Article Text
PDF
Abstract
Aim Patients with severe mitral stenosis (MS) and their clinicians typically choose percutaneous transvenous mitral commissurotomy (PTMC) over surgical commissurotomy (SC). However, the durability of PTMC relative to SC is uncertain. We compared the efficacy, safety and durability of PTMC with SC for the treatment of MS.
Methods We searched EMBASE, MEDLINE and WHO ICTRP registers for randomised controlled trials (RCTs) comparing PTMC, and open and/or closed mitral commissurotomy. The principal outcomes were rate of re-intervention and symptomatic improvement as inferred from the surrogate measures of immediate postprocedural mitral valve area (MVA), MVA at ≥6 month follow-up, incidence of mitral regurgitation (MR) and restenosis. We calculated weighted mean differences (WMD) for continuous outcomes, relative risks (RR) for binary outcomes and pooled outcomes using random-effects models and assessed the quality of evidence using the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach.
Results Seven RCTs with 553 patients proved eligible. Pooled estimates showed no convincing difference in the risk of restenosis or re-intervention (15/100 fewer with PTMC, 95% CI (−20 to +8); quality of evidence: moderate) or in symptoms as inferred from immediate MVA (WMD 0.15, 95% CI (−0.18 to 0.48): very low), from the incidence of postprocedural severe MR (3/100 more with PTMC, 95% CI (−1 to +10): moderate) or from MVA at 30 months.
Conclusion Until data demonstrating convincing superiority of SC over PTMC become available, our results support the current practice of recommending PTMC to young patients with MS and favourable valve morphology, as it is associated with lower peri-procedural morbidity.
PROSPERO registration number PROSPERO 2017 (CRD42017079512).
- mitral stenosis
- transcatheter valve interventions
- cardiac procedures and therapy
- valvular heart disease
Statistics from Altmetric.com
- mitral stenosis
- transcatheter valve interventions
- cardiac procedures and therapy
- valvular heart disease
Introduction
Rheumatic heart disease (RHD), a sequelae of acute rheumatic fever (ARF) is an important preventable cause of cardiovascular morbidity and mortality in the developing world.1 There are an estimated 33.4 million RHD cases globally.2 A recent multi-centre, hospital-based study of over three thousand patients with clinical RHD showed that these patients are in poor functional class and have a high mortality rate (2-year case fatality rate of 16.9%).3 Nearly half the patients with RHD presenting to hospitals in developing countries have mitral stenosis (MS), and about 73% of these have moderate to severe stenosis.4 MS, when untreated, often leads to atrial fibrillation (AF), congestive heart failure (CHF) and stroke.5
The treatment of rheumatic MS involves mechanically relieving mitral valve obstruction, either by surgical or percutaneous means. Medical therapy is restricted to the management of complications such as AF and CHF, and secondary prophylaxis for the prevention of recurrent ARF.6 Interventions aimed at widening the mitral valve orifice and diminishing the trans-valvular gradients slow the natural progression of MS.7 Surgical open mitral commissurotomy (OMC), closed mitral commissurotomy (CMC) and percutaneous transvenous mitral commissurotomy (PTMC) are the principal methods of treating MS. PTMC is likely to confer less procedural risk than surgical commissurotomy (SC), and is associated with a shorter duration of hospitalisation, less up-front costs and is generally the preferred form of treatment.8 9 However, despite similar initial treatment outcomes following PTMC, the durability of results over time, when compared with surgical methods has not been clearly established.
A systematic review performed in 2011 compared SC and PTMC for immediate and long-term outcomes.10 This review demonstrated a higher risk of mitral regurgitation (MR) and re-intervention with PTMC and the authors concluded that surgery was the superior option. That conclusion is inconsistent with current recommendations.8 This review was, however, limited by the inclusion of quasi-randomised studies, and did not compare the two surgical modalities (OMC and CMC) individually with PTMC. Unlike CMC, OMC allows for direct visualisation of the valve apparatus and the immediate assessment of anatomical and functional results.11 This differentiation between the surgical techniques may influence the outcomes. Further, this prior review did not rate the quality of evidence.12
We therefore performed a systematic review and meta-analyses of all the available randomised controlled trials (RCTs) comparing PTMC and SC with the objective of determining the relative efficacy, safety and durability of the procedures. We assessed the quality of evidence using the GRADE approach in order to provide evidence-based guidance for clinical practice and further research.12
Methods
Search strategy
We searched MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials and trial registers (ClinicalTrials.gov, ISRCTN Register and WHO ICTRP) up to August 2018, for randomised trials comparing PTMC and SC. We used the key words ‘(percutaneous, transvenous, balloon valvuloplasty, commissurotomy or valvotomy) and (mitral stenosis or mitral valve stenosis)’. The search strategy is detailed in the online supplementary appendix 1. No language or publication date restrictions were imposed. The review was conducted and is reported according to the Preferred Reporting Items for Systematic Reviews And Meta-Analysis guidelines.13 The study protocol is registered in PROSPERO.14
Supplemental material
Study screening and selection
Two reviewers (ADS and AM) independently screened titles and abstracts of all retrieved articles and selected potentially eligible studies. Full-text articles of all relevant studies were retrieved and assessed for inclusion based on the following eligibility criteria: RCTs of patients (>12 years of age) with severe MS; comparing PTMC with either surgical OMC or CMC, or both, and reporting at least one of the principal outcomes of interest. Traditionally, the efficacy of mitral valve interventions is inferred from the surrogate measures of mitral valve area (MVA) and mitral valve gradient (MVG) following the procedures. An improvement in either of these measures may result in improvement in symptoms and heart failure. While evidence suggests this is generally true for large increases in MVA, or increases in MVA to ≥1.5 cm2, the significance of smaller differences for symptoms remains uncertain. Other outcomes that are of more direct importance to patients are the occurrence of severe MR postprocedure (which almost invariably needs valve replacement surgery), and the need for re-intervention due to symptomatic restenosis.
We categorised outcomes as those measured immediately following the procedure and at follow-up. The immediate outcomes were: (1) percentage increase in the MVA; (2) final MVA in cm2; (3) postprocedure MVG in mm Hg; (4) new onset severe and non-severe MR following the intervention; (5) incidence of successful outcome following the procedures. Outcomes measured at follow-up (≥6 months) were, (1) MVA and MVG on echocardiography; (2) mitral restenosis and/or re-intervention, (3) persistent MR. All-cause mortality at any time following the procedure was a secondary outcome. A successful procedure was defined as an immediate postprocedural MVA of ≥1.5 cm2 or ≥50% increase from the preprocedural MVA, with no more than grade 2 MR after the procedure.7 8 Studies were excluded on the following grounds: non-randomised design, absence of an active comparator arm (OMC or CMC) or if none of the outcomes of interest were reported.
Data extraction
Two reviewers (ADS and AM) extracted data in duplicate using prepiloted forms in Microsoft Excel including study and population characteristics, procedural details, key outcomes, mode, duration and completeness of follow-up, incidence of complications, ongoing drug therapy and any other additional co-morbidities. Reviewers extracted the main outcomes of MVA, restenosis, mortality, NYHA status, MR and complications at all time points available and grouped outcomes as—immediate postprocedure, or intermediate (≥6 months). In the event of missing outcome data, additional information was requested from the authors of the papers. Authors’ failure to respond after two contact attempts over a 2-week period, the data were considered as unavailable. Reviewers resolved disagreements in study selection and data extraction through discussion (ADS, AM) and if necessary by consultation with the corresponding author (GK).
Quality assessment
Each reviewer assessed risk of bias of the included studies independently using the Cochrane Risk of Bias 2.0 tool.15 This tool assigns risk of bias to RCTs at three levels: low risk, some concerns and high risk in each of six domains: randomisation process, deviation from intended intervention, missing outcome data, measurement of the outcome, selection of the reported result and overall bias. The biases were assessed for individual outcomes. The GRADE approach that involves study design (RCTs and non-randomised studies) and consideration of risk of bias, directness of evidence, consistency, precision and publication bias informed assessment of the quality of evidence for each outcome.12 The ‘Summary of findings’ table and ‘Evidence profile’ were constructed using the GRADEpro Guideline Development Tool.16
Statistical analysis
We calculated pooled weighted mean differences (WMD) for the outcomes of MVA and MVG and crude relative risks (RR) for the categorical outcomes: MR, mitral restenosis and re-intervention, using DerSimonian and Laird random effects models for pooling. We calculated absolute risk differences (RD) for the categorical outcomes from the relative risk estimates using the formula RD=ESC RR×ESC , where ESC is the event rate in the surgical arm expressed as the number of events per 100 patients. We assessed between study heterogeneity using χ2 and I2 statistics, with an I2 >50% implying substantial heterogeneity and performed subgroup analyses to identify and explain potential sources of heterogeneity. A priori, we postulated that differences in the techniques used to perform PTMC (Inoue, double-balloon, or others), methods used to assess the postprocedural valve area (catheterisation or echocardiography), or the surgical techniques used (OMC or CMC) might explain heterogeneity anticipating that studies assessing MVA by echocardiography may favour PTMC and that open commissurotomy may prove superior to CMC in comparison to PTMC. Analyses were performed using Stata V.15.
Results
Study selection
Our search yielded 210 articles, of which 26 proved eligible for full-text review and seven ultimately proved eligible.17–23 An updated search in August 2018 did not retrieve any additional studies. The author of the study by Cordoso et al was unsuccessfully contacted for unpublished data regarding the severity of MR in the surgical group. Figure 1 summarises the search process. In case of multiple publications18 24 25 by the same group of authors involving the same group of patients, the report with more complete information was included.18
PRISMA flow diagram. PRISMA, Preferred Reporting Items for Systematic Reviews And Meta-Analysis.
Characteristics of eligible studies
In the seven eligible studies, all single-centre, 261 patients were treated with PTMC and 292 with surgery (192 patients—CMC, 100 patients—OMC). One study contributed one-third of the patients.17 Table 1 presents the patient and study characteristics. One study evaluated only immediate outcomes.19 Two studies compared PTMC to OMC,18 20 four studies compared PTMC to CMC17 19 22 23; one compared all the three interventions.21
Study characteristics of the randomised controlled trial
Included patients were almost all under 40, with a female preponderance, usually presented in New York Heart Association (NYHA) functional classes 3 and 4, and were followed from 7 to 99 months. Most studies used the Wilkins score to assess baseline anatomical features of the mitral valve.26
No study reported change from baseline MVA, or adjusted RRs or OR for the outcomes of interest. Publication bias could not be assessed formally as we had fewer than 10 studies.
Risk of bias
Most of the studies did not report the randomisation and allocation concealment methods17–19 23 leading to ‘some concerns’ for the randomisation process domain. Outcome assessment using an operator-dependent method (echocardiography) by non-blinded researchers also led to ‘some concerns’ for the measurement of the primary outcome.19 None of the trials were registered prospectively, or had a published protocol, so it was not possible to assess selective reporting of results. Most of the included studies had ‘some concerns’ with regard to the overall bias. Online supplementary table 1 presents the risk of bias assessment.
Outcomes
Immediate outcomes
Six studies reported the immediate postprocedural MVA.17–20 22 23 The mean MVA (figure 2) achieved immediately after the procedure was similar between PTMC and SC (WMD 0.15, 95% CI (−0.18 to 0.48)). There was high statistical heterogeneity (I2=88.8%) between the studies reporting this outcome. The immediate mean MVG (online supplementary figure 1) was significantly lower in the patients undergoing PTMC than SC (WMD −1.12, 95% (CI −1.76 to 0.48); I2=0%). The incidence of severe MR immediately (figure 3) following the procedure differed between groups but CIs proved very wide (4.1% with PTMC vs 1.3% with SC, RR 2.12, 95% CI (0.50 to 8.92); I2=0%). This translates to a RD of three cases of severe MR per 100 more with PTMC (95% CI (1 fewer to 10 more)). The incidence of non-severe MR immediately following the procedures (figure 4) was similar between the groups (15.8% with PTMC vs 14.3% for SC, RR 1.16, 95% CI (0.75 to 1.81)). Peri-procedural morbidity (severe bleeding needing transfusion or re-operation, tachycardia and bradycardia) was greater with SC (table 2). Data regarding the improvement in NYHA class and clinical features of heart failure following the procedures was not consistently reported and could not be compared. Table 3 summarises these results and presents the quality of evidence in relation to inferences regarding symptoms of heart failure.
Forest plot of mitral valve areas after the procedure. CMC, closed mitral commissurotomy; OMC, open mitral commissurotomy; PTMC, percutaneous transvenous mitral commissurotomy; WMD, weighted mean differences.
Forest plot of severe mitral regurgitation immediately after the procedures. PTMC, percutaneous transvenous mitral commissurotomy; RR, relative risks.
Forest plot of non-severe mitral regurgitation after the procedures. PTMC, percutaneous transvenous mitral commissurotomy; RR, relative risks.
Immediate and late complications reported in the included studies
GRADE summary of findings table
Intermediate-term outcomes
Six studies reported outcomes at a median follow-up of 30 months (7–99 months).17 18 20–23 There was no statistically significant difference in any of the assessed outcomes: MVA (figure 2) (WMD 0.13, 95% CI (−0.09 to 0.35)); I2=81.2%), non-severe MR (figure 4), mitral restenosis (figure 5), re-intervention (figure 5) (8.3% with PTMC vs 23.3%, with SC, RR 0.42, 95% CI (0.13 to 1.34); I2=27%, an absolute reduction of 15% with PTMC (95% CI (−20% less to 8% more)), or mortality (online supplementary figure 2) across the groups. Improvement in functional class and clinical heart failure were not consistently reported. Table 3 summarises these results and presents the quality of evidence in relation to inferences regarding symptoms of heart failure.
Forest plot of mitral restenosis and mitral re-intervention after intermediate follow-up. PTMC, percutaneous transvenous mitral commissurotomy; RR, relative risks.
Subgroup analyses
Subgroup analyses were performed to explain the high between-study heterogeneity observed for both immediate and intermediate MVA outcomes. The heterogeneity in the differences between SC and PTMC was partly explained by the type of surgery, with OMC faring better than PTMC (while CMC fared worse than PTMC) (figure 2). However, these results were based on the results of just one study, and the difference was not observed at intermediate follow-up (figure 2 lower panel). The mean WMDs tended to favour PTMC when MVA was assessed by cardiac catheterisation19–21 both immediately after the procedure and at intermediate follow-up, compared with when echocardiography was used17 18 22 23 (online supplementary figure 3). Again the differences were accounted for by the results of a single study. Outcomes did not differ between studies with different follow-up durations (online supplementary figure 4). Due to small numbers of patients, we were unable to analyse the effect of different PTMC techniques on outcomes. Moreover, some of the included studies reported using multiple techniques (table 1).
Quality of evidence
All outcomes were rated down for serious imprecision. We rated the quality of evidence as moderate for the severe MR and re-intervention outcomes. The remaining outcomes (improvement in valve area and non-severe MR) are surrogates for patient-important outcomes of functional class and heart failure, and therefore were also rated down for indirectness, and inconsistency due to substantial heterogeneity. The quality of evidence for individual outcomes is detailed in the ‘Summary of findings table’ (table 3) and the ‘Evidence profile’ in online supplementary table 2.
Discussion
This systematic review and meta-analysis failed to demonstrate differences between immediate and intermediate-term MVAs achieved by PTMC compared with either of the surgical modalities. The complication rates, restenosis and re-intervention rates between the two groups differed substantially in relative—though not in absolute—terms, but with very small numbers of events, CIs were very wide, and chance could easily explain the differences. The quality of evidence regarding symptomatic improvement proved low due to significant imprecision, inconsistency and indirectness.
The strengths of this systematic review include protocol prespecification and registration27 duplicate eligibility assessment, data extraction and the assessment of risk of bias. We also rated the quality of evidence using the GRADE approach with particular attention to limitations of surrogate outcomes.
One previously published systematic review and meta-analysis compared PTMC and SC in severe MS.10 This meta-analysis by Hu et al excluded an RCT22 assuming—incorrectly—that it was a duplicate publication of Reyes et al.20 Hu et al also included a non-randomised study in their analysis.28 They found significantly higher rates of MR (pooled RR 1.66, 95% CI (1.08 to 2.58); I2=0%) and re-intervention (pooled RR 2.88, 95% CI (1.97 to 4.2); I2=0%) following PTMC as compared with SC. We found none of the differences immediately post-procedure or at intermediate-term follow-up, largely because of our restriction to randomised trials and thus exclusion of an observational study (n=580).
The inconsistency in results may be due to the differences in the surgical method used (OMC or CMC), or due to methods used to assess MVA and MVG (catheterisation or echocardiography). However, our analysis was underpowered to definitively identify the reasons for the heterogeneity. The included studies did not have a consistent definition of success for the tested procedures. Change in the MVA from the baseline (combined with symptomatic improvement) is arguably the best measure of the efficacy of interventions for MS in the short term. However, except for the study by Ben Farhat et al,21 no other study reported this measure. These authors found that both PTMC and OMC resulted in an absolute increase in MVA at 6 months, which was significantly greater than that produced by CMC (mean change 1.3 cm2 vs 0.7 cm2; p<0.01). Symptomatic improvement (NYHA class) was not consistently reported in any of the included studies.
Restenosis and re-intervention rates at 30 months were low (8.3%) and did not differ between the two groups. The lower re-intervention rates in RCTs may reflect the inclusion of younger patients with favourable valve morphology. Long-term observational studies including older patients, with valve morphology less favourable for PTMC, have indicated that re-intervention rates may be considerably higher.29 30 Jneid et al reported an annualised rate of re-do PTMC or mitral valve replacement of 20.7% among older patients (average age 52 years).29 Likewise, Wang et al found that the 5-year restenosis rate among patients with an echo score >8 was 61% compared with 20% among those with an echo score ≤8. Moreover, unsuccessful PTMC (which was rare in the included RCTs) was seen in a quarter to two-thirds of older, all-comer populations.30 Therefore caution is required in extrapolating the results of this meta-analysis to a contemporary all-comer population.31
Complications of PTMC and SC
Complication rates were low and similar between the two groups. The incidence of immediate and intermediate-term MR (of any severity) were similar in both the groups, based on moderate quality evidence. Various studies reported severe MR differently. But we nevertheless pooled the results for severe MR, as it has immediate implications for treatment and prognosis. Severe MR is reported in about 1.4%–3% of the patients undergoing PTMC.32 In studies with long-term follow-up included in this review, the prevalence of non-severe MR reduced over follow-up, consistent with previous reports32 33 suggesting that milder degrees of MR following PTMC may regress over time. Other known complications of PTMC such as embolism and cardiac tamponade were infrequently observed precluding meaningful comparisons.
Study limitations
Limitations were largely those of the existing evidence, namely the small number of available trials, the inconsistent assessment of outcomes, the small number of patients and the limited follow-up duration that was, on average, about 30 months. Differences in restenosis and re-intervention rates may emerge over a longer time-frame (>5 years).34 Finally, the randomised trials in this review enrolled highly selected patients (young patients with favourable valve morphology). Therefore, these results may not apply to more diverse populations undergoing PTMC.
Implications for further research
Procedural success may be lower, and the rates of restenosis and re-intervention may be higher in contemporary older patients with MS undergoing PTMC. Adequately powered randomised studies comparing SC with PTMC in ‘all-comers’ with severe MS are needed before strong recommendations for preferring one modality over the other can be made. Assessment of MVA should be performed consistently, in both arms and during all follow-up time points, using the same modality. We also suggest that the change in MVA from baseline to postprocedure, or at follow-up, be used as the primary measure of efficacy, and that patients be followed-up for a minimum of 5 years. Patient-important outcomes such as improvement in functional class, symptoms of heart failure, and quality of life should be assessed along with MVA. The current preference for PTMC over SC is partly guided by up-front cost considerations. Formal economic evaluations of the competing strategies, over an extended time frame, are needed to guide the efficient allocation of resources in developing countries.
Conclusion
Though the quality and quantity of evidence is insufficient to confidently assert the superiority of either PTMC or SC for long-term outcomes of importance to patients, given the greater procedural morbidity associated with cardiac surgery, PTMC is the procedure of choice for young patients with favourable valve morphology.
Key questions
What is already known about this subject?
Percutaneous mitral commissurotomy (PTMC) is preferred over surgical commissurotomy (SC) for the treatment of severe rheumatic mitral stenosis. A previous systematic review however suggests that SC may be better than PTMC
What might this study add?
In this rigorously performed systematic review of randomised trials, we found that though the quality of the available evidence is low (assessed using the GRADE approach), there appears to be no difference between PTMC and SC with respect to key patient important outcomes
How might this impact on clinical practice?
Given the higher peri-procedural morbidity associated with SC, PTMC should remain the procedure of choice in young patients with favourable valve morphology
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Twitter @drkarthik2010
ADS and AM contributed equally.
Contributors Study concept, design: GK, ND, GG. Data collection: ADS, AM. Statistical analysis: ND, ADS, AM, GG, GK. Writing the first draft: ADS, AM. Analysis and interpretation of results, critical revision of the manuscript for important intellectual content, approval of final version: GK, ND, ADS, AM, GG.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information.