edwards valve) Thrombosed Cutter valve
60
Thrombosed prostheses Intraoperative cerebral embolism Postoperative myocardial infarction Uncontrolled arrhythmias Low output Mediastinitis from persistent wound infection
1 1 1 1 2 1
Total
7 (9%)
Total
75
5 7 1 1 1
Table IV. Isolated mitral valve replacement: Late deaths
Table II. Isolated mitral valve replacement: Sizes of the prosthetic valves Valve size
No. of patients
16 18 20 22 25
4 13 20 19 21
Total
No. of patients
77
of operation and autopsy demonstrated a thrombosed prosthesis. Late deaths. There were ten late deaths among the 68 patients who survived the operation, and the causes are given in Table IV. Two patients died suddenly and 2 died of pulmonary edema; no autopsy was performed and the cause is undetermined. Frequency of thrombosis of the prosthesis. This occurred in 7 patients and involved eight of 77 valves (10 per cent). Three of these valves were size 20, three were size 22, and two were size 25. Four patients were admitted with acute pulmonary edema including a man whose prostheses thrombosed twice despite permanent anticoagulation therapy; all but one survived emergency operations in which inverted homograft aortic valves were inserted. One patient had systemic embolism and was also found to have thrombus on the valve at a second operation. Two patients died suddenly at home, and medicolegal autopsy demonstrated thrombosis of the prostheses (Fig. 1). Of the 77 valves there were thus nine thrombosed prostheses, one early and eight late (12 per cent), and at least four deaths, one early and three late (5 per cent), attributable to this cause. Thromboembolism. A minor cerebral embolus occurred in one patient who made a satisfactory recovery. Periprosthetic mitral insufficiency. This occurred
in 68
survivors Cause
No.
Thrombosed prosthesis Sudden (undetermined cause) Pulmonary edema (undetermined cause) Myocardial infarction Renal failure
2 2 3 2 1
Total
10(15r/r)
in 3 patients and was sufficiently severe to require a second operation in 2 of them. Postoperative myopathy. Three patients had impaired left ventricular function with ejection fractions of less than 35 per cent. Two have since died of intractable cardiac failure. Hemolysis. Clinical hemolysis has been insignificant and only one patient required iron and folic acid to correct anemia. Postoperative auscultatory findings. In only 3 of the 68 patients could an opening click be heard; the closing click was invariably single. A mid-diastolic murmur was detected in 20, and these findings were confirmed by phonocardiograms (Fig. 2). There was no correlation between the occurrence of a middiastolic murmur and the magnitude of the gradient across the prosthesis. Postoperative hemodynamic evaluation. Cardiac catheterization and cineangiography was performed in 25 patients 3 to 9 months after operation, and the findings are detailed in Table V. On the basis that the Gorlin formula usually underestimates the area of a prosthetic valve by 20 per cent, the values obtained are satisfactory. Isolated aortic valve replacement Aortic valve replacement with the Lillehei-Kaster valve was performed in 49 patients during the same period; 34 were men, 15 were women, and their
Volume 72 Number 3 September, 1976
Lillehei-Kaster
prosthesis
403
Fig. 1. /I, Interior of left atrium (Z*4) viewed from above showing organized thrombus on the Lillehei-Kaster prosthesis (LK). B, Section through heart showing organized thrombus occluding the disc of the prosthesis. Ao, aorta. LV and RV, left and right ventricles, respectively. average age was 44 years. Ninety per cent of the patients were placed in Group IV of the New York Heart Association classification, and there was an approximately equal distribution of aortic stenosis and aortic incompetence. The clinicopathological features are detailed in Table VI. Surgical results Early deaths. In the uncomplicated group, there were two deaths; one patient died of intractable ventricular fibrillation and the other could not be weaned from cardiopulmonary bypass. In the group of 10 complicated cases there were five deaths, resulting from persistent infective endocarditis, ventricular fibrillation, respiratory failure, progressive aortic dissec-
tion, and intractable hemmorrhage, respectively. There were no deaths in those who had a previous operation, and the over-all mortality rate was 14 per cent (Table VI). Late deaths and complications. Among the 42 survivors there were two deaths (5 per cent), both the result of thrombosis of the prostheses. The mean follow-up period was 14 months. One patient had a minor cerebral embolus with resultant hemiparesis. Hemolysis was not a complication in any of the survivors. Postoperative hemodynamic evaluation. Six of the 42 patients underwent cardiac catherization, and the gradients as assessed by transeptal puncture are given in Table VII.
404
The Journal of Thoracic and Cardiovascular Surgery
Mitha et al.
Fig. 2. Simultaneously recorded electrocardiogram (ECG) and phonocardiogram (Phono) from an asymptomatic patient with a Lillehei-Kaster valve, showing initial low frequency sound "M" followed by a single closing sound (C.C). The opening click (O.C.) is followed by a mid-diastolic rumble (MDM). A2 and P2, Aortic and pulmonic components of the second sound, respectively. Table V. Isolated mitral valve replacement: Valve size
No. of patients
25 22 20 16
7 7 9 2
Postoperative
Mean gradient (mm. Hg) 7 6.57 8.33 9
± ± ± ±
hemodynamic findings in 25 patients Mean valve area (sq. cm.) 2.33 1.87 1.81 1.45
3.2 2.7 5.2 4.24
Combined mitral and aortic valve replacement There were 31 patients (15 men, and 16 women) and their average age was 31 years. All the patients were in Grade IV of the New York Heart Association classification. Early deaths. There was one early death from intraoperative cerebral damage, resulting in an operative mortality rate of 3 per cent. Late deaths and complications. Among the 30 survivors there were eight deaths of which four were sudden, the causes were undetermined because autopsy was not performed (Table VIII). The causes of death in the remaining four instances were bacterial endocarditis, pulmonary embolism, postoperative paraplegia, and a thrombosed mitral prosthesis, respectively. The latter occurred in a patient in whom the mitral prosthesis thrombosed twice; attempted replacement of the second prosthesis with a homograft was unsuccessful. Significant hemolysis occurred in one patient and another sustained major cerebral embolism. Anticoagulant therapy Of the entire series of 155 patients, there were 140 survivors after operation. The majority of these patients were black people living in rural areas under poor socioeconomic circumstances, so that anticoagulant control was impossible. Sixteen patients living in urban areas were maintained on permanent anticoagulation therapy, and thrombosis of the prosthesis occurred in 4
± ± ± ±
.74 .42 .81 .07
In vitro area (sq. cm.) 4.91 3.8 3.14 2.01
of them (25 per cent). Among the remaining 124 patients who did not receive anticoagulants, there were seven instances of thrombosed prostheses (6 per cent). Of the 11 thrombosed prostheses, the mitral was involved in 9 and the aortic in 2 patients, respectively. Comment Lillehei and associates 1, 2 recently reported their results in 139 patients who underwent heart valve replacement with the Lillehei-Kaster prosthesis. Seventy patients had isolated aortic valve replacement, 9 of whom (14.3 per cent) died within the first month. During the next 24 months 5 patients died, and no deaths were attributed to malfunction of the valve or thromboembolism. Forty-nine patients had their mitral valve replaced, and 3 (6.1 per cent) died in the first postoperative month. An additional 7 patients (14 per cent) died during a follow-up period of 2 to 24 months. Except for one patient who died after a second operation for replacement of a malfunctioning prosthesis, improperly positioned at the first operation, no late postoperative deaths were related to valve malfunction or thromboembolism. In both the aortic and mitral positions, intravascular hemolysis was minimal and measured gradients and calculated valve areas at catherterization were satisfactory. In terms of absence of significant intravascular hemolysis and satisfactory hemodynamics assessed at cardiac catherization, our results are similar (Tables V
Volume 72 Number 3 September, 1976
Lillehei-Kaster prosthesis
405
Table VI. Isolated aortic valve replacement: Clinicopathological material No.
Operative deaths
Uncomplicated AVD Complicated cases AS + VSD AI (+ aneurysm of aorta, ostial stenosis) Impaired function of LV Chronic lung disease Aneurysm of sinus of Valsalva Dissecting aneurysm of aorta Valsalva fistula into RV Renal failure Previous surgery Starr-Edwards valve with periprosthetic leak Starr-Edwards valve with hemolysis Leaking homograft Starr-Edwards with SBE
34
2
Totals
49
1 1 1 1 1 1 3 1 1 1 2 1
■itiS-.
5
0
7 (14%)
Legend: AVD, Aortic valve disease. AS. Aortic stenosis. VSD, Ventricular septal defect. AI, Aortic insufficiency. LV, Left ventricle. RV, Right ventricle. SBE, Subacute bacterial endocarditis.
Table VII. Isolated aortic valve replacement: Hemodynamic data of 6 patients
Fig. 3. Left ventricular cineangiogram, right anterior oblique projection, showing afillingdefect produced by thrombus (T) below the Lillehei-Kaster prosthesis. Ao, Aorta. LV, Left ventricle. Table VIII. Combined aortic and mitral valve replacement: Late deaths in 30 survivors
Valve size
PSG (mm. Hg)
Valve area (sq. cm.)
Cause
22 20 20 18 16 16
4 15 19 12 23 40
2.4 1.2 1.9 1.1 1.03 0.93
Thombosis of mitral prosthesis Sudden undetermined Infective endocarditis Pulmonary embolism Paraplegia
Total
No.
: (27%)
Legend: PSG, Pulmonary-systemic gradient.
and VII). The short follow-up period of 28 months precludes any comment on the frequency of systemic thromboembolism. In our series, the incidence of definite prosthetic thrombosis was 10 per cent in the mitral position and 5 per cent in the aortic position. However, if those patients who died suddenly and in who autopsies were not performed are included, the incidence could be much higher. Because of these disappointing results the use of the prosthesis was abandoned in 1974. Clinical evaluation of the Lillehei-Kaster prosthesis postoperatively is difficult, and valve dysfunction is not readily detected by auscultation. In contrast to cagedball or non-tilting disc mitral valve prostheses, which have clearly defined opening and closing sounds on auscultation, we could detect and document opening clicks in only 4 per cent of normally functioning mitral
prostheses. Additionally, mid-diastolic murmurs were audible in 30 per cent of cases, irrespective of the gradient across the valve; persumably this murmur is related to turbulence around the disc. Similar findings have been reported by other observers. In our experience the closing click is invariably single (Fig. 2). Gibson and associates3 reported that the closing sound is always twofold, with a small initial high-frequency component and a second high-frequency component separated by less than 0.03 second. Previous work has clearly demonstrated, however, that this initial component of the first heart sound ("M") is present in normal subjects and also in patients with Starr-Edwards prosthetic valves whether they are in sinus rhythm or atrial filbrillation.4 The nature of origin of " M , " which occurs at the time of crossover of the left atrial and left ventricular pressures,5 is uncertain, but it is probably
4 0 6 Mitha et al.
The Journal of Thoracic and Cardiovascular Surgery
Fig. 4. Chest roentgenograms in an asymptomatic patient with a Lillehei-Kaster prosthesis showing dramatic change in heart size. A, Preoperative. B and C, Six weeks and 28 months after operation, respectively. produced by resonance of the myocardium and atrioventricular ring and is unrelated to prosthetic events. The presentation of cases of abrupt valve dysfunction was typically one of acute pulmonary edema or congestive cardiac failure with marked tachycardia or uncontrolled atrial fibrillation. Because of the difficulty in assessing valve dysfunction by auscultation, we found chest roentgenograms particularly helpful in diagnosis by demonstrating increasing cardiomegaly and, in particular, sudden left atrial enlargement and pulmonary venous hypertension. Fluoroscopy is of no value because the disc is not radiopaque. Cardiac catherization and cineangiography demonstrated significant gradients across the prosthesis, with filling defects produced by thromboses on the atrial and ventricular aspects of the valve (Fig. 3). The problem of dysfunction of disc prostheses has been reviewed recently by Roberts and colleagues.6 Important factors are (1) intrinsic stenosis of the prosthesis, (2) inadequate space in the ventricles or
ascending aorta to freely accomodate the prosthesis when disproportion exists or develops, (3) thrombus formation on the prostheses, and (4) prosthetic variance. It would appear from our findings that prosthetic dysfunction was in no way related to anticoagulant therapy. Prosthetic thrombosis occurred more frequently in those patients who received anticoagulants (25 per cent) than in those who did not (4 per cent). Prosthetic disproportion, that is a prosthesis which is too large for the aortic or ventricular cavity, is probably the most important mechanism pertinent to our series. Most of the mitral prostheses used were large—60 of the 77 prostheses were sizes 20 to 25, and these were the valves involved by thrombosis. The smaller prostheses (sizes 16 and 18) were not affected in this manner. It is likely that disproportion was present, with minimal impingement and cocking of the disc on the ventricular endocardium resulting in thrombosis of the prosthesis and the left atrium. Apart from one case in which thrombosis occurred within 24 hours of operation, it is extremely improbable that dispropor-
Volume 72 Number 3 September, 1976
tion and cocking of the disc was present at the conclusion of the operation. In the majority, some other factor must have been responsible for the disproportion. Most of our patients are young, having had virulent, acute rheumatic fever at an early age. Their hearts are usually markedly dilated, and we have frequently observed dramatic shrinkage in heart size following correction of the hemodynamics 7 (Fig. 4). It is probable that disproportion is related to this change, with resultant impingement of the prosthetic disc on the endocardium of the left ventricle as the chamber decreases in size. Other factors, such as unsuspected severe aortic incompetence with the regurgitant jet interfering with disc opening, were not present. 6 It is now well appreciated that the caged-ball type of prosthesis may obstruct left ventricular outflow or inflow after valve replacement, and thrombus formation on these valves is not uncommonly found at autopsy. 8 Usually the amount of thrombus on the struts or primary orifice is insufficient to interfere with ball or poppet movement. Prostheses of the disc type, however, are much more vulnerable to abrupt valve dysfunction with an immobile disc than is the caged-ball variety. Our disturbing experience cannot be attributed to faulty valve design but more appropriately to failure to recognize the potential for subsequent disproportion. Deliberate insertion of the smaller prostheses may avoid this complication, particularly in those cases in which the left ventricle is dilated and volume overloaded from mitral or aortic incompetence. REFERENCES I Lillehei, C. W., Kaster, R. L., Coleman, M., and Bloch, J. H.: Heart Valve Replacement with the Lillehei-Kaster
Lillehei-Kaster
2
3
4 5
6
7
8
prosthesis
407
Pivoting Disc Prosthesis, N.Y. State J. Med 74: 1426, 1974. Lillehei, C. W., Carlson, R. G., and Kaster, R. L., et al.: Heart Valve Replacement with the Lillehei-Kaster Pivoting Disc Prosthesis, Presented before the Society for Vascular Surgery, Carmel, Calif., June, 1972. Cited by Brawley, R. K., Donahoo, J. D., andGott, V. L.: Current Status of the Beall, Bjork-Shiley, Braunwald-Cutter Cardiac Valve Prostheses, Am. J. Cardiol. 35: 855, 1975. Gibson, T. C , Starek, P. J. K., Roos, S., and Craige, E.: Echocardiographic and Phonocardiographic Characteristics of the Lillehei-Kaster Mitral Valve Prostheses, Circulation 49: 434, 1974. Armstrong, T. G., and Gotsman, M. S.: Initial Low Frequency Vibrations of the First Heart Sound, Br. Heart J. 35: 691, 1973. Lakier, J. B., Pocock, W. A., Gale, G. E., and Barlow, J. B.: Haemodynamic and Sound Events Preceding First Heart Sound in Mitral Stenosis, Br. Heart J. 34: 1 152, 1972. Roberts, W. C , Fishbein, M. C , and Golden, A.: Cardiac Pathology after Valve Replacement by Disc Prosthesis. A Study of 61 Necropsy Patients, Am. J. Cardiol. 35: 740, 1975. Van der Horst, R. L., Joshi, M. A., le Roux, B. T., and Rogers, N. M. A., and Gotsman, M. S.: The Chest X-ray After Mitral Valve Replacement in Childhood, S. Afr. Med. J. 46: 1933, 1972. Kalke, B., Korns, M. E., Goott, B., Lillehei, C. W., and Edwards, J. E.: Engagement of Ventricular Myocardium by Open-Cage Atrioventricular Valvular Prosthesis, J. THORAC. CARDIOVASC. SURG. 58: 92,
1969.
9 Roberts, W. C , Buckely, B. A., and Morrow, A. C : Pathologic Anatomy of Cardiac Valve Replacement: A Study of 24 Necropsy Patients, Progr. Cardiovasc. Dis. 15: 539, 1973.