Long-term outcomes following transcatheter aortic valve implantation with the Portico self-expanding valve

Transcatheter aortic valve implantation (TAVI) is currently the standard of care in Europe for patients aged 75 and higher with severe symptomic aortic valve stenosis [1‐3]. Ongoing studies are evaluating whether patients younger than 75 years might also benefit from TAVI [3‐6], and indeed in the US, patients are eligible for TAVI from 65 years. These advances towards lower risk patients increases the awareness of life time management of patients with severe aortic stenosis [7‐10]. However, limited data is available on the long-term clinical and hemodynamic performance of these transcatheter heart valves.

Several devices for TAVI are currently available, with specific differences in terms of delivery system and THV design [11]. The Portico valve (Abbott Vascular, Santa Clara, CA, USA) is a self-expanding intra-annular device which is characterized by the ease of delivery and satisfactory early and mid-term outcomes, but limited evidence on long-term effectiveness [12‐15].

We aimed at evaluating the performance of Portico beyond the early and mid-term timeframes, and conducted an international retrospective observational study focusing on long-term (≥ 3 years) clinical and imaging outcomes in patients undergoing TAVI with Portico.

The current study is a spontaneous, not-for-profit, international, multicenter, observational and retrospective cohort study. Baseline clinical data, preprocedural transthoracic echocardiography with aortic valve gradients and procedural details were collected from 7 high-volume European TAVI centers (> 100 cases per year). The details on patients undergoing TAVI with the Portico valve were obtained by querying institutional databases, with ethical approval whenever appropriate. In particular, details on Italian centers were obtained from the ongoing the Registro Italiano GISE sull’impianto di Valvola Aortica Percutanea (RISPEVA) study, approved at Pineta Grande Hospital, Castel Volturno, Italy, as well as elsewhere (ClinicalTrials.gov Identifier: NCT02713932). Thus, all methods were carried out in accordance with relevant guidelines and regulations, all experimental protocols were approved by Pineta Grande Hospital, Castel Volturno, Italy, and informed consent was obtained from all subjects and/or their legal guardians. Notably, all patients receiving Portico at participating centers could be included, without any formal exclusion criteria, save for exclusion of individuals unwilling to allow anonymized data collection.

Briefly, we collected several baseline data including clinical history, comorbidities, and surgical risk scores. Baseline imaging data included transthoracic echocardiography details, such as left ventricular ejection fraction, aortic valve area, peak aortic valve gradient, mean aortic valve gradient, and aortic regurgitation. Patient-prosthesis mismatch, while not systematically sought, was approximated using a mean aortic valve gradient > 10 mmHg definition. Procedural features were systematically sought, including sheathless strategy, access site, predilation and postdilation. Clinical outcomes included all-cause death, cardiovascular death, stroke, myocardial infarction, major vascular complication, life-threatening bleeding, permanent pacemaker implantation, rehospitalization for cardiovascular causes, valve-related dysfunction requiring repeat procedure, endocarditis, leaflet thrombosis, and New York Heart Association class, applying whenever appropriate current Valve Academic Research Consortium 3 definitions [16]. The primary outcomes of interest were all-cause death and major adverse events (the composite of all-cause death, stroke, myocardial infarction, and reintervention for valve degeneration).

Descriptive analysis was based on the computing mean ± standard deviation for continuous variables, supplemented by median (1st quartile; 3rd quartile) whenever appropriate, and count (%) for categorical variables. Percentile bootstrapping techniques was used (1000 repetitions) to compute 95% confidence intervals of event rates. Failure curves were generated according to Kaplan and Meier, with hypothesis testing based on the log-rank test. Comparisons of interest for these log-rank tests included those based on the tertiles of age, and those based on the tertiles of surgical risk. We also identified independent predictors of death and major adverse events by conducting an iterative series of bivariate Cox proportional hazard analyses, and then including in the final model only predictors associated with p < 0.05 for the outcome of interest. The results of Cox proportional hazard analyses were reported as point estimate of hazard ratio, with corresponding 95% confidence intervals and p values. Competing risk analyses were performed for sensitivity purposes. Exploratory analyses focusing on Portico valve size were performed using unpaired Student’s t tests for continuous variables, and Fisher’s exact tests for dichotomous variables. In addition, we explored the impact of institutional learning phase (first 30 cases vs subsequent ones). Complete case analysis was used throughout. Statistical significance was set at the 2-tailed 0.05 level. Computations were performed with Stata 13 (StataCorp, College Station, TX, USA).

A total of 803 patients undergoing TAVI with Portico implantation were included (Table 1), with most procedures performed in 2016, 2017, and 2018 (Table 1S), and 3 centers contributing more than 100 cases each (Table 2S). The mean age was 82.2 ± 5.6 years, 504 (62.8%) were women, and 386 (48.1%) patients were at low surgical risk (EuroSCORE II < 3.0%), with median median EuroSCORE II of 3.1% (1^st quartile: 2.0%; 3^rd quartile: 5.8%), and median STSPROM score of 4.0% (2.8; 6.0%). Left ventricular systolic function was preserved in most patients and at least moderate aortic regurgiation was present in 162 (21.0%) subjects (Table 3S). Femoral access and local anesthesia were used in the vast majority of cases (Table 2) and predilation was used liberally (677 [84.6%]) cases. Device success was achieved in 778 (97.1%) procedures, with peri-procedural death occurring in 8 (1.0%), need for emergency surgery in 7 (0.9%), cardiac tamponade in 11 (1.4%), valve embolization in 14 (1.8%), and bailout TAVI-in-TAVI in 15 (1.9%).

After a minimum follow-up of 3 years (median 3.0 years [3.0; 4.0]), major adverse events occurred in 301 (37.5% [95% confidence interval: 34.1%; 40.9%]) patients (Fig. 1). Death was the most common adverse event (35.1% [31.8%; 38.4%]) (Table 3; Fig. 1S). Although cardiovascular death was more frequent in the first 2 years after the procedure, noncardiovascular death occurred more frequently thereafter (Table 4S). The major vascular complications occurred in 40 patients (8.0% [5.6%; 10.3%]), permanent pacemaker implantation rate was 23.1% [19.4%; 26.7%], reintervention for valve degenaration was necessary in 9 (1.1% [0.6%; 2.8%]) cases, and, finally, 4 cases of endocarditis were reported, for an incidence of 0.5% (0.0%; 1.0%).

Long-term hemodynamic performance was assessed using echocardiography after a median follow-up of 3.1 years (2.0; 4.8). Mean aortic valve gradient at follow-up was 8.1 ± 4.6 mmHg, and at least moderate aortic regurgitation was present in 9.1% (6.7%; 12.3%) (Table 5S). Bioprosthetic valve failure, i.e. the composite of death, reintervention, and severe aortic regurgitation, occurred in 35.4% (32.0%; 38.8%). Mild-to-moderate or moderate aortic regurgitation at discharge was stable over time, with no case of worsening and improvement in most patients. Notably, smaller devices were associated with higher gradients, more patient-prosthesis mismatch, less regurgitation, and more adverse outcomes, even if this latter finding was largely due to worse risk profile (Tables 6S, 7S, and 8S).

Clinical outcomes were not significantly impacted by age (Fig. 2S), but surgical risk was a significant bivariate predictor of events (Fig. 2). Further exploratory bivariate analysis assessing the impact of the device size suggested that TAVI with Portico valve size 25 mm or smaller was associated with an increase in risk of major adverse events (p = 0.026), stroke (p = 0.042), and higher peak aortic valve gradient (p = 0.045) (Table 8S; Fig. 3), but these findings were not confirmed at multivariable analysis. Indeed, the baseline differences in patient risk according to device size were most likely the cause of such discrepancies in event rates (e.g. EuroSCORE II was 4.8 ± 0.3 in patients receiving smaller devices in comparison to 4.0 ± 0.2 in those receiving larger devices, p = 0.044).

Transcatheter aortic valve implantation has seen a momentous expansion since the first human case description by Cribier et al. in [17]. The success of TAVI rests on many aspects, including refinements in patient selection, preprocedural imaging, devices, ancillary equipment (e.g. delivery systems), accompanying techniques (e.g., transaxillary access), and concomitant medical therapy [3]. Developments in devices have been quite substantial, and range from improvements of first-generation devices to the advent of completely original second-generation devices, such as Portico.

In summary, TAVI with the Portico valve, when performed in selected patients and by experienced structural heart teams, appears associated with favorable clinical outcomes over long-term (> 3 years) follow-up, with few cases of valve deterioration or infection. Clinical outcomes were largely impacted by baseline risk factors and surgical risk, whereas valve size did not impact independently on clinical or imaging endpoints.