Transcatheter Mitral Valve-in-Valve and Valve-in-Ring Replacement in High-Risk Surgical Patients: Feasibility, Safety, and Longitudinal Outcomes in a Single-Center Experience

Abstract: Background. Transcatheter mitral valve-in-valve (TMVIV) or valve-in-ring (TMVIR) replacement has shown early promise in patients deemed poor surgical candidates as a less invasive alternative to conventional reoperative mitral valve (MV) replacement. Objective. This retrospective, single-center study reviewed the procedural outcomes of all TMVIV and TMVIR procedures between 2013-2018 at a large tertiary referral center serving the southeastern United States. An analysis of patient safety measures was also performed, with a retrospective assessment of relative procedural safety that included preoperative risk stratification and postoperative mortality predictors, operative time, average blood loss, length of hospital stay, and readmission rates. Methods. This study included 24 patients with severe MV disease and medical comorbidities who were considered too high risk for conventional MV replacement. All patients underwent TMVIV or TMVIR with Edwards Sapien XT or S3 transcatheter valves (Edwards Lifesciences). A secure database of patient demographics, preoperative risk assessment, and procedural data was created and included technical success rates, blood loss, operative time, and intraoperative and immediate postoperative complications. Subsequent follow-up of patient outcomes reported here include those collected at 30 days, 180 days, and 1 year. Results. Of the 24 patients in our study, 16 received TMVIV and 8 received TMVIR implantation. Each procedure was performed successfully under general anesthesia via transseptal approach (n = 17) or transapical approach (n = 7), with only 1 patient (8.2%) requiring late operative reintervention 252 days post op. Average procedural time was 76 min and average blood loss was <75 mL, with 20/24 patients (83%) successfully extubated on postoperative day 0. Length of Intensive Care Unit stay was 1.7 ± 1.4 days and length of total inpatient stay was 2.8 ± 1.8 days. Echocardiograms were collected immediately post op, at 30 days, at 180 days, and subsequently at yearly intervals; follow-up demonstrated excellent prosthetic valve function with low transvalvular gradients, and no evidence of valve embolization or thrombosis. In patients with follow-up data available at 1 year (n = 13), there were no readmissions at 30 days or 180 days, and only 1 admission (8.3%) during the first postoperative year for symptoms related to congestive heart failure (CHF). Conclusion. TMVIV and TMVIR can be safe and effective in a patient population considered at prohibitive risk for conventional surgery. These procedures can be performed efficiently in a hybrid operating room, with relatively short procedural times and high rates of early extubation. Procedural complications, mortality, and readmission rates for CHF at 30 days, 180 days, and 1 year were very low in this high-risk cohort.  

J INVASIVE CARDIOL 2018;30(9):324-328. Epub 2018 June 15.

Key words: mitral valve, transcatheter aortic valve replacement, minimally invasive, procedural outcomes

Over the last 5 years, a minimally invasive transcatheter approach to the management of hemodynamically significant valve disease has developed into an excellent alternative for patients whose medical comorbidities place them at prohibitive risk for conventional invasive valve replacement surgery. Experience in transcatheter valve replacement thus far has largely involved the aortic valve (TAVR), which has been approved for patients with intermediate surgical risk¹ and high surgical risk, and for those who are inoperable.² Furthermore, transcatheter valve replacement has proven successful in the treatment of degenerated surgical prosthetic valves, and has been approved for treatment of stenotic or regurgitant prosthetic aortic valves in higher-risk patients.^3,4 

Reoperative mitral valve replacement (MVR) is a complex and invasive procedure; the technical challenges of reentering the chest are often compounded by the frailty and medical complexities of the patient. A growing interest in a transcatheter approach to management of mitral valve disease in patients with failed mitral valves after previous cardiac surgery^5,6 has evolved, including transcatheter mitral valve-in-valve (TMVIV) implantation and valve-in-ring (TMVIR) replacement.^5,6 TMVIV procedures (but not TMVIR procedures) have been approved by the United States Food and Drug Administration for the treatment of degenerated surgical prostheses, but remain in the early stages of clinical experience in this high-risk patient population.

Early experiences with transcatheter mitral valve replacement (TMVR). Thus far, procedural experiences with TMVR have been limited to small case studies^7,8 and ongoing clinical trials. The MITRAL (Mitral Implantation of Transcatheter Valves) trial recently demonstrated early outcomes in patients who underwent TMVR with Edwards Sapien XT and Sapien 3 transcatheter heart valves (Edwards Lifesciences) for severe mitral annular calcification, failing surgical rings, or bioprosthetic valves.⁹ 

At our center, we have performed TMVIV and TMVIR procedures on prohibitive-risk patients since 2013. In this study, we sought to retrospectively examine and report the relative safety of these procedures, with the goal of adding to and expanding current early-outcomes data. We performed a longitudinal, multiyear, retrospective quality review of all TMVRs at a large referral center serving the southeastern United States from 2013-2018.                               

Methods

Registry design. A secure registry of 24 patients who underwent TMVIV/TMVIR between the years of 2013-2018 was created as part of a retrospective quality review study of TMVRs and subsequent patient outcomes. All patients had been previously evaluated preoperatively by a multidisciplinary structural heart team at our institution consisting of structural heart cardiologists, cardiothoracic surgeons, and advanced practice clinicians. As part of this evaluation, patients underwent preoperative cardiac catheterization to screen for preexisting coronary artery disease, transthoracic echocardiogram (TTE) for characterization of valvular pathology, pulmonary functions tests, and panoramic chest x-rays. Patients received preoperative cardiac computed tomography (CT) scans for the purpose of optimal preoperative characterizations of chamber, septal, and valvular anatomy and movement; preoperative screening for neo left ventricular outflow tract (LVOT)¹⁰ by cardiac CT was available since early 2017. 

Retrospective quality review. Patient charts were reviewed with attention to the following: (1) demographics, including age and gender; and (2) past medical history, with regard to previous diagnoses of type II diabetes, chronic obstructive pulmonary disease, need for home oxygen, chronic renal insufficiency, and previous cardiac history including atrial fibrillation, previous coronary artery bypass graft (CABG) surgery, and  previous valve replacement surgeries. Additional noted preoperative factors included New York Heart Association (NYHA) failure classification, preoperative ejection fraction (EF) and pulmonary artery pressure, and mechanism of mitral valve pathology necessitating transcatheter replacement. Assessments of preoperative risk of mortality from valve replacement were made using the online Society of Thoracic Surgeons (STS) Adult Cardiac Surgery Risk Calculator (https://riskcalc.sts.org/stswebriskcalc/#/calculate). An additional four-point preoperative frailty assessment^11,12 was used to predict risk of mortality in TMVR procedures. Lastly, number of admissions for heart-failure related reasons was quantified by chart review for 1 year and 5 years prior to the index procedure for the purpose of comparisons following the index procedure.

Procedure descriptions. All TMVIV and TMVIR procedures were performed at our institution in a specially equipped hybrid suite with a heart-lung machine available in the room in case of hemodynamic compromise. Procedures were performed under general anesthesia, as previously described.^13,14 Measurements of the inner diameter of the bioprostheses were obtained by preoperative CT angiography. Details of prior surgical prostheses were obtained through medical records and confirmed with vendors. Valve sizing was done in accordance with published recommendations using the Valve in Valve (Mitral) app developed by Vinayak Bhapat, MD (UBQO Limited). Sapien XT or S3 valves were used in all cases. The valve was introduced through a transapical or transseptal approach, positioned, and then deployed into the failed mitral bioprosthesis under fluoroscopic and echocardiographic guidance.¹⁴ A final transesophageal echocardiogram (TEE) was then performed prior to conclusion of the case to assess final position, presence of central or paravalvular leaks, transmitral gradients, and motion of valvular leaflets after valve placement. This strategy allowed for the detection of potential acute complications (eg, right-to-left interatrial shunt, LVOT obstruction, or tamponade). Follow-up TTEs were employed prior to discharge and at each outpatient appointment to screen for late onset of these complications, in addition to evaluating for valve dehiscence, thrombosis, or embolism. 

Procedural quality review and analysis of postoperative outcomes. Included within a secure, HIPAA-compliant database were procedural details specific to the TMVR procedure, including type of valve procedure (TMVIV or TMVIR), approach employed (transapical vs transseptal), operative time (defined as the amount of time that elapsed between the documented first inoperative percutaneous needle stick and the final removal of the catheter sheath after device delivery), and average blood loss during the course of the procedure. Postprocedural TEE was also employed to assess position of the valve, to measure valvular associated gradients, and for EF documentation. Acute procedural success was determined by delivery of the Sapien valve into position via primary method chosen preoperatively. Additional intraoperative events and postoperative complications were assessed during the hospital stay, and included need for operative valvular reintervention at any time, either during initial postoperative inpatient stay and/or during the subsequent follow-up (up to 3 years) for this study. Postoperative TTEs were performed at 24-48 hours post procedure, and were performed at outpatient follow-up exams at 1-month and 6-month intervals initially, and at annual office visits thereafter. Longitudinal outcomes were documented at each visit, and included hospital readmissions, evidence of LVOT obstruction, valve embolization or thrombosis by TTE, myocardial or cerebrovascular insults, acute kidney risk, injury, or renal failure with or without a new dialysis requirement, and documentation of hemolytic anemia. All causes of mortality were also noted.

Study design and statistical analysis. This retrospective quality review study was previously approved by the institutional review board at Eastern Virginia Medical School. Categorical data presented herein were tabulated and are presented as frequencies; continuous variables are expressed as mean ± standard deviation or median (range) where appropriate. Estimates of overall survival were determined using Kaplan-Meier non-parametrics. P-values <.05 were considered statistically significant. 

Results

Demographics and preoperative clinical assessments. During the period of this assessment, there were no patients excluded from consideration of TMVR secondary to valve or ring type. Extensive preoperative assessments of medical comorbidities for each patient were reviewed (Table 1A). In this cohort, the average age was 68.7 ± 14.2 years (range, 23-91 years at time of initial presentation), 50.0% were female, and average STS score was 8.9 ± 5.2. The symptomatic severity of all patients in our study met NYHA class II or higher. A four-part assessment of frailty^11,12 was also performed to assess postoperative prognostics and mortality risk in each patient; the majority of patients in this study (95.6%) met functional criteria in two or fewer categories of all categories presented (Table 1A), indicating increased risk of all-cause mortality at 30 days.^11,12 Additionally, in the 12 months preceding intervention, all patients had at least 2 admissions for heart-failure related causes (range, 2-8 admissions, with 4.8 ± 1.3 admissions/patient; in the 5 years preceding intervention, there were 7.5 ± 2.8 admissions/patient) (Table 1A; n = 24).

Procedural outcomes. Of the 24 patients who underwent TMVR at our institution between 2013-2018, the majority (65.2%) underwent TMVIV implantation for a failed bioprosthetic mitral valve, while the remaining patients underwent TMVIR replacements secondary to failed annuloplasty rings (Table 1B). Two patients underwent planned TAVR for coexisting aortic stenosis during the same procedure as the planned TMVIR or TMVIV replacement via transseptal approach through the right femoral artery. The majority of patients (78.1%) were treated transseptally, with the remainder treated transapically (Table 1B). The first 7 patients underwent transapical approach; subsequently, we converted to a transseptal approach in April of 2015. Acute procedural success (defined as the completion of the initial TVR procedure via access route planned without conversion to an open approach) was achieved in 100% of patients, with all cases performed under general anesthesia. There were no transseptal cases that were converted to a transapical approach. Additional procedural assessments included mean postoperative gradient (3.5 ± 2.1 mm Hg; n = 24) and postoperative ejection fraction (50.6 ± 11.1%; n = 24) measured immediately by TEE.  

Intraoperative and inpatient safety and quality measures. Average operative time was 76 ± 11.2 min and average procedural blood loss was 72 ± 50.6 mL. There were no significant differences between operating times for TMVIV or TMVIR procedures (77.5 ± 30.3 min vs 69.4 ± 34.5 min, respectively; P>.05) or amount of blood lost (62.2 ± 25.9 mL vs 69.2 ± 21.5 mL, respectively; P>.05). All but 4 patients (15.9%) were successfully extubated on postoperative day 0. Patients with extended ventilator dependence beyond this had known respiratory comorbidities, with 3 of the 4 patients requiring continuous home oxygen, as noted preoperatively. Average length of Intensive Care Unit stay was 1.7 ± 1.4 days, and total inpatient stay was 2.8 ± 1.8 days (Table 1B; n = 24). During the course of postoperative inpatient stay, there were no deaths, no patient required operative reintervention, and 3 patients (12.5%) developed stage 1 acute kidney injury (defined as 0.3 mg/dL increase in creatinine or 50% increase over baseline creatinine within 48 hours).¹⁵ All cases of acute kidney injury resolved with conservative management. Additionally, 2 patients (8.2%) developed postoperative anemia that required transfusion and 1 patient (4.1%) developed a new arrhythmia (atrial fibrillation). There were no new myocardial infarctions, ischemic or hemorrhagic cerebrovascular accidents, or acute hemodialysis requirements for any patient during the postoperative inpatient recovery period (Table 1B; n = 24).                                                                 

Assessment of postoperative readmissions for heart failure. Following intervention, there were no readmissions for any reason in the first 30 days (Table 2). For the 19 patients with data for 180 days, we saw no heart-failure related readmissions (Table 2). Three admissions in different patients were seen for non-heart-failure related reasons (sepsis at postoperative day 50, gastrointestinal bleed at postoperative day 67, small bowel obstruction at postoperative day 118). At 1 year, data were available for 13 patients. Only 1 additional patient was readmitted for heart-failure related reasons from 160-365 days. This patient had done quite well after index TMVIV procedure, improving dramatically and regaining full clinical function. She had an abrupt decompensation and was found to have partial dehiscence of the sewing ring with 4+ paravalvular leak on postoperative day 201. She had clinically improved to the point where she was now a surgical candidate, and underwent successful repeat MVR. At the conclusion of year 1 post procedure, the readmission rate for heart-failure related reasons was 8.3% in the 13 patients for whom data was available.

Mortality rates and long-term postoperative outcomes. Survival data were available on all patients at the end of the study; no patients were lost to follow-up. Six-month data were available in 19 patients. At this point, the mortality rate remained at 0%, there were no cerebrovascular events or myocardial infarctions, and operative reintervention rates remained at 0%. At 1 year, follow-up data were available for 13 patients, and the mortality rate remained at 0%. The first patient death was on postoperative day 412; the patient had been readmitted for complications from an orthopedic procedure (septic patellar joint), and later expired at an outpatient rehabilitation facility. The second patient death occurred on postoperative day 462, secondary to cardiopulmonary arrest. To date, there have been no other deaths in the study.

Longitudinal assessments of stability in valvular function. TTE results for 24 patients at 30 days post procedure showed mean EF of 50.1 ± 13.2% and mean pulmonary artery pressure of 56.3 ± 18.5 mm Hg, with no change in mean valve area (MVA) compared to that measured postoperatively (reported as DMVA), and mean gradient of 8.9 ± 1.8 mm Hg. These values were not significantly different (P>.05) from those documented at subsequent follow-up in 19 patients at 180 days, where EF was 50.6 ± 16.8%, pulmonary artery pressure was 63.5 ± 20.1 mm Hg, DMVA was 0.96 mm Hg, and mean gradient was 10.4 ± 2.4 mm Hg. Data collected from 13 patients at 1 year post procedure reiterated a trend of stable EF (48.3 ± 12.2%), with improved pulmonary artery pressure 52.8 ± 12.1 mm Hg, DMVA of 0.94 mm Hg, and mean gradient of 9.0 ± 5.4 mm Hg. 

LVOT obstruction. Transcatheter replacement of the mitral valve can lead to obstruction of the LVOT, which can lead to significant morbidity or mortality.¹⁰ At our institution, preoperative and postoperative TEEs are performed at each TMVR procedure for structural and hemodynamic assessments, with most patients in our study additionally receiving preoperative cardiac CT for optimal anatomical assessment and prediction of LVOT obstruction risk. Of the 24 patients in our study, 2 (8%) were noted to have a minimal increase above a previously established standard^10,16 of >10 mm Hg peak LVOT gradient (13.8 mm Hg and 12.2 mm Hg) at 30-day follow-up. There were no quantifiable changes in LVOT gradients as noted by serial assessments by TTE over time during outpatient follow-up. No patient in our study required alcohol septal ablation as a bail-out strategy for LVOT obstruction at any time. 

Discussion

In this multiyear study of procedural quality and technical outcomes at a large tertiary referral center, we show that TMVIV/TMVIR can be done safely and effectively in patients with prior mitral valve surgery who were considered at prohibitively high surgical risk for an invasive approach to reoperative valve replacement. We demonstrate that TMVIV and TMVIR procedures can be performed in a little more than 1 hour with minimal blood loss and very low intraprocedural mortality (0.0%). Readmission rates at 30 days and at 180 days for valve thrombosis or embolization, congestive heart failure, myocardial infarction, or cerebrovascular insult were also very low (0.0%), which is consistent with good clinical efficacy in this high-risk population. Survival rates at 30 days, 180 days, and 1 year were 100%. 

Study limitations. This study contains a number of limitations. Sample size was small, with available follow-up data at 1 year post procedure for 13 patients. Additional data are needed as this procedure becomes more widespread. The retrospective non-randomized assessment schema does not allow for a comparison between either surgical or conservative medical approaches. All of the patients in this study were considered inoperable for conventional MVR secondary to significant medical comorbidities; therefore, there is a lack of applicability to other patient populations with hemodynamically significant mitral valve disease, but low, intermediate, or high (but operable) surgical risk. Lastly, while assessment of neo LVOT has become a critical step in the assessment of these patients in current practice, it was not assessed in the earlier stage of this trial; fortunately, the retrospective analysis of these patients identified no significant LVOT obstruction.  

Conclusion

In this single-center study of patients at prohibitive risk for reoperative mitral valve surgery, we demonstrate immediate and 1-year procedural safety and efficacy of TMVIV and TMVIR. 

References

1.    Arora S, Misenheimer J, Jones W, et al. Transcatheter versus surgical aortic valve replacement in intermediate risk patients: a meta-analysis. Cardiovasc Diagn Ther. 2016;6:241-249.

2.     Deutsch MA, Bleiziffer S, Elhmidi Y, et al. Beyond adding years to life: health-related quality-of-life and functional outcomes in patients with severe aortic valve stenosis at high surgical risk undergoing transcatheter aortic valve replacement. Curr Cardiol Rev. 2013;9:281-294. 

3.     Praz F, Windecker S, Huber C, Carrel T, Wenaweser P. Expanding indications of transcatheter heart valve interventions. JACC Cardiovasc Interv. 2015;8:1777-1796.

4.    Reddy G, Wang Z, Nishimura RA, et al. Transcatheter aortic valve replacement for stenotic bicuspid aortic valves: systematic review and meta analyses of observational studies. Catheter Cardiovasc Interv. 2018;91:975-983. Epub 2017 Sep 30.

5.    Cheung A, Webb JG, Barbanti M, et al. 5-year experience with transcatheter transapical mitral valve-in-valve implantation for bioprosthetic valve dysfunction. J Am Coll Cardiol. 2013;61:1759-1766.

6.     Bouleti C, Fassa AA, Himbert D, et al. Transfemoral implantation of transcatheter heart valves after deterioration of mitral bioprosthesis or previous ring annuloplasty. JACC Cardiovasc Interv. 2015;8:83-91.

7.     Seiffert M, Conradi L, Baldus S, et al. Transcatheter mitral valve-in-valve implantation in patients with degenerated bioprostheses. JACC Cardiovasc Interv. 2012;5:341-349. 

8.     Muller DW, Farivar RS, Jansz P, et al. Tendyne global feasibility trial investigators. Transcatheter mitral valve replacement for patients with symptomatic mitral regurgitation: a global feasibility trial. J Am Coll Cardiol. 2017;69:381-391. 

9.     Guerrero M. Mitral implantation of transcatheter valves (MITRAL) clinical trial. 2016. Available at: https://clinicaltrials.gov/ct2/show/NCT02370511. Accessed March 29, 2017.

10.     Wang DD, Eng MH, Greenbaum AB, et al. Validating a prediction modeling tool for left ventricular outflow tract (LVOT) obstruction after transcatheter mitral valve replacement (TMVR). Catheter Cardiovasc Interv. 2017 Dec 11 (Epub ahead of print).

11.     Green P, Woglom AE, Genereux P, et al. The impact of frailty status on survival after transcatheter aortic valve replacement in older adults with severe aortic stenosis: a single-center experience.  JACC Cardiovasc Interv. 2012;5:974-981.

12.     Forcillo J, Condado JF, Ko YA, et al. Assessment of commonly used frailty markers for high- and extreme-risk patients undergoing transcatheter aortic valve replacement. Ann Thorac Surg. 2017;104:1939-1946.

13.     Wilbring M, Alexiou K, Tugtekin SM, et al. Pushing the limits – further evolutions of transcatheter valve procedures in the mitral position, including valve-in-valve, valve-in-ring, and valve-in-native-ring. J Thorac Cardiovasc Surg. 2014;147:210-219.

14.    Urena M, Himbert D, Brochet E, et al. Transseptal transcatheter mitral valve replacement using balloon-expandable transcatheter heart valves: a step-by-step approach. JACC Cardiovasc Interv. 2017;10:1905-1919.

15.     Garrido JM, Candela-Toha AM, Parise-Roux D, et al. Impact of a new definition of acute kidney injury based on creatinine kinetics in cardiac surgery patients: a comparison with the RIFLE classification. Interact Cardiovasc Thorac Surg. 2015;20:338-344.

16.     Murphy D, Ge Y, Creighton WD, et al. Use of cardiac computerized tomography to predict neo–left ventricular outflow tract obstruction before transcatheter mitral valve replacement. J Am Heart Assoc. 2017;6(11).

From the ¹Department of Internal Medicine Residency, Eastern Virginia Medical School, Norfolk, Virginia; and ²Sentara Heart Hospital, Heart Valve and Structural Disease Center, Norfolk, Virginia.

Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. The authors report no conflicts of interest regarding the content herein.

Manuscript submitted March 14, 2018, provisional acceptance given April 23, 2018, final version accepted May 7, 2018.

Address for correspondence: Paul Mahoney, MD, Sentara Heart Hospital, Heart Valve and Structural Disease Center, 600 Gresham Drive, Norfolk, VA 23507. Email: paul.mahoney.md@gmail.com