Long‐Term Follow‐Up After Direct‐Flow Transcatheter Aortic Valve Implantation: A Single Center Experience

ABSTRACT

Aims

Valve degeneration after DirectFlow implantation poses a challenge for valve‐in‐valve procedures due to its fragile polymeric structure and embolization risk. This study analyzes long‐term follow‐up of patients who underwent TAVI with DF to understand valve degeneration tendencies.

Methods

In this single‐center registry, we included all patients who underwent DF implantation in our center. Clinical characteristics and echocardiographic parameters were assessed at baseline and during the latest follow‐up. Long‐term overall survival was analyzed. Outcome data was compared with a matched cohort of patients who underwent TAVI with other commercial devices during the same period.

Results

From 2014 to 2017, 67 patients with significant aortic stenosis were treated with DF; 36 (54%) were male, mean age 83 ± 7 years. Left ventricular ejection fraction was 50 ± 13% with a mean gradient of 46 ± 15 mmHg. Post‐procedural echocardiography showed a mean gradient of 16 ± 8 mmHg, markedly higher than the reference group (8 ± 3 mmHg; p < 0.01). This residual gradient showed no progression during follow‐up at 20 [9; 39] months (16 ± 9 mmHg; p = 0.52). Overall survival was comparable between DF and the reference group at 12 months (6% vs. 4%, HR 1.37, 95% CI 0.31–6.02) and maximal follow‐up of 39 ± 25 months (31% vs. 25%, HR 1.44, 95% CI 0.76–2.73).

Conclusions

According to this single center experience, DF devices resulted in higher residual gradients; however, no signs of faster degeneration or worse long‐term outcomes were observed compared to other devices.

1. Background

Aortic stenosis (AS) is the most common valvular diseases worldwide, with up to 10% of all patients over 80 years being affected by this condition [1, 2]. Given the increasingly aging population, calcific degeneration of the native aortic valve represents the most common etiology of AS in Europe and North America [3]. While traditionally surgical aortic valve replacement (SAVR) was considered the standard treatment for severe AS, transcatheter aortic valve implantation (TAVI) has become the standard of care for the elderly population over 75 years old [4, 5, 6, 7, 8, 9, 10, 11]. Bioprosthetic valves have been more frequently used in recent years, both after SAVR and after TAVI [12]. Biological valves have a lower thromboembolic risk; however, because their durability is limited, they tend to degenerate over a period of 8–10 years [13]. This leads to valve dysfunction, in which case, due to typical age and perioperative risks, percutaneous treatment is often preferred. Although concrete data about re‐do TAVI is lacking, the frequency of valve‐in‐valve (ViV) procedures is steadily increasing [14].

The Direct Flow Medical (DF) transcatheter aortic valve (DF Medical Inc., Santa Rosa, CA, USA) was a non‐metallic percutaneous valve system made of bovine pericardium, with a Dacron‐polyester structure, available worldwide between 2013 and 2017. It was manufactured in four sizes (23/25/27/29 mm) with two independently inflatable rings, in aortic and ventricular positions (Figure 1). The rings were initially filled with saline, which was exchanged for a polymer after the final position was confirmed. Despite initially favorable short‐term outcomes and a decreased incidence of post‐interventional permanent pacemaker implantation, the DF valve failed to achieve a sustainable paradigm shift and was eventually withdrawn from the market [15]. This device had a unique polymeric structure, designed to offer safety improvements through its ability to be inflated, deflated, retrieved, and repositioned, features that differentiated it from all other available prostheses. While the DF valve was designed to offer procedural advantages and favorable acute outcomes, its unique features may pose challenges when the valve degenerates and a ViV procedure becomes necessary.

Figure 1.

Implantation of a 25 mm DirectFlow Valve. Upper/aortic ring (red); lower/ventricular ring (yellow); interconnection tubular bridging system (blue arrow). [Color figure can be viewed at wileyonlinelibrary.com]

It is estimated that around 3000 DF valves were implanted worldwide before this type of prosthesis was withdrawn from the market. Considering the increasing survival rates of TAVI patients, a large number treated with the DF valve are still alive and encountered nowadays during routine follow‐up [16].

Data regarding long‐term durability and valve degeneration after DF implantation is lacking. The aim of the present work is to analyze the long‐term follow‐up of patients who have undergone TAVI with the DF valve, to better understand the valve degeneration tendencies associated with this device.

2. Case Report

We report a case of severe aortic regurgitation due to DF valve degeneration, which, in our experience, was the only case of significant valve failure requiring invasive correction.

A 75‐year‐old patient presented to our center with symptoms of heart failure, particularly dyspnea, classified as NYHA (New York Heart Association) class II. His medical history included COPD (chronic obstructive pulmonary disease) and moderately impaired renal function. Co‐morbidities included hypertension and dyslipidemia.

At the time of admission, severe AS (mean gradient 48 mmHg) combined with mild regurgitation in a degenerated, calcified aortic valve was diagnosed. After Heart‐Team evaluation and discussion, he was scheduled for transfemoral TAVI with a DF valve system. This system is composed of two rings, an upper (aortic) and a lower (ventricular) ring, connected by a tubular bridging system (Figure 1). Independent pressurization of the two rings could be performed through position‐fill lumens. The DF design offered the possibility of optimal positioning before final deployment due to its capacity for retrieval and repositioning. This feature was considered important to prevent potential complications, such as valve dislocation, obstruction of the coronary ostia, or paravalvular leakage [17]. After predilation of the aortic ring with a 22 mm balloon, a 25 mm DF device was delivered (Figure 1). The patient was discharged after a few days with normalized postprocedural gradients (mean gradient 15 mmHg, maximum gradient 19 mmHg) (Figure 2).

Figure 2.

Postprocedural gradients at discharge (mean gradient 15 mmHg, max gradient 19 mmHg). [Color figure can be viewed at wileyonlinelibrary.com]

After 7 years, he presented to the emergency department with progressive dyspnea and was found to have severe aortic regurgitation due to valve degeneration (Figures 3 and 4), without a significant stenotic component (mean gradient 19 mmHg). Transthoracic echocardiography as well as transesophageal echocardiography (TEE) revealed morphologic deterioration of the DF valve with leaflet wear, in the absence of significant calcification, and a central holodiastolic regurgitation jet. The Heart Team decided to perform a repeat TAVI using a ViV approach, selecting a balloon‐expandable 23 mm Edwards Sapien XT valve based on preprocedural CT measurements and with consideration for future coronary access (Figure 5). The valve appeared underexpanded on fluoroscopy, so postdilation with a 21 mm True balloon was performed (Figure 6). The invasively measured gradient was 7 mmHg, while postprocedural echocardiography confirmed a good result, with no aortic insufficiency, mild residual AS (mean gradient 12 mmHg), and preserved left ventricular function.

Figure 3.

Transthoracic echocardiogram showing a severe aortic regurgitation due to DirectFlow valve degeneration. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 4.

Transesophageal echocardiogramm showing a severe aortic regurgitation due to DirectFlow valve degeneration. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 5.

Valve‐in‐valve implantation of a 23 mm Edwards Sapien XT Valve.

Figure 6.

Postdilatation of the Edwards Sapien XT Valve with a 21 mm True balloon.

At 1‐year follow‐up, the patient remained in NYHA class II with a mean transvalvular gradient of 9 mmHg.

3. Methods

3.1. Patient Selection

This was a single‐center retrospective registry conducted at the University Heart Center Graz, Austria, over a consecutive 3‐year period (i.e., between 2014 and 2017). The registry enrolled all consecutive patients who underwent elective TAVI procedures with the DF Medical transcatheter aortic valve system for severe AS (DF group). In the Reference group, we included a cohort of patients matched by age, echocardiographic parameters (LVEF and AS severity), and overall risk status (calculated by EuroScore II and STS), who underwent TAVI with other devices during the same period.

The DF valves were among the workhorse TAVI devices used at our center during that period. Consequently, patients were not specifically selected for this device over others. The valves implanted in the Reference group included other transcatheter aortic valve systems routinely used at our center at that time, namely the CoreValve Evolut R (Medtronic, Minneapolis, MN, USA), the Edwards Sapien XT valve (Edwards Lifesciences, Irvine, CA, USA), the ACURATE TA (Symetis SA, Ecublens, Switzerland), and the Portico TAVI system (St. Jude Medical, St. Paul, MN, USA).

Before the intervention, severe AS was confirmed by echocardiography according to the guidelines of the European Society of Cardiology [18], performed by trained physicians. The preferred valve was chosen based on the dimensions and morphology of the aorta and aortic valve, assessed by multi‐slice computed tomography. The 3mensio planning software (PIE Medical Imaging, Maastricht, Netherlands) was used for calcification measurements. Each case was independently discussed in the Heart Team, while the EuroScore II and STS score were used for estimation of surgical risk and to guide further management.

Patient data, demographics, and follow‐up data were collected using the electronic medical records from the hospital database. We analyzed postprocedural and long‐term aortic gradients, as well as 1‐year and long‐term survival outcomes.

The Ethics Committee of the Medical University of Graz, Austria (EK 31–323 ex 18/19), granted approval for the study, which adhered to the 1964 Declaration of Helsinki and its subsequent revisions.

3.2. Statistical Analysis

All analyses were performed with Prism GraphPad 9.0 (GraphPad Software Inc., California, US). Summary descriptive statistics are reported as mean ± standard deviation or n (%), as appropriate. Normal distribution was tested by D'Agostino‐Pearson omnibus normality test. Continuous variables were compared by Mann−Whitney tests or Kruskal−Wallis test and categorical variables were compared with Fisher's exact or chi‐square tests, as appropriate. A probability value of p < 0.05 was considered as significant.

4. Results

All patients who received a DF valve at our center were included in the study (DF group, n = 67). The obtained data were compared with an age‐, echocardiographic parameters‐, and overall risk status‐matched cohort of patients who underwent TAVI with another commercially available device during the same period (Reference group, n = 67).

In the DF group, the mean age was 82.6 ± 6.5 years, compared to 82.5 ± 6.3 years in the Reference group (p = 0.83). Significantly more male patients received a DF valve compared to the Reference group (54% vs. 34%, respectively; p < 0.05). Arterial hypertension was more frequent in the Reference group (98% vs. 82%, respectively; p < 0.05), while patients in the DF group had a significantly higher glomerular filtration rate (61 ± 20 mL/min vs. 52 ± 18 mL/min, respectively; p < 0.05). Other cardiovascular risk factors (dyslipidemia and diabetes mellitus), baseline characteristics, overall health condition, and operative risk status (as calculated by the EuroScore II and STS score) were comparable between the two groups.

No significant differences were observed regarding the history of percutaneous coronary artery interventions (21% vs. 25%, respectively; p = 0.68). However, significantly more patients with a history of coronary artery bypass grafting received a DF valve (11% vs. 2%, respectively; p < 0.05). Patient clinical characteristics and comorbidities are detailed in Table 1.

Table 1.

Clinical characteristics and comorbidities.

	DF group	n = 67	Reference group	n = 67	p
	n/mean	% or SD	n/mean	% or SD
Age	82.6	±6.5	82.5	±6.3	0.83
Male gender	36	54	23	34	< 0.05
Hypertension	55	82	66	98	< 0.05
Dyslipidemia	34	51	42	63	0.22
Diabetes mellitus	18	27	11	16	0.20
GFR	61	20	52	18	< 0.05
History of PCI	14	21	17	25	0.68
History of CABG	11	16	2	3	< 0.05
EuroScore II	9.26	4.48	10.23	4.54	0.224
STS Score	8.01	3.78	8.28	4.51	0.754

Abbreviations: CABG, coronary artery bypass graft; DF, Direct Flow; GFR, Glomerular filtration rate; PCI, percutaneous coronary intervention; SD, standard deviation; STS, Society of Thoracic Surgeons.

At presentation, considerably more patients in the Reference group were diagnosed with more than mild aortic insufficiency on echocardiography compared to the DF group (18% vs. 6%, respectively; p < 0.05). However, there was only one case of more‐than‐moderate aortic insufficiency in the Reference group before the intervention.

After the TAVI procedure, more‐than‐mild intravalvular aortic regurgitation occurred in five cases (7%) in the DF group, comparable to three cases (4%) in the Reference group (p = 0.71). There were no cases of severe paravalvular regurgitation in either group (Table 2).

Table 2.

Echocardiographic Parameters at presentation, Post TAVI‐Implantation and 1‐year follow‐up.

	DF group	n = 67	Reference group	n = 67	p
	n/mean	% or SD	n/mean	% or SD
Baseline echocardiographic parameters
LVEF	50	13	51	13	0.51
Mean gradient	46	15	47	18	0.91
More then mild AR	6	10	18	27	< 0.05
Postprocedural echocardiographic parameters
LVEF	51	11	56	10	< 0.05
Mean gradient	16	8	8	3	< 0.01
More then mild AR	5	7	3	4	0.71
Echocardiographic parameters at 12‐months FU
LVEF	55	14	56	9	0.78
Mean gradient	16	6	8	3	< 0.05
More then mild AR	1	4	0	0	0.41

Abbreviations: AR, aortic regurgitation; DF, Direct Flow; FU, follow‐up; LVEF, left ventricular ejection fraction; SD, standard deviation.

The postprocedural mean gradient was significantly higher in the DF group (16 ± 8 mmHg vs. 8 ± 3 mmHg, respectively; p < 0.01) (Figure 7), while the left ventricular ejection fraction was lower (51 ± 11% vs. 56 ± 10%, respectively; p < 0.05). Furthermore, only one patient in the DF group and three patients in the Reference group (two treated with CoreValve and one with an Edwards valve) required permanent pacemaker implantation for high‐grade or complete AV block (1.5% vs. 4.5%, respectively; p = 0.62).

Figure 7.

Postprocedural mean gradient in the DF group versus Reference Group. DF, Direct Flow; TAVI, transcatheter aortic valve implantation. [Color figure can be viewed at wileyonlinelibrary.com]

At 1‐year follow‐up, the mean gradient remained significantly higher in the DF group (16 ± 6 mmHg vs. 8 ± 3 mmHg, respectively; p < 0.05); however, neither group showed relevant progression compared to their postprocedural values (p = 0.60 and p = 0.55, respectively).

Survival at 12 months was comparable between the DF and Reference groups (6% vs. 4% mortality, respectively; HR 1.37, 95% CI 0.31–6.02), as shown in the Kaplan‐Meier survival analysis (Figure 8). Survival remained comparable at maximal follow‐up of 39 ± 25 months (31% vs. 25% mortality, respectively; HR 1.44, 95% CI 0.76–2.73).

Figure 8.

Kaplan−Meier survival curve for 12‐months survival in the DF and reference group. [Color figure can be viewed at wileyonlinelibrary.com]

At maximal follow‐up, in the DF group, death from any cause occurred in 30% (n = 20) of cases, while thre patients (4%) presented with stroke, five patients (7%) experienced myocardial infarction, and one patient (1%) underwent elective post‐TAVI PCI. In the Reference group, similar rates were observed: death from any cause occurred in 17 patients (25%; p = 0.69), stroke in one patient (1%; p = 0.61), myocardial infarction in two patients (3%; p = 0.44), and elective PCI was performed in four cases (3%; p = 0.36) (Table 3).

Table 3.

Postprocedural complications and outcomes at maximal Follow‐up. Mechanical and hemodynamic complication include ventricular rupture, hemodynamic collapse and need for resuscitation.

	DF group	n = 67	Reference group	n = 67	p
	n/mean	% or SD	n/mean	% or SD
Postprocedural complications permanent pacemaker	1	1.5	3	4.5	0.62
Mechanical/hemodynamic complications	3	4.5	4	6	≈1
Long‐term outcomes at maximal FU Death of any cause	20	30	17	25	0.69
Stroke	3	4	1	1	0.61
Myocardial infarction	5	7	2	3	0.44
Elective post‐TAVI PCI	1	1	4	3	0.36

Abbreviations: FU, Follow‐up; PCI, percutaneous coronary intervention; TAVI, transcatheter aortic valve implantation.

5. Discussion

As TAVI extends into younger and lower‐risk populations, the long‐term fate of early‐generation devices has become critically important. Among these, the DF prosthesis, unique for its dual polymeric rings and fully retrievable, non‐metallic frame, presents particular challenges when structural valve deterioration occurs.

Degeneration of bioprosthetic valves is inevitable, driven by factors such as heavy leaflet calcification, incomplete expansion at implant, and mechanical fatigue [17]. In surgical bioprostheses, redo‐SAVR was most widely accepted treatment for a long time, but carries substantial perioperative risk, 4.6% in‐hospital mortality for repeat SAVR versus 2.2% for first‐time procedures [15]. As ViV‐TAVI emerged, registry data demonstrated lower 30‐day mortality versus redo‐SAVR, though surgery may afford fewer long‐term heart‐failure readmissions [19, 20]. US approval of Sapien and CoreValve ViV is limited to high‐risk patients [21], underscoring the need for careful heart‐team decision‐making when DF fails.

Although DF's design facilitated precise initial placement, its lower radial force and variable ring expansion can complicate subsequent ViV interventions [15, 22]. DF's polymer rings, once filled, may recoil under external force, deforming leaflet geometry and reducing effective orifice. Successful ViV‐TAVI in DF prostheses depends on precise coaxial alignment between the new transcatheter heart valve and the DF's ventricular ring [23]. Karaduman et al. report underexpansion of a 23 mm Sapien XT in a 25 mm DF, remedied by 25 mm balloon post‐dilation to lower gradients [21]. Butter et al. similarly corrected early restenosis by deploying a 20 mm Sapien XT and aggressive post‐dilation [24]. In our experience, considering the structure and fail mechanism of the DF, using a balloon‐expendable valve for the ViV intervention as well as performing balloon postdilatation under rapid pacing is indispensable to overcome DF's lower radial force and achieve optimal THV expansion.

Furthermore, DF's epoxy‐polymer rings may be prone to faster structural fatigue than bovine pericardial leaflets on metal frames [15]. Yap et al. underscored that ring expansion variability and leaflet wear can precipitate both stenosis and regurgitation within 3 years [22]. Our case, degeneration at 7 years dominated by regurgitation, highlights that DF failure mechanisms may vary (leaflet degeneration vs ring recoil), necessitating individualized evaluation.

A propensity‐matched analysis by Edlinger et al. demonstrated significantly higher DF gradients at discharge and at 6 months with worse 2‐year MACE and survival in DF patients [15]. In our DF cohort of patients, we observed higher residual gradients immediately post‐TAVI and at 12 months, yet comparable survival at 12 months and through 39 ± 25 months follow‐up. These performance limitations likely stem from DF's lower radial force and perimeter‐based sizing, which may predispose to prosthesis mismatch and accelerated degeneration [15].

Another important concern when performing Redo‐TAVI, as well as for the initial TAVI procedure, is avoiding obstruction of the coronary ostia. Factors that may play an important role and contribute to this issue, are patient related, such as female gender, low ostia position, valve‐to‐coronary distance or sinus of Valsalva width, as well as procedure related, like oversizing or malpositioning of the catheter valve. All this contribute to a three times higher likelihood of coronary obstruction for reinterventions, in comparison to the baseline TAVI procedure [25]. Although neither in our experience nor prior case series reported coronary compromise, prophylactic coronary wiring and on‐table angiography are prudent, especially when CT identifies bulky calcification near coronary cusps or low ostial heights [24, 26]. In our described case, given the importance of not losing coronary access or obstructing the LM ostia while delivering the ViV system, we opted to protect the LM ostia by placing a workhorse wire in the LCx, during the inflation of the valve balloon, but also for the postdilatation (Figures 5 and 6). The wire could safely be removed after measuring invasive transvalvular gradients.

Although DF prostheses are no longer in commercial use, clinicians must be prepared to manage their failure. Nevertheless, with meticulous consideration of all described aspects, along with considering potential challenges regarding coronary access in case of required future consecutive interventions and also the risk of pacemaker dependency, redo‐TAVI is the intervention of choice in selected cases, especially when supported by preprocedural CT imaging, careful annular sizing, and readiness for high‐pressure post‐dilation.

The main limitations to our study are the relatively small sample size of a single center setting, and the retrospective, non‐randomized design, which are reducing the statistical power and strength of the study findings. Also, we recognize the need of longer follow‐up for drawing conclusive clinical outcomes.

6. Summary

DF valves show comparable degeneration pattern and long‐term outcomes, compared to other TAVI prostheses. Considering the time of its on‐market availability, patients with degenerated DF can be expected recently. In that case, ViV procedure after DF implantation requires special precautions and then it can be safely performed.

Ethics Statement

The study was conducted in full conformity with the 1964 Declaration of Helsinki and all subsequent revisions, as well as in accordance with the guidelines of the International Conference on Harmonization for Good Clinical Practice (ICH GCP E6 guidelines).

Consent

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy restrictions.