Dedicated Closure Device for Transcaval Access Closure: from Concept to First-in-human Testing

Abstract

Objectives:

To test safety and exploratory effectiveness of a dedicated transcaval closure device.

Background:

Transcaval access enables delivery of large-caliber devices to the aorta in patients ineligible for transfemoral arterial access. Closure of aorto-caval fistulae using off-label nitinol cardiac occluders has been shown to be safe, but persistent aorto-caval fistulae at exit from the catheterization lab and bleeding complications were common in a prospective study.

Methods:

Pre-clinical testing of the Transcaval Closure Device (TCD) was performed in 24 Yorkshire swine, including 10 under Good Laboratory Practice (GLP) conditions. Subsequently, subjects undergoing transcatheter aortic valve replacement (TAVR) for symptomatic severe aortic stenosis, ineligible for transfemoral arterial access, were enrolled in a United States Food and Drug Administration (FDA) approved Early Feasibility Study (EFS) of the TCD (Transmural Systems, Andover, MA). Independently-adjudicated endpoints included technical, device and procedural success, incorporating in-hospital and 30-day clinical and imaging follow up.

Results:

Transcaval access and closure in swine confirmed that at 30 days, TCDs were almost entirely endothelialized. Subsequently, twelve subjects were enrolled in the EFS. Transcaval access, TAVR and aorto-caval fistula closure was successful in all 12 subjects. The primary endpoint of technical success was met in 100% of subjects. Complete closure of the transcaval access tract was achieved in 75% of subjects at exit from the catheterization lab and in 100% of subjects at 30 days. There were zero modified VARC-2 major vascular complications and zero VARC-2 life-threatening or major bleeding complications related to transcaval access or the TCD.

Conclusions:

The TCD achieved complete closure of the transcaval access tract in most subjects at exit from the catheterization lab and essentially eliminated transcaval-related bleeding. Dedicated devices for transcaval access and closure could enable more widespread adoption of transcaval without fear of bleeding complications. (https://clinicaltrials.gov/ct2/show/NCT03432494)

CONDENSED ABSTRACT

Transcaval closure using off-label nitinol cardiac occluders is safe, but persistent aorto-caval fistulae and bleeding remain common. Twelve subjects ineligible for transfemoral arterial access undergoing transcatheter aortic valve replacement (TAVR) were enrolled in a United States Food and Drug Administration (FDA) approved Early Feasibility Study to test the safety and exploratory effectiveness of a dedicated closure device (). Complete fistula closure was achieved in 75% of subjects at exit from the catheterization lab and in 100% of subjects at 30 days, effectively eliminating transcaval-related bleeding and shunting. Dedicated devices should enable more widespread adoption of transcaval access without fear of bleeding complications.

INTRODUCTION

Transcaval access is a percutaneous, transfemoral, transvenous approach for patients with small or diseased iliofemoral arteries who require large-bore access to the aorta. In current clinical practice, the aorto-caval tract is closed using off-label nitinol cardiac occluder devices – typically the first-generation Amplatzer Duct Occluder (Abbott/St. Jude Medical, St Paul, MN). The National Heart, Lung and Blood Institute (NHLBI) sponsored prospective transcaval trial demonstrated safety and feasibility using Amplatzer cardiac occluder devices in 100 subjects undergoing transcatheter aortic valve replacement (TAVR) who were deemed to be high or extreme risk for surgery(1). However, the Amplatzer device is incompletely hemostatic, prone to pull-through, and may result in need for blood transfusion or covered stent usage. Consequently, complete closure of the transcaval tract was only achieved in 36% of subjects at the exit from the catheterization lab and the rate of transcaval-related life-threatening or major bleeding was 12%, largely driven by retroperitoneal hematoma detected on systematic pre-discharge computed tomography (CT) scans. Furthermore, persistent left-to-right shunting was occasionally poorly tolerated in patients with severely impaired right ventricular systolic function. Dedicated closures devices are needed to simplify procedural steps and achieve rapid hemostasis. Herein, we describe the pre-clinical development and first-in-human testing of a dedicated transcaval closure device in patients undergoing TAVR who are ineligible for transfemoral artery access.

METHODS

Transcaval Closure Device (TCD)

The Transcaval Closure Device (TCD) (Transmural Systems, Andover, MA) has a braided nitinol double-disc design with an 8mm outer diameter (OD) neck and a retention paddle and interconnecting spring to flatten the intravascular disc and prevent inadvertent pull through the aortic wall during deployment (Central Illustration). Both discs are covered with polyester woven fabric (polyethylene terephthalate) to promote rapid hemostasis. The TCD is connected to a flexible cable and delivered through a beveled catheter over a central 0.014” stainless steel guidewire to maintain aortic access throughout the closure procedure. The TCD is recapturable and repositionable while connected to its flexible delivery cable. So long as the central guidewire remains in place, the cable can be reattached and the TCD can be recaptured, repositioned or withdrawn. The techniques for CT-based transcaval procedure planning(2) and transcaval access have been described elsewhere(3).

Central illustration -. Transmural Systems Transcaval Closure Device (TCD).

(A) The TCD has a braided nitinol double-disc design with an interconnecting spring to flatten the intravascular disc and a retention paddle to prevent inadvertent pull through the aortic wall during deployment. The TCD is delivered on a flexible cable through a beveled catheter and over a central guidewire. (B) Deployment of the TCD in the transcaval tract. The TCD is delivered on a flexible cable through a beveled delivery catheter over a central guidewire. (C) Digital subtraction angiography confirms complete closure of the transcaval tract. (D) 30-day CT scan shows the TCD (highlighted in green) in the wall of the abdominal aorta. (E) Transcaval outcomes comparing the TCD and the Amplatzer cardiac occluder from the NHLBI sponsored prospective transcaval trial(1).

Pre-clinical study

The institutional animal care and use committee approved all procedures, which were performed according to contemporary NIH guidelines. Yorkshire swine with median bodyweight 48kg (range 38–53kg) were anesthetized with ketamine (25mg/kg), midazolam (0.5mg/kg) and glycopyrrolate (0.01mg/kg), and maintained with isoflurane (1–3%) and mechanical ventilation. Fourteen animals underwent non-survival technical development procedures and are not described further here. Ten additional animals underwent procedures under Good Laboratory Practice (GLP) conditions to support an application to the US Food and Drug Administration (FDA) for the Investigational Device Exemption (IDE) Early Feasibility Study (EFS), and are described in the results. All received aspirin loading dose (162mg) prior to the procedure and maintenance dose (81mg) for 30 days post-procedure. Vascular access was obtained with ultrasound guidance.

For the GLP experiments, nine animals underwent transcaval access and closure using the TCD. Three were survived for 30 days and 6 for 90 days before repeat angiography, euthanasia and necropsy, with high resolution X-ray imaging and histopathology. One animal underwent transcaval access and closure using a commercially available nitinol cardiac occluder (Amplatzer Duct Occluder) and was survived for 30 days for histology comparison.

Early Feasibility Study (EFS)

Subjects undergoing TAVR for symptomatic severe aortic stenosis using commercially available transcatheter heart valves (THV) were enrolled into an investigator-initiated FDA-approved EFS sponsored by NHLBI at three centers in the United States (‘Transmural Systems Transcaval Closure Device for Transcaval Access Ports During Transcatheter Aortic Valve Replacement’ ())(4). Table S1 (Supplemental Materials) summarizes trial inclusion and exclusion criteria. In brief, subjects were enrolled if they were ineligible for transfemoral artery access and eligible for transcaval access according to the institutional heart team and the study clinical review committee. The primary endpoint was technical success, measured at exit from the catheterization laboratory, requiring all of the following to be present: (1) alive; (2) successful delivery of the TCD, and retrieval of the TCD delivery system; (3) deployment and correct positioning of a single intended TCD (recapture and repositioning of the device, if needed, was not classified as failure); and (4) no additional unplanned or emergency surgery or reintervention related to the TCD or delivery system (adjunctive balloon aortic tamponade was permissible and consistent with technical success). A key performance endpoint was device success (Table S2, Supplemental Materials). The primary safety endpoint was procedural success, measured at 30 days, requiring (1) device success; and (2) no device-related Serious Adverse Events, defined as VARC-2 life-threatening bleeding, major vascular or cardiac complications related to the TCD requiring unplanned reintervention or surgery (such as covered stent implantation at the transcaval access site). Additional secondary endpoints included individual components of the technical and procedural success endpoints. Modified VARC 2 definitions were used for all clinical endpoints(5).

Animal experiments were performed by the first author with manufacturer representatives in attendance, and the non-clinical portions of the IDE license application, including histology images, were prepared by the manufacturer.

The clinical protocol was prepared and executed by the investigators, who collected and analyzed the data independently from the manufacturer. The study investigators have custody of all data and are responsible for the findings. Categorical variables are presented as n (percentage) and continuous variables are presented as median (interquartile range). To assure data integrity, all case report forms were independently verified with source data on-site, clinical events were independently monitored, all computed tomography (CT) images were analyzed by a central laboratory, and the NHLBI Data and Safety Monitoring Board provided oversight along with participating site Institutional Review Boards. An independent Clinical Event Adjudication Committee classified the primary and clinical endpoints including death and stroke and determined relatedness to the transcaval procedure and to the TCD. The manuscript was prepared by the investigators without participation by the manufacturer.

RESULTS

Preclinical study

Figure 1 summarizes the key findings of the GLP preclinical study. Transcaval access was successful in all 10 GLP animals. Deployment of a single TCD was successful with complete closure of the transcaval tract after median 9min (8–10min) in 9/9 animals. On gross pathology, all TCDs appeared well-deployed with no evidence of local tissue injury. There were no nitinol wire fractures assessed by high-resolution micro-CT (Figure 1). Already at 30 days, the TCDs were mostly incorporated by neointima consisting of organized layers of smooth muscle cells covered by endothelium and the guidewire lumen was filled with fibrous tissue (Figure 1). This was similar to the fibrosis pattern surrounding the comparator Amplatzer Duct Occluder (Data Not Shown). There was persistent lymphocytic inflammation around the device seen up to 90-days, also present in the comparator animal that received an Amplatzer Duct Occluder (Data Not Shown). Histological analysis of non-target organs was unremarkable.

Figure 1 -. Preclinical findings.

Angiography (A) immediately post-deployment and (B) after 90 days. Gross pathology of (C) IVC surface and (D) aorta surface after 90 days shows endothelialization of the TCD. Micro-CT images (E-M) confirms structural integrity of the TCD and lack of nitinol wire braid fracture. Hematoxylin and Eosin (H&E) stains: (N) low power view shows the central hub of the TCD with the aorta at the top of the image. (O) and (P) correspond to the green and red boxes in (N) respectively. The central guidewire lumen is completely filled with fibrous tissue in (P). Low power (Q) and (R) and corresponding higher power (q) and (r) views of the aortic surface show the TCD fully incorporated by a smooth muscle cell and proteoglycan rich neointima with polyester material surrounded by fibrous tissue, scattered lymphocytic infiltrates, giant cells, and golden-brown hemosiderin pigment. (S) corresponds to the gray box in (N), with giant cells incorporated by fibrous tissue and scattered lymphocytes at high power (s). Low power (T) and corresponding high power (t) of the central hub on the IVC surface shows the TCD fully incorporated by a smooth muscle cell and proteoglycan rich neointima. (T) corresponds to the blue box in (S). (U) and (V) and corresponding high power (u) and (v) views of the IVC surface show polyester material surrounded by fibrous tissue, with giant cells and a few scattered lymphocytes. Scale bar in (N) and (S) = 1mm; low power = 200μm; high power = 100μm.

First-in-human Early Feasibility Study

Baseline characteristics and operative risk assessment of the 12 enrolled subjects are summarized in Table 1. Table 2 summarizes transcaval TAVR procedure details. Successful transcaval closure with a single TCD was achieved in 100% (12/12) subjects. Transient hypotension was observed after administration of protamine and after TAVR sheath withdrawal back into the inferior vena cava (Figure S1, Supplemental materials), at which point there is expected shunting of blood from aorta to cava prior to deployment of the TCD. There was immediate return to baseline blood pressure after deployment of the TCD. The primary endpoint of technical success was met in 100% (12/12) subjects enrolled and treated with the TCD. Contrast extravasation was observed after deployment of the TCD in 1 subject with extreme aortic calcification and was successfully treated with adjunctive balloon aortic tamponade. No subject required a covered stent. There were zero modified VARC-2 major vascular complications and zero VARC-2 life-threatening or major bleeding complications related to transcaval access or closure with the TCD. Overall, complete closure of the transcaval tract (Type 0 closure pattern) was achieved at exit from the catheterization lab in 75% (9/12) subject (Figure S2, Supplemental materials)(1). One subject had Type 1 (funnel-shaped) and two subjects had Type 2 (cruciform or swirling) angiographic closure patterns at exit from the cath lab. Complete closure was confirmed in 100% of subjects at 30 days based on contrast-enhanced CT scans (Figure 2). 30-day CT scans confirmed stability of TCDs with no evidence of migration or erosion.

Table 1 -.

Baseline characteristics

Variable	n=12
Age (years)	82 (75, 88)
Male sex	50 % (6/12)
Body surface area (m2)	1.7 (1.6, 1.8)
Diabetes	25% (3/12)
Hypertension	83% (10/12)
Severe pulmonary disease	25% (3/12)
Coronary artery disease	67% (8/12)
Atrial fibrillation	17% (2/12)
Prior permanent pacemaker/defibrillator	8% (1/12)
Prior stroke/transient ischemic attack	8% (1/12)
Prior percutaneous coronary intervention	25% (3/12)
Prior coronary artery bypass surgery	25% (3/12)
Frail	25% (3/12)
Hemoglobin (g/dL)	12.6 (11.7, 13.1)
STS-PROM score (%)	3.7 (3.1, 4.5)
Operative risk	Extreme (2/12), High (6/12), Intermediate (4/12)
Ineligible for transfemoral access	100% (12/12)

Continuous variables are presented as median (interquartile range). Categorical variables are presented as percentages. STS-PROM: Society of Thoracic Surgeons Predicted Risk of Mortality.

Table 2 -.

Procedure details

	n=12
Moderate sedation	75% (9/12)
Transcaval sheath nominal size (French)	15.1±2.3
Transcaval sheath maximum outer diameter	7.76mm (determined by valve type, valve size and sheath)
TAVR device	83% (10/12) Edwards Sapien 3
	17% (2/12) Medtronic Evolut R
TAVR success using the first intended valve	100% (12/12)
Adjunctive aortic balloon tamponade	8% (1/12)
Covered stent	0% (0/12)
Successful transcaval closure with a single TCD	100% (12/12)
Final angiographic transcaval closure score *	75% (9/12) Type 0
	8% (1/12) Type 1
	17% (2/12) Type 2
	0% (0/12) Type 3

TAVR: transcatheter aortic valve replacement; TCD: transcaval closure device.

^*Angiographic transcaval closure score: see Figure S2, Supplemental materials.

Figure 2 -. Confirmed 100% aorto-caval tract closure at 30 days.

Final digital subtraction angiogram at the end of the TAVR procedure (subject 1) and contrast-enhanced arterial-phase CT scans at 30 days (subjects 2–12). CT: computed tomography; TAVR: transcatheter aortic valve replacement.

Table 3 summarizes additional 30-day clinical outcomes. One subject died on post-procedure day 15 from respiratory failure attributed to exacerbation of previously unrecognized fibrotic pulmonary disease, adjudicated as non-cardiovascular mortality and unrelated to transcaval access or TAVR. Transcaval access, TAVR and deployment of the TCD were successful in this subject, with complete closure of the transcaval tract on the final angiogram at exit from the catheterization lab. One subject with prior coronary artery bypass surgery developed chest pain and new left bundle branch block after TAVR, and troponin levels were elevated. Cardiac catheterization demonstrated stenosis of the right brachiocephalic artery (supplying a right internal mammary graft) for which he underwent successful stenting with resolution of his chest pain. Two subjects suffered strokes on post-procedure day 1 and day 3 respectively. Brain MRI in both subjects confirmed multiple acute infarcts consistent with cardioembolic source. Both were adjudicated as unrelated to the TCD. One subject had a major bleed following a mechanical fall complicated by a fractured neck of femur on post-procedure day 12. This subject underwent hemiarthroplasty during which hemoglobin fell by 3g/dL and 2 units red blood cells were transfused. This bleed was adjudicated as unrelated to the TCD. One other subject was transfused 1 unit of packed red blood cells, administered during the TAVR procedure without overt bleeding, that was adjudicated as not clinically indicated. One subject developed a cold left lower limb on post-procedure day 23. CT angiography demonstrated occlusion of the left common and external iliac arteries, with thrombosis on top of pre-existing laminar thrombus in a left internal iliac artery aneurysm. The subject subsequently underwent successful percutaneous endovascular intervention with thrombectomy and stenting. The CT angiogram did not reveal any thrombus on the aortic aspect of the TCD and the subject was in atrial fibrillation but was not anticoagulated. This was adjudicated as possibly related to the TCD because it was not possible to determine with certainty the origin of the presumed thromboembolism. There was one addition minor vascular complication: one subject had a groin hematoma at the contralateral 6Fr femoral artery access site. Overall, the performance endpoint of device success was met in 10/12 subjects at 30 days and the safety endpoint of procedural success was also met in 10/12 subjects at 30 days, including follow up CT analysis. The 2 failures for both device and procedural success endpoints were due to the 1 mortality, and the 1 lower limb thromboembolism.

Table 3 -.

Outcomes through 30 days

Outcome	n=12	Relatedness to TCD*
Post-TAVR hospital length of stay	2 (2,8)	—
Death
Cardiovascular	0	—
Non-cardiovascular	1	Unrelated
Stroke	2	Unrelated
Myocardial infarction	1	Unrelated
Acute kidney injury (stage II or III)	0	—
Transfusion	1 (1 unit)	Unrelated
VARC-2 bleeding
Life-threatening	0	—
Major	1	Unrelated
VARC-2 vascular complications
Major	0	—
Minor	2	1 unrelated / 1 possibly related

VARC-2: Second valve academic research consortium consensus.

^*relatedness of event to the TCD device adjudicated by independent clinical events adjudication committee

DISCUSSION

Herein we describe the pre-clinical experiments that supported submission to FDA for the IDE and EFS. The key findings from the EFS were (1) the TCD achieved complete closure of the transcaval tract in 75% of subjects at exit from the catheterization lab and in 100% of subjects at 30 days; (2) Transcaval closure was safe with no modified VARC-2 major vascular or bleeding complications; and (3); no subject required transfusion for transcaval-related bleeding. Transcaval access has several advantages over other extrathoracic non-transfemoral artery approaches such as transaxillary, subclavian and transcarotid. Firstly, the procedure is performed via right femoral access and therefore the catheterization laboratory ergonomics of the TAVR procedure are identical to transfemoral artery access. Secondly, operator radiation exposure is similar to transfemoral artery approach, and less than transaxillary or transcarotid access that require the operator to stand close to the X-ray source. Thirdly, unlike transcarotid and subclavian, transcaval does not require general anesthesia and is therefore compatible with the prevailing minimalist TAVR approach. Finally, because transcaval is a transvenous procedure, patients are able to ambulate soon after the procedure, similar to other large-bore transvenous procedures such as edge-to-edge mitral valve repair or atrial septal defect closure. Among interventional cardiologists, the principal apprehension with transcaval is the risk of bleeding. In the NHLBI-sponsored prospective transcaval trial, complete closure of the transcaval tract was achieved at exit from the catheterization lab in only 36% of subjects. At 12 months, there were no late major vascular complications and CT demonstrated spontaneous closure of almost all fistulas(6). Clinical transcaval experience has shown that persistent aorto-caval fistulae are tolerable in the vast majority of patients and will close within hours or days. Persistent left-to-right shunting is poorly tolerated in a small subset of patients with severely impaired right ventricular systolic function. The rates of peri-procedural life-threatening or major bleeding and major vascular complications in the NHLBI study were 18% and 19% respectively. VARC-2 life-threatening bleeding and modified VARC-2 major vascular complications possibly related to transcaval were 7% and 13%, respectively. Thirty-five percent of subjects received median 2 units (2–4 units) blood transfusions(1).

In contrast, complete closure of the transcaval tract was achieved in 75% of subjects at exit from the catheterization lab in this EFS. Only 1 subject received a blood transfusion, which was not clinically indicated. No procedure-related life-threatening or major bleeding occurred. These differences are likely multifactorial. Firstly, the polyester fabric wrap on the TCD promotes rapid occlusion. Secondly, the internal spring and paddle resist pull through and enforce optimal expansion of the braided nitinol within the hole in the aortic wall. Thirdly, subjects enrolled in the EFS were lower risk than those enrolled in the NHLBI prospective transcaval trial (STS-PROM score 4.7% vs. 9.6%, respectively) with fewer comorbidities and potentially lower bleeding risk. Fourthly, protamine was used systematically in the EFS to fully reverse heparinization prior to closure. Finally, operator experience with transcaval has increased since the NHLBI prospective trial, potentially resulting in improved technical outcomes. Other perceived disadvantages of transcaval access include procedural complexity, reluctance to use devices off-label and cost of necessary equipment (specifically, deflectable guiding sheath and nitinol occluder device). Simplifying closure with dedicated devices such as the TCD would address many of these concerns. Additional equipment such as dedicated guidewires for transcaval access would further simplify the procedure. In our experience, as with most structural heart interventions there is a learning curve with transcaval, but once mastered comparable procedure times to transfemoral artery access can be achieved. The TCD has a number of unique features that improve upon off-label Amplatzer nitinol cardiac occluder devices. Firstly, the central spring optimizes apposition against the aortic wall, thus preventing unintentional elongation and pull through. The paddle also helps prevent pull through by apposing with the endoluminal wall of the aorta. Secondly, the TCD is wrapped in polyester fabric to promote rapid occlusion – unlike the Amplatzer Duct Occluder which is designed to be porous and close by thrombosis over hours to days – and early endothelialization. Pre-clinical histological analysis demonstrated that TCDs were almost entirely endothelialized at 30 days. Thirdly, it has a central guidewire lumen that is maintained throughout deployment. Finally, the TCD is recapturable and repositionable even after release from the cable: so long as the central guidewire is in place, the cable can be re-advanced over the guidewire and reattached to the TCD. These novel features are not solely applicable to transcaval closure, and indeed may simplify closure of atrial septal defects, paravalvular leaks and other intracardiac or endovascular holes.

Limitations

This was a first-in-human study and based on these promising results, a further trial with larger sample size is warranted. There is no approved device for transcaval closure to serve as a comparator: the Amplatzer Duct occluder is used off-label in clinical practice. Subjects were selected for having an appropriate calcium-free window in the abdominal aorta for transcaval access.

CONCLUSIONS

The TCD achieved complete closure of the transcaval access tract in the majority of subjects at exit from the catheterization lab in this first-in-human Early Feasibility Study. Transcaval-related bleeding was eliminated. Dedicated devices for transcaval access and closure should enable more widespread adoption of this approach without fear of bleeding complications.

PERSPECTIVES.

WHAT IS KNOWN?

Transcaval access and closure using off-label nitinol cardiac occluder devices has been shown to be safe and effective to deliver large-caliber devices to the aorta in patients with small or diseased iliofemoral arteries, but many patients had persistent aorto-caval fistulae at exit from the catheterization lab and transcaval-related bleeding remains a concern.

WHAT IS NEW?

First-in-human experience with a dedicated closure device achieved a high rate of complete aorto-caval fistula closure at exit from the catheterization lab and eliminated transcaval-related bleeding.

WHAT IS NEXT?

After completion of this Early Feasibility Study, a larger study is warranted to test safety and effectiveness of the transcaval closure device.

ABBREVIATIONS AND ACRONYMS

CT

Computed tomography

EFS

Early Feasibility Study

GLP

Good Laboratory Practice

IDE

Investigational Device Exemption

NHLBI

National Heart, Lung, and Blood Institute

NIH

National Institutes of Health

STS-PROM

Society of Thoracic Surgeons Predicted Risk of Mortality

TAVR

Transcatheter aortic valve replacement

TCD

Transcaval closure device

THV

Transcatheter heart valve

VARC-2

Second valve academic research consortium criteria