Surgical Treatment of Mitral Valve Endocarditis in North America

James S. Gammie, MD, Sean M. O’Brien, PhD, Bartley P. Griffith, MD, and Eric D. Peterson, MD Division of Cardiac Surgery, University of Maryland Medical Center, Baltimore, Maryland, and Duke Clinical Research Institute, Durham, North Carolina

Background. Several single-institution series have suggested the feasibility and effectiveness of mitral valve repair for infective endocarditis (IE). Methods. We examined 6627 patients with IE undergoing mitral valve surgery at 661 Society of Thoracic Surgeons–participating centers in 1994 to 2003. Results. The diagnosis of IE was assigned to 5.8% (6,627 of 114,934) of patients having mitral valve surgery. The overall frequency of mitral valve repair for IE was 29.7% (1,965 of 6,627). Mitral valve repair was less frequently used for patients with active IE (423 of 2,654; 15.9%) than those with treated IE (1,459 of 3,570; 40.9%). Operative mortality was 3.7% (72 of 1,965) for mitral valve repair and 10.8% (502 of 4,662) for mitral valve replacement. Mortality rates were lower for patients with treated IE compared with active IE. After adjusting for

multiple preoperative risk factors, mitral valve repair (odds ratio, 0.67; 95% confidence interval, 0.51 to 0.88) was associated with a significantly lower risk of death. Active (versus treated) IE (odds ratio, 2.12; 95% confidence interval, 1.68 to 2.68) and recent cerebrovascular accident (odds ratio, 1.71; 95% confidence interval, 1.28 to 2.31) were independent predictors of mortality. Conclusions. Mitral valve repair is less commonly applied for IE compared with other indications for mitral valve surgery. Patients with active IE were less likely to receive repair than those with treated IE. Mitral valve repair was associated with a lower risk of mortality. These results provide support for performing mitral valve repair when technically feasible in the setting of IE. (Ann Thorac Surg 2005;80:2199 –204) © 2005 by The Society of Thoracic Surgeons

B

outcomes of mitral valve repair and characterize the current state of mitral valve surgery for IE in a large patient cohort, we examined the clinical characteristics and outcomes of all patients having surgery for native mitral valve endocarditis during the last decade in North American institutions participating in The Society of Thoracic Surgeons (STS) National Cardiac Database.

efore the advent of antibiotics and cardiac surgery, infective endocarditis (IE) was a uniformly fatal disease [1]. Despite improvements in diagnosis and therapy, the in-hospital mortality rate for IE remains between 20% and 25% [1–3], and the incidence has not changed during the last two decades [4]. Appropriately timed and performed surgical therapy of the infected heart valve contributes substantially to reduced early and late mortality [5]. Although the infected aortic valve is uncommonly amenable to repair, established repair techniques can be applied to patients with mitral valve endocarditis. The advantages of mitral valve repair compared with replacement are well established for noninfectious mitral valve disease and include better preservation of left ventricular function, low perioperative mortality, avoidance of long-term anticoagulation, low long-term rate of thromboembolic complications, low risk of IE, excellent freedom from reoperation, and improved long-term survival [6]. Several small single-institution series have established the feasibility, effectiveness, and durability of mitral valve repair for IE [7–10]. To determine the incidence and Accepted for publication May 12, 2005. Presented at the Forty-first Annual Meeting of The Society of Thoracic Surgeons, Tampa, FL, Jan 24 –26, 2005. Address correspondence to Dr Gammie, Division of Cardiac Surgery, University of Maryland Medical Center, N4W94, 22 S Greene St, Baltimore, MD 21201; e-mail: [email protected].

© 2005 by The Society of Thoracic Surgeons Published by Elsevier Inc

Material and Methods The Society of Thoracic Surgeons National Cardiac Database The STS National Cardiac Database (NCD) was initiated in 1986 and now contains detailed clinical information on more than 1.5 million patients undergoing cardiac surgery at 661 sites in North America [11]. The STS National Cardiac Database includes more than 409,000 patients having valvular heart surgery between 1994 and 2003.

Patients The study population consists of all adult patients having first-time mitral valve operations between 1994 and 2003. Patients having concomitant coronary artery bypass graft surgery or tricuspid valve repair were included, whereas combined aortic and mitral valve operations were excluded. We examined the character0003-4975/05/$30.00 doi:10.1016/j.athoracsur.2005.05.036

CARDIOVASCULAR

Surgical Treatment of Mitral Valve Endocarditis in North America

2200

GAMMIE ET AL SURGERY FOR MITRAL VALVE ENDOCARDITIS

istics of patients having mitral valve surgery with a diagnosis of IE as well as those having mitral valve operations without a diagnosis of IE. CARDIOVASCULAR

Definitions Infective endocarditis was defined as active if the patient was receiving antibiotic therapy for endocarditis at the time of surgery. Treated IE patients were not receiving antibiotic therapy other than perioperative prophylactic therapy. All other clinical definitions were standardized based on specifications of core data elements that can be accessed at http://www.sts.org/doc/8428. Operative mortality was defined as death occurring during the hospitalization in which surgery took place, as well as those occurring after discharge but within 30 days of surgery.

Statistical Analysis The frequency of selected clinical characteristics and operative outcomes was compared across subgroups of patients, with subgroups defined by IE status (yes/no), treatment status (active/treated, not infected), procedure type (repair, replacement), and year of surgery. Differences between subpopulations with respect to a binary (yes/no) variable were tested using ␹2 tests. Changes in the frequency of a binary (yes/no) variable with time were assessed using the Cochran-Armitage ␹2 trend test. Multivariable logistic regression was used to assess the association between patient preoperative characteristics and risk of mortality among patients undergoing mitral surgery for IE. Patients without IE or with unknown treatment status (active/treated) were excluded. Explanatory variables were chosen based on existing STS risk adjustment models for isolated aortic or mitral valve replacement and for combined aortic or mitral valve replacement plus coronary artery bypass grafting [12]. Preoperative variables included age, body surface area, cerebrovascular accident, diabetes, renal failure, dialysis, hypertension, chronic lung disease, peripheral vascular disease, reoperation, New York Heart Association class IV, pulmonary hypertension, myocardial infarction, status, preoperative intraaortic balloon pump, immunosuppressive therapy, sex, triple-vessel disease, and ejection fraction. Additional covariates included treatment status (active/treated) and an indicator variable for concomitant coronary artery bypass grafting. Initially, separate models were fit for patients undergoing concomitant CABG surgery and for patients undergoing isolated valve surgery. These models were subsequently combined because the results were similar. For presentation purposes, the model was simplified using a backward selection algorithm.

Results Incidence of Infective Endocarditis Among Patients Having Mitral Valve Surgery Between 1994 and 2003, 6,627 patients with IE meeting the inclusion criteria had mitral valve repair or replacement at 661 STS-participating centers. The diagnosis of

Ann Thorac Surg 2005;80:2199 –204

Table 1. Characteristics of Patients Having Mitral Valve Surgery Based on Presence or Absence of Infective Endocarditis and Treatment Status (Active or Treated)a Variable

Treated

Active

All IE

No IE

N Male (%) Age (median) African American (%) BMI (median) DM (%) Hypercholesterolemia (%) Renal failure (%) Renal failure hemodialysis (%) CVA (%) MI (%) CHF (%) Angina (%) Cardiogenic shock (%) NYHA class IV (%) Triple-vessel disease (%) EF (median) Mitral stenosis (%) Status, emergent (%) Status, elective (%) CABG (%)

3,570 61.2 58 10.0 24.8 12.5 22.4 11.3 5.4

2,654 58.3 54 15.1 24.6 21.5 17.3 28.9 15.4

6,627 59.7 56 12.3 24.7 16.1 20.2 18.7 9.5

103,974 48.4 67 6.4 26.0 19.0 35.3 6.6 1.7

14.2 6.9 47.3 15.5 1.8 16.6 6.6 60 13.7 1.9 77.5 23.5

25.5 10.8 54.1 13.0 10.7 37.3 5.4 60 10.6 15.4 36.1 19.9

19.0 8.6 50.5 14.6 5.5 25.3 6.1 60 12.5 7.7 60.1 22.0

a

7.1 24.4 54.1 35.8 4.2 22.8 24.7 50 21.1 4.1 72.8 47.2

p value ⱕ 0.0001 for all comparisons of IE versus no IE.

BMI ⫽ body mass index; CABG ⫽ coronary artery bypass grafting; CHF ⫽ congestive heart failure; CVA ⫽ cerebrovascular accident; DM ⫽ diabetes mellitus; EF ⫽ ejection fraction; IE ⫽ infective endocarditis; MI ⫽ myocardial infarction; NYHA ⫽ New York Heart Association.

infective endocarditis was assigned to 5.8% (6,627 of 114,934) of patients having mitral valve operations during the study period. IE status was missing for 3.8% of patients (4,333 of 114,934).

Characteristics of Infective Endocarditis Patients Preoperative characteristics of patients having mitral valve surgery are detailed in Table 1. Compared with patients without IE, patients having mitral valve surgery for IE were younger (median age, 56 versus 67 years), were more likely to be in renal failure before surgery (18.7% versus 6.6%), were more likely to have had a cerebrovascular accident (19.0% versus 7.1%), were less likely to have a history of myocardial infarction, angina, or three-vessel disease, and were less likely to require coronary artery bypass grafting at the time of surgery (22.0% versus 47.2%; all p ⬍ 0.0001). Among patients with IE, 40% were currently taking antibiotics (active), 54% had completed antibiotic therapy (treated), and 6% were of unknown treatment status. Compared with patients with treated IE, patients with active IE were more likely to have comorbidities including renal failure requiring hemodialysis (15.4% versus 5.4%), cerebrovascular accident (25.5% versus 14.2%), and diabetes (21.5% versus 12.5%). Active IE patients

Ann Thorac Surg 2005;80:2199 –204

GAMMIE ET AL SURGERY FOR MITRAL VALVE ENDOCARDITIS

2201

Variable

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

Number Repair (%)a Treateda Activeb Active IE (%)a

400 26.8 38.8 11.2 35.8

497 27.8 34.7 16.2 38.4

512 23.4 34.1 10.3 39.7

632 26.1 36.4 12.9 36.9

702 27.8 40.0 12.7 36.0

639 27.1 38.5 17.0 35.8

671 31.6 40.2 20.5 41.4

728 30.9 40.7 16.4 40.1

863 33.7 47.0 16.3 43.3

983 34.5 48.4 19.2 46.6

a

Test of trend p ⬍ 0.0001.

b

p ⫽ 0.0017.

IE ⫽ infective endocarditis.

were more likely than treated IE patients to have emergent or salvage surgery (15.4% versus 1.9%), were less likely to undergo elective surgery (36.1% versus 77.5%), and were more likely to be assigned New York Heart Association class IV before surgery (37.3% versus 16.6%; all p ⬍ 0.0001). The ratio of active to all IE patients undergoing operation increased from 35.8% to 46.6% during the period of the study, a relative increase of 31% (p ⬍ 0.0001) (Table 2).

Use of Mitral Valve Repair Performance of mitral valve repair among patients with IE during this study was 29.7% (1,965 of 6,627), whereas the repair rate among patients without IE during the same period was 45.7% (47,561 of 103,974; p ⬍ 0.0001). During the decade studied, the frequency of mitral valve repair for IE increased from 27% to 34% (p ⬍ 0.0001). The use of repair for non-IE patients in the same study period increased from 33.26% to 59.28% (p ⬍ 0.0001). Among patients having mitral valve surgery for IE, those that were treated with mitral valve repair (versus replacement) had less renal failure requiring hemodialysis (5.0% versus 11.4%), fewer strokes (14.3% versus 21.0%), less congestive heart failure (38.8% versus 55.5%), and less diabetes (11.1% versus 18.2%; all p ⬍ 0.0001). Mitral valve repair was less frequently used for patients with active IE (423 of 2,654, 15.9%) than those with treated IE (1,459 of 3,570, 40.9%; p ⬍ 0.0001).

Outcomes For patients with mitral valve IE undergoing repair, mortality was 3.7% (72 of 1,965) versus 10.8% (502 of 4,662) for mitral valve replacement (p ⬍ 0.0001). The overall mortality rate for patients having mitral valve operations for IE ranged from 6.5% to 10.2% during the decade studied (mean, 8.6%), without a significant change with time. Mortality rates were lower in patients with treated IE compared with active IE (Table 3). Cross-clamp and cardiopulmonary bypass times were clinically similar for repair and replacement. Postoperatively, blood products were required less frequently (46.6% versus 67.3%), length of stay was shorter (6 versus 9 days), and renal failure was less common (10.5% versus 4.7%) in the repair group (p ⬍ 0.0001).

Multivariable Analysis Multivariable logistic regression was used to identify important independent risk factors for operative mortal-

ity in the context of mitral surgery for IE (Table 4). In addition to risk factors previously identified as predictors of mortality in existing STS risk adjustment models for isolated valve replacement, active (versus treated) IE (odds ratio, 2.12; 95% confidence interval, 1.68 to 2.68) and recent (⬍2 weeks) cerebrovascular accident (odds ratio, 1.71; 95% confidence interval, 1.28 to 2.31) were independent predictors of mortality. The association between repair versus replacement and operative mortality was assessed by adding a variable for repair (yes/no) to the model that contained patient preoperative characteristics. Mitral valve repair (versus replacement) remained associated with a significantly lower risk of death (odds ratio, 0.67; 95% confidence interval, 0.51 to 0.88).

Comment The results of this study characterize the application of mitral valve surgery for IE during the last decade. Table 3. Outcomes of Mitral Valve Surgery Among Infective Endocarditis Patients Variable Mortality Treated Active Bleeding Treated Active Sternal infection Treated Active Stroke Treated Active Prolonged ventilation Treated Active Renal failure Treated Active PLOS ⬎ 14 d Treated Active PLOS ⫽ patient length of stay.

Repair (%)

Replace (%)

All (%)

3.7 1.8 10.6 4.0 3.6 5.2 0.2 0.1 0.2 2.9 1.9 6.4 8.6 5.1 20.8 4.7 2.6 11.6 10.1 6.2 24.8

10.8 5.6 15.7 6.8 5.7 7.7 0.6 0.7 0.4 3.7 1.9 5.3 21.3 13.1 29.4 10.5 6.7 14.2 26.0 16.1 35.5

8.7 4.0 14.9 6.0 4.8 7.3 0.5 0.5 0.4 3.5 1.9 5.5 17.5 9.8 28.0 8.8 5.0 13.8 21.3 12.0 33.8

CARDIOVASCULAR

Table 2. Rate of Mitral Valve Repair and Percentage of Patients With Active Infective Endocarditis by Year of Surgery

2202

GAMMIE ET AL SURGERY FOR MITRAL VALVE ENDOCARDITIS

Ann Thorac Surg 2005;80:2199 –204

Table 4. Multivariate Predictors of Operative Mortalitya Variable CARDIOVASCULAR

Renal failure: w/ dialysis Renal failure: w/o dialysis Age (per 10 years) Status: salvage Status: emergent Status: urgent Active IE Preop IABP/inotropes Diabetes NYHA class IV CVA ⬍ 2 weeks Peripheral vascular disease Chronic lung disease

ChiSquare

Odds Ratio

95% CI

107.8 — 82.7 45.4 — — 40.1 21.7 16.9 14.0 12.8 10.7 7.0

2.13 1.78 1.41 4.65 2.53 1.36 2.12 1.88 1.59 1.50 1.71 1.55 1.41

1.56–2.91 1.36–2.32 1.31–1.52 2.59–8.35 1.84–3.46 1.07–1.74 1.68–2.68 1.44–2.45 1.28–1.99 1.21–1.86 1.28–2.31 1.19–2.01 1.09–1.81

a

When repair is added to the model, the odds ratio for repair versus replacement is 0.670 (95% confidence interval, 0.51 to 0.88). CI ⫽ confidence interval; CVA ⫽ cerebrovascular accident; IABP ⫽ intraaortic balloon pump; NYHA ⫽ New York Heart Association; preop ⫽ preoperative; w/ ⫽ with; w/o ⫽ without.

Infective endocarditis was the indication for operation in approximately 5% of patients having mitral valve surgery. The proportion of mitral valve surgery patients with IE remained constant for the duration of this study, an observation consistent with reports documenting that the incidence of IE has remained static during the last two decades [4]. Mitral valve repair was less commonly applied to patients with IE compared with non-IE patients. Approximately one third of patients with IE studied received mitral valve repair, in contrast to the non-IE population, in which repair was performed in 45%. Although we observed a slight increase in rates of repair for IE patients during the course of our study, it was far outpaced by increases in repair in the non-IE group. In fact, during the final year of the study the repair rate for non-IE mitral valves (59%) was nearly double that of the IE valves (34%). There were important differences in the rate of repair between active and treated IE patients, with repair significantly more likely in patients with treated

IE. From published experience, surgery is necessary in 25% to 30% of cases of IE during acute infection, and in 20% to 40% in later phases [13, 14]. The degree of leaflet destruction is an important correlate of the likelihood of repair, with greater leaflet tissue destruction associated with a lower likelihood of repair [15]. It is likely that the difference in repair rates reflects the fact that patients with greater leaflet destruction were more likely to require surgery in the acute phase, whereas those with less tissue damage were more likely to undergo surgery in the chronic phase. Several small single-institution series have reported experience with mitral valve repair for IE (Table 5). In 1990 Dreyfus and associates [10] described 40 patients with successful mitral valve repair for active IE. Repair rates in the published series have ranged from 37% to 100%. Groups from Cleveland, France, and Michigan with a strong interest in mitral valve repair have reported repair rates of 70%, 81%, and 100%, respectively [7–9]. The lower rate of application of repair in the present study is reflective of the broader North American experience with mitral valve surgery for IE. During the period of this study, there was a significant (31%) increase in the percentage of active IE patients having mitral valve surgery. Despite this trend toward earlier operation on more acutely ill patients, the rates of mitral valve repair increased and operative mortality was not adversely affected. Although we have demonstrated better outcomes among patients having surgery (both repair and replacement) for chronic mitral valve IE, it is likely that there is a strong selection bias that results in good outcomes among those patients who can complete a course of antibiotics for IE and survive to have later operation. Our data suggest that aggressive early operative intervention in the presence of indications for surgery in the patient with active mitral valve IE can be carried out with reasonable morbidity and mortality, particularly in comparison to outcomes of the universal population diagnosed with IE. We used multivariable analysis to adjust for recognized preoperative risk factors and demonstrated that mitral valve repair remained associated with a lower risk of death. Although it is possible that repair confers a mortality benefit strictly as a result of physiologic and

Table 5. Reported Series of Mitral Valve Repair for Infective Endocarditis Author

Year

N

Repaired (%)

F/U (y)

Active IE (%)

Mortality (%)

Reop Rep Failure

Reinfection

Iung [7] Mihaljevic [18] Wilhelm [19] Muehrcke [9] Pagani [8] Dreyfus [10]

2004 2004 2004 1997 1996 1990

78 53 154 146 22 40

81 40 37 70 100 100a

4.1 4.0 — 3.7 1.8 2.5

40 100 74 40 32 100

3.2% 7.5 (4% repair) 3.9 6.8% (3.2 repair) 0 3.2

4.9% 0 1 ? 0 0

1.6% 10 — 1% 0 0

a

Only reported repair patients.

F/U ⫽ follow-up;

IE ⫽ infective endocarditis;

Reop Rep ⫽ reoperation for repair failure.

procedural advantages, it is likely that a component of the observed benefit arises from the selective application of repair to patients with less tissue destruction and therefore a less aggressive infectious process. The presence of active IE and the presence of a cerebrovascular accident within 2 weeks of surgery were also independently associated with a higher risk of operative mortality. In addition to its recognized advantages, mitral valve repair is particularly attractive in the setting of IE because it reduces the incidence of recurrent valvular infection [9]. Published rates of prosthetic valve endocarditis after mitral valve replacement for IE range from 8% to 27% at long-term follow-up [9, 16]. In contrast, reinfection of the repaired mitral valve is uncommon (Table 5). Gillinov and colleagues [17] reported a series of only 22 patients diagnosed with infected mitral valve repairs at the Cleveland Clinic during a 14-year period, with only 3 having had IE as the original indication for operation. The nature of endocarditis of the repaired valve was an infection of the leaflet in 15 of 22 and a vegetation on the annuloplasty ring in only 4 patients. Taken together, these data strongly support the notion that the risk of late endocarditis after mitral valve repair for IE is very low. The limitations of this study reflect the limitations of a large retrospective clinical database: participation is voluntary with data supplied by participants, and as a result quality and completeness of data are imperfect. The STS National Cardiac Database does not capture data on microbiology, details of disease seen at the time of valve operation, the presence or absence of underlying mitral valve disease, nor details regarding the type of repair performed. Outcome observations are limited to the perioperative period, with no longterm follow-up. This study has documented that IE is present in approximately 5% of patients requiring mitral valve surgery in North America. Repair is less commonly applied to patients with IE compared with non-IE patients. Although mitral valve repair has been increasingly adopted for non-IE mitral valve disease, its use in the setting of IE has increased only slightly. Operation on patients with active IE was associated with a lower likelihood of mitral valve repair compared with patients with treated IE. During the decade studied, there was a significant increase in the percentage of patients with active IE having mitral valve surgery, suggesting a more aggressive approach to this disease. Independent predictors of mortality included replacement (versus repair), active (versus treated) IE, and a recent (⬍2 weeks) preoperative cerebrovascular accident. Rates of mitral valve repair for IE in the STS National Cardiac Database were lower than those in single-institution reports, suggesting an opportunity

GAMMIE ET AL SURGERY FOR MITRAL VALVE ENDOCARDITIS

2203

for increased application of mitral valve repair for the patient with IE. These results provide support for performing mitral valve repair rather than replacement when technically feasible in the setting of mitral valve IE.

References 1. Cabell CH, Abrutyn E. Progress toward a global understanding of infective endocarditis. Lessons from the International Collaboration on Endocarditis. Cardiol Clin 2003;21:147–58. 2. Chu VH, Cabell CH, Benjamin DK Jr, et al. Early predictors of in-hospital death in infective endocarditis. Circulation 2004;109:1745–9. 3. Mylonakis E, Calderwood SB. Infective endocarditis in adults. N Engl J Med 2001;345:1318 –30. 4. Moreillon P, Que YA. Infective endocarditis. Lancet 2004;363: 139 – 49. 5. Vikram HR, Buenconsejo J, Hasbun R, Quagliarello VJ. Impact of valve surgery on 6-month mortality in adults with complicated, left-sided native valve endocarditis: a propensity analysis. JAMA 2003;290:3207–14. 6. Lawrie GM. Mitral valve repair versus replacement. Current recommendations and long-term results. Cardiol Clin 1998;16: 437– 48. 7. Iung B, Rousseau-Paziaud J, Cormier B, et al. Contemporary results of mitral valve repair for infective endocarditis. J Am Coll Cardiol 2004;43:386 –92. 8. Pagani FD, Monaghan HL, Deeb GM, Bolling SF. Mitral valve reconstruction for active and healed endocarditis. Circulation 1996;94(Suppl 2):II-133– 8. 9. Muehrcke DD, Cosgrove DM 3rd, Lytle BW, et al. Is there an advantage to repairing infected mitral valves? Ann Thorac Surg 1997;63:1718 –24. 10. Dreyfus G, Serraf A, Jebara VA, et al. Valve repair in acute endocarditis. Ann Thorac Surg 1990;49:706 –13. 11. Ferguson TB Jr, Dziuban SW Jr, Edwards FH, et al. The STS National Database: current changes and challenges for the new millennium. Committee to Establish a National Database in Cardiothoracic Surgery, The Society of Thoracic Surgeons. Ann Thorac Surg 2000;69:680 –91. 12. Edwards FH, Peterson ED, Coombs LP, et al. Prediction of operative mortality after valve replacement surgery. J Am Coll Cardiol 2001;37:885–92. 13. Jault F, Gandjbakhch I, Rama A, et al. Active native valve endocarditis: determinants of operative death and late mortality. Ann Thorac Surg 1997;63:1737– 41. 14. Larbalestier RI, Kinchla NM, Aranki SF, Couper GS, Collins JJ Jr, Cohn LH. Acute bacterial endocarditis. Optimizing surgical results. Circulation 1992;86(Suppl 2):II-68 –74. 15. Frater RW. Surgery for bacterial endocarditis. In: Baue AG, Geha AS, Hammond GL, Lahs H, Naunheim KS, eds. Glenn’s Thoracic and Cardiovascular Surgery, 6th ed. Stanford, CT: Appleton & Lange; 1996:1915– 41. 16. Yacoub M, Halim M, Radley-Smith R, McKay R, Nijveld A, Towers M. Surgical treatment of mitral regurgitation caused by floppy valves: repair versus replacement. Circulation 1981;64(Suppl 2):II-210 – 6. 17. Gillinov AM, Faber CN, Sabik JF, et al. Endocarditis after mitral valve repair. Ann Thorac Surg 2002;73:1813– 6. 18. Mihaljevic T, Paul S, Leacche M, et al. Tailored surgical therapy for acute native mitral valve endocarditis. J Heart Valve Dis 2004;13:210 – 6. 19. Wilhelm MJ, Tavakoli R, Schneeberger K, et al. Surgical treatment of infective mitral valve endocarditis. J Heart Valve Dis 2004;13:754 –9.

CARDIOVASCULAR

Ann Thorac Surg 2005;80:2199 –204

2204

GAMMIE ET AL SURGERY FOR MITRAL VALVE ENDOCARDITIS

Ann Thorac Surg 2005;80:2199 –204

DISCUSSION CARDIOVASCULAR

DR STEVEN F. BOLLING (Ann Arbor, MI): I think this paper is an excellent example of data mining and clearly shows that we should be striving for repair in this difficult group of patients. Is there any way in this database to get a sense of the number of failed repairs or the amount of residual or recurrent mitral regurgitation with longer term follow-up? DR GAMMIE: That is a great question, and obviously the strength of the database is the numbers. The weakness as it applies to mitral valve surgery is that it does not capture those specific pieces of information, so we can’t answer those questions. We do know from the literature that the overall re-repair rate in this patient population is low, but we cannot comment on it from this study. DR SOON PARK (San Francisco, CA): Doctor Gammie, thank you for the wonderful presentation, again. I think, though, that when you look at repair versus replacement, the type of procedure may reflect more of the extent of endocarditis; for example, patients with a more extensive mitral annular abscess would be less likely to get a repair. So it may not be the procedure that was done but the extent of the disease that determines outcome. Could you comment? DR GAMMIE: I agree with you. I think that the ability to repair the valve depends on the amount of tissue destruction that you face in the operating room. So it is possible that repair confers a physiologic and procedural advantage, but it is also possible that you are selecting patients with less amount of tissue destruction

and probably a less aggressive infectious process, and in fact it is probably a combination of those two things that provides the benefit that we have seen. DR ALAIN F. CARPENTIER (Paris, France): I enjoyed this presentation very much. I have one question. Have you ever used Gore-Tex chordae in your experience with bacterial endocarditis? DR GAMMIE: Doctor Carpentier, it is an honor to have you in the audience. We have found Gore-Tex chordal repair quite helpful, among other techniques. DR CARPENTIER: I like to say that the aim of valve repair is not to correct a mitral valve insufficiency, it is to correct a mitral valve insufficiency for the rest of the life of the patient. The more you use living tissue to prevent recurrence of endocarditis the better for the patient. Nowadays, there is a general tendency of replacing rather than repairing chordae, and I am not sure that it is going to challenge the results that have been obtained with using native chordae. There is a lack of long-term results with PTFE (polytetrafluoroethylene) chordae, and I am personally surprised that after more than 15 years of use of Gore-Tex chords, we haven’t seen long-term results published with this material. I do use Gore-Tex chordae, but only whenever I don’t have native chordae available to be transposed, that is to say very rarely. Thank you very much.