Impact of Lesion Localization on Durability of Mitral Valve Repair in Infective Endocarditis

Journal Pre-proof Impact of Lesion Localization on Durability of Mitral Valve Repair in Infective Endocarditis Takashi Miura, MD, PhD, Kikuko Obase, MD, PhD, Tsuneo Ariyoshi, MD, PhD, Ichiro Matsumaru, MD, PhD, Shogo Yokose, MD, PhD, Shun Nakaji, MD, PhD, Yuichi Tasaki, MD, Takashi Shimada, MD, Junya Miyamoto, Kiyoyuki Eishi, MD, PhD PII:

S0003-4975(19)31741-2

DOI:

https://doi.org/10.1016/j.athoracsur.2019.10.010

Reference:

ATS 33249

To appear in:

The Annals of Thoracic Surgery

Received Date: 25 January 2019 Revised Date:

1 October 2019

Accepted Date: 2 October 2019

Please cite this article as: Miura T, Obase K, Ariyoshi T, Matsumaru I, Yokose S, Nakaji S, Tasaki Y, Shimada T, Miyamoto J, Eishi K, Impact of Lesion Localization on Durability of Mitral Valve Repair in Infective Endocarditis, The Annals of Thoracic Surgery (2019), doi: https://doi.org/10.1016/ j.athoracsur.2019.10.010. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 by The Society of Thoracic Surgeons

Impact of Lesion Localization on Durability of Mitral Valve Repair in Infective Endocarditis Short title: Repair for Mitral Endocarditis

Takashi Miura, MD, PhDa; Kikuko Obase, MD, PhDa; Tsuneo Ariyoshi, MD, PhDa; Ichiro Matsumaru, MD, PhDa; Shogo Yokose, MD, PhDa; Shun Nakaji, MD, PhDa; Yuichi Tasaki, MDa; Takashi Shimada, MDa; Junya Miyamotob; Kiyoyuki Eishi, MD, PhDa

Department of Cardiovascular Surgery, Nagasaki University Hospital, Nagasaki, Japana; Nagasaki University Hospital Clinical Research Center, Nagasaki, Japanb

Word Count: 248 words in the abstract and 4,666 words in the text Key words: Infective endocarditis; Mitral valve repair; Feasibility; Durability Classifications: Infection; Mitral valve repair; Reoperation

Corresponding author: Takashi Miura, MD, PhD Department of Cardiovascular Surgery, Nagasaki University Hospital 1-7-1 Sakamoto, Nagasaki City, Nagasaki 852-8501, Japan E-mail address: [email protected]

1

Abstract Background. Mitral valve (MV) repair for infective endocarditis (IE) is sometimes challenging. We investigated the durability of repair associated with the location of the infected lesion. Methods. Eighty-three patients (55 ± 18 years, active: 66, healed: 17) who underwent MV repair at our institution were studied. Patients were categorized into five types based on the location of the main lesion: Type I, posterior leaflet (n = 36); Type IIC, clear zone of anterior leaflet (n = 12); Type IIR, rough zone of anterior leaflet (n = 28); and Type III, annulus (n = 7). Type IIR was divided into two subgroups: IIR-large (more than one segment involvement, n = 9) and IIR-small (one segment, n = 19). Results. Follow-up was 95.2% completed for 7.2 ± 5.2 years. The rate of recurrent moderate or severe mitral regurgitation (MR) was 2.8% in Type I, 0% in Type IIC, 28.6% in Type IIR, and 14.3% in Type III. The freedom from recurrent moderate or severe MR was significantly lower in Type IIR compared with the other types (63.1 ± 10.6% vs. 96.8 ± 3.2% at 5 years, p = 0.0003). Moreover, the recurrence rate was higher in Type IIR-large (55.6%) than in Type IIR-small (15.8%, p = 0.0684). Type IIR was an independent predictor of recurrent MR (hazard ratio: 11.1, 95% confidence interval: 2.52-78.2, p = 0.0010). Conclusions. The durability of MV repair for IE was satisfactory in posterior leaflet infection without annulus invasion and in clear zone infection of the anterior leaflet. However, rough zone infection of the anterior leaflet, especially with more than one segment involvement, was associated with a high risk of recurrent MR.

2

Abbreviations and Acronyms MV = mitral valve IE = infective endocarditis MR = mitral regurgitation PML = posterior mitral leaflet AML = anterior mitral leaflet

3

Efforts are made to choose repair as the preferred surgical treatment for mitral valve (MV) infective endocarditis (IE) because the long-term outcomes of repair are superior to those of replacement.1,2 However, some patients who underwent MV repair may subsequently develop recurrent mitral regurgitation (MR) and require reoperation. It is therefore important to know preoperatively which lesions are favorable for repair and which are not in order to avoid repair failure. The aim of this study is to develop an algorithm that can direct the surgical decision about the feasibility and durability of valve repair based on the main location of the MV vegetation. Accordingly, we have classified the main infected lesions into five types, and investigated the feasibility and durability of MV repair according to these types.

Patients and Methods Between January 1999 and June 2018, 116 patients underwent MV surgery for IE at our institution. Of these, 83 patients (71.6%) who underwent MV repair (66 patients with active IE and 17 with healed IE) were studied. A diagnosis of IE was made using the Duke or the modified Duke criteria.3, 4 Active IE was defined as the completion of antibiotic therapy. Table 1 presents a comparison of patient characteristics between the MV repair group and the MV replacement group. None of the patients were intravenous drug users in either of the groups. The MV replacement group included five patients who were converted from MV repair to replacement during surgery. All patients with healed IE underwent MV repair. In contrast, all IE patients with rheumatic MV stenosis

4

underwent MV replacement. The percentages of patients with intubation, acute renal failure, steroid use, concomitant aortic valve IE, and uncontrolled infection were significantly higher in the MV replacement group. This retrospective study was approved by the institutional review board at Nagasaki University Hospital (#16082208).

Classification of infected lesions Patients were categorized into five types: (1) Type I, posterior mitral leaflet (PML) (n = 36); (2) Type IIC, clear zone of anterior mitral leaflet (AML) (n = 12); (3) Type IIR-small, rough zone of AML (one segment) (n = 19); (4) Type IIR-large, rough zone of AML (more than one segment involvement) (n = 9), and (5) Type III, annulus invasion (n = 7). Table 2 shows the summary of the classification, and Figure 1 presents a schematic drawing of the classification and photographs of typical cases. Supplementary Video 1 shows surgery for those cases. The following rules were applied for the classification (Figure 2). If the annulus was involved, the classification of the lesion was Type III regardless of its primary location. A lesion without annular involvement was classified into either Type II (AML) or Type I (PML). If the AML lesion that was in need of repair was present, the classification of the lesion was Type II regardless of the PML lesion. Because in cases with PML lesion where the AML is also diseased and in need of repair, the result of repair rests on whether the AML lesion can be successfully repaired or not. The rest were Type I. Type II was further sub-categorized into the rough zone as Type IIR or the

5

clear zone as Type IIC based on the main infected lesion. Type IIR was divided into two subgroups: IIR-large (more than one segment involvement) and IIR-small (one segment). Type IIC classification was established for the following three reasons. First, an isolated AML clear zone lesion exists in clinical settings, such as aortic regurgitant jet-related or accelerated flow in hypertrophic obstructive cardiomyopathy-related endocarditis. Second is the presence of AML clear zone main lesion without rough zone lesion accompanied by PML lesion in both leaflet infection (see photograph of Type IIC in Figure 1). These lesions were classified as Type IIC. Third, MV repair for Type IIC lesions sometimes requires special techniques, such as auto-pericardial patch reconstruction. Supplementary Figure 1 presents a summary of the classification based on infected leaflets.

Our treatment strategy for active IE Our treatment strategy for active IE conforms to the published guidelines of the American Heart Association5 and the Japanese Circulation Society (2017)6 for management of IE. Antibiotics were selected on the basis of the results of blood cultures or MV culture, and were used for 4 or 6 weeks based on the bacterial species. Empiric antibiotic therapy was administered in patients with negative blood cultures. We performed emergent or urgent surgery in patients at high risk of embolism (vegetation size ≥10 mm), heart failure, and/or uncontrolled infection. Surgery in patients with brain complications also followed the above guidelines’ recommendation.

6

Surgery Following our years of experience, we have begun to attempt MV repairs for Type IIR lesions in recent years. The frequency of Type IIR was similar between the first-half group (January 1999 to December 2010) and the second-half group (January 2011 to June 2018) from among the 116 total cases (23/58 = 39.7% vs. 25/58 = 43.1%, p = 0.8506). The repair rate for Type IIR was significantly higher in the second-half group (19/25 = 76.0%) than in the first-half group (9/23 = 39.1%, p = 0.0182). Surgical approaches and techniques are shown in Table 3. Surgery was performed via a standard full sternotomy or right mini-thoracotomy approach.7 The right mini-thoracotomy approach was mainly used in young patients. After aortic cross-clamping, the MV was exposed through a right-sided left atriotomy under cardiac arrest. Prior to leaflet repair, the mitral annulus was sutured with 2-0 Ticron (Ethicon Inc., Somerville, NJ) in order to precisely evaluate the MV. After inspection of the MV, vegetation was removed by forceps and Metzenbaum scissors. Debridement of infected tissue was performed by electric cautery and a small osteotrite. Valve repair was mainly performed with the following five techniques: (1) resection and suture, (2) neochordae reconstruction, (3) auto-pericardial patch repair, (4) edge-to-edge fixation, and/or (5) chordal transfer. Leaflet perforation was repaired by direct suture, or was reconstructed with an auto-pericardium. A fresh auto-pericardial patch was applied in small perforation repair, such as perforation of the clear zone of the AML or the PML. As shown in Supplementary Video 2, a glutaraldehyde-treated auto-pericardial patch was used for reconstruction of large defects, such as missing A2 medial and A3 segments. Annulus

7

reconstruction was performed via mattress annular plication, auto-pericardium patch, or bovine pericardium patch. Mitral annuloplasty was performed using a fresh auto-pericardial band, or a prosthetic partial band or complete ring. Partial band annuloplasty was employed for isolated PML infection. A complete ring was used for AML or both leaflet infections in order to secure the coaptation of the MV. Some young patients or females hoping for conception and birth underwent a commissure band annuloplasty, or no annuloplasty out of concern for functional mitral stenosis after repair. These complex MV repairs were performed by a single experienced surgeon (K.E.) who has performed over 50 MV repair cases per year for the last 20 years. Supplementary Figure 2 and Video 3 show typical cases of MV replacement.

Follow-up Patients were followed up by our outpatient clinic or the referring cardiologists. Medical interview was conducted by telephone for patients lacking outpatient clinic follow-up. Postoperative event evaluations were strictly defined according to the published guidelines of the American Association for Thoracic Surgery and the Society of Thoracic Surgeons.8 Follow-up for this study was closed on August 31, 2018. Four patients were lost to follow-up, and the follow-up rate was 95.2% completed for 7.2 ± 5.2 years, for a total of 599.9 patient-years. Maximum follow-up was 19.3 years, and 28 patients (33.7%) were tracked for over 10 years. The grade of post-operative mitral regurgitation (MR) was semi-quantitatively classified according to the colored regurgitation jet areas as none or trace (0–2.0 cm2), mild (2–4.0 cm2), moderate (4.0–8.0

8

cm2), or severe (> 8 cm2). The echocardiographic data collected at an outside hospital were read as follows: “trace to mild MR” was entered into the database as mild, “mild to moderate MR” as moderate, and “moderate to severe MR” as severe.9 All patients had anticoagulation therapy with warfarin sodium for 3 months after surgery. Warfarin treatment was continued in patients with chronic atrial fibrillation.

Statistical analysis Continuous data were reported as the mean ± standard deviation or median (interquartile range), and categorical variables as percentages. Categorical variables were analyzed using Fisher’s exact test, and continuous variables were analyzed using the Student’s t-test or the Mann-Whitney U-test. The actuarial survival, cardiac-related mortality, and freedom from reoperation and recurrent moderate or severe MR were estimated using Kaplan-Meier survival analyses. The estimates were reported as the mean ± standard error. A significant result was defined by log-rank test. Late valve-related events were expressed in percent per patient-years. Cox proportional hazard regression analyses were performed to identify the predictors of recurrent moderate or severe MR. We included types of infection (Type IIR vs. the other three types), active IE, and Staphylococcus aureus infection in the model. The proportional hazards assumption was confirmed by testing the correlation of Schoenfeld residuals with time. P-values less than 0.05 were considered significant. Analyses were conducted with JMP version 12.0 (SAS Institute Inc., Berkeley, CA, USA) and R version 3.3.1.10

9

Results As shown in Figure 3A, the feasibility of MV repair is highest in Type I (94.7%), followed by Type IIR-small (86.4%), Type IIC (75.0%), Type III (50%), and Type IIR-large (34.6%). The success of MV repair for Type IIR-small lesion was significantly higher than that of Type IIR-large lesion (p=0.0004). Table 4 shows the outcomes of MV repair. In-hospital mortality was noted in three patients (3.6%). Causes of death were sepsis in one, renal failure in one, and cerebral hemorrhage in one. Late death occurred in ten patients (cardiac-related in two and non-cardiac-related in eight). The actuarial survival rate and cardiac death-free rate 10 years after surgery were 82.8 ± 5.4% and 93.6 ± 3.4%, respectively (Supplementary Figure 3). No patients developed recurrent endocarditis. Recurrent moderate or severe MR occurred in 10 patients (12.0%). Table 5 shows the clinical findings and the mechanisms of recurrent MR. Nine patients had active IE and the other one had healed IE. Nine of the 10 patients underwent reoperation (all replacements). No patients died after reoperation. Freedom from reoperation and recurrent moderate or severe MR rates at 10 years were 86.4 ± 4.3% and 79.2 ± 7.0%, respectively (Figure 4A). The mechanisms of recurrent MR were classified into procedure-related failure (seven patients) or valve-related failure (three patients). Moreover, procedure-related failure was classified into three sub-groups: leaflet suture dehiscence, ring detachment, and patch failure. The timing of MR recurrence was early in the procedure-related failure group, and five of the seven patients had recurrent MR within 6 months after repair. As shown in Supplementary Figure 4, the patch failure

10

mechanism was dehiscence or shrinkage of the patch. The cause of the valve-related failure was progressive degeneration of the MV in two and mitral leaflet tethering due to dilated cardiomyopathy in one. Of the 10 patients with recurrent MR, 8 patients were Type IIR (large: 5 and small: 3), 1 patient was Type I, and 1 patient was Type III. Figure 3B shows the incidence of recurrent moderate or severe MR according to the classification. The rates were 2.8% in Type I, 0% in Type IIC, 28.6% in Type IIR, and 14.3% in Type III. Moreover, the recurrence rate was higher in Type IIR-large (55.6%) than in Type IIR-small (15.8%, p = 0.0684). As shown in Figure 4B, the freedom from recurrent moderate or severe MR was significantly lower in Type IIR compared with the other three types (p = 0.0003). Type IIR was the only independent predictor of recurrent moderate or severe MR (hazard ratio: 11.1, 95% confidence interval: 2.52-78.2, p = 0.0010).

Comment The present study showed that the success of MV repair for IE is closely related to the main lesion localization. The feasibility and durability were excellent in the PML main lesion without annulus invasion, suggesting that this lesion is a good candidate for MV repair. In contrast, the rough zone infection of the AML, especially with more than one-segment involvement, was associated with low durability of MV repair although the rate of successful MV repair for AML rough zone lesion increased after we acquired years of experience. The reason for repair failure in Type IIR may be explained by

11

comparing it to the repair of degenerative AML prolapse. Generally, the culprit lesion in degenerative prolapse is in the rough zone, which is the important segment in which to secure leaflet coaptation. A rough zone lesion can be repaired by simple leaflet resection in PML cases, but in AML cases, it carries a risk of leaflet distortion, which can lead to residual or recurrent MR, especially in cases with extensive prolapse. The same mechanism may be considered in the treatment of Type IIR lesions in IE, especially when there is more than one-segment involvement. In the present study, we divided Type IIR lesions into two subgroups (small and large) to investigate the influence of the extent of infected lesions on the outcomes of MV repair. The results suggested that in the IIR lesion, it is critical for surgeons to carefully consider the number of involved segments since the outcomes differ between IIR-small and IIR-large categories. To improve the feasibility of MV repair, it is necessary to reconstruct the leaflet defects with auto-pericardium or other material. 11-15 De Kerchove et al. reported that the success of MV repair for active MV-IE could be increased to 80% by applying a glutaraldehyde-treated auto-pericardial patch.13 The patch repair had the reoperation risk of approximately 20% at 8 years and 25% at 15 years.13,14 In this study, we reconstructed the MV leaflet using an auto-pericardial patch for 14 patients, and 12 of them were able to recover MV morphology and function without recurrent MR. The remaining two patients (14.3%) belonged to the Type IIR group, and developed recurrent MR. The auto-pericardial patches in both cases were applied to the reconstruction of the rough zone of the AML. This finding was in contrast to the application sites in the other 12 patients without recurrent MR (PML in 4 patients and

12

AML clear zone in 8 patients). The previous clinical study reported that the durability of the small patches applied to the leaflet perforation repair was superior to that of patches used for the free margin repair.13 Another experimental study reported that the mechanical stress on mitral leaflets was greater in the AML than in the PML, and the stress of the non-coapted leaflet area depended on the size of the leaflet surface area (i.e., the degree of billowing).16 Thus, the durability of the auto-pericardial patch may be explained by the location, stress distribution, or size of the patch itself. In the present study, MV repair was performed by only a single surgeon who is an expert in MV repair methods with a sufficient caseload to maintain his skills. This technical ability, and experience dealing with a multitude of MV lesions are the critical elements in successful repair surgery, which might lead to limitations in the generalization of our results. The main limitation of the present study is that it involves a retrospective design at a single institution, which may therefore contain an inherent bias.

Conclusions The durability of MV repair for IE was satisfactory in PML lesions without annulus invasion, and clear zone infection of the AML. In contrast, the rough zone infection of the AML, especially with more than one segment involvement, was associated with low feasibility and durability of MV repair.

13

References 1. Feringa HH, Shaw LJ, Poldermans D, Hoeks S, van der Wall EE, Dion RA, et al. Mitral valve repair and replacement in endocarditis: a systematic review of literature. Ann Thorac Surg. 2007;83:564-70. 2. Toyoda N, Itagaki S, Egorova NN, Tannous H, Anyanwu AC, El-Eshmawi A, et al. Real-world outcomes of surgery for native mitral valve endocarditis. J Thorac Cardiovasc Surg. 2017;154:1906-1912. 3. Durack DT, Lukes AS, Bright DK. New criteria for diagnosis of infective endocarditis: utilization of specific echocardiographic findings. Duke Endocarditis Service. Am J Med. 1994;96:200-9. 4. Lamas CC, Eykyn SJ. Suggested modifications to the Duke criteria for the clinical diagnosis of native valve and prosthetic valve endocarditis: analysis of 118 pathologically proven cases. Clin Infect Dis. 1997;25:713-9. 5. Baddour LM, Wilson WR, Bayer AS, Fowler VG Jr, Tleyjeh IM, Rybak MJ, et al. Infective Endocarditis in Adults: Diagnosis, Antimicrobial Therapy, and Management of Complications: A Scientific Statement for Healthcare Professionals From the American Heart Association. Circulation. 2015;132(15):1435-86. 6. Nakatani S, Ohara T, Ashihara K, Izumi C, Iwanaga S, Eishi K, et al. JCS 2017 Guideline on Prevention and Treatment of Infective Endocarditis. Circ J. 2019 Jul 5.

14

doi: 10.1253/circj.CJ-19-0549. 7. Miura T, Tanigawa K, Matsukuma S, Matsumaru I, Hisatomi K, Hazama S, et al. A right thoracotomy approach for mitral and tricuspid valve surgery in patients with previous standard sternotomy: comparison with a re-sternotomy approach. Gen Thorac Cardiovasc Surg. 2016;64:315-24. 8. Akins CW, Miller DC, Turina MI, Kouchoukos NT, Blackstone EH, Grunkemeier GL, et al. Guidelines for reporting mortality and morbidity after cardiac valve interventions. J Thorac Cardiovasc Surg. 2008;135:732-8. 9. David TE, Ivanov J, Armstrong S, Christie D, Rakowski H. A comparison of outcomes of mitral valve repair for degenerative disease with posterior, anterior, and bileaflet prolapse. J Thorac Cardiovasc Surg. 2005;130:1242-9. 10. R Core Team (2016). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/. 11. Zegdi R, Debièche M, Latrémouille C, Lebied D, Chardigny C, Grinda JM, et al. Long-term results of mitral valve repair in active endocarditis. Circulation. 2005;111:2532-6. 12. Shang E, Forrest GN, Chizmar T, Chim J, Brown JM, Zhan M, et al. Mitral valve infective endocarditis: benefit of early operation and aggressive use of repair. Thorac Surg. 2009;87:1728-33. 13. de Kerchove L, Price J, Tamer S, Glineur D, Momeni M, Noirhomme P, et al.

15

Ann

Extending the scope of mitral valve repair in active endocarditis. J Thorac Cardiovasc Surg. 2012;143(4 Suppl):S91-5. 14. Solari S, De Kerchove L, Tamer S, Aphram G, Baert J, Borsellino S, et al. Active infective mitral valve endocarditis: is a repair-oriented surgery safe and durable? Eur J Cardiothorac Surg. 2019;55:256-262. 15. Ng CK, Nesser J, Punzengruber C, Pachinger O, Auer J, Franke H, et al. Valvuloplasty with glutaraldehyde-treated autologous pericardium in patients with complex mitral valve pathology. Ann Thorac Surg. 2001;71:78-85. 16. Xu C, Jassar AS, Nathan DP, Eperjesi TJ, Brinster CJ, Levack MM, et al. Augmented mitral valve leaflet area decreases leaflet stress: a finite element simulation. Ann Thorac Surg. 2012;93:1141-5.

16

Figure legends Figure 1. Typical cases of classification Type I: PML lesion without annulus invasion (P3 and posteromedial commissure) Type IIC: AML clear zone lesion Type IIR-small: AML rough zone lesion (A2 medial) Type IIR-large: AML rough zone lesion (A2 medial and A3) Type III: Annulus invasion (P1 and P2)

Figure 2. Rules of classification If the annulus was involved, the classification of the lesion was Type III regardless of its primary location. A lesion without annular involvement was classified into either Type II (AML) or Type I (PML). If the AML lesion that was in need of repair was present, the classification of the lesion was Type II regardless of the PML lesion. The rest were Type I.

Figure 3. Feasibility and durability of mitral valve repair (A) Feasibility The rate of mitral valve repair was highest in Type I (94.7%), followed by Type IIR-small (86.4%), Type IIC (75.0%), Type III (50%), and Type IIR-large (34.6%). (B) Incidence of recurrent moderate or severe MR The incidence rate was highest in Type IIR-large (55.6%), followed by Type IIR-small (15.8%), Type III (14.3%), Type I (2.8%), and Type IIC (0%).

17

Figure 4. Reoperation-free rate, and recurrent moderate or severe MR-free rate (A) Freedom from reoperation and recurrent moderate or severe MR (B) Freedom from recurrent moderate or severe MR: Type IIR vs. the other three types

Supplementary Figure 1. Summary of the classification based on infected leaflets PML = posterior mitral leaflet, BML = both mitral leaflets, AML = anterior mitral leaflet

Supplementary Figure 2. Typical cases of mitral valve replacement (A) Type IIR-large: A1, A2, and A3 rough zone infection (B) Type III: Infection within mitral annular calcification

Supplementary Figure 3. Comparisons of survival rate and cardiac-related death-free rate between the mitral valve repair group and the replacement group

Supplementary Figure 4. Auto-pericardial patch failures (A) The patch was applied for reconstruction of the A2 and A3 rough and clear zones. The suturing site was ruptured 2 months after surgery. (B) The patch with glutaraldehyde treatment was used for extensive reconstruction of the rough and clear zones in the A2 medial and A3 segments. The patient underwent mitral valve replacement due to the shrinkage of the patch. Supplementary Video 2 shows this patient’s initial operation.

18

Supplementary Video legends Video 1: Mitral valve repair Video 2: Extensive anterior leaflet reconstruction with the auto-pericardial patch Video 3: Mitral valve replacement

19

Table 1 Patient characteristics

Age, years old Female Infective endocarditis Active Healed Bacterial species Streptococcus Staphylococcus Enterococcus Others Unknown Hospital-acquired infection Intubated Comorbidities Diabetes mellitus Vertebral osteomyelitis Atrial fibrillation Atopic dermatitis Steroid user Hemodialysis Acute renal failure Primary heart disease Hypertrophied obstructive cardiomyopathy Idiopathic dilated cardiomyopathy Mitral annulus calcification Rheumatic mitral stenosis Infected leaflets Posterior Bi-leaflet Anterior MR grade Aortic valve endocarditis Brain complications Indication for surgery in active endocarditis Mitral mobile vegetation >10 mm Heart failure (NYHA class III or IV) Uncontrolled infection

20

Repair n = 83

Replacement n = 33

p value

60 (44–70) 23 (27.7)

63 (51–75.5) 17 (51.5)

.0753 .0183

66 (79.5) 17 (20.5)

33 (100) 0 (0)

.0028

39 (47.0) 24 (28.9) 7 (8.4) 3 (3.6) 10 (12.1) 9 (10.8) 4 (4.8)

11 (33.3) 12 (36.4) 0 (0) 8 (24.2) 2 (6.1) 5 (15.2) 8 (24.2)

10 (12.1) 6 (7.2) 5 (6.0) 4 (4.8) 4 (4.8) 3 (3.6) 1 (1.2)

6 (18.2) 2 (6.1) 4 (12.1) 0 (0) 7 (21.2) 3 (9.1) 6 (18.2)

.3854 1.0 .2716 .5765 .0118 .3501 .0021

1 (1.2) 1 (1.2) 1 (1.2) 0 (0)

2 (6.1) 0 (0) 2 (6.1) 3 (9.1)

.1944 1.0 .1944 .0215

23 (27.7) 37 (44.6) 23 (27.7) 3 (3–4) 5 (6.0) 42 (50.6)

4 (12.1) 13 (39.4) 16 (48.5) 3 (2–4) 15 (45.5) 17 (51.5)

.3204 < .0001 1.0

51 (77.3) 25 (37.9) 20 (30.3)

28 (84.8) 20 (60.6) 17 (51.5)

.4369 .0532 .0489

< .0048

.5366 .0043

.0595

Values represent n (%), or mean ± SD. NYHA = New York Heart Association

21

Table 2 Classification of the lesion localization Repair n = 83

Type I

PML without annulus invasion

36

IIC

Clear zone of AML

12

IIR

Rough zone of AML

28

small

One segment

19

large

Two or three segments

9

III

Annulus invasion

7*

PML = posterior mitral leaflet, AML = anterior mitral leaflet * Type III included two AML annulus cases and five PML annulus cases.

22

Table 3 Surgery Repair n = 83 Approach Median sternotomy

69 (83.1)

Right mini-thoracotomy

14 (16.9)

Operation time, min

240 (195–289)

Cardiopulmonary bypass time, min

124 (96–152)

Aortic cross-clamp time, min

74 (58–100)

Concomitant surgery*

12 (14.5)

Leaflet repair techniques Resection and sutures Anterior leaflet

24 (28.9)

Posterior leaflet

42 (50.6)

Neochordae Anterior leaflet

15 (18.1)

Posterior leaflet

4 (4.8)

Autologous pericardial patch repair Anterior leaflet

10 (12.1)

Posterior leaflet

4 (4.8)

Edge-to-edge fixation

19 (22.9)

Chordal translocation

4 (4.8)

Aorto-mitral curtain repair by bovine pericardium

1 (1.2)

Annuloplasty Partial rings

35 (42.2)

Twisted auto-pericardium

23 (27.7)

Prosthetic

11 (13.3)

Bovine pericardium

1 (1.2)

Complete rings

40 (48.2)

Commissure bands

2 (2.4)

Kay

1 (1.2)

23

None

5 (6.0)

Values represent n (%), or mean ± SD.

*Concomitant surgery included aortic valve replacement in six patients, aortic valve repair in one, coronary artery bypass grafting in two, pulmonary isolation in three, aortic arch aneurysmal patch repair in one, implantation of permanent pacemaker in one, and extended septal myectomy in one.

24

Table 4 Outcomes

Repair n = 83 Mortality In-hospital

3 (3.6)

Late

10 (12.0)

MR grade after repair None or trace

55 (66.2)

Mild

18 (21.7)

Moderate

2 (2.4)

Severe

8 (9.6)

Morbidity Recurrence of IE, % patient-year

0

Thromboembolic events, % patient-year Major bleeding events, % patient-year Values represent n (%), or mean ± SD.

0.5 0.5

MR = mitral regurgitation, IE = infective endocarditis

25

Table 5 Mechanisms of recurrent MR and clinical findings in 10 patients

Mechanisms of recurrent MR

Numbers

Active or Healed

Timing of recurrence

Reoperation

Types

Pa re

2

Active

12 days

Yes

IIR-small

Y

Active

5 months

Yes

IIR-small

Y

Active

2 months

Yes

IIR-large

A

Active

3 months

Yes

IIR-large

A

Active

2 months

Yes

IIR-large

D

Active

9 months

Yes

IIR-large

Sh

1

Active

3.8 years

Yes

III

A B

Procedure-related 7 pts 1.

2.

3.

4.

Leaflet suture dehiscence

Ring detachment

Auto-pericardial patch failure

Aorto-mitral continuity patch failure

2

2

Valve-related 3 pts 1.

New chordae rupture

1

Active

3 months

Yes

IIR-small

O

2.

Sclerosis of anterior mitral leaflet

1

Healed

4.4 years

Yes

IIR-large

T

3.

Left ventricular dilatation

1

Active

9.8 years

No

I

D

MR = mitral regurgitation

26