- Email: [email protected]
Graft Selection for Aortic Root Replacement in Complex Active Endocarditis: Does It Matter?
Graft Selection for Aortic Root Replacement in Complex Active Endocarditis: Does It Matter? ADULT CARDIAC
Arminder Singh Jassar, MBBS, Joseph E. Bavaria, MD, Wilson Y. Szeto, MD, Patrick J. Moeller, BS, Jon Maniaci, Rita K. Milewski, MD, PhD, Joseph H. Gorman III, MD, Nimesh D. Desai, MD, PhD, Robert C. Gorman, MD, and Alberto Pochettino, MD Division of Cardiovascular Surgery, University of Pennsylvania Medical Center, Philadelphia, Pennsylvania
Background. Endocarditis affecting the aortic valve, with abscess formation and root destruction, remains a challenge to treat. Aortic root homografts have been advocated because of a perceived lower risk of infective complications than with other root replacement grafts. However, the theoretical advantage of homografts has not been re-evaluated in the modern era. This report is based on an examination of our results for all aortic root replacements in complex, active endocarditis affecting the aortic valve. Methods. From 2000 to 2010, 134 patients (70.9% male; mean age 58.3 ⴞ 14.8 years) at our institution underwent aortic root replacement for active endocarditis. Ninety of the patients (67.2%) had a previously implanted prosthetic aortic valve. Our findings for these patients included one or more of the following: abscess (n ⴝ 110, 82.1%), valve vegetation (n ⴝ 98, 73.1%), and pseudoaneurysm or rupture or both (n ⴝ 62, 46.3%). We retrospectively reviewed data for the patients from hospital records and the social security data base.
Results. A mechanical composite graft (MC) was used in 43 of the patients (32.1%), a non-homograft biologic valve conduit (BC) in 55 patients (41.0%), and a homograft (HG) valve in 36 patients (26.9%). There was no significant difference among the groups in the incidence of major complications or in-hospital mortality. During a mean follow-up of 32.1 ⴞ 29.4 months, the rates of readmission, reinfection, and reoperation were similar for the three groups. The mean 5-year survival in the study was 58 ⴞ 9% for the MC group, 62 ⴞ 7% for the BC group, and 58 ⴞ 9% for the HG group, respectively (p ⴝ 0.48). Conclusions. Aortic root replacement in the presence of complex active infection is associated with significant morbidity and mortality. We report that the rates of major complications and late mortality were similar among MC, BC, and HG groups in our study.
I
they pose for reoperation because of structural deterioration represent drawbacks to their use. This has led to the use of other prosthetic materials to replace infected aortic roots. A number of previous studies have examined the outcomes of xenografts and mechanical prosthetic implants for aortic valves with IE. Most of these studies have examined the role of valve choice in isolated aortic valve replacement (AVR) [6, 9 –11]. The role of prosthetic material in the setting of advanced, destructive endocarditis requiring complete aortic root reconstruction is less well studied. In this report we describe a large series of patients with active advanced IE of the aortic valve that necessitated complete reconstruction of the aortic root. The report includes an examination of short-term morbidity and mortality data for these patients, as well as of long-term rates of reinfection and reoperation stratified by the choice of graft utilized for aortic root reconstruction.
nfective endocarditis (IE) effects about 15,000 patients yearly in the United States [1]. Despite advances in antimicrobial therapies, outcomes in IE remain poor [2, 3]. Surgery is indicated in patients with heart failure or cardiogenic shock caused by valvular dysfunction. Surgery should also be undertaken in hemodynamically stable patients with aggressive infection (ie, aortic or periannular abscess, heart block, recurrent emboli despite appropriate antibiotic therapy), infections resistant to antibiotic therapy, and fungal endocarditis [4]. The optimal type of aortic prosthesis for patients with IE remains unclear, with cryopreserved human allograft (homograft) prostheses often recommended [5–7]. However, although homografts may confer a theoretical advantage in resisting infection through their lack of artificial materials [8], their limited availability and the need
Accepted for publication Sept 28, 2011. Presented at the Poster Session of the Forty-seventh Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 31–Feb 2, 2011. Address correspondence to Dr Pochettino, Division of Cardiovascular Surgery, University of Pennsylvania School of Medicine, Hospital of the University of Pennsylvania, 3400 Spruce St, 6 Silverstein Pavilion, Philadelphia, PA 19104-4283; e-mail: [email protected].
© 2012 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2012;93:480 – 8) © 2012 by The Society of Thoracic Surgeons
Patients and Methods Patients Between January 2000 and March 2010, 1,337 aortic root replacements were performed at the University of Penn0003-4975/$36.00 doi:10.1016/j.athoracsur.2011.09.074
sylvania Health System (valve-sparing aortic root reconstructions were excluded). Of the patients who received these root replacements, 192 had a diagnosis of aortic valve endocarditis. Endocarditis was classified according to criteria of The Society of Thoracic Surgeons (STS) as being active if the patient was being treated for endocarditis at the time of operation, or as treated if no antibiotic medication other than prophylactic medication was being given at the time of surgery. Fifty-eight patients with treated endocarditis were excluded; the data for 134 patients with active aortic valve endocarditis were analyzed for the purpose of this report.
Operative Procedure All 134 patients in the study underwent aortic root replacement through a sternotomy. A repeat sternotomy was required in 93 patients (90 of whom had a previous AVR or aortic root replacement). In patients with a high risk of aortic injury on surgical entry (eg, those with an aneurysm or pseudoaneurysm adherent to the posterior aspect of the sternum), peripheral cannulation was performed before sternotomy. In all other patients arterial cannulation was established through the ascending aorta, with venous cannulation performed via the right atrium, superior vena cava, and inferior vena cava as necessary. If reconstruction of the aortic arch was performed concomitantly, deep hypothermic circulatory arrest with adjunctive retrograde cerebral perfusion was used. Aggressive debridement was done to remove all infected, devitalized, or suspicious tissue. Any cardiac structural defects caused by infection or debridement (eg, defects in aorto-mitral continuity; defects in the left ventricle, left atrium, or mitral valve; or ventricular septal defects) were repaired with pericardial tissue (autologous or bovine) or a with a Dacron patch at the discretion of the surgeon. Complete replacement of the aortic root was done in all of the patients, with coronary ostia reimplanted as buttons. Composite grafts with a mechanical valve (MC) were used in 43 patients (32.1%), nonhomograft conduits with biologic valves (BC) were used in 55 patients (41.0%) (39 porcine full roots and 16 “biologic composite” grafts, consisting of a stented tissue valve sewn to a Dacron tube graft), and homografts (HG) were used in 36 patients (26.9%).
Postoperative Care Intravenous antibiotics were continued for at least 6 weeks postoperatively. The antibiotic regimen for each patient was based on microbiologic data, with a combination of broad-spectrum antibiotic agents used in the absence of positive cultures. In patients with fungal endocarditis or recurrent prosthetic valve endocarditis, longer periods of suppression with oral antibiotics were used. The antibiotic regimens did not depend on the type of root replacement done in a patient.
Statistical Analysis The University of Pennsylvania aortic surgery database, inpatient and outpatient charts, echocardiography re-
JASSAR ET AL AORTIC ROOT REPLACEMENT FOR ENDOCARDITIS
481
ports, and records of any readmissions or reoperations were reviewed as sources for the study data. Follow-up data were available for 84 (80.8 %) of the 104 patients who were discharged alive. The mean follow-up period in the study was 32.1 ⫾ 29.4 months (MC ⫽ 27.4 ⫾ 22.1 months [median, 24.1 months], BC ⫽ 31.7 ⫾ 28.5 months [median, 20.2 months], HG ⫽ 40.5 ⫾ 39.7 months [median, 27.2 months], p ⫽ 0.33). Only 5 patients (5.95%) were found to have definite recurrent or reinfective endocarditis (according to the modified Duke Criteria [12]) (n ⫽ 1, MC group; n ⫽ 2, BC group; n ⫽ 2, HG group). However, instances of possible endocarditis and some cases in which a diagnosis of endocarditis was rejected (eg, root pseudoaneurysm, or fever with mediastinal collection, or sepsis, and death due to cardiac arrest prior to completion of workup) were included, to encompass the total burden of infection. Similarly, for readmissions and reoperations, all occurrences related to patients’ cardiovascular status (infective or not) were indexed. Survival data were supplemented from online Social Security Death Index database [13]. Continuous variables were compared through analysis of variance (ANOVA) with post hoc analysis. Categorical variables were compared through Pearson’s 2 test and Fisher’s exact test. Adjustments were not made for multiple testings. A multivariate logistic regression model was used to determine predictors of in-hospital and late mortality. Input covariates for the logistic regression model for in-patient mortality and the Cox regression model for late mortality included patient age, sex, presence of diabetes mellitus, renal failure, abscess, positive cultures, and the type of prosthesis used. The odds ratio (OR) and 95% confidence interval (CI) were calculated for each risk factor. The percent rates of readmission, reinfection, reoperation, and survival were analyzed with the Kaplan–Meier method, and between-group comparisons were made with the log-rank test and presented as the mean ⫾ standard error.
Results Preoperative Patient Characteristics and Graft Selection Preoperative characteristics of the study patients are shown in Table 1. The types and numbers of native and prosthetic aortic valve grafts for the patients in the study are shown in Table 2. The presence of a bicuspid or tricuspid, native, or prosthetic aortic valve at the time of surgery did not influence graft selection. Incidences of abscess in the aortic root or surrounding regions, and of aortic valve vegetation, dehiscence, pseudoaneurysm, fistula, or rupture are shown in Table 3. Microbiologic data (preoperative blood cultures or intraoperative tissue cultures) are shown in Table 4. A causative organism was identified in 115 patients (89.1%). Cultures from 24 patients (18.6%) grew more than one organism. Intraoperative tissue-culture data were available for 117 patients (87.3%), with the causative organism cultured in 58 patients (49.6%).
ADULT CARDIAC
Ann Thorac Surg 2012;93:480 – 8
482
JASSAR ET AL AORTIC ROOT REPLACEMENT FOR ENDOCARDITIS
Ann Thorac Surg 2012;93:480 – 8
Table 1. Preoperative Characteristics of Patients Undergoing Aortic Root Replacement Patient Characteristics ADULT CARDIAC
Age (years) Male Dyslipidemia CVD CVA Chronic lung disease DM HTN Renal failure/HD Smoker Myocardial infarction Heart failure by NYHA Class Class I Class II Class III Class IV Ejection fraction ⱕ 30 % 31% – 40% 41% – 50% 51% – 60% ⬎60% Cardiogenic shock CPR Urgent/emergent a
All (n ⫽ 134)
Mechanical Composite Graft (n ⫽ 43)
Biologic Conduit (n ⫽ 55)
Homograft (n ⫽ 36)
58.3 ⫾ 14.8 95 66 46 37 27 29 91 24/9 57 19
51.8 ⫾ 12.7 38 15 16 13 4 9 26 3/0 18 7
62.7 ⫾ 15.1 35 34 21 17 16 10 39 8/2 22 6
59.4 ⫾ 14.3a,b 22a,b 17 9 7 7a 10 26 13/7b,c 17 6
22 15 49 30
7 7 19 4
2 2 16 18
3 6 14 8
12 13 29 47 20 11 3 96
3 4 13 17 3 3 0 30
6 5 8 19 11 4 2 36
3 4 8 11 6 4 1 30c
p ⬍ 0.05 mechanical conduit vs biologic graft.
CPR ⫽ cardiopulmonary resuscitation; hemodialysis; HTN ⫽ hypertension;
b
p ⬍ 0.05 mechanical conduit vs homograft.
c
p ⬍ 0.05 biologic graft vs homograft.
CVA ⫽ cerebrovascular accident; CVD ⫽ cerebrovascular disease; NYHA ⫽ New York Heart Association.
Utilization of MC and BC conduits became more frequent during the course of this study (Fig 1). The use of HG declined throughout the course of the study (n ⫽ 27, 42.2% from 2000 to 2005, vs n ⫽ 9, 12.9% from 2006 to 2010).
Operative Procedures Details of concomitant procedures and mean operative times are listed in Table 5. A hemiarch replacement was performed in 28 patients (13 with a previous AVR, 10 with previous root replacements, 5 with native bicuspid aortic
DM ⫽ diabetes mellitus;
HD ⫽
valves). Coronary artery bypass was required in 16 patients (7 because of involvement or destruction of coronary ostia by infection, 6 because of occlusive native coronary disease, 1 because of stenosis of a previous aortocoronary bypass, 1 because of native right coronary artery kinking, and 1 because of injury to a left internal mammary artery graft upon re-entry. Total cardiopulmonary bypass time and aortic crossclamp times were longer in the MC group than in the HG group (p ⫽ 0.027 and p ⫽ 0.01, respectively), probably
Table 2. Graft Selection in Replacement of Native and Prosthetic Aortic Valves Valve Type Native AV BAV Prosthetic AV Previous AVR Previous root replacement
All (n ⫽ 134)
Mechanical Composite Graft (n ⫽ 43)
Biologic Conduit (n ⫽ 55)
Homograft (n ⫽ 36)
44 18 90 69 21
12 7 31 20 11
19 9 36 31 5
13 2 23 18 5
There was no significant difference between groups by pairwise 2 comparison. AV ⫽ aortic valve;
AVR ⫽ aortic valve replacement;
BAV ⫽ biologic aortic valve.
Ann Thorac Surg 2012;93:480 – 8
483
JASSAR ET AL AORTIC ROOT REPLACEMENT FOR ENDOCARDITIS
Anatomic Findings
All (n ⫽ 134)
Mechanical Composite Graft (n ⫽ 43)
Biologic Conduit (n ⫽ 55)
Homograft (n ⫽ 36)
110 98 62 114
36 28 17 37
46 42 30 47
28 28 15 30
Abscess Vegetations Dehiscence/rupture/pseudoaneurysm/fistula Abscess or dehiscence/rupture/pseudoaneurysm/fistula
There was no significant difference between groups by pairwise 2 comparison. IE ⫽ infective endocarditis.
because of the higher incidence of concomitant procedures in the former group.
In-Hospital Complications and Mortality In-hospital mortality, postoperative length of stay, and the incidences of major complications are presented in Table 6. Overall perioperative mortality was 22.4%, and did not differ significantly by type of root replacement.
Readmission, Reinfection, Reoperation and Long-Term Survival Of the 104 discharged patients, long-term follow-up data were available for 84 (80.9%). Of these 84 patients, 18 (21.4%) were readmitted, of whom 8 (27.6%) were in the MC group (n ⫽ 5 for reinfection; n ⫽ 1 for pseudoaneurysm because of a dehisced coronary button; n ⫽ 1 for heart failure; and n ⫽ 1 for bilateral pleural effusions requiring chest-tube drainage); 5 (13.5%) were in the BC group (n ⫽ 3 for reinfection; n ⫽ 1 for hemorrhagic stroke; and n ⫽ 1 for pleural effusion requiring pleurodesis); and 5 (27.7%) were in the HG group (n ⫽ 5, for reinfection). Freedom from readmission was 76% ⫾ 8% at 1 year of follow-up and 60% ⫾ 13% at 5 years of follow-up for patients in the MC group; 88% ⫾ 6% at 1 year and 83% ⫾ 7% at 5 years of follow-up for patients in the BC group; and 63% ⫾ 12% at both 1 and 5 years of follow-up for patients in the HG group (p ⫽ 0.16) (Fig 2A). Overall, 14 patients (16.6 %) had a diagnosis of reinfection (n ⫽ 5,
MC group; n ⫽ 3, BC group; n ⫽ 6, HG group; 1 patient with infection in the HG group was treated as an outpatient). Freedom from reinfection was 84% ⫾ 7% at 1 year and 74% ⫾ 10% at 5 years for patients in the MC group; 94% ⫾ 4% at 1 year and 89% ⫾ 6% at 5 years for patients in the BC group; and 75% ⫾ 11% at 1 year and 64% ⫾ 14% at 5 years for patients in the HG group (p ⫽ 0.10) (Fig 2B). Six patients (7.1%) required cardiac-related reoperation. One patient in the MC group underwent drainage of a mediastinal collection and muscle flap coverage of substernal wound, and 1 patient underwent repeat sternotomy and repair of a dehisced coronary button 1 month postoperatively. Two patients in the BC group underwent repeat aortic root replacement, the first with a homograft at 3 months after the first root replacement operation, because of a root abscess, valve vegetation, and stroke; and the second with an MC conduit at 2.1 years after initial root replacement, because of aortic stenosis and endocarditis. Two patients in the HG group underwent reoperation. The first of these patients had a repeat HG implantation at 7 months postoperatively, because of infection and dehiscence of the first homograft; and the second underwent mitral valve repair for regurgitation at 11 months after initial placement of an HG root. Freedom from reoperation at 1 and 5 years was 96% ⫾ 4% and 89% ⫾ 8% respectively in the MC group; 97% ⫾ 3% and 90% ⫾ 7% respectively in the BC group; and 86% ⫾ 9% at both 1 and 5 years in the HG group (p ⫽ 0.67) (Fig 3A).
Table 4. Organisms Cultured From Aortic Roots Affected by Infectious Endocarditis Preoperative Blood Cultures or Operative Tissue Cultures Culture-negative Fungus MRSA MSSA Coagulase-negative Staphylococcus Enterococcus Streptococcus Gram-negative rods Others Positive operative tissue culturesb
All (n ⫽ 129)
Mechanical Composite Graft (n ⫽ 42)
Biologic Conduit (n ⫽ 55)
Homograft (n ⫽ 33)
14 12 16 18 21 19 30 10 15 58 (den ⫽ 117)
4 6 4 5 8 5 13 5 5 22 (den ⫽ 39)
7 4 9 4 9 8 14 2 8 23 (den ⫽ 50)
3 2 3 9a 5 6 3 3 2 13 (den ⫽ 28)
p ⬍ 0.05 biologic conduit vs homograft. Other comparisons did not show significant difference by pairwise 2 comparison. 117 of 134 patients. Denominators (den) are listed for each group.
a
MRSA ⫽ methicillin-resistant Staphylococcus aureus;
MSSA ⫽ methicillin-sensitive Staphylococcus aureus.
b
Data were available for
ADULT CARDIAC
Table 3. Anatomic Findings in Cases of IE Affecting the Aortic Valve
484
JASSAR ET AL AORTIC ROOT REPLACEMENT FOR ENDOCARDITIS
Ann Thorac Surg 2012;93:480 – 8
Fig 1. Yearly trend in prosthetic graft selection.
ADULT CARDIAC
The overall cumulative survival (including in-hospital mortality) of the patients in the study at 1 year and 5 years was 67.9% and 58.7%, respectively. The 1-year and 5-year rates of survival were 67% ⫾ 7% and 58% ⫾ 9% respectively for patients in the MC group; 65% ⫾ 7% and 62% ⫾ 7% respectively for patients in the BC group; and 61% ⫾ 8% and 58% ⫾ 9% respectively for patients in the HG group (p ⫽ 0.48) (Fig 3B).
Predictors of In-Hospital and Late Mortality The results of risk-factor analysis for predictors of inhospital mortality are shown in Table 7. Abscess (OR ⫽ 9.15; 95%CI, 1.05 to 80.18) and a history of diabetes (OR ⫽ 4.65; 95%CI, 1.64 to 13.17) significantly increased the risk of in-hospital mortality. Significant multivariate predictors of late mortality included diabetes (OR ⫽ 2.25; 95%CI, 1.27 to 3.99) and renal failure (OR ⫽ 2.17; 95%CI, 1.17 to 4.04). The type of prosthesis used was not predictive of either in-hospital or late mortality.
Comment Endocarditis often requires surgical intervention, which needs to be safe and effective, with a minimal risk of reinfection [2, 14]. Lytle and Sabik reported excellent results after homograft replacement for the treatment of prosthetic aortic valve endocarditis, with a mortality ranging from 3% to 13% [5, 6, 8]. Most other reports show much higher mortality rates. Several factors can explain the higher mortality rate in our series. One of these is that the patients in our study had extensive disease, as evidenced by abscess cavities in most of them (82.1%), and pseudoaneurysm, rupture, dehiscence, or a fistula in a significant fraction (46.2%). Seventy one percent of our patients underwent urgent or emergent operations, and about two thirds of our patients had prosthetic valve infection, which has been shown to worsen outcomes [9]. Our results are similar to results recently reported by Musci and colleagues [15], who presented data for aortic root replacement with homograft roots in a patient pop-
Table 5. Concomitant Procedures Done With Aortic Root Replacement
Procedures Any additional procedures Hemiarch CABG Mitral valve replacement Mitral valve repair Tricuspid valve repair/replacement Perfusion time (minutes) Aortic clamp time (minutes) Total circulatory arrest time (minutes) p ⬍ 0.05 mechanical vs homograft. comparison. a
CABG ⫽ coronary artery bypass grafting.
b
All (n ⫽ 134)
Mechanical Composite Graft (n ⫽ 43)
Biologic Conduit (n ⫽ 55)
Homograft (n ⫽ 36)
90 28 16 20 16 11 293 ⫾ 99.7 227.1 ⫾ 75.3 29.0 ⫾ 15.5 (n ⫽ 29)
33 13 4 7 9 1 308.3 ⫾ 98.2 244.1 ⫾ 78.7 31.6 ⫾ 19.5 (n ⫽ 14)
38 13 5 10 4 7 300.8 ⫾ 109.5 230.3 ⫾ 76.5 27.5 ⫾ 10.8 (n ⫽ 13)
19a,b 2a,b 7 3 3 3 262.6 ⫾ 80.0a 201.9 ⫾ 63.6a 21.0 ⫾ 9.9 (n ⫽ 2)
p ⬍ 0.05 biologic vs homograft. Other comparisons did not show significant difference by pairwise 2
Ann Thorac Surg 2012;93:480 – 8
JASSAR ET AL AORTIC ROOT REPLACEMENT FOR ENDOCARDITIS
485
Complications
All (n ⫽ 134)
Mechanical Composite Graft (n ⫽ 43)
Biologic Conduit (n ⫽ 55)
Homograft (n ⫽ 36)
Freedom from all major complications In-hospital mortality Length of stay Septicemia Deep sternal infection Permanent stroke Reoperation for bleeding/tamponade Renal failure/hemodialysis Cardiac arrest Heart block Multisystem organ failure Prolonged ventilation (⬎24 hours)
36 30 18.2 ⫾ 16.6 18 3 5 12 26/12 10 27 16 51
15 8 19.7 ⫾ 21.3 9 1 1 5 6/4 4 9 8 17
12 13 15.7 ⫾ 13.8 7 1 1 4 14/6 2 15 5 23
9 9 20.1 ⫾ 13.4 2 1 3 3 6/2 4 3a 3 11
a
p ⬍ 0.05 biologic conduit vs homograft. Other comparisons did not show significant difference by pairwise 2 comparison.
ulation similar to ours (having an 86% incidence of aortic root abscess). They experienced a 30-day mortality rate of 21.2%, which increased significantly for operations done in the setting of prosthetic valve endocarditis (25.4% vs 16.1% for native valve endocarditis). Leyh and associates [16] compared outcomes after aortic root replacement for prosthetic valve endocarditis with use of either aortic allografts or mechanical composite grafts. In their small series (n ⫽ 29), they reported an 18.5% overall rate of inhospital mortality (18.7% for patients given allografts and 15.4% for those given composite grafts). As a frame of reference, our in-hospital mortality for aortic root replacement in lower risk patients is approximately 2.5% [17], emphasizing the significant effect of infection on outcome. In our study, the HG group failed to show any difference from the MC or BC group in the rates of in-hospital complications, mortality, or long term reinfection. Hagl and colleagues [18] have previously reported an inhospital mortality of 11% in 28 patients who underwent aortic root replacement (25 of whom received mechanical composite grafts and 3 of whom received homografts) for prosthetic valve endocarditis. Abscess was found in 57% of these patients, and recurrent endocarditis occurred in 4% over a median follow-up of 44.5 months. Hagl and colleagues used historical data from the literature to conclude that the use of artificial prosthetic material
yields results comparable to those with homografts. David and coworkers [19] reported outcomes after surgery for paravalvular abscess (aortic, mitral, tricuspid, or pulmonary) in 135 patients. The overall operative mortality among these patients was 15.5%, with no statistically significant difference in the mortality of patients who received a mechanical valve (13.6%), a bioprosthetic valve (16.3%), or a homograft (21.4%). The 10-year survival in this patient group was 57% ⫾ 5% and the rate of recurrent endocarditis was 14%, and no predictors for recurrent endocarditis could be identified. David and coworkers concluded that the recognition and extirpation of all infected aortic valve tissue was critical, but that a homograft was preferred because of ease of handling and the possibility of using homograft mitral leaflet tissue to patch defects created by abscess resection. Similar results were reported by Moon and associates [9] in a retrospective review of data for 306 patients with endocarditis (211 with active disease, 95 with treated disease) who underwent replacement of the aortic or mitral valve with a bioprosthesis (n ⫽ 221), mechanical valve (n ⫽ 65), or homograft (n ⫽ 20) in the setting of either a native (n ⫽ 209) or a prosthetic valve (n ⫽ 97) infection. They reported an overall mortality of 18% ⫾ 2% for these patients. Even though 10-year survival was better for patients with native valve endocarditis than for those Fig 2. Kaplan–Meier analysis demonstrating (A) freedom from readmission and (B) freedom from reinfection after mechanical valve conduit (M), biologic valve conduit (B), or homograft (H) replacement of infected aortic root.
ADULT CARDIAC
Table 6. Postoperative Complications
486
JASSAR ET AL AORTIC ROOT REPLACEMENT FOR ENDOCARDITIS
Ann Thorac Surg 2012;93:480 – 8
Fig 3. Kaplan–Meier analysis demonstrating (A) freedom from reoperation and (B) long term survival after mechanical valve conduit (M), biologic valve conduit (B), or homograft (H) replacement of infected aortic root. ADULT CARDIAC
similar rates of abscess formation. Our study focused on patients with extensive native or prosthetic valve endocarditis, which could not be treated with valve replacement alone and requires replacement of the entire aortic root. A significant number of patients in our study underwent concomitant surgical procedures, mainly hemiarch replacement, and mitral valve interventions. Although the differences were not statistically significant, the number of these procedures were higher in the MC group than in the other two groups in the study. Without available donor mitral tissue, as in a homograft, the degree of distortion in the anterior leaflet of the mitral valve after the implantation of a rigid mechanically valved conduit may require further reparative techniques to ensure competency of the mitral valve. Our study has limitations inherent in any retrospective nonrandomized study. The evolution of surgical techniques and of the conduits available for implantation may have affected the outcomes of the study, with this effect potentially more prominent in the HG group. Graft selection was based on the preference and judgment of the surgeon. A greater fraction of patients in the MC group than in the BC or HG groups were younger males; patients in the MC group had a lower incidence of preoperative renal dysfunction. We have tried to adjust for some of these variables using the multivariate logistic and Cox regression
with prosthetic valve endocarditis (54% vs 41%) in their patient population, there was no difference in relation to the type of valve implanted. A recent study by Klieverik and colleagues [20] compared the outcome after AVR for active endocarditis in 138 patients, of whom 106 received allografts and 32 received mechanical prostheses). There was no survival benefit for patients who had an allograft over those who had a mechanical valve replacement (59% vs 66%, respectively). The rate of recurrent infection did not differ for the two groups. The 15-year freedom from reoperation was significantly higher in the mechanical valve group (93% vs 76%). However, in their study, only 18% of patients in the mechanical valve group had root abscesses, as compared with 38% in the allograft group, with homografts having been selected for more advanced disease. Similarly, Avierinos and associates [21], in a study of 127 patients with active endocarditis (of whom 50% had abscesses) who had either a homograft (n ⫽ 54) or a conventional prosthesis (n ⫽ 54, a bioprosthesis; n ⫽ 23, a mechanical prosthesis), found no significant difference between the two groups in operative mortality or event-free 5-year survival. However, the majority of patients with annular abscesses (65%) in their study were treated with homografts, which suggests a bias towards using homografts in patients with more severe disease [22]. In our study, the HG, MC, and BC groups had
Table 7. Multivariate Risk Factor Analysis for Predictors of In-Hospital and Late Mortality In-Hospital Mortality
Late Mortality
95% CI for OR Effect Group Age Sex Diabetes Renal failure Abscess Culture-positive Mechanical composite graft Biologic conduit
OR
Lower Limit
Upper Limit
OR
Lower Limit
Upper Limit
1.005 2.515 4.912 2.761 9.153 0.995 1.331 1.434
0.970 0.883 1.747 0.893 1.045 0.219 0.351 0.4444
1.041 7.166 13.811 8.535 80.184 4.519 5.037 4.633
1.014 1.542 2.253 2.177 1.291 1.293 1.142 0.888
0.993 0.870 1.270 1.170 0.610 0.552 0.556 0.467
1.035 2.731 3.998 4.048 2.731 3.029 2.343 1.690
For prosthetic type, homograft was considered the reference. CI ⫽ confidence interval;
95% CI for OR
OR ⫽ odds ratio.
analysis. It is important, however, to recognize that all three of the groups in the study had infection of similar severity (similar abscess rates). The frequencies of concomitant procedures were significantly lower for patients in the HG group, which resulted in shorter operating times for this group. An incomplete long-term follow-up (80.8%) may have failed to capture patients who sought care at other facilities, and the rates of reinfection and reoperation in our study may have been underestimated. The mortality data for the study are complete because the social security death index was used to determine this. Our approach to graft selection has evolved over time, as shown in Figure 1. Homografts were preferentially used early in our experience owing to the early reports of a greater freedom from reinfection with their use, and before 2000, homografts were used exclusively for aortic roots affected by the aggressive infections described in this study. Over time we have introduced alternate conduits for treating extensive aortic valve endocarditis requiring root replacement. At our center, the choice of conduit is now mainly governed by the age of the patient and the presence of relative or absolute contraindications to warfarin. Additionally, MC conduits have become the most frequent choice for root replacement in patients who have had a previous prosthetic root (11 of the 21 such patients in our study [52%]), to avoid a third root replacement related to structural deterioration of a BC or the HG replacement. In our practice, HG may continue to play a role in patients with contraindications to warfarin or those who require extensive mitral valve augmentation as part of their treatment for severe root abscess. In conclusion, we found that active infective aortic valve endocarditis with root destruction remains a challenging condition, with significant morbidity and mortality. In our experience, the type of aortic root prosthesis used for patients with roots affected by IE did not affect the outcomes of root replacement. As other authors have stated before us, we believe that the complete eradication of IE through aggressive debridement, and not the choice of conduit used for root replacement, is the critical factor in determining the ultimate outcome of such replacement. The type of prosthetic root to be implanted should be selected on the basis of the complete clinical picture for a given patient, including the technical requirements for reconstruction, the patient’s age and comorbidities, an assessment of compliance to medication, and follow-up, and the availability and hemodynamic performance of a conduit of the appropriate size. The authors acknowledge the help and support provided by Mary E. Mckenna, BSN, toward data collection for this study.
JASSAR ET AL AORTIC ROOT REPLACEMENT FOR ENDOCARDITIS
4.
5.
6. 7.
8.
9. 10. 11. 12. 13. 14. 15.
16.
17.
18.
19. 20.
References 1. Bashore TM, Cabell C, Fowler V, Jr. Update on infective endocarditis. Curr Probl Cardiol 2006;31:274 –352. 2. Gaca JG, Sheng S, Daneshmand MA, et al. Outcomes for endocarditis surgery in North America: a simplified risk scoring system. J Thorac Cardiovasc Surg 2011;141:98 –106 e2. 3. David TE, Gavra G, Feindel CM, Regesta T, Armstrong S, Maganti MD. Surgical treatment of active infective endocar-
21. 22.
487
ditis: a continued challenge. J Thorac Cardiovasc Surg 2007;133:144 –9. Bonow RO, Carabello BA, Chatterjee K, et al. 2008 focused update incorporated into the ACC/AHA 2006 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 (Writing Committee to revise the 1998 guidelines for the management of patients with valvular heart disease). J Am Coll Cardiol 2008;52:e1–142. Lytle BW, Sabik JF, Blackstone EH, Svensson LG, Pettersson GB, Cosgrove DM III. Reoperative cryopreserved root and ascending aorta replacement for acute aortic prosthetic valve endocarditis. Ann Thorac Surg 2002;74:S1754 –7; discussion: S1792–9. Lytle BW, Priest BP, Taylor PC, et al. Surgical treatment of prosthetic valve endocarditis. J Thorac Cardiovasc Surg 1996;111:198 –210. Perrotta S, Aljassim O, Jeppsson A, Bech-Hanssen O, Svensson G. Survival and quality of life after aortic root replacement with homografts in acute endocarditis. Ann Thorac Surg 2010;90:1862–7. Sabik JF, Lytle BW, Blackstone EH, Marullo AG, Pettersson GB, Cosgrove DM. Aortic root replacement with cryopreserved allograft for prosthetic valve endocarditis. Ann Thorac Surg 2002;74:650 –9. Moon MR, Miller DC, Moore KA, et al. Treatment of endocarditis with valve replacement: the question of tissue versus mechanical prosthesis. Ann Thorac Surg 2001;71:1164 –71. Sweeney MS, Reul GJ, Jr., Cooley DA, et al. Comparison of bioprosthetic and mechanical valve replacement for active endocarditis. J Thorac Cardiovasc Surg 1985;90:676 – 80. Newton S, Hunter S. What type of valve replacement should be used in patients with endocarditis? Interact Cardiovasc Thorac Surg 2010;11:784 – 8. Li SL, Sexton DJ, Mick N, et al. Proposed modifications to the Duke criteria for the diagnosis of infective endocarditis. Clin Infect Dis 2000;30:633– 8. Social Security Death Index database. Available at: http:// ssdi.rootsweb.ancestry.com. Accessed January 9, 2011. Leontyev S, Borger MA, Modi P, et al. Redo aortic valve surgery: Influence of prosthetic valve endocarditis on outcomes. J Thorac Cardiovasc Surg 2011;142:99 –105. Musci M, Weng Y, Hubler M, et al. Homograft aortic root replacement in native or prosthetic active infective endocarditis: twenty-year single-center experience. J Thorac Cardiovasc Surg 2010;139:665–73. Leyh RG, Knobloch K, Hagl C, et al. Replacement of the aortic root for acute prosthetic valve endocarditis: prosthetic composite versus aortic allograft root replacement. J Thorac Cardiovasc Surg 2004;127:1416 –20. Desai ND, McCarthy F, Moser W, et al. Durability of porcine bioroots in younger patients with aortic root pathology: a propensity-matched comparison with composite mechanical roots. Ann Thorac Surg. 2011 Aug 11. [Epub ahead of print] Hagl C, Galla JD, Lansman SL, et al. Replacing the ascending aorta and aortic valve for acute prosthetic valve endocarditis: is using prosthetic material contraindicated? Ann Thorac Surg 2002;74:S1781–5; discussion: S1792–9. David TE, Regesta T, Gavra G, Armstrong S, Maganti MD. Surgical treatment of paravalvular abscess: long-term results. Eur J Cardiothorac Surg 2007;31:43– 8. Klieverik LM, Yacoub MH, Edwards S, et al. Surgical treatment of active native aortic valve endocarditis with allografts and mechanical prostheses. Ann Thorac Surg 2009;88:1814 –21. Avierinos JF, Thuny F, Chalvignac V, et al. Surgical treatment of active aortic endocarditis: homografts are not the cornerstone of outcome. Ann Thorac Surg 2007;84:1935– 42. Mihaljevic T. Invited commentary. It would not be wise to completely abandon the use of homograft root replacement in particular in patients with prosthetic valve endocarditis and aortic root abscess. Ann Thorac Surg 2007;84:1942.
ADULT CARDIAC
Ann Thorac Surg 2012;93:480 – 8
Arminder Singh Jassar, MBBS, Joseph E. Bavaria, MD, Wilson Y. Szeto, MD, Patrick J. Moeller, BS, Jon Maniaci, Rita K. Milewski, MD, PhD, Joseph H. Gorman III, MD, Nimesh D. Desai, MD, PhD, Robert C. Gorman, MD, and Alberto Pochettino, MD Division of Cardiovascular Surgery, University of Pennsylvania Medical Center, Philadelphia, Pennsylvania
Background. Endocarditis affecting the aortic valve, with abscess formation and root destruction, remains a challenge to treat. Aortic root homografts have been advocated because of a perceived lower risk of infective complications than with other root replacement grafts. However, the theoretical advantage of homografts has not been re-evaluated in the modern era. This report is based on an examination of our results for all aortic root replacements in complex, active endocarditis affecting the aortic valve. Methods. From 2000 to 2010, 134 patients (70.9% male; mean age 58.3 ⴞ 14.8 years) at our institution underwent aortic root replacement for active endocarditis. Ninety of the patients (67.2%) had a previously implanted prosthetic aortic valve. Our findings for these patients included one or more of the following: abscess (n ⴝ 110, 82.1%), valve vegetation (n ⴝ 98, 73.1%), and pseudoaneurysm or rupture or both (n ⴝ 62, 46.3%). We retrospectively reviewed data for the patients from hospital records and the social security data base.
Results. A mechanical composite graft (MC) was used in 43 of the patients (32.1%), a non-homograft biologic valve conduit (BC) in 55 patients (41.0%), and a homograft (HG) valve in 36 patients (26.9%). There was no significant difference among the groups in the incidence of major complications or in-hospital mortality. During a mean follow-up of 32.1 ⴞ 29.4 months, the rates of readmission, reinfection, and reoperation were similar for the three groups. The mean 5-year survival in the study was 58 ⴞ 9% for the MC group, 62 ⴞ 7% for the BC group, and 58 ⴞ 9% for the HG group, respectively (p ⴝ 0.48). Conclusions. Aortic root replacement in the presence of complex active infection is associated with significant morbidity and mortality. We report that the rates of major complications and late mortality were similar among MC, BC, and HG groups in our study.
I
they pose for reoperation because of structural deterioration represent drawbacks to their use. This has led to the use of other prosthetic materials to replace infected aortic roots. A number of previous studies have examined the outcomes of xenografts and mechanical prosthetic implants for aortic valves with IE. Most of these studies have examined the role of valve choice in isolated aortic valve replacement (AVR) [6, 9 –11]. The role of prosthetic material in the setting of advanced, destructive endocarditis requiring complete aortic root reconstruction is less well studied. In this report we describe a large series of patients with active advanced IE of the aortic valve that necessitated complete reconstruction of the aortic root. The report includes an examination of short-term morbidity and mortality data for these patients, as well as of long-term rates of reinfection and reoperation stratified by the choice of graft utilized for aortic root reconstruction.
nfective endocarditis (IE) effects about 15,000 patients yearly in the United States [1]. Despite advances in antimicrobial therapies, outcomes in IE remain poor [2, 3]. Surgery is indicated in patients with heart failure or cardiogenic shock caused by valvular dysfunction. Surgery should also be undertaken in hemodynamically stable patients with aggressive infection (ie, aortic or periannular abscess, heart block, recurrent emboli despite appropriate antibiotic therapy), infections resistant to antibiotic therapy, and fungal endocarditis [4]. The optimal type of aortic prosthesis for patients with IE remains unclear, with cryopreserved human allograft (homograft) prostheses often recommended [5–7]. However, although homografts may confer a theoretical advantage in resisting infection through their lack of artificial materials [8], their limited availability and the need
Accepted for publication Sept 28, 2011. Presented at the Poster Session of the Forty-seventh Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 31–Feb 2, 2011. Address correspondence to Dr Pochettino, Division of Cardiovascular Surgery, University of Pennsylvania School of Medicine, Hospital of the University of Pennsylvania, 3400 Spruce St, 6 Silverstein Pavilion, Philadelphia, PA 19104-4283; e-mail: [email protected].
© 2012 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2012;93:480 – 8) © 2012 by The Society of Thoracic Surgeons
Patients and Methods Patients Between January 2000 and March 2010, 1,337 aortic root replacements were performed at the University of Penn0003-4975/$36.00 doi:10.1016/j.athoracsur.2011.09.074
sylvania Health System (valve-sparing aortic root reconstructions were excluded). Of the patients who received these root replacements, 192 had a diagnosis of aortic valve endocarditis. Endocarditis was classified according to criteria of The Society of Thoracic Surgeons (STS) as being active if the patient was being treated for endocarditis at the time of operation, or as treated if no antibiotic medication other than prophylactic medication was being given at the time of surgery. Fifty-eight patients with treated endocarditis were excluded; the data for 134 patients with active aortic valve endocarditis were analyzed for the purpose of this report.
Operative Procedure All 134 patients in the study underwent aortic root replacement through a sternotomy. A repeat sternotomy was required in 93 patients (90 of whom had a previous AVR or aortic root replacement). In patients with a high risk of aortic injury on surgical entry (eg, those with an aneurysm or pseudoaneurysm adherent to the posterior aspect of the sternum), peripheral cannulation was performed before sternotomy. In all other patients arterial cannulation was established through the ascending aorta, with venous cannulation performed via the right atrium, superior vena cava, and inferior vena cava as necessary. If reconstruction of the aortic arch was performed concomitantly, deep hypothermic circulatory arrest with adjunctive retrograde cerebral perfusion was used. Aggressive debridement was done to remove all infected, devitalized, or suspicious tissue. Any cardiac structural defects caused by infection or debridement (eg, defects in aorto-mitral continuity; defects in the left ventricle, left atrium, or mitral valve; or ventricular septal defects) were repaired with pericardial tissue (autologous or bovine) or a with a Dacron patch at the discretion of the surgeon. Complete replacement of the aortic root was done in all of the patients, with coronary ostia reimplanted as buttons. Composite grafts with a mechanical valve (MC) were used in 43 patients (32.1%), nonhomograft conduits with biologic valves (BC) were used in 55 patients (41.0%) (39 porcine full roots and 16 “biologic composite” grafts, consisting of a stented tissue valve sewn to a Dacron tube graft), and homografts (HG) were used in 36 patients (26.9%).
Postoperative Care Intravenous antibiotics were continued for at least 6 weeks postoperatively. The antibiotic regimen for each patient was based on microbiologic data, with a combination of broad-spectrum antibiotic agents used in the absence of positive cultures. In patients with fungal endocarditis or recurrent prosthetic valve endocarditis, longer periods of suppression with oral antibiotics were used. The antibiotic regimens did not depend on the type of root replacement done in a patient.
Statistical Analysis The University of Pennsylvania aortic surgery database, inpatient and outpatient charts, echocardiography re-
JASSAR ET AL AORTIC ROOT REPLACEMENT FOR ENDOCARDITIS
481
ports, and records of any readmissions or reoperations were reviewed as sources for the study data. Follow-up data were available for 84 (80.8 %) of the 104 patients who were discharged alive. The mean follow-up period in the study was 32.1 ⫾ 29.4 months (MC ⫽ 27.4 ⫾ 22.1 months [median, 24.1 months], BC ⫽ 31.7 ⫾ 28.5 months [median, 20.2 months], HG ⫽ 40.5 ⫾ 39.7 months [median, 27.2 months], p ⫽ 0.33). Only 5 patients (5.95%) were found to have definite recurrent or reinfective endocarditis (according to the modified Duke Criteria [12]) (n ⫽ 1, MC group; n ⫽ 2, BC group; n ⫽ 2, HG group). However, instances of possible endocarditis and some cases in which a diagnosis of endocarditis was rejected (eg, root pseudoaneurysm, or fever with mediastinal collection, or sepsis, and death due to cardiac arrest prior to completion of workup) were included, to encompass the total burden of infection. Similarly, for readmissions and reoperations, all occurrences related to patients’ cardiovascular status (infective or not) were indexed. Survival data were supplemented from online Social Security Death Index database [13]. Continuous variables were compared through analysis of variance (ANOVA) with post hoc analysis. Categorical variables were compared through Pearson’s 2 test and Fisher’s exact test. Adjustments were not made for multiple testings. A multivariate logistic regression model was used to determine predictors of in-hospital and late mortality. Input covariates for the logistic regression model for in-patient mortality and the Cox regression model for late mortality included patient age, sex, presence of diabetes mellitus, renal failure, abscess, positive cultures, and the type of prosthesis used. The odds ratio (OR) and 95% confidence interval (CI) were calculated for each risk factor. The percent rates of readmission, reinfection, reoperation, and survival were analyzed with the Kaplan–Meier method, and between-group comparisons were made with the log-rank test and presented as the mean ⫾ standard error.
Results Preoperative Patient Characteristics and Graft Selection Preoperative characteristics of the study patients are shown in Table 1. The types and numbers of native and prosthetic aortic valve grafts for the patients in the study are shown in Table 2. The presence of a bicuspid or tricuspid, native, or prosthetic aortic valve at the time of surgery did not influence graft selection. Incidences of abscess in the aortic root or surrounding regions, and of aortic valve vegetation, dehiscence, pseudoaneurysm, fistula, or rupture are shown in Table 3. Microbiologic data (preoperative blood cultures or intraoperative tissue cultures) are shown in Table 4. A causative organism was identified in 115 patients (89.1%). Cultures from 24 patients (18.6%) grew more than one organism. Intraoperative tissue-culture data were available for 117 patients (87.3%), with the causative organism cultured in 58 patients (49.6%).
ADULT CARDIAC
Ann Thorac Surg 2012;93:480 – 8
482
JASSAR ET AL AORTIC ROOT REPLACEMENT FOR ENDOCARDITIS
Ann Thorac Surg 2012;93:480 – 8
Table 1. Preoperative Characteristics of Patients Undergoing Aortic Root Replacement Patient Characteristics ADULT CARDIAC
Age (years) Male Dyslipidemia CVD CVA Chronic lung disease DM HTN Renal failure/HD Smoker Myocardial infarction Heart failure by NYHA Class Class I Class II Class III Class IV Ejection fraction ⱕ 30 % 31% – 40% 41% – 50% 51% – 60% ⬎60% Cardiogenic shock CPR Urgent/emergent a
All (n ⫽ 134)
Mechanical Composite Graft (n ⫽ 43)
Biologic Conduit (n ⫽ 55)
Homograft (n ⫽ 36)
58.3 ⫾ 14.8 95 66 46 37 27 29 91 24/9 57 19
51.8 ⫾ 12.7 38 15 16 13 4 9 26 3/0 18 7
62.7 ⫾ 15.1 35 34 21 17 16 10 39 8/2 22 6
59.4 ⫾ 14.3a,b 22a,b 17 9 7 7a 10 26 13/7b,c 17 6
22 15 49 30
7 7 19 4
2 2 16 18
3 6 14 8
12 13 29 47 20 11 3 96
3 4 13 17 3 3 0 30
6 5 8 19 11 4 2 36
3 4 8 11 6 4 1 30c
p ⬍ 0.05 mechanical conduit vs biologic graft.
CPR ⫽ cardiopulmonary resuscitation; hemodialysis; HTN ⫽ hypertension;
b
p ⬍ 0.05 mechanical conduit vs homograft.
c
p ⬍ 0.05 biologic graft vs homograft.
CVA ⫽ cerebrovascular accident; CVD ⫽ cerebrovascular disease; NYHA ⫽ New York Heart Association.
Utilization of MC and BC conduits became more frequent during the course of this study (Fig 1). The use of HG declined throughout the course of the study (n ⫽ 27, 42.2% from 2000 to 2005, vs n ⫽ 9, 12.9% from 2006 to 2010).
Operative Procedures Details of concomitant procedures and mean operative times are listed in Table 5. A hemiarch replacement was performed in 28 patients (13 with a previous AVR, 10 with previous root replacements, 5 with native bicuspid aortic
DM ⫽ diabetes mellitus;
HD ⫽
valves). Coronary artery bypass was required in 16 patients (7 because of involvement or destruction of coronary ostia by infection, 6 because of occlusive native coronary disease, 1 because of stenosis of a previous aortocoronary bypass, 1 because of native right coronary artery kinking, and 1 because of injury to a left internal mammary artery graft upon re-entry. Total cardiopulmonary bypass time and aortic crossclamp times were longer in the MC group than in the HG group (p ⫽ 0.027 and p ⫽ 0.01, respectively), probably
Table 2. Graft Selection in Replacement of Native and Prosthetic Aortic Valves Valve Type Native AV BAV Prosthetic AV Previous AVR Previous root replacement
All (n ⫽ 134)
Mechanical Composite Graft (n ⫽ 43)
Biologic Conduit (n ⫽ 55)
Homograft (n ⫽ 36)
44 18 90 69 21
12 7 31 20 11
19 9 36 31 5
13 2 23 18 5
There was no significant difference between groups by pairwise 2 comparison. AV ⫽ aortic valve;
AVR ⫽ aortic valve replacement;
BAV ⫽ biologic aortic valve.
Ann Thorac Surg 2012;93:480 – 8
483
JASSAR ET AL AORTIC ROOT REPLACEMENT FOR ENDOCARDITIS
Anatomic Findings
All (n ⫽ 134)
Mechanical Composite Graft (n ⫽ 43)
Biologic Conduit (n ⫽ 55)
Homograft (n ⫽ 36)
110 98 62 114
36 28 17 37
46 42 30 47
28 28 15 30
Abscess Vegetations Dehiscence/rupture/pseudoaneurysm/fistula Abscess or dehiscence/rupture/pseudoaneurysm/fistula
There was no significant difference between groups by pairwise 2 comparison. IE ⫽ infective endocarditis.
because of the higher incidence of concomitant procedures in the former group.
In-Hospital Complications and Mortality In-hospital mortality, postoperative length of stay, and the incidences of major complications are presented in Table 6. Overall perioperative mortality was 22.4%, and did not differ significantly by type of root replacement.
Readmission, Reinfection, Reoperation and Long-Term Survival Of the 104 discharged patients, long-term follow-up data were available for 84 (80.9%). Of these 84 patients, 18 (21.4%) were readmitted, of whom 8 (27.6%) were in the MC group (n ⫽ 5 for reinfection; n ⫽ 1 for pseudoaneurysm because of a dehisced coronary button; n ⫽ 1 for heart failure; and n ⫽ 1 for bilateral pleural effusions requiring chest-tube drainage); 5 (13.5%) were in the BC group (n ⫽ 3 for reinfection; n ⫽ 1 for hemorrhagic stroke; and n ⫽ 1 for pleural effusion requiring pleurodesis); and 5 (27.7%) were in the HG group (n ⫽ 5, for reinfection). Freedom from readmission was 76% ⫾ 8% at 1 year of follow-up and 60% ⫾ 13% at 5 years of follow-up for patients in the MC group; 88% ⫾ 6% at 1 year and 83% ⫾ 7% at 5 years of follow-up for patients in the BC group; and 63% ⫾ 12% at both 1 and 5 years of follow-up for patients in the HG group (p ⫽ 0.16) (Fig 2A). Overall, 14 patients (16.6 %) had a diagnosis of reinfection (n ⫽ 5,
MC group; n ⫽ 3, BC group; n ⫽ 6, HG group; 1 patient with infection in the HG group was treated as an outpatient). Freedom from reinfection was 84% ⫾ 7% at 1 year and 74% ⫾ 10% at 5 years for patients in the MC group; 94% ⫾ 4% at 1 year and 89% ⫾ 6% at 5 years for patients in the BC group; and 75% ⫾ 11% at 1 year and 64% ⫾ 14% at 5 years for patients in the HG group (p ⫽ 0.10) (Fig 2B). Six patients (7.1%) required cardiac-related reoperation. One patient in the MC group underwent drainage of a mediastinal collection and muscle flap coverage of substernal wound, and 1 patient underwent repeat sternotomy and repair of a dehisced coronary button 1 month postoperatively. Two patients in the BC group underwent repeat aortic root replacement, the first with a homograft at 3 months after the first root replacement operation, because of a root abscess, valve vegetation, and stroke; and the second with an MC conduit at 2.1 years after initial root replacement, because of aortic stenosis and endocarditis. Two patients in the HG group underwent reoperation. The first of these patients had a repeat HG implantation at 7 months postoperatively, because of infection and dehiscence of the first homograft; and the second underwent mitral valve repair for regurgitation at 11 months after initial placement of an HG root. Freedom from reoperation at 1 and 5 years was 96% ⫾ 4% and 89% ⫾ 8% respectively in the MC group; 97% ⫾ 3% and 90% ⫾ 7% respectively in the BC group; and 86% ⫾ 9% at both 1 and 5 years in the HG group (p ⫽ 0.67) (Fig 3A).
Table 4. Organisms Cultured From Aortic Roots Affected by Infectious Endocarditis Preoperative Blood Cultures or Operative Tissue Cultures Culture-negative Fungus MRSA MSSA Coagulase-negative Staphylococcus Enterococcus Streptococcus Gram-negative rods Others Positive operative tissue culturesb
All (n ⫽ 129)
Mechanical Composite Graft (n ⫽ 42)
Biologic Conduit (n ⫽ 55)
Homograft (n ⫽ 33)
14 12 16 18 21 19 30 10 15 58 (den ⫽ 117)
4 6 4 5 8 5 13 5 5 22 (den ⫽ 39)
7 4 9 4 9 8 14 2 8 23 (den ⫽ 50)
3 2 3 9a 5 6 3 3 2 13 (den ⫽ 28)
p ⬍ 0.05 biologic conduit vs homograft. Other comparisons did not show significant difference by pairwise 2 comparison. 117 of 134 patients. Denominators (den) are listed for each group.
a
MRSA ⫽ methicillin-resistant Staphylococcus aureus;
MSSA ⫽ methicillin-sensitive Staphylococcus aureus.
b
Data were available for
ADULT CARDIAC
Table 3. Anatomic Findings in Cases of IE Affecting the Aortic Valve
484
JASSAR ET AL AORTIC ROOT REPLACEMENT FOR ENDOCARDITIS
Ann Thorac Surg 2012;93:480 – 8
Fig 1. Yearly trend in prosthetic graft selection.
ADULT CARDIAC
The overall cumulative survival (including in-hospital mortality) of the patients in the study at 1 year and 5 years was 67.9% and 58.7%, respectively. The 1-year and 5-year rates of survival were 67% ⫾ 7% and 58% ⫾ 9% respectively for patients in the MC group; 65% ⫾ 7% and 62% ⫾ 7% respectively for patients in the BC group; and 61% ⫾ 8% and 58% ⫾ 9% respectively for patients in the HG group (p ⫽ 0.48) (Fig 3B).
Predictors of In-Hospital and Late Mortality The results of risk-factor analysis for predictors of inhospital mortality are shown in Table 7. Abscess (OR ⫽ 9.15; 95%CI, 1.05 to 80.18) and a history of diabetes (OR ⫽ 4.65; 95%CI, 1.64 to 13.17) significantly increased the risk of in-hospital mortality. Significant multivariate predictors of late mortality included diabetes (OR ⫽ 2.25; 95%CI, 1.27 to 3.99) and renal failure (OR ⫽ 2.17; 95%CI, 1.17 to 4.04). The type of prosthesis used was not predictive of either in-hospital or late mortality.
Comment Endocarditis often requires surgical intervention, which needs to be safe and effective, with a minimal risk of reinfection [2, 14]. Lytle and Sabik reported excellent results after homograft replacement for the treatment of prosthetic aortic valve endocarditis, with a mortality ranging from 3% to 13% [5, 6, 8]. Most other reports show much higher mortality rates. Several factors can explain the higher mortality rate in our series. One of these is that the patients in our study had extensive disease, as evidenced by abscess cavities in most of them (82.1%), and pseudoaneurysm, rupture, dehiscence, or a fistula in a significant fraction (46.2%). Seventy one percent of our patients underwent urgent or emergent operations, and about two thirds of our patients had prosthetic valve infection, which has been shown to worsen outcomes [9]. Our results are similar to results recently reported by Musci and colleagues [15], who presented data for aortic root replacement with homograft roots in a patient pop-
Table 5. Concomitant Procedures Done With Aortic Root Replacement
Procedures Any additional procedures Hemiarch CABG Mitral valve replacement Mitral valve repair Tricuspid valve repair/replacement Perfusion time (minutes) Aortic clamp time (minutes) Total circulatory arrest time (minutes) p ⬍ 0.05 mechanical vs homograft. comparison. a
CABG ⫽ coronary artery bypass grafting.
b
All (n ⫽ 134)
Mechanical Composite Graft (n ⫽ 43)
Biologic Conduit (n ⫽ 55)
Homograft (n ⫽ 36)
90 28 16 20 16 11 293 ⫾ 99.7 227.1 ⫾ 75.3 29.0 ⫾ 15.5 (n ⫽ 29)
33 13 4 7 9 1 308.3 ⫾ 98.2 244.1 ⫾ 78.7 31.6 ⫾ 19.5 (n ⫽ 14)
38 13 5 10 4 7 300.8 ⫾ 109.5 230.3 ⫾ 76.5 27.5 ⫾ 10.8 (n ⫽ 13)
19a,b 2a,b 7 3 3 3 262.6 ⫾ 80.0a 201.9 ⫾ 63.6a 21.0 ⫾ 9.9 (n ⫽ 2)
p ⬍ 0.05 biologic vs homograft. Other comparisons did not show significant difference by pairwise 2
Ann Thorac Surg 2012;93:480 – 8
JASSAR ET AL AORTIC ROOT REPLACEMENT FOR ENDOCARDITIS
485
Complications
All (n ⫽ 134)
Mechanical Composite Graft (n ⫽ 43)
Biologic Conduit (n ⫽ 55)
Homograft (n ⫽ 36)
Freedom from all major complications In-hospital mortality Length of stay Septicemia Deep sternal infection Permanent stroke Reoperation for bleeding/tamponade Renal failure/hemodialysis Cardiac arrest Heart block Multisystem organ failure Prolonged ventilation (⬎24 hours)
36 30 18.2 ⫾ 16.6 18 3 5 12 26/12 10 27 16 51
15 8 19.7 ⫾ 21.3 9 1 1 5 6/4 4 9 8 17
12 13 15.7 ⫾ 13.8 7 1 1 4 14/6 2 15 5 23
9 9 20.1 ⫾ 13.4 2 1 3 3 6/2 4 3a 3 11
a
p ⬍ 0.05 biologic conduit vs homograft. Other comparisons did not show significant difference by pairwise 2 comparison.
ulation similar to ours (having an 86% incidence of aortic root abscess). They experienced a 30-day mortality rate of 21.2%, which increased significantly for operations done in the setting of prosthetic valve endocarditis (25.4% vs 16.1% for native valve endocarditis). Leyh and associates [16] compared outcomes after aortic root replacement for prosthetic valve endocarditis with use of either aortic allografts or mechanical composite grafts. In their small series (n ⫽ 29), they reported an 18.5% overall rate of inhospital mortality (18.7% for patients given allografts and 15.4% for those given composite grafts). As a frame of reference, our in-hospital mortality for aortic root replacement in lower risk patients is approximately 2.5% [17], emphasizing the significant effect of infection on outcome. In our study, the HG group failed to show any difference from the MC or BC group in the rates of in-hospital complications, mortality, or long term reinfection. Hagl and colleagues [18] have previously reported an inhospital mortality of 11% in 28 patients who underwent aortic root replacement (25 of whom received mechanical composite grafts and 3 of whom received homografts) for prosthetic valve endocarditis. Abscess was found in 57% of these patients, and recurrent endocarditis occurred in 4% over a median follow-up of 44.5 months. Hagl and colleagues used historical data from the literature to conclude that the use of artificial prosthetic material
yields results comparable to those with homografts. David and coworkers [19] reported outcomes after surgery for paravalvular abscess (aortic, mitral, tricuspid, or pulmonary) in 135 patients. The overall operative mortality among these patients was 15.5%, with no statistically significant difference in the mortality of patients who received a mechanical valve (13.6%), a bioprosthetic valve (16.3%), or a homograft (21.4%). The 10-year survival in this patient group was 57% ⫾ 5% and the rate of recurrent endocarditis was 14%, and no predictors for recurrent endocarditis could be identified. David and coworkers concluded that the recognition and extirpation of all infected aortic valve tissue was critical, but that a homograft was preferred because of ease of handling and the possibility of using homograft mitral leaflet tissue to patch defects created by abscess resection. Similar results were reported by Moon and associates [9] in a retrospective review of data for 306 patients with endocarditis (211 with active disease, 95 with treated disease) who underwent replacement of the aortic or mitral valve with a bioprosthesis (n ⫽ 221), mechanical valve (n ⫽ 65), or homograft (n ⫽ 20) in the setting of either a native (n ⫽ 209) or a prosthetic valve (n ⫽ 97) infection. They reported an overall mortality of 18% ⫾ 2% for these patients. Even though 10-year survival was better for patients with native valve endocarditis than for those Fig 2. Kaplan–Meier analysis demonstrating (A) freedom from readmission and (B) freedom from reinfection after mechanical valve conduit (M), biologic valve conduit (B), or homograft (H) replacement of infected aortic root.
ADULT CARDIAC
Table 6. Postoperative Complications
486
JASSAR ET AL AORTIC ROOT REPLACEMENT FOR ENDOCARDITIS
Ann Thorac Surg 2012;93:480 – 8
Fig 3. Kaplan–Meier analysis demonstrating (A) freedom from reoperation and (B) long term survival after mechanical valve conduit (M), biologic valve conduit (B), or homograft (H) replacement of infected aortic root. ADULT CARDIAC
similar rates of abscess formation. Our study focused on patients with extensive native or prosthetic valve endocarditis, which could not be treated with valve replacement alone and requires replacement of the entire aortic root. A significant number of patients in our study underwent concomitant surgical procedures, mainly hemiarch replacement, and mitral valve interventions. Although the differences were not statistically significant, the number of these procedures were higher in the MC group than in the other two groups in the study. Without available donor mitral tissue, as in a homograft, the degree of distortion in the anterior leaflet of the mitral valve after the implantation of a rigid mechanically valved conduit may require further reparative techniques to ensure competency of the mitral valve. Our study has limitations inherent in any retrospective nonrandomized study. The evolution of surgical techniques and of the conduits available for implantation may have affected the outcomes of the study, with this effect potentially more prominent in the HG group. Graft selection was based on the preference and judgment of the surgeon. A greater fraction of patients in the MC group than in the BC or HG groups were younger males; patients in the MC group had a lower incidence of preoperative renal dysfunction. We have tried to adjust for some of these variables using the multivariate logistic and Cox regression
with prosthetic valve endocarditis (54% vs 41%) in their patient population, there was no difference in relation to the type of valve implanted. A recent study by Klieverik and colleagues [20] compared the outcome after AVR for active endocarditis in 138 patients, of whom 106 received allografts and 32 received mechanical prostheses). There was no survival benefit for patients who had an allograft over those who had a mechanical valve replacement (59% vs 66%, respectively). The rate of recurrent infection did not differ for the two groups. The 15-year freedom from reoperation was significantly higher in the mechanical valve group (93% vs 76%). However, in their study, only 18% of patients in the mechanical valve group had root abscesses, as compared with 38% in the allograft group, with homografts having been selected for more advanced disease. Similarly, Avierinos and associates [21], in a study of 127 patients with active endocarditis (of whom 50% had abscesses) who had either a homograft (n ⫽ 54) or a conventional prosthesis (n ⫽ 54, a bioprosthesis; n ⫽ 23, a mechanical prosthesis), found no significant difference between the two groups in operative mortality or event-free 5-year survival. However, the majority of patients with annular abscesses (65%) in their study were treated with homografts, which suggests a bias towards using homografts in patients with more severe disease [22]. In our study, the HG, MC, and BC groups had
Table 7. Multivariate Risk Factor Analysis for Predictors of In-Hospital and Late Mortality In-Hospital Mortality
Late Mortality
95% CI for OR Effect Group Age Sex Diabetes Renal failure Abscess Culture-positive Mechanical composite graft Biologic conduit
OR
Lower Limit
Upper Limit
OR
Lower Limit
Upper Limit
1.005 2.515 4.912 2.761 9.153 0.995 1.331 1.434
0.970 0.883 1.747 0.893 1.045 0.219 0.351 0.4444
1.041 7.166 13.811 8.535 80.184 4.519 5.037 4.633
1.014 1.542 2.253 2.177 1.291 1.293 1.142 0.888
0.993 0.870 1.270 1.170 0.610 0.552 0.556 0.467
1.035 2.731 3.998 4.048 2.731 3.029 2.343 1.690
For prosthetic type, homograft was considered the reference. CI ⫽ confidence interval;
95% CI for OR
OR ⫽ odds ratio.
analysis. It is important, however, to recognize that all three of the groups in the study had infection of similar severity (similar abscess rates). The frequencies of concomitant procedures were significantly lower for patients in the HG group, which resulted in shorter operating times for this group. An incomplete long-term follow-up (80.8%) may have failed to capture patients who sought care at other facilities, and the rates of reinfection and reoperation in our study may have been underestimated. The mortality data for the study are complete because the social security death index was used to determine this. Our approach to graft selection has evolved over time, as shown in Figure 1. Homografts were preferentially used early in our experience owing to the early reports of a greater freedom from reinfection with their use, and before 2000, homografts were used exclusively for aortic roots affected by the aggressive infections described in this study. Over time we have introduced alternate conduits for treating extensive aortic valve endocarditis requiring root replacement. At our center, the choice of conduit is now mainly governed by the age of the patient and the presence of relative or absolute contraindications to warfarin. Additionally, MC conduits have become the most frequent choice for root replacement in patients who have had a previous prosthetic root (11 of the 21 such patients in our study [52%]), to avoid a third root replacement related to structural deterioration of a BC or the HG replacement. In our practice, HG may continue to play a role in patients with contraindications to warfarin or those who require extensive mitral valve augmentation as part of their treatment for severe root abscess. In conclusion, we found that active infective aortic valve endocarditis with root destruction remains a challenging condition, with significant morbidity and mortality. In our experience, the type of aortic root prosthesis used for patients with roots affected by IE did not affect the outcomes of root replacement. As other authors have stated before us, we believe that the complete eradication of IE through aggressive debridement, and not the choice of conduit used for root replacement, is the critical factor in determining the ultimate outcome of such replacement. The type of prosthetic root to be implanted should be selected on the basis of the complete clinical picture for a given patient, including the technical requirements for reconstruction, the patient’s age and comorbidities, an assessment of compliance to medication, and follow-up, and the availability and hemodynamic performance of a conduit of the appropriate size. The authors acknowledge the help and support provided by Mary E. Mckenna, BSN, toward data collection for this study.
JASSAR ET AL AORTIC ROOT REPLACEMENT FOR ENDOCARDITIS
4.
5.
6. 7.
8.
9. 10. 11. 12. 13. 14. 15.
16.
17.
18.
19. 20.
References 1. Bashore TM, Cabell C, Fowler V, Jr. Update on infective endocarditis. Curr Probl Cardiol 2006;31:274 –352. 2. Gaca JG, Sheng S, Daneshmand MA, et al. Outcomes for endocarditis surgery in North America: a simplified risk scoring system. J Thorac Cardiovasc Surg 2011;141:98 –106 e2. 3. David TE, Gavra G, Feindel CM, Regesta T, Armstrong S, Maganti MD. Surgical treatment of active infective endocar-
21. 22.
487
ditis: a continued challenge. J Thorac Cardiovasc Surg 2007;133:144 –9. Bonow RO, Carabello BA, Chatterjee K, et al. 2008 focused update incorporated into the ACC/AHA 2006 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 (Writing Committee to revise the 1998 guidelines for the management of patients with valvular heart disease). J Am Coll Cardiol 2008;52:e1–142. Lytle BW, Sabik JF, Blackstone EH, Svensson LG, Pettersson GB, Cosgrove DM III. Reoperative cryopreserved root and ascending aorta replacement for acute aortic prosthetic valve endocarditis. Ann Thorac Surg 2002;74:S1754 –7; discussion: S1792–9. Lytle BW, Priest BP, Taylor PC, et al. Surgical treatment of prosthetic valve endocarditis. J Thorac Cardiovasc Surg 1996;111:198 –210. Perrotta S, Aljassim O, Jeppsson A, Bech-Hanssen O, Svensson G. Survival and quality of life after aortic root replacement with homografts in acute endocarditis. Ann Thorac Surg 2010;90:1862–7. Sabik JF, Lytle BW, Blackstone EH, Marullo AG, Pettersson GB, Cosgrove DM. Aortic root replacement with cryopreserved allograft for prosthetic valve endocarditis. Ann Thorac Surg 2002;74:650 –9. Moon MR, Miller DC, Moore KA, et al. Treatment of endocarditis with valve replacement: the question of tissue versus mechanical prosthesis. Ann Thorac Surg 2001;71:1164 –71. Sweeney MS, Reul GJ, Jr., Cooley DA, et al. Comparison of bioprosthetic and mechanical valve replacement for active endocarditis. J Thorac Cardiovasc Surg 1985;90:676 – 80. Newton S, Hunter S. What type of valve replacement should be used in patients with endocarditis? Interact Cardiovasc Thorac Surg 2010;11:784 – 8. Li SL, Sexton DJ, Mick N, et al. Proposed modifications to the Duke criteria for the diagnosis of infective endocarditis. Clin Infect Dis 2000;30:633– 8. Social Security Death Index database. Available at: http:// ssdi.rootsweb.ancestry.com. Accessed January 9, 2011. Leontyev S, Borger MA, Modi P, et al. Redo aortic valve surgery: Influence of prosthetic valve endocarditis on outcomes. J Thorac Cardiovasc Surg 2011;142:99 –105. Musci M, Weng Y, Hubler M, et al. Homograft aortic root replacement in native or prosthetic active infective endocarditis: twenty-year single-center experience. J Thorac Cardiovasc Surg 2010;139:665–73. Leyh RG, Knobloch K, Hagl C, et al. Replacement of the aortic root for acute prosthetic valve endocarditis: prosthetic composite versus aortic allograft root replacement. J Thorac Cardiovasc Surg 2004;127:1416 –20. Desai ND, McCarthy F, Moser W, et al. Durability of porcine bioroots in younger patients with aortic root pathology: a propensity-matched comparison with composite mechanical roots. Ann Thorac Surg. 2011 Aug 11. [Epub ahead of print] Hagl C, Galla JD, Lansman SL, et al. Replacing the ascending aorta and aortic valve for acute prosthetic valve endocarditis: is using prosthetic material contraindicated? Ann Thorac Surg 2002;74:S1781–5; discussion: S1792–9. David TE, Regesta T, Gavra G, Armstrong S, Maganti MD. Surgical treatment of paravalvular abscess: long-term results. Eur J Cardiothorac Surg 2007;31:43– 8. Klieverik LM, Yacoub MH, Edwards S, et al. Surgical treatment of active native aortic valve endocarditis with allografts and mechanical prostheses. Ann Thorac Surg 2009;88:1814 –21. Avierinos JF, Thuny F, Chalvignac V, et al. Surgical treatment of active aortic endocarditis: homografts are not the cornerstone of outcome. Ann Thorac Surg 2007;84:1935– 42. Mihaljevic T. Invited commentary. It would not be wise to completely abandon the use of homograft root replacement in particular in patients with prosthetic valve endocarditis and aortic root abscess. Ann Thorac Surg 2007;84:1942.
ADULT CARDIAC
Ann Thorac Surg 2012;93:480 – 8