Cardiac operation during active infective endocarditis

J

THORAC CARDIOVASC SURG

84:579-584, 1982

Cardiac operation during active infective endocarditis Results of aortic, mitral, and double valve replacement in 94 patients Cardiac valve replacement was performed in 94 patients (95 operations) in the presence of active infective endocarditis. Most of the patients were extremely ill. The operation was performed as an emergency or semiemergency lifesaving procedure in 88% of them, and more than half received little or no antibiotic treatment prior to the operation. The hospital mortality was 16%-14%for aortic valve replacement (AVR) and ll%for double valve replacement (DVR) but 31%for isolated mitral valve replacement (MVR). The mortality was not higher in patients operated on urgently (emergency or semiemergency), nor was it higher in patients who had aortic annular abscesses or aneurysms. Prosthetic valve endocarditis (PVE) (in each case occurring more than 60 days after the previous valve operation) carried a higher mortality (33%) than native valve endocarditis (NVE) (14%). The relatively high early mortality for MVR may have been related to the fact that we operated upon MVR patients after intensive medical treatment had failed. The late results were good: Sixty-six patients are alive and well, 51 of them in Functional Class 1. Six patients were reoperated upon for aortic periprosthetic leaks, and five are now well. Eight patients died late (9%), one of them because of a periprosthetic leak and one because of a clotted valve. 1n seven of the eight deaths, the cause of death was probably not related to the timing of the original operation. We recommend early valve replacement for patients with infective endocarditis. We believe that early operation reduces mortality, prevents emboli, and is associated with excellent long-term results.

Basil S. Lewis, M.D., M.R.C.P., F.C.P. (S.A.), Nearchos E. Agathangelou, M.D., Peter R. Colsen, M.B., F.R.C.S. (Ed.), Manuel Antunes, M.D., and Robin H. Kinsley, M.B., F.C.S. (S.A.), F.A.C.C., Johannesburg, South Africa

Cardiac valve replacement may be lifesaving in patients with infective endocarditis.v t! but the indications for surgical intervention are not altogether clear and the mortality is high in many centers. Since rheumatic valve disease is still common in southern Africa and many of these patients do not receive proper medical treatment until late in the course of their disease, we have had an unusual opportunity to see and treat a large From the Cardiovascular Research Unit, Departments of Cardiology and Cardiothoracic Surgery, University of the Witwatersrand and Baragwanath Hospital, Johannesburg, South Africa. Received for publication Oct. I, 1981. Accepted for publication Dec. 21, 1981. Supported in part by the Stella and Paul Loewenstein Cardiac Fund of the University of the Witwatersrand. Address for reprints: Professor B. S. Lewis, Cardiology Department, Hadassah Hospital, Jerusalem, Israel. 0022-5223/82/100579+06$00.60/0

© 1982 The

C. V. Mosby Co.

number of patients with severe, life-threatening infective endocarditis. Many patients have low cardiac output and severe cardiac failure on admission, and others are virtually moribund and require resuscitation soon after admission. rs We have operated upon a large number of patients with active endocarditis, most of them within hours of admission to the hospital. This paper describes our experience in 94 patients with infective endocarditis who underwent 95 valve replacement operations during the period 1975 to 1980.

Patients and methods During the 6 year period from January, 1975, to December, 1980, 95 valve operations were performed in 94 patients who had active infective endocarditis. Fifty-six patients were male and 38 female; their ages ranged from 8 to 65 years (mean 31 years). The crite579

The Journal of Thoracic and Cardiovascular Surgery

580 Lewis et ai.

Table I. Indications for operation Indications

Table III. Operative mortality No.ojoperations

Hemodynamic Severe cardiac failure Cardiac failure plus low cardiac output Cardiac failure plus cardiac arrest Cardiac failure plus conduction defect Hemodynamic lesion plus major embolization Uncontrolled sepsis

NVE (N = 86)

82 60 17

No. AVR (N = 43) DVR (N = 36) MVR (N = 16)

3 2 11 2

Totals

95

Total

I

%

PVE (N = 9) No.

I

%

Totals (N = 95) No·1

5 4

14 11

3

23

2

67

: 5 31

12

14

3

33

15 16

17

:~} ns

%

ns

Ins

ns Legend: Native valve endocarditis. PVE. Prosthetic valve endocarditis. AVR.

Aortic valve replacement. DVR. Double valve replacement. MVR. Mitral valve replacement.

Table II. Timing of operation Hospital deaths No.ojoperations

No.

Emergency Semiemergency Elective

61

IO

23

3

11

2

Totals

95

15

I

%

16 13 18

rion for inclusion in the study was a diagnosis of infective endocarditis made or confirmed a~ operation by the finding of florid macroscopic infection on the valves, chordal apparatus, or adjacent endocardial or myocardial structures. Fresh vegetations were present in 78 patients, leaflet or chordal rupture in 20, and total destruction of the valve(s) in 17 patients. Annular abscesses were present in 15 patients and a mycotic aneurysm of the ascending aorta was found in three. One patient had an aortico-right ventricular fistula and frank purulent pericarditis. Eighty-six operations were performed for native valve endocarditis (NVE) and nine for prosthetic valve endocarditis (PVE). One patient (included twice in the series) had PVE 35 months after a previous operation for NVE. All cases of PVE occurred more than 6 months after the previous valve replacement. In most patients, infective endocarditis had been diagnosed preoperatively on the basis of a typical clinical picture and/or positive blood cultures. In a few, infective endocarditis had not been suspected before operation. Positive blood cultures were obtained in 17 patients (Streptococcus in 11, Staphylococcus in three, and Candida, Corynebacterium, and Bacteroides in one each). In the clinical setting of Baragwanath Hospital, however, where an unusually large number of very ill patients receive early antibiotic treatment, many patients with florid clinical infective endocarditis (confirmed at operation and by histologic examination) have negative blood cultures. It has also been reported

that patients with infective endocarditis and negative blood cultures are more ill and have a higher incidence of major emboli and congestive heart failure. 16 For this reason we based our definitive diagnosis of infective endocarditis and our criterion for inclusion in this study on the finding of florid active infection at operation, whereas patients with old, healed lesions and/or a history of successfully treated infective endocarditis were excluded. During the last 18 months of the study period, real-time two-dimensional echocardiography was available, and large vegetations were seen on the valve(s) in 18 of 20 patients studied prior to operation. The indications for cardiac operation are given in Table I. In most patients, operation was undertaken for a severe hemodynamic lesion, almost always complicated by severe cardiac failure (60 patients) and in 17 patients by low cardiac output and shock. In three patients, cardiac arrest occurred and successful resuscitation was performed before the operation. Seventy-six patients were in New York Heart Association Functional Class IV and 19 in Class III at the time of operation. Emboli had occurred in 11 patients (12%), but these patients were also in severe cardiac failure. In two patients, operation was undertaken after aggressive antibiotic treatment failed to control septicemia. Preoperative treatment for infective endocarditis was limited. Seventeen patients had received no antibiotic treatment, 33 less than a week of treatment for infective endocarditis, 17 had had 1 to 3 weeks, and 13 had had 4 to 6 weeks of therapy. (In 15 the records from a referring hospital were not altogether clear, but it was likely that little or no treatment had been given to these patients.) Patients who had received antibiotic treatment were included in this series of "active" infective endocarditis for two reasons: (1) medical treatment had failed and septicemia persisted clinically and/or (2) there was florid uncontrolled infection at operation.

Volume 84 Number 4 October, 1982

Valve replacement was performed as an extreme emergency in 61 patients (64%) (within hours of admission to the hospital), as a "semiemergency" in 23 patients (24%) (on the next operating list), and "electively" in II patients (12%) (at a preselected date). Thus 84 patients (88%) were operated upon as a matter of urgency within 24 to 36 hours of admission to our hospital (Table II). Forty-three patients underwent aortic valve replacement (AVR), 36 double (aortic and mitral) valve replacement (DVR), and 15 isolated mitral valve replacement (MYR) (one patient twice). The type of valve inserted depended on the preference of the surgeon and the policy of the unit at the time. In the aortic position, 59 Bjork-Shiley prostheses were inserted, 17 Hall-Kaster valves, one St. Jude Medical and two Ionescu-Shiley pericardial valves. In the mitral position, eight Bjork-Shiley valves were used, seven HallKaster valves, and 37 tissue valves (23 Capentier-Edwards, seven Hancock, four Angell-Shiley, and three Ionescu-Shiley valves). All patients undergoing DVR in this series had predominant aortic valve endocarditis with lesser involvement of the mitral valve. The operative technique included cardiopulmonary bypass with moderate hypothermia in all patients. Intermittent cross-clamping or coronary perfusion was used prior to 1978, after which cold cardioplegia with iced saline in the pericardium was used routinely on all patients for myocardial protection during bypass. All valves in the aortic position were inserted with interrupted sutures, as were all tissue valves in the mitral position. The majority of mechanical valves in the mitral position were inserted with a continuous suturing technique. Care was taken to remove all infected tissue, and both the anulus and prosthesis (mechanical) were bathed in neomycin solution. All patients were ventilated electively after the operation, and inotropic agents were used when necessary. Statistical analysis. Statistical analysis of the difference between groups was performed by the chi square test with Yates' correction for continuity.

Results Operative mortality. The early mortality (30 day hospital mortality) was low (Table III). Fifteen patients died: six after AVR (14%), four after DVR (11%), and five after MVR (31%) (ns). Four patients died of continuing septicemia, three of pump failure and low cardiac output, two of multiple organ failure, and two of brain damage. Cardiac arrhythmias caused two deaths, and a massive pulmonary embolus caused a third. In one patient the cause of sudden death was unknown. It

Active infective endocarditis

58 I

is noteworthy that three of the five deaths after MVR were due to continuing sepsis with organ failure. The early mortality for PVE (33%) tended to be higher that than for NVE (14%) (ns) (Table III). This was especially true after MVR, for two of three patients operated upon for PVE died. Nonetheless, the mortality of MVR for NVE was also higher (23%) than that of the other groups. There was no difference in early mortality in the emergency (16% mortality), semiemergency (13% mortality), and elective groups (18% mortality) (Table II). Similarly, the operative mortality was not related to the organism involved, the duration of preoperative antibiotic treatment, the functional class at the time of operation, or the presence of annular abscesses or mycotic aneurysms of the aorta. Late results. Eighty patients were discharged from the hospital after valve replacement. Five have been lost to follow-up, but 75 have been followed for a period ranging from 2 to 74 months (mean 17.5 months). The mean follow-up periods for MVR, AVR, and DVR were similar. Eight patients died after discharge from the hospital: four of the AVR patients available for follow-up, three DYR patients, and one MVR patient. Two patients died of causes probably related to the presence of infective endocarditis at the time of the initial operation, one of recurrent periprosthetic aortic incompetence 9 months after operation and the other of septicemia 2 months postoperatively. One late death was due to a clotted Bjork-Shiley aortic prosthesis 30 months after operation. The remaining five deaths occurred at home and the cause remains unknown. Actuarial survival curves for AYR, DVR, and MVR patients are shown in Fig. 1. After an initially greater operative mortality for MVR, the three curves are similar. The secondary late fall in the curves 2 to 4 years after operation was probably not related to the fact that the original operation was performed in the presence of active infective endocarditis. Ring leaks occurred in 10 patients (11%), all but one of whom had an aortic prosthesis. The incidence of ring leaks was higher in patients who had had annular aortic abscesses. Leaks appeared within 2 months of operation in seven ofthe 10 patients and within the first year in the other three. Six patients (all in the isolated AYR group) required reoperation to close the leak. It is interesting that the original operation had been performed on an emergency basis in all six patients. One patient had two reoperations and one patient three. Both these patients are now well, but there has been one late death due to severe cardiac failure in a patient

The Journal of Thoracic and Cardiovascular Surgery

582 Lewis et at.

100

80

15 I ... 7 2 3 10 - (H) I 18 2 8 5 I tr--6 80 40 13 5 3 23 - ....... II 7 I I 5 , , , , 20 ...... , ----.-----.-----,-----,-----,-----,-----, j 7 6 4 o 2 3 5 Years

40

32

1

37

,

DVR AVR Totol MVR

Fig. 1. Actuarial survival curves for patients undergoing aortic (AVR), double (DVR), and mitral (MVR) valve replacement, and for the total group. Operative mortality is higher for MVR, but the curves are parallel thereafter. There is a 75% overall survival rate 2 years after operation. The numbers on the graph refer to the number of patients at risk in each group for each year.

who had a ring leak closed but subsequently developed recurrent aortic regurgitation and progressive cardiac failure. Sixty-six patients are alive and well, 51 in Functional Class I and the remainder in Functional Class II. Discussion

The present series comprised an unusual group of patients. Most presented very late in the course of their disease and many were moribund at the time of operation: In 88% of the patients, valve replacement was performed as a dire emergency or as a semiemergency within 36 hours of admission to the hospital, and many patients received either no preoperative antibiotic treatment or antibiotic treatment for less than 1 week prior to operation. In almost half of them, DVR was necessary. Nonetheless, the operative mortality was low compared to that of other published series (Table IV), both for patients undergoing AVR and for the relatively large group undergoing DVR. In contrast to the study of Young and associates, 13 DVR carried a low (11 %) early mortality in our hands, similar to that for AYR. Our mortality in these groups was not much higher than the 6% to 8% overall operative mortality for valve replacement in our hospital; again, we show that an aggressive surgical policy produces promising results in active infective endocarditis. 7, 10 Isolated MVR was associated with an increased

hospital mortality. This was true both for NVE and PVE and may be related to two factors: First, three of the five early deaths after MYR were due to continuing sepsis. In patients requiring MVR, the valvular involvement by infective endocarditis was usually gross and valvular destruction severe, so that in some patients the valve was removed piecemeal. It is probable, then, that adequate curettage of septic tissue was not always possible in MVR patients. Sepsis and abscesses in relation to the aortic valve and aortic root were more easily removed and repaired at operation. In contrast to the study by Richardson and associates, lOwe found no increased mortality in patients who had annular aortic abscesses. Second, it is our policy to treat ill patients with mitral valve lesions conservatively at first, with intensive antifailure therapy. Patients with mitral valve disease often respond dramatically to ventilation and can be operated upon subsequently in an elective or semielective procedure. Thus the small group of MVR patients with active infective endocarditis in this series are probably a selected group of very ill patients not responding to aggressive medical treatment. It is possible that the mortality for MVR would be lower if we had operated upon these patients soon after admission, as we do those with aortic and double valve disease. We are at present reconsidering our previous policy because of the results of this study. PVE also carried a higher operative mortality then NVE, although the difference was not significant statistically. In our patients, PVE developed late after the previous valve operation, and the mortality was expected to be closer to that of NVE.IO Because infective endocarditis did not develop early (less than 2 months after operation) in any of our patients, we cannot add to the data of Richardson and associates'? in this regard. The late results of operation were good, with a 75% survival rate 2 years after operation. The secondary fall in the actuarial survival curves after 2 years was probably not related to the circumstances of the original operation. Sixty-six patients are alive and well, all in Functional Class I or II. Almost all would have died had they not been operated upon in the acute phase of their disease. The relatively low incidence of emboli (12%) is probably also related to our aggressive surgical policy. The late complication rate was low. The 11% incidence of ring leaks (with 1% mortality) was acceptable in view of the severity of infection in our patients. Reinfection has not been a problem. One patient developed PVE 35 months after operation and died of septicemia following his second MVR, but no other pa-

Volume 84

583

Active infective endocarditis

Number 4 October, 1982

Table IV. Recent literature relating to surgery for infective endocarditis

Ref. No. 5

No. of cases

16 293

Emergency or semiemergency procedures (%)*

Periprosthetic leads (%)*

Operative mortality (%)*

No.

I

Requiring reoperation

I

Fatal outcome

63

31 22

15

6

4

?

20

?

?

?

?

53

19 14

4 11

4 4

2

80

43

23

6

0

100

27 16 43 37 16

27

27

9

? 13

?

?

7

7

(review)

6

239 (review)

7 10

54 81 (NYE)

35

II

12 13

14 Present report

(PYE) 11

27 30 19 95

? ?

95 88

4

None reported 11

6

Legend: NVE. Native valve endocarditis. PYE. Prosthetic valve endocarditis.

'Expressed as a percentage of total No. of cases.

tient in our series has had new infective endocarditis. Recurrent infective endocarditis is common in drug abusers. The infection often is right-sided and carries a high late mortality." The results of the study are of importance in relation to the management of patients with infective endocarditis. Our low operative mortality is probably due to our aggressive surgical policy in patients with infective endocarditis, as suggested by Boyd and associates. 7 Much higher mortality rates (30% to 40%) are reported in the literature (Table IV) and are almost certainly due to delay in operation. ra, 14 We believe that a good clinical examination, electrocardiogram, chest roentgenogram, and high-quality two-dimensional echocardiogram obviate the need for cardiac catheterization in most patients. Most of our patients go from the admission ward to the noninvasive laboratory and then directly to the operating room. In this series, there was no difference in the mortality of patients undergoing such emergency valve replacement compared with those operated upon electively. Echocardiography (and especially real-time twodimensional echocardiography) is a useful noninvasive method for detecting the presence, size, and location of vegetations in patients with infective endocarditis. 17 The echo is "positive" for vegetations in 40% to 80% of patients with infective endocarditis. In our hospital, where patients present for medical treatment late in the course of their disease, 80% to 90% are "echo positive, " and the vegetations are very large, often 20 mm or greater. "Echo positive" infective endocarditis car-

ries a high morbidity and mortality, and most patients require early cardiac operation to avoid death. 17- 19 Major emboli are frequent.": 19 We operate immediately on patients who have a severe hemodynamic lesion of the aortic valve and prefer to operate early (next convenient operating list) on patients with aortic valve vegetations and a significant hemodynamic lesion which is stable. We have not as yet operated upon patients with large aortic vegetations and a mild or trivial hemodynamic lesion, but this situation is unusual, since vegetations are usually associated with important valve destruction and regurgitation. The presence of associated mitral valve disease does not change this policy. On the other hand, mitral valve endocarditis appears to be a different problem, with a higher operative mortality. It is possible that a more aggressive policy may have produced better surgical results, but as yet we have not operated early for mitral vegetations unless the hemodynamic lesion was severe and refractory to medical treatment. REFERENCES Wallace AG, Young WG, Osterhout S: Treatment of acute bacterial endocarditis by valve excision and replacement. Circulation 31:450-453, 1965 2 Manhas DR, Mohri H, Hessel EA, Merendino KA: Experience with surgical management of primary infective endocarditis. Am Heart J 84:738-747, 1972 3 Weinstein L, Schlesinger J: Treatment of infective endocarditis. Prog Cardiovasc Dis 16:275-302, 1973 4 Black S, O'Rouke R, Karliner J: Role of surgery in the

584 Lewis et al,

5

6

7

8

9

10

II

12

treatment of primary infective endocarditis. Am J Med 56:357-369, 1974 Jung JY, Saab SB, Almond CH: The case for early surgical treatment of left-sided primary infective endocarditis. A collective review. J THoRAc CARDIOVASC SURG 70:509-518, 1975 Parrott JCW, Hill JD, Kerth WJ, Gerbode F: The surgical management of bacterial endocarditis. A review. Ann Surg 183:289-292, 1976 Boyd AD, Spencer FC, Isom OW, Cunningham IN, Reed GE, Acinapura AJ, Tice DA: Infective endocarditis. An analysis of 54 surgically treated patients. J THORAC CARDIOVASC SURG 73:23-30, 1977 Saffle JR, Gardner P, Schoenbaum SC, Wild W: Prosthetic valve endocarditis. The case for prompt valve replacement. J THoRAc CARDIOVASC SURG 73:416-420, 1977 Kinsley RH, Colsen PR, Bakst A: Emergency valve replacement for primary infective endocarditis. S Afr Med J 53:86-88, 1978 Richardson J V, Karp RB, Kirklin JW, Dismukes WE: Treatment of infective endocarditis. A 10 year comparative analysis. Circulation 58:589-597, 1978 Wilson WR, Danielson GK, Giuliani ER, Washington JA, Jaumin PM, Geraci JE: Valve replacement in patients with active infective endocarditis. Circulation 58:585588, 1978 Rapaport E: The changing role of surgery in the management of infective endocarditis (Editorial). Circulation 58:598-599, 1978

The Journal of Thoracic and Cardiovascular Surgery

13 Young JB, Welton DE, Raizner AE, Ishimori T, Montero A, Guinn GA, Mattox K, Gentry LO, Alexander JK, Miller RR: Surgery in active infective endocarditis. Circulation 60:Suppl:77-81, 1979 14 Davis RS, Strom JA, Frishman W, Becker R, Matsumoto M, Lejemtel TH, Sonnenblick EH, Frater RWM: The demonstration of vegetations by echocardiography in bacterial endocarditis. An indication for early surgical intervention. Am J Med 69:57-63, 1980 15 Louw JWK, Kinsley RH, Dion RAE, Colsen PR, Girdwood RW: Emergency heart valve replacement. An analysis of 170 patients. Ann Thorac Surg 29:415-422, 1980 16 Pesanti EL, Smith 1M: Infective endocarditis with negative blood cultures. An analysis of 52 cases. Am J Med 66:43-50, 1979 17 Mintz GS, Kotler MN: Clinical value and limitations of echocardiography. Its use in the study of patients with infectious endocarditis. Arch Intern Med 140: 1022-1027, 1980 18 Wann LS, Hallam CC, Dillon JC, Weyman AE, Feigenbaum H: Comparison of M-mode and cross-sectional echocardiography in infective endocarditis. Circulation 60:728-733, 1979 19 Stewart JA, Silimperi D, Harris P, Wise NK, Fraker TD, Kisslo JA: Echocardiographic documentation of vegetative lesions in infective endocarditis. Clinical implications. Circulation 61:374-380, 1980