Repair versus replacement for the surgical correction of tricuspid regurgitation: a meta-analysis

Repair versus replacement for the surgical correction of tricuspid regurgitation: a meta-analysis Abstract OBJECTIVES Tricuspid valve repair (TVr) has both a theoretical advantage in preserving right ventricular function and disadvantages such as a higher risk of repair failure in the long-term compared with tricuspid valve replacement (TVR). This study was conducted to compare the results of TVr with those of TVR to find an optimal treatment option. METHODS A literature search of 5 databases was performed. The primary outcome was all-cause mortality. Secondary outcomes were tricuspid reoperation and valve-related events. Subgroup analyses were performed according to the risk of bias, year of publication and proportions of patients with tricuspid regurgitation Grade ≥3, functional aetiology and isolated tricuspid valve surgery. Publication bias was explored using the funnel plot and Egger’s test. RESULTS Seventeen retrospective studies involving 4561 patients (TVr group = 3432 patients and TVR group = 1129 patients) were included. A pooled analysis showed that the risk of all-cause mortality was significantly higher in the TVR group than in the TVr group [hazard ratio (95% confidence interval) 1.59 (1.26–2.00)]. There were no significant differences in tricuspid valve reoperation in 6 studies and valve-related events in 5 studies between the TVR and TVr groups [hazard ratio (95% confidence interval) 1.30 (0.88–1.91) and 1.47 (0.91–2.38), respectively]. None of the subgroup analyses demonstrated a significant difference in the hazard ratio of all-cause mortality. No publication bias was identified for the primary and secondary outcomes. CONCLUSIONS This meta-analysis indicates that TVr is more beneficial compared with TVR in terms of all-cause mortality. From the available data, TVr is not associated with an increased risk of tricuspid reoperation compared with TVR. Valve disease surgery, Tricuspid valve disease, Meta-analysis INTRODUCTION Current guidelines recommend tricuspid valve (TV) surgery for (i) patients with severe primary tricuspid regurgitation (TR) accompanying symptoms or right heart failure and (ii) patients with severe or progressive functional TR who underwent left-sided valve surgery [1]. The surgical treatment of choice for patients with TR is, however, still controversial; although previous studies have demonstrated favourable long-term outcomes after TV repair (TVr) compared with TV replacement (TVR) [2–4], other studies have argued that worse clinical outcomes after TVR might be caused by worse baseline characteristics of TVR patients such as advanced heart failure and a history of frequent cardiac surgery [5, 6]. Therefore, this meta-analysis was conducted to compare results of TVr with those of TVR in patients with TR. This study hypothesized that despite a long-term risk of repair failure, long-term clinical outcomes, particularly survival rates, would be better after TVr than after TVR. MATERIALS AND METHODS Multiple comprehensive databases were used to find literature comparing the results of TVr with those of TVR. This systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [7]. Data source and literature searches Full-text articles were searched using MEDLINE, EMBASE, the Cochrane Central Register of Controlled Trials (CENTRAL), Web of Science and Scopus on 6 February 2017. No restrictions were placed on the language or year of publication. The following keywords and Medical Subject Headings (MeSH) were searched through Medline: ((((‘tricuspid valve’[MeSH Terms] OR ‘tricuspid valve insufficiency’[MeSH Terms]) OR ‘tricuspid valve prolapse’[MeSH Terms]) OR ((((tricuspid valve disease[Title/Abstract] OR tricuspid regurgitation[Title/Abstract]) OR tricuspid valve regurgitation[Title/Abstract]) OR tricuspid insufficiency[Title/Abstract]) OR tricuspid valve insufficiency[Title/Abstract])) AND (‘cardiac valve annuloplasty’[MeSH Terms] OR ((annuloplasty[Title/Abstract] OR valvuloplasty[Title/Abstract]) OR valve repair[Title/Abstract]))) AND (‘heart valve prosthesis implantation’[MeSH Terms] OR (replacement[Title/Abstract] OR implantation[Title/Abstract])) (Supplementary Material, Table S1). Search strategies were adapted for other databases based on the MEDLINE strategy. Study selection The inclusion of studies was independently decided by 2 reviewers (J.W.C. and H.Y.H.) based on the selection criteria. Any disagreements between the reviewers were resolved through discussion or review from the third author (M.-J.J.). The study selection was made following 2 levels of screening: at the first level, the titles and abstracts of the identified studies were screened. At the second level, the full texts were reviewed. Studies were included in this meta-analysis if they (i) enrolled patients undergoing TV surgery and (ii) compared the survival rates of patients receiving TVr with those receiving TVR as time-to-event data. When duplicated publications with overlapping study populations were found, the study with the largest sample size was selected. Data extraction Study characteristics and the patient baseline data were extracted independently by 2 reviewers (J.W.C. and H.Y.H.). Data regarding primary and secondary outcomes were also independently extracted by 2 reviewers (M.-J.J and H.Y.H.). Any disagreements between reviewers were resolved through discussion. If the outcome data were unclear, the authors were contacted via email. Assessment of study quality The overall study quality was assessed independently by two reviewers (M.-J.J. and H.Y.H.) using Risk Of Bias In Non-randomized Studies of Interventions (ROBINS-I) [8]. Seven domains of risk of bias (ROB) were assessed: bias due to confounding, bias in selection of participants into the study, bias in classification of interventions, bias due to departures from intended interventions, bias due to missing data, bias in measurement of outcomes and bias in selection of the reported result. Based on the assessment of each domain, an overall ROB was determined as low, moderate, serious or critical. Any disagreements between the reviewers were resolved through discussion. Statistical analysis The primary outcome was to investigate the association between TVR versus TVr and all-cause mortality in patients with TR. The secondary outcomes were time-to-TV reoperation and time-to-valve-related events (VREs). The outcomes were measured in terms of the hazard ratio (HR) of using TVR compared with TVr. For each study, the HR and 95% confidence interval (CI) were directly extracted from the research article. When the HR and its 95% CI were not reported, their values were calculated using other available statistical information (log-rank test, Kaplan–Meier curve) [9]. In 1 study [10], their values for all-cause mortality were drawn from both 30-day morality and the Kaplan–Meier curve for 30-day survivors. Pooled HRs, 95% CIs and P-values were calculated using the random-effect model. Two-sided P-values <0.05 were considered statistically significant. Heterogeneity was assessed with the χ2 tests and quantified with I2 statistics to evaluate the extent of variability that was attributed to statistical heterogeneity between studies. The I2 values of 25%, 50% and 75% have been suggested as indicators of low, moderate and high heterogeneity, respectively [11]. A sensitivity analysis was conducted to assess the influence of an individual study on the results by omitting 1 study in each turn [12]. Subgroup analyses were performed to evaluate the influences of ROB and patient characteristics on the study results; a stratified analysis was conducted based on the result of the ROB assessment (serious or moderate), and tertile-group comparisons were performed using all studies demonstrating available data according to the proportions of patients with TR Grade 3 or 4, functional TR and isolated TV surgery in each study. Additional subgroup analysis was performed by dividing studies according to the year of publication (before 2000 vs from 2000 to 2009 vs after 2010). A funnel plot and the Egger’s test for asymmetry were applied to assess the possibility of publication bias among the studies. All analyses were performed using STATA version 12.0 (Stata Corporation, College Station, TX, USA). RESULTS Identification of studies The database searches yielded 2942 articles. Of these, 2891 publications were excluded as it was clear from the title and abstract that they did not fulfil the selection criteria. For the remaining 51 articles, full manuscripts were obtained, and following the scrutiny of these manuscripts, 34 publications were excluded due to unmet inclusion criteria (n = 28), duplicated data (n = 5) or the inability to extract HRs (n = 1). Therefore, the total number of studies included in this review was 17 (Fig. 1) [2–6, 10, 13–23]. Figure 1: View largeDownload slide Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) flow diagram. Figure 1: View largeDownload slide Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) flow diagram. Study characteristics and patient population All studies were non-randomized and involved 4561 patients (3432 patients in the TVr group and 1129 patients in the TVR group). Nine studies [4, 5, 10, 14, 17, 19, 21, 22] exclusively included TR patients, and another 5 studies [3, 6, 15, 16, 20] included 85.2–98.9% of patients with TR. In 1 study [2], the authors demonstrated that only 77.6% of study patients had TR of Grade 3 or 4, while only 0.7% of patients had tricuspid stenosis of Grade 3 or 4. The proportion of patients with TR was unclear in the other 2 studies (Table 1) [13, 23]. The patients were, on average or median, in their late 40s to early 60s and were predominantly women (range 52.9–84.7%). In the TVr group, the use of a flexible band or a rigid ring was present in all but 1 study; 1596 (46.9%) patients among 3405 patients underwent ring annuloplasty. In the 1004 TVR patients for whom the type of prosthesis was presented, mechanical and bioprosthetic TVRs were performed in 311 (31.0%) patients and 693 (69.0%) patients, respectively (Table 2). Table 1: Study characteristics Study  Operative era  Country  Study population   Proportion of TR patients (%)  Follow-up   Total  TVr  TVR  Duration (years)a  Completeness (%)  Kratz et al. [10]  1972–1983  US  41  17  24    mean 7.3 (1–12)  100  Cohen et al. [11]  Unclear  US  17  6  11  100  mean 5.9 (0.2–14.7)  92  Kay et al. [14]  1975–1988  US  156  127  29  100  0–12    McGrath et al. [15]  1961–1987  US  530  351  179  92.3  mean 4.2 (0–26.3)  98.6  Bajzer et al. [2]  1988–1997  US  401  392  9  >77.6b  0–8.3    Singh et al. [3]  1979–2003  Canada  250  178  72  85.2c  5.2 ± 4.1  99  Guenther et al. [16]  1974–2003  Germany  416  310  106  90.6c  5.9 ± 6.3  97  Moraca et al. [23]  1986–2006  US  315  222  93    5.4 ± 5.4  100  Chen et al. [17]  2005–2010  Taiwan  77  44  33  100  2.7 ± 1.5  100  Marquis-Gravel et al. [6]  1977–2008  Canada  926  792  134  98.3  median 4.3  92  Kim et al. [18]  1997–2010  Korea  449  397  52  100  5.3 (0–17.5)  93.5  Raikhelkar et al. [19]  1998–2010  US  56  27  29  100c  2.3 (0–11.6)  95  De Meester et al. [20]  1995–2011  Belgium  92  33  59  98.9  0–17    Hwang et al. [4]  1996–2012  Korea  175  108  67  100  6.9 (0.3–17.7)  96  Oh et al. [21]  1965–2011  New Zealand  72  38  34  100  mean 13.7  100  Jintao et al. [22]  2003–2008  China  228  127  101  100  mean 5.4  96.1  Chang et al. [5]  1994–2012  Korea  360  263  97  100  6.8 ± 4.6  99.7  Study  Operative era  Country  Study population   Proportion of TR patients (%)  Follow-up   Total  TVr  TVR  Duration (years)a  Completeness (%)  Kratz et al. [10]  1972–1983  US  41  17  24    mean 7.3 (1–12)  100  Cohen et al. [11]  Unclear  US  17  6  11  100  mean 5.9 (0.2–14.7)  92  Kay et al. [14]  1975–1988  US  156  127  29  100  0–12    McGrath et al. [15]  1961–1987  US  530  351  179  92.3  mean 4.2 (0–26.3)  98.6  Bajzer et al. [2]  1988–1997  US  401  392  9  >77.6b  0–8.3    Singh et al. [3]  1979–2003  Canada  250  178  72  85.2c  5.2 ± 4.1  99  Guenther et al. [16]  1974–2003  Germany  416  310  106  90.6c  5.9 ± 6.3  97  Moraca et al. [23]  1986–2006  US  315  222  93    5.4 ± 5.4  100  Chen et al. [17]  2005–2010  Taiwan  77  44  33  100  2.7 ± 1.5  100  Marquis-Gravel et al. [6]  1977–2008  Canada  926  792  134  98.3  median 4.3  92  Kim et al. [18]  1997–2010  Korea  449  397  52  100  5.3 (0–17.5)  93.5  Raikhelkar et al. [19]  1998–2010  US  56  27  29  100c  2.3 (0–11.6)  95  De Meester et al. [20]  1995–2011  Belgium  92  33  59  98.9  0–17    Hwang et al. [4]  1996–2012  Korea  175  108  67  100  6.9 (0.3–17.7)  96  Oh et al. [21]  1965–2011  New Zealand  72  38  34  100  mean 13.7  100  Jintao et al. [22]  2003–2008  China  228  127  101  100  mean 5.4  96.1  Chang et al. [5]  1994–2012  Korea  360  263  97  100  6.8 ± 4.6  99.7  a Data were represented as mean ± standard deviation, median (range) or range if unspecified. b 77.6% of study patients had TR of Grade III or IV. c Including patients with mixed tricuspid stenoinsufficiency. TR: tricuspid regurgitation; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Table 2: Patients characteristics Study  Age (years), mean ± SD   Female (%)   Type of surgery   TR of Grade 3 or 4 (%)  Functional TR (%)  Isolated TV surgery (%)  Previous cardiac surgery (%)   TVr  TVR  TVr  TVR  Use of ring in TVr (%)  Mechanical valve in TVR (%)  TVr  TVR  Kratz et al. [10]  48.0 ± 14.5  63.4  41.2      56.1  7.3      Cohen et al. [11]  41 ± 21.9  53 ± 6.2  50  54.5  16.7  9.1  64.7    0  0  0  Kay et al. [14]  61 ± 11  69.2  0  37.9      0  9.6  McGrath et al. [15]  mean 56.9 (10.1–79.9)  79.2  6.8  13.4  75.7    3.2  42.2  55.9  Bajzer et al. [2]  mean 60  69.6  76.8  66.7  77.6      0  0  Singh et al. [3]  51 ± 15  51 ± 14  68.0  63.9  33.1  25.0    0    43.8  56.9  Guenther et al. [16]  60.6 ± 12.5  49.9 ± 11.3  64.8  79.2  13.9  35.9      12.0  44.2  43.4  Moraca et al. [23]  58.4 ± 16.4  52.2 ± 16.2  62.2  62.4  45.5  22.6    51.1  41.0  36.0  65.6  Chen et al. [17]  58.5 ± 12.4  52.9 ± 5.8  47.7  60.6  100  30.3    48.1  33.8  100  100  Marquis-Gravel et al. [6]  63 (54–70)a  57 (44–64)a  72.6  67.9  46.6  11.9  73.0  71.8  14.7      Kim et al. [18]  56.4 ± 12.3  54.6 ± 12.7  65.5  71.2  67.8  63.5  100  80.6  17.6  11.8  34.6  Raikhelkar et al. [19]  51.8 ± 18.2  60.2 ± 13  63.0  51.7    3.4    14.2  100  29.6  69.0  De Meester et al. [20]  56 ± 14  62.0  75.8  44.1  75.0  41.3  100  62.0  Hwang et al. [4]  57 ± 12  57 ± 10  70.4  71.6  28.7  44.8  100  100  22.9  30.6  89.6  Oh et al. [21]  48.9 ± 16  46.9 ± 17  63.2  79.4  55.3  20.6  84.3    100  65.8  73.5  Jintao et al. [22]  58 ± 9.2  57.9 ± 9.4  59.8  68.3  90.6    47.8    0  0  0  Chang et al. [5]  59 ± 12  54 ± 13  64.6  76.3  70.7  71.1  100  70.3  15.3  17.1  59.8  Study  Age (years), mean ± SD   Female (%)   Type of surgery   TR of Grade 3 or 4 (%)  Functional TR (%)  Isolated TV surgery (%)  Previous cardiac surgery (%)   TVr  TVR  TVr  TVR  Use of ring in TVr (%)  Mechanical valve in TVR (%)  TVr  TVR  Kratz et al. [10]  48.0 ± 14.5  63.4  41.2      56.1  7.3      Cohen et al. [11]  41 ± 21.9  53 ± 6.2  50  54.5  16.7  9.1  64.7    0  0  0  Kay et al. [14]  61 ± 11  69.2  0  37.9      0  9.6  McGrath et al. [15]  mean 56.9 (10.1–79.9)  79.2  6.8  13.4  75.7    3.2  42.2  55.9  Bajzer et al. [2]  mean 60  69.6  76.8  66.7  77.6      0  0  Singh et al. [3]  51 ± 15  51 ± 14  68.0  63.9  33.1  25.0    0    43.8  56.9  Guenther et al. [16]  60.6 ± 12.5  49.9 ± 11.3  64.8  79.2  13.9  35.9      12.0  44.2  43.4  Moraca et al. [23]  58.4 ± 16.4  52.2 ± 16.2  62.2  62.4  45.5  22.6    51.1  41.0  36.0  65.6  Chen et al. [17]  58.5 ± 12.4  52.9 ± 5.8  47.7  60.6  100  30.3    48.1  33.8  100  100  Marquis-Gravel et al. [6]  63 (54–70)a  57 (44–64)a  72.6  67.9  46.6  11.9  73.0  71.8  14.7      Kim et al. [18]  56.4 ± 12.3  54.6 ± 12.7  65.5  71.2  67.8  63.5  100  80.6  17.6  11.8  34.6  Raikhelkar et al. [19]  51.8 ± 18.2  60.2 ± 13  63.0  51.7    3.4    14.2  100  29.6  69.0  De Meester et al. [20]  56 ± 14  62.0  75.8  44.1  75.0  41.3  100  62.0  Hwang et al. [4]  57 ± 12  57 ± 10  70.4  71.6  28.7  44.8  100  100  22.9  30.6  89.6  Oh et al. [21]  48.9 ± 16  46.9 ± 17  63.2  79.4  55.3  20.6  84.3    100  65.8  73.5  Jintao et al. [22]  58 ± 9.2  57.9 ± 9.4  59.8  68.3  90.6    47.8    0  0  0  Chang et al. [5]  59 ± 12  54 ± 13  64.6  76.3  70.7  71.1  100  70.3  15.3  17.1  59.8  a Median with interquartile range. TR: tricuspid regurgitation; TV: tricuspid valve; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Quality of the included studies Regarding all-cause mortality, of the 17 eligible studies that were included, the appropriate methods to adjust for measured confounders were used in 7 studies, and those were rated as moderate ROB due to confounding. Propensity score matching was performed in 1 study [23], propensity score adjusted was performed in 2 studies [6, 18], inverse probability of treatment weighting was performed in 2 studies [4, 5] and multivariable Cox proportional hazard model was performed in 2 studies [2, 3]. The other 10 studies reported only the results of univariate analysis and were considered as a serious ROB (Supplementary Material, Table S2). Two studies demonstrated late mortality instead of all-cause mortality by excluding patients with early mortality [6] or 1-year mortality [19]. These 2 studies were found to have moderate and low ROB in the selection of study participants respectively, because the former study showed different operative mortality between the TVR and TVr groups (21% vs 13%), whereas the latter study demonstrated similar 1-year mortality between the 2 groups (21.4% vs 20%). Regarding secondary outcomes, for the TV reoperation, all the 6 studies were found to have serious ROB due to confounding, because they only reported the results from univariate analysis. For the VRE, all the 5 studies were rated as having a moderate ROB due to confounding (Supplementary Material, Table S3). All-cause mortality A pooled analysis of 17 studies in 4382 patients showed that the risk of all-cause mortality was significantly higher in the TVR group than in the TVr group (HR 1.59, 95% CI =  1.26–2.00; Fig. 2). There was a moderate-to-high heterogeneity between the studies (I2 = 62.0%, P < 0.001). A sensitivity analysis showed that 1 study by Singh et al. [3] had the greatest influence on the results: the pooled HR without this study was 1.42 (95% CI 1.19–1.69) with an I2 of 31.7% (P = 0.108; Fig. 3). Figure 2: View largeDownload slide The HR of all-cause mortality for TVR compared with TVr from 17 studies including 4382 patients. CI: confidence interval; HR: hazard ratio; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Figure 2: View largeDownload slide The HR of all-cause mortality for TVR compared with TVr from 17 studies including 4382 patients. CI: confidence interval; HR: hazard ratio; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Figure 3: View largeDownload slide Sensitivity analysis to identify the impact of each study on study results. CI: confidence interval; HR: hazard ratio. Figure 3: View largeDownload slide Sensitivity analysis to identify the impact of each study on study results. CI: confidence interval; HR: hazard ratio. The subgroup analysis according to ROB grade showed that this benefit was statistically significant for both the serious and the moderate ROB groups; the pooled HRs were 1.38 (95% CI 1.10–1.73) and 1.82 (95% CI 1.16–2.84) in the 10 studies with serious ROB and 7 studies with moderate ROB, respectively (Fig. 4). There were no statistically significant differences in the HRs between the subgroups when the subgroup analyses were performed according to the proportions of patients with a TR of Grade ≥3 (in 10 studies), patients with functional TR (in 10 studies), patients with isolated TV surgery (in 15 studies) and year of publication (P = 0.868, 0.401, 0.416 and 0.607, respectively; Fig. 5). All the subgroup analyses were also performed after excluding 1 study by Singh et al. [3]; there were still no statistically significant differences in the HRs between the subgroups. Figure 4: View largeDownload slide The HR of all-cause mortality for TVR compared with TVr from 10 studies showing the results of univariate analysis and 7 studies presenting the results of adjusted analysis. CI: confidence interval; HR: hazard ratio; ROB: risk of bias; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Figure 4: View largeDownload slide The HR of all-cause mortality for TVR compared with TVr from 10 studies showing the results of univariate analysis and 7 studies presenting the results of adjusted analysis. CI: confidence interval; HR: hazard ratio; ROB: risk of bias; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Figure 5: View largeDownload slide Subgroup analyses between the 3 groups based on the tertile values according to the proportions of patients with TR of Grade 3 or 4, functional TR, isolated tricuspid valve surgery and year of publication. CI: confidence interval; HR: hazard ratio; OP: operation; TR: tricuspid regurgitation; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Figure 5: View largeDownload slide Subgroup analyses between the 3 groups based on the tertile values according to the proportions of patients with TR of Grade 3 or 4, functional TR, isolated tricuspid valve surgery and year of publication. CI: confidence interval; HR: hazard ratio; OP: operation; TR: tricuspid regurgitation; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Tricuspid valve reoperation Seven studies [3, 6, 14–16, 18, 21] compared time-to-TV reoperation between the TVr and the TVR groups using univariate analysis. The HR could not be calculated from 1 study [18] in which non-significant differences (P = 0.15) in TV reoperation rates were demonstrated in the 3 groups of mechanical TVR, bioprosthetic TVR and TVr patients instead of in the TVR and the TVr groups; therefore, it was excluded from the analysis. A pooled analysis of 6 studies including 2350 patients demonstrated that there was no statistically significant difference in the risk of TV reoperation between the TVR and the TVr groups (HR 1.30, 95% CI 0.88–1.91; Fig. 6A). The I2 statistics showed a low-to-moderate heterogeneity among the studies (I2 = 27.5%, P = 0.229). Figure 6: View largeDownload slide The HR of (A) tricuspid valve reoperation and (B) valve-related events for TVR compared with TVr from 6 studies demonstrating the results of univariate analysis and 5 studies showing the results of adjusted analysis, respectively. CI: confidence interval; HR: hazard ratio; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Figure 6: View largeDownload slide The HR of (A) tricuspid valve reoperation and (B) valve-related events for TVR compared with TVr from 6 studies demonstrating the results of univariate analysis and 5 studies showing the results of adjusted analysis, respectively. CI: confidence interval; HR: hazard ratio; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Valve-related events The VRE rates were demonstrated in 5 studies [2–5, 18]. All the HRs were derived from the adjusted analysis, including inverse probability of treatment weighting-adjusted HRs in 2 studies [4, 5], propensity score-adjusted HRs in 1 study [18] and HRs from the Cox proportional hazard model in 2 studies [2, 3]. Three studies [3–5] demonstrated VRE according to the current guidelines [24]; however, 2 of the 3 studies included cardiac mortality [4] or all-cause mortality [3] instead of valve-related mortality. The VRE was defined as death, TV reoperation and readmission due to congestive heart failure in another study [18] and as death, recurrence of symptoms that prompted the initial TV surgery, the occurrence cardinal symptoms attributed to congestive heart failure and TV repair failure that led to a second TV surgery in the other study [2]. A pooled analysis of 5 studies including 1635 patients showed that there was no statistically significant difference in the risk of VRE between the TVR and TVr groups (HR 1.47, 95% CI 0.91–2.38; Fig. 6B). A pooled analysis of 3 studies including 785 patients in which VRE was defined according to the guidelines also showed a non-significant difference in VRE between the 2 groups (HR 1.27, 95% CI 0.81–1.99). Publication bias There was no evidence of publication bias in the funnel plots of all-cause mortality, TV reoperation and VRE (Supplementary Material, Fig. S1). The Egger’s tests also showed that publication bias was non-significant in each outcome (P = 0.150, 0.320 and 0.533, respectively). DISCUSSION This meta-analysis demonstrated 2 main findings. First, TVR was associated with increased all-cause mortality compared with TVr in patients with TR. Second, there were no differences in morbidity, such as TV reoperation and VRE, between patients who underwent TVR and those who underwent TVr. The current guidelines suggest several indications for TV surgery in patients with TR, such as symptomatic severe primary TR, asymptomatic severe primary TR with right ventricular dysfunction and functional TR of a severe degree or progressive nature during left-sided valve surgery [1]. Even in these guidelines, TVr is recommended as the treatment of choice only for patients who undergo TV surgery combined with a left-sided valve surgery for mild or moderate functional TR in a progressive nature. A theoretical disadvantage of TVR over TVr has been suggested to be progressive right ventricular dysfunction caused by a large, rigid prosthesis in a deformable, low-pressure cavity [3]. A recent study demonstrated, using cardiac magnetic resonance study, that right ventricular ejection fraction decreased after TVR, but it was preserved after TVr [25]. However, there have been controversial results regarding the long-term clinical outcomes after TVR versus TVr, and no definite conclusions could be drawn [5, 6]. To the best of our knowledge, this is the first meta-analysis comparing the results of TVR with those of TVr. This study clearly demonstrated that TVR significantly increased all-cause mortality after surgery compared with TVr. The subgroup analyses demonstrated that this finding was consistent regardless of the severity of TR, the proportion of patients with functional TR and the proportion of patients with isolated TV surgery. In the pooled analysis for all-cause mortality, I2 statistics demonstrated a significant heterogeneity among studies with an I2 value of 66.0%. Although no statistically significant differences in the HRs were found in the subgroup analyses as demonstrated above, sensitivity analyses showed that this heterogeneity among studies was significantly affected in 1 study by Singh et al. [3]. Possible explanations for this included the following: (i) all patients included in that study had organic TV disease and (ii) a substantial proportion (35%) of patients had congenital heart disease. Two of the 17 studies demonstrated late mortality instead of all-cause mortality by excluding early mortality [6] or 1-year mortality [19]. The operative mortality after TVR was significantly higher than that after TVr (21% vs 13%) in the former study [6], and this might attenuate the true HR of all-cause mortality in that study. One of the major concerns when performing TVr is a risk of recurrent TR after repair, because the subsequent TV reoperation is still a high-risk procedure [26, 27]. However, a pooled analysis of 6 studies demonstrated that the risk of TV reoperation was not higher in the TVr group than in the TVR group. In addition, the risk of VRE was not significantly different between the 2 groups. However, further data might be needed to draw definite conclusions, because the HR for reoperation was drawn from only 6 of the 17 studies using the results of univariate analyses, and the HR of VRE was calculated using 5 studies in which the VREs were defined differently. Limitations There are limitations to this study that must be acknowledged. First, all of the included studies were non-randomized studies, and the confounding variables could affect the results of our analysis. Second, the definition of events and the duration of follow-up were different among studies. Finally, a publication bias could not be ruled out, because meta-analyses can only report on the outcomes of published studies. However, we tried to minimize the publication bias by not placing a restriction on the language or year of publication. CONCLUSION This meta-analysis indicates that TVr is more beneficial compared with TVR regarding all-cause mortality. From the available data, TVr is not associated with an increased risk of tricuspid reoperation compared with TVR. Surgeons should try to repair the valve regardless of the aetiology or severity when surgical correction is needed in patients with TR. SUPPLEMENTARY MATERIAL Supplementary material is available at EJCTS online. ACKNOWLEDGEMENTS The authors greatly appreciate Dr Keitaro Okamoto's (Faculty of Medicine, Oita University, Oita, Japan) contribution in reviewing the articles that were written in Japanese. The authors also thank Drs Pieter De Meester, Werner Budts, Pyo Won Park and Hyoung Woo Chang for confirming the data regarding their papers via mail. Conflict of interest: none declared. REFERENCES 1 Nishimura RA, Otto CM, Bonow RO, Carabello BA, Erwin JP3rd, Guyton RA et al.   2014 AHA/ACC guideline for the management of patients with valvular heart disease: executive summary: a report of the American College of Cardiology/American Heart Association task force on practice guidelines. J Am Coll Cardiol  2014; 63: 2438– 88. Google Scholar CrossRef Search ADS PubMed  2 Bajzer CT, Stewart WJ, Cosgrove DM, Azzam SJ, Arheart KL, Klein AL. Tricuspid valve surgery and intraoperative echocardiography. J Am Coll Cardiol  1998; 32: 1023– 31. Google Scholar CrossRef Search ADS PubMed  3 Singh SK, Tang GHL, Maganti MD, Armstrong S, Williams WG, David TE et al.   Midterm outcomes of tricuspid valve repair versus replacement for organic tricuspid disease. Ann Thorac Surg  2006; 82: 1735– 41. Google Scholar CrossRef Search ADS PubMed  4 Hwang HY, Kim KH, Kim KB, Ahn H. Treatment for severe functional tricuspid regurgitation: annuloplasty versus valve replacement. Eur J Cardiothorac Surg  2014; 46: e21– 7. Google Scholar CrossRef Search ADS PubMed  5 Chang HW, Jeong DS, Cho YH, Sung K, Kim WS, Lee YT et al.   Tricuspid valve replacement vs. repair in severe tricuspid regurgitation. Circ J  2017; 81: 330– 8. Google Scholar CrossRef Search ADS PubMed  6 Marquis-Gravel G, Bouchard D, Perrault LP, Page P, Jeanmart H, Demers P et al.   Retrospective cohort analysis of 926 tricuspid valve surgeries: clinical and hemodynamic outcomes with propensity score analysis. Am Heart J  2012; 163: 851– 8. Google Scholar CrossRef Search ADS PubMed  7 Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Int J Surg  2010; 8: 336– 41. Google Scholar CrossRef Search ADS PubMed  8 Sterne JA, Hernan MA, Reeves BC, Savovic J, Berkman ND, Viswanathan M et al.   Robins-I: A tool for assessing risk of bias in non-randomised studies of interventions. BMJ  2016; 355: i4919. Google Scholar CrossRef Search ADS PubMed  9 Tierney JF, Stewart LA, Ghersi D, Burdett S, Sydes MR. Practical methods for incorporating summary time-to-event data into meta-analysis. Trials  2007; 8: 16. Google Scholar CrossRef Search ADS PubMed  10 Kratz JM, Crawford FA, Stroud MR, Appleby DC, Hanger KH. Trends and results in tricuspid valve surgery. Chest  1985; 88: 837– 40. Google Scholar CrossRef Search ADS PubMed  11 Cohen SR, Sell JE, McIntosh CL, Clark RE. Tricuspid regurgitation in patients with acquired, chronic, pure mitral regurgitation. II. Nonoperative management, tricuspid valve annuloplasty, and tricuspid valve replacement. J Thorac Cardiovasc Surg  1987; 94: 488– 97. Google Scholar PubMed  12 Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ  2003; 327: 557– 60. Google Scholar CrossRef Search ADS PubMed  13 Copas J, Shi JQ. Meta-analysis, funnel plots and sensitivity analysis. Biostatistics  2000; 1: 247– 62. Google Scholar CrossRef Search ADS PubMed  14 Kay GL, Morita S, Mendez M, Zubiate P, Kay JH. Tricuspid regurgitation associated with mitral valve disease: repair and replacement. Ann Thorac Surg  1989; 48: S93– 5. Google Scholar CrossRef Search ADS PubMed  15 McGrath LB, Gonzalez-Lavin L, Bailey BM, Grunkemeier GL, Fernandez J, Laub GW. Tricuspid valve operations in 530 patients. Twenty-five-year assessment of early and late phase events. J Thorac Cardiovasc Surg  1990; 99: 124– 33. Google Scholar PubMed  16 Guenther T, Noebauer C, Mazzitelli D, Busch R, Tassani-Prell P, Lange R. Tricuspid valve surgery: a thirty-year assessment of early and late outcome. Eur J Cardiothorac Surg  2008; 34: 402– 9. Google Scholar CrossRef Search ADS PubMed  17 Chen SW, Tsai FC, Tsai FC, Chao YK, Huang YK, Chu JJ et al.   Surgical risk and outcome of repair versus replacement for late tricuspid regurgitation in redo operation. Ann Thorac Surg  2012; 93: 770– 5. Google Scholar CrossRef Search ADS PubMed  18 Kim JB, Jung SH, Choo SJ, Chung CH, Lee JW. Surgical outcomes of severe tricuspid regurgitation: predictors of adverse clinical outcomes. Heart  2013; 99: 181– 7. Google Scholar CrossRef Search ADS PubMed  19 Raikhelkar J, Lin HM, Neckman D, Afonso A, Scurlock C. Isolated tricuspid valve surgery: predictors of adverse outcome and survival. Heart Lung Circ  2013; 22: 211– 20. Google Scholar CrossRef Search ADS PubMed  20 De Meester P, Van De Bruaene A, Voigt JU, Herijgers P, Budts W. Outcome and determinants of prognosis in patients undergoing isolated tricuspid valve surgery: retrospective single center analysis. Int J Cardiol  2014; 175: 333– 9. Google Scholar CrossRef Search ADS PubMed  21 Oh TH, Wang TK, Sidhu K, Haydock DA. Isolated tricuspid valve surgery at a single centre: the 47-year Auckland experience, 1965–2011. Interact CardioVasc Thorac Surg  2014; 18: 27– 32. Google Scholar CrossRef Search ADS PubMed  22 Jintao F, Qingliang C, Dong X, Feng Z, Zihou L, Nan J. Efficacies of tricuspid valve replacement versus repair for moderate-to-severe tricuspid regurgitation with right heart failure. Zhonghua Yi Xue Za Zhi  2015; 95: 1396– 400. Google Scholar PubMed  23 Moraca RJ, Moon MR, Lawton JS, Guthrie TJ, Aubuchon KA, Moazami N et al.   Outcomes of tricuspid valve repair and replacement: a propensity analysis. Ann Thorac Surg  2009; 87: 83-8. Google Scholar CrossRef Search ADS   24 Akins CW, Miller DC, Turina MI, Kouchoukos NT, Blackstone EH, Grunkemeier GL et al.   Guidelines for reporting mortality and morbidity after cardiac valve interventions. J Thorac Cardiovasc Surg  2008; 135: 732– 8. Google Scholar CrossRef Search ADS PubMed  25 Choi JW, Park EA, Lee W, Kim KH, Kim KB, Ahn H et al.   Changes in right ventricular volume and function after tricuspid valve surgery—Tricuspid annuloplasty vs. tricuspid valve replacement. Circ J  2016; 80: 1142– 7. Google Scholar CrossRef Search ADS PubMed  26 Bernal JM, Morales D, Revuelta C, Llorca J, Gutierrez-Morlote J, Revuelta JM. Reoperations after tricuspid valve repair. J Thorac Cardiovasc Surg  2005; 130: 498– 503. Google Scholar CrossRef Search ADS PubMed  27 Jeganathan R, Armstrong S, Al-Alao B, David T. The risk and outcomes of reoperative tricuspid valve surgery. Ann Thorac Surg  2013; 95: 119– 24. Google Scholar CrossRef Search ADS PubMed  © The Author 2017. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png European Journal of Cardio-Thoracic Surgery Oxford University Press

Repair versus replacement for the surgical correction of tricuspid regurgitation: a meta-analysis

Loading next page...
 
/lp/ou_press/repair-versus-replacement-for-the-surgical-correction-of-tricuspid-bQzyZcSIlV
Publisher
Oxford University Press
ISSN
1010-7940
eISSN
1873-734X
D.O.I.
10.1093/ejcts/ezx420
Publisher site
See Article on Publisher Site

Abstract

Abstract OBJECTIVES Tricuspid valve repair (TVr) has both a theoretical advantage in preserving right ventricular function and disadvantages such as a higher risk of repair failure in the long-term compared with tricuspid valve replacement (TVR). This study was conducted to compare the results of TVr with those of TVR to find an optimal treatment option. METHODS A literature search of 5 databases was performed. The primary outcome was all-cause mortality. Secondary outcomes were tricuspid reoperation and valve-related events. Subgroup analyses were performed according to the risk of bias, year of publication and proportions of patients with tricuspid regurgitation Grade ≥3, functional aetiology and isolated tricuspid valve surgery. Publication bias was explored using the funnel plot and Egger’s test. RESULTS Seventeen retrospective studies involving 4561 patients (TVr group = 3432 patients and TVR group = 1129 patients) were included. A pooled analysis showed that the risk of all-cause mortality was significantly higher in the TVR group than in the TVr group [hazard ratio (95% confidence interval) 1.59 (1.26–2.00)]. There were no significant differences in tricuspid valve reoperation in 6 studies and valve-related events in 5 studies between the TVR and TVr groups [hazard ratio (95% confidence interval) 1.30 (0.88–1.91) and 1.47 (0.91–2.38), respectively]. None of the subgroup analyses demonstrated a significant difference in the hazard ratio of all-cause mortality. No publication bias was identified for the primary and secondary outcomes. CONCLUSIONS This meta-analysis indicates that TVr is more beneficial compared with TVR in terms of all-cause mortality. From the available data, TVr is not associated with an increased risk of tricuspid reoperation compared with TVR. Valve disease surgery, Tricuspid valve disease, Meta-analysis INTRODUCTION Current guidelines recommend tricuspid valve (TV) surgery for (i) patients with severe primary tricuspid regurgitation (TR) accompanying symptoms or right heart failure and (ii) patients with severe or progressive functional TR who underwent left-sided valve surgery [1]. The surgical treatment of choice for patients with TR is, however, still controversial; although previous studies have demonstrated favourable long-term outcomes after TV repair (TVr) compared with TV replacement (TVR) [2–4], other studies have argued that worse clinical outcomes after TVR might be caused by worse baseline characteristics of TVR patients such as advanced heart failure and a history of frequent cardiac surgery [5, 6]. Therefore, this meta-analysis was conducted to compare results of TVr with those of TVR in patients with TR. This study hypothesized that despite a long-term risk of repair failure, long-term clinical outcomes, particularly survival rates, would be better after TVr than after TVR. MATERIALS AND METHODS Multiple comprehensive databases were used to find literature comparing the results of TVr with those of TVR. This systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [7]. Data source and literature searches Full-text articles were searched using MEDLINE, EMBASE, the Cochrane Central Register of Controlled Trials (CENTRAL), Web of Science and Scopus on 6 February 2017. No restrictions were placed on the language or year of publication. The following keywords and Medical Subject Headings (MeSH) were searched through Medline: ((((‘tricuspid valve’[MeSH Terms] OR ‘tricuspid valve insufficiency’[MeSH Terms]) OR ‘tricuspid valve prolapse’[MeSH Terms]) OR ((((tricuspid valve disease[Title/Abstract] OR tricuspid regurgitation[Title/Abstract]) OR tricuspid valve regurgitation[Title/Abstract]) OR tricuspid insufficiency[Title/Abstract]) OR tricuspid valve insufficiency[Title/Abstract])) AND (‘cardiac valve annuloplasty’[MeSH Terms] OR ((annuloplasty[Title/Abstract] OR valvuloplasty[Title/Abstract]) OR valve repair[Title/Abstract]))) AND (‘heart valve prosthesis implantation’[MeSH Terms] OR (replacement[Title/Abstract] OR implantation[Title/Abstract])) (Supplementary Material, Table S1). Search strategies were adapted for other databases based on the MEDLINE strategy. Study selection The inclusion of studies was independently decided by 2 reviewers (J.W.C. and H.Y.H.) based on the selection criteria. Any disagreements between the reviewers were resolved through discussion or review from the third author (M.-J.J.). The study selection was made following 2 levels of screening: at the first level, the titles and abstracts of the identified studies were screened. At the second level, the full texts were reviewed. Studies were included in this meta-analysis if they (i) enrolled patients undergoing TV surgery and (ii) compared the survival rates of patients receiving TVr with those receiving TVR as time-to-event data. When duplicated publications with overlapping study populations were found, the study with the largest sample size was selected. Data extraction Study characteristics and the patient baseline data were extracted independently by 2 reviewers (J.W.C. and H.Y.H.). Data regarding primary and secondary outcomes were also independently extracted by 2 reviewers (M.-J.J and H.Y.H.). Any disagreements between reviewers were resolved through discussion. If the outcome data were unclear, the authors were contacted via email. Assessment of study quality The overall study quality was assessed independently by two reviewers (M.-J.J. and H.Y.H.) using Risk Of Bias In Non-randomized Studies of Interventions (ROBINS-I) [8]. Seven domains of risk of bias (ROB) were assessed: bias due to confounding, bias in selection of participants into the study, bias in classification of interventions, bias due to departures from intended interventions, bias due to missing data, bias in measurement of outcomes and bias in selection of the reported result. Based on the assessment of each domain, an overall ROB was determined as low, moderate, serious or critical. Any disagreements between the reviewers were resolved through discussion. Statistical analysis The primary outcome was to investigate the association between TVR versus TVr and all-cause mortality in patients with TR. The secondary outcomes were time-to-TV reoperation and time-to-valve-related events (VREs). The outcomes were measured in terms of the hazard ratio (HR) of using TVR compared with TVr. For each study, the HR and 95% confidence interval (CI) were directly extracted from the research article. When the HR and its 95% CI were not reported, their values were calculated using other available statistical information (log-rank test, Kaplan–Meier curve) [9]. In 1 study [10], their values for all-cause mortality were drawn from both 30-day morality and the Kaplan–Meier curve for 30-day survivors. Pooled HRs, 95% CIs and P-values were calculated using the random-effect model. Two-sided P-values <0.05 were considered statistically significant. Heterogeneity was assessed with the χ2 tests and quantified with I2 statistics to evaluate the extent of variability that was attributed to statistical heterogeneity between studies. The I2 values of 25%, 50% and 75% have been suggested as indicators of low, moderate and high heterogeneity, respectively [11]. A sensitivity analysis was conducted to assess the influence of an individual study on the results by omitting 1 study in each turn [12]. Subgroup analyses were performed to evaluate the influences of ROB and patient characteristics on the study results; a stratified analysis was conducted based on the result of the ROB assessment (serious or moderate), and tertile-group comparisons were performed using all studies demonstrating available data according to the proportions of patients with TR Grade 3 or 4, functional TR and isolated TV surgery in each study. Additional subgroup analysis was performed by dividing studies according to the year of publication (before 2000 vs from 2000 to 2009 vs after 2010). A funnel plot and the Egger’s test for asymmetry were applied to assess the possibility of publication bias among the studies. All analyses were performed using STATA version 12.0 (Stata Corporation, College Station, TX, USA). RESULTS Identification of studies The database searches yielded 2942 articles. Of these, 2891 publications were excluded as it was clear from the title and abstract that they did not fulfil the selection criteria. For the remaining 51 articles, full manuscripts were obtained, and following the scrutiny of these manuscripts, 34 publications were excluded due to unmet inclusion criteria (n = 28), duplicated data (n = 5) or the inability to extract HRs (n = 1). Therefore, the total number of studies included in this review was 17 (Fig. 1) [2–6, 10, 13–23]. Figure 1: View largeDownload slide Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) flow diagram. Figure 1: View largeDownload slide Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) flow diagram. Study characteristics and patient population All studies were non-randomized and involved 4561 patients (3432 patients in the TVr group and 1129 patients in the TVR group). Nine studies [4, 5, 10, 14, 17, 19, 21, 22] exclusively included TR patients, and another 5 studies [3, 6, 15, 16, 20] included 85.2–98.9% of patients with TR. In 1 study [2], the authors demonstrated that only 77.6% of study patients had TR of Grade 3 or 4, while only 0.7% of patients had tricuspid stenosis of Grade 3 or 4. The proportion of patients with TR was unclear in the other 2 studies (Table 1) [13, 23]. The patients were, on average or median, in their late 40s to early 60s and were predominantly women (range 52.9–84.7%). In the TVr group, the use of a flexible band or a rigid ring was present in all but 1 study; 1596 (46.9%) patients among 3405 patients underwent ring annuloplasty. In the 1004 TVR patients for whom the type of prosthesis was presented, mechanical and bioprosthetic TVRs were performed in 311 (31.0%) patients and 693 (69.0%) patients, respectively (Table 2). Table 1: Study characteristics Study  Operative era  Country  Study population   Proportion of TR patients (%)  Follow-up   Total  TVr  TVR  Duration (years)a  Completeness (%)  Kratz et al. [10]  1972–1983  US  41  17  24    mean 7.3 (1–12)  100  Cohen et al. [11]  Unclear  US  17  6  11  100  mean 5.9 (0.2–14.7)  92  Kay et al. [14]  1975–1988  US  156  127  29  100  0–12    McGrath et al. [15]  1961–1987  US  530  351  179  92.3  mean 4.2 (0–26.3)  98.6  Bajzer et al. [2]  1988–1997  US  401  392  9  >77.6b  0–8.3    Singh et al. [3]  1979–2003  Canada  250  178  72  85.2c  5.2 ± 4.1  99  Guenther et al. [16]  1974–2003  Germany  416  310  106  90.6c  5.9 ± 6.3  97  Moraca et al. [23]  1986–2006  US  315  222  93    5.4 ± 5.4  100  Chen et al. [17]  2005–2010  Taiwan  77  44  33  100  2.7 ± 1.5  100  Marquis-Gravel et al. [6]  1977–2008  Canada  926  792  134  98.3  median 4.3  92  Kim et al. [18]  1997–2010  Korea  449  397  52  100  5.3 (0–17.5)  93.5  Raikhelkar et al. [19]  1998–2010  US  56  27  29  100c  2.3 (0–11.6)  95  De Meester et al. [20]  1995–2011  Belgium  92  33  59  98.9  0–17    Hwang et al. [4]  1996–2012  Korea  175  108  67  100  6.9 (0.3–17.7)  96  Oh et al. [21]  1965–2011  New Zealand  72  38  34  100  mean 13.7  100  Jintao et al. [22]  2003–2008  China  228  127  101  100  mean 5.4  96.1  Chang et al. [5]  1994–2012  Korea  360  263  97  100  6.8 ± 4.6  99.7  Study  Operative era  Country  Study population   Proportion of TR patients (%)  Follow-up   Total  TVr  TVR  Duration (years)a  Completeness (%)  Kratz et al. [10]  1972–1983  US  41  17  24    mean 7.3 (1–12)  100  Cohen et al. [11]  Unclear  US  17  6  11  100  mean 5.9 (0.2–14.7)  92  Kay et al. [14]  1975–1988  US  156  127  29  100  0–12    McGrath et al. [15]  1961–1987  US  530  351  179  92.3  mean 4.2 (0–26.3)  98.6  Bajzer et al. [2]  1988–1997  US  401  392  9  >77.6b  0–8.3    Singh et al. [3]  1979–2003  Canada  250  178  72  85.2c  5.2 ± 4.1  99  Guenther et al. [16]  1974–2003  Germany  416  310  106  90.6c  5.9 ± 6.3  97  Moraca et al. [23]  1986–2006  US  315  222  93    5.4 ± 5.4  100  Chen et al. [17]  2005–2010  Taiwan  77  44  33  100  2.7 ± 1.5  100  Marquis-Gravel et al. [6]  1977–2008  Canada  926  792  134  98.3  median 4.3  92  Kim et al. [18]  1997–2010  Korea  449  397  52  100  5.3 (0–17.5)  93.5  Raikhelkar et al. [19]  1998–2010  US  56  27  29  100c  2.3 (0–11.6)  95  De Meester et al. [20]  1995–2011  Belgium  92  33  59  98.9  0–17    Hwang et al. [4]  1996–2012  Korea  175  108  67  100  6.9 (0.3–17.7)  96  Oh et al. [21]  1965–2011  New Zealand  72  38  34  100  mean 13.7  100  Jintao et al. [22]  2003–2008  China  228  127  101  100  mean 5.4  96.1  Chang et al. [5]  1994–2012  Korea  360  263  97  100  6.8 ± 4.6  99.7  a Data were represented as mean ± standard deviation, median (range) or range if unspecified. b 77.6% of study patients had TR of Grade III or IV. c Including patients with mixed tricuspid stenoinsufficiency. TR: tricuspid regurgitation; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Table 2: Patients characteristics Study  Age (years), mean ± SD   Female (%)   Type of surgery   TR of Grade 3 or 4 (%)  Functional TR (%)  Isolated TV surgery (%)  Previous cardiac surgery (%)   TVr  TVR  TVr  TVR  Use of ring in TVr (%)  Mechanical valve in TVR (%)  TVr  TVR  Kratz et al. [10]  48.0 ± 14.5  63.4  41.2      56.1  7.3      Cohen et al. [11]  41 ± 21.9  53 ± 6.2  50  54.5  16.7  9.1  64.7    0  0  0  Kay et al. [14]  61 ± 11  69.2  0  37.9      0  9.6  McGrath et al. [15]  mean 56.9 (10.1–79.9)  79.2  6.8  13.4  75.7    3.2  42.2  55.9  Bajzer et al. [2]  mean 60  69.6  76.8  66.7  77.6      0  0  Singh et al. [3]  51 ± 15  51 ± 14  68.0  63.9  33.1  25.0    0    43.8  56.9  Guenther et al. [16]  60.6 ± 12.5  49.9 ± 11.3  64.8  79.2  13.9  35.9      12.0  44.2  43.4  Moraca et al. [23]  58.4 ± 16.4  52.2 ± 16.2  62.2  62.4  45.5  22.6    51.1  41.0  36.0  65.6  Chen et al. [17]  58.5 ± 12.4  52.9 ± 5.8  47.7  60.6  100  30.3    48.1  33.8  100  100  Marquis-Gravel et al. [6]  63 (54–70)a  57 (44–64)a  72.6  67.9  46.6  11.9  73.0  71.8  14.7      Kim et al. [18]  56.4 ± 12.3  54.6 ± 12.7  65.5  71.2  67.8  63.5  100  80.6  17.6  11.8  34.6  Raikhelkar et al. [19]  51.8 ± 18.2  60.2 ± 13  63.0  51.7    3.4    14.2  100  29.6  69.0  De Meester et al. [20]  56 ± 14  62.0  75.8  44.1  75.0  41.3  100  62.0  Hwang et al. [4]  57 ± 12  57 ± 10  70.4  71.6  28.7  44.8  100  100  22.9  30.6  89.6  Oh et al. [21]  48.9 ± 16  46.9 ± 17  63.2  79.4  55.3  20.6  84.3    100  65.8  73.5  Jintao et al. [22]  58 ± 9.2  57.9 ± 9.4  59.8  68.3  90.6    47.8    0  0  0  Chang et al. [5]  59 ± 12  54 ± 13  64.6  76.3  70.7  71.1  100  70.3  15.3  17.1  59.8  Study  Age (years), mean ± SD   Female (%)   Type of surgery   TR of Grade 3 or 4 (%)  Functional TR (%)  Isolated TV surgery (%)  Previous cardiac surgery (%)   TVr  TVR  TVr  TVR  Use of ring in TVr (%)  Mechanical valve in TVR (%)  TVr  TVR  Kratz et al. [10]  48.0 ± 14.5  63.4  41.2      56.1  7.3      Cohen et al. [11]  41 ± 21.9  53 ± 6.2  50  54.5  16.7  9.1  64.7    0  0  0  Kay et al. [14]  61 ± 11  69.2  0  37.9      0  9.6  McGrath et al. [15]  mean 56.9 (10.1–79.9)  79.2  6.8  13.4  75.7    3.2  42.2  55.9  Bajzer et al. [2]  mean 60  69.6  76.8  66.7  77.6      0  0  Singh et al. [3]  51 ± 15  51 ± 14  68.0  63.9  33.1  25.0    0    43.8  56.9  Guenther et al. [16]  60.6 ± 12.5  49.9 ± 11.3  64.8  79.2  13.9  35.9      12.0  44.2  43.4  Moraca et al. [23]  58.4 ± 16.4  52.2 ± 16.2  62.2  62.4  45.5  22.6    51.1  41.0  36.0  65.6  Chen et al. [17]  58.5 ± 12.4  52.9 ± 5.8  47.7  60.6  100  30.3    48.1  33.8  100  100  Marquis-Gravel et al. [6]  63 (54–70)a  57 (44–64)a  72.6  67.9  46.6  11.9  73.0  71.8  14.7      Kim et al. [18]  56.4 ± 12.3  54.6 ± 12.7  65.5  71.2  67.8  63.5  100  80.6  17.6  11.8  34.6  Raikhelkar et al. [19]  51.8 ± 18.2  60.2 ± 13  63.0  51.7    3.4    14.2  100  29.6  69.0  De Meester et al. [20]  56 ± 14  62.0  75.8  44.1  75.0  41.3  100  62.0  Hwang et al. [4]  57 ± 12  57 ± 10  70.4  71.6  28.7  44.8  100  100  22.9  30.6  89.6  Oh et al. [21]  48.9 ± 16  46.9 ± 17  63.2  79.4  55.3  20.6  84.3    100  65.8  73.5  Jintao et al. [22]  58 ± 9.2  57.9 ± 9.4  59.8  68.3  90.6    47.8    0  0  0  Chang et al. [5]  59 ± 12  54 ± 13  64.6  76.3  70.7  71.1  100  70.3  15.3  17.1  59.8  a Median with interquartile range. TR: tricuspid regurgitation; TV: tricuspid valve; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Quality of the included studies Regarding all-cause mortality, of the 17 eligible studies that were included, the appropriate methods to adjust for measured confounders were used in 7 studies, and those were rated as moderate ROB due to confounding. Propensity score matching was performed in 1 study [23], propensity score adjusted was performed in 2 studies [6, 18], inverse probability of treatment weighting was performed in 2 studies [4, 5] and multivariable Cox proportional hazard model was performed in 2 studies [2, 3]. The other 10 studies reported only the results of univariate analysis and were considered as a serious ROB (Supplementary Material, Table S2). Two studies demonstrated late mortality instead of all-cause mortality by excluding patients with early mortality [6] or 1-year mortality [19]. These 2 studies were found to have moderate and low ROB in the selection of study participants respectively, because the former study showed different operative mortality between the TVR and TVr groups (21% vs 13%), whereas the latter study demonstrated similar 1-year mortality between the 2 groups (21.4% vs 20%). Regarding secondary outcomes, for the TV reoperation, all the 6 studies were found to have serious ROB due to confounding, because they only reported the results from univariate analysis. For the VRE, all the 5 studies were rated as having a moderate ROB due to confounding (Supplementary Material, Table S3). All-cause mortality A pooled analysis of 17 studies in 4382 patients showed that the risk of all-cause mortality was significantly higher in the TVR group than in the TVr group (HR 1.59, 95% CI =  1.26–2.00; Fig. 2). There was a moderate-to-high heterogeneity between the studies (I2 = 62.0%, P < 0.001). A sensitivity analysis showed that 1 study by Singh et al. [3] had the greatest influence on the results: the pooled HR without this study was 1.42 (95% CI 1.19–1.69) with an I2 of 31.7% (P = 0.108; Fig. 3). Figure 2: View largeDownload slide The HR of all-cause mortality for TVR compared with TVr from 17 studies including 4382 patients. CI: confidence interval; HR: hazard ratio; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Figure 2: View largeDownload slide The HR of all-cause mortality for TVR compared with TVr from 17 studies including 4382 patients. CI: confidence interval; HR: hazard ratio; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Figure 3: View largeDownload slide Sensitivity analysis to identify the impact of each study on study results. CI: confidence interval; HR: hazard ratio. Figure 3: View largeDownload slide Sensitivity analysis to identify the impact of each study on study results. CI: confidence interval; HR: hazard ratio. The subgroup analysis according to ROB grade showed that this benefit was statistically significant for both the serious and the moderate ROB groups; the pooled HRs were 1.38 (95% CI 1.10–1.73) and 1.82 (95% CI 1.16–2.84) in the 10 studies with serious ROB and 7 studies with moderate ROB, respectively (Fig. 4). There were no statistically significant differences in the HRs between the subgroups when the subgroup analyses were performed according to the proportions of patients with a TR of Grade ≥3 (in 10 studies), patients with functional TR (in 10 studies), patients with isolated TV surgery (in 15 studies) and year of publication (P = 0.868, 0.401, 0.416 and 0.607, respectively; Fig. 5). All the subgroup analyses were also performed after excluding 1 study by Singh et al. [3]; there were still no statistically significant differences in the HRs between the subgroups. Figure 4: View largeDownload slide The HR of all-cause mortality for TVR compared with TVr from 10 studies showing the results of univariate analysis and 7 studies presenting the results of adjusted analysis. CI: confidence interval; HR: hazard ratio; ROB: risk of bias; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Figure 4: View largeDownload slide The HR of all-cause mortality for TVR compared with TVr from 10 studies showing the results of univariate analysis and 7 studies presenting the results of adjusted analysis. CI: confidence interval; HR: hazard ratio; ROB: risk of bias; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Figure 5: View largeDownload slide Subgroup analyses between the 3 groups based on the tertile values according to the proportions of patients with TR of Grade 3 or 4, functional TR, isolated tricuspid valve surgery and year of publication. CI: confidence interval; HR: hazard ratio; OP: operation; TR: tricuspid regurgitation; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Figure 5: View largeDownload slide Subgroup analyses between the 3 groups based on the tertile values according to the proportions of patients with TR of Grade 3 or 4, functional TR, isolated tricuspid valve surgery and year of publication. CI: confidence interval; HR: hazard ratio; OP: operation; TR: tricuspid regurgitation; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Tricuspid valve reoperation Seven studies [3, 6, 14–16, 18, 21] compared time-to-TV reoperation between the TVr and the TVR groups using univariate analysis. The HR could not be calculated from 1 study [18] in which non-significant differences (P = 0.15) in TV reoperation rates were demonstrated in the 3 groups of mechanical TVR, bioprosthetic TVR and TVr patients instead of in the TVR and the TVr groups; therefore, it was excluded from the analysis. A pooled analysis of 6 studies including 2350 patients demonstrated that there was no statistically significant difference in the risk of TV reoperation between the TVR and the TVr groups (HR 1.30, 95% CI 0.88–1.91; Fig. 6A). The I2 statistics showed a low-to-moderate heterogeneity among the studies (I2 = 27.5%, P = 0.229). Figure 6: View largeDownload slide The HR of (A) tricuspid valve reoperation and (B) valve-related events for TVR compared with TVr from 6 studies demonstrating the results of univariate analysis and 5 studies showing the results of adjusted analysis, respectively. CI: confidence interval; HR: hazard ratio; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Figure 6: View largeDownload slide The HR of (A) tricuspid valve reoperation and (B) valve-related events for TVR compared with TVr from 6 studies demonstrating the results of univariate analysis and 5 studies showing the results of adjusted analysis, respectively. CI: confidence interval; HR: hazard ratio; TVR: tricuspid valve replacement; TVr: tricuspid valve repair. Valve-related events The VRE rates were demonstrated in 5 studies [2–5, 18]. All the HRs were derived from the adjusted analysis, including inverse probability of treatment weighting-adjusted HRs in 2 studies [4, 5], propensity score-adjusted HRs in 1 study [18] and HRs from the Cox proportional hazard model in 2 studies [2, 3]. Three studies [3–5] demonstrated VRE according to the current guidelines [24]; however, 2 of the 3 studies included cardiac mortality [4] or all-cause mortality [3] instead of valve-related mortality. The VRE was defined as death, TV reoperation and readmission due to congestive heart failure in another study [18] and as death, recurrence of symptoms that prompted the initial TV surgery, the occurrence cardinal symptoms attributed to congestive heart failure and TV repair failure that led to a second TV surgery in the other study [2]. A pooled analysis of 5 studies including 1635 patients showed that there was no statistically significant difference in the risk of VRE between the TVR and TVr groups (HR 1.47, 95% CI 0.91–2.38; Fig. 6B). A pooled analysis of 3 studies including 785 patients in which VRE was defined according to the guidelines also showed a non-significant difference in VRE between the 2 groups (HR 1.27, 95% CI 0.81–1.99). Publication bias There was no evidence of publication bias in the funnel plots of all-cause mortality, TV reoperation and VRE (Supplementary Material, Fig. S1). The Egger’s tests also showed that publication bias was non-significant in each outcome (P = 0.150, 0.320 and 0.533, respectively). DISCUSSION This meta-analysis demonstrated 2 main findings. First, TVR was associated with increased all-cause mortality compared with TVr in patients with TR. Second, there were no differences in morbidity, such as TV reoperation and VRE, between patients who underwent TVR and those who underwent TVr. The current guidelines suggest several indications for TV surgery in patients with TR, such as symptomatic severe primary TR, asymptomatic severe primary TR with right ventricular dysfunction and functional TR of a severe degree or progressive nature during left-sided valve surgery [1]. Even in these guidelines, TVr is recommended as the treatment of choice only for patients who undergo TV surgery combined with a left-sided valve surgery for mild or moderate functional TR in a progressive nature. A theoretical disadvantage of TVR over TVr has been suggested to be progressive right ventricular dysfunction caused by a large, rigid prosthesis in a deformable, low-pressure cavity [3]. A recent study demonstrated, using cardiac magnetic resonance study, that right ventricular ejection fraction decreased after TVR, but it was preserved after TVr [25]. However, there have been controversial results regarding the long-term clinical outcomes after TVR versus TVr, and no definite conclusions could be drawn [5, 6]. To the best of our knowledge, this is the first meta-analysis comparing the results of TVR with those of TVr. This study clearly demonstrated that TVR significantly increased all-cause mortality after surgery compared with TVr. The subgroup analyses demonstrated that this finding was consistent regardless of the severity of TR, the proportion of patients with functional TR and the proportion of patients with isolated TV surgery. In the pooled analysis for all-cause mortality, I2 statistics demonstrated a significant heterogeneity among studies with an I2 value of 66.0%. Although no statistically significant differences in the HRs were found in the subgroup analyses as demonstrated above, sensitivity analyses showed that this heterogeneity among studies was significantly affected in 1 study by Singh et al. [3]. Possible explanations for this included the following: (i) all patients included in that study had organic TV disease and (ii) a substantial proportion (35%) of patients had congenital heart disease. Two of the 17 studies demonstrated late mortality instead of all-cause mortality by excluding early mortality [6] or 1-year mortality [19]. The operative mortality after TVR was significantly higher than that after TVr (21% vs 13%) in the former study [6], and this might attenuate the true HR of all-cause mortality in that study. One of the major concerns when performing TVr is a risk of recurrent TR after repair, because the subsequent TV reoperation is still a high-risk procedure [26, 27]. However, a pooled analysis of 6 studies demonstrated that the risk of TV reoperation was not higher in the TVr group than in the TVR group. In addition, the risk of VRE was not significantly different between the 2 groups. However, further data might be needed to draw definite conclusions, because the HR for reoperation was drawn from only 6 of the 17 studies using the results of univariate analyses, and the HR of VRE was calculated using 5 studies in which the VREs were defined differently. Limitations There are limitations to this study that must be acknowledged. First, all of the included studies were non-randomized studies, and the confounding variables could affect the results of our analysis. Second, the definition of events and the duration of follow-up were different among studies. Finally, a publication bias could not be ruled out, because meta-analyses can only report on the outcomes of published studies. However, we tried to minimize the publication bias by not placing a restriction on the language or year of publication. CONCLUSION This meta-analysis indicates that TVr is more beneficial compared with TVR regarding all-cause mortality. From the available data, TVr is not associated with an increased risk of tricuspid reoperation compared with TVR. Surgeons should try to repair the valve regardless of the aetiology or severity when surgical correction is needed in patients with TR. SUPPLEMENTARY MATERIAL Supplementary material is available at EJCTS online. ACKNOWLEDGEMENTS The authors greatly appreciate Dr Keitaro Okamoto's (Faculty of Medicine, Oita University, Oita, Japan) contribution in reviewing the articles that were written in Japanese. The authors also thank Drs Pieter De Meester, Werner Budts, Pyo Won Park and Hyoung Woo Chang for confirming the data regarding their papers via mail. Conflict of interest: none declared. REFERENCES 1 Nishimura RA, Otto CM, Bonow RO, Carabello BA, Erwin JP3rd, Guyton RA et al.   2014 AHA/ACC guideline for the management of patients with valvular heart disease: executive summary: a report of the American College of Cardiology/American Heart Association task force on practice guidelines. J Am Coll Cardiol  2014; 63: 2438– 88. Google Scholar CrossRef Search ADS PubMed  2 Bajzer CT, Stewart WJ, Cosgrove DM, Azzam SJ, Arheart KL, Klein AL. Tricuspid valve surgery and intraoperative echocardiography. J Am Coll Cardiol  1998; 32: 1023– 31. Google Scholar CrossRef Search ADS PubMed  3 Singh SK, Tang GHL, Maganti MD, Armstrong S, Williams WG, David TE et al.   Midterm outcomes of tricuspid valve repair versus replacement for organic tricuspid disease. Ann Thorac Surg  2006; 82: 1735– 41. Google Scholar CrossRef Search ADS PubMed  4 Hwang HY, Kim KH, Kim KB, Ahn H. Treatment for severe functional tricuspid regurgitation: annuloplasty versus valve replacement. Eur J Cardiothorac Surg  2014; 46: e21– 7. Google Scholar CrossRef Search ADS PubMed  5 Chang HW, Jeong DS, Cho YH, Sung K, Kim WS, Lee YT et al.   Tricuspid valve replacement vs. repair in severe tricuspid regurgitation. Circ J  2017; 81: 330– 8. Google Scholar CrossRef Search ADS PubMed  6 Marquis-Gravel G, Bouchard D, Perrault LP, Page P, Jeanmart H, Demers P et al.   Retrospective cohort analysis of 926 tricuspid valve surgeries: clinical and hemodynamic outcomes with propensity score analysis. Am Heart J  2012; 163: 851– 8. Google Scholar CrossRef Search ADS PubMed  7 Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Int J Surg  2010; 8: 336– 41. Google Scholar CrossRef Search ADS PubMed  8 Sterne JA, Hernan MA, Reeves BC, Savovic J, Berkman ND, Viswanathan M et al.   Robins-I: A tool for assessing risk of bias in non-randomised studies of interventions. BMJ  2016; 355: i4919. Google Scholar CrossRef Search ADS PubMed  9 Tierney JF, Stewart LA, Ghersi D, Burdett S, Sydes MR. Practical methods for incorporating summary time-to-event data into meta-analysis. Trials  2007; 8: 16. Google Scholar CrossRef Search ADS PubMed  10 Kratz JM, Crawford FA, Stroud MR, Appleby DC, Hanger KH. Trends and results in tricuspid valve surgery. Chest  1985; 88: 837– 40. Google Scholar CrossRef Search ADS PubMed  11 Cohen SR, Sell JE, McIntosh CL, Clark RE. Tricuspid regurgitation in patients with acquired, chronic, pure mitral regurgitation. II. Nonoperative management, tricuspid valve annuloplasty, and tricuspid valve replacement. J Thorac Cardiovasc Surg  1987; 94: 488– 97. Google Scholar PubMed  12 Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ  2003; 327: 557– 60. Google Scholar CrossRef Search ADS PubMed  13 Copas J, Shi JQ. Meta-analysis, funnel plots and sensitivity analysis. Biostatistics  2000; 1: 247– 62. Google Scholar CrossRef Search ADS PubMed  14 Kay GL, Morita S, Mendez M, Zubiate P, Kay JH. Tricuspid regurgitation associated with mitral valve disease: repair and replacement. Ann Thorac Surg  1989; 48: S93– 5. Google Scholar CrossRef Search ADS PubMed  15 McGrath LB, Gonzalez-Lavin L, Bailey BM, Grunkemeier GL, Fernandez J, Laub GW. Tricuspid valve operations in 530 patients. Twenty-five-year assessment of early and late phase events. J Thorac Cardiovasc Surg  1990; 99: 124– 33. Google Scholar PubMed  16 Guenther T, Noebauer C, Mazzitelli D, Busch R, Tassani-Prell P, Lange R. Tricuspid valve surgery: a thirty-year assessment of early and late outcome. Eur J Cardiothorac Surg  2008; 34: 402– 9. Google Scholar CrossRef Search ADS PubMed  17 Chen SW, Tsai FC, Tsai FC, Chao YK, Huang YK, Chu JJ et al.   Surgical risk and outcome of repair versus replacement for late tricuspid regurgitation in redo operation. Ann Thorac Surg  2012; 93: 770– 5. Google Scholar CrossRef Search ADS PubMed  18 Kim JB, Jung SH, Choo SJ, Chung CH, Lee JW. Surgical outcomes of severe tricuspid regurgitation: predictors of adverse clinical outcomes. Heart  2013; 99: 181– 7. Google Scholar CrossRef Search ADS PubMed  19 Raikhelkar J, Lin HM, Neckman D, Afonso A, Scurlock C. Isolated tricuspid valve surgery: predictors of adverse outcome and survival. Heart Lung Circ  2013; 22: 211– 20. Google Scholar CrossRef Search ADS PubMed  20 De Meester P, Van De Bruaene A, Voigt JU, Herijgers P, Budts W. Outcome and determinants of prognosis in patients undergoing isolated tricuspid valve surgery: retrospective single center analysis. Int J Cardiol  2014; 175: 333– 9. Google Scholar CrossRef Search ADS PubMed  21 Oh TH, Wang TK, Sidhu K, Haydock DA. Isolated tricuspid valve surgery at a single centre: the 47-year Auckland experience, 1965–2011. Interact CardioVasc Thorac Surg  2014; 18: 27– 32. Google Scholar CrossRef Search ADS PubMed  22 Jintao F, Qingliang C, Dong X, Feng Z, Zihou L, Nan J. Efficacies of tricuspid valve replacement versus repair for moderate-to-severe tricuspid regurgitation with right heart failure. Zhonghua Yi Xue Za Zhi  2015; 95: 1396– 400. Google Scholar PubMed  23 Moraca RJ, Moon MR, Lawton JS, Guthrie TJ, Aubuchon KA, Moazami N et al.   Outcomes of tricuspid valve repair and replacement: a propensity analysis. Ann Thorac Surg  2009; 87: 83-8. Google Scholar CrossRef Search ADS   24 Akins CW, Miller DC, Turina MI, Kouchoukos NT, Blackstone EH, Grunkemeier GL et al.   Guidelines for reporting mortality and morbidity after cardiac valve interventions. J Thorac Cardiovasc Surg  2008; 135: 732– 8. Google Scholar CrossRef Search ADS PubMed  25 Choi JW, Park EA, Lee W, Kim KH, Kim KB, Ahn H et al.   Changes in right ventricular volume and function after tricuspid valve surgery—Tricuspid annuloplasty vs. tricuspid valve replacement. Circ J  2016; 80: 1142– 7. Google Scholar CrossRef Search ADS PubMed  26 Bernal JM, Morales D, Revuelta C, Llorca J, Gutierrez-Morlote J, Revuelta JM. Reoperations after tricuspid valve repair. J Thorac Cardiovasc Surg  2005; 130: 498– 503. Google Scholar CrossRef Search ADS PubMed  27 Jeganathan R, Armstrong S, Al-Alao B, David T. The risk and outcomes of reoperative tricuspid valve surgery. Ann Thorac Surg  2013; 95: 119– 24. Google Scholar CrossRef Search ADS PubMed  © The Author 2017. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.

Journal

European Journal of Cardio-Thoracic SurgeryOxford University Press

Published: Apr 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Unlimited reading

Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.

Stay up to date

Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.

Organize your research

It’s easy to organize your research with our built-in tools.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve Freelancer

DeepDyve Pro

Price
FREE
$49/month

$360/year
Save searches from
Google Scholar,
PubMed
Create lists to
organize your research
Export lists, citations
Read DeepDyve articles
Abstract access only
Unlimited access to over
18 million full-text articles
Print
20 pages/month
PDF Discount
20% off