The incorporated aortomitral homograft for double-valve endocarditis: the ‘hemi-Commando’ procedure. Early and mid-term outcomes†

Abstract

OBJECTIVES

Surgical management of invasive double-valve infective endocarditis (IE) involving the intervalvular fibrosa (IVF) is a technical challenge that requires extensive debridement followed by complex reconstruction. In this study, we present the early and mid-term outcomes of the hemi-Commando procedure and aortic root replacement with reconstruction of IVF using an aortomitral allograft.

METHODS

From 2010 to 2017, 37 patients with IE involving the IVF underwent the hemi-Commando procedure. Postoperative clinical data and echocardiograms were reviewed for the assessment of cardiac structural integrity and clinical outcomes.

RESULTS

Twenty-nine (78%) cases were redo surgery and 15 (41%) were emergency surgery. Preoperatively, 70% (n = 26) of patients were admitted to the intensive care unit and 11% (n = 4) of patients were in septic shock. Ten (27%) patients had native aortic valve IE, while 27 (73%) patients had prosthetic valve IE. Hospital death occurred in 8% (n = 3) of patients due to multisystem organ failure. Postoperative echocardiogram showed no aortic regurgitation in 86% (n = 32) and mild regurgitation in 14% (n = 5) of patients, while mitral regurgitation prevalence was none/trivial in 62% (n = 23), mild in 32% (n = 12) and moderate in 5%. Intact IVF reconstruction was confirmed in all patients with no abnormal communication between the left heart chambers. One-year survival was 91%, while 3-year survival was 82%. Mid-term follow up revealed 1 death secondary to recurrent IE.

CONCLUSIONS

Compared to double-valve replacement with IVF reconstruction (‘Commando operation’), the early and mid-term outcomes of the hemi-Commando procedure proved to be a feasible treatment option for IVF reconstruction, enabling preservation of the mitral valve and the subvalvular apparatus in high-risk patients with invasive double-valve IE.

Endocarditis, Intervalvular fibrosa, Mitral valve, Hemi-Commando

INTRODUCTION

Invasive double-valve infective endocarditis (IE) involving the fibrous skeleton of the heart requires complex, high-risk surgical management [1, 2]. The pathology of biofilm-associated IE necessitates extensive debridement of all necrotic and infected tissues and the removal of prosthetic material to provide the best chance for cure [3, 4]. Subsequently, a complex reconstruction is often required, particularly in cases of aortic root abscess extending to the intervalvular fibrosa (IVF) and the mitral valve (MV). Reconstruction of the IVF can be accomplished during double-valve replacement procedures using pericardial or Dacron patches or by implanting a monobloc aortomitral prosthesis or an allograft [2, 5, 6]. This is a challenging operation, which we refer to as the ‘Commando operation’ at our institution, a name that denotes its technical complexity [2, 7]. Recently, our group described a modified approach to double-valve IE that extends only to the anterior mitral leaflet; this entails implantation of a trimmed aortomitral allograft into the aortic root and the edge of the anterior mitral leaflet as the ‘hemi-Commando operation’ [8]. This technique represents a feasible and relatively less complex alternative with the advantage of preservation of the major portion of the MV and its subvalvular apparatus. In this report, we present the early and mid-term outcomes of our experience with the hemi-Commando operation.

METHODS

Patients

From January 2010 to July 2017, 37 adults with invasive IE (aortic root abscess involving the IVF and the anterior mitral leaflet) underwent the hemi-Commando procedure as described previously [8]. Inclusion criteria for this retrospective analysis include aortic valve replacement for IE using aortomitral homograft with or without MV repair. Exclusion criteria included MV replacement (to exclude Commando patients) and aortic valve replacement using aortomitral homograft for non-infective pathology. Mean age at surgery was 55 ±17 years, and 84% were men. Fifteen patients were followed up until 30 June or until the time of death. The median follow-up period was 305 days.

Clinical data

Patient characteristics, procedural information and perioperative echocardiographic data were obtained from the Cardiovascular Information Registry, echocardiographic database and medical record review. These data were approved for use in research by the Cleveland Clinic Institutional Review Board, with patient consent waived. The indication for surgery was aortic valve and MV IE. All patients underwent intraoperative transoesophageal echocardiogram and routine postoperative 2D transthoracic echocardiography with Doppler colour flow mapping. Similar imaging studies were performed during surveillance in the subset of patients followed at the Cleveland Clinic.

Surgery

All patients underwent the following hemi-Commando procedures: radical debridement of an aortic root abscess, the IVF and the anterior mitral leaflet followed by implantation of a trimmed aortomitral homograft. The clear proximal zone (atrial zone) of the anterior mitral leaflet is resected in continuity with the infected aortomitral curtain tissue, whereas the free rough edge of the anterior mitral leaflet (coaptation zone) that receives the primary and secondary chordates is preserved. Preoperative or intraoperative echocardiographic measurements of the aortic and the mitral annuli are used in determining the size of the homograft and the annuloplasty ring. However, selection of the homograft size is only done after completing radical debridement using Hegar dilators. While the homograft is being prepared, an incomplete mitral annuloplasty ring is sized and is inserted from trigone-to-trigone. The aortomitral homograft is trimmed in a fashion that preserves the anterior mitral leaflet in continuity with the aortic valve through the IVF. The trimmed anterior mitral leaflet of the homograft was sutured to the free edge of the native mitral leaflet base [8] (Fig.1). An annuloplasty ring is implanted in the mitral annulus to stabilize the intertrigonal area connected to the new IVF and avoid hypermobility of the anterior mitral leaflet into the left ventricular (LV) outflow tract (LVOT) causing systolic anterior motion. If the trimmed anterior mitral leaflet of the homograft is not large enough for reconstruction of the IVF and the base of the native anterior leaflet, we used a patch of bovine pericardium (n = 6) or Dacron patch (n = 1). (Video published on ctsnet.org: https://www.ctsnet.org/article/incorporated-aortomitral-homograft-double-valve-endocarditis-hemi-commando-procedure).

Figure 1:

(A) Infective endocarditis involving the aortic root and extending to the anterior mitral leaflet. (B) Aortomitral homograft. (C) Debridement of all infected tissues and the material. (D) Homograft anterior mitral leaflet sutured to the native anterior mitral leaflet edge. (E) Implantation of the aortomitral homograft. (F) Patch closure of the left atrium roof.

The technique for homograft implantation includes multiple running 3-0 Prolene sutures (≈ 4 or 5), which allow for maintaining adequate distance for optimum visualization of the LVOT when placing the sutures. A strip of bovine pericardium is used to support the proximal suture line, as most often the tissues are friable. A running 3-0 Prolene suture is used for the anastomosis between the homograft anterior mitral leaflet and the native anterior mitral leaflet. In case of significant destruction of the LVOT tissues, the post-debridement defect is patched with bovine pericardium before implanting the homograft. To avoid tension on the suture line, a patch of bovine pericardium can be used to reconstruct the roof of the left atrium when necessary (Fig. 1).

INDICATIONS FOR THE HEMI-COMMANDO VERSUS COMMANDO OPERATION

The degree of MV involvement in the pathological process is the key factor in choosing the hemi-Commando procedure over the Commando procedure. Based on the findings in the preoperative and intraoperative echocardiograms, the hemi-Commando procedure is considered in the following situations:

• An aortic root pathology that requires root replacement (e.g. aortic root abscess, fistula between the aortic root and another heart chamber).

• Involvement of the IVF and the anterior MV leaflet, shown as thickening of the tissue, increased echocardiographic density or communication with aortic root abscess.

• Mitral regurgitation caused solely by anterior leaflet pathology but spares the subvalvular apparatus and posterior leaflet/annulus.

On the other hand, extensive damage to the free edge of the anterior mitral leaflet, subvalvular apparatus and/or posterior leaflet/annulus that require complex repair or replacement would preclude the hemi-Commando procedure and favour the full Commnando procedure. The preoperative and intraoperative imaging data provide clues about the extent of debridement and whether mitral repair would be feasible. However, the final decision on the surgical approach is made intraoperatively after radical debridement of all infected tissues/materials and careful inspection of the MV.

Statistical analysis

All continuous data were expressed as mean with standard deviation or median with range. Categorical data were shown as units and percentages. The Kaplan–Meier survival plot analysis was performed in R statistical computing (Foundation for Statistical Computing, Vienna, Austria).

RESULTS

Cohort characteristics at the time of surgery

All patients had invasive IE extending to IVF, 27% (n = 10) with native and 73% (n = 27) with prosthetic aortic valve IE (Table 1). Preoperatively, 70% (n = 26) of patients were admitted to the intensive care unit (ICU), 11% (n = 4) were in septic shock, 14% (n = 5) were on mechanical ventilatory support, 27% (n = 10) suffered from cerebral embolic complications and 30% (n = 11) suffered from peripheral embolic complications. Twenty-nine (78%) patients had prior heart surgery and 15 (41%) presented for emergency surgery. Baseline transthoracic echocardiography showed more than moderate aortic regurgitation (AR) in 68% (n = 25) and more than moderate mitral regurgitation (MR) in 46% (n = 17) of patients. LV function was preserved in 73% of patients, while mild, moderate and moderately severe LV dysfunction were observed in 8, 11 and 8% of patients, respectively (Table 1).

Table 1:

Preoperative characteristics

Variables	n (%) or mean ± SD
Demographics
Age (years)	55.08 ± 16.59
Male	31 (84)
Body surface area (m2)	2.05 ± 0.27
Non-cardiac comorbidities
Hypertension	21 (57)
Diabetes	9 (16)
COPD	5 (14)
Stroke	15 (41)
CKD	57 (19)
IVDU	8 (22)
Cardiac comorbidities
Prior myocardial infarction	5 (14)
CAD	12 (32)
Atrial fibrillation	10 (27)
Prior cardiac surgery	29 (78)
First reoperation	21 (57)
Aortic valve replacement	18 (49)
Aortic valve repair	2 (5)
Aortic root replacement	4 (11)
Tetralogy of Fallot repair	1 (3)
Second reoperation	8 (22)
Aortic valve replacement	4 (11)
Mitral valve repair	1 (3)
CABG	2 (5)
Type A dissection repair	1 (3)
Endocarditis
Native valve	10 (27)
Prosthetic valve IE	27 (73)
Mitral valve pathology
Mitral annular dilation	4 (11)
Mitral flail leaflet	2 (5)
Mitral restricted leaflet	8 (22)
Mitral leaflet perforation	4 (11)
Mitral regurgitation
None/trivial	8 (22)
Mild	12 (32)
Moderate	11 (30)
Severe	6 (16)
Left ventricular dysfunction
None	27 (73)
Mild	3 (8)
Moderate	4 (11)
Severe	3 (8)
Ejection fraction	52.2 ± 12.3
Endocarditis-related comorbidities
Admission to ICU	26 (70)
LOS ICU	4 ± 2.35
Mechanical ventilation	5 (14)
Septic shock	4 (11)
Cerebral embolic complication	10 (27)
Peripheral embolic complication	11 (30)
Mycotic aneurysm	3 (8)
Timing of surgery
Emergency	15 (41)
Elective	22 (59)

CABG: coronary artery bypass grafting; CAD: coronary artery disease; COPD: chronic obstructive pulmonary disease; CKD: chronic kidney disease; ICU: intensive care unit; IE: infective endocarditis; IVDU: intravenous drug abuser; LOS: length of stay; SD: standard deviation.

Hemi-Commando procedure

The trimmed aortomitral homograft (size 25 ± 2 mm) was used for aortic root replacement in all patients. After debridement of the IVF and the native mitral leaflet, reconstruction was done using the trimmed homograft anterior mitral leaflet in 81% (n = 30) of the cohort, while a pericardial patch was used in 16% (n = 6) and a Dacron patch in 1 (3%) patient. Twenty-four (65%) patients required MV repair, with 15 (41%) patients receiving a prosthetic annuloplasty ring (Table 2). Additional procedures included tricuspid valve repair in 22% (n = 8) of the cohort, coronary bypass grafting in 14% (n = 5) and pacemaker lead replacement in 11% (n = 4). The total cardiopulmonary bypass and aortic cross-clamp times were 247 ± 73 min and 197 ± 55 min, respectively. Mechanical circulatory support including an intra-aortic balloon was required in 4 (11%) patients and extracorporeal membrane oxygenation (ECMO) in 1 (3%) patient. Seven (19%) patients were transferred to the ICU after surgery with an open chest for severe mediastinal tissue oedema or for haemodynamic instability. There were no intraoperative deaths (Table 2).

Table 2:

Operative data

Variables	n (%) or mean ± SD
Homograft sizes (mm)	25.44 ± 1.50
Reconstruction of IVF with homograft leaflet	30 (81)
Reconstruction of IVF with pericardial patch	6 (16)
Reconstruction of IVF with Dacron patch	1 (3)
MV repair	24 (65)
MV repair with annuloplasty ring	15 (41)
MV repair annuloplasty ring sizes	30.07 ± 1.53
TV repair	8 (22)
CABG (min)	5 (14)
PPM lead replacement (min)	4 (11)
Total CPB	247.56 ± 72.79
Total aortic cross clamp	197.72 ± 55.48
IABP	4 (11)
ECMO	1 (3)
Open chest to ICU	7 (19)

CABG: coronary artery bypass grafting; CPB: cardiopulmonary bypass; ECMO: extracorporeal membrane oxygenator; IABP: intra-aortic balloon pump; ICU: intensive care unit; IVF: intervalvular fibrosa; MV: mitral valve; PPM: permanent pacemaker; SD: standard deviation; TV: tricuspid valve.

Microbiology

The most common causative pathogens were the genera Staphylococcus (46%), Streptococcus (22%) and Enterococcus (8%). Six patients had negative preoperative blood cultures, while 18 valves were negative for microbial colonization according to standard cultures (Table 3). Molecular analysis (polymerase chain reaction, followed by sequencing) revealed the causative micro organism in 16 culture-negative valves while 1 valve was negative and 3 valves were not sent for analysis by polymerase chain reaction. Pathological examination of the sterile valve (negative culture and polymerase chain reaction) showed heavy calcification and pseudoaneurysm of the IVF, whereas the majority of the other valves showed microbial clusters embedded within a fibrinous exudate or vegetations (Table 3).

Table 3:

Microbiology data

Blood cultures	n	Valve cultures	n	Valve cultures + PCR	n
MRSA	8	MRSA	6	MRSA	6
MSSA	3	MSSA	2	MSSA	4
CoNS	7	CoNS	3	CoNS	7
Streptococcus	8	Streptococcus	3	Streptococcus	8
Enterococcus	3	Enterococcus	2	Enterococcus	3
Brucella	1	Propionibacterium	2	Propionibacterium	2
Erysipelothrix	1	Erysipelothrix	1	Erysipelothrix	1
Negative cultures	6	Negative cultures	18	Negative cultures	1
				Brucella	1
				Bartonella	1
				PCR test was not performed	3

CoNS: coagulase-negative Staphylococcus; MRSA: methicillin-resistant Staphylococcus aureus; MSSA: methicillin-sensitive Staphylococcus aureus; PCR: polymerase chain reaction.

Early outcomes

Postoperatively, the incidence of prolonged ventilation (>24 h) was 65% (n = 24), tracheostomy 14% (n = 5), stroke 5% (n = 2), renal dysfunction requiring dialysis 14% (n = 5), atrial fibrillation 49% (n = 18) and pacemaker insertion 33% (n = 12). There were 3 re-explorations for bleeding (9%) and no reoperations for valve failure. There were no perioperative myocardial ischaemic events. In addition to a single ECMO insertion case intraoperatively, there were 4 additional cases requiring ECMO insertion in the ICU. Two ECMO patients were weaned successfully from mechanical support and were discharged after prolonged hospital stays. There were 3 hospital deaths (8%) due to postoperative sepsis and multiorgan failure (2 patients received ECMO insertion and 1 patient had large stroke). The total ICU stay and hospital length of stay were 12 ± 12 days and 24 ± 13 days, respectively (Table 4).

Table 4:

Postoperative complications and early outcomes

Variables	n (%) or mean ± SD
Hospital death	3 (8)
Stroke	2 (5)
Prolonged ventilation >24 h	24 (65)
ECMO	4 (11)
AKI requiring dialysis	5 (14)
Reoperation for bleeding	3 (9)
AF postoperatively	18 (49)
Tracheostomy postoperatively	5 (14)
PPM implant	12 (33)
Sepsis	3 (9)
Hospital length of stay (days)	24.3 ± 13.49
ICU length of stay (days)	11.97 ± 11.76
Echocardiography
Aortic regurgitation
None/trivial	32 (86)
Mild	5 (14)
Moderate	0
Severe	0
Mitral regurgitation
None/trivial	23 (62)
Mild	12 (32)
Moderate	2 (5)
Severe	0
LV dysfunction
None	24 (65)
Mild	7 (19)
Moderate	4 (11)
Severe	2 (5)

AF: atrial fibrillation; AKI: acute kidney injury; ECMO: extracorporeal membrane oxygenator; IE: infective endocarditis; ICU: intensive care unit; LV: left ventricular; PPM: permanent pacemaker; SD: standard deviation.

Postoperative echocardiograms showed no AR in 86% (n = 32) of patients and mild regurgitation in 14% (n = 5), while MR prevalence was none/trivial in 62% (n = 23), mild in 32% (n = 12) and moderate in 5% (n = 2, treated medically and did not require intervention). Intact IVF reconstruction was confirmed in all patients with no abnormal flow communication between the left heart chambers (Fig. 2 and Video 1). The cohort had 65% of patients without LV dysfunction, 19% with mild dysfunction, 11% with moderate dysfunction and 5% with severe dysfunction (Table 4).

Figure 2:

(A) Preoperative transoesophageal echocardiogram showing aortic root abscess with thickening of the intervalvular fibrosa and the base of the anterior mitral leaflet. (B) Postoperative transoesophageal echocardiogram showing successful implantation of the aortomitral homograft (yellow arrow refers to the suture line between the base of the native anterior mitral leaflet and the homograft anterior mitral leaflet). Abs: abscess; AML: anterior mitral leaflet; AoR: aortic root; LA: left atrium; LV: left ventricle; MV: mitral valve.

Late outcomes

The mean follow-up period was 581 ± 729 days (10 patients followed up for 1 year). At the 1-month follow-up, 73% of patients had none/trivial AR, 27% had mild AR, 64% had none/trivial MR, 27% had mild MR and 1 patient had moderate MR. There was no LV dysfunction in 64% of patients, while 9% had mild, 18% had moderate and 1 patient had severe LV dysfunction. At the 1-year follow-up, 86% of patients had none/trivial AR, 14% had mild AR, 63% had none/trivial MR, 13% had mild MR and 25% had moderate MR. There was no LV dysfunction in 75% of patients but mild dysfunction in 25% (Table 5). There was 1 reoperation for aortic root pseudoaneurysm. One patient suffered from a stroke, and there were 2 post-discharge deaths: 1 patient from recurrent IE (1.5 years after surgery) and 1 patient at age 91 years who lived for 8 years after surgery. The 1-year survival was 91%, while 3-year survival was 82% (Fig. 3).

Table 5:

Follow-up

Variables	n/N (%) or mean ± SD
Patients discharged	34/37 (92)
Death after discharge	2 (5)
Reoperation	1 (3)
Patients followed up at 1-month visit	15/37 (41)
Patients followed up at 1-year visit	10/37 (27)
Patient follow-up beyond 3 years	5/37 (14)
Mean follow-up	580.9 ± 728.9 (n = 28)
Mean follow-up for patients followed up at the 1-year mark	824.9 ± 797.1 (n = 8)
Echocardiographic data during follow-up	1-month follow-up	1-year- follow up
Aortic regurgitation
None/trivial	8 (73)	6 (86)
Mild	3 (27)	1 (14)
Moderate	0	0
Severe	0	0
Mitral regurgitation
None/trivial	7 (64)	5 (63)
Mild	3 (27)	1 (13)
Moderate	1 (9)	2 (25)
Severe	0	0
LV dysfunction
None	7 (64)	6 (75)
Mild	1 (9)	2 (25)
Moderate	2 (18)	0
Severe	1 (9)	0
Ejection fraction	54.7 ± 9.3	56.83 ± 6.55

LV: left ventricle; SD: standard deviation.

Figure 3:

The Kaplan–Meier overall survival curve for all patients. The coloured area represents the 95% confidence limits.

Video 1:

Dynamic chest computed tomography showing intact reconstruction of the intervulvular fibrosa using the incorporated aortomitral homograft technique 2 months postoperatively.

COMMENTS

Invasive infective endocarditis

Recently, with a substantial increase in cardiac device implantations, there has been a paradigm shift in the causative organisms of IE; Staphylococcus species have become the leading cause, rather than Streptococcus species [9]. It has been shown in previous reports that Staphylococcus species are an independent predictor of poor outcomes [10–12]. In line with the other reports of double-valve IE, the most common pathogen in this series was Staphylococcus species, which are known as virulent, biofilm-forming micro-organisms causing persistent infection, tissue invasion and disintegration [4, 13]. These properties explain the development of aortic root abscesses extending to the MV through the IVF even while patients are receiving antimicrobial therapy. Hence, extensive debridement of all infected tissues and materials followed by complex reconstruction is mandatory to control the infection [14].

Reconstruction of the fibrous skeleton

Several surgical techniques have been described for the treatment of invasive double-valve IE. These include aortic root replacement using allograft, bioprosthetic or mechanical composite graft, reconstruction of the IVF using pericardial or Dacron graft patch along with MV replacement [1, 2, 15, 16]. Other groups used aortomitral homograft, mechanical, or bioprosthestic monobloc [5, 6]. This procedure is referred as the Commando operation at our institution, reflecting the extreme technical challenges it entails [2]. The principle of using the allograft anterior mitral leaflet to cover the aortic root abscess extending to the IVF has been described by other groups [17]. Our group has developed a modified surgical technique (hemi-Commando) for the treatment of double-valve IE that extends to the anterior mitral leaflet but spares the posterior leaflet and annulus [8]. In the former setting, an aortic homograft is implanted in the aortic root, while the trimmed homograft anterior mitral leaflet is used to reconstruct the IVF and the native anterior MV. This provides the advantages of reconstructing the fibrous skeleton using almost the same quality tissue as the native tissue with higher resistance to recurrent infection when compared with using the prosthetic material [3, 8, 18, 19]. Also, in terms of technicality, it is relatively less complex than a Commando operation. The aortic cross-clamp time for the hemi-Commando operation is comparable to most, if not shorter, of the reported Commando operation times, which range from 124 ± 42 min to 297 ± 71 min [1, 15, 16] (Table 2). More importantly, this approach preserves the MV and its subvalvular apparatus in high-risk surgical group. In the early cases of this series, the IVF and the anterior mitral leaflet were reconstructed without using an annuloplasty ring. Adding a mitral annuloplasty ring was found to be technically important to stabilize the intertrigonal area connected to the new IVF and to avoid hypermobility of the anterior mitral leaflet into the LVOT causing systolic anterior motion. The early and mid-term results of MV repair in this cohort were satisfactory; however, long-term follow-up is needed to better evaluate the durability of the repair technique.

Outcomes of the hemi-Commando procedure

Patients referred for surgical management of invasive IE, often have had failed medical therapy and present with embolic complications, shock or congestive heart failure in addition to their underlying comorbidities, as presented in our cohort (Table 1). Additionally, the majority of patients in our series had prior cardiac surgery and required redo sternotomy. Thus, this group of patients is a high-risk surgical group with a significantly morbid perioperative course. The reported early mortality for the Commando operation ranges from 6% to 32%, and the mid-term survival ranges from 30% to 80% [1, 2, 15, 16]. In this hemi-Commando series, we had 3 operative deaths (8%), low early mortality and comparable mid-term survival for this morbid group of patients (Table 5). Although aortic allograft has a theoretical advantage of resistance to infection and thus has been the preferred option for aortic root IE at our institute [2, 18, 19], the rate of recurrent infection was similar to other prostheses [1, 15, 16]. However, these results represent only mid-term outcomes of a small series.

David et al. [1, 20] reported initial high rates of valve-related reoperation for the Commando procedure (double-valve replacement for IE); however, this has improved with their growing experience. Other groups reported low incidence of valve-related reoperation [15, 16]. We had 1 late reoperation for aortic root pseudoaneurysm in our cohort. With regard to the Commando postoperative complications, the reported rate of reoperation for bleeding was 12–32%, neurological events 6–17%, acute renal failure 8–56% and pacemaker implantation 10–20% [1, 15, 16]. This hemi-Commando series had lower rates of reoperation for bleeding and neurological events while comparable rates of acute renal failure and pacemaker implantation (Table 4). Davierwala et al. [16] reported a 20% postoperative sepsis rate in their series, with 8% associated mortality (2 of 5 patients died due to multiorgan failure). David et al. [1] reported 4% postoperative sepsis in their series but no associated mortality. On the other hand, they had 6% early mortality secondary to cardiogenic shock and 3% secondary to multiorgan failure. No mechanical circulatory support was required in either report. In our series, 3 patients developed postoperative sepsis and multiorgan failure that accounts for the hospital deaths.

One of the key proposed advantages of the hemi-Commando procedure is preservation of the MV and its subvalvular apparatus and, therefore, conservation of LV function. Echocardiographic assessment of the MV function showed satisfactory results of the anterior leaflet reconstruction technique at discharge and during follow-up. Additionally, LV function was maintained through the perioperative period and improved during follow-up. This is an extremely important benefit for high-risk surgical patients with multiple comorbidities undergoing extensive and long surgical procedures, and this factor may have contributed to the low mortality in this cohort.

Limitations

This is a single-institution retrospective study of a relatively small cohort of endocarditis patients undergoing surgical treatment with only mid-term follow-up. Because patients are referred to us nationally and internationally, only a subgroup is followed up routinely by cardiologists at our institution, limiting information from echocardiographic surveillance. Additionally, this is a relatively new technique, so long-term outcomes are not available at this time.

CONCLUSIONS

The early and mid-term outcomes of the hemi-Commando operation suggest that it is a good alternative treatment for double-valve endocarditis, neither involving the posterior mitral leaflet nor the MV annulus. Despite the preoperative morbidities in this cohort, outcomes following the hemi-Commando were comparable, if not superior in some aspects, to the double-valve endocarditis surgery (Commando operation). More importantly, preserving the mitral subvalvular apparatus and subsequently the LV function provides an additional advantage in these high-risk patients. Long-term follow-up is needed to better evaluate and compare the hemi-Commando operation outcomes with the conventional double-valve replacement with IVF reconstruction (Commando operation).

Conflict of interest: José L. Navia discloses a financial relationship with Edwards Lifesciences, St. Jude Medical, MAQUET and NaviGate Cardiac Structures and A. Marc Gillinov with On-X, Edwards Lifesciences, Medtronic, St. Jude Medical, Abbott, AtriCure and Tendyne.

Author notes

Presented as a poster at the Annual Meeting of the American Association for Thoracic Surgery Mitral Conclave, New York, NY, USA, 28 April 2017.