- Email: [email protected]
Annular destruction in acute bacterial endocarditis
J
THORAC CARDIOVASC SURG
1989;97:755-63
Annular destruction in acute bacterial endocarditis Surgical techniques to meet the challenge Destruction and disruption of ventricular-aortic or mitral-aortic continuity in the presence of acute infection of the annular tissue is a significant surgical challenge. Among 82 patients who underwent surgical treatment for acute endocarditis over a lo-year period, 15 (18.2 %) had extensive destruction of the anulus necessitating special reconstructive techniques for treatment. Surgical treatment involved removal of all infected tissue including annular elements followed by appropriate restoration of the anulus for safe anchoring of the prosthetic valve. The reconstruction of the anulus consisted of the following: a Teflon felt patch inside and outside the aorta or ventricle, or both, for secure attachment of the prosthesis (felt aortic root, in three patients with native valveendocarditis), valvedcomposite graft replacement of the aortic root for ventricular-aortic discontinuity (Bentall procedure, in eight patients with prosthetic valve endocarditis), composite patch reconstruction of the mitral anulus and the ascending aorta to restore mitral-aortic continuity (mitral-aortic composite patch in two patients with mitral-aortic prosthetic valve endocarditis), and direct suture of the sewing skirts of the mitral and aortic prostheses to restore the defect (attached skirts, in one patient with mitral-aortic native valveendocarditis). There was one hospital death caused by multiple organ failure. The most common complication was heart block. Two late deaths were due to reinfection resulting from continued intravenous drug abuse. One patient with a felt aortic root repair required late reoperation for subannular aneurysm. Eleven patients were followed up from 7 months to 66 months and are alive and well without complications. This experience indicates that these seemingly radical surgical techniques can be used in these desperately ill patients with safety and good long-term results. They offer the only lasting solution for major disruption in cardiac anatomy in the presence of infection.
M. Arisan Ergin, MD, Sharo Raissi, MD, Fabrizio Follis, MD, Steven L. Lansman, MD, and Randall B. Griepp, MD, New York, N.Y.
h e role of surgical treatment in the management of bacterial endocarditis is well established. Timely valve replacement during the course of the disease is expected to lead to an excellent functional outcome with a low prevalenceof residual or recurrent infection." However, extension of the infectious process into the annular structures remains a major determinant of both the early and late results of surgical treatment of endocarditis.3-8 Varying degrees of annular involvement present a substantial challenge to the cardiac surgeon's ingenu-
From the Division of Cardiothoracic Surgery, The Mount Sinai Medical Center, New York, N.Y.
ity in restoring the anatomic and functional integrity of the cardiac structures after removal of infected tissues. Several surgical techniques have been described and used effectively, mostly in the treatment of aortic annular involvement. There appears to be a paucity of information in the literature regarding the management of destruction of the mitral anulus or the left fibrous trigone. This report is a summary of our clinical experience in a group of patients with extensive annular destruction requiring special reconstruction of the anulus for treatment. On the basis of this experience, indications for various techniques in the management of this increasingly common clinical problem are suggested.
Read at the Fourteenth Annual Meeting of The Western Thoracic Surgical Association in Hawaii, June 22-25, 1988.
Patients and methods
Address for reprints: M. Arisan Ergin, MD, Divisionof Cardiothoracic Surgery, The Mount Sinai Medical Center, One Gustave L. Levy Place, Box 1028, New York, NY 10029.
Among 82 patients who underwent surgical treatment for acute bacterial endocarditis between 1977 and 1987, 15 patients had extensive destruction of the anulus and required
755
The Journal of Thoracic and Cardiovascular Surgery
7 5 6 Ergin et al.
Table I. Prevalence of infecting organisms and culture results Cultures
Organism Staphylococcus aureus Streptococcus viridans Staphylococcus epidermidis Streptococcus faeca/is Pneumococcus Actinobacillus Gram-positive cocci Total
No. of patients
Blood
Valve
+
+
5 4
5 4
2
2
I I I
I I I
0 0 0 0 0 0
-.l
....Q
1
15
14
special reconstructive techniques for treatment. There were 13 men and two women aged from 20 to 74 years. Five patients had native valve endocarditis and 10 had prosthetic valve endocarditis. In 10 patients the infection was localized to the aortic valve and the anulus, in four it involved the mitral and aortic valves and anuli, and in one only the mitral anulus was affected. Eight patients were current or former drug abusers, and five had prosthetic endocarditis with only two thought to be active users. In two patients with prosthetic endocarditis the present procedure represented a third valve replacement. In patients with prosthetic valve endocarditis the interval from the original valve replacement was from 1.5 to 84 months (mean 17.3 ± 23.4 months). Fourteen patients had positive blood cultures, and the same organism was cultured from the excised valves in seven. Organisms were seen in the pathology specimen in 14 patients. One patient had negative blood and valve cultures, but organisms were found on pathologic examination. The most common organism was Staphylococcus aureus (Table I). All patients received varying courses of preoperative antibiotic treatment, the duration of which varied from 7 to 42 days (mean 16.9 ± 10.7 days). In all patients the diagnosis was based on clinical and bacteriologic findings supported by echocardiography. Five patients underwent cardiac catheterization, as well. The indication for operation in nine patients was deteriorating hemodynamics and persistent sepsis. Six patients were operated on because of increasing heart failure alone. One patient with double valve endocarditis was admitted with cardiogenic shock and renal shutdown, and another patient had peripheral embolization. All operations were performed with standard cardiopulmonary bypass and hypothermia. Single-dose cardioplegia supplemented with topical profound hypothermia was used for myocardial protection. Operative findings in patients with prosthetic endocarditis having isolated aortic annular involvement showed more than 50% destruction of the anulus, with subannular aneurysms measuring 2 to 4 cm. Ventricular-aortic continuity was completely disrupted in two of these patients, and two others had associated ventricular septal defects (Fig. I, A and B). All eight of these patients received a valved composite graft for the repair. In three patients with native valve endocarditis, the destruction of the aortic anulus was confined to less than 50% of the anulus in a contiguous manner, with subannular aneurysms measuring lessthan 2 ern. In all these a felt aortic root repair was possiblewith orthotopic
Pathology
I
+
0
I 3 I I
I
0
4 I I
5 3
0
2
0 0 0 0
Q
0
I
I I I
Q
1
-.l
7
8
14
I
I
implantation of the new prosthesis. Three patients with combined mitral-aortic endocarditis had significant destruction of the left fibrous trigone, -with aorta-left atrial fistulization in two (Fig. 2). In the patient with native valve endocarditis, the defect was repaired by direct suture of the mitral and aortic prostheses. In the other two with prosthetic endocarditis, extensive reconstruction with a composite mitral-aortic patch was necessary. In the only patient with isolated mitral annular destruction, partial ventricular translocation of the prosthesis was used (Table II). Surgical technique. In all patients, all infected tissue was removed and thorough debridement was done. This included removal of the valve and all annular elements that appeared to be involved. All abscesses were curetted until firm tissue was encountered. At times it was difficult to discern between infection and sterile destruction. Sutures for repair of the resultant defect were taken away from the edge in viable tissue, preferably with transmural buttressing. Inside and outside patches were used for tension-free closure of the defects. Five types of repair were used in this group of patients for specific indications. I. Felt aortic root. The felt aortic root technique was used in patients with native valve endocarditis with less than 50% destruction of the aortic anulus. After adequate debridement of the infected tissues, the resultant defect in the anulus was lined with a patch of Teflon felt or pericardium, which was sutured in place circumferentially with mattress sutures through healthy tissue. These sutures were tied over another Teflon strip on the outside, which in effect obliterated the defect in the anulus between two patches (one inside and the other outside) secured with through-and-through sutures. Teflon felt was used for its compliance for effectivebuttressing of sutures. Pericardium was preferred for inside patches to reduce the amount of prosthetic material facing the bloodstream. This technique lends itself best for defects along the noncoronary and left coronary cusps. In one patient the partially detached anterior leaflet of the mitral valve was also reattached to this patch. This method allows secure closure of limited subannular aneurysms and provides a solid base for securing the new prosthesis. 2. Valved composite graft replacement ofthe aortic root. A valved composite graft was used in patients with extensive destruction of the aortic anulus (more than 50%) with subannular aneurysms measuring more than 2 em or in noncontiguous areas of the anulus leading to partial or total
Volume 97 Number 5 May 1989
Acute bacterial endocarditis
757
Fig. 1. A, Aortic root injection showing complete ventricular-aortic discontinuity in patient 7 with prosthetic endocarditis. Note the separation of the ventricle from the aortic root with formation of large circumferential subannular aneurysm. B, Postoperative aortogram in the same patient after valved composite graft insertion. The distance between the prosthesisand the coronary orifices is apparent.
ventricular-aortic discontinuity. Interrupted pledget-backed mattress sutures were used for safe anchoring of the prosthesis on the ventricular side. In patients with an accompanying ventricular septal defect, transmural sutures were taken along the septum. The subannular aneurysmal area was completely excluded from the bloodstream. Coronary anastomoses and distal anastomosis were done as described for the original Bentall procedure. The periprosthetic space was closed with the native aorta and a patch of bovine pericardium for effective hemostasis (Fig. 3). 3. Attached skirts. The erosion of the left fibrous trigone in
a patient with native mitral and aortic endocarditis was treated by direct suture of the valve skirts in this area . The bulky sewing skirt of the Hancock mitral prosthesis was especially well suited for this purpose (Fig. 4). 4. Mitral-aortic composite patch. Modified adaptation of the method described for the enlargement of mitral and aortic anuli in double valve replacement was used in two patients with extensive destruction of the left fibrous trigone because of bivalvular prosthetic endocarditis. The aortotomy was extended into the left atrium through the destroyed left fibrous trigone, and after adequate excision of all infected tissue the
The Journal of Thoracic and Cardiovascular Surgery
7 5 8 Ergin et al.
Fig. 2. Aortic root injection in patient 13 with prosthetic double valve endocarditis and destruction of the left fibrous trigone. Ready opacification of the left atrium indicates aorta-left atrial fistula.
defect in the mitral anulus was restored with a composite patch of autogenous pericardium facing the bloodstream and knitted Dacron graft on the outside (Fig. 5, A, B, and C). This triangular patch was used to anchor the mitral prosthesis in this region and the aortic prosthesis slightly higher up at its base. The rest of the patch was then extended and incorporated into the closure of the aortotomy (Fig. 5, D and E). 5. Partial translocation of the mitral value. In a patient with destruction of the posterior mitral anulus, the anchoring sutures for the prosthesis were taken in the ventricle. The valve was partially translocated into the ventricle along the posterior anulus, and the weakened area of the anulus was left on the atrial side of the repair facing low pressure.
Results Complete reconstruction of the anulus with implantation of the appropriate prosthesis was accomplished in all patients with the described techniques. There were no intraoperative deaths. The only hospital death occurred in a patient with native aortic and mitral endocarditis who was admitted in cardiogenic shock and renal shutdown. He required felt repair of the noncoronary cusp and part of the left coronary cusp area with aortic and mitral valve replacement. He died 45 days after the operation with multiple organ failure. At autopsy both the repair and the prosthetic valves were intact. The most common complication was heart block. Only one patient had preoperative complete heart block, whereas seven patients had postoperative block. Six patients had heart block among the group of eight receiving composite grafts. Five patients received pacemakers at the same operation and two later. Two serious
gastrointestinal tract complications were encountered. Ischemic colitis in one patient responded to conservative treatment, and upper gastrointestinal tract bleeding required gastrectomy in another. The same patient also had a postoperative cerebrovascular accident but recovered without sequelae. There were no residual infections. All patients received appropriate antibiotic treatment for from 3 to 8 weeks postoperatively. Fourteen patients were discharged from the hospital and were available for follow-up, ranging from 7 to 66 months (mean 23.7 ± 15.0 months). Eleven patients are alive and well without complications. There were two late deaths resulting from reinfection caused by continued drug abuse. One occurred in a patient with a valved conduit at 12 months, and the other occurred in the patient with double valve replacement and attached skirts at 35 months. Neither patient had a repeat operation. In both of these patients the original offending organism was Staphylococcus aureus, and the long interval between repair and recurrence of endocarditis suggests reinfection because of intravenous drug abuse rather than recurrence. There was one late failure of felt repair of the aortic root at 12 months. This patient had a large subannular aneurysm as a result of dehiscence of the lower aspect of the inside patch without evidence of reinfection and was treated with a valved conduit replacement (Table II).
Discussion Annular involvement during the course of endocarditis carries important diagnostic, therapeutic, and prognostic significance.2-6, 9,10 The prevalence of annular destruction in native valve endocarditis is related to the virulence of the responsible organism and the duration of the process.2, 5, II Varying degrees of annular involvement are a constant feature of prosthetic endocarditiS. 3, 12-17 The spectrum of annular destruction extends from simple localized annular abscess formation to larger subannular aneurysms, with or without perforation into other cardiac chambers, extension into the pericardial space, or total disruption of ventricularaortic continuity and the mitral-aortic fibrous trigone." 18-23 Early surgical treatment in patients with suspected annular involvement is indicated to limit the annular destruction and its sequelae." 3, 22 Advanced annular destruction necessitates specific surgical techniques for effective treatment. In this situation the surgeon is confronted with the challenge of removing all infected tissue while preserving the functional and anatomic integrity of the left ventricular inlet and outlet.
Volume 97 Number 5
Acute bacterial endocarditis
May 1989
759
Table II. Operative findings, surgical techniques, and results Patient No.
Age (yr)
Annular involvement
2
74
LV-Ao discontinuity + VSD 75%
3
36
50%
4
24
75%
5
29
50%
6
52
75%
7
29
LV-Ao discontinuity
8
29
75%+ VSD
9
24
LCS, 2 ern
10
57
LCS, 1.5 em
11
23
NCS 2.0 crn
12
20
Left fibrous trigone
13
50
Left fibrous trigone
14
25
Left fibrous trigone
15
29
Post mitral anulus
27
Technique
Conduit, 23 mm B-S valve Conduit, 23 mm B-S valve Conduit, 23 mm B-S valve Conduit, 23 mm B-S valve Conduit, 23 mm B-S valve Conduit, 27 mm B-S valve Conduit, 23 mm B-S valve Conduit, 27 mm B-S valve Felt outside and inside, AVR Felt outside, pericardium inside, AVR, MVR Felt outside and inside, AVR Attached skirts, AVR, MVR Mitral-aortic patch, AVR, MVR Mitral-aortic patch, AVR, MVR Partial translocation
Complications
Follow-up
Heart block
Alive, 16 mo
Gastrectomy, CVA
Alive, 17 mo
Heart block
Late death, 12 mo Alive, 27 mo
Heart block Heart block
Alive, about 12 mo Alive, 66 mo
Heart block
Alive, 12 mo
Heart block
Alive, 30 rno Alive, 19 mo
Multiorgan failure
Hospital death
Reop, 12 rno
Ischemic colitis
Late death, 36 mo Alive, 36 rno Alive, 32 rno Alive, reop at 30 rno
LV-Ao, Left ventricular-aortic; VSD, ventricular septal defect; LCS, left coronary sinus; NCS, noncoronary sinus; B-S, Bjork-Shiley; A VR, aortic valve replacement; MVR, mitral valve replacement; CV A, cerebrovascular accident.
The aortic anulus is the most frequently involved structure of the fibrous skeleton of the heart." Several effectivesurgical techniques have been described for the repair of aortic annular disruption. Small abscesses of the aortic anulus may be debrided and closed or incorporated into the valve sutures. 3, 25 Larger subannular aneurysms have been treated by exclusion with prosthetic or pericardial patches.v" 10, 18, 19, 26 For more severe forms of destruction of the aortic anulus with ventricular-aortic discontinuity, aortic root replacement with a valved composite graft or homograft or distal translocation of the aortic valve with saphenous vein grafts has been used." 22, 25, 27-30 We believe it is important to excise all grossly infected structures to reduce the risk of residual infection or late valve dehiscence. Familiarity with these techniques leads to more confident use of debridement. The defect created in the
anulus after removal of all infected tissue dictates the choice of the reconstructive technique. It is best to reserve simple plication for only small cavities. Larger cavities, those less than 2 em in contiguous areas of the aortic anulus, should be treated with one of the patch methods. This approach allows tension-free closure of the aneurysm with sutures placed well beyond the limits of the excision in viable tissue. We have used a double patch technique with felt or preferably pericardium on the blood-contacting surface and compliant prosthetic material on the outside (felt aortic root). This method is best suited for defects along the better portions of the left and noncoronary sinuses. Defects along other areas of the aortic anulus can be closed by techniques described by others.v'? The patch provides a sturdy structure for the fixation of the new prosthesis while excluding the aneurysmal area from the high-pressure
760
Ergin et al.
Fig. 3. Artist's concept of ventricular-aortic discontinuity and its treatment withvalved composite graft. Note buttressed transmural sutures along the crest of the left ventricle for secure fixation of the prosthesis. The closure of the native aorta helps with immediate hemostasis.
system. The double patch technique allows use of autogenous pericardium inside with the prosthetic patch outside for secure fixation of the valve. In the more extensive annular destructions that we encountered with prosthetic endocarditis, secure annular repair could not be accomplished with patch techniques. In patients with large subannular aneurysms and more than 50% destruction of the circumference of the anulus, implantation of a valved composite graft was the most expeditious and secure repair. This method in the treatment of ventricular-aortic discontinuity was first described by Frantz, Murray, and Wilcox" and since then has been used by others. 10, 25, 27 The conduit totally excludes the attenuated area from the systemic pressure and has proved to be a safe and durable repair. Transmural sutures, as described by others, help secure fixation of the composite graft to the ventricular crest.": 28 The anastomosis of the coronary ostia is not difficult, since the subannular separation leaves adequate distance between the prosthesis and the coronary ostia. A composite homograft, as described by Lau and associates," may be preferable on theoretical grounds. However, the availability of the prosthetic conduit and familiarity with the technique playa role in this choice. The alternative method of distal translocation of the prosthesis with closure of the coronary ostia and saphenous vein grafting is a more complex operation that leaves the patient at risk for aneurysmal dilatation of the annular area, as well as dependent on vein grafts.": 29,30
The Journal of Thoracic and Cardiovascular Surgery
Fig. 4. Attached skirtsusedfor repairof simple disruption of the mitral-aortic trigone. Direct suture of the sewing skirts of the prostheses wassufficient to bridge the defect in the patient with native mitral-aortic endocarditis,
Both of these problems have been seen in patients with translocation of the aortic valve.v" Complete heart block was the most frequent complication of valved conduit repair of the aortic root in our experience, the price paid for secure fixation of the prosthesis. There is a paucity of information in the literature about the management of the mitral anulus and aorticmitral continuity in endocarditis." 25 We have encountered four patients with aortic-mitral discontinuity, Destruction of the left fibrous trigone, especially in the presence of prosthetic endocarditis, led to a large defect at the junction of the aortic and mitral anuli with serious physiologic and anatomic consequences. Adaptation of the technique that was originally described by Rastan and associates" for the enlargement of the combined anuli for mitral and aortic valve replacement proved to be an effective solution to this otherwise devastating defect. The composite patch fashioned from autogenous pericardium and knitted graft material easily bridges the defect in the mitral anulus and provides for secure fixation of the mitral prosthesis in this area. The aortic prosthesis is likewise secured to the patch distally. The remainder of the patch is then incorporated into the closure of the aortotomy, so that the left fibrous trigone is completely reconstructed. Pericardium leaves autogenous tissue facing the bloodstream and provides for absolute hemostasis, while the compliant knitted graft material serves as structural support against systemic pressure and provides for solid fixation of the prostheses.
Volume 97 Number 5 May 1989
Acute bacterial endocarditis
76 I
Fig. 5. Steps in the reconstruction of extensive destruction of the left fibrous trigone with a composite patch. See text for details.
We have not found a description of previous use of this technique in the management of endocarditis. Isolated destruction of the mitral anulus is rare in endocarditis." Rocchiccioli and associates" described attachment of a Dacron flange to the mitral prosthesis for atrial implantation of the prosthesis in cases in which the mitral anulus could not be used for fixation of the valve. This technique leaves the attenuated portion of the anulus in the high-pressure chamber at risk for aneurysmal dilatation or rupture as a late complication, especially along the posterior anulus. Therefore, in the only patient with destruction of the posterior .mitral anulus whom we encountered, a partial ventricular translocation of the prosthesis was used. At reoperation for degeneration of the porcine valve, the sewing skirt was found to be firmly attached without any evidence of dehiscence. This experience indicates that serious complications of annular destruction in endocarditis can be treated effectively by adhering to sound surgical principles of thorough debridement, followed by tension-free reconstruction of the anulus that excludes all attenuated structures from systemic pressure and provides a solid fixation point for the prosthesis. These seemingly radical surgical techniques were used with an acceptable opera-
tive risk and good long-term results and seem to be the only lasting solution to the major disruption of cardiac anatomy in the presence of infection. REFERENCES I. Young JB, Welton DE, Raizner AE, et ai. Surgery in active infective endocarditis. Circulation 1979;60(Pt 2):177-8 I. 2. D'Agostino RS, Miller DC, Stinson EB, et ai. Valve replacement in patients with native valve endocarditis: What really determines operative outcome? Ann Thorac Surg 1985;40:429-38. 3. Nelson RJ, Harley, DP, French WJ, Bayer AS. Favorable ten-year experience with valve procedures for active infective endocarditis. J THORAC CARDIOVASC SURG 1984;87:493-502. 4. Richardson JV, Karp RB, Kirklin JW, Dismukes WE. Treatment of infective endocarditis: a lO-year comparative analysis. Circulation 1978;58:589-97. 5. Arnett E, Roberts W. Valve ring abscess in active infective endocarditis: frequency, location and clues to clinical diagnosis from the study of 95 necropsy patients. Circulation 1976;54:140-5. 6. Baumgartner WA, Miller CD, Reitz BA, et ai. Surgical treatment of prosthetic valve endocarditis. Ann Thorac Surg 1983;35:87-104.
762
The Journal of Thoracic and Cardiovascular Surgery
Ergin et al.
7. Wilcox BR, Murray GF, Starek PJK. The long-term outlook for valve replacement in active endocarditis. J THORAC CARDIOVASC SURG 1977;74:860-3. 8. Calderwood SB, Swinski RN, Karchmer AW, Waternaus CM, Buckley MJ. Prosthetic valve endocarditis. J THoRAC CARDIOVASC SURG 1986;92:776-83. 9. Lau JM, Guinn GA, Beall AC Jr, et al. Operative techniques in infective endocarditis. Ann Thorac Surg 1981;32:351-6. 10. Fiore AC, Ivey TD, McKeown PP, et al. Patch closure of aortic annulus mycotic aneurysms. Ann Thorac Surg 1986;42:372-8. II. Rapaport E. The changing role of surgery in the management of infective endocarditis [Editorial]. Circulation 1978;58(Pt 2):II-IV598-9. 12. Arnett EN, Roberts We. Prosthetic valve endocarditis: clinicopathologic analysis of 22 necropsy patients with comparison of observations in 74 necropsy patients with active infective endocarditis involving natural left-sided cardiac valves. Am J Cardiol 1976;38:281-92. 13. Dismukes WE, Karchmer AW, Buckley MJ, et al. Prosthetic valve endocarditis: analysis of 38 cases. Circulation 1973;48:365-77. 14. Slaughter L, Morris JE, Starr A. Prosthetic valvular endocarditis: a 12-year review. Circulation 1973;47:131926. 15. Wilson WR, Jaumin PM, Danielson GK, et al. Prosthetic valve endocarditis. Ann Intern Med 1975;82:751-6. 16. Stinson EB, Griepp RB, Vosti K, et al. Operative treatment of active endocarditis. J THORAC CARDIOVASC SURG 1976;71:659-64. 17. Symbas PN, Vlasis SE, Zacharopoulos L, et al. Immediate and longterm outlook for valve replacement in acute bacterial endocarditis. Ann Surg 1982;195:721-5. 18. Todd EP, Hubbard SG, Zeok JV, et al. Repair of mycotic aneurysms of the aorta involving the aortic valve. Ann Thorac Surg 1980;30:160-6. 19. Bailey WW, Ivey TD, Miller DW Jr. Dacron patch closure of aortic annulus mycotic aneurysms. Circulation 1982;66(Pt 2):1127-30. 20. Lau JKH, Robles A, Cherian A, et al. Surgical treatment of prosthetic endocarditis. J THORAC CARDIOVASC SURG 1984;87:712-6. 21. Danielson GK, Titus JL, DuShane JW. Successful treatment of aortic valve endocarditis and aortic root abscesses by insertion of prosthetic valve in ascending aorta and placement of bypass grafts to coronary arteries. J THoRAc CARDIOVASC SURG 1974;67:443-9. 22. Frantz PT, Murray GF, Wilcox BR. Surgical management of left ventricular-aortic discontinuity complicating bacterial endocarditis. Ann Thorac Surg 1987;29:1-7. 23. Chesler E, Korns ME, Porter GE, Reyes CN, Edwards JE. False aneurysm of the left ventricle secondary to bacterial endocarditis with perforation of the mitral-aortic intervalvular fibrosa. Circulation 1968;37:518-23. 24. Cowgill DL, Addonizio P, Hopeman AR, Harken AH. A
25.
26.
27.
28.
29.
30.
31.
practical approach to prosthetic valve endocarditis. Ann Thorac Surg 1987;43:450-7. Rocchiccioli C, Chastre J, Lecompte Y, Gandjbakhch I, Gibert C. Prosthetic valve endocarditis. J THoRAc CAROlOVASC SURG 1986;92:784-9. Olinger GN, Maloney JV Jr. Repair of left ventricularaortic discontinuity complicating endocarditis from an aortic valve prosthesis. Ann Thorac Surg 1977;23:576-7. VanHooser DW, Johnson RG, Hein RA, Elkins nc Successful management of aortic valve endocarditis with associated periannular abscess and aneurysm. Ann Thorae Surg 1986;42:148-51. Rumisek JD, Berry WR, Barry MJ, Clarke JS. Transseptal control of the difficult aortic annulus. Ann Thorae Surg 1985;39:385-6: Stinson EB. Surgical treatment of infective endocarditis. Progress in cardiovascular diseases. Vol 22, New York: Grone & Stratton, 1979. Reitz BA, Stinson EB, Watson· DC, et al. Translocation of the aortic valve for prosthetic valve endocarditis. J THoRAC CARDIOVASC SURG 1981;81:212-8. Rastan H, Atai M, Hadi H, Yazdanyar A. Enlargement of mitral valvular ring. J THORAC CARDIOVASC SURG 1981;81:106-11.
Discussion Dr. Tom D. hey (Seattle. Wash.). This paper is an important milestone in the treatment of endocarditis. The Stanford group has shown us that early intervention is important in achieving acceptable results in these difficult patients. They have also, at times and in certain situations, advocated translocation of the aortic valve and coronary artery bypass grafting. Dr. Warren Bailey and Dr. Andy Fiore have previously reported our techniques of closing annular abscesses with patches, allowing normal positioning of the aortic prosthesis. The authors have presented an extraordinary series of 15 patients with bacterial endocarditis and with only one perioperative death. I might add that this was before the era of substrate cardioplegia and modern myocardial protection techniques. Ten of 15 patients had prosthetic endocarditis and, in four of them, both the aortic and mitral anuli were destroyed. Staphyloccocal organisms were most commonly present. We strongly agree with the authors' approach to radical debridement of all infected tissue, and we believe this is the key element in their success. Heart block is certainly acceptable in this situation. The mean duration of antibiotic treatment in this series was 17 days before surgical intervention. We recently saw a patient with a Starr-Edwards prosthesis who was inadequately anticoagulated and had been referred to the neurology service with an embolic stroke. The patient had a heparin lock placed and 3 days after the institution of this intravenous therapy, had a fever spike and subsequently a positive blood culture for Staphylococcus aureus. Within 48 hours, the patient had prolongation of the PR interval and was immediately operated on. In this short period of time, the anulus under the prosthesis was totally destroyed. The prosthesis was still seated, but roughly 90% of the annular tissue was
Volume 97 Number 5 May 1989
eroded with purulent material. Thus annular destruction can certainly occur within 48 hours with staphyloccocal organisms if a prosthesis is in place. Unfortunately, the decision for early operation, as I am sure the authors would agree, is not often in the hands of the surgeon. The treatment of combined aortic and mitral prosthetic endocarditis, as outlined by the authors, is also an important contribution. I have several questions: What type of suture technique do you use to anchor the composite grafts on the ventricular septum? If I understand your manuscript, this is not an annular location but is subannular. Am I correct in this conclusion? Dr. Ergin. Yes, you are correct in this conclusion. In most patients with annular aortic discontinuity the better portion of the anulus is totally destroyed. The composite graft was sutured to the firm tissue underneath the destroyed anulus, consisting of the ventricular crest along the septum and a portion of the anterior leaflet of the mitral valve. When feasible, these sutures were taken transmurally for solid fixation of the conduit. Dr. Ivey. In this instance, we have always tried to reconstruct the anulus and have resorted to composite replacement only when the actual aortic wall had to be radically debrided. My second question: Why do you think that you and your group can be successful with such radical treatment in the placement of massive amounts of foreign body in an infected bed, when our colleagues in vascular surgery have been much less successful in the same setting? Dr. Ergin. Placement of the conduit in an "infected field" has been cause for concern. However, reinfection has not been a problem in practice. This is in keeping with the experience in prosthetic valve replacment during active endocarditis. The reinfection rate remains low even in the face of active infection. I believe the key element in the surgical treatment of these patients is excision of all infected tissue. Granted, it is sometimes difficult to tell the difference between infected tissues and sterile, attenuated areas. We would like to err on the side of wider excision and routinely go beyond all friable tissue until firm structure is encountered for the fixation of the conduit. I should also reemphasize the importance of longterm antibacterial treatment with appropriate agents in the postoperative period. Dr. hey. Is there any situation in which you would resort to translocation of the valve with coronary artery bypass grafting? Dr. Ergin. We have not encountered such a situation, and it is difficult for me to imagine one so long as the conduit is available and the surgeon is comfortable with the technique. The conduit proved to be an excellent solution in the eight patients in this series. There are two problems associated with translocation, which I believe are significant in the long run. The first one is related to the saphenous vein bypass grafts.
Acute bacterial endocarditis
763
Leaving a fairly young patient dependent on vein grafts, in our view, is less than an optimal solution. The second problem involves the fact that translocation of the aortic prosthesis leaves the subannular aneurysm open to systemic pressures and future complications. In fact, the original patient described by Dr. Danielson died of a rupture of the subannular aneurysm. Dr. Daniel A. Reid (Concord. Calif). With all the popularity with the allograft, at least all the publicity, is this a good place to use an allograft as a composite replacement? Is there any indication in which the allograft may be a preferential graft? Dr. Ergin. As I mentioned, the allograft has been used for this purpose. There is a theoretical advantage to the allograft, in that it may be more resistant to infection than a totally prosthetic conduit in a potentially infected field. However, some of these patients were treated before the allograft was widely available, and we have not used it in this setting. The allograft may prove to be the graft of choice when it is readily available. Dr. Reid. It seems to me we are starting to see some unusual infections with prosthetics, such as Q fever and aspergillosis. Have you had any experience with either of these two organisms and, if so, how have you treated them? Dr. Ergin. No, we have not seen either of these organisms. The only organism that was sort of unusual in this series was Actinobacillus in a patient with prosthetic endocarditis. Dr. Bruce W. Lytle (Cleveland. Ohio). Recently we treated two patients with infections involving the left fibrous trigone. Our approach was very similar to what you described, although we carried the patch down over the dome of the left atrium. We found that dividing the superior vena cava not only provided us with excellent exposure to that area, but also encouraged us to perform the kind of radical debridement of infected tissue that needs to be done. Have you used this technique and, if you haven't, what tricks do you have in terms of exposure of this area? Dr. Ergin. Unfortunately, the illustration does not show very clearly the approach to the left fibrous trigone. Of necessity, one has to go through the dome of the left atrium in this area. This incision can beclosed separately or incorporated into the patch repair. The incision through the base of the aorta and dome of the left atrium into the mitral anulus opens this area widely. We did not have to resort to additional maneuvers such as division of the superior cava for better exposure. In closing, annular destruction is an ever-increasing clinical problem because of the increasing number of patients with prosthetic valves at risk of endocarditis. Familiarity with these techniques will encourage the cardiac surgeon facing this difficult problem to perform the necessary extensive debridement for improved long-term results.
THORAC CARDIOVASC SURG
1989;97:755-63
Annular destruction in acute bacterial endocarditis Surgical techniques to meet the challenge Destruction and disruption of ventricular-aortic or mitral-aortic continuity in the presence of acute infection of the annular tissue is a significant surgical challenge. Among 82 patients who underwent surgical treatment for acute endocarditis over a lo-year period, 15 (18.2 %) had extensive destruction of the anulus necessitating special reconstructive techniques for treatment. Surgical treatment involved removal of all infected tissue including annular elements followed by appropriate restoration of the anulus for safe anchoring of the prosthetic valve. The reconstruction of the anulus consisted of the following: a Teflon felt patch inside and outside the aorta or ventricle, or both, for secure attachment of the prosthesis (felt aortic root, in three patients with native valveendocarditis), valvedcomposite graft replacement of the aortic root for ventricular-aortic discontinuity (Bentall procedure, in eight patients with prosthetic valve endocarditis), composite patch reconstruction of the mitral anulus and the ascending aorta to restore mitral-aortic continuity (mitral-aortic composite patch in two patients with mitral-aortic prosthetic valve endocarditis), and direct suture of the sewing skirts of the mitral and aortic prostheses to restore the defect (attached skirts, in one patient with mitral-aortic native valveendocarditis). There was one hospital death caused by multiple organ failure. The most common complication was heart block. Two late deaths were due to reinfection resulting from continued intravenous drug abuse. One patient with a felt aortic root repair required late reoperation for subannular aneurysm. Eleven patients were followed up from 7 months to 66 months and are alive and well without complications. This experience indicates that these seemingly radical surgical techniques can be used in these desperately ill patients with safety and good long-term results. They offer the only lasting solution for major disruption in cardiac anatomy in the presence of infection.
M. Arisan Ergin, MD, Sharo Raissi, MD, Fabrizio Follis, MD, Steven L. Lansman, MD, and Randall B. Griepp, MD, New York, N.Y.
h e role of surgical treatment in the management of bacterial endocarditis is well established. Timely valve replacement during the course of the disease is expected to lead to an excellent functional outcome with a low prevalenceof residual or recurrent infection." However, extension of the infectious process into the annular structures remains a major determinant of both the early and late results of surgical treatment of endocarditis.3-8 Varying degrees of annular involvement present a substantial challenge to the cardiac surgeon's ingenu-
From the Division of Cardiothoracic Surgery, The Mount Sinai Medical Center, New York, N.Y.
ity in restoring the anatomic and functional integrity of the cardiac structures after removal of infected tissues. Several surgical techniques have been described and used effectively, mostly in the treatment of aortic annular involvement. There appears to be a paucity of information in the literature regarding the management of destruction of the mitral anulus or the left fibrous trigone. This report is a summary of our clinical experience in a group of patients with extensive annular destruction requiring special reconstruction of the anulus for treatment. On the basis of this experience, indications for various techniques in the management of this increasingly common clinical problem are suggested.
Read at the Fourteenth Annual Meeting of The Western Thoracic Surgical Association in Hawaii, June 22-25, 1988.
Patients and methods
Address for reprints: M. Arisan Ergin, MD, Divisionof Cardiothoracic Surgery, The Mount Sinai Medical Center, One Gustave L. Levy Place, Box 1028, New York, NY 10029.
Among 82 patients who underwent surgical treatment for acute bacterial endocarditis between 1977 and 1987, 15 patients had extensive destruction of the anulus and required
755
The Journal of Thoracic and Cardiovascular Surgery
7 5 6 Ergin et al.
Table I. Prevalence of infecting organisms and culture results Cultures
Organism Staphylococcus aureus Streptococcus viridans Staphylococcus epidermidis Streptococcus faeca/is Pneumococcus Actinobacillus Gram-positive cocci Total
No. of patients
Blood
Valve
+
+
5 4
5 4
2
2
I I I
I I I
0 0 0 0 0 0
-.l
....Q
1
15
14
special reconstructive techniques for treatment. There were 13 men and two women aged from 20 to 74 years. Five patients had native valve endocarditis and 10 had prosthetic valve endocarditis. In 10 patients the infection was localized to the aortic valve and the anulus, in four it involved the mitral and aortic valves and anuli, and in one only the mitral anulus was affected. Eight patients were current or former drug abusers, and five had prosthetic endocarditis with only two thought to be active users. In two patients with prosthetic endocarditis the present procedure represented a third valve replacement. In patients with prosthetic valve endocarditis the interval from the original valve replacement was from 1.5 to 84 months (mean 17.3 ± 23.4 months). Fourteen patients had positive blood cultures, and the same organism was cultured from the excised valves in seven. Organisms were seen in the pathology specimen in 14 patients. One patient had negative blood and valve cultures, but organisms were found on pathologic examination. The most common organism was Staphylococcus aureus (Table I). All patients received varying courses of preoperative antibiotic treatment, the duration of which varied from 7 to 42 days (mean 16.9 ± 10.7 days). In all patients the diagnosis was based on clinical and bacteriologic findings supported by echocardiography. Five patients underwent cardiac catheterization, as well. The indication for operation in nine patients was deteriorating hemodynamics and persistent sepsis. Six patients were operated on because of increasing heart failure alone. One patient with double valve endocarditis was admitted with cardiogenic shock and renal shutdown, and another patient had peripheral embolization. All operations were performed with standard cardiopulmonary bypass and hypothermia. Single-dose cardioplegia supplemented with topical profound hypothermia was used for myocardial protection. Operative findings in patients with prosthetic endocarditis having isolated aortic annular involvement showed more than 50% destruction of the anulus, with subannular aneurysms measuring 2 to 4 cm. Ventricular-aortic continuity was completely disrupted in two of these patients, and two others had associated ventricular septal defects (Fig. I, A and B). All eight of these patients received a valved composite graft for the repair. In three patients with native valve endocarditis, the destruction of the aortic anulus was confined to less than 50% of the anulus in a contiguous manner, with subannular aneurysms measuring lessthan 2 ern. In all these a felt aortic root repair was possiblewith orthotopic
Pathology
I
+
0
I 3 I I
I
0
4 I I
5 3
0
2
0 0 0 0
Q
0
I
I I I
Q
1
-.l
7
8
14
I
I
implantation of the new prosthesis. Three patients with combined mitral-aortic endocarditis had significant destruction of the left fibrous trigone, -with aorta-left atrial fistulization in two (Fig. 2). In the patient with native valve endocarditis, the defect was repaired by direct suture of the mitral and aortic prostheses. In the other two with prosthetic endocarditis, extensive reconstruction with a composite mitral-aortic patch was necessary. In the only patient with isolated mitral annular destruction, partial ventricular translocation of the prosthesis was used (Table II). Surgical technique. In all patients, all infected tissue was removed and thorough debridement was done. This included removal of the valve and all annular elements that appeared to be involved. All abscesses were curetted until firm tissue was encountered. At times it was difficult to discern between infection and sterile destruction. Sutures for repair of the resultant defect were taken away from the edge in viable tissue, preferably with transmural buttressing. Inside and outside patches were used for tension-free closure of the defects. Five types of repair were used in this group of patients for specific indications. I. Felt aortic root. The felt aortic root technique was used in patients with native valve endocarditis with less than 50% destruction of the aortic anulus. After adequate debridement of the infected tissues, the resultant defect in the anulus was lined with a patch of Teflon felt or pericardium, which was sutured in place circumferentially with mattress sutures through healthy tissue. These sutures were tied over another Teflon strip on the outside, which in effect obliterated the defect in the anulus between two patches (one inside and the other outside) secured with through-and-through sutures. Teflon felt was used for its compliance for effectivebuttressing of sutures. Pericardium was preferred for inside patches to reduce the amount of prosthetic material facing the bloodstream. This technique lends itself best for defects along the noncoronary and left coronary cusps. In one patient the partially detached anterior leaflet of the mitral valve was also reattached to this patch. This method allows secure closure of limited subannular aneurysms and provides a solid base for securing the new prosthesis. 2. Valved composite graft replacement ofthe aortic root. A valved composite graft was used in patients with extensive destruction of the aortic anulus (more than 50%) with subannular aneurysms measuring more than 2 em or in noncontiguous areas of the anulus leading to partial or total
Volume 97 Number 5 May 1989
Acute bacterial endocarditis
757
Fig. 1. A, Aortic root injection showing complete ventricular-aortic discontinuity in patient 7 with prosthetic endocarditis. Note the separation of the ventricle from the aortic root with formation of large circumferential subannular aneurysm. B, Postoperative aortogram in the same patient after valved composite graft insertion. The distance between the prosthesisand the coronary orifices is apparent.
ventricular-aortic discontinuity. Interrupted pledget-backed mattress sutures were used for safe anchoring of the prosthesis on the ventricular side. In patients with an accompanying ventricular septal defect, transmural sutures were taken along the septum. The subannular aneurysmal area was completely excluded from the bloodstream. Coronary anastomoses and distal anastomosis were done as described for the original Bentall procedure. The periprosthetic space was closed with the native aorta and a patch of bovine pericardium for effective hemostasis (Fig. 3). 3. Attached skirts. The erosion of the left fibrous trigone in
a patient with native mitral and aortic endocarditis was treated by direct suture of the valve skirts in this area . The bulky sewing skirt of the Hancock mitral prosthesis was especially well suited for this purpose (Fig. 4). 4. Mitral-aortic composite patch. Modified adaptation of the method described for the enlargement of mitral and aortic anuli in double valve replacement was used in two patients with extensive destruction of the left fibrous trigone because of bivalvular prosthetic endocarditis. The aortotomy was extended into the left atrium through the destroyed left fibrous trigone, and after adequate excision of all infected tissue the
The Journal of Thoracic and Cardiovascular Surgery
7 5 8 Ergin et al.
Fig. 2. Aortic root injection in patient 13 with prosthetic double valve endocarditis and destruction of the left fibrous trigone. Ready opacification of the left atrium indicates aorta-left atrial fistula.
defect in the mitral anulus was restored with a composite patch of autogenous pericardium facing the bloodstream and knitted Dacron graft on the outside (Fig. 5, A, B, and C). This triangular patch was used to anchor the mitral prosthesis in this region and the aortic prosthesis slightly higher up at its base. The rest of the patch was then extended and incorporated into the closure of the aortotomy (Fig. 5, D and E). 5. Partial translocation of the mitral value. In a patient with destruction of the posterior mitral anulus, the anchoring sutures for the prosthesis were taken in the ventricle. The valve was partially translocated into the ventricle along the posterior anulus, and the weakened area of the anulus was left on the atrial side of the repair facing low pressure.
Results Complete reconstruction of the anulus with implantation of the appropriate prosthesis was accomplished in all patients with the described techniques. There were no intraoperative deaths. The only hospital death occurred in a patient with native aortic and mitral endocarditis who was admitted in cardiogenic shock and renal shutdown. He required felt repair of the noncoronary cusp and part of the left coronary cusp area with aortic and mitral valve replacement. He died 45 days after the operation with multiple organ failure. At autopsy both the repair and the prosthetic valves were intact. The most common complication was heart block. Only one patient had preoperative complete heart block, whereas seven patients had postoperative block. Six patients had heart block among the group of eight receiving composite grafts. Five patients received pacemakers at the same operation and two later. Two serious
gastrointestinal tract complications were encountered. Ischemic colitis in one patient responded to conservative treatment, and upper gastrointestinal tract bleeding required gastrectomy in another. The same patient also had a postoperative cerebrovascular accident but recovered without sequelae. There were no residual infections. All patients received appropriate antibiotic treatment for from 3 to 8 weeks postoperatively. Fourteen patients were discharged from the hospital and were available for follow-up, ranging from 7 to 66 months (mean 23.7 ± 15.0 months). Eleven patients are alive and well without complications. There were two late deaths resulting from reinfection caused by continued drug abuse. One occurred in a patient with a valved conduit at 12 months, and the other occurred in the patient with double valve replacement and attached skirts at 35 months. Neither patient had a repeat operation. In both of these patients the original offending organism was Staphylococcus aureus, and the long interval between repair and recurrence of endocarditis suggests reinfection because of intravenous drug abuse rather than recurrence. There was one late failure of felt repair of the aortic root at 12 months. This patient had a large subannular aneurysm as a result of dehiscence of the lower aspect of the inside patch without evidence of reinfection and was treated with a valved conduit replacement (Table II).
Discussion Annular involvement during the course of endocarditis carries important diagnostic, therapeutic, and prognostic significance.2-6, 9,10 The prevalence of annular destruction in native valve endocarditis is related to the virulence of the responsible organism and the duration of the process.2, 5, II Varying degrees of annular involvement are a constant feature of prosthetic endocarditiS. 3, 12-17 The spectrum of annular destruction extends from simple localized annular abscess formation to larger subannular aneurysms, with or without perforation into other cardiac chambers, extension into the pericardial space, or total disruption of ventricularaortic continuity and the mitral-aortic fibrous trigone." 18-23 Early surgical treatment in patients with suspected annular involvement is indicated to limit the annular destruction and its sequelae." 3, 22 Advanced annular destruction necessitates specific surgical techniques for effective treatment. In this situation the surgeon is confronted with the challenge of removing all infected tissue while preserving the functional and anatomic integrity of the left ventricular inlet and outlet.
Volume 97 Number 5
Acute bacterial endocarditis
May 1989
759
Table II. Operative findings, surgical techniques, and results Patient No.
Age (yr)
Annular involvement
2
74
LV-Ao discontinuity + VSD 75%
3
36
50%
4
24
75%
5
29
50%
6
52
75%
7
29
LV-Ao discontinuity
8
29
75%+ VSD
9
24
LCS, 2 ern
10
57
LCS, 1.5 em
11
23
NCS 2.0 crn
12
20
Left fibrous trigone
13
50
Left fibrous trigone
14
25
Left fibrous trigone
15
29
Post mitral anulus
27
Technique
Conduit, 23 mm B-S valve Conduit, 23 mm B-S valve Conduit, 23 mm B-S valve Conduit, 23 mm B-S valve Conduit, 23 mm B-S valve Conduit, 27 mm B-S valve Conduit, 23 mm B-S valve Conduit, 27 mm B-S valve Felt outside and inside, AVR Felt outside, pericardium inside, AVR, MVR Felt outside and inside, AVR Attached skirts, AVR, MVR Mitral-aortic patch, AVR, MVR Mitral-aortic patch, AVR, MVR Partial translocation
Complications
Follow-up
Heart block
Alive, 16 mo
Gastrectomy, CVA
Alive, 17 mo
Heart block
Late death, 12 mo Alive, 27 mo
Heart block Heart block
Alive, about 12 mo Alive, 66 mo
Heart block
Alive, 12 mo
Heart block
Alive, 30 rno Alive, 19 mo
Multiorgan failure
Hospital death
Reop, 12 rno
Ischemic colitis
Late death, 36 mo Alive, 36 rno Alive, 32 rno Alive, reop at 30 rno
LV-Ao, Left ventricular-aortic; VSD, ventricular septal defect; LCS, left coronary sinus; NCS, noncoronary sinus; B-S, Bjork-Shiley; A VR, aortic valve replacement; MVR, mitral valve replacement; CV A, cerebrovascular accident.
The aortic anulus is the most frequently involved structure of the fibrous skeleton of the heart." Several effectivesurgical techniques have been described for the repair of aortic annular disruption. Small abscesses of the aortic anulus may be debrided and closed or incorporated into the valve sutures. 3, 25 Larger subannular aneurysms have been treated by exclusion with prosthetic or pericardial patches.v" 10, 18, 19, 26 For more severe forms of destruction of the aortic anulus with ventricular-aortic discontinuity, aortic root replacement with a valved composite graft or homograft or distal translocation of the aortic valve with saphenous vein grafts has been used." 22, 25, 27-30 We believe it is important to excise all grossly infected structures to reduce the risk of residual infection or late valve dehiscence. Familiarity with these techniques leads to more confident use of debridement. The defect created in the
anulus after removal of all infected tissue dictates the choice of the reconstructive technique. It is best to reserve simple plication for only small cavities. Larger cavities, those less than 2 em in contiguous areas of the aortic anulus, should be treated with one of the patch methods. This approach allows tension-free closure of the aneurysm with sutures placed well beyond the limits of the excision in viable tissue. We have used a double patch technique with felt or preferably pericardium on the blood-contacting surface and compliant prosthetic material on the outside (felt aortic root). This method is best suited for defects along the better portions of the left and noncoronary sinuses. Defects along other areas of the aortic anulus can be closed by techniques described by others.v'? The patch provides a sturdy structure for the fixation of the new prosthesis while excluding the aneurysmal area from the high-pressure
760
Ergin et al.
Fig. 3. Artist's concept of ventricular-aortic discontinuity and its treatment withvalved composite graft. Note buttressed transmural sutures along the crest of the left ventricle for secure fixation of the prosthesis. The closure of the native aorta helps with immediate hemostasis.
system. The double patch technique allows use of autogenous pericardium inside with the prosthetic patch outside for secure fixation of the valve. In the more extensive annular destructions that we encountered with prosthetic endocarditis, secure annular repair could not be accomplished with patch techniques. In patients with large subannular aneurysms and more than 50% destruction of the circumference of the anulus, implantation of a valved composite graft was the most expeditious and secure repair. This method in the treatment of ventricular-aortic discontinuity was first described by Frantz, Murray, and Wilcox" and since then has been used by others. 10, 25, 27 The conduit totally excludes the attenuated area from the systemic pressure and has proved to be a safe and durable repair. Transmural sutures, as described by others, help secure fixation of the composite graft to the ventricular crest.": 28 The anastomosis of the coronary ostia is not difficult, since the subannular separation leaves adequate distance between the prosthesis and the coronary ostia. A composite homograft, as described by Lau and associates," may be preferable on theoretical grounds. However, the availability of the prosthetic conduit and familiarity with the technique playa role in this choice. The alternative method of distal translocation of the prosthesis with closure of the coronary ostia and saphenous vein grafting is a more complex operation that leaves the patient at risk for aneurysmal dilatation of the annular area, as well as dependent on vein grafts.": 29,30
The Journal of Thoracic and Cardiovascular Surgery
Fig. 4. Attached skirtsusedfor repairof simple disruption of the mitral-aortic trigone. Direct suture of the sewing skirts of the prostheses wassufficient to bridge the defect in the patient with native mitral-aortic endocarditis,
Both of these problems have been seen in patients with translocation of the aortic valve.v" Complete heart block was the most frequent complication of valved conduit repair of the aortic root in our experience, the price paid for secure fixation of the prosthesis. There is a paucity of information in the literature about the management of the mitral anulus and aorticmitral continuity in endocarditis." 25 We have encountered four patients with aortic-mitral discontinuity, Destruction of the left fibrous trigone, especially in the presence of prosthetic endocarditis, led to a large defect at the junction of the aortic and mitral anuli with serious physiologic and anatomic consequences. Adaptation of the technique that was originally described by Rastan and associates" for the enlargement of the combined anuli for mitral and aortic valve replacement proved to be an effective solution to this otherwise devastating defect. The composite patch fashioned from autogenous pericardium and knitted graft material easily bridges the defect in the mitral anulus and provides for secure fixation of the mitral prosthesis in this area. The aortic prosthesis is likewise secured to the patch distally. The remainder of the patch is then incorporated into the closure of the aortotomy, so that the left fibrous trigone is completely reconstructed. Pericardium leaves autogenous tissue facing the bloodstream and provides for absolute hemostasis, while the compliant knitted graft material serves as structural support against systemic pressure and provides for solid fixation of the prostheses.
Volume 97 Number 5 May 1989
Acute bacterial endocarditis
76 I
Fig. 5. Steps in the reconstruction of extensive destruction of the left fibrous trigone with a composite patch. See text for details.
We have not found a description of previous use of this technique in the management of endocarditis. Isolated destruction of the mitral anulus is rare in endocarditis." Rocchiccioli and associates" described attachment of a Dacron flange to the mitral prosthesis for atrial implantation of the prosthesis in cases in which the mitral anulus could not be used for fixation of the valve. This technique leaves the attenuated portion of the anulus in the high-pressure chamber at risk for aneurysmal dilatation or rupture as a late complication, especially along the posterior anulus. Therefore, in the only patient with destruction of the posterior .mitral anulus whom we encountered, a partial ventricular translocation of the prosthesis was used. At reoperation for degeneration of the porcine valve, the sewing skirt was found to be firmly attached without any evidence of dehiscence. This experience indicates that serious complications of annular destruction in endocarditis can be treated effectively by adhering to sound surgical principles of thorough debridement, followed by tension-free reconstruction of the anulus that excludes all attenuated structures from systemic pressure and provides a solid fixation point for the prosthesis. These seemingly radical surgical techniques were used with an acceptable opera-
tive risk and good long-term results and seem to be the only lasting solution to the major disruption of cardiac anatomy in the presence of infection. REFERENCES I. Young JB, Welton DE, Raizner AE, et ai. Surgery in active infective endocarditis. Circulation 1979;60(Pt 2):177-8 I. 2. D'Agostino RS, Miller DC, Stinson EB, et ai. Valve replacement in patients with native valve endocarditis: What really determines operative outcome? Ann Thorac Surg 1985;40:429-38. 3. Nelson RJ, Harley, DP, French WJ, Bayer AS. Favorable ten-year experience with valve procedures for active infective endocarditis. J THORAC CARDIOVASC SURG 1984;87:493-502. 4. Richardson JV, Karp RB, Kirklin JW, Dismukes WE. Treatment of infective endocarditis: a lO-year comparative analysis. Circulation 1978;58:589-97. 5. Arnett E, Roberts W. Valve ring abscess in active infective endocarditis: frequency, location and clues to clinical diagnosis from the study of 95 necropsy patients. Circulation 1976;54:140-5. 6. Baumgartner WA, Miller CD, Reitz BA, et ai. Surgical treatment of prosthetic valve endocarditis. Ann Thorac Surg 1983;35:87-104.
762
The Journal of Thoracic and Cardiovascular Surgery
Ergin et al.
7. Wilcox BR, Murray GF, Starek PJK. The long-term outlook for valve replacement in active endocarditis. J THORAC CARDIOVASC SURG 1977;74:860-3. 8. Calderwood SB, Swinski RN, Karchmer AW, Waternaus CM, Buckley MJ. Prosthetic valve endocarditis. J THoRAC CARDIOVASC SURG 1986;92:776-83. 9. Lau JM, Guinn GA, Beall AC Jr, et al. Operative techniques in infective endocarditis. Ann Thorac Surg 1981;32:351-6. 10. Fiore AC, Ivey TD, McKeown PP, et al. Patch closure of aortic annulus mycotic aneurysms. Ann Thorac Surg 1986;42:372-8. II. Rapaport E. The changing role of surgery in the management of infective endocarditis [Editorial]. Circulation 1978;58(Pt 2):II-IV598-9. 12. Arnett EN, Roberts We. Prosthetic valve endocarditis: clinicopathologic analysis of 22 necropsy patients with comparison of observations in 74 necropsy patients with active infective endocarditis involving natural left-sided cardiac valves. Am J Cardiol 1976;38:281-92. 13. Dismukes WE, Karchmer AW, Buckley MJ, et al. Prosthetic valve endocarditis: analysis of 38 cases. Circulation 1973;48:365-77. 14. Slaughter L, Morris JE, Starr A. Prosthetic valvular endocarditis: a 12-year review. Circulation 1973;47:131926. 15. Wilson WR, Jaumin PM, Danielson GK, et al. Prosthetic valve endocarditis. Ann Intern Med 1975;82:751-6. 16. Stinson EB, Griepp RB, Vosti K, et al. Operative treatment of active endocarditis. J THORAC CARDIOVASC SURG 1976;71:659-64. 17. Symbas PN, Vlasis SE, Zacharopoulos L, et al. Immediate and longterm outlook for valve replacement in acute bacterial endocarditis. Ann Surg 1982;195:721-5. 18. Todd EP, Hubbard SG, Zeok JV, et al. Repair of mycotic aneurysms of the aorta involving the aortic valve. Ann Thorac Surg 1980;30:160-6. 19. Bailey WW, Ivey TD, Miller DW Jr. Dacron patch closure of aortic annulus mycotic aneurysms. Circulation 1982;66(Pt 2):1127-30. 20. Lau JKH, Robles A, Cherian A, et al. Surgical treatment of prosthetic endocarditis. J THORAC CARDIOVASC SURG 1984;87:712-6. 21. Danielson GK, Titus JL, DuShane JW. Successful treatment of aortic valve endocarditis and aortic root abscesses by insertion of prosthetic valve in ascending aorta and placement of bypass grafts to coronary arteries. J THoRAc CARDIOVASC SURG 1974;67:443-9. 22. Frantz PT, Murray GF, Wilcox BR. Surgical management of left ventricular-aortic discontinuity complicating bacterial endocarditis. Ann Thorac Surg 1987;29:1-7. 23. Chesler E, Korns ME, Porter GE, Reyes CN, Edwards JE. False aneurysm of the left ventricle secondary to bacterial endocarditis with perforation of the mitral-aortic intervalvular fibrosa. Circulation 1968;37:518-23. 24. Cowgill DL, Addonizio P, Hopeman AR, Harken AH. A
25.
26.
27.
28.
29.
30.
31.
practical approach to prosthetic valve endocarditis. Ann Thorac Surg 1987;43:450-7. Rocchiccioli C, Chastre J, Lecompte Y, Gandjbakhch I, Gibert C. Prosthetic valve endocarditis. J THoRAc CAROlOVASC SURG 1986;92:784-9. Olinger GN, Maloney JV Jr. Repair of left ventricularaortic discontinuity complicating endocarditis from an aortic valve prosthesis. Ann Thorac Surg 1977;23:576-7. VanHooser DW, Johnson RG, Hein RA, Elkins nc Successful management of aortic valve endocarditis with associated periannular abscess and aneurysm. Ann Thorae Surg 1986;42:148-51. Rumisek JD, Berry WR, Barry MJ, Clarke JS. Transseptal control of the difficult aortic annulus. Ann Thorae Surg 1985;39:385-6: Stinson EB. Surgical treatment of infective endocarditis. Progress in cardiovascular diseases. Vol 22, New York: Grone & Stratton, 1979. Reitz BA, Stinson EB, Watson· DC, et al. Translocation of the aortic valve for prosthetic valve endocarditis. J THoRAC CARDIOVASC SURG 1981;81:212-8. Rastan H, Atai M, Hadi H, Yazdanyar A. Enlargement of mitral valvular ring. J THORAC CARDIOVASC SURG 1981;81:106-11.
Discussion Dr. Tom D. hey (Seattle. Wash.). This paper is an important milestone in the treatment of endocarditis. The Stanford group has shown us that early intervention is important in achieving acceptable results in these difficult patients. They have also, at times and in certain situations, advocated translocation of the aortic valve and coronary artery bypass grafting. Dr. Warren Bailey and Dr. Andy Fiore have previously reported our techniques of closing annular abscesses with patches, allowing normal positioning of the aortic prosthesis. The authors have presented an extraordinary series of 15 patients with bacterial endocarditis and with only one perioperative death. I might add that this was before the era of substrate cardioplegia and modern myocardial protection techniques. Ten of 15 patients had prosthetic endocarditis and, in four of them, both the aortic and mitral anuli were destroyed. Staphyloccocal organisms were most commonly present. We strongly agree with the authors' approach to radical debridement of all infected tissue, and we believe this is the key element in their success. Heart block is certainly acceptable in this situation. The mean duration of antibiotic treatment in this series was 17 days before surgical intervention. We recently saw a patient with a Starr-Edwards prosthesis who was inadequately anticoagulated and had been referred to the neurology service with an embolic stroke. The patient had a heparin lock placed and 3 days after the institution of this intravenous therapy, had a fever spike and subsequently a positive blood culture for Staphylococcus aureus. Within 48 hours, the patient had prolongation of the PR interval and was immediately operated on. In this short period of time, the anulus under the prosthesis was totally destroyed. The prosthesis was still seated, but roughly 90% of the annular tissue was
Volume 97 Number 5 May 1989
eroded with purulent material. Thus annular destruction can certainly occur within 48 hours with staphyloccocal organisms if a prosthesis is in place. Unfortunately, the decision for early operation, as I am sure the authors would agree, is not often in the hands of the surgeon. The treatment of combined aortic and mitral prosthetic endocarditis, as outlined by the authors, is also an important contribution. I have several questions: What type of suture technique do you use to anchor the composite grafts on the ventricular septum? If I understand your manuscript, this is not an annular location but is subannular. Am I correct in this conclusion? Dr. Ergin. Yes, you are correct in this conclusion. In most patients with annular aortic discontinuity the better portion of the anulus is totally destroyed. The composite graft was sutured to the firm tissue underneath the destroyed anulus, consisting of the ventricular crest along the septum and a portion of the anterior leaflet of the mitral valve. When feasible, these sutures were taken transmurally for solid fixation of the conduit. Dr. Ivey. In this instance, we have always tried to reconstruct the anulus and have resorted to composite replacement only when the actual aortic wall had to be radically debrided. My second question: Why do you think that you and your group can be successful with such radical treatment in the placement of massive amounts of foreign body in an infected bed, when our colleagues in vascular surgery have been much less successful in the same setting? Dr. Ergin. Placement of the conduit in an "infected field" has been cause for concern. However, reinfection has not been a problem in practice. This is in keeping with the experience in prosthetic valve replacment during active endocarditis. The reinfection rate remains low even in the face of active infection. I believe the key element in the surgical treatment of these patients is excision of all infected tissue. Granted, it is sometimes difficult to tell the difference between infected tissues and sterile, attenuated areas. We would like to err on the side of wider excision and routinely go beyond all friable tissue until firm structure is encountered for the fixation of the conduit. I should also reemphasize the importance of longterm antibacterial treatment with appropriate agents in the postoperative period. Dr. hey. Is there any situation in which you would resort to translocation of the valve with coronary artery bypass grafting? Dr. Ergin. We have not encountered such a situation, and it is difficult for me to imagine one so long as the conduit is available and the surgeon is comfortable with the technique. The conduit proved to be an excellent solution in the eight patients in this series. There are two problems associated with translocation, which I believe are significant in the long run. The first one is related to the saphenous vein bypass grafts.
Acute bacterial endocarditis
763
Leaving a fairly young patient dependent on vein grafts, in our view, is less than an optimal solution. The second problem involves the fact that translocation of the aortic prosthesis leaves the subannular aneurysm open to systemic pressures and future complications. In fact, the original patient described by Dr. Danielson died of a rupture of the subannular aneurysm. Dr. Daniel A. Reid (Concord. Calif). With all the popularity with the allograft, at least all the publicity, is this a good place to use an allograft as a composite replacement? Is there any indication in which the allograft may be a preferential graft? Dr. Ergin. As I mentioned, the allograft has been used for this purpose. There is a theoretical advantage to the allograft, in that it may be more resistant to infection than a totally prosthetic conduit in a potentially infected field. However, some of these patients were treated before the allograft was widely available, and we have not used it in this setting. The allograft may prove to be the graft of choice when it is readily available. Dr. Reid. It seems to me we are starting to see some unusual infections with prosthetics, such as Q fever and aspergillosis. Have you had any experience with either of these two organisms and, if so, how have you treated them? Dr. Ergin. No, we have not seen either of these organisms. The only organism that was sort of unusual in this series was Actinobacillus in a patient with prosthetic endocarditis. Dr. Bruce W. Lytle (Cleveland. Ohio). Recently we treated two patients with infections involving the left fibrous trigone. Our approach was very similar to what you described, although we carried the patch down over the dome of the left atrium. We found that dividing the superior vena cava not only provided us with excellent exposure to that area, but also encouraged us to perform the kind of radical debridement of infected tissue that needs to be done. Have you used this technique and, if you haven't, what tricks do you have in terms of exposure of this area? Dr. Ergin. Unfortunately, the illustration does not show very clearly the approach to the left fibrous trigone. Of necessity, one has to go through the dome of the left atrium in this area. This incision can beclosed separately or incorporated into the patch repair. The incision through the base of the aorta and dome of the left atrium into the mitral anulus opens this area widely. We did not have to resort to additional maneuvers such as division of the superior cava for better exposure. In closing, annular destruction is an ever-increasing clinical problem because of the increasing number of patients with prosthetic valves at risk of endocarditis. Familiarity with these techniques will encourage the cardiac surgeon facing this difficult problem to perform the necessary extensive debridement for improved long-term results.