CADTH Health Technology Review

Transcatheter Mitral Edge-to-Edge Repair in Patients with Primary or Secondary Mitral Regurgitation

Rapid Review

Authors: Kwakye Peprah, Holly Gunn, Melissa Walter

Key Messages

• Four systematic reviews (SRs) and 6 retrospective cohort studies provided evidence for the clinical effectiveness of transcatheter mitral valve repair (TMVR) versus open heart conventional surgical mitral valve repair or replacement (SMVR) in patients with primary or secondary mitral regurgitation (MR). No relevant evidence regarding the cost-effectiveness of TMVR versus SMVR in patients with primary or secondary MR was identified; therefore, no summary can be provided.

• There was evidence indicating a statistically significant difference in favour of TMVR over SMVR regarding the odds of post-procedure bleeding, need for permanent pacemaker implantation, 30-day readmission, and a shorter duration of hospitalization.

• There was evidence suggesting a statistically significant difference in favour of SMVR over TMVR regarding the odds of recurrent MR, the need for reoperation, and mortality rate (i.e., during hospitalization, at 1 year, and > 3 years). Also, compared with TMVR, the likelihood of residual MR grade > 2 or freedom from MR grade ≥ 2 or ≥ 3 at 4 years was statistically significantly lower or higher, respectively, with SMVR.

• Evidence regarding the comparative clinical effectiveness of TMVR versus SMVR concerning stroke, acute kidney injury (AKI), cardiogenic shock, and death during hospitalization was conflicting and inconclusive.

• There was no evidence of a significant difference between the 2 interventions regarding overall mortality or mortality at 5 years, overall survival, freedom from cardiac death at 4 years, cardiac arrest, acute myocardial infarction (MI), and respiratory or vascular complications.

• A major limitation of the evidence was that it derives from studies of low or unknown quality and risk of bias, Furthermore, all the findings are confounded by differences in patient selection, which reflect the approved indications for the interventions but prevent a direct comparison between the TMVR and SMVR groups.

Context and Policy Issues

The mitral valve is a valve located between the heart’s 2 left chambers where its 2 leaflets or flaps open and close in a manner that ensures a unidirectional blood flow from the atrium to the ventricle. Mitral regurgitation (MR) is a condition of the malfunctioning mitral valve that allows a backward blood leak in the heart because the mitral valve does not close properly.¹ The disease is categorized into primary (also called degenerative) or secondary (also called functional), where primary MR is caused by an abnormality of 1 or more components of the valve apparatus (e.g., leaflets, papillary muscles, annulus); whereas, secondary MR is due to another heart disease such as coronary heart disease, cardiac ischemia, and heart failure.^2,3

MR is a progressive disease and considered the second most frequent valve disease.^1,4 The incidence of MR increases with age, with a prevalence of over 9% in patients ≥ 75 years of age.⁴ Common symptoms of MR include fatigue, shortness of breath, coughing, an irregular heartbeat, and swollen feet or ankles.¹ The disease is frequently associated with decreased quality of life, increased difficulty performing regular daily activities,¹ ventricular dysfunction, and high mortality.⁵ Without appropriate treatment the annual mortality rate in symptomatic patients exceeds 5% and may be up to 20% in those with severe MR.⁶

The goal of treatment is to decrease MR severity and improve quality of life.¹ The main treatment options include guideline-directed medical therapy, surgical mitral valve repair or replacement (SMVR), as well as heart transplant and left ventricular assist devices for advanced heart failure.⁵ Medical management of MR has been reported to have poor outcomes, with a high hospitalization rate due to heart failure, and mortality rates of 20% at 1-year and 50% at 5-years following treatment.⁷ In contrast, SMVR is considered the gold-standard treatment for patients with symptomatic, severe MR and is associated with high rates of MR relief (> 95%) and mortality rates of 1% to 3%.^6,8 However, approximately 50% of patients with severe MR are considered to have a prohibitive risk for surgery and are not qualified as candidates for SMVR.^4,6-8 High-risk factors leading to SMVR ineligiblity include frailty, old age (> 80 years), left ventricular dysfunction with severely reduced (< 30%) left ventricular ejection fraction, and other associated comorbidities. Therapeutic alternatives for patients with severe MR who are ineligible for SMVR are limited, and most of them traditionally receive medical management although it has been proven to be less effect.^7,8

In 2014 Health Canada–approved transcatheter mitral valve repair (TMVR) using MitraClip for percutaneous reduction of severe symptomatic primary MR in patients with prohibitive risk for SMVR.⁹ The device had received the US FDA approval for a similar indication in 2013.^4,6,8 In 2019, the FDA approved MitraClip for use in patients with moderate-to-severe or severe functional MR, who have heart failure symptoms despite optimal medical therapy.

A CADTH report on TMVR published in June 2020 reviewed the clinical effectiveness of the MitraClip device for the treatment of tricuspid regurgitation but did not include any evidence about its clinical or cost-effectiveness in MR.¹⁰ The objective of the current report is to identify and summarize evidence of clinical and cost-effectiveness of TMVR versus open heart conventional SMVR in patients with primary or secondary MR.

Research Questions

1. What is the clinical effectiveness of transcatheter edge-to-edge repair in patients with primary or secondary mitral regurgitation?

2. What is the cost-effectiveness of transcatheter edge-to-edge repair in patients with primary or secondary mitral regurgitation?

Methods

Literature Search Methods

A limited literature search was conducted by an information specialist on key resources including MEDLINE, the Cochrane Database of Systematic Reviews, the international HTA database, the websites of Canadian and major international health technology agencies, as well as a focused internet search. The search strategy comprised controlled vocabularies, such as the National Library of Medicine’s MeSH (Medical Subject Headings), and keywords. The main search concepts were transcatheter mitral edge-to-edge repair and mitral valves. No filters were applied to limit the retrieval by study type. Comments, newspaper articles, editorials, letters, and case reports were excluded. Where possible, retrieval was limited to the human population. The search was also limited to documents in English published between January 1, 2016, and June 4, 2021.

Selection Criteria and Methods

In the first screening level, titles and abstracts were reviewed and potentially relevant articles were retrieved and assessed for inclusion. One reviewer screened citations and selected studies. The final selection of full-text articles was based on the inclusion criteria presented in Table 1.

Exclusion Criteria

Articles were excluded if they did not meet the selection criteria outlined in Table 1 or if they were duplicate publications or were published before 2016. Primary studies retrieved by the search were excluded if captured in 1 or more of the included SRs. SRs in which all relevant studies were included in other more recent or more comprehensive systematic reviews were excluded.

Critical Appraisal of Individual Studies

The included publications were critically appraised by 1 reviewer using the following tools as a guide: A MeaSurement Tool to Assess Systematic Reviews 2 (AMSTAR 2)¹¹ for SRs and the Downs and Black checklist¹² for non-randomized studies. Summary scores were not calculated for the included studies; instead, a narrative review of the strengths and limitations of each included study was provided.

Summary of Evidence

Quantity of Research Available

A total of 1,083 citations were identified in the literature search. After screening of titles and abstracts, 1,050 citations were excluded, and 33 potentially relevant reports from the electronic search were retrieved for full-text review. The grey literature search did not identify any additional potentially relevant studies. Of 33 articles reviewed in full-text, 23 were excluded for various reasons, while 10 that met the inclusion criteria were included in this report. These comprised 4 SRs^3,4,8,13 with meta-analyses and 6 retrospective cohort studies^6,14-18 relevant for clinical effectiveness of TMVR. No studies relevant for cost-effectiveness of TMVR were identified. Appendix 1 presents the PRISMA¹⁹ flow chart of the study selection.

Summary of Study Characteristics

Study Design

Systematic reviews

Three of the included SRs were published in 2021^3,4,8 while 1 was published in 2020.¹³ Each SR was based on relevant literature retrieved by systematic searches conducted in multiple databases (i.e., 2 or more) between December 2018 and June 2020. The number of primary studies included in the individual SRs ranged from 9 to 14. In 1 SR of 12 studies,¹³ which had a broader scope with an additional comparator of interest, 7 retrospective cohort studies were relevant to this report, while 5 were out of the scope. The 4 included SRs reported including 42 primary studies. When overlapping included studies were accounted for, the 4 SRs involved 22 unique studies published from 2011 to 2020. Thus, there was a considerable overlap of primary studies across 2 or more SRs. A table describing the overlap is available in Appendix 5. The individual studies included in the SRs were retrospective cohort studies except for 2 RCTs, 1 of which was common to 3 of the SRs.^3,4,8 Additional details regarding the characteristics of the included SRs have been provided in Appendix 2.

Primary clinical studies

Four of the included retrospective cohort studies were published in 2020^6,14,17,18 while 1 was published in 2018¹⁵ and another in 2016.¹⁶ Three of the studies^14,17,18 used data from the National Inpatient Sample database of the US, and another study⁶ was based on data from the Nationwide Readmission Database of the US. Each of the remaining 2 studies^15,16 was conducted using patient data retrieved from records of 2 separate hospitals (one per each study). One study¹⁷ analyzed data for propensity score-matched patients from the TMVR and SMVR groups. Additional details regarding the characteristics of the included retrospective cohort studies have been provided in Appendix 2.

Country of Origin

Each of the 4 included SRs had lead authors from the UK,³ the US,⁴ China,⁸ or Portugal.¹³ Four of the 6 retrospective cohort studies were conducted in the US^6,14,17,18 while 1 was conducted in Germany¹⁵ and another 1 in Italy.¹⁶

Patient Population

Additional details regarding the populations of the studies included in this report are available in Appendix 2.

Systematic reviews

The number of patients involved in the individual SRs ranged from 930 to 4,219. One SR³ included patients who underwent an intervention for any defined MR, whereas 2 another SRs included patients with severe MR.^8,13 One SR⁴ did not specify the severity of MR, and another SR broadly classified MR grade as ≥ 2. The mean age of patients in each SR was between 67.0.and 67.3 years for those treated with TMVR and 53.4 and 63.0 for those treated with SMVR. All 4 SRs indicated that patients treated with TMVR were older and had more comorbidities. Three SRs evaluated patients’ surgical risk and found that patients in the TMVR group had a higher surgical risk than those in the SMVR group as measured by logistic Euroscore^3,4,8 or the Society of Thoracic Surgeons (STS)⁴ scores.

Primary clinical studies

The number of patients in the retrospective cohort studies varied from hundreds for the single-centre studies^15,16 (148 to 378) to thousands for the national database-dependent studies^6,14,17,18 (2,910 to 19,580). One study used data for patients with severe MR¹⁴ and another was conducted in patients with severe or moderately severe secondary MR.¹⁶ The other studies did not provide information regarding the MR severity of the patients. The authors indicated that overall, patients in the TMVR groups were older and had more comorbidities or surgical risk than those in the SMVR group, except in 1 study¹⁷ that used a propensity score matching approach to match patients from the 2 groups whose data were used in analyses. The mean age of patients in 4 of the studies^6,15,16,18 was between 67.0 and 77.5 years for those treated with TMVR compared with 61.2 and 63.7 for those treated with SMVR. One study¹⁸ reported age using median instead of mean, with a median age of 79.0 years and 62.0 years for the TMVR and SMVR groups, respectively. The study¹⁷ that used a propensity score matching approach involved elderly patients at least 80 years old. The mean age in that study was 83.7 years in both the TMVR and SMVR groups.¹⁷ Two studies evaluated comorbidity using the Charlson Comorbidity Index (CCI),^6,14 and another 2 studies assessed surgical risk using logistic Euroscore¹⁶ or STS.¹⁵ In contrast, the other 2 studies reported itemized varying conditions considered risks factors for poor surgical outcomes.^17,18

Interventions and Comparators

The relevant study comparisons were between TMVR and SMVR. One SR³ and 1 retrospective cohort study¹⁶ stated that the MitraClip procedural techniques were standard, with the clip device positioned at the place where regurgitation is occurring. In both publications, it was indicated that more than 1 clip was implanted when necessary. According to the SR,³ 6 of its primary studies that provided SMVR procedure details used annuloplasty while 1 performed surgical replacement of the valve. In the cohort study,¹⁶ the SMVR approach was a surgical edge-to-edge repair together with annuloplasty using undersized rigid or semirigid rings. Nine of the included studies did not provide procedural details about TMVR or SMVR. All 4 SRs^3,4,8,13 and 5 of the retrospective cohort studies^6,14-16,18 stated directly or implied that the TMVR procedure used the MitraClip devices. However, the type of TMVR device was unclear from the retrospective cohort study conducted in Germany.¹⁵

Outcomes

Reported outcomes of interest to this review included, but not limited to, post-operative mortality,^{3,4,8,13,17,18} survival,^3,16 length of stay (LOS) in hospital,^3,4,6,14-18 recurrent/residual MR,^3,4,8,15 readmission rate,⁶ reoperation rate,^3,4 need for a pacemaker,^14,17 and incidence of complications (e.g., major bleeding,^13,14,17 stroke,^3,4,6,13-15 acute kidney injury [AKI],^4,14,15,17 and acute myocardial infarction [MI]).^6,14

Summary of Critical Appraisal

Overall, the quality of evidence from the included SRs^3,4,8,13 and retrospective cohort studies^6,14-18 was limited by factors discussed in the appraisal below. Details regarding the strengths and limitations of individual studies are available in Appendix 3.

Systematic Reviews

Two SRs^4,13 reported establishing review protocols a priori and registering them on PROSPERO; whereas, for the other 2 SRs,^3,8 there was no indication that a protocol was prepared before the review started. Thus, in the absence of a published protocol for the latter 2,^3,8 it could not be independently ascertained if there had been any significant deviations from the protocol that could have introduced risks of bias. Each SR^3,4,8,13 had a clear objective and well-defined populations, interventions, and comparators under study, as well as the outcomes of interest. All the SRs^3,4,8,13 included relevant primary studies retrieved from comprehensive literature searches of multiple databases. In 1 SR,⁴ the database search was supplemented by a handsearch of all references of relevant reviews and eligible articles for potentially missed eligible studies.

All the SRs^3,4,8,13 performed abstract and citation screening and final study selection in duplicate according to prespecified inclusion and exclusion criteria to minimized selection bias. In 2 SRs,^4,13 2 reviewers independently rated the quality or risk of bias of included primary studies with appropriate tools (Newcastle Ottawa Scale and the Cochrane method), resolving disagreements by consensus. One SR⁴ also evaluated the certainty in evidence for each outcome using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach. However, 2 other SRs^3,8 neither assessed the quality or risk of bias of their primary studies nor the certainty in evidence for the outcomes they measured, making it difficult to judge the quality and reliability of their reported estimates. It is noteworthy that across the 4 included SRs,^3,4,8,13 all the primary studies were non-randomized studies, except 2 RCTs of unknown quality and risk of bias, 1 of which was common to 3 SRs,^3,4,8 while the other was included in another SR.¹³ Therefore, for each SR,^3,4,8,13 the individual studies providing input for analysis had inherent high-risk of selection bias. Indeed, there were imbalances in some patient characteristics, such as age, comorbidities, and surgical risk status, that could potentially bias the reported findings against TMVR. The authors did not explain their selection of the study designs for inclusion in the SRs. However, it appeared that the lack of relevant studies limited the options for study selection.

In 2 SRs,^3,8 extracted data were independently checked for accuracy by other reviewers; however, it was unclear whether data extraction in 2 other SRs^4,13 was performed in duplicate or independently confirmed for accuracy. All the SRs^3,4,8,13 conducted meta-analyses using random-effects models on pooled data from multiple studies. Thus, it was expected that their reported outcomes, representing a summary of findings combined across multiple studies, would offer increased precision of effect estimates than the individual primary studies would. However, in 2 SRs,^8,13 the overall effect estimates for some comparisons were derived from pooled data involving an RCT and non-randomized studies, meaning that the estimates may have been biased. For the meta-analyses, each of the SRs^3,4,8,13 assessed between-study heterogeneity with the appropriate statistical methods.^3,4,8,13 Three SRs^3,4,13 assessed publication bias; whereas, 1 SR⁸ did not. For the SRs that reported assessing publication bias, 2 reported no evidence of significant publication bias; ^,13 whereas, 1 found significant publication bias regarding 5 years mortality but not any other outcome.⁴

Primary Clinical Studies

All the retrospective cohort studies^6,14-18 stated objectives and described patients’ characteristics, the interventions and comparators of interest, and outcomes to be measured clearly. Four studies analyzed data from large nationwide databases (the US NIS^14,17,18 and the Nationwide Readmissions Database⁶). Thus, the data were likely to represent the targeted population of patients treated for MR, and the settings and staff who performed the procedures and provided care were likely to be representative of care settings and treatment received by most patients who underwent TMVR or SMVR across the country. The analytical methods were appropriate, and overall, the findings were reported clearly along with estimates of the random variability and P values reported. However, none of the cohort studies adjusted for multiplicity, even though they all conducted multiple statistical tests. As a result, there may be potential inflation of the type I error rate in these studies. Two studies^6,17 reported results from propensity-score-matched populations across the comparing groups to minimize bias due to inter-group differences.

A key source of uncertainty in the included primary studies was that they were all retrospective cohort studies lacking randomization that reduces the risk of selection bias. Indeed, there were imbalances in some patient characteristics, such as age and comorbidities, surgical risk status, and general medical history that could potentially bias the reported findings against TMVR. The limitation applied even for the 2 studies^6,17 that used a propensity score matching approach since the method adjusts only for parameters considered in a chosen model and does not rule out significant residual or unknown confounders. For 4 of the primary studies^6,14,17,18 data were limited to the period of patients’ hospitalization. Therefore, information after discharged from the hospital, including intermediate and long-term outcomes, was not available for assessment. Also, 4 studies^6,14,17,18 analyzed data pooled and coded from multiple hospitals across the US, increasing the likelihood that the differences in expertise and facilities at the various settings could influence their results. In 3 studies,^15,16,18 patients in the TMVR group were compared to a historical cohort of SMVR-treated patients. Thus, it was unclear whether the results had been impacted by changes in techniques and health care delivery over time, and if the findings could be different had the procedures been performed within the same period. Relatedly, with a few exceptions, most of the data used in all the included studies came from the early years (up to 2016) after MitraClip received approval for use in patients with MR in the US (2013) and Canada (2014). Therefore, it is unclear if increased practitioner experience and skills with the procedure over time may have resulted in significant improvements in current outcomes compared to the earlier period. Moreover, all the studies were based on data from outside Canada, including 2 single-site studies conducted in Germany and Italy. Thus, the generalizability of the findings in the Canadian context is unknown.

Summary of Findings

Clinical Effectiveness of Transcatheter Mitral Edge-to-Edge Repair

The findings from the included studies have been summarized below by outcome. The main study-level findings and authors’ conclusions for the individual studies are available in Appendix 4 4. The included SRs had overlapping primary studies; therefore, the pooled estimates from the separate reviews contain much of the same data. A citation matrix illustrating the degree of overlap is presented in Appendix 5.

Complications

Post-procedural complications reported by the included studies include stroke, AKI, major bleeding, pacemaker implantation, cardiac arrest, cardiogenic shock, myocardia infraction, and respiratory and vascular complications. Details for each outcome are summarized below.

Stroke

Three SRs^3,4,13 and 3 retrospective cohort studies^6,14,15 reported on the odds ratio for stroke (Table 2 and Table 3). While all the SRs^3,4,13 and 1 cohort study¹⁵ found no statistically significant difference between TMVR and SMVR regarding this outcome, 2 cohort studies^6,14 reported a statistically significantly lower odds of having a stroke with TMVR than with SMVR. Thus, the results were inconclusive due to the inconsistency across the reporting SRs and studies.

Acute kidney injury

One SR⁴ and 4 retrospective cohort studies^6,14,15,17 reported on the odds ratio for AKI. Whereas the SR⁴ and 2 of the cohort studies^6,14 found no statistically significant difference between TMVR and SMVR regarding this outcome, 2 other cohort studies^15,17 reported a statistically significantly lower odds of having an AKI with TMVR than with SMVR (P < 0.001) Table 4 and Table 5). Thus, the evidence was inconclusive due to the inconsistency across the reporting SRs and studies.

Bleeding

One SR¹³ and 3 retrospective cohort studies^6,14,17 reported on the odds ratio for bleeding. They were consistent in finding a statistically significantly lower (P < 0.001) bleeding with TMVR than with SMVR (Table 6 and Table 7). The results were consistent with the understanding that TMVR is a less invasive procedure compared with SMVR.

Permanent Pacemaker implantation

Two retrospective cohort studies^14,17 reported on implantation of permanent pacemakers. They were consistent in finding that the odds for needing a permanent pacemakers implantation was statistically significantly lower with TMVR than with SMVR (Table 8).

Cardiac arrest

One retrospective cohort study¹⁴ reporting on the odds ratio for cardiac arrest found no statistically significant difference in the odds of having a cardiac arrest in patients treated with TMVR compared with those who had SMVR (Table 9).

Cardiogenic shock

Two retrospective cohort studies^6,17 reporting on the odds ratio for cardiogenic shock found a statistically significantly lower odds of having a cardiogenic shock with TMVR than with SMVR. However, results from another retrospective cohort study¹⁴ did not show a statistically significant difference in this outcome between TMVR and SMVR Table 10).

Acute myocardial infarction

Two retrospective cohort study^6,14 reported on the odds ratio for MI. They found no statistically significant difference between TMVR and SMVR regarding this outcome (Table 11).

Respiratory and vascular complications

One retrospective cohort study¹⁷ reporting on the percentage of patients who had respiratory or vascular complications found no statistically significant difference between TMVR and SMVR regarding these outcomes (Table 12).

Length of hospital stay

Two SRs^3,4 and 6 retrospective cohort studies^6,14-18 reported on the duration of hospital stay after TMVR versus SMVR. The units of measurement were days with the results presented as standardized mean difference,³ mean difference,⁴ median with interquartile rage (IQR),^14-16 and mean with standard deviation (SD).^6,17 All the SRs^3,4 and cohort studies^6,14-18 were consistent in showing that the LOS was statistically significantly shorter with TMVR than SMVR (Table 13 and Table 14).

Mortality

Three SRs^3,4,8 and 3 retrospective cohort studies^14,17,18 reported on the odds ratio for mortality after TMVR versus SMVR. While all the SRs^3,4,8 and 2 cohort studies^14,18 found no statistically significant difference in mortality during hospitalization between TMVR and SMVR, 1 cohort study¹⁷ reported a statistically significantly lower odds of dying for this outcome with TMVR than with SMVR (Table 15 and Table 16). Thus, the results concerning mortality during post-procedure hospitalization were inconclusive due to the inconsistency across the reporting SRs and cohort studies. Two SRs^4,8 reporting 1 year outcomes were consistent in showing that the odds of dying were statistically significantly higher with TMVR than with SMVR. One of the SR⁸ also found that that the odds of dying were statistically significantly higher with TMVR than with SMVR at > 3 years (Table 15). Another SR⁴ found no statistically significant difference in the odds for overall mortality or mortality at 5 years between TMVR and SMVR. It should be noted that the SR⁴ reported a significant publication bias regarding the 5-year mortality, but not for other outcomes.

Survival

One SR³ and 1 retrospective cohort study¹⁶ reported survival outcomes for TMVR versus SMVR. Of note, the SR³ reported on the odds ratio for survival both at midterm (1 to 2 years) and long-term (4 to 5 years) outcomes; whereas, the cohort study¹⁶ reported on the overall survival and freedom from cardiac death at 4 years, analyzed by Kaplan–Meier method with differences among groups evaluated with the log-rank test. However, mean percentage (SD) results were reported for the 2 outcomes in the cohort study.¹⁶ Both the SR³ and the cohort study¹⁶ were consistent in showing no statistically significant difference between the 2 procedures regarding all the assessed outcomes (Table 17 and Table 18). In the retrospective cohort study,¹⁶ the duration of follow-up was significantly shorter for the TMVR group compared with the SMVR group (median 2.5 years versus 7.2 years; P = 0.0001) (Table 25). Thus, it is unknown if the difference between the 2 groups will be reproducible in follow-ups of equal duration.

Recurrent MR

Three SRs^3,4,8 reported on the on the odds ratio for recurrent MR following TMVR versus SMVR, with post-procedural findings at the time of discharge from hospital,⁴ and at 30 days,⁸ 1 year,⁸ midterm (i.e., 1 to 2 years),³ greater than 3 years,⁸ and 5 years.⁴ All the SRs^3,4,8 were consistent in showing that the odds of recurrent MR were statistically significantly higher following treatment with TMVR compared with SMVR. The OR (95% CI) at the various assessment points were between 2.46 (1.54 to 3.94) and 9.51 (5.38 to 16.83) (Table 19). Also, 1 retrospective cohort study¹⁵ reporting the mean (SD) for post-operative residual MR greater than grade 2 found that the outcome was statistically significantly higher with TMVR than with SMVR (Table 20). Similarly, another retrospective cohort study¹⁶ reported that the mean percentage (SD) of patients with freedom from MR grade greater than or equal to 2 or greater than or equal to 3 at 4 years was statistically significantly lower in the TMVR group than in the SMVR group (Table 21).

Reoperation

Two SRs^3,4 reported on the odds ratio for reoperation after TMVR versus SMVR. All the studies were consistent in showing that the odds of having a reoperation was statistically significantly higher with TMVR compared with the SMVR (Table 22).

Readmission rate

One retrospective cohort study⁶ reported on the odds ratio for 30-day readmission rates between patients who underwent TMVR versus those who had SMVR. The unadjusted results showed no statistically significant difference in the odds of readmission between the 2 procedures. However, after a propensity score matching adjustment, TMVR was associated with statistically significantly lower odds for 30-day readmission compared with SMVR (Table 23). There were no data provided to independently assess the extent to which the adjustment reduced the imbalance between the TMVR and SMVR groups.

Cost-Effectiveness of Transcatheter Mitral Edge-to-Edge Repair

No relevant evidence regarding the cost-effectiveness of TMVR versus SMVR for the treatment of primary or secondary MR was identified; therefore, no summary can be provided.

Limitations

A fundamental limitation involves limited information about the included studies and differences in study characteristics. For example, most of the included studies provided insufficient information about the type (i.e., degenerative or functional MR) and severity of MR before the interventions were given. However, 1 study¹⁶ restricted enrolment to patients who had severe or moderately severe functional MR, while another study¹⁷ recruited only elderly patients who were 80 years or older (Table 24 and Table 29 for detailed patient characteristics), which could inform patient selection criteria. Nevertheless, the results from those studies may not be generalizable to patients who are different from the study participants. Five of the 6 included studies did not report a specific follow-up period. Also, there were inadequate details about how the TMVR and SMVR procedures were applied in the various studies, with the available information suggesting a lack of standardization for the interventions. For instance, while some patients in the TMVR groups were implanted with a single clip, others were treated with multiple clips at the discretion of the attending surgeon. Also, the reported SMVR methods differed between the studies that provided information about the procedure. Taken together with other limitations discussed in the critical appraisal section, the generalizability of the findings is unclear, and it is difficult to draw conclusions from the included studies with any certainty. Some retrospective cohort studies identified patients with relevant data using the International Classification of Disease, ninth revision, Clinical Modifications (ICD-9-CM); whereas others used the 10th revision (ICD-10-CM). Thus, there was a potential for misalignment in the definition of patient’s conditions. However, the impact of this on the reported outcomes was unknown. None of the included studies reported on health-related quality of life outcomes, an essential critical end point, especially for patients with chronic diseases. No relevant evidence on cost-effectiveness was identified.

Conclusions and Implications for Decision- or Policy-Making

Four SRs^3,4,8,13 and 6 retrospective cohort studies^6,14-18 provided evidence for the clinical effectiveness of TMVR versus SMVR in patients with MR. Most of the clinical effectiveness outcomes of interest fell under the category of complications. They included the incidence of stroke,^3,4,6,13-15 AKI,^4,6,14,15,17 major bleeding,^6,13,14,17 cardiac arrest,¹⁴ cardiogenic shock,^6,14,17 acute MI,^6,14 as well as respiratory¹⁷ and vascular¹⁷ complications. The other outcomes were implantation of permanent pacemakers,^14,17 post-procedure LOS^3,4,6,14-18 mortality,^{3,4,8,14,17,18} survival,^3,16 recurrent/residual MR,^3,4,8,15 freedom from grade greater than or equal to 2 or greater than or equal to 3 MR at 4 years,¹⁶ reoperation rate^3,4 and readmission rate.⁶ No relevant evidence regarding the cost-effectiveness of TMVR versus SMVR for the treatment of primary or secondary MR was identified; therefore, no summary can be provided.

Evidence from 1 SR¹³ and 3 retrospective cohort studies^6,14,17 indicated that TMVR was associated with less frequent post-procedural bleeding than SMVR. Also, there was consistent evidence from 2 SRs and 6 retrospective cohort studies showing a shorter hospitalization duration following TMVR than after SMVR. Furthermore, in an analysis that applied a propensity score matching adjustment, 1 retrospective cohort study⁶ found that TMVR was associated with a statistically significantly lower odds of 30-day readmission compared with SMVR. However, there were no data provided to independently assess the extent to which the adjustment reduced the imbalance between the TMVR and SMVR groups. These findings may be explained because SMVR is a more invasive procedure likely to result in more bleeding and a longer recovery time than TMVR. Evidence from 2 retrospective cohort studies suggested that TMVR was associated with a significantly less frequent need for permanent pacemaker implantation compared with SMVR. This finding is in agreement with reports in the literature showing that the rate of new-onset atrial fibrillation was 3.8% after TMVR²⁰ and 18.0% after SMVR,²¹ suggesting that patients who underwent SMVR are more likely to require permanent pacemaker implantation.

However, evidence from 3 SRs^3,4,8 suggested a higher incidence of recurrent MR following TMVR compared with SMVR at the time of discharge from hospital,⁴ and at 30 days,⁸ 1 year,⁸ midterm (i.e., 1 to 2 years),³ greater than 3 years,⁸ and 5 years⁴ after the procedures. Similarly, 1 retrospective cohort study¹⁵ found that postoperative residual MR greater than grade 2 occurred significantly more frequently after TMVR than SMVR. Another retrospective cohort study¹⁶ reported that the percentage of patients with freedom from MR grade greater than or equal to 2 or greater than or equal to 3 at 4 years was statistically significantly lower in the TMVR group than in the SMVR group. Two SRs^3,4 found evidence of a significantly higher odds of reoperation after TMVR than after SMVR. However, the death rate at 1 year was statistically significantly higher with TMVR than SMVR in in 2 SRs^4,8 including 1 SR⁸ with a similarly higher mortality rate for TMVR than SMVR at greater than 3 years. One SR⁴ found no statistically significant difference in overall mortality or mortality at 5 years between the 2 procedures. Also, there was no evidence of a statistically significant difference between TMVR and SMVR regarding overall survival,^3,16 freedom from cardiac death at 4 years,¹⁶ cardiac arrests,¹⁴ acute MI^6,14 and respiratory¹⁷ and vascular¹⁷ complications.

The evidence from comparisons between TMVR and SMVR regarding the incidence of stroke, AKI, cardiogenic shock, and death during hospitalization were inconclusive due to conflicting findings from the reporting SRs and studies. Three SRs^3,4,13 and 1 retrospective cohort study found no difference between TMVR and SMVR in the incidence of stroke. In contrast, 2 retrospective cohort studies reported a significantly lower incidence of stroke with TMVR. For AKI, evidence from 1 SR⁴ and 2 retrospective cohort studies^6,14 showed no significant difference between the 2 procedures, whereas 2 retrospective cohort studies^15,17 reported a significantly lower incidence of stroke with TMVR. Similarly, evidence from 3 SRs^3,4,8 and 2 retrospective studies suggested no significant difference in mortality rates during hospitalization after TMVR compared with SMVR. In contrast, the results from 1 retrospective cohort study indicated a statistically significantly lower percentage of in-hospital deaths among patients who underwent TMVR compared with those who had SMVR.

A key source of uncertainty in the SRs^3,4,8,13 is the use of mostly retrospective cohort studies and 2 RCTs of unknown quality and risk of bias as primary studies. This implies that all the individual studies providing input for analyses in the SRs^3,4,8,13 had an inherently high or unknown risk of bias, making it difficult to draw conclusions. Other limitations in the SRs include failure to evaluate the quality and risk of bias of their primary studies,^3,8 and potentially biased pooled estimates due to meta-analyses combining data from an RCT and retrospective cohort studies.^8,13 For the included primary studies, the major source of uncertainty was that they were all retrospective cohort studies lacking randomization that reduces the risk of selection bias. Thus, imbalances in baseline patient characteristics, such as surgical risk status and general medical history, potentially biased their reported findings against TMVR. Four of the retrospective cohort studies^6,14,17,18 used data limited to patients’ hospitalization period. Therefore, information after hospital discharge, including intermediate and long-term outcomes, was not available for assessment. Also, 4 studies^6,14,17,18 analyzed data pooled and coded from multiple hospitals across the US. Although that provides a more representative sample of practice in real life, it also increases the likelihood that the differences in expertise and facilities at the various settings could impact their results. Moreover, patients in the TMVR group in 3 studies^15,16,18 were compared to a cohort of SMVR-treated patients from a different historical period, creating uncertainty about whether their findings would be replicated if the cohorts underwent the procedures performed within the same period.

Furthermore, all the included studies^{3,4,6,8,13-18} used data from outside Canada, and the generalizability of the findings in the Canadian context is unknown.

Future studies using more rigorous designs such as RCTs or other types of studies employing appropriate statistical methods to minimize confounding may provide a more definitive answer concerning the clinical effectiveness of TMVR versus SMVR in the treatment of patients with primary or secondary MR. Economic evaluations are also needed to provide cost-effectiveness evidence.

References

1.University of Ottawa Heart Institute. MitraClip. 2021; https://www.ottawaheart.ca/test-procedure/mitraclip. Accessed 2021 Jul 5.

2.Gaasch WH. Management of chronic primary mitral regurgitation. In: Post TW, ed. UpToDate. Waltham (MA): UpToDate; 2021: www.uptodate.com. Accessed 2021 Jul 5.

3.Khader AA, Allaf M, Lu OW, et al. Does the clinical effectiveness of Mitraclip compare with surgical repair for mitral regurgitation? J Card Surg. 2021;36(3):1103-1119. PubMed

4.Oh NA, Kampaktsis PN, Gallo M, et al. An updated meta-analysis of MitraClip versus surgery for mitral regurgitation. Ann Cardiothorac Surg. 2021;10(1):1-14. PubMed

5.Ali M, Shreenivas SS, Pratt DN, Lynch DR, Kereiakes DJ. Percutaneous interventions for secondary mitral regurgitation. Circ Cardiovasc Interv. 2020;13(8):e008998. PubMed

6.Lima FV, Kolte D, Rofeberg V, et al. Thirty-day readmissions after transcatheter versus surgical mitral valve repair in high-risk patients with mitral regurgitation: Analysis of the 2014-2015 Nationwide readmissions databases. Catheter Cardiovasc Interv. 2020;96(3):664-674. PubMed

7.Yandrapalli S, Biswas M, Kaplan J. Mitral valve minimally invasive surgical treatment. StatPearls. 2021.

8.Yuan H, Wei T, Wu Z, et al. Comparison of transcatheter mitral-valve repair and surgical mitral-valve repair in elderly patients with mitral regurgitation. Heart Surg Forum. 2021;24(1):E108-E115. PubMed

9.Health Canada. Summary Basis of Decision (SBD) for MitraClip Clip Delivery System. 2020; https://hpr-rps.hres.ca/reg-content/summary-basis-decision-medical-device-detailThree.php?linkID=SBD00402. Accessed 2021 Jul 6.

10.Li KX, Wright M-D. Transcatheter mitral valve repair device for the treatment of tricuspid or tricuspid and mitral regurgitation: a review of clinical effectiveness and cost-effectiveness. (CADTH Rapid response report: summary with critical appraisal). Ottawa (ON): CADTH; 2020: https://cadth.ca/sites/default/files/pdf/htis/2020/RC1282%20MitraClip%20Tricuspid%20Regurgitation%20Final.pdf. Accessed 2021 Jul 5.

11.Shea BJ, Reeves BC, Wells G, et al. AMSTAR 2: a critical appraisal tool for systematic reviews that include randomised or non-randomised studies of healthcare interventions, or both. BMJ. 2017;358:j4008. PubMed

12.Downs SH, Black N. The feasibility of creating a checklist for the assessment of the methodological quality both of randomised and non-randomised studies of health care interventions. J Epidemiol Community Health. 1998;52(6):377-384. PubMed

13.Barros da Silva P, Sousa JP, Oliveiros B, et al. Stroke after transcatheter edge-to-edge mitral valve repair: a systematic review and meta-analysis. EuroIntervention. 2020;15(16):1401-1408. PubMed

14.Jogu HR, Arora S, Strassle PD, et al. Impact of age and comorbidities on the effect of transcatheter versus surgical mitral valve repair on inpatient outcomes. Catheter Cardiovasc Interv. 2020;95(6):1195-1201. PubMed

15.Korber MI, Scherner M, Kuhr K, et al. Acute kidney injury following percutaneous edge-to-edge vs. minimally invasive surgical mitral valve repair: incidence, predictors and prognostic value. EuroIntervention. 2018;13(14):1645-1651. PubMed

16.De Bonis M, Lapenna E, Buzzatti N, et al. Optimal results immediately after MitraClip therapy or surgical edge-to-edge repair for functional mitral regurgitation: are they really stable at 4 years? Eur J Cardiothorac Surg. 2016;50(3):488-494. PubMed

17.Malik AH, Zaid S, Yandrapalli S, et al. Trends and outcomes with transcatheter versus surgical mitral valve repair in patients >=80 years of age. Am J Cardiol. 2020;125(7):1083-1087. PubMed

18.Ullah W, Sattar Y, Mukhtar M, et al. Outcomes of open mitral valve replacement versus transcatheter mitral valve repair; insight from the National Inpatient Sample Database. Int J Cardiol Heart Vasc. 2020;28:100540. PubMed

19.Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. J Clin Epidemiol. 2009;62(10):e1-e34. PubMed

20.Arora S, Vemulapalli S, Stebbins A, et al. The prevalence and impact of atrial fibrillation on 1-year outcomes in patients undergoing transcatheter mitral valve repair: results from the Society of Thoracic Surgeons/American College of Cardiology Transcatheter Valve Therapy Registry. JACC Cardiovasc Interv. 2019;12(6):569-578. PubMed

21.Kang MK, Joung B, Shim CY, et al. Post-operative left atrial volume index is a predictor of the occurrence of permanent atrial fibrillation after mitral valve surgery in patients who undergo mitral valve surgery. Cardiovasc Ultrasound. 2018;16(1):5. PubMed

Appendix 1: Selection of Included Studies

Figure 1: Selection of Included Studies

Appendix 2: Characteristics of Included Publications

Note that this appendix has not been copy-edited.

Table 24: Characteristics of Included Systematic Reviews and Meta-Analyses

AF = atrial fibrillation; HF = heart failure; LOS = length of stay at the hospital; MR = mitral regurgitation; NR = not reported; NA = not applicable; NMA = network meta-analysis; NYHA = New York Heart Association; OMT = optimal medical treatment; RCT = randomized controlled trial; SD = standard deviation; SMVR = surgical mitral valve repair, STS = Society of Thoracic Surgeons: TMVR = transcatheter mitral valve repair.

Appendix 3: Critical Appraisal of Included Publications

Note that this appendix has not been copy-edited.

Table 26: Strengths and Limitations of Systematic Reviews and Meta-Analyses Using AMSTAR 2¹¹

AMSTAR 2 = A MeaSurement Tool to Assess systematic Reviews 2; GRADE = Grading of Recommendations Assessment, Development and Evaluation; NOS = Newcastle Ottawa Scale; PICO = population, intervention, control group and outcome; PRISMA = Preferred Reporting Items for Systematic Reviews and Meta-Analyses; PROSPERO = International prospective register of systematic reviews; RoB = risk of bias; RCT randomized controlled trial; SR = systematic review.

Table 27: Strengths and Limitations of Clinical Studies Using the Downs and Black checklist¹²

CM = clinical modification; ICD = International Classification of Diseases; NA = not applicable; NIS = national inpatient sample; SMVR = surgical mitral valve repair; TMVR = transcatheter mitral valve repair.

Appendix 4: Main Study Findings and Authors’ Conclusion

Note that this appendix has not been copy-edited.

Summary of Findings from Included Studies by outcome

Table 28: Summary of Stroke Outcome

CI = confidence interval; OR = odds ratio; RCT = randomized controlled trial; SMVR = surgical mitral valve repair, TMVR = transcatheter mitral valve repair.

Table 29: Summary of Acute Kidney Injury Outcome

AKI = acute kidney injury; CI = confidence interval; NR = not reported; OR = odds ratio; RCT = randomized controlled trial; SMVR = surgical mitral valve repair, TMVR = transcatheter mitral valve repair.

Table 30: Summary of Bleeding Outcomes

CI = confidence interval; HR = hazard ratio; NR = not reported; OR = odds ratio; SMVR = surgical mitral valve repair, TMVR = transcatheter mitral valve repair.

Table 31: Summary of Heart-Related Outcomes

CI = confidence interval; NR = not reported; OR = odds ratio; RCT = randomized controlled trial; SMVR = surgical mitral valve repair, TMVR = transcatheter mitral valve repair.

Table 32: Summary of Respiratory and Vascular Complications Outcomes

SMVR = surgical mitral valve repair, TMVR = transcatheter mitral valve repair.

Table 33: Summary of Length of Hospital Stay Outcomes

CI = confidence interval; CIE = change in estimate; IQR = interquartile range; MD = mean difference; RCT = randomized controlled trial; SD = standard deviation; SMVR = surgical mitral valve repair, TMVR = transcatheter mitral valve repair.

Table 34: Summary of Mortality and Survival Outcomes

CI = confidence interval; HR = hazard ratio; MR = mitral regurgitation; NR = not reported; OR = odds ratio; RCT = randomized controlled trial; SD = standard deviation; SMVR = surgical mitral valve repair, TMVR = transcatheter mitral valve repair.

Table 35: Summary of Recurrent or Residual MR Outcomes

CI = confidence interval; HR = hazard ratio; MR = mitral regurgitation; OR = odds ratio; RCT = randomized controlled trial; SD = standard deviation; SMVR = surgical mitral valve repair, TMVR = transcatheter mitral valve repair.

Table 36: Summary of Re-operation and Readmission Outcomes

CI = confidence interval; MR = mitral regurgitation; OR = odds ratio; RCT = randomized controlled trial; SMVR = surgical mitral valve repair, SR = systematic review; TMVR = transcatheter mitral valve repair.

Appendix 5: Overlap Between Included Systematic Reviews

Note that this appendix has not been copy-edited.

Table 37: Overlap in Relevant Primary Studies Between Included Systematic Reviews