Transcatheter Valve-in-Band Implantation in a Swine Model of Failed Mitral Repair: A Technical Note and Proof of Feasibility

1. Introduction

Mitral regurgitation (MR) is among the most common valvular lesions, with the prevalence of clinically significant valve disease rising with age, reaching 13.3% in those aged ≥75 years [1]. Surgical mitral valve repair is the gold standard for severe primary MR, offering superior long-term outcomes to valve replacement [2], including lower operative mortality (3.9% versus 8.9%) and superior 1-, 5-, and 10-year survival [3,4]. At high volume centers, roughly three quarters of degenerative valves are repaired rather than replaced [5]. Repair is completed with an annuloplasty device, either a complete ring or a partial band, and device choice varies by surgeon and center. Partial posterior bands, favored to preserve anterior annular and leaflet mobility and the physiologic saddle-shaped dynamics of the annulus, predominate at some centers (roughly two-thirds of repairs in one series [6]), whereas complete rings remain the majority for degenerative mitral repair in the United States [7]. Repairs are durable but not always permanent. They can fail through a variety of mechanisms, including incomplete correction of MR at the baseline operation, continued disease progression, or leaflet calcification [8]. Significant regurgitation recurs progressively in the years after surgery [9], affecting approximately 16.5% of patients with reported reoperation rates ranging from 4.5-10% [10]. Because repair volumes are large and many affected patients are elderly or otherwise poor candidates for redo sternotomy, a substantial and growing population lives with a failing repair, yet minimally invasive options to avoid redo sternotomy remain limited.

For patients who are poor surgical candidates, transcatheter mitral valve replacement (TMVR) has emerged as a reintervention strategy, with established valve-in-valve (ViV) and valve-in-ring (ViR) options for failed bioprostheses and complete rings. Yet, for failed bands, valve-in-band (ViB) intervention remains largely unexplored, limited to an isolated case report [11], with no US Food and Drug Administration (FDA)-approved option. The principal concern for valve-in-band is incomplete annular support; partial bands reinforce only the posterior and lateral annuli, leaving the anterior annulus unsupported and theoretically predisposing to insufficient fixation, embolization, left ventricular outflow tract (LVOT) obstruction, and paravalvular leak (PVL) [12]. Although the expanding valve-in-ring experience offers transferable lessons in device selection, sizing, and pre-procedural LVOT-obstruction screening, no comparable framework yet exists for valve-in-band. The aim of our work is to begin building such a framework. Our model is part of a broader research program (IACUC Protocol #00002378) evaluating percutaneous ring completion and transcatheter valve deployment following atrioventricular-valve band placement in both the mitral and tricuspid positions. Building on a swine model that recreates failed repair by generating MR within an annuloplasty band, we deployed a TMVR device within the band in a large animal.

2. Materials and Methods

Study design and animal. This was a single-animal, non-survival, first-in-animal feasibility study conducted under a protocol approved by the Dartmouth College Institutional Animal Care and Use Committee (IACUC Protocol #00002378; approved 4 November 2025; Animal Welfare Assurance D16-00166). The study protocol, including its objectives and design, was defined a priori and approved by the IACUC; it was not entered in a public registry. Reporting follows the ARRIVE 2.0 guidelines. One 20-month-old, 68 kg female Yucatan mini pig (Sinclair Bio Resources, Auxvasse, MO, USA) was studied and was housed and acclimated in the Dartmouth College Center for Comparative Medicine and Research. A single animal was used because the objective was proof of technical feasibility rather than a statistically powered comparison. The experimental unit was the individual animal; no control group was included, and no a priori sample-size calculation was performed, consistent with a single-arm feasibility design. Randomization, blinding, and formal inclusion/exclusion criteria were not applicable to a single-animal study, and there were no exclusions. The animal was healthy at enrollment, with no prior experimental procedures.

Anesthesia and monitoring. Following a 12 h fast, the animal was premedicated with intramuscular tiletamine-zolazepam (Telazol, 4.4 mL) and xylazine (0.44 mL), intubated (7.5 mm endotracheal tube), and maintained on isoflurane (2-3%) in oxygen (2 L/min). Neuromuscular blockade was provided with rocuronium, and heparin, calcium chloride, magnesium sulfate, lidocaine, and sodium bicarbonate were administered intravenously as indicated. Monitoring included pulse oximetry, end-tidal CO2, invasive arterial pressure, temperature, and serial arterial blood gases. Anesthesia and analgesia were titrated to minimize pain and distress; as a terminal procedure performed entirely under general anesthesia, the study had no humane endpoints or recovery-monitoring schedule, and no unexpected adverse events occurred.

Surgical preparation and cardiopulmonary bypass. The animal was aseptically prepped and draped in the hybrid operating suite (Center for Surgical Innovation). The chest was opened via clamshell incision with division of the internal mammary arteries; to maximize exposure, the sternum and ribs were divided to approximately the midaxillary line, a retractor was placed, and the pericardium was opened. A hypoplastic aorta (~1.5 cm) was noted without other structural abnormalities. After heparinization to an activated clotting time > 400 s, central cannulation was performed (16 Fr arterial, 36 Fr venous) and cardiopulmonary bypass was initiated using a Medtronic oxygenator with a Plasma-Lyte prime (~1400 mL). The aorta was cross-clamped and the heart was arrested with antegrade cold del Nido cardioplegia (4:1; ~1.9 L), with mild systemic hypothermia (~34-35 °C) during the arrest.

Model creation. Through a left atriotomy (accessed by following the right pulmonary vein), the mitral valve was exposed and circumferential 2-0 Ethibond sutures were placed around the posterior annulus from trigone to trigone. A 28 mm mitral annuloplasty band, oversized by 2 mm to generate mitral regurgitation, was seated and secured with core knots (Figure 1). To enable subsequent transcatheter access, a Gelweave mesh conduit with an 8 mm side-arm was anastomosed to the left atrium to simulate a dilated atrium. The cross-clamp was removed after 75 min of ischemia and the heart resumed sinus rhythm.

Device selection and transcatheter deployment. A 23 mm Edwards Sapien Ultra was selected on the basis of ex vivo measurement of the implanted band. With the animal on partial cardiopulmonary bypass for hemodynamic support, guidewire access to the left ventricle was established through the left atrial graft under fluoroscopic guidance (Figure 2), and the valve was deployed within the band using the Commander delivery system (Figure 3). Fluoroscopy and echocardiography guided and documented positioning.

Assessments and euthanasia. The primary outcome was technical success, defined as deployment of the transcatheter valve within the band with stable fixation and without device embolization. Secondary outcomes were resolution of mitral regurgitation, paravalvular leak, LVOT obstruction, valve thrombus, and periprocedural survival. Valve position, apposition, mitral regurgitation, paravalvular leak, thrombus, and left ventricular outflow tract patency were assessed by intraoperative echocardiography with color Doppler and corroborated by direct inspection and postmortem dissection. Given the single-animal design, outcomes are reported descriptively, without formal statistical analysis. At completion of this non-survival protocol, euthanasia was performed by the attending veterinarian with intravenous pentobarbital-phenytoin (Euthasol), and death was confirmed by the absence of cardiac function.

3. Results

The animal tolerated the procedure without perioperative complications. After band placement, severe MR was confirmed by intraoperative echocardiography, and fluoroscopy confirmed correct TMVR positioning within the band. Following deployment, the valve frame was stable with appropriate contact, without redundant leaflet motion, frame prolapse, or embolization. The native anterior leaflet and annulus contributed circumferential radial support, complementing the posterior and lateral fixation provided by the band. Echocardiography demonstrated a well-functioning valve with resolution of MR (Figure 4) without thrombus or paravalvular leak on color Doppler. No LVOT obstruction was identified on echocardiography. The pig remained hemodynamically stable, and no obstruction was evident on postmortem dissection. Direct visualization confirmed apposition of the valve frame against the band without focal gaps or LVOT obstruction (Figure 5).

4. Discussion

Failed partial-band mitral repair leaves a large and growing population without a minimally invasive option. A partial posterior band is often chosen because it stabilizes the dilated posterior annulus while preserving anterior leaflet and annular mobility [13], with durability comparable to a complete ring [14]. The trade-off is that structural support is confined to the posterior and lateral annulus. Valve-in-valve and valve-in-ring succeed because the bioprosthetic valve or complete ring gives the transcatheter valve a circumferential landing zone [15,16]. A partial band does not; thus, valve-in-band has remained mainly theoretical until this point [12]. This study asked whether a band can nonetheless anchor a transcatheter valve. The valve seated stably within the band without embolization or paravalvular leak, and the preserved native anterior leaflet and annulus supplied the circumferential radial support the band lacks (Figure 4 and Figure 5). The finding is consistent with the preserved native anterior annulus and leaflet supplying the anterior radial support a partial band omits, completing a circumferential anchor for the valve in the acute setting.

The valve-in-valve and valve-in-ring experience, though built on a different anchoring substrate, offers a template for preoperative planning and screening for valve-in-band candidates. Recent literature has found that valve-in-ring and valve-in-mitral annular calcification are feasible, but not without risk, citing 22.4% 1-year mortality, a 9.1% reintervention rate, and left ventricular outflow tract obstruction as a leading cause of death [15,16]. Preprocedural strategies for ViR can be adapted to the ViB setting, such as preprocedural computed tomography sizing of the annulus and modeling the potential LVOT obstruction. This strategy can also assess individual anterior leaflet anatomy and flag patients at risk of obstruction [17,18] and inform whether adjunctive leaflet modification is warranted [19]. Valve-in-band can adopt this planning and screening playbook.

Building on the success of transcatheter deployment of the valve in the swine model, two strategies now warrant further evaluation (Figure 6). The first is direct valve-in-band, relying on the native anterior leaflet for anterior support. The second is staged percutaneous ring completion, bridging the unsupported anterior annulus between the fibrous trigones to convert the band into a complete ring before deployment; this strategy restores circumferential support and reframes the procedure as the better characterized valve-in-ring [16,17]. Comparing the two is the explicit aim of the program under which this study was performed (IACUC Protocol #00002378, approved for up to 30 animals). Definitive translation will require a fully percutaneous, transseptal approach, survival data, and adequately powered comparison against redo surgery [10] using standardized Mitral Valve Academic Research Consortium endpoints [20].

These conclusions must be tempered by the limitations inherent to a single, acute, non-survival experiment. With one animal, we cannot assess reproducibility, chronic valve function, durability, or the delayed risk of LVOT obstruction, and outcomes are necessarily descriptive rather than statistical. The preparation also carries model-specific features, including a surgically simulated dilated left atrium for access and an incidentally hypoplastic aorta, that may not reflect human anatomy or a fully percutaneous procedure. These constraints define the immediate next steps: survival studies in additional animals to characterize durability and the safety profile before any clinical translation.

5. Conclusions

In a large-animal model of failed band annuloplasty, transcatheter valve-in-band was technically feasible and acutely safe, achieving stable fixation with resolution of mitral regurgitation and without embolization or paravalvular leak. Critically, the preserved native anterior leaflet supplied the circumferential support that a partial band cannot, providing the first direct evidence that a failed band, long regarded as an unsuitable landing zone, can anchor a transcatheter valve. These findings establish a platform for survival and comparative studies of direct valve-in-band versus staged percutaneous ring completion, using standardized endpoints, to determine whether the approach can mature into a durable, minimally invasive alternative to redo mitral surgery.