Anticoagulation therapy and clinical outcomes following transcatheter mitral valve repair for patients with mitral regurgitation: A meta‐analysis

Abstract Transcatheter mitral valve repair (TMVR) using MitraClip (MC) is now an established technique in the interventional treatment of mitral regurgitation. Common complications after MC procedure are bleeding and ischemic events. However, 2017 ESC/EACTS and 2020 ACC/AHA did not give a clear antithrombotic protocol, the policy has been based on clinical experience. Here, we performed a meta‐analysis comparing outcomes with and without the addition of anticoagulants after TMVR. We searched the Cochrane Library, EMBASE, PubMed, and Web of Science from inception to October 6, 2022 to identify studies with or without the use of anticoagulants after TMVR. From each study, we extracted the number of people with bleeding, stroke, combined endpoints, and all‐cause death. Five observational cohort studies were included, enrolling a total of 1892 patients undergoing TMVR who were assigned to either the anticoagulation group (n = 1209) or the no‐anticoagulation group (n = 683). Pooled analysis showed a significantly lower stroke rate in the anticoagulated group (at least 4 weeks duration) compared with the non‐anticoagulated group (RR [95% CI] = 0.14 [0.0−0.77], p = 0.02), and similar rates of bleeding, combined endpoints, and all‐cause death in both groups (RR [95% CI] = 0.76 [0.48−1.22], p = 0.26), (RR [95% CI] = 0.52 [0.10−2.63], p = 0.43), and (RR [95% CI] = 0.89 [0.58−1.35], p = 0.58). We observed a reduced risk of stroke without elevated risk of bleeding, combined endpoints, or all‐cause death in patients using anticoagulants (at least 4 weeks duration) after TMVR compared to no anticoagulants.


| INTRODUCTION
The mitral valve, also known as the left atrioventricular valve, is attached to the left fibrous atrioventricular ring and is formed by folds of the endocardium. Mitral valve disease mainly includes mitral regurgitation, mitral stenosis, and mitral valve prolapse, with mitral regurgitation accounting for the majority of all mitral valve disease and even affecting more than 10% of the population over 75 years of age. 1 Severe mitral regurgitation that affects the quality of life requires aggressive treatment. Pharmacological treatment only improves symptoms and does not improve survival. 2 In the absence of surgical intervention, patients with severe mitral regurgitation have an annual mortality rate of 5% 3,4 and a 5-year mortality rate of up to 60% when combined with severe heart failure. 5,6 Therefore, if the situation permits, surgical treatment is recommended, 7 it improves survival, morbidity, and the quality of life. However, patients treated with mitral valve repair or replacement often exhibit advanced age, frailty, and substantial complications and are considered to be at high risk for or unable to tolerate the procedure, and transcatheter mitral valve repair (TMVR) emerged as an alternative treatment. Over the past two decades, TMVR using the MitraClip (MC) system has been increasingly and successfully used to treat degenerative and functional mitral regurgitation. 8 As TMVR causes endothelial damage and the implanted MC increases the risk of thrombosis until endothelialization is complete, this process carries the risk of thromboembolic stroke and requires intensive anticoagulation to prevent thrombosis. Cases of postoperative left atrial or left ventricular thrombosis have been reported. 9,10 However, patients often exhibit risk factors for bleeding such as congestive heart failure, hypertension, advanced age, renal insufficiency, and diabetes mellitus.
ESC/EACTS 2017 and ACC/AHA 2020 did not give a clear antithrombotic protocol, 11 current policy is largely dependent on the experience of the clinician. Here, we conducted a meta-analysis of clinical studies investigating the clinical outcomes of patients receiving TMVR with or without anticoagulant therapy, drawing conclusions that we expect will help clinicians' decisions.

| Protocol and registration
This systematic review was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) criteria, 12 and registered with PROSPERO (International Prospective Register of Systematic Reviews, CRD42022369486).

| Search strategy and source
Two broad search terms ("TMVR" and "antithrombotic") were used and combined using the Boolean operator "AND." For the search term "TMVR," the combination of the following text words, entry terms, and Mesh were used: TMVR, transcatheter mitral valve repair, percutaneous mitral valve repair, MitraClip, mitral clip, mitral valve transcatheter edgeto-edge repair. For the search theme "antithrombotic," the combination of the following text words was used: antiplatelet, antithrombotic, anticoagulation, anticoagulant, aspirin, clopidogrel, P2Y receptor antagonist, edoxaban, apixaban, rivaroxaban, dabigatran, vitamin K antagonist, non-vitamin K oral anticoagulant. The above keywords were used and combined using the Boolean operator "OR." The references listed in all eligible studies were checked to identify further citations.

| Study selection and screening criteria
The following selection and screening criteria were used to perform

| Data extraction
Two reviewers (J. N. S. and A. X.) independently extracted and recorded all data using a standard form. Data were checked by a third reviewer (J. Z.). Information extracted included: first author, year of publication, country, study design, duration of follow-up, total number of patients, age, sex, number of patients per intervention, outcomes (bleeding, stroke, combined endpoints and all-cause death, the combined endpoints being a combination of all-cause death, all strokes, all bleeding), hypertension, diabetes, previous myocardial infarction, peripheral vascular disease, chronic kidney disease, chronic obstructive lung disease, heart failure, and so forth.  | 599 categorical variables were extracted from baseline features of participants enrolled in each included study. RR (risk ratio) and its 95% CI (confidence intervals) were calculated to combinate the categorical variable data in this meta-analysis. Among combined study results, the degree of inconsistency (I 2 ) were used to assess the heterogeneity. I 2 values of approximately 25%, 50%, and 75% were considered to indicate low, moderate, or high heterogeneity, respectively. A Mantel-Haenszel fixed-effects model was used when I 2 was less than 50%. Otherwise, the Mantel-Haenszel randomeffects model was used. p < 0.05 indicated statistical significance. We also performed a subgroup analysis stratifying study design, according to country. Forest plots were generated to show the relative effect size for each clinical outcome. Publication bias was assessed by examining funnel plot asymmetry and Egger's regression test.

| Quality assessment standard
Sensitivity analysis was used to judge the stability of the ultimate results.

| Search results
After screening 569 articles, 5 studies published before October 6, 2022, were included in this meta-analysis. The flow chart of the literature screening and the selection process is shown in Figure 1 Table 1. In brief, all studies enrolled patients with moderate to severe mitral regurgitation treated with TMVR. The treatment regimen in the control group did not use anticoagulants but SAPT or DAPT. The treatment regimen in the intervention group used anticoagulants, with or without the addition of antiplatelet agents, mainly OAC, except for the use of low molecular weight heparin in some patients in the Geis study. 13

| Quality assessments
The quality of cohort studies was determined according to the NOS scoring system and quality scores all varied from 7 to 8, indicate a low risk of bias. Specific data are shown in Supporting Information: Table 1.

| Data analysis, subgroup analysis
All outcomes were presented in Supporting Information: Table 2.
During the follow-up period, there were 33 cases of bleeding in the intervention group and 31 cases in the control group. There was no statistically significant difference in the incidence between the two groups (RR = 0.76, 95% CI: [0.48−1.22], p = 0.26), with moderate heterogeneity (I 2 = 44%) (details available in Figure 2A).   There were 0 cases of stroke in the intervention group and 4 cases in the control group. There was a statistically significant difference in the incidence between the two groups (RR = 0.14, 95% CI: [0.02−0.77], p = 0.02), with low heterogeneity (I 2 = 0%) (details available in Figure 2B).
There were 6 cases of combined endpoints in the intervention group and 15 cases in the control group. There was no statistically significant difference in the incidence between the two groups (RR = 0.52, 95% CI: [0.10−2.63], p = 0.43), with high heterogeneity (I 2 = 53%) (details available in Figure 2C).
There were 41 cases of all-cause death in the intervention group and 40 cases in the control group. There was no statistically significant difference in the incidence between the two groups (RR = 0.89, 95% CI: [0.58−1.35], p = 0.58), with low heterogeneity (I 2 = 0%) (details available in Figure 2D).  Figure 1A).
There were 0 case of stroke in the OAC group and 1 case in the control group. There was no statistically significant difference in the  Figure 1D).

| Sensitivity analysis and publication bias
In sensitivity analyses (details available in Supporting Information: - Figure 2A,B,D), removing any single study did not change the outcomes for the rates of bleeding, stroke, and all-cause death.
Supporting Information: Figure 2C suggests that the Cammalleri study may have contributed to one of the causes of instability in the combined endpoint analysis model.
Publication bias was assessed using funnel plots (details available in Supporting Information: Figure 3A,B,C,D), Egger's test and Begg's test showed no evidence of publication bias in all the analyses (p = 0.799 and 1.000, 0.846 and 1.000, N/A and 1.000, 0.530 and 0.296 for bleeding, stroke, combined endpoints, all-cause death, respectively).

| DISCUSSION
This meta-analysis demonstrated that anticoagulants (at least 4 weeks duration) following TMVR for patients reduced the rates of stroke without increasing the rates of bleeding compared with no anticoagulants. There were no differences in combined endpoints and all-cause death between groups.
TMVR with the MC system has been a safe and effective treatment for mitral regurgitation in patients who cannot tolerate conventional surgery or are at high risk for surgery, and has been recommended by the 2020 ACC/AHA guidelines 18  Patients with TMVR are at high risk for thrombosis. The three elements of thrombosis, known as the Virchow's triad, include: blood stasis, endothelial injury or vessel wall injury, and hypercoagulability.
Patients with mitral regurgitation have an increased prevalence of atrial fibrillation, with an incidence of up to 63%, 27 and then significant intra-atrial blood stasis occurs in atrial fibrillation. 28  We believe that in patients with a low risk of bleeding after TMVR, it is possible that the application of short-term anticoagulants (at least 4 weeks duration) is the best therapy to avoid thrombosis and reduce the incidence of stroke and bleeding events. However, the optimal therapy (type and duration of anticoagulant drugs) still needs to be further evaluated in larger studies.

| Limitations
Our study has several limitations. First, this analysis contains five observational cohort studies, therefore, it is subject to possible selection bias and confounding, inherent to this type of study designs. Second, the control group was receiving SAPT or DAPT, the intervention group was receiving OACs with SAPT/DAPT/ASA in some patients and only OAC in some. The intervention group was stratification by OAC, OAC plus DAPT, and OAC plus SAPT, but the sample size was small. The lack of detail in the stratification of the intervention group and control group as well as the small sample size led to low confidence in the conclusions. The finding on statistically significant lower risk of stroke is informed by a total of 4 lower stroke events. This also leads to less credible conclusion. Third, the baseline characteristics of the included population were incomplete, regarding prevalence of atrial fibrillation, risk of stroke and information on use of anticoagulants before the procedure Forth, none of the results from the included studies were proper time-to-event analyses.
Simple measures such as RR based on time to event data and studies with variable follow-up cannot be reliably compared.

| CONCLUSION
We observed a reduced risk of stroke without elevated risk of bleeding, combined endpoints, or all-cause death in patients using anticoagulants (at least 4 weeks duration) after TMVR compared to no anticoagulants.
The included studies were small, which could explain the nonsignificant results for the bleeding and mortality endpoints. So the findings in this study are based on observational studies and have to be considered as hypothesis generating. A rigorous, well-designed, large-scale RCT is needed to further validate this finding and guide anticoagulant treatment strategies in this population.