Cost-effectiveness of percutaneous mitral repair for patients with severe secondary mitral regurgitation: an updated evaluation using a modelling approach based on COAPT final data after 5-year follow-up

Abstract

Objectives

To evaluate the cost-effectiveness of percutaneous repair (PR) for secondary mitral regurgitation.

Design

An economic evaluation using a time-varying Markov model comprising three states to assess the cost and effectiveness of PR added to guideline-directed medical treatment (GDMT) compared with GDMT alone. Clinical outcomes considered within the model were overall survival and heart failure (HF) hospitalisations (HFH), and the incremental cost-effectiveness ratio (ICER) was calculated. Cost data were derived from a literature search. Sensitivity analyses were undertaken.

Setting

The French healthcare system perspective assuming a lifetime horizon.

Participants

Published data at 5 years obtained from patients enrolled in the Cardiovascular Outcomes Assessment of the MitraClip Percutaneous Therapy for Heart Failure Patients with Functional Mitral Regurgitation study.

Results

In our base case, we chose cubic spline models to extrapolate overall survival, and we used log-logistic models to estimate cumulative HFH. After discounting, the model generated life-years of 3.843 years and 3.055 years for PR+GDMT and GDMT, respectively. Discounted total quality-adjusted life-year (QALY) values were 2.572 and 1.945 for PR+GDMT and GDMT, respectively (incremental 0.627 QALY). Discounted total costs were €42 709 and €20 732 for the intervention and the control groups, respectively (incremental €21,977), resulting in an ICER of €35,068/QALY. At a threshold of €50 000 per QALY, PR had a 0.85 probability of being cost-effective.

Conclusion

Updated trial data have enabled investigators to provide a more reliable estimation of the ICER, which suggests that PR has good value for money compared with GDMT alone.

Correspondence to Dr Xavier Armoiry; [email protected]

STRENGTHS AND LIMITATIONS OF THIS STUDY

• Our updated analysis, using the most recent survival outcomes, provides a reliable estimation of the cost-effectiveness of percutaneous repair (PR) for decision-making purposes.

• Parametric models of survival and hospitalisation for heart failure were developed to extrapolate beyond observed data.

• Crossover observed in the trial may lead to an overestimation of survival benefit for people in the control arm.

• The observed changing mortality trajectory in the 5-year follow-up after PR required the development of cubic spline models to fit the observed data.

Introduction

Riding the wave of transcatheter aortic valve implantation (TAVI), percutaneous repair (PR) of the mitral valve (MV) has become an increasingly popular procedure to enable minimally invasive treatment of patients presenting mitral regurgitation (MR). While originally confined to patients with primary MR who were deemed at too high risk for MV surgery,^{1 2} the indications of PR have expanded dramatically targeting patients with HF with severe secondary MR (SMR). Following the publication of the Cardiovascular Outcomes Assessment of the MitraClip Percutaneous Therapy for Heart Failure Patients with Functional Mitral Regurgitation (COAPT) study,³ demonstrating a benefit of PR added to medical treatment alone, the MitraClip system (Abbott Vascular, CA, USA) was granted with the approval of the Food and Drug Administration (FDA), which constituted a major step forward as it opened the technique to the largest market worldwide.

A report published in late 2023 reported that the global market of transcatheter MV repair was as high as US$1.22 billion in 2023 and suggested an impressive compound annual growth rate of more than 15% in the next decade.⁴

These recent trends and forecasts represent a major stake for healthcare systems which advocates for a more systematic use of economic evaluation to support decision-making.

Using the 3-year follow-up data from the COAPT study, we reported an economic evaluation of PR with the MitraClip system that emphasised the importance of updating cost-effectiveness analysis as more mature trial data become available.⁵

With the final results from the COAPT study being published,⁶ we have evaluated the cost-effectiveness of PR with the MitraClip system based on the 5-year follow-up data.

Methods

Overview

The study population corresponds to patients enrolled in the COAPT study. Briefly, patients enrolled in the COAPT study had HF and moderate-to-severe SMR, and remained symptomatic despite optimal medical treatment. Unlike patients included in the Mitra-FR trial, who displayed no improved clinical outcomes after PR,⁷ patients enrolled in the COAPT study were characterised by a more disproportionately severe MR compared with the degree of left-ventricle dilatation.⁸

The two compared interventions were PR with the MitraClip system (Abbott Vascular) + guideline-directed medical treatment (GDMT) relative to GDMT alone.

The MitraClip system is a CE-marked class III medical device that has received a positive recommendation for reimbursement for patients with severe MR in several European jurisdictions, including France and the United Kingdom.

Like our previous analyses,^{5 9} we adopted the perspective of the French Health Insurance. Given the absence of long-term outcomes beyond 60 months, lifelong costs and clinical/utility outcomes were modelled. We chose a lifetime horizon of 30 years to ensure that our modelling approach for overall survival would capture a duration sufficiently important so that the entire population of the cohort entering the model reaches the death state.

The model, which was executed using Microsoft Excel, had the same structure (online supplemental figure A) compared with that previously reported.^{5 9} Briefly, it comprised three mutually exclusive health states: alive free of HF hospitalisation (HFH), alive with HFH and dead.

The entire hypothetical cohort enters the model in the alive nonhospitalised state and after each 1-month length cycle, patients could transition to death or to HFH. After HFH, patients would transition back to the alive nonhospitalised state.

Monthly transition probabilities from alive nonhospitalised to death state were derived from overall survival while those from alive nonhospitalised to HFH state were obtained from cumulative HFH curves as reported in the COAPT main papers.^{3 6 10} Guyot et al’s¹¹ method was used to reconstruct individual patient data (IPD) for cumulative HFH in the COAPT trial at 5-year follow-up.⁶ The method to fit parametric models for HFH was as previously described.⁹

We also reconstructed IPD for overall survival (all-cause mortality) at 5-year follow-up and developed Kaplan-Meier plots for each arm (figure 1). Next, we fit parametric models using Stata V.15.0 (StataCorp, College Station, TX, USA) and determined the best model based on goodness of fit as well as on clinical plausibility. In the GDMT-only arm, crossover that occurs from GDMT to the GDMT+MitraClip regimen makes modelling difficult. We therefore used the base-case survival model as a guide to what was clinically plausible.¹² Baron et al modelled survival in the GDMT arm using an observation period before crossover could influence since nearly all crossovers occur during subsequent years of observation. Baron et al’s model suggests few survivors beyond 14 years. Clinical plausibility for modelling overall survival for the GDMT+PR arm was based on the survival seen in two real-world studies^{13 14} and their modelled extrapolation beyond 5-year observation (online supplemental figures S1a–c). Both these studies suggest few survivors beyond 10–14 years. This is also in line with the base-case economic model¹² (online supplemental figures S2a,b).

Figure 1. Reconstructed KM of cumulative mortality to 30 months and 60 months based on 5- and 3-year follow-up of COAPT. COAPT, Cardiovascular Outcomes Assessment of the MitraClip Percutaneous Therapy for Heart Failure Patients with Functional Mitral Regurgitation; GDMT, guideline-directed medical treatment; KM, Kaplan-Meier; PR, percutaneous repair.

Given these estimates of clinical plausibility that were confirmed by our clinical expert, we then fit parametric models for overall survival in each arm rejecting models that were clinically implausible and judging between those that were plausible according to Akaike information criterion (AIC) Bayesian Information Criterion (BIC) values and visually fit to the observed data. The Kaplan-Meier plot for the GDMT+PR arm has a distinct downturn in survival at about 26 months and standard parametric models score poorly on ACI BIC values, but cubic spline models score better on AIC BIC scores and some provide good visual fit to the observed data while also providing clinically plausible extrapolation beyond the observed data (online supplemental figures S3a–c); this led to the final selection of the spline model shown in figure 2. To be equitable and consistent with other economic models, we also applied this spline model for the GDMT arm (first 30 months of data). We further constrained the GDMT model by not allowing survival in the GDMT arm to be better than that in the GDMT+PR arm (figure 2).

Figure 2. Cubic spline models fit to 5-year follow-up of the PR+GDMT arm and to the first 30 months of the 5-year follow-up of the GDMT arm. GDMT, guideline-directed medical treatment; PR, percutaneous repair.

Costs and utilities

In addition to the resource utilisation that we considered in our previous analyses (MitraClip system and index stay, hospitalisation for HF, outpatient medical treatment and medical visit), the occurrence of left ventricular assistance device (LVAD) implantation and heart transplantation was taken into account based on the values reported by Stone et al.⁶ The corresponding costs were valued based on the 2023 official fares (online supplemental table A).

We used the same utility inputs that we used previously^{5 9} which were differentiated clinical outcomes depending on the occurrence of unplanned HFH.

Analysis

The economic analysis was undertaken according to the best practice guidelines.^{15 16} We calculated the incremental cost-effectiveness ratio (ICER) which was expressed as monetary costs (€) per quality-adjusted life-year (QALY) gained.

A 2.5% annual discount was applied for both costs and benefits consistent with the national guidelines.¹⁷

We undertook univariate sensitivity analyses based on the cost per QALY, with the results reported using a tornado diagram.

We undertook a multivariate sensitivity analysis by widely modifying model inputs as follows: (a) overall survival, 95% lower confidence interval (LCI) and upper confidence interval (UCI) for parametric models of overall survival for each arm; (b) ±20% on the number of monthly hospitalisations for HFH for each arm; (c) ±20% on the major costs (consisting of the monthly cost of new HFH and the cost of the MitraClip device+index stay) and (d) ±20% on the utility values. This procedure generated up to 16 pairs of costs (euros) and benefits (life years gained (LYG) or QALYs) for each arm. These were then used in bootstrapping¹⁸ with 1000 iterations to generate the joint distributions of incremental costs and QALYs. The results were graphed using the Stata Ellip package¹⁹ and used to generate cost-effectiveness acceptability curves (CEAC).

Patient and public involvement

No patient was involved.

Results

Parametric Modelling of overall survival

Figure 1 shows the reconstructed all-cause mortality plot for each arm of COAPT at 5-year follow-up and that based on 3-year follow-up (see also online supplemental figures S4a,b).

The PR+GDMT plot is characterised by a gradually decreasing mortality rate to about 26 months followed by a distinct upturn in mortality trajectory that extends to 5 years. In contrast, the plot for the GDMT arm exhibits a fairly linear trajectory for about 2–3 years followed by a gradually decreasing mortality rate to 5 years thereafter.

These trajectories imply that with extrapolation beyond 5 years, the survival in the GDMT arm will become superior to that in the PR+GDMT arm. However, such interpretation of the 5-year data is complicated because crossover to PR+GDMT from GDMT after 2 years of treatment⁶ benefits the GDMT arm, so the decreasing mortality rate in the GDMT arm seen over 5-year follow-up is influenced by crossover.

Modelling all-cause mortality over 5 years and beyond to a lifetime horizon presents challenges because of the changing trajectory in the PR+GDMT plot and the occurrence of crossover in the GDMT arm. We examined two different modelling approaches. In our base case, we used parametric modelling as it has been used in previously published cost-effectiveness studies.^{5 20 21} As an exploratory analysis, we used the nonparametric method described by Baron et al,¹² based on COAPT 2-year observed data with extrapolation of the GDMT using age- and gender-matched US general population life tables followed by the application of an HR to the GDMT arm to obtain extrapolation for the PR+GDMT arm (see online supplemental file 2 for details).

Modelling for the GDMT arm

Standard parametric models or cubic splines fit to the 5-year GDMT data generate unrealistic survivors beyond 20 years (data available on request). The GDMT survival plot appears linear to 30 months (online supplemental figure S5). From the 2- to 5-year follow-up of the GDMT arm, 64 crossover events occurred at various time intervals over 3 years (online supplemental 6 and figure S6). Any beneficial influence on survival will take time to materialise; in the absence of contradictory evidence, we assumed that during the first 30 months, survival in the GDMT arm has been minimally influenced by crossovers. Parametric or spline models fit well to 30-month data and they predict similar survival on extrapolation with very few survivors beyond 15 years. Notably, this result aligns with 2-year follow-up base-case GDMT model of Baron et al’s publication of additional material (figure 3)¹² that generates very few survivors beyond 15 years and in which crossover is minimal.

Figure 3. Tornado diagram (one-way sensitivity analyses). The range of variation for the cost of PR was +-20% (ie, €16 880 to 25 320). GDMT, guideline-directed medical treatment; HF, heart failure; LVAD, left ventricular assistance device; PR, percutaneous repair.

Modelling for the PR+GDMT arm

The change in trajectory in the PR+GDMT arm at approximately 26 months (online supplemental figure S4a,b) means that the usual battery of parametric models fails to provide a good visual fit to one or other or both phases of trajectory (data available on request). On extrapolation (data available on request), most standard parametric models predict unlikely proportions of survivors beyond 20 years. Exceptions are models with a single parameter (Rayleigh and exponential), but these suffer from poor visual fit, and models with three parameters (bathtub and ggamma) with slightly improved visual fit. These results suggest that models with more than three parameters, as provided by cubic splines, would generate superior models and better visual fit to the 5-year COAPT trial results and will generate clinically plausible lifetime extrapolation for the PR+GDMT arm. Cubic spline models with increasing df fit to the PR+GDMT in-trial data are shown in online supplemental figure S3b. With three df, the spline model resembles that obtained with the three-parameter bathtub and ggamma models, splines with 4 and 5 df generate overgenerous survivors at 20-year follow-up; with further df (6, 7, 8 and 9), very similar extrapolations are obtained with few survivors beyond 10-year follow-up. We, therefore, selected the spline model with 6 df and applied this also to the first 30 months of the 5-year follow-up of the GDMT arm (to minimise the influence of crossovers). The resulting models are shown in figure 2. After approximately 85 months, the survival in the GDMT arm becomes superior to that in the PR+GDMT arm; for base-case economic modelling, survival in the GDMT arm beyond 85 months was constrained to equal that in the PR+GDMT arm and generates 0.863 LYG by PR+GDMT over GDMT (0% discounting; 0.773 with 3% discounting).

Our exploratory analysis following the nonparametric method of Baron et al¹² used Baron’s base-case, best-case and worst-case scenarios for the PR+GDMT models. Of these, the worst-case scenario model aligns remarkably well with the observed COAPT data for 5 years (online supplemental figure S2a,b) whereas base case and best case infer very optimistic benefit from PR when compared with 5-year follow-up survival.

Parametric modelling of HFH

As mentioned previously, we used log-logistic models to estimate cumulative hospitalisation for HF (online supplemental figure S7).

Cost-effectiveness estimates

After discounting, the model generated LYs of 3.843 years and 3.055 years for the PR+GDMT and GDMT groups, respectively, resulting in an incremental 0.788 LYs.

Employing utility values for health states, discounted total QALY were 2.572 and 1.945 for PR+GDMT and GDMT, respectively, resulting in an incremental 0.627 QALY.

Total discounted costs were €42 709 and €20 732 for the intervention and the control groups, respectively (incremental of €21 977 between the two strategies), resulting in a base-case ICER of €35 068/QALY gained (discounted value).

Excluding inputs for LVAD implantation and heart transplantation in the model resulted in a decrease of total costs (€32 278 and €6861 for the intervention and the control groups, respectively, incremental of €25 417) and slightly increased the ICER (€40 557/QALY).

Univariate sensitivity analyses (figure 3) showed that the two most influential variables in the model were the overall survival in the PR strategy and the cost of PR.

In multivariate sensitivity analysis, the 1000 bootstrap iterations were dispersed in the northeast quadrant mostly below the threshold willingness to pay (WTP) of €50 000/QALY; the CEAC indicated that at this threshold PR+GDMT had a 0.85 probability of being cost-effective relative to GDMT (figure 4). The bootstrap ICER was €39 127/QALY.

Figure 4. Cost-effectiveness plane after multivariate sensitivity analysis and comparison with deterministic ICER. ICER, incremental cost-effectiveness ratio; QALY, quality-adjusted life-year.

Discussion

Our cost-effectiveness analysis showed that using the most up-to-date survival outcome from COAPT, PR added to GDMT compared with GDMT alone was associated with an ICER of approximately €35 000/QALY gained and approximately €39 000/QALY based on deterministic and multivariate analyses, respectively.

These estimations are consistent with our previous analyses⁵ which relied on less mature data (3 years) in contrast to the current study, where COAPT data at 5 years were used.

At the commonly admitted WTP of €50 000/QALY, our analyses indicate a high likelihood of PR being cost-effective, despite the considerable cost of the MitraClip System (€21 100 as of March 2024, with an additional cost of approximately €5000 for the index stay during implantation).

Results from the COAPT study indicate a significant benefit of PR in reducing the incidence of HFH, which from the economic viewpoint could translate into reduced hospital costs. However, the incremental cost between the two strategies remained substantial, approximately €22 000, indicating that the cost associated with the intervention (MitraClip system+index stay) is only marginally offset by reduced HFH costs. The economic impact of PR added to GDMT may become more evident when considering LVAD and heart transplantation occurring over time. Indeed, these events were observed less frequently in the device group at 5 years, which means that the inclusion of these events in the economic analysis contributed to slightly reduced incremental costs (approximately €22 000/€25 400, respectively, including/excluding these events).

The overall acceptable level of ICER is therefore primarily attributed to the substantial incremental effectiveness that we modelled at approximately 0.79 LY and 0.63 QALY, representing a gain of +26% and +32% in terms of life expectancy and QALY, respectively.

The Baron 2019 base-case economic model estimated a best-case ICER of US$27 733/QALY, a base-case ICER of US$55 600/QALY and a worst-case scenario ICER of US$70 592/QALY, indicating the strong influence of survival modelling on CE estimates. Although US costs incurred over 5 years will differ from those over the 2-year follow-up, it seems clear from the analyses of survival that ICER estimates will probably be higher when the 5-year follow-up mortality is taken into account.

Our study has strengths and limitations.

The major strength is that our cost-effectiveness estimate was derived from the most recent survival outcomes available. Since COAPT 5-year follow-up corresponds to the maximal duration that was planned, this reduces the likelihood of new survival analyses emerging and provides a robust foundation for our cost-effectiveness analysis.

Our analysis does support the cost-effectiveness of PR using the MitraClip system compared with GDMT alone in patients with severe SMR who present echocardiographic characteristics similar to those enrolled in the COAPT study. A limitation is that crossover may lead to overestimation of survival benefit for the GDMT arm; however, because we used observed survival to 2.5 years only, we believe this bias is essentially nullified. A

further potential limitation, dictated by the observed changing mortality trajectory in the 5-year follow-up after MitraClip, is that cubic spline models may overfit the data. However, our results in terms of LYG are similar to those from the German perspective by Estler et al,²² who made use of 2-year follow-up data from the COAPT study unlikely to be influenced by crossover.

Finally, as with any economic evaluation using a modelling approach, our model provides a simplified representation of the patient’s clinical pathway. For instance, a patient who has experienced a first HFH is more likely to have a second episode compared with a patient who has not, which could lead to reduced accuracy compared with an ideal scenario with access to IPD. However, from a cost-effectiveness standpoint, our model does incorporate the actual number of HFH events observed in the COAPT trial.

On submission, we became aware of the interesting study of Yao et al²³ that points to the obvious risk of bias arising from crossover in COAPT and uses 5-year follow-up from COAPT to reanalyse cost-effectiveness from the US, German and UK healthcare system perspectives (online supplemental table 2). The model structure in Yao et al differs from ours in being based on New York Heart Association (NYHA) subclass occupancy and in employing different survival modelling. Because of the post-2-year upturn in mortality in the PR MitraClip arm of COAPT, we would expect Yao et al’s estimates of LYG for this arm to be less than in the original US, UK and German studies referenced in Yao et al (ie, Baron et al,¹² Cohen et al²⁴ and Estler et al,²² respectively) that use 2-year follow-up data with favourable extrapolation relative to what is observed for 5 years in COAPT. Yao et al’s estimates (online supplemental table 2) of LY gained in the PR arm can be compared with the worst-case scenarios of the original studies for the US perspective by Baron et al¹² and similarly for the UK perspective by Cohen et al²⁴ because these survival models conform most closely to the survival observed for 5 years in COAPT (online supplemental figure S2b) while base-case and best-case models greatly overestimate benefit. From the US perspective, Yao et al’s estimate is slightly larger (by 3%) than the original study (worst case); from the UK perspective, Yao et al’s estimate is about 2% smaller than the Cohen worst-case scenario, but the latter overestimates survival relative to the 5-year follow-up of COAPT (online supplemental figure S7). In the case of the German perspective, Yao et al’s estimate is larger (by 36%) than that of the original German study (Estler et al²²).

In short, relative to the observed 5-year follow-up data from COAPT, Yao et al’s estimates for the PR arm conform reasonably to the worst-case scenarios of Cohen and Baron, but seem at odds in the case of the German perspective. Estler’s base-case ICER of €59 728/QALY becomes €25 824/QALY in the Yao reanalysis. Yao et al ICERs for the US and UK perspectives (£28 910/QALY and $71 199/QALY) are almost identical to the worst-case ICERs of the original studies (UK£28 607/QALY and US$70 592/QALY, respectively).

Compared with our preliminary analyses which used COAPT follow-up data at 2 years,⁹ this work provides more reliable estimates on the cost-effectiveness of PR. This highlights the importance of adopting a continuous approach to updating economic evaluations as more mature survival outcomes become publicly available, ensuring that analyses remain reflective of the most current evidence.

We would like to emphasise that this work does only apply to patients with severe SMR as enrolled in the COAPT trial, which means that our conclusions cannot be generalised to all patients with SMR. A similar work could be undertaken based on the recent findings of the RESHAPE-HF2 trial.²⁵

The follow-up of patients enrolled in COAPT who underwent PR emphasised an increase of mortality rate from 26 months post-randomisation, a trend that persisted up to 5 years. This shift in trajectory was also noted in other studies^{13 14} (online supplemental 1 and 4), raising concerns on the durability of the benefit of PR observed in terms of overall survival. To address this concern, we are planning to extend our research as part of a systematic review that will comprehensively examine the long-term survival outcomes of patients following PR.²⁶

Data availability statement

Data sharing not applicable as no datasets generated and/or analysed for this study. We used data available in the public domain.

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

This study involves human participants but was not approved by This study used data available in the public domain. This study used data available in the public domain.

Footnote

Contributors MC acted as guarantor. MC: (1) substantial contributions to conception and design, or acquisition of data, and interpretation of data; (2) drafting the article and (3) final approval of the version to be published. PA: (1) substantial contributions to analysis and interpretation of data; (2) revising the draft critically for important intellectual content and (3) final approval of the version to be published. JF-O: (1) substantial contributions to analysis and interpretation of data; (2) revising the article critically for important intellectual content and (3) final approval of the version to be published. XA: (1) substantial contributions to conception and design, or acquisition of data, and interpretation of data; (2) drafting the article and (3) final approval of the version to be published.

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 JFO is the principal investigator of the Mitra-FR study and reports personal fees from Abbott Vascular, Delacroix-Chevalier, Landanger, and Medtronic, outside the submitted work. XA is member of the steering committee of the Mitra-FR study.

Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Provenance and peer review Not commissioned; externally peer reviewed.

Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

References

¹ Feldman T, Kar S, Elmariah S, et al. Randomized Comparison of Percutaneous Repair and Surgery for Mitral Regurgitation: 5-Year Results of EVEREST II. J Am Coll Cardiol 2015; 66: 2844–54. doi:10.1016/j.jacc.2015.10.018

² Kar S, Feldman T, Qasim A, et al. Five-year outcomes of transcatheter reduction of significant mitral regurgitation in high-surgical-risk patients. Heart 2019; 105: 1622–8. doi:10.1136/heartjnl-2017-312605

³ Stone GW, Lindenfeld J, Abraham WT, et al. Transcatheter Mitral-Valve Repair in Patients with Heart Failure. N Engl J Med 2018; 379: 2307–18. doi:10.1056/NEJMoa1806640

⁴ Transcatheter mitral valve repair (TMVR) market size by segments, share, regulatory, reimbursement, procedures and forecast to 2033. Available: https://www.globaldata.com/store/report/transcatheter-mitral-valve-repair-devices-market-analysis/ [Accessed 29 Feb 2024 ].

⁵ Connock M, Auguste P, Obadia J-F, et al. Impact of updated trial data on the cost-effectiveness of percutaneous mitral repair. PLoS One 2023; 18: e0280554. doi:10.1371/journal.pone.0280554

⁶ Stone GW, Abraham WT, Lindenfeld J, et al. Five-Year Follow-up after Transcatheter Repair of Secondary Mitral Regurgitation. N Engl J Med 2023; 388: 2037–48. doi:10.1056/NEJMoa2300213

⁷ Obadia J-F, Messika-Zeitoun D, Leurent G, et al. Percutaneous Repair or Medical Treatment for Secondary Mitral Regurgitation. N Engl J Med 2018; 379: 2297–306. doi:10.1056/NEJMoa1805374

⁸ Messika-Zeitoun D, Iung B, Armoiry X, et al. Impact of Mitral Regurgitation Severity and Left Ventricular Remodeling on Outcome After MitraClip Implantation: Results From the Mitra-FR Trial. JACC Cardiovasc Imaging 2021; 14: 742–52. doi:10.1016/j.jcmg.2020.07.021

⁹ Armoiry X, Obadia J-F, Auguste P, et al. Conflicting findings between the Mitra-Fr and the Coapt trials: Implications regarding the cost-effectiveness of percutaneous repair for heart failure patients with severe secondary mitral regurgitation. PLoS One 2020; 15: e0241361. doi:10.1371/journal.pone.0241361

¹⁰ Mack MJ, Lindenfeld J, Abraham WT, et al. 3-Year Outcomes of Transcatheter Mitral Valve Repair in Patients With Heart Failure. J Am Coll Cardiol 2021; 77: 1029–40. doi:10.1016/j.jacc.2020.12.047

¹¹ Guyot P, Ades AE, Ouwens MJNM, et al. Enhanced secondary analysis of survival data: reconstructing the data from published Kaplan-Meier survival curves. BMC Med Res Methodol 2012; 12: 9. doi:10.1186/1471-2288-12-9

¹² Baron SJ, Wang K, Arnold SV, et al. Cost-Effectiveness of Transcatheter Mitral Valve Repair Versus Medical Therapy in Patients With Heart Failure and Secondary Mitral Regurgitation: Results From the COAPT Trial. Circulation 2019; 140: 1881–91. doi:10.1161/CIRCULATIONAHA.119.043275

¹³ Velu JF, Kortlandt FA, Hendriks T, et al. Comparison of Outcome After Percutaneous Mitral Valve Repair With the MitraClip in Patients With Versus Without Atrial Fibrillation. Am J Cardiol 2017; 120: 2035–40. doi:10.1016/j.amjcard.2017.08.022

¹⁴ Adamo M, Fiorelli F, Melica B, et al. COAPT-Like Profile Predicts Long-Term Outcomes in Patients With Secondary Mitral Regurgitation Undergoing MitraClip Implantation. JACC Cardiovasc Interv 2021; 14: 15–25. doi:10.1016/j.jcin.2020.09.050

¹⁵ Husereau D, Drummond M, Augustovski F, et al. Consolidated Health Economic Evaluation Reporting Standards 2022 (CHEERS 2022) Statement: Updated Reporting Guidance for Health Economic Evaluations. Value Health 2022; 25: 3–9. doi:10.1016/j.jval.2021.11.1351

¹⁶ Philips Z, Ginnelly L, Sculpher M, et al. Review of guidelines for good practice in decision-analytic modelling in health technology assessment. Health Technol Assess 2004; 8: ix–xi,. doi:10.3310/hta8360

¹⁷ Haute Autorité de Santé. Guide Méthodologique: Choix méthodologiques pour l’évaluation de l’efficience à la HAS Version soumise à la consultation publique. 2019.

¹⁸ Gallacher D, Achana F. Assessing the Health Economic Agreement of Different Data Sources. Stata J Promoting comm on stats and Stata 2018; 18: 223–33. doi:10.1177/1536867X1801800114

¹⁹ Alexandersson A. Graphing Confidence Ellipses: An Update of Ellip for Stata 8. Stata J Promoting comm on stats and Stata 2004; 4: 242–56. doi:10.1177/1536867X0400400302

²⁰ Shore J, Russell J, Frankenstein L, et al. An analysis of the cost-effectiveness of transcatheter mitral valve repair for people with secondary mitral valve regurgitation in the UK. J Med Econ 2020; 23: 1425–34. doi:10.1080/13696998.2020.1854769

²¹ National Guideline Centr (NICE). Cost-utility analysis: transcatheter mitral edgeto-edge repair for inoperable patients. NICE guideline NG208. 2021.

²² Estler B, Rudolph V, Seleznova Y, et al. Cost-effectiveness of the MitraClip device in German heart failure patients with secondary mitral regurgitation. Eur J Health Econ 2023; 24: 349–58. doi:10.1007/s10198-022-01476-4

²³ Yao Y, Zhou Z, Geng T, et al. Long-term cost-effectiveness of transcatheter mitral valve repair in HF patients with secondary mitral regurgitation. ESC Heart Fail 2024; 11: 2023–32. doi:10.1002/ehf2.14729

²⁴ Cohen DJ, Wang K, Magnuson E, et al. Cost-effectiveness of transcatheter edge-to-edge repair in secondary mitral regurgitation. Heart 2022; 108: 717–24. doi:10.1136/heartjnl-2021-320005

²⁵ Anker SD, Friede T, von Bardeleben R-S, et al. Transcatheter Valve Repair in Heart Failure with Moderate to Severe Mitral Regurgitation. N Engl J Med 2024; 391: 1799–809. doi:10.1056/NEJMoa2314328

²⁶ Armoiry X, Connock M, Madani R, et al. Assessing long-term survival outcomes after percutaneous repair for patients with severe mitral regurgitation. PROSPERO 2023 CRD42023429527.