Author + information
- Received April 8, 2003
- Revision received July 4, 2003
- Accepted September 8, 2003
- Published online February 4, 2004.
- Bernard Iung, MD*,* (, )
- Juliette Rousseau-Paziaud, MD†,
- Bertrand Cormier, MD*,
- Eric Garbarz, MD‡,
- Olivier Fondard, MD*,
- Eric Brochet, MD*,
- Christophe Acar, MD§,
- Jean-Paul Couëtil, MD∥,
- Ulrik Hvass, MD∥ and
- Alec Vahanian, MD*
- ↵*Reprint requests and correspondence:
Dr. Bernard Iung, Cardiology Department, Bichat Hospital, 46 rue Henri Huchard, 75018 Paris, France.
Objectives We sought to evaluate the feasibility and immediate and late results of mitral valve repair (MVRep) for acute and healed endocarditis.
Background Improvements in techniques of MVRep have extended its feasibility in complex lesions, but experience with endocarditis is limited.
Methods Among 78 patients operated on for mitral endocarditis between 1990 and 1999, 63 underwent MVRep. The repair was performed for acute endocarditis in 25 patients (40%) at a median of 20 days after the onset of treatment and in 38 patients (60%) for healed endocarditis after a median of 11 months.
Results Repair of the mitral valve was feasible in 63 patients (81%). This repair involved annuloplasty in 61 patients (97%), valve resection in 49 (78%), shortening or transposition of chordae in 29 (46%), suture of perforation in 18 (29%), a pericardial patch in 12 (19%), and a partial mitral homograft in 7 (11%). Associated procedures were aortic valve replacement in 11 patients, bypass grafting in 3, and tricuspid repair in 2. Early complications were two deaths (3.2%), one re-operation for severe mitral regurgitation and one re-operation for subsequent aortic endocarditis. The seven-year rate of event-free survival was 78 ± 6% in the global series. Multivariate predictors of event-free survival were hypertension (p < 0.006) and intervention for acute endocarditis (p < 0.026). Five-year survival rates were 96 ± 4% after MVRep for acute endocarditis and 91 ± 5% for healed endocarditis.
Conclusions Mitral valve repair is frequently feasible and gives good results in patients with infective endocarditis. Patients operated on for acute endocarditis experience more events during follow-up than those operated on after healed endocarditis but have excellent late survival.
Mitral valve repair (MVRep) is now considered as the preferred treatment of degenerative mitral regurgitation (MR), and technical improvements have enabled more complex valve disease to be successfully repaired. In the particular case of infective endocarditis, the possibility to intervene with a low operative risk and to avoid the implantation of a prosthetic device is particularly attractive, but this raises different concerns. The first is the feasibility of MVRep given the wide variety of lesions that can be encountered in mitral endocarditis. The second is related to the efficacy and durability of complex MVRep, particularly when performed on infected tissues during acute endocarditis.
Published series on MVRep for endocarditis have generally included a limited number of patients, and the feasibility and late results have seldom been addressed (1–9). This explains why indications for MVRep in patients with infective endocarditis remain debated.
To assess the immediate and mid-term results of MVRep for infective endocarditis, we report here a contemporary series in which we separately analyzed the outcome of the patients who were operated on during acute endocarditis or after the healing process.
Between January 1990 and December 1999, 78 consecutive patients underwent surgery in our institution because of native mitral endocarditis: 32 patients were operated on during the course of antibiotic treatment (acute endocarditis) and 46 after the completion of antibiotic treatment (healed endocarditis). Patient characteristics are summarized in Table 1.
Of them, 63 patients underwent MVRep (81%) and constitute the study population. Mitral valve repair was performed for acute endocarditis in 25 patients (40%) and for healed endocarditis in 38 (60%). Patients had bacteriologically documented endocarditis (n = 47) or a history of endocarditis with compatible valvular lesions (n = 16). Ten patients (16%) were ≥70 years old. Previous cardiac interventions were one aortic valve replacement and one closure of an atrial septal defect.
Causative micro-organisms were identified in 47 patients (Table 2).
Transthoracic echocardiography and transesophageal echocardiography (TEE) were performed in all patients before surgery, and the echocardiograms were reviewed by two observers (Table 3).
Coronary angiography was performed in 33 patients. Of the 30 patients who had no coronary angiography, 21 were <50 years old and nine needed surgery for acute endocarditis.
In patients with acute endocarditis, surgery was indicated mainly in cases with complications. In patients with healed endocarditis, surgery was performed in cases of severe MR associated with symptoms or left ventricular (LV) dysfunction (Table 4).
Mitral valve repair was performed using cardiopulmonary bypass with cold crystalloid or blood cardioplegia. In the 25 patients operated on for acute endocarditis, surgery was performed after a median of 20 days (range 1 to 60) after the onset of antibiotic treatment; six of these patients had an operation within the first week. In the 38 patients operated on for healed endocarditis, surgery was performed after a median of 11 months (range 2 to 348) after endocarditis.
The distribution of anatomic lesions according to surgical findings is detailed in Table 5.
During surgery, infected tissues were largely resected. Valve reconstruction combined various Carpentier techniques and is detailed in Table 6. Leaflet perforation, either caused by endocarditis or as a result of the resection of vegetation or leaflet abscess, was treated by direct suture in 18 patients who had small lesions. In 12 patients, larger perforations were closed using an autologous pericardial patch, which was tanned intraoperatively in a glutaraldehyde solution (10). A cryopreserved partial mitral homograft was used in seven patients operated on for acute endocarditis, whose leaflet destructions precluded the use of a pericardial patch (11). Ring annuloplasty was performed in 61 patients (97%). Combined procedures were aortic valve replacement for aortic endocarditis (n = 11), tricuspid repair (n = 2; one patient with tricuspid endocarditis and one with functional regurgitation), and coronary artery bypass grafting (n = 3). Intraoperative TEE was used in 28 patients. The mean cross-clamp time was 66 ± 23 min (range 30 to 120). The mean duration of cardiopulmonary bypass was 80 ± 28 min (range 35 to 169).
Valve cultures were positive in 17 of the patients (68%) operated on for acute endocarditis and in none of those operated on for healed endocarditis. All patients operated on for acute endocarditis received postoperative antibiotic treatment for a median duration of 33 days.
Postoperative echocardiography was performed in all patients at a mean of 15 days after surgery, including TEE in 61 patients.
Clinical follow-up was assessed by visits to the department or by a standardized questionnaire sent to the patient's cardiologist. It was concluded in 2000 and was complete in 59 patients (94%). Mean follow-up was 49 ± 38 months (range 1 to 118).
Quantitative variables are expressed as the mean value ± SD. Comparisons between groups were made using the Mann-Whitney Utest for quantitative variables and the chi-square or Fisher exact test for qualitative variables. Analysis of late results was performed on survival and three composite end points: survival without re-operation; good functional results, defined as survival without re-operation and in New York Heart Association (NYHA) functional class I or II; and event-free survival, defined as survival without re-operation and in NYHA class I or II without thromboembolism, bleeding, and endocarditis. Cumulative survival was determined for these four end points, according to the Kaplan-Meier method. Univariate analysis of the predictive factors of event-free survival used a log-rank test for qualitative variables and a univariate Cox model for quantitative variables. It included 28 variables: 13 were related to clinical characteristics or microbiology, 12 to anatomy, and 3 to surgery. Variables with p < 0.20 were entered in a multivariate Cox model and selected backward with a threshold of p < 0.05. The adjusted relative risk (RR) and 95% confidence interval (CI) are given. All analyses were performed with SAS statistical software (SAS Institute Inc., release 6.11).
During the study period, MVRep was feasible in 63 patients (81%) (i.e., 25 [78%] of 32 patients with acute endocarditis and 38 [83%] of 46 patients with healed endocarditis). Fifteen patients (19%) underwent mitral valve replacement because of the extent of tissue destruction: 13 had a prosthetic valve replacement, eight of whom were operated on between 1990 and 1992, and two of whom received a total homograft between 1994 and 1999. There was no significant difference between the characteristics of the 15 patients who had mitral valve replacement and the 63 patients who underwent MVRep (Table 1).
Of the 63 patients who had MVRep, two (3.2%) died during the postoperative period (30 days). Of the 15 patients who had valve replacement, there was no postoperative death.
One death occurred on the first postoperative day afterMVRep in a 78-year-old woman who was operated on while in NYHA class IV because of acute aortic and mitral endocarditis and who developed low-output syndrome. The other death was caused by multi-organ failure 12 days after MVRep for healed mitral endocarditis in a 71-year-old woman.
Two patients (3.2%) who had undergone initial MVRep for acute endocarditis underwent re-operation during the postoperative period. A 37-year-old patient developed severe MR on the second postoperative day because of a nonseptic annular dehiscence, which was successfully treated by repeat repair. The second patient developed aortic endocarditis, and he underwent successful aortic valve replacement using a homograft 21 days after MVRep, though no procedure was needed on the previously repaired mitral valve.
Other postoperative complications were one transient ischemic attack, one hemorrhagic stroke, and three cases of bleeding requiring transfusion (gastrointestinal in two cases and hematuria in one). These five complications did not leave any sequel.
The mean mitral valve area, as assessed by planimetry on postoperative echocardiography, was 2.4 ± 0.6 cm2(range 1.3 to 3.5). One patient had a residual stenosis with a valve area <1.5 cm2. The mean Doppler gradient was 5.0 ± 2.1 mm Hg (range 1.5 to 12). Color Doppler imaging did not detect residual MR in 33 patients (53%). Residual MR was graded 1/4 in 24 patients (39%) and 2/4 in 5 (3 in the acute group and 2 in the healed group) (8%). No patient had residual MR graded >2/4.
Late results after MVRep are shown on Figure 1. Seven-year rates were 93 ± 4% for survival, 84 ± 5% for survival with no re-operation, 80 ± 5% for good functional results, and 78 ± 6% for event-free survival. Among the 59 patients who were alive and without re-operation one month after surgery, the rates were 95 ± 3%, 90 ± 4%, 86 ± 5%, and 84 ± 5%, respectively.
The two late deaths occurred in patients who had been operated on for healed endocarditis. One sudden death occurred 12 months after MVRep in a 71-year-old patient who had a normal preoperative coronary angiogram. The other patient died at the age of 67 years because of hepatic carcinoma 24 months after surgery.
Re-operation was needed in four patients after the postoperative period. One patient had aortic dissection seven months after MVRep for healed endocarditis and underwent replacement of the ascending aorta. The three other patients who had MVRep for acute endocarditis experienced deterioration of MVRep, requiring valve replacement. The first patient had renal failure on dialysis and presented with postoperative grade 2 MR; he had a re-operation after four months because of grade 3 MR related to retraction of the pericardial patch in the commissural area. The second patient who had a postoperative valve area of 1.3 cm2was re-operated on after 10 months, when he became symptomatic. The third patient had a good immediate result of MVRep for acute streptococcal endocarditis and was re-operated on after 15 months because of staphylococcal endocarditis with grade 2 MR and uncontrolled sepsis. There was no other case of recurrent endocarditis during follow-up.
Three other patients developed NYHA class III dyspnea. One was a 78-year-old man who underwent MVRep for acute endocarditis on a rheumatic valve disease; he developed mitral stenosis after one year (valve area 1.1 cm2) and was judged to be too high a risk for re-operation. In the two other patients, there was neither residual stenosis nor MR grade >2, and the cause of dyspnea was multifactorial, related to obesity, hypertension, and atrial fibrillation.
Only one stroke and one gastrointestinal bleeding requiring transfusion occurred. Both events happened in the same patient, who was in atrial fibrillation and had poor compliance to oral anticoagulation. Sixteen patients received oral anticoagulant therapy at last the follow-up visit.
Multivariate analysis identified two predictors of poor event-free survival: hypertension (RR 5.1, 95% CI 1.6 to 16.1; p < 0.006) and surgery for acute endocarditis (RR 4.0, 95% CI 1.2 to 13.3; p < 0.026).
Figure 2represents survival and event-free survival according to whether patients underwent MVRep for acute or healed endocarditis. Five-year survival was 96 ± 4% for acute endocarditis and 91 ± 5% for healed endocarditis (p < 0.66). Five-year event-free survival was 62 ± 11% for acute endocarditis and 88 ± 5% for healed endocarditis (p < 0.02). This was mainly the consequence of a higher frequency of re-operations in patients with acute endocarditis, most of which occurred during the first postoperative year.
This contemporary series shows that MVRep can be performed with good results in a majority of patients who present with a wide range of lesions caused by acute or healed endocarditis.
A specificity of the present series is the inclusion of patients who were consecutively admitted to a medical department for mitral endocarditis and who underwent surgery during the 1990s. The period considered is of importance because epidemiology, diagnosis, and treatment of endocarditis have changed over time. Patient characteristics and the distribution of microorganisms are consistent with other recent series (12–15). Native mitral valve endocarditis, either isolated or associated with aortic endocarditis, accounts for 35% to 50% of native valve endocarditis (13–16). Infective endocarditis still carries a poor prognosis, regardless of the valve affected, as even in recent series, in-hospital mortality ranged from 16% to 21% (12–17).
Another particularity of the 1990s is the wide use of TEE, which enables anatomic lesions to be accurately analyzed, and this has had profound implications in the management of patients (18). In the present series, TEE was systematically performed for preoperative evaluation and, mainly in the recent cases, for intraoperative control.
Finally, the last decade has been associated with a trend toward an increased use of surgery in patients with endocarditis (17), and the growing experience with MVRep has led to refinements in the technique and extension of its indications.
Published series reporting on surgery, mainly valve replacement, for acute endocarditis are retrospective and frequently concern long time periods during which patient characteristics and management are subject to changes. Most series originated from surgical teams, and this might induce a referral bias. Operative mortality for acute infective endocarditis ranged from 7% to 30% (12,14,19–23), late survival was 79% at seven years and ranged from 58% to 73% at 10 years (15,20–23). Comparisons between different series are particularly subject to flaws in the case of infective endocarditis, as it is a very heterogeneous disease.
Experience is limited specifically with regard to MVRep. The first detailed description was published in 1990 by Dreyfus et al. (1), who reported MVRep for acute endocarditis in 40 patients. Eight other series included 7 to 26 patients who had MVRep for acute endocarditis (2–9), whereas MVRep for healed endocarditis has been studied only once (5).
The present series is characterized by the maturity of the technique of MVRep, which is particularly complex in the case of infective endocarditis. The different Carpentier techniques used in endocarditis have been described by Dreyfus et al. (1); their purpose is to remove all pathologic tissues and then to restore normal valve function. Valve prolapse was treated by leaflet resection, and there was also a frequent use of chordal transposition and chordal shortening when valve prolapse occurred in the anterior leaflet and/or commissural area. Thirty percent of patients in our series had valve prolapse involving one commissural area, and this further increases the complexity of MVRep (1). Most leaflet perforations were treated using a direct suture or a pericardial patch, according to the size of the lesions. In the case of extensive destruction of the anterior leaflet, it is possible to combine a large pericardial patch with transposition of chordae (1), whereas the use of a cryopreserved partial mitral homograft is another possibility, as in the case of destruction of the commissural area (11). Pericardial patches and partial homografts were more frequently used in patients operated on for acute endocarditis, as a consequence of more extensive valvular destruction.
Given the complexity of the techniques used, feasibility is an important concern, particularly in acute endocarditis, because of the fragility of infected tissues. However, feasibility has seldom been studied. During the 10-year period of our series, MVRep could be performed in 81% of operated patients by surgeons who have a high level of expertise in MVRep. There was no difference in feasibility between acute and healed endocarditis. Half of the procedures of valve replacement were performed during the first three years of this experience. The reason for precluding repair to be performed was the extent of tissue destruction. The absence of differences in clinical characteristics makes it unlikely that MVRep was restricted to a selected group of low-risk patients.
Operative mortality was low: 4% in the acute group and 2.6% in the healed group. In other series, operative mortality ranged from 0% to 9.1% for acute endocarditis (1–9)and was 2.6% for healed endocarditis (5). Despite the complexity of MVRep, postoperative echocardiography showed good valve function in 55 (90%) of the 61 survivors (i.e., no residual stenosis and no or trivial regurgitation). No patient had severe residual MR.
Seven-year survival was 93 ± 4% in the present series, and the most frequent events were re-operations in six patients, which involved the previously repaired mitral valve in four, all of whom had an operation for acute endocarditis. The re-operation rate was higher in patients who had undergone surgery for acute endocarditis than for healed endocarditis. Most re-operations occurred during the first postoperative year in patients who had suboptimal immediate results. Besides regurgitation, the possibility of residual stenosis should be stressed; it was observed in two patients of our series after MVRep for acute endocarditis. In patients with healed endocarditis, five-year event-free survival was 91%. In the other series, survival after MVRep for acute endocarditis was 88% at four years (4), 87% at five years (7), and 74% at six years (5). Healed endocarditis is also associated with a wide range of lesions, but there is a lower likelihood to have to repair extensive valve destruction, because such patients frequently require surgery at an earlier stage. There was a low rate of re-operations after MVRep for healed endocarditis. Finally, there was a particularly low incidence of late valve-related events in the whole series, particularly regarding thromboembolic events and endocarditis, as is the case after MVRep for other etiologies, and oral anticoagulation was not needed in most patients (24,25).
Multivariate analysis identified MVRep for acute (vs. healed) endocarditis and hypertension as predictive factors of event-free survival. This is mainly the consequence of the higher rate of re-operations in patients who were initially operated on for acute endocarditis. The identification of hypertension as a predictor of poor late functional results may be related to a strong link with age and to the occurrence of multifactorial functional deterioration, which occurred frequently in elderly patients who had good continuing valve function.
Whatever the etiology of MR, the advantages of MVRep over prosthetic valve replacement include better preservation of LV function and a lower incidence of valve-related events. Both mechanical and biological prostheses are associated with high rates of valve-related events, and there is no difference between the two substitutes in the incidence of endocarditis in randomized trials (26). Prosthetic valve replacement for acute endocarditis is associated with a re-operation rate of ∼10% (15,19–23). A more recent alternative has emerged with the use of total mitral homografts (11). Homografts are widely used in patients with aortic endocarditis, but valve replacement with a homograft is a technically demanding technique in the mitral position, and experience is limited, particularly with regard to durability.
Overall, the good immediate and mid-term results observed in the present series are not only the consequence of the surgical technique, but they should also be analyzed in the light of the changes that occurred during the last decade in the global management of patients with infective endocarditis, particularly regarding earlier indications for surgery (27). In our practice, a combined medical and surgical approach was based on clinical characteristics, the type of microorganism, and the analysis of valve lesions after echocardiography, and this led to earlier indications of surgery when there was a high likelihood of performing MVRep (28). The main indication for surgery was heart failure in only 32% of the cases in the acute endocarditis group and 26% in the healed endocarditis group. In recent series reporting on surgery for acute endocarditis, pre-operative hemodynamic conditions were generally more severe: 46% to 91% of the patients were in NYHA class III or IV (4,7,19–21,23), and 46% to 67% had congestive heart failure (3,15,23). In the present series, the low proportion of atrial fibrillation and the moderate dilation of the LV are other characteristics that reflect early indications for surgery. Besides MVRep, per se, these factors contribute to a low operative risk and good late results. With regard to the patients with healed endocarditis who presented as subacute or chronic MR, the experience acquired with other etiologies of chronic MR led us to consider surgery in those with few or no symptoms at the onset of impairment of LV systolic function, provided MVRep was judged to be feasible (29).
Our study included only a limited number of patients, and this limits the accuracy of the estimation of the results. This problem is encountered in the majority of series on endocarditis, which is not a frequent disease; its annual incidence was recently estimated at 31 cases per million in France (17). The number of patients was also reduced by the deliberate restriction of the analysis to the 1990s, in order to consider a relatively homogeneous time period concerning patient management.
It could be argued that the surgical techniques used in the present study are highly specialized and cannot be widely extrapolated. This is true, but the good results obtained with MVRep should lead to its diffusion, as it has been previously the case in degenerative etiologies.
Our purpose was not to compare the results of MVRep with valve replacement. We mentioned patients who had valve replacement during the same period only to assess the feasibility of MVRep. The small sample size of the group of patients treated by valve replacement and the multiple factors intervening in the decision for surgery limit the relevance of such comparisons. It is unlikely that randomized trials will become available in this field, and current series are only observational.
There is a consensus to recommend surgery for native mitral valve endocarditis in the case of severe MR with acute heart failure, infection persisting after 7 to 10 days, or recurrent embolic events (29,30). The restriction of surgery to severe situations is mainly the consequence of the high risk of valve surgery for acute infective endocarditis. However, this might lead to intervention in high-risk patients with refractory heart failure or after cerebral embolism leaving permanent sequelae. The feasibility, moderate operative risk, and good late event-free survival observed with MVRep in the present series are incentives to consider earlier indications for surgery.
During acute endocarditis, MVRep can be considered in patients with severe MR and moderate or transient heart failure. An indication for surgery solely on the basis of vegetation size is controversial but may be discussed in cases of large and mobile vegetations before the occurrence of an embolic event, particularly during the first week, because of the high embolic risk, and in staphylococcal infections. The possibility of performing MVRep with good results may also be an incentive for early surgery in patients with severe MR caused by healed endocarditis. In all cases, an indication for surgery should take into account the presumed operative risk according to age and co-morbidities and the feasibility of MVRep according to echocardiographic findings and surgical advice.
A combined medical and surgical approach by experienced teams in patients with infective endocarditis enables MVRep to be frequently performed with good immediate and late results. Although further evaluation is needed, these findings may lead to wider and earlier indications for MVRep in patients with mitral endocarditis.
- left ventricle/ventricular
- mitral regurgitation
- mitral valve repair
- New York Heart Association
- relative risk
- transesophageal echocardiography
- Received April 8, 2003.
- Revision received July 4, 2003.
- Accepted September 8, 2003.
- American College of Cardiology Foundation
- Acar C.,
- Tolan M.,
- Berrebi A.,
- et al.
- Schulz R.,
- Werner G.S.,
- Fuchs J.B.,
- et al.
- Castillo J.C.,
- Anguita M.P.,
- Ramirez A.,
- et al.
- Wallace S.M.,
- Walton B.I.,
- Kharbanda R.K.,
- Hardy R.,
- Wilson A.P.,
- Swanton R.H.
- Netzer R.O.,
- Altwegg S.C.,
- Zollinger E.,
- Tauber M.,
- Carrel T.,
- Seiler C.
- Shively B.K.,
- Gurule F.T.,
- Roldan C.A.,
- Leggett J.H.,
- Schiller N.B.
- D'Udekem Y.,
- David T.E.,
- Feindel C.M.,
- Armstrong S.,
- Sun Z.
- Guerra J.M.,
- Tornos M.P.,
- Permanyer-Miralda G.,
- Almirante B.,
- Murtra M.,
- Soler-Soler J.
- Hammermeister K.,
- Sethi G.K.,
- Henderson W.G.,
- Grover F.L.,
- Oprian C.,
- Rahimtoola S.H.
- Cormier B.,
- Vahanian A.
- ↵Bonow RO, Carabello B, de Leon AC Jr., et al. ACC/AHA guidelines for the management of patients with valvular heart disease: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol 1998;32:1486–588
- Bayer A.S.,
- Bolger A.F.,
- Taubert K.A.,
- et al.