Systematic Review Systematic review and meta-analysis of surgical outcomes comparing mechanical valve replacement and bioprosthetic valve replacement in infective endocarditis Campbell D. Flynn1, Neil P. Curran1, Stephanie Chan1, Isabel Zegri-Reiriz2, Manel Tauron3, David H. Tian4, Gosta B. Pettersson5, Joseph S. Coselli6,7, Martin Misfeld8, Manuel J. Antunes9, Carlos A. Mestres10,11 Eduard Quintana12 1 Department of Cardiothoracic Surgery, St George Hospital, Sydney, Australia; 2Heart Failure and Heart Transplant Unit, Cardiology Department, 3 Cardiac Surgery Department, Hospital de la Santa Creu i Santa Pau, Barcelona, Spain; 4Collaborative Research Group, Macquarie University, Sydney, Australia; 5Department of Thoracic and Cardiovascular Surgery, Cleveland Clinic, Cleveland, OH, USA; 6Division of Cardiothoracic Surgery, Baylor College of Medicine, Houston, TX, USA; 7Texas Heart Institute, Houston, TX, USA; 8University Department for Cardiac Surgery, Leipzig Heart Center, Leipzig, Germany; 9Clinic of Cardiothoracic Surgery, University of Coimbra Medical School, Coimbra, Portugal; 10 Department of Cardiac Surgery, University Hospital Zürich, Zürich, Switzerland; 11Department of Cardiothoracic Surgery, The University of the Free State, Bloemfontein, South Africa; 12Department of Cardiovascular Surgery, Hospital Clinic, University of Barcelona, Barcelona, Spain Correspondence to: Dr. Campbell D. Flynn. Department of Cardiothoracic Surgery, St George Hospital, Sydney, Australia. Email: campbellf@doctors.org.uk. Background: Infective endocarditis (IE) is an infection involving either native or prosthetic heart valves, the endocardial surface of the heart or any implanted intracardiac devices. IE is a rare condition affecting 3–15 patients per 100,000 population. In-hospital mortality rates in patients with IE remain high at around 20% despite treatment advances. There is no consensus recommendation favoring either bioprosthetic valve or mechanical valve implantation in the setting of IE; patient age, co-morbidities and preferences should be considered selecting the replacement prosthesis. Methods: A systematic review and meta-analysis of studies reporting the outcomes of patients undergoing bioprosthetic or mechanical valve replacement for infective endocarditis with data extracted for overall survival, valve reinfection rates and valve reoperation. Results: Eleven relevant studies were identified, with 2,336 patients receiving a mechanical valve replacement and 2,057 patients receiving a bioprosthetic valve replacement. There was no significant difference for overall survival between patients treated with mechanical valves and those treated with bioprosthetic valves [hazard ratio (HR) 0.94, 95% confidence interval (CI): 0.73–1.21, P=0.62]. There was no significant difference in reoperation rates between patients treated with a bioprosthetic valve and those treated with a mechanical valve (HR 0.82, 95% CI: 0.34–1.98, P=0.66) and there was no significant difference in the rate of valve reinfection rates (HR 0.95, 95% CI: 0.48–1.89, P=0.89). Conclusions: The presence of infective endocarditis alone should not influence the decision of which type of valve prosthesis that should be implanted. This decision should be based on patient age, co-morbidities and preferences. Keywords: Valve replacement; tissue valve; bioprosthesis; bioprosthetic; mechanical; infective endocarditis Submitted Sep 02, 2019. Accepted for publication Sep 27, 2019. doi: 10.21037/acs.2019.10.03 View this article at: http://dx.doi.org/10.21037/acs.2019.10.03 © Annals of Cardiothoracic Surgery. All rights reserved. Ann Cardiothorac Surg 2019;8(6):587-599 | http://dx.doi.org/10.21037/acs.2019.10.03 588 Flynn et al. Meta-analysis: mechanical valve versus bioprosthetic valve replacement in IE Introduction comparing mechanical valve replacement and bioprosthetic valve replacement in infective endocarditis. Infective endocarditis (IE) is an infection involving either native or prosthetic heart valves, the endocardial surface of the heart or any implanted intracardiac devices (1). The Methods intact endocardium is an effective barrier to infection, with Literature search strategy the risk of endocarditis rising for individuals with congenital intracardiac defects, acquired degenerative valvular lesions This meta-analysis was performed in accordance with and implanted intra-cardiac prostheses (2); however, it PRISMA recommendations and guidance. The search remains an important condition with significant associated strategy was employed to search electronic databases morbidity and mortality. IE is a rare condition affecting EMBASE, Ovid Medline, the entire Cochrane Central 3–15 patients per 100,000 population (3-5). Infective Register of Controlled Trails (CCRCT), Cochrane endocarditis is commonly diagnosed according to the Database of Systematic reviews (CDSR), the Database modified Duke criteria (6) which determines the likelihood of Abstracts of Reviews of Effects (DARE) and the ACP of a patient having IE based on organisms detected on journal club from their inception to 03 March 2019. The blood culture, echocardiographic images and clinical search strategy included search terms for “(tissue valve OR examination. In patients with prosthetic valve replacement, bioprosthesis OR biological valve OR bioprosthetic valve the risk of developing prosthetic endocarditis is 2.2–3.7% OR xenograft OR mechanical valve) AND endocarditis”. (7-9). Aggressive surgical treatment is often required with The references of previous systematic reviews were assessed surgery indicated in up to half of patients with native valve to ensure no additional publications were missed. endocarditis (10,11). In-hospital mortality rates in patients with IE remain high at around 20% despite treatment Selection criteria advances (11). Indications for surgical intervention for left-sided Eligibility for inclusion in this systematic review and meta- IE are relatively well defined with urgent intervention analysis included papers that assessed the outcomes of suggested for patients without adequate control of sepsis patients undergoing conventional valve replacement with despite appropriate antibiotics, for patients with evidence a bioprosthetic or mechanical valve prosthesis undertaken of intra-cardiac abscess formation or other paravalvular in adult patients for the treatment of infective endocarditis. extension, for patients with heart failure secondary to In order to ensure sufficient center experience, papers were valve destruction or fistula formation, and to prevent only included if more than 15 cases were reported in each catastrophic embolic phenomena in the presence of arm. English language papers with at least 1-year follow- large mobile vegetations (12,13). Where possible, valve up were considered for assessment. Studies were excluded if repair is preferred over replacement, particularly for there was inadequate data regarding the outcomes and for right-sided and mitral valve IE. There is no consensus patients treated with homograft as a bioprosthesis. If centers recommendation for either bioprosthetic valve or reported outcomes of overlapping patient series, then mechanical valve in the setting of IE and patient age, the most contemporary series was analyzed. Conference co-morbidities and preferences should be considered in abstracts, case reports, editorials, expert opinion, reviews making the decision for valve selection (12,14). and expert opinion were excluded. In the general population, the European Society of Cardiology suggest treatment with mechanical valves Data extraction under the age of 60 for a valve in the aortic position or under the age 65 for a valve in the mitral position (15). For the assessed papers, data was extracted from the The American Heart Association guidelines recommend reviewed text, tables and figures. Data was extracted a mechanical valve prosthesis in patients under the age of independently by two of the authors (CD Flynn, NP 50 and a bioprosthetic valve prosthesis over the age of 70. Curran) and any discrepancies were reviewed and discussed In between these ages, patient factors and preferences will until consensus was reached. The recorded parameters determine prosthesis type (16). This meta-analysis was were: number of cases in series, procedure undertaken, therefore undertaken to evaluate the surgical outcomes urgency of procedure, average age, average follow-up, early © Annals of Cardiothoracic Surgery. All rights reserved. Ann Cardiothorac Surg 2019;8(6):587-599 | http://dx.doi.org/10.21037/acs.2019.10.03 Annals of cardiothoracic surgery, Vol 8, No 6 November 2019 589 death, late death, valve reinfection, valve reoperation. CI: 6.6–9.0, P<0.001). Ten studies reported overall survival data (19,21-29), drawing outcomes data from 3,884 patients (1,792 patients Statistical analysis receiving bioprosthetic valves and 2,092 patients receiving Meta-analysis of incidence rates of post-operative mechanical valves). Meta-analysis of derived hazard ratio complications including valve reinfection, reoperation (HR) demonstrated no significant difference for overall and overall survival. Hazard ratios were calculated from survival between patients treated with mechanical valves source Kaplan-Meier curves according to Tierney et al. and those treated with bioprosthetic valves (HR 0.94, 95% methods (17) with mechanical valve being the reference CI: 0.73–1.21; P=0.62) (Figure 1). There was significant group. Incidence data was assessed using Comprehensive heterogeneity Cochrane Q-statistic P value <0.001, I2 =75%. Meta-analysis v3.3 (Biostat, Englewood, USA). Comparative Post-hoc subgroup analysis performed to explore sources outcomes were assessed using Review Manager v5.3 of heterogeneity demonstrated that analyzing outcomes (Cochrane Collaboration, Copenhagen, Denmark). Due to of left sided infective endocarditis (21-23,25-29) showed the varied patient populations, a random effects model was no significant difference in overall survival, regardless of chosen for analysis. Data heterogeneity was assessed with valve type, HR 0.97 (95% CI: 0.67–1.41, P=0.89), there was the Cochrane Q statistic with P value <0.05 being significant still significant heterogeneity in data, Cochrane q-statistic and the I 2 test statistic an I 2 value greater than 50% P=0.005, I2 =68% (Figure 1). There was a possible benefit denoting significant heterogeneity. Publication bias was for overall survival with bioprosthetic valve replacement in assessed by generation of funnel plot and assessment using the treatment of prosthetic valve endocarditis (21,25,27) Egger’s test produced with Comprehensive Meta-analysis. (HR 1.37, 95% CI: 1.01–1.85, P=0.04) (Figure 1). Sensitivity analysis was performed using a “leave-one- Composite survival curves demonstrated overall survival for out” analysis. Individual patient survival data, reinfection bioprosthetic valves of 71.8%, 57.3%, 49.2%, and 37.5% and reoperation data were reconstructed using an iterative at 1-, 3-, 5-, and 10-year, respectively, and 73.2%, 56.8%, algorithm that was applied to digitized source Kaplan- 49.3%, and 34.0% at 1-, 3-, 5-, and 10-year, respectively, Meier curves and subsequently aggregated and graphed (18) for mechanical valves (Figure 2). Leave-one-out sensitivity using R (R Foundation for Statistical Computing, Vienna, analysis did not demonstrate any significant impact on Austria). overall outcomes with any one study being removed. There was no evidence of publication bias on visual inspection of a funnel plot of the standard error by log hazard ration for Results overall survival (Figure S3), nor by Egger’s test (P=0.358). The search strategy revealed 4,053 citations for review after Four studies reported data on reoperation on valve duplications were removed with an additional two studies prosthesis (22,23,25,29), drawing outcomes data from 531 added for full text review on reviewing reference lists. After patients (302 patients receiving bioprosthetic vales and full review, 11 papers met the pre-determined inclusion 229 patients receiving mechanical valves). Meta-analysis criteria (19-29) (Figure S1). The included publications had of derived HR demonstrated no significant difference a total analyzed patient population of 4,393 patients with between reoperation rates between patients treated with 2,336 patients receiving a mechanical valve replacement and a bioprosthetic valve and those treated with a mechanical 2,057 patients receiving a bioprosthetic valve replacement. valve (HR 0.82, 95% CI: 0.34–1.98, P=0.66) There was The overall study characteristics are detailed in Table 1. moderate heterogeneity with the Cochrane q statistic An assessment of bias was made for all studies which P=0.12, I 2 =49% (Figure 3). Composite survival curves demonstrated significant risks of bias in all included studies, demonstrated freedom from reoperation for bioprosthetic largely based on study design (Figure S2). valves of 90.1%, 80.7%, 79.2%, and 56.9% at 1-, 3-, 5-, and The mean age of patients receiving bioprosthetic valves 10-year, respectively, and 94.1%, 91.3%, 86.0%, and 75.6% calculated from six studies reporting mean age of patients at 1-, 3-, 5-, and 10-year, respectively, for mechanical (19,20,22,23,28,29) was 59.2 years old (95% CI: 58.2–60.1). valves (Figure 4). Leave-one-out sensitivity analysis did not The mean age of patients receiving mechanical valves demonstrate any significant difference in overall effect on was 52.1 years old (95% CI: 51.4–52.9). Overall, patients outcomes. There was no evidence of publication bias on receiving a mechanical valve were younger by 7.8 years (95% visual inspection of a funnel plot of the standard error by © Annals of Cardiothoracic Surgery. All rights reserved. Ann Cardiothorac Surg 2019;8(6):587-599 | http://dx.doi.org/10.21037/acs.2019.10.03 590 Flynn et al. Meta-analysis: mechanical valve versus bioprosthetic valve replacement in IE Table 1 Study characteristics Recruitment Location of Number of Follow up Patient age Study Study design Patient population years institution patients (n) (median) [range] (years) [range] Reul 1979–1984 Houston, Texas Single-centre NVE and PVE Total =185 BV =22 months BV =49 [8–77] 1989 (USA) retrospective [6–68 months] BV =88 MV =18 months MV =46.6 [6–70 months] [11–78] MV =97 Lytle 1975–1992 Cleveland, Single-centre PVE Total =128 – – 1996 Ohio (USA) retrospective BV =52 (registry data) Mv =76 Moon 1964–1995 Stanford, Single-centre Left-sided valve IE Total =286 Overall =15.3 Overall =49 2001 California retrospective (NVE and PVE) years [0–29 BV =221 (USA) years] MV =65 Fedoruk 1975–2000 British Multi-centre NVE Total =358 Overall Overall 2009 Columbia, retrospective =5.5±5.5 years =48.8±15.9 Vancouver (registry data) [18–88] (Canada) BV =189 BV =6.1±6.1 BV =51.6±17.3 years MV =169 MV =4.9±4.5 MV =45.6±13.5 years Musci 1986–2008 Berlin Single-centre PVE Total =210 – – 2010 (Germany) retrospective BV =165 Mv =45 Nguyen 1998–2000 France Multicentre Aortic valve IE Total =140 – Overall 2010 prospective (NVE and PVE) =57.9±13.0 BV =31 BV =63.2±13.6 MV =109 MV =57.3±11.6 Jassar 2000–2010 Philadelphia, Single-centre ARR for IE Total =98 BV =31.7±28.5 BV =62.7±15.1 2012 Pennsylvania retrospective months (USA) BV =43 MV =27.4±22.1 MV =51.8±12.7 months MV =55 Leither 2004–2007 USA Multi-centre Renal-replacement Total =1267 – – 2013 retrospective therapy patients BV =561 (registry data)1 with NVE affecting aortic or mitral valve MV =706 Greason 1974–2009 Rochester, Single-centre PVE affecting the Total =36 Overall =5.7 Overall =68 2014 Minnesota retrospective MV years [0.6–24 [15.8–82.3] BV =18 (USA) (registry data) years] MV =18 Delahaye 2000–2006 Global Multicentre Patients with IE Total =1,467 – Overall 2015 prospective undergoing valve =56.6±15.7 surgery in the acute MV =917 BV =61.6±15.2 phase MV =917 MV =53.6±15.2 Kim 2016 2002–2014 Boston, Multicentre Adult patients with Total =218 Overall =29.4 BV =59.8±14.6 Massachusetts prospective AVE months [4.7– BV =139 MV =47.2±14.5 (USA) 72.6 months] MV =79 NVE, native valve endocarditis; PVE, prosthetic valve endocarditis; IE, infective endocarditis; BV, bioprosthetic valve; MV, mechanical valve. © Annals of Cardiothoracic Surgery. All rights reserved. Ann Cardiothorac Surg 2019;8(6):587-599 | http://dx.doi.org/10.21037/acs.2019.10.03 Annals of cardiothoracic surgery, Vol 8, No 6 November 2019 591 Figure 1 Forest plot detailing survival outcomes. 100 75 Overall survival (%) Bioprosthetic valve 50 Mechanical valve 25 Numbers at risk: 1,792 Bio 764 591 460 369 247 201 170 145 122 104 0 2,092 Mech 704 553 436 353 196 82 60 55 36 33 0 1 2 3 4 5 6 7 8 9 10 Years after surgery Figure 2 Composite Kaplan-Meyer curve detailing overall survival for patients receiving bioprosthetic and mechanical valve for infective endocarditis (shaded region represents 95% confidence intervals). © Annals of Cardiothoracic Surgery. All rights reserved. Ann Cardiothorac Surg 2019;8(6):587-599 | http://dx.doi.org/10.21037/acs.2019.10.03 592 Flynn et al. Meta-analysis: mechanical valve versus bioprosthetic valve replacement in IE A B Figure 3 Forest plots detailing outcomes of (A) valve reoperation; (B) valve reinfection. 100 Mechanical valve Freedom from reoperation (%) 75 Bioprosthetic valve 50 25 Numbers at risk: 302 Bio 228 168 145 130 52 0 229 Mech 193 164 148 136 32 0 1 2 3 4 5 Years after surgery Figure 4 Composite Kaplan-Meyer curve detailing freedom from reoperation for patients receiving bioprosthetic and mechanical valve for infective endocarditis (shaded region represents 95% confidence intervals). log hazard ration for reoperation rates (Figure S4). Egger’s I2 =5% (Figure 3B). Composite survival curves demonstrated test was not applied due to the small number of studies a freedom from valve reinfection for bioprosthetic valves being assessed. of 97.0%, 91.4%, 88.3%, and 79.0% at 1-, 3-, 5-, and Three studies reported data on valve reinfections 10-year, respectively, and 97.1%, 93.8%, 92.3%, and 86.6% (20,22,23) drawing on data from 652 patients (365 patients at 1-, 3-, 5-, and 10-year, respectively, for mechanical receiving bioprosthetic valves and 287 patients receiving valves (Figure 5). Leave-one-out sensitivity analysis did not mechanical valves). Meta-analysis of derived hazard ratios demonstrate any significant difference in overall effect on demonstrated no significant difference in the rate of valve outcomes. There was no evidence of publication bias on reinfection rates (HR 0.95, 95% CI: 0.48–1.89; P=0.89). visual inspection of a funnel plot of the standard error by There was minimal heterogeneity in this small group of log hazard ration for valve reinfection rates (Figure S5). studies with the Cochrane Q-statistic P value of 0.35, Egger’s test was not applied due to the small number of © Annals of Cardiothoracic Surgery. All rights reserved. Ann Cardiothorac Surg 2019;8(6):587-599 | http://dx.doi.org/10.21037/acs.2019.10.03 Annals of cardiothoracic surgery, Vol 8, No 6 November 2019 593 100 Mechanical valve Freedom from valve reinfection (%) 75 Bioprosthetic valve 50 25 Numbers at risk: 365 Bio 296 223 179 145 116 98 81 68 54 47 0 287 Mech 239 168 157 128 105 90 73 56 34 19 0 1 2 3 4 5 6 7 8 9 10 Years after surgery Figure 5 Composite Kaplan-Meyer curve detailing freedom from valve reinfection for patients receiving bioprosthetic and mechanical valve for infective endocarditis (shaded region represents 95% confidence intervals). studies being assessed. which was most pronounced in the small number of patients who received a bioprosthetic valve under the age of 65. These patients had a markedly increased 5-year mortality Discussion compared to patients receiving a mechanical valve (HR 4.14; The evidence supporting indications for valve selection 95% CI: 1.27–13.45, P=0.018). However, patients receiving in infective endocarditis is currently limited. This meta- bioprosthetic valves in that study were older overall and analysis suggests that at present there is no clear evidence had more medical co-morbidities including liver disease that one prosthetic type should be selected over any (OR 21.86) and severe left ventricle (LV) impairment other, which is nonetheless an important finding. There (OR 14.11), therefore confounding the interpretation of was significant heterogeneity when reviewing data the results. Musci et al. (27) also note age at operation as from all analyses, which may relate to the varied patient independent risk factor for early mortality. In contrast populations being examined with variation in valve choices to Nguyen et al., Moon et al. (26) found no difference in and operations. It was not possible to correct for severity overall mortality between mechanical and bioprosthetic of infective endocarditis, urgency status, concomitant valves in patients under the age of 60. Unfortunately, there operations or co-morbidities. An important consideration was insufficient data to perform adequate meta-regression is that two of the three prospective studies (19,28) to correct for patient age, co-morbidities or pre-operative demonstrated a significant survival advantage for patients status in this meta-analysis. receiving mechanical valves, suggesting there is possibly an The choice of valve in the general population is area of future research to develop more robust evidence. determined by balancing the risk of anticoagulation-related It may have been expected that bioprosthetic valves and thromboembolic complications that may be associated would have worse overall survival as this patient cohort was with mechanical valves versus the risk of structural valve significantly older compared to those receiving mechanical degeneration and the requirement for redo valve surgery valves. Several studies reported that increasing age was an for patients receiving bioprosthetic valves. The cut-off for independent risk factor for overall mortality in patients patients receiving a mechanical valve versus a bioprosthetic with endocarditis (19,21,27,28). Nguyen et al. (28) note that valve is yet to be firmly established with conflicting advancing age increased the risk for overall 5-year mortality evidence regarding long-term outcomes for the two valves. in infective endocarditis (HR 1.03; CI: 1.00–1.05, P=0.037), Some evidence refutes an excess of mortality in middle- © Annals of Cardiothoracic Surgery. All rights reserved. Ann Cardiothorac Surg 2019;8(6):587-599 | http://dx.doi.org/10.21037/acs.2019.10.03 594 Flynn et al. Meta-analysis: mechanical valve versus bioprosthetic valve replacement in IE aged patients receiving bioprosthetic valves, including a There are several limitations that must be considered large propensity-matched study of non-elderly patients. when interpreting results from the present review. Data McClure et al. (30) demonstrated no significant difference has been collected from a relatively small number of studies in late mortality at 18 years with a cumulative survival that have been drawn from markedly different patient estimated at 18 years of 60% for bioprosthetic valve and populations. Furthermore, all studies are at a high risk of 51% for mechanical valves, though not unsurprisingly they selection bias given the non-randomized and non-blinded found an increase in the requirement for reoperation for nature of the study. In particular, many studies have stark bioprosthetic valves. Similarly, although Kulick et al. (31) differences between the bioprosthetic and mechanical valve demonstrated no difference in survival for patients aged patient populations making accurate analysis challenging. between 50 to 65 years requiring aortic or mitral valve Unfortunately, there is insufficient published data to replacement, mechanical valves were associated with a attempt to correct for some of these differences. significant increased risk of thromboembolism (HR 4.1; CI: 1.3–12.7, P=0.01) and an increased risk of reoperation Conclusions associated with bioprosthetic valves (HR 7.1; CI: 1.8-27.8 P=0.005). However, a recent meta-analysis of five studies This meta-analysis did not identify any evidence of a by Diaz et al. (32), assessed valve replacement in patients significant survival advantage, reduction in valve reinfection between the age of 50 to 70, demonstrated a statistically or reoperation rate for any particular conventional significant reduction in long-term mortality for patients prosthesis choice in the treatment of patients with infective receiving mechanical valves (HR 0.86; 95% CI: 0.76–0.97, endocarditis. It is clear that evidence guiding our decision P=0.01). They did note an increased risk of major bleeding making in this area is currently limited and is an area that with mechanical valve recipients. would benefit from future research. The choice of prosthesis There remains no simple answer for valve choice, should continue to be made with the conscientious particularly in middle aged patients. It appears that the consideration to the patients age, co-morbidities and patient presence of IE alone should not influence the decision preferences. outside other patient factors. Newer-generation bioprosthetic valves with novel preservation techniques could optimistically reduce the rate of structural valve Expert opinion: choice of prosthetic valve in degeneration, with ongoing trials currently at 4 years follow- infective endocarditis up and no episodes of structural valve degeneration to Zegri-Reiriz, Tauron date (33), although long-term follow-up is required to determine if this truly is an improvement on current In patients with infective endocarditis and indication for technologies. Another addition to the decision-making surgery, the general recommendations regarding the choice algorithm is the ever-increasing experience with valve- of prosthesis (biological or mechanical) do not differ with in-valve transcatheter aortic valve replacement which has respect to the general population. Both have shown similar offered equivalent mid-term survival rates with reduced rates of post-operative mortality and reinfection. early morbidity compared to conventional reoperative The European guidelines recommend mechanical surgery (34-36). However, an option for valve-in-valve prostheses for those patients under 60 years in the aortic replacement is questionable for valves smaller than 21 mm, position and for those under 65 years in the mitral suggesting that expected valve size and the consequences for position (12). According to the American Heart Association, future management should play a role in prosthesis choice. mechanical prostheses are recommended in patients <50 On the other hand, the latest generation mechanical valves years and biological prostheses in <70 years (13). (On-x valve) have been proven to be safe with reduced doses However, it should be noted that these patients deserve of anticoagulation and have an approved target INR range of some important considerations. In many cases, they have a 1.5–2.0. This reduction in anticoagulation has significantly serious clinical condition which is the result of uncontrolled reduced major and minor bleeding events without infection, acute heart failure or septic embolism. causing a significant increase in thromboemboli (37). In this scenario, the risk of bleeding and the implications This may address some of the major drawbacks associated of anticoagulation in the short and medium term should with mechanical valve replacement. be carefully evaluated. Thus, some young patients with a © Annals of Cardiothoracic Surgery. All rights reserved. Ann Cardiothorac Surg 2019;8(6):587-599 | http://dx.doi.org/10.21037/acs.2019.10.03 Annals of cardiothoracic surgery, Vol 8, No 6 November 2019 595 high risk of bleeding could benefit from a bioprosthesis if The main limitation with Flynn and coworkers’ meta- carefully selected. analysis is that it doesn’t take into consideration all the In worse cases, there is a risk of severe coagulopathy, nuances, reasons for one choice over the other, information hepatopathy or in an embolic stroke, due to the risk of about the pathology and complications (e.g., stroke or hemorrhagic transformation. In this population of critically intracranial bleed), and other factors guiding the surgeons’ ill patients who may need mechanical circulatory support, and patients’ choice of replacement valve. It is still true a bioprosthesis would be preferable. On the other hand, a that “Patients with IE are often very sick and have suffered sutureless bioprosthesis could be a reasonable alternative strokes, so using allografts or bioprosthetic valves simplifies for surgical treatment of prosthetic aortic valve endocarditis management and avoids postoperative anticoagulation, with significant involvement of the aortic annulus (38). lowering the risk of hemorrhagic conversion of strokes These decisions should be individualized in each case, and other bleeding complications.” When the disease is balancing the risk of bleeding due to anticoagulation against invasive and the annulus destroyed, extensive debridement, the possible benefit of implanting a mechanical prosthesis in reconstruction and usually root replacement are required. young patients. In this situation many surgeons, including us believe that an allograft is a better choice than a prosthetic valve conduit, be it mechanical or bioprosthetic valve. Expert opinion: choice of valve prosthesis in Flynn and coworkers should be thanked and patients undergoing valve surgery for infective congratulated for confirming that current guidelines endocarditis and practices for choice of prosthetic type for patients Pettersson, Coselli with infective endocarditis are adequate and produce very comparable outcomes with both mechanical and In this issue of the Annals of Cardiothoracic Surgery, Flynn bioprosthetic valves. and coworkers present a systematic review and meta- analysis comparing outcomes after surgery for left sided infective endocarditis (IE) using mechanical prosthesis Expert opinion: individuality2 versus bioprosthesis. Both choices produced very similar Misfeld outcomes with slight trends favoring the mechanical valve with regard to freedom from reoperation and risk of In the current issue of the Annals of Cardiothoracic Surgery, recurrent endocarditis; but a possible benefit for overall Flynn et al. present a systematic review and meta-analysis survival with bioprosthetic valve replacement. of eleven studies, published between 1989 and 2016, Current AATS Guidelines (39) for surgical treatment about outcomes of mechanical versus bioprosthetic valve of endocarditis makes recommendations related to the substitutes in infective endocarditis (IE). The analysis did choice of replacement valve: valve repair is preferable when not identify a significant survival advantage, reduction in possible. When replacement is required and the disease is valve reinfection or reoperation with one of the valve types. confined to cusps, leaflets, or valve prosthesis, choice of There is no clear agreement on the best prosthetic replacement valve—mechanical or tissue prosthesis should valve choice, neither for non-infective nor for IE. This be based on usual criteria. is represented by different recommendations of the With regard to risk of recurrent endocarditis the European and American guidelines (39,42). Despite guidelines say: “For patients requiring valve replacement, current recommendations, mechanical valves compared there is little evidence that risk of recurrent infection to bioprosthetic valve substitutes are probably used in the is different between mechanical and tissue prostheses.” ratio 1:10, with more tissue valves being implanted even Also, in Flynn and coworkers’ paper there is no significant in younger patients (43,44). Bioprostheses are increasingly difference in risk of recurrent endocarditis, however, in implanted not only because anticoagulation and valve noise recent registry studies from Sweden and Denmark when with mechanical valves remaining important aspects for comparing incidence of endocarditis following valve most patients, but also because interventional options, such replacement of non-infected pathologies, bioprostheses as valve-in-valve procedures in tissue valves, have become were associated with higher risk of endocarditis than an attractive second step when biological valves fail (43). mechanical prostheses (40,41). Whatever guidelines recommend, neither the patient, © Annals of Cardiothoracic Surgery. All rights reserved. Ann Cardiothorac Surg 2019;8(6):587-599 | http://dx.doi.org/10.21037/acs.2019.10.03 596 Flynn et al. Meta-analysis: mechanical valve versus bioprosthetic valve replacement in IE nor the surgeon can foresee, what kind of impact the choice a question of which of these two prostheses but if there of a valve (biological valve: risk of reoperation, but patient are alternatives; and, in my view, there is an alternative, will “forget” about having a valve substitute in between; especially useful in the aortic position with significant mechanical valve: valve noise and regular reminder because disruption of the annulus—the homograft. of regular anticoagulation checks) has for the patient during Several works appear to indicate that homografts daily life. may be more resistant to infection than either biological Therefore, it remains an individual decision, which valve or mechanical prostheses (45,46). My own experience type should be used, independent of the underlying disease also appears to demonstrate that superiority (47). (valve surgery with/without IE). This individual decision is However, others have shown no difference in the rates of not only a patient related aspect, but also surgeon related. re-infection (23). Implantation of homografts may be Patients with IE, especially if complicated by cerebral technically more demanding and, similarly to bioprostheses, embolism, will most likely have a tissue valve, despite they have the problem of biodegradation. In any case, the patient age. On the other hand, surgeon related factors, i.e., pliability of the homograft facilitates the treatment of the not being experienced with the use of allografts, may also extensively disrupted aortic annuli often seen in PVE (48). influence the type of valve used. Some authors have also proposed the use of stentless As none of the valve substitutes available for the xenografts for the same indications (49). Finally, the Ross treatment of IE has been shown to have a clear advantage operation has also been advocated in these situations (50), over the others, surgeons should use the valve substitute but it is a surgically challenging procedure to be used in an they are most familiar and experienced with under the already complex anatomical situation. background of patient related factors. Expert opinion: do not radically position yourself Expert opinion: prosthetic valve endocarditis, in the surgery for infective Endocarditis still a dilemma Mestres, Quintana Antunes In this issue of the Annals of Cardiothoracic Surgery, Flynn In the paper published in this issue of the Annals of et al. used meta-analysis to understand surgical outcomes Cardiothoracic Surgery, Flynn et al. analyse the incidence comparing mechanical and bioprosthetic valve replacement of prosthetic valve endocarditis (PVE) in patients with in infective endocarditis (IE). They found no significant biological and mechanical prostheses. They performed a difference in overall survival, reoperation and valve meta-analysis of studies reporting the outcomes of patients reinfection rates between patients with mechanical or undergoing valve replacement for infective endocarditis (IE), bioprosthetic valves. both native and prosthetic. The authors found that at 10- The decision of what type of valve to use in the setting year follow-up, there were no differences in overall survival, of IE has been discussed for over sixty years since the first and freedom from valve reinfection and reoperation surgical replacements (51,52), and remains a challenging between the two types of prostheses. However, the series and controversial topic. In addition to considerations included in the study are mostly historical and include of expected anticoagulation-related events and device both native and prosthetic endocarditis, and many types of reoperations (53), IE is another factor to consider when prostheses of both groups. choosing a prosthesis as this population has higher surgical The results are not surprising, but I cannot completely risk. Current IE guidelines do not support one device over agree with the conclusion derived by the authors that “the another due to the lack of evidence on superiority (54). presence of infective endocarditis alone should not influence In their analysis, the authors found substantial the decision of which type of valve prosthesis that should heterogeneity in the studies and were unable to correct be implanted”. Although the data is not provided by the for severity of IE, urgency status, concomitant operations authors, the fact is that there was a significant incidence of or co-morbidities. There are additional limitations in PVE after valve replacement in IE. There is no consensus the Flynn et al. meta-analysis (51). It was impossible to with regard to favouring either bioprosthetic or mechanical separate by valve positions and type of device. It is difficult valve implantation in the setting of IE. Hence, it is not to understand if under the term “bioprostheses”, stentless © Annals of Cardiothoracic Surgery. All rights reserved. Ann Cardiothorac Surg 2019;8(6):587-599 | http://dx.doi.org/10.21037/acs.2019.10.03 Annals of cardiothoracic surgery, Vol 8, No 6 November 2019 597 valves and homografts were included. Most homograft Circulation 2013;127:1647-55. implants in IE correspond to the aortic position. Patients 8. Hellgren L, Granath F, Ekbom A, et al. Biological versus implanted with a homograft are at higher risk than mechanical prosthesis in 3279 patients from the Swedish those implanted with stented bioprosthesis as they have in-patients register. Scand Cardiovasc J 2011;45:223-8. extravalvular spread of the infection, meaning the infection 9. Agnihotri AK, McGiffin DC, Galbraith AJ, et al. The is more destructive, requiring extensive reconstruction, prevalence of infective endocarditis after aortic valve especially in reoperations for prosthetic valve IE (55,56). replacement. J Thorac Cardiovasc Surg 1995;110:1708-20; Considering the complexity of IE and the process of discussion 1720-4. choosing a valve, and difficulty in producing high quality 10. Prendergast BD, Tornos P. Surgery for infective endocarditis: clinical data for this surgical population, we should who and when? Circulation 2010;121:1141-52. not adopt radical positions, with or without practice 11. Murdoch DR, Corey GR, Hoen B, et al. Clinical recommendations. Guidelines are recommendations, not presentation, etiology, and outcome of infective the law, and all options must be carefully individualized. endocarditis in the 21st century: the International The treating team makes considered decisions based Collaboration on Endocarditis-Prospective Cohort Study. on preoperative condition, comorbidities, age among Arch Intern Med 2009;169:463-73. other variables, integrating this and acting according to 12. Habib G, Lancellotti P, Antunes MJ, et al. 2015 ESC experience, expertise and local resources. Guidelines for the management of infective endocarditis: The Task Force for the Management of Infective Endocarditis of the European Society of Cardiology (ESC). Acknowledgments Endorsed by: European Association for Cardio-Thoracic None. Surgery (EACTS), the European Association of Nuclear Medicine (EANM). Eur Heart J 2015;36:3075-128. 13. Baddour LM, Wilson WR, Bayer AS, et al. Infective Footnote Endocarditis in Adults: Diagnosis, Antimicrobial Therapy, Conflicts of Interest: The authors have no conflicts of interest and Management of Complications: A Scientific Statement to declare. for Healthcare Professionals From the American Heart Association. Circulation 2015;132:1435-86. 14. Byrne JG, Rezai K, Sanchez JA, et al. Surgical management References of endocarditis: the society of thoracic surgeons clinical 1. Cahill TJ, Prendergast BD. Infective endocarditis. Lancet practice guideline. Ann Thorac Surg 2011;91:2012-9. 2016;387:882-93. 15. Martin AK, Mohananey D, Ranka S, et al. The 2017 2. Que YA, Moreillon P. Infective endocarditis. Nat Rev European Society of Cardiology (ESC)/European Cardiol 2011;8:322. Association of Cardiothoracic Surgeons (EACTS) 3. Cahill TJ, Baddour LM, Habib G, et al. Challenges in Guidelines for Management of Valvular Heart Disease- Infective Endocarditis. J Am Coll Cardiol 2017;69:325-44. Highlights and Perioperative Implications. J Cardiothorac 4. Hoen B, Alla F, Selton-Suty C, et al. Changing profile of Vasc Anesth 2018;32:2810-6. infective endocarditis: results of a 1-year survey in France. 16. Nishimura RA, Otto CM, Bonow RO, et al. 2017 AHA/ JAMA 2002;288:75-81. ACC Focused Update of the 2014 AHA/ACC Guideline 5. Dzupova O, Machala L, Baloun R, et al. Incidence, for the Management of Patients With Valvular Heart predisposing factors, and aetiology of infective endocarditis Disease: A Report of the American College of Cardiology/ in the Czech Republic. Scand J Infect Dis 2012;44:250-5. American Heart Association Task Force on Clinical 6. Li JS, Sexton DJ, Mick N, et al. Proposed modifications Practice Guidelines. J Am Coll Cardiol 2017;70:252-89. to the Duke criteria for the diagnosis of infective 17. Tierney JF, Stewart LA, Ghersi D, et al. Practical methods endocarditis. Clin Infect Dis 2000;30:633-8. for incorporating summary time-to-event data into meta- 7. Brennan JM, Edwards FH, Zhao Y, et al. Long-term safety analysis. Trials 2007;8:16. and effectiveness of mechanical versus biologic aortic valve 18. Guyot P, Ades AE, Ouwens MJ, et al. Enhanced secondary prostheses in older patients: Results from the society of analysis of survival data: reconstructing the data from thoracic surgeons adult cardiac surgery national database. published Kaplan-Meier survival curves. BMC Med Res © Annals of Cardiothoracic Surgery. All rights reserved. Ann Cardiothorac Surg 2019;8(6):587-599 | http://dx.doi.org/10.21037/acs.2019.10.03 598 Flynn et al. Meta-analysis: mechanical valve versus bioprosthetic valve replacement in IE Methodol 2012;12:9. 31. Kulik A, Bedard P, Lam BK, et al. Mechanical versus 19. Delahaye F, Chu VH, Altclas J, et al. One-year outcome bioprosthetic valve replacement in middle-aged patients. following biological or mechanical valve replacement for Eur J Cardiothorac Surg 2006;30:485-91. infective endocarditis. Int J Cardiol 2015;178:117-23. 32. Diaz R, Hernandez-Vaquero D, Alvarez-Cabo R, et al. 20. Fedoruk LM, Jamieson WR, Ling H, et al. Predictors Long-term outcomes of mechanical versus biological aortic of recurrence and reoperation for prosthetic valve valve prosthesis: Systematic review and meta-analysis. J endocarditis after valve replacement surgery for Thorac Cardiovasc Surg 2019;158:706-14.e18. native valve endocarditis. J Thorac Cardiovasc Surg 33. Bartus K, Litwinowicz R, Bilewska A, et al. Intermediate- 2009;137:326-33. term outcomes after aortic valve replacement with a 21. Greason KL, Thomas M, Steckelberg JM, et al. Outcomes novel RESILIA(TM) tissue bioprosthesis. J Thorac Dis of surgery in the treatment of isolated nonnative mitral 2019;11:3039-46. valve infective endocarditis. J Thorac Cardiovasc Surg 34. Webb JG, Mack MJ, White JM, et al. Transcatheter 2014;147:349-54. Aortic Valve Implantation Within Degenerated Aortic 22. Jassar AS, Bavaria JE, Szeto WY, et al. Graft selection for Surgical Bioprostheses: PARTNER 2 Valve-in-Valve aortic root replacement in complex active endocarditis: Registry. J Am Coll Cardiol 2017;69:2253-62. Does it matter? Ann Thorac Surg 2012;93:480-7. 35. Neupane S, Singh H, Lammer J, et al. Meta-Analysis 23. Kim JB, Ejiofor JI, Yammine M, et al. Are homografts of Transcatheter Valve-in-Valve Implantation Versus superior to conventional prosthetic valves in the setting of Redo Aortic Valve Surgery for Bioprosthetic Aortic Valve infective endocarditis involving the aortic valve? J Thorac Dysfunction. Am J Cardiol 2018;121:1593-600. Cardiovasc Surg 2016;151:1239-46, 1248.e1-2. 36. Silaschi M, Wendler O, Seiffert M, et al. Transcatheter 24. Leither MD, Shroff GR, Ding S, et al. Long-term survival valve-in-valve implantation versus redo surgical aortic valve of dialysis patients with bacterial endocarditis undergoing replacement in patients with failed aortic bioprostheses. valvular replacement surgery in the United States. Interact Cardiovasc Thorac Surg 2017;24:63-70. Circulation 2013;128:344-51. 37. Puskas J, Gerdisch M, Nichols D, et al. Reduced 25. Lytle BW, Priest BP, Taylor PC, et al. Surgical treatment anticoagulation after mechanical aortic valve replacement: of prosthetic valve endocarditis. J Thorac Cardiovasc Surg interim results from the prospective randomized on-X 1996;111:198-207; discussion 207-10. valve anticoagulation clinical trial randomized Food and 26. Moon MR, Miller DC, Moore KA, et al. Treatment of Drug Administration investigational device exemption endocarditis with valve replacement: The question of trial. J Thorac Cardiovasc Surg 2014;147:1202-10; tissue versus mechanical prosthesis. Ann Thorac Surg discussion 1210-1. 2001;71:1164-71. 38. Roselló-Díez E, Cuerpo G, Estevez F, et al. Use of 27. Musci M, Hubler M, Amiri A, et al. Surgical treatment thePerceval Sutureless Valve in Active Prosthetic Aortic for active infective prosthetic valve endocarditis: 22- Valve Endocarditis. Ann Thorac Surg 2018;105:1168-74. year single-center experience. Eur J Cardiothorac Surg 39. Pettersson GB, Coselli JS, Hussain ST, et al. 2016 The 2010;38:528-38. American Association for Thoracic Surgery (AATS) 28. Nguyen DT, Delahaye F, Obadia JF, et al. Aortic valve consensus guidelines: Surgical treatment of infective replacement for active infective endocarditis: 5-year endocarditis: Executive summary. J Thorac Cardiovasc survival comparison of bioprostheses, homografts Surg 2017;153:1241-1258.e29. and mechanical prostheses. Eur J Cardiothorac Surg 40. Glaser N, Jackson V, Holzmann MJ, et al. Prosthetic 2010;37:1025-32. valve endocarditis after surgical aortic valve replacement. 29. Reul GJ, Sweeney MS. Bioprosthetic versus mechanical Circulation 2017;136:329-31. valve replacement in patients with infective endocarditis. J 41. Østergaard L, Valeur N, Ihlemann N, et al. Incidence and Card Surg 1989;4:348-51. factors associated with infective endocarditis in patients 30. McClure RS, McGurk S, Cevasco M, et al. Late undergoing left-sided heart valve replacement. Eur Heart J outcomes comparison of nonelderly patients with stented 2018;39:2668-75. bioprosthetic and mechanical valves in the aortic position: 42. Falk V, Baumgartner H, Bax JJ, et al. 2017 ESC/EACTS a propensity-matched analysis. J Thorac Cardiovasc Surg Guidelines form the management of valvular heart disease. 2014;148:1931-9. Eur J Cardiothorac Surg 2017;52:616-64. © Annals of Cardiothoracic Surgery. All rights reserved. Ann Cardiothorac Surg 2019;8(6):587-599 | http://dx.doi.org/10.21037/acs.2019.10.03 Annals of cardiothoracic surgery, Vol 8, No 6 November 2019 599 43. Head SJ, Celik M, Kappetein AP. Mechanical versus term Evaluation of the Ross Procedure in Acute bioprosthetic aortic valve replacement. Eur Heart J Infective Endocarditis. Semin Thorac Cardiovasc Surg 2017;38:2183-91. 2017;29:494-501. 44. Savage EB, Saha-Chaudhuri P, Asher CR, et al. Outcomes 51. Braunwald NS, Cooper T, Morrow AG. Complete and prosthesis choice for active aortic valve infective replacement of the mitral valve. Successful clinical endocarditis: Analysis of The Society of Thoracic application of a flexible polyurethane prosthesis. J Thorac Surgeons Adult Cardiac Surgery Database. Ann Thorac Cardiovasc Surg 1960;40:1-11. Surg 2014;98:806-14. 52. Starr A, Edwards ML. Mitral replacement: clinical 45. Musci M, Weng Y, Hübler M, et al. Homograft aortic experience with a ball-valve prosthesis. Ann Surg root replacement in native or prosthetic active infective 1961;154:726-40. endocarditis: twenty-year single-center experience. J 53. Falk V, Baumgartner H, Bax JJ, et al. 2017 ESC/EACTS Thorac Cardiovasc Surg 2010;139:665-73. Guidelines for the management of valvular heart disease. 46. Solari S, Mastrobuoni S, De Kerchove L, et al. Over 20 Eur J Cardiothorac Surg 2017;52:616-64. years experience with aortic homograft in aortic valve 54. Habib G, Lancellotti P, Antunes MJ, et al. 2015 ESC replacement during acute infective endocarditis. Eur J Guidelines for the management of infective endocarditis: Cardiothorac Surg 2016;50:1158-64. The Task Force for the Management of Infective 47. Lopes S, Calvinho P, de Oliveira F, et al. Allograft aortic Endocarditis of the European Society of Cardiology (ESC). root replacement in complex prosthetic endocarditis. Eur J Endorsed by: European Association for Cardio-Thoracic Cardiothorac Surg 2007;32:126-30. Surgery (EACTS), the European Association of Nuclear 48. Zwischenberger JB, Shalaby TZ, Conti VR. Viable Medicine (EANM). Eur Heart J 2015;36:3075-128. Cryopreserved Aortic Homograft for Aortic Valve 55. Yankah AC, Klose H, Petzina R, et al. Surgical Endocarditis and Annular Abscesses. Ann Thorac Surg management of acute aortic root endocarditis with viable 1989;48:365-9; discussion 369-70. homograft: 13-year experience. Eur J Cardiothorac Surg 49. Schneider AW, Hazekamp MG, Versteegh MI, et al. 2002;21:260-7. Stentless bioprostheses: a versatile and durable solution 56. Leontyev S, Borger MA, Modi P, et al. Surgical in extensive aortic valve endocarditis. Eur J Cardiothorac management of aortic root abscess: a 13-year experience Surg 2016;49:1699-704. in 172 patients with 100% follow-up. J Thorac Cardiovasc 50. Ratschiller T, Sames-Dolzer E, Paulus P. Long- Surg 2012;143:332-7. Cite this article as: Flynn CD, Curran NP, Chan S, Zegri- Reiriz I, Tauron M, Tian DH, Pettersson GB, Coselli JS, Misfeld M, Antunes MJ, Mestres CA, Quintana E. Systematic review and meta-analysis of surgical outcomes comparing mechanical valve replacement and bioprosthetic valve replacement in infective endocarditis. Ann Cardiothorac Surg 2019;8(6):587-599. doi: 10.21037/acs.2019.10.03 © Annals of Cardiothoracic Surgery. All rights reserved. Ann Cardiothorac Surg 2019;8(6):587-599 | http://dx.doi.org/10.21037/acs.2019.10.03 Supplementary Records identified through Additional records identified search strategy after on bibliography review =2 duplicates removed =4,053 Number of titles and Number of records excluded on title and abstracts reviewed =4055 abstract review =3985 Number of full-text articles Number of records excluded after full text assessed for eligibility =70 review =40 • Conference abstract =12 records • No focus on prosthesis type =7 records • Non-comparative study =9 records • No focus on IE =12 records Number of records Included Number of studies excluded due to in qualitative synthesis =30 insufficient data available for extraction =18 Number of studies excluded due to insufficient follow up =1 Number of studies excluded in qualitative synthesis =11 Figure S1 Study selection. PRISMA flow diagram detailing the systematic review to identify studies reporting the outcomes of bioprosthetic valve replacement compared to mechanical valve replacement for infective endocarditis. Blinding of participants and personnel (performance bias) Blinding of outcome assessment (detection bias) Random sequence generation ( selection bias ) Incomplete outcome data (attrition bias) Allocation concealment ( selection bias) Selective reporting (reporting bias) Other bias Delahaye 2015 Fedoruk 2009 Greason 2014 Jassar 2012 Kim 2016 Leither 2013 Lytle 1996 Moon 2001 Musci 2010 Nguyen 2010 Reul 1989 Figure S2 Risk of bias table. Thorough assessment of biases for all studies (red = high risk, yellow = uncertain risk, green = low risk). Funnel plot of standard error of log hazard ratio 0.0 0.1 0.2 Standard error 0.3 0.4 0.5 0.6 –2.0 –1.5 –1.0 –0.5 0.0 0.5 1.0 1.5 2.0 log hazard ratio Figure S3 Funnel plot of standard error of log hazard ratio for overall survival. Funnel plot of standard error by log hazard ratio 0.0 0.2 Standard error 0.4 0.6 0.8 1.0 –2.0 –1.5 –1.0 –0.5 0.0 0.5 1.0 1.5 2.0 log hazard ratio Figure S4 Funnel plot of standard error of log hazard ratio for reoperation rates. Funnel plot of standard error by log hazard ratio 0.0 0.5 Standard error 1.0 1.5 2.0 –2.0 –1.5 –1.0 –0.5 0.0 0.5 1.0 1.5 2.0 log hazard ratio Figure S5 Funnel plot of standard error of log hazard ratio for valve reinfection rates.