- Email: [email protected]
Prosthetic valve endocarditis
Prosthetic valve endocarditis The case for prompt valve replacement Treatment of patients with prosthetic valve endocarditis with existing guidelines has failed to reduce the over-all mortality rate to below 50 per cent. However, subgroups with high or lower risk of death can be delineated on the basis of risk factors such as early onset following surgery, high-risk causative organisms, cardiodynamic failure, and septic emboli. High-risk patients comprise more than 70 per cent of those with prosthetic valve endocarditis in recent series. Analysis of previously reported series indicates that the mortality rate for high-risk patients with late onset of prosthetic valve endocarditis treated by valve replacement was less than that of patients receiving only medical therapy. We have treated 6 consecutive patients with prosthetic valve endocarditis (3 early onset, 3 late onset) by valve replacement before completion of a course of antibiotics. All patients survived surgery but one patient died after 4Vi months of noninfectious causes. Prompt valve replacement is technically feasible and should become the standard therapy for patients with prosthetic valve endocarditis who do not fall into the lower risk group.
Jeffrey R. Saffle, M.D.,* Pierce Gardner, M.D., Stephen C. Schoenbaum, M.D., and Wayne Wild, M.D., Chicago, III., and Boston, Mass.
/ a l t h o u g h the incidence of prosthetic valve endocarditis appears to be decreasing,1 the over-all mortality rate for patients with this infection remains high, averaging 56 per cent in recently reported series. 1-4 Factors correlating with high mortality rate include onset within 60 days of operation and infection with certain pathogens, especially Staphylococcus aureus, gramnegative bacilli, and fungi. Existing therapeutic guidelines 1-7 rely primarily on high doses of antimicrobial agents and reserve surgical intervention for patients who have congestive heart failure, valve dysfunction, or multiple septic emboli or those in whom the infection cannot be eradicated by medical therapy. Given the unhappy results with prosthetic valve endocarditis, alternative therapeutic approaches must be From the Infectious Disease Program at the University of Chicago and Michael Reese Medical Center, Chicago, 111., the Departments of Medicine and Pediatrics, Pritzker School of Medicine, the University of Chicago, Chicago, 111., and from the Departments of Medicine, New England Deaconess Hospital, Peter Bent Brigham Hospital, and Harvard Medical School, Boston, Mass. Received for publication Aug. 23, 1976. Accepted for publication Sept. 28, 1976. Address for reprints: Pierce Gardner, M.D., Billings Hospital, 950 E. 59th Street, Chicago, 111. 60637. *Present address: Department of Surgery, University of Utah College of Medicine, Salt Lake City, Utah.
416
considered. Three observations have led us to consider prompt replacement of infected prosthetic valves as a potential therapeutic improvement: (1) the favorable experience in patients requiring valve replacement for complications of primary infective endocarditis,8-11 (2) the reduction in mortality rate when prompt valve replacement became the accepted therapy for Candida endocarditis,12,13 and (3) our own favorable experience in managing infected prosthetic devices in other areas (cerebrospinal fluid shunts) by prompt surgical replacement. 14 This report is an analysis of valve replacement versus medical therapy alone in three recent series of prosthetic valve endocarditis, for which additional information was kindly supplied by the authors. In addition, we describe our own experience with 6 consecutive patients with prosthetic valve endocarditis managed by prompt valve replacement. Analysis of previously reported cases Recent series from three groups, the Mayo Clinic,1 the University of Oregon,2 and the Massachusetts General Hospital,3, 4 have reported a total of 134 cases of prosthetic valve endocarditis in 131 patients. The authors of these reports have kindly provided additional details on management and outcome of individual cases, which form the basis for this analysis. Smaller
Volume 73 Number 3 March, 1977
series5-7, 15 were examined but are not included in this report because of incomplete data. The mortality rate in the three series ranged from 51 to 60 per cent, averaging 56 per cent (Table I). This represents moderate improvement over the 72 per cent mortality rate observed in an earlier summary of experience with prosthetic valve endocarditis.6 Differences have been observed between those infections occurring within 60 days of prosthetic valve insertion—designated "early"—and those occurring later. 1-4,6 Prosthetic valve endocarditis of early onset carries a higher mortality rate (78 per cent) and is chiefly caused by staphylococci (5. aureus and 5. epidermidis), gram-negative bacilli, or Candida sp. (Table II). Prosthetic valve endocarditis of late onset carries a lower mortality rate (38 per cent) and is more commonly caused by streptococci (usually Streptococcus, viridans group, or group D streptococci), although staphylococci and gram-negative bacilli remain major pathogens (Table II). Over-all mortality rate greater than 60 per cent resulted from infections caused by 5. aureus, gramnegative bacilli, and Candida sp., whereas infections with Streptococcus, viridans group, and other streptococci had a better than average outcome (Table II). Although numbers are small, prosthetic valve endocarditis caused by group D streptococci gram-negative bacilli, and perhaps Streptococcus, viridans group, appeared to cause a greater mortality rate in early onset cases than in late onset cases. The mortality rate of prosthetic valve endocarditis caused by S. aureus and 5. epidermidis was not dramatically influenced by time of onset of the infection. Twenty-two per cent of patients were treated by valve replacement. The indications for operation were (1) presence of significant valvular leak or congestive heart failure or both (64 per cent), (2) persistent or recurrent bacteremia (35 per cent), and (3) peripheral emboli (22 per cent). Several cases met more than one such criterion. Whereas the over-all mortality rate was 62 per cent (63 of 102 cases) for patients treated with antibiotics alone, it was 41 per cent (12 of 29 cases) for patients treated by valve replacement plus antibiotics (p < 0.05). Because patients who die shortly after diagnosis are often too ill to benefit from any therapy and therefore might unfairly prejudice the treatment results in either group, further analysis was carried out on the subgroup of 115 patients who survived at least one week following diagnosis (Table III). Interestingly, the apparent advantage of surgical intervention became more pronounced (23 per cent mortality rate versus 60 per cent for medical treatment alone; p < 0.05). This difference was due entirely to the re-
Prosthetic valve endocarditis
417
Table I. Prosthetic valve endocarditis: Onset and mortality rate* Early onset of infection t
Late onset of infection^:
Mortality rate (%) No.
Total cases
Mortality Mortality rate (%) No. rate (%)
Series
No.
Mayo Clinic1 University of Oregon2 Mass. General Hospital3,4
16 23
88 87
29 25
38 36
45 48
55 60
21
62
20
40
41
51
60
78
74
38
134
56
Totals
*One hundred thirty-four cases of endocarditis occurred in 131 patients in the three reviewed series. 1-4 tWithin 60 days of prosthetic valve insertion. $More than 60 days from prosthetic valve insertion.
Table II. Prosthetic valve endocarditis: Mortality rate by onset and causative organism*
Organism Staphylococcus epidermidis Staphylococcus aureus Streptococcus, viridans group Group D streptococci Other streptococci! Gram-negative bacilli Candida species Othert Totals
Early onset (deaths/ cases)
Late onset (deaths/ cases)
Total (deaths/ cases)
7/13
3/7
10/20 (50%)§
13/15
8/10
21/25 (84%)
2/2
4/18
6/20 (30%)
3/4
3/12
6/16 (37%)
3/9
3/9
-
(33%)
13/15
3/12
16/27 (59%)
7/8 2/3
2/2 2/4
9/10 (90%) 4/7 (56%)
47/60 (78%)
28/74 (38%)
75/134 (56%)
*One hundred thirty-four cases of endocarditis occurred in 131 patients in three reviewed series. 1-4 tOther streptococci consisted of the following: group G Streptococcus, one case; group K Streptococcus, one case; group B Streptococcus, 2 cases; unclassified Streptococcus, 5 cases. tOther organisms: Corynebacterium xerosis, 2 cases, diphtheroids, 2 cases; Micrococcus, 3 cases. §Figures in parentheses are mortality rate.
suits in the treatment of prosthetic valve endocarditis of late onset. Too few patients with early onset of the disease were treated surgically to allow a valid comparison of results. A comparison of medical therapy and valve replacement for each organism considered separately was inconclusive owing to small numbers, although in no case was a disadvantage to valve replacement apparent.
The Journal of Thoracic and Cardiovascular Surgery
4 1 8 Saffle et al.
Table III. Outcome of valve replacement versus antibiotic therapy alone for prosthetic valve endocarditis in patients surviving at least one week after diagnosis* Late onset
Early onset
Series Mayo Clinic University of Oregon Mass. General Hospital Totals
Medical (deaths/cases)
Valve replacement (deaths leases)
11/13 18/21 7/11 36/45 (80%)
1/1 2/4 3/5 (60%)
*As compiled from three reviewed series.1 t p < 0.05.
-
Total
Medical (deaths leases)
Valve replacement (deaths/cases)
Medical (deaths/cases)
Valve replacement (deaths leases)
11/21 5/16 4/11 20/48 (42%)t
1/7 0/5 1/5 2/17(12%)t
22/34 23/37 11/22 56/93 (60%)t
1/7 1/6 3/9 5/22 (23%)t
4
New England Deaconess Hospital cases Table IV summarizes our own experience with 6 consecutive patients treated by valve replacement before the completion of antibiotic therapy and before the development of complications which would require surgical intervention under existing guidelines.1-7 All patients received appropriate drugs from the time positive blood cultures were reported to the time of operation. Although each patient showed clinical improvement and lessening of fever, none was entirely afebrile at the time of operation. All had infected prosthetic valves at surgery, and all survived the operation despite complicated postoperative periods which in 2 cases required immediate reoperation for mediastinal bleeding. One patient had postpericardiotomy syndrome. All patients received antimicrobial drugs for at least 4 weeks postoperatively. Four patients have survived long term (>one year), and one patient is alive and well 4 months following valve replacement. One death occurred in a patient who had a significant paravalvular leak which would have required yet another operation; surgery was deferred once because of an anticoagulant-related cerebral hemorrhage and later because of a false-positive culture report of Mycobacterium tuberculosis in the pleural fluid. Before operation could be rescheduled, a low output state and renal and bowel ischemia developed, and he died. At autopsy the valve ring was fibrotic, but there was no evidence of active prosthetic valve infection or of a mycotic cerebral aneurysm. Discussion Experiences with prosthetic valve endocarditis in three centers (Massachusetts General Hospital, the Mayo Clinic, and the University of Oregon Hospital) have yielded remarkably similar results. The over-all mortality rate remains high, 56 per cent, and even in late onset cases, when organisms with a "favorable" prognosis are included, almost 40 per cent of patients die. These results mandate a search for a more effective
approach to therapy. One strategy would be to replace the infected valve in patients for whom high mortality rate can be predicted once the diagnosis of valve infection is established, without waiting for treatment failure or complications as an indication for surgical intervention. The goals of prompt surgical intervention are to excise the infected area before it has extended into vital or inaccessible myocardial tissue and to lessen opportunity for development of congestive heart failure caused by valve incompetence. In addition, the risk of embolic or vasculitic lesions, including mycotic aneurysm, might be reduced. The potential risks of prompt surgical intervention include death during or immediately after operation, recurrent infection, and development of new paravalvular leaks which might necessitate yet another operation. Although replacement of an infected prosthetic valve may entail significant technical difficulties, these problems do not appear prohibitive as evidenced by our 6 cases and other reported series. 1-4 One of the consequences of recommending prompt surgery for patients with prosthetic valve endocarditis is that valve replacement will be done in some patients who could have been cured with antibiotics alone. It is likely that the infection in this group is limited to the prosthetic valve and has not extended into adjacent myocardial tissue; therefore, a high cure rate would also be anticipated with valve replacement. The operative mortality rate for patients undergoing cardiac surgery for bacterial endocarditis is correlated chiefly with the degree of congestive heart failure at the time of operation.16 A policy of prompt valve replacement would probably result in patients being operated upon during periods of more favorable hemodynamic function. Thus it might be anticipated that the operative mortality rate for patients undergoing prompt replacement of infected prosthetic valves would approach that of valve replacement for noninfected valves (one to 20
Volume 73 Number 3 March, 1977
Prosthetic valve endocarditis
419
Table IV. Six consecutive cases of prosthetic valve endocarditis managed by early valve replacement* Interval from first symptoms to surgery
Interval from report of positive culture to surgery
Congestive heart failure before surgery
Highest temperature on day before surgery
Case No.
Sex, age
1
M, 43
Aortic 42 days (early)
Candida spp.
36 days
17 days
No
101.0° F.
2
M, 72
Aortic 25 mo. (late)
Staphylococcus epidermidis
35 days
12 days
No
101.5° F.
3
F, 45
Aortic Immediate (early)
Diphtheroids
35 days
20 days
No
99.6° F.
4
M, 48
24 days
No
100.2° F.
M, 55
Staphylococcus epidermidis Diphtheroids
33 days
5
Aortic 37 mo. (late) Aortic 40 days (early)
24 days
2 days
101.8° F.
Died, 4'/2 mo. postop. (see text)
6
M, 48
Aortic 86 mo. (late)
Staphylococcus aureus
12 days
8 days
Yes (controlled medically before surgery) No
101.0° F.
Alive, 3 mo. postop.
Valve
Onset of symptoms postop.
Organism
Outcome Large paravalvular leak causing refractory failure; valve resutured 10 mo. later; alive and well 4 yr. after infection Immediate reop. for mediastinal bleeding; >1 yr. survival Postpericardiotomy syndrome; >1 yr. survival >1 yr. survival
♦Prosthetic valve infection was documented at surgery in all cases. Prior to surgery no patient had the commonly accepted indications for valve replacement, i.e., congestive heart failure, valve dysfunction, multiple septic emboli, and/or relapse of infection.
per cent depending on the degree of cardiac failure, valve involved, and general clinical status16). This range of mortality rate compares favorably with the mortality rate associated with even the best prognostic groups in Table II. Analysis of recent series of prosthetic valve endocarditis1-4 shows a significantly lower mortality rate for those patients who were treated by valve replacement compared to patients managed entirely medically, even though the surgically treated patients were operated upon following complications (i.e., congestive heart failure) which are associated with increased operative risk. This difference is due to favorable results of surgery in late onset cases. In late onset prosthetic valve endocarditis, results of therapy with antimicrobial drugs alone approach the success rate of valve replacement only for patients who have infections caused by less virulent organisms (i.e., Streptococcus, viridans group, Enterococcus) and who do not have any of the other indications for surgical intervention (valve dysfunction, congestive heart failure, multiple septic emboli, or refractory infection). This lower risk group comprised less than 30 per cent of the patients in recent series. 1-4 The unexpectedly high survival rate of patients with late onset infections caused by gram-negative bacilli should await further confirmation before these patients are included in the lower risk group.
Experience with valve replacement for prosthetic valve endocarditis of early onset is too limited to judge its value. However, the prohibitive mortality rate (80 per cent) of medically managed patients with the early onset of prosthetic valve endocarditis mandates a search for different therapeutic approaches, and prompt valve replacement would seem the most logical alternative. In our small series, 3 patients with early onset of prosthetic valve endocarditis were successfully treated by prompt valve replacement. All survived the operation, although one patient died 4'/2 months later without evidence of recurrent infection. It is our opinion that enough experience with prosthetic valve endocarditis has accumulated to predict reliably which patients are at high risk of dying on the basis of onset infection, causative organism, hemodynamic status, and presence or absence of septic emboli. It is possible that other prognostic factors may emerge to allow finer delineation of mortality risk for individual patients. At least 70 per cent of patients in recently reported series were at high risk, yet only 22 per cent of patients in these series were treated by valve replacement. Analysis of previous reports and our own experience with 6 consecutive patients with prosthetic valve endocarditis bolsters our opinion that prompt valve replacement is feasible and should be the standard therapy for prosthetic valve endocarditis in all patients who do not fall into the lower risk group, i.e.,
4 20
The Journal of Thoracic and Cardiovascular Surgery
Saffle et al.
patients with late-onset infections caused by organism^) associated with low mortality rates and in whom there are no other indications for surgery such as congestive failure, valve dysfunction, emboli, or refractory infections. We are extremely grateful to Drs. William E. Dismukes, Adolf W. Karchmer, Laura Slaughter, and Walter R. Wilson, who provided us with additional data from their previously reported cases, and to Drs. F. Henry Ellis, Jr., and Wilford B. Neptune, who performed the surgery at the New England Deaconess Hospital. REFERENCES 1 Wilson, W. R., Jaumin, P. M., Danielson, G. K., et al.: Prosthetic Valve Endocarditis, Ann. Intern. Med. 82: 751, 1975. 2 Slaughter, L., Morris, J. F., and Starr, A.: Prosthetic Valve Endocarditis: A 12-Year Review, Circulation 47: 1319, 1973. 3 Dismukes, W. E., Karchmer, A. W., Buckley, M. J., et al.: Prosthetic Valve Endocarditis: Analysis of 38 Cases, Circulation 48: 365, 1973. 4 Block, P. C , DeSanctis, R. W., Weinberg, A. N., and Austen, W. G.: Prosthetic Valve Endocarditis, J. THORAC. CARDIOVASC. SURG. 60: 540,
1970.
5 Bircks, W., Reidemeister, C , Sadony, V., et al.: Diagnostic and Therapeutic Problems of Septicemia After Valve Replacement, J. Cardiovasc. Surg. (Torino) 13: 385, 1972. 6 Shafer, R. B., and Hall, W. H.: Bacterial Endocarditis Following Open Heart Surgery, Am. J. Cardiol. 25: 602, 1970.
7 Killen, D. A., Collins, H. A., Koenig, M. G., et al.: Prosthetic Cardiac Valves and Bacterial Endocarditis, Ann. Thorac. Surg. 9: 238, 1970. 8 Black, S., O'Rourke, R. A., and Karliner, J. S.: Role of Surgery in the Treatment of Primary Infective Endocarditis, Am. J. Med. 57: 357, 1974. 9 Sarot, I. A., Weber, D., and Schechter, D. C : Cardiac Surgery in Active, Primary Infective Endocarditis, Chest 57: 58, 1970. 10 Walker, S. R., Shumway, N. E., and Merigan, T. C : Management of Infected Cardiac Valve Prostheses, J. A. M. A. 208: 531, 1969. 11 Wilcox, B. R., Proctor, H. J., Rackley, C. E., et al.: Early Surgical Treatment of Valvular Endocarditis, J. A. M. A. 200: 820, 1967. 12 Turnier, E., Kay, J. H., Bernstein, S., et al.: Surgical Treatment of Candida Endocarditis, Chest 67: 262, 1975. 13 Rubinstein, E., Noriega, E. R., Simberkoff, M. S.,etal.: Fungal Endocarditis: Analysis of 24 Cases and Review of the Literature, Medicine 54: 331, 1975. 14 Schoenbaum, S. C , Gardner, P., and Shillito, J.: Infections of Cerebrospinal Fluid Shunts: Epidemiology, Clinical Manifestations, and Therapy, J. Infect. Dis. 131: 543, 1975. 15 Sande, M. A., Johnson, W. D., Jr., Hook, E. W., et al.: Sustained Bacteremia in Patients With Prosthetic Cardiac Valves, N. Engl. J. Med. 286: 1067, 1972. 16 Wilson, W. R., Jaumin, P. M., Danielson, G. K., et al.: Cardiac Valve Replacement in Patients With Congestive Heart Failure Caused by Infective Endocarditis. Fifteenth Interscience Conference on Antimicrobial Agents and Chemotherapy, 1975, Abstract 195.
Jeffrey R. Saffle, M.D.,* Pierce Gardner, M.D., Stephen C. Schoenbaum, M.D., and Wayne Wild, M.D., Chicago, III., and Boston, Mass.
/ a l t h o u g h the incidence of prosthetic valve endocarditis appears to be decreasing,1 the over-all mortality rate for patients with this infection remains high, averaging 56 per cent in recently reported series. 1-4 Factors correlating with high mortality rate include onset within 60 days of operation and infection with certain pathogens, especially Staphylococcus aureus, gramnegative bacilli, and fungi. Existing therapeutic guidelines 1-7 rely primarily on high doses of antimicrobial agents and reserve surgical intervention for patients who have congestive heart failure, valve dysfunction, or multiple septic emboli or those in whom the infection cannot be eradicated by medical therapy. Given the unhappy results with prosthetic valve endocarditis, alternative therapeutic approaches must be From the Infectious Disease Program at the University of Chicago and Michael Reese Medical Center, Chicago, 111., the Departments of Medicine and Pediatrics, Pritzker School of Medicine, the University of Chicago, Chicago, 111., and from the Departments of Medicine, New England Deaconess Hospital, Peter Bent Brigham Hospital, and Harvard Medical School, Boston, Mass. Received for publication Aug. 23, 1976. Accepted for publication Sept. 28, 1976. Address for reprints: Pierce Gardner, M.D., Billings Hospital, 950 E. 59th Street, Chicago, 111. 60637. *Present address: Department of Surgery, University of Utah College of Medicine, Salt Lake City, Utah.
416
considered. Three observations have led us to consider prompt replacement of infected prosthetic valves as a potential therapeutic improvement: (1) the favorable experience in patients requiring valve replacement for complications of primary infective endocarditis,8-11 (2) the reduction in mortality rate when prompt valve replacement became the accepted therapy for Candida endocarditis,12,13 and (3) our own favorable experience in managing infected prosthetic devices in other areas (cerebrospinal fluid shunts) by prompt surgical replacement. 14 This report is an analysis of valve replacement versus medical therapy alone in three recent series of prosthetic valve endocarditis, for which additional information was kindly supplied by the authors. In addition, we describe our own experience with 6 consecutive patients with prosthetic valve endocarditis managed by prompt valve replacement. Analysis of previously reported cases Recent series from three groups, the Mayo Clinic,1 the University of Oregon,2 and the Massachusetts General Hospital,3, 4 have reported a total of 134 cases of prosthetic valve endocarditis in 131 patients. The authors of these reports have kindly provided additional details on management and outcome of individual cases, which form the basis for this analysis. Smaller
Volume 73 Number 3 March, 1977
series5-7, 15 were examined but are not included in this report because of incomplete data. The mortality rate in the three series ranged from 51 to 60 per cent, averaging 56 per cent (Table I). This represents moderate improvement over the 72 per cent mortality rate observed in an earlier summary of experience with prosthetic valve endocarditis.6 Differences have been observed between those infections occurring within 60 days of prosthetic valve insertion—designated "early"—and those occurring later. 1-4,6 Prosthetic valve endocarditis of early onset carries a higher mortality rate (78 per cent) and is chiefly caused by staphylococci (5. aureus and 5. epidermidis), gram-negative bacilli, or Candida sp. (Table II). Prosthetic valve endocarditis of late onset carries a lower mortality rate (38 per cent) and is more commonly caused by streptococci (usually Streptococcus, viridans group, or group D streptococci), although staphylococci and gram-negative bacilli remain major pathogens (Table II). Over-all mortality rate greater than 60 per cent resulted from infections caused by 5. aureus, gramnegative bacilli, and Candida sp., whereas infections with Streptococcus, viridans group, and other streptococci had a better than average outcome (Table II). Although numbers are small, prosthetic valve endocarditis caused by group D streptococci gram-negative bacilli, and perhaps Streptococcus, viridans group, appeared to cause a greater mortality rate in early onset cases than in late onset cases. The mortality rate of prosthetic valve endocarditis caused by S. aureus and 5. epidermidis was not dramatically influenced by time of onset of the infection. Twenty-two per cent of patients were treated by valve replacement. The indications for operation were (1) presence of significant valvular leak or congestive heart failure or both (64 per cent), (2) persistent or recurrent bacteremia (35 per cent), and (3) peripheral emboli (22 per cent). Several cases met more than one such criterion. Whereas the over-all mortality rate was 62 per cent (63 of 102 cases) for patients treated with antibiotics alone, it was 41 per cent (12 of 29 cases) for patients treated by valve replacement plus antibiotics (p < 0.05). Because patients who die shortly after diagnosis are often too ill to benefit from any therapy and therefore might unfairly prejudice the treatment results in either group, further analysis was carried out on the subgroup of 115 patients who survived at least one week following diagnosis (Table III). Interestingly, the apparent advantage of surgical intervention became more pronounced (23 per cent mortality rate versus 60 per cent for medical treatment alone; p < 0.05). This difference was due entirely to the re-
Prosthetic valve endocarditis
417
Table I. Prosthetic valve endocarditis: Onset and mortality rate* Early onset of infection t
Late onset of infection^:
Mortality rate (%) No.
Total cases
Mortality Mortality rate (%) No. rate (%)
Series
No.
Mayo Clinic1 University of Oregon2 Mass. General Hospital3,4
16 23
88 87
29 25
38 36
45 48
55 60
21
62
20
40
41
51
60
78
74
38
134
56
Totals
*One hundred thirty-four cases of endocarditis occurred in 131 patients in the three reviewed series. 1-4 tWithin 60 days of prosthetic valve insertion. $More than 60 days from prosthetic valve insertion.
Table II. Prosthetic valve endocarditis: Mortality rate by onset and causative organism*
Organism Staphylococcus epidermidis Staphylococcus aureus Streptococcus, viridans group Group D streptococci Other streptococci! Gram-negative bacilli Candida species Othert Totals
Early onset (deaths/ cases)
Late onset (deaths/ cases)
Total (deaths/ cases)
7/13
3/7
10/20 (50%)§
13/15
8/10
21/25 (84%)
2/2
4/18
6/20 (30%)
3/4
3/12
6/16 (37%)
3/9
3/9
-
(33%)
13/15
3/12
16/27 (59%)
7/8 2/3
2/2 2/4
9/10 (90%) 4/7 (56%)
47/60 (78%)
28/74 (38%)
75/134 (56%)
*One hundred thirty-four cases of endocarditis occurred in 131 patients in three reviewed series. 1-4 tOther streptococci consisted of the following: group G Streptococcus, one case; group K Streptococcus, one case; group B Streptococcus, 2 cases; unclassified Streptococcus, 5 cases. tOther organisms: Corynebacterium xerosis, 2 cases, diphtheroids, 2 cases; Micrococcus, 3 cases. §Figures in parentheses are mortality rate.
suits in the treatment of prosthetic valve endocarditis of late onset. Too few patients with early onset of the disease were treated surgically to allow a valid comparison of results. A comparison of medical therapy and valve replacement for each organism considered separately was inconclusive owing to small numbers, although in no case was a disadvantage to valve replacement apparent.
The Journal of Thoracic and Cardiovascular Surgery
4 1 8 Saffle et al.
Table III. Outcome of valve replacement versus antibiotic therapy alone for prosthetic valve endocarditis in patients surviving at least one week after diagnosis* Late onset
Early onset
Series Mayo Clinic University of Oregon Mass. General Hospital Totals
Medical (deaths/cases)
Valve replacement (deaths leases)
11/13 18/21 7/11 36/45 (80%)
1/1 2/4 3/5 (60%)
*As compiled from three reviewed series.1 t p < 0.05.
-
Total
Medical (deaths leases)
Valve replacement (deaths/cases)
Medical (deaths/cases)
Valve replacement (deaths leases)
11/21 5/16 4/11 20/48 (42%)t
1/7 0/5 1/5 2/17(12%)t
22/34 23/37 11/22 56/93 (60%)t
1/7 1/6 3/9 5/22 (23%)t
4
New England Deaconess Hospital cases Table IV summarizes our own experience with 6 consecutive patients treated by valve replacement before the completion of antibiotic therapy and before the development of complications which would require surgical intervention under existing guidelines.1-7 All patients received appropriate drugs from the time positive blood cultures were reported to the time of operation. Although each patient showed clinical improvement and lessening of fever, none was entirely afebrile at the time of operation. All had infected prosthetic valves at surgery, and all survived the operation despite complicated postoperative periods which in 2 cases required immediate reoperation for mediastinal bleeding. One patient had postpericardiotomy syndrome. All patients received antimicrobial drugs for at least 4 weeks postoperatively. Four patients have survived long term (>one year), and one patient is alive and well 4 months following valve replacement. One death occurred in a patient who had a significant paravalvular leak which would have required yet another operation; surgery was deferred once because of an anticoagulant-related cerebral hemorrhage and later because of a false-positive culture report of Mycobacterium tuberculosis in the pleural fluid. Before operation could be rescheduled, a low output state and renal and bowel ischemia developed, and he died. At autopsy the valve ring was fibrotic, but there was no evidence of active prosthetic valve infection or of a mycotic cerebral aneurysm. Discussion Experiences with prosthetic valve endocarditis in three centers (Massachusetts General Hospital, the Mayo Clinic, and the University of Oregon Hospital) have yielded remarkably similar results. The over-all mortality rate remains high, 56 per cent, and even in late onset cases, when organisms with a "favorable" prognosis are included, almost 40 per cent of patients die. These results mandate a search for a more effective
approach to therapy. One strategy would be to replace the infected valve in patients for whom high mortality rate can be predicted once the diagnosis of valve infection is established, without waiting for treatment failure or complications as an indication for surgical intervention. The goals of prompt surgical intervention are to excise the infected area before it has extended into vital or inaccessible myocardial tissue and to lessen opportunity for development of congestive heart failure caused by valve incompetence. In addition, the risk of embolic or vasculitic lesions, including mycotic aneurysm, might be reduced. The potential risks of prompt surgical intervention include death during or immediately after operation, recurrent infection, and development of new paravalvular leaks which might necessitate yet another operation. Although replacement of an infected prosthetic valve may entail significant technical difficulties, these problems do not appear prohibitive as evidenced by our 6 cases and other reported series. 1-4 One of the consequences of recommending prompt surgery for patients with prosthetic valve endocarditis is that valve replacement will be done in some patients who could have been cured with antibiotics alone. It is likely that the infection in this group is limited to the prosthetic valve and has not extended into adjacent myocardial tissue; therefore, a high cure rate would also be anticipated with valve replacement. The operative mortality rate for patients undergoing cardiac surgery for bacterial endocarditis is correlated chiefly with the degree of congestive heart failure at the time of operation.16 A policy of prompt valve replacement would probably result in patients being operated upon during periods of more favorable hemodynamic function. Thus it might be anticipated that the operative mortality rate for patients undergoing prompt replacement of infected prosthetic valves would approach that of valve replacement for noninfected valves (one to 20
Volume 73 Number 3 March, 1977
Prosthetic valve endocarditis
419
Table IV. Six consecutive cases of prosthetic valve endocarditis managed by early valve replacement* Interval from first symptoms to surgery
Interval from report of positive culture to surgery
Congestive heart failure before surgery
Highest temperature on day before surgery
Case No.
Sex, age
1
M, 43
Aortic 42 days (early)
Candida spp.
36 days
17 days
No
101.0° F.
2
M, 72
Aortic 25 mo. (late)
Staphylococcus epidermidis
35 days
12 days
No
101.5° F.
3
F, 45
Aortic Immediate (early)
Diphtheroids
35 days
20 days
No
99.6° F.
4
M, 48
24 days
No
100.2° F.
M, 55
Staphylococcus epidermidis Diphtheroids
33 days
5
Aortic 37 mo. (late) Aortic 40 days (early)
24 days
2 days
101.8° F.
Died, 4'/2 mo. postop. (see text)
6
M, 48
Aortic 86 mo. (late)
Staphylococcus aureus
12 days
8 days
Yes (controlled medically before surgery) No
101.0° F.
Alive, 3 mo. postop.
Valve
Onset of symptoms postop.
Organism
Outcome Large paravalvular leak causing refractory failure; valve resutured 10 mo. later; alive and well 4 yr. after infection Immediate reop. for mediastinal bleeding; >1 yr. survival Postpericardiotomy syndrome; >1 yr. survival >1 yr. survival
♦Prosthetic valve infection was documented at surgery in all cases. Prior to surgery no patient had the commonly accepted indications for valve replacement, i.e., congestive heart failure, valve dysfunction, multiple septic emboli, and/or relapse of infection.
per cent depending on the degree of cardiac failure, valve involved, and general clinical status16). This range of mortality rate compares favorably with the mortality rate associated with even the best prognostic groups in Table II. Analysis of recent series of prosthetic valve endocarditis1-4 shows a significantly lower mortality rate for those patients who were treated by valve replacement compared to patients managed entirely medically, even though the surgically treated patients were operated upon following complications (i.e., congestive heart failure) which are associated with increased operative risk. This difference is due to favorable results of surgery in late onset cases. In late onset prosthetic valve endocarditis, results of therapy with antimicrobial drugs alone approach the success rate of valve replacement only for patients who have infections caused by less virulent organisms (i.e., Streptococcus, viridans group, Enterococcus) and who do not have any of the other indications for surgical intervention (valve dysfunction, congestive heart failure, multiple septic emboli, or refractory infection). This lower risk group comprised less than 30 per cent of the patients in recent series. 1-4 The unexpectedly high survival rate of patients with late onset infections caused by gram-negative bacilli should await further confirmation before these patients are included in the lower risk group.
Experience with valve replacement for prosthetic valve endocarditis of early onset is too limited to judge its value. However, the prohibitive mortality rate (80 per cent) of medically managed patients with the early onset of prosthetic valve endocarditis mandates a search for different therapeutic approaches, and prompt valve replacement would seem the most logical alternative. In our small series, 3 patients with early onset of prosthetic valve endocarditis were successfully treated by prompt valve replacement. All survived the operation, although one patient died 4'/2 months later without evidence of recurrent infection. It is our opinion that enough experience with prosthetic valve endocarditis has accumulated to predict reliably which patients are at high risk of dying on the basis of onset infection, causative organism, hemodynamic status, and presence or absence of septic emboli. It is possible that other prognostic factors may emerge to allow finer delineation of mortality risk for individual patients. At least 70 per cent of patients in recently reported series were at high risk, yet only 22 per cent of patients in these series were treated by valve replacement. Analysis of previous reports and our own experience with 6 consecutive patients with prosthetic valve endocarditis bolsters our opinion that prompt valve replacement is feasible and should be the standard therapy for prosthetic valve endocarditis in all patients who do not fall into the lower risk group, i.e.,
4 20
The Journal of Thoracic and Cardiovascular Surgery
Saffle et al.
patients with late-onset infections caused by organism^) associated with low mortality rates and in whom there are no other indications for surgery such as congestive failure, valve dysfunction, emboli, or refractory infections. We are extremely grateful to Drs. William E. Dismukes, Adolf W. Karchmer, Laura Slaughter, and Walter R. Wilson, who provided us with additional data from their previously reported cases, and to Drs. F. Henry Ellis, Jr., and Wilford B. Neptune, who performed the surgery at the New England Deaconess Hospital. REFERENCES 1 Wilson, W. R., Jaumin, P. M., Danielson, G. K., et al.: Prosthetic Valve Endocarditis, Ann. Intern. Med. 82: 751, 1975. 2 Slaughter, L., Morris, J. F., and Starr, A.: Prosthetic Valve Endocarditis: A 12-Year Review, Circulation 47: 1319, 1973. 3 Dismukes, W. E., Karchmer, A. W., Buckley, M. J., et al.: Prosthetic Valve Endocarditis: Analysis of 38 Cases, Circulation 48: 365, 1973. 4 Block, P. C , DeSanctis, R. W., Weinberg, A. N., and Austen, W. G.: Prosthetic Valve Endocarditis, J. THORAC. CARDIOVASC. SURG. 60: 540,
1970.
5 Bircks, W., Reidemeister, C , Sadony, V., et al.: Diagnostic and Therapeutic Problems of Septicemia After Valve Replacement, J. Cardiovasc. Surg. (Torino) 13: 385, 1972. 6 Shafer, R. B., and Hall, W. H.: Bacterial Endocarditis Following Open Heart Surgery, Am. J. Cardiol. 25: 602, 1970.
7 Killen, D. A., Collins, H. A., Koenig, M. G., et al.: Prosthetic Cardiac Valves and Bacterial Endocarditis, Ann. Thorac. Surg. 9: 238, 1970. 8 Black, S., O'Rourke, R. A., and Karliner, J. S.: Role of Surgery in the Treatment of Primary Infective Endocarditis, Am. J. Med. 57: 357, 1974. 9 Sarot, I. A., Weber, D., and Schechter, D. C : Cardiac Surgery in Active, Primary Infective Endocarditis, Chest 57: 58, 1970. 10 Walker, S. R., Shumway, N. E., and Merigan, T. C : Management of Infected Cardiac Valve Prostheses, J. A. M. A. 208: 531, 1969. 11 Wilcox, B. R., Proctor, H. J., Rackley, C. E., et al.: Early Surgical Treatment of Valvular Endocarditis, J. A. M. A. 200: 820, 1967. 12 Turnier, E., Kay, J. H., Bernstein, S., et al.: Surgical Treatment of Candida Endocarditis, Chest 67: 262, 1975. 13 Rubinstein, E., Noriega, E. R., Simberkoff, M. S.,etal.: Fungal Endocarditis: Analysis of 24 Cases and Review of the Literature, Medicine 54: 331, 1975. 14 Schoenbaum, S. C , Gardner, P., and Shillito, J.: Infections of Cerebrospinal Fluid Shunts: Epidemiology, Clinical Manifestations, and Therapy, J. Infect. Dis. 131: 543, 1975. 15 Sande, M. A., Johnson, W. D., Jr., Hook, E. W., et al.: Sustained Bacteremia in Patients With Prosthetic Cardiac Valves, N. Engl. J. Med. 286: 1067, 1972. 16 Wilson, W. R., Jaumin, P. M., Danielson, G. K., et al.: Cardiac Valve Replacement in Patients With Congestive Heart Failure Caused by Infective Endocarditis. Fifteenth Interscience Conference on Antimicrobial Agents and Chemotherapy, 1975, Abstract 195.