This was an observational, single center, prospective study. We included all patients who underwent TAVI in our university hospital from November 2008 to June 2017 (n = 399) were included. All patients were selected for transcatheter replacement according to the clinical practice guideline recommendations available at the time; only patients with a life expectancy of more than 1 year and severe symptomatic aortic stenosis were included. The indication for TAVI was made according to the guidelines available at the time of inclusion. All patients had been previously discussed by a Heart Team consisting of clinical cardiologists, interventional cardiologists, and cardiac surgeons. All patients voluntarily signed consent forms before the procedure.

The major factors that contributed to the decision to perform the transcatheter procedure vs surgical aortic valve replacement were high or unacceptable surgical risk, frailty associated with older age, and technical contraindications for surgery (most frequently the presence of porcelain aorta).

In most patients, TAVI was performed under local anesthesia and conscious light sedation. General anesthesia was used in 8% (n = 30) of procedures, when the nonfemoral arterial approach was preferred. The femoral approach was used in most procedures. When this was not feasible, axillar artery was the selected approach. We used the standard technique as described in the literature.12

A Medtronic biological prosthesis were implanted in most patients: CoreValve, CoreValve Evolut R or CoreValve Evolut Pro (Medtronic Inc., Minneapolis, MN, USA), depending on the availability at the time of the procedure. In a small percentage of patients, ACURATE-Neo (Symetis S.A., Ecublens, Switzerland) devices were used.

Before the intervention, 2-dimensional echocardiography and diagnostic coronary angiography were performed in all patients. As part of our routine protocol, computed tomography was used to determine aortic root anatomy and adequacy of vascular accesses. Any complication during the procedure was registered according to Valve Academic Research Consortium-2 (VARC-2) consensus document.13

This study was performed in accordance with the principles of the Helsinki Declaration.

All data related to the event were registered in the patients’ electronic medical records. In our TAVI Registry, follow-up was performed using previous registries by trained cardiologists. Our protocol includes telephone calls and review of the electronic medical records. All medical interventions, hospital admissions and pharmacological treatments were reviewed. Vital status was determined by telephone calls in the absence of medical records. No patient was lost to follow-up.

Following the established protocol at least 1 follow-up echocardiogram was performed in all patients after discharge and another one 3 months later. Subsequently an annual echocardiogram was performed.

AHF was defined following the available practice guidelines at the time of recruitment,14 based on clinical and radiological data. The nutritional risk index (NRI) was calculated as 1.519 × serum albumin (g/L) + 41.7 × (actual body weight [kg]/ideal body weight [kg]), using the modified formula for the elderly by Bouillanne et al.15 Ideal body weight was determined using the Lorentz formula:15 height (cm) −100 −([height (cm) − 150]/4) for men, or height (cm) −100 −([height (cm) −150]/2.5) for women. If the ratio of measured body weight (kg) to ideal body weight (kg) was ≥ 1, the assigned value was 1, as previously described.15,16 The NRI was calculated using the body weight measured on the day of the TAVI procedure, and the albumin value was obtained from the blood sample performed the day before the procedure. Based on NRI values, we classified the patients into 4 groups: no nutritional risk (NRI > 100), mild nutritional risk (97.5 ≤ NRI < 100), moderate nutritional risk (83.5 ≤ NRI < 97.4), and severe nutritional risk (NRI < 83.5). To simplify the model, we obtained 2 risk categories after the combination of the following: no and mild nutritional risk, and moderate and severe nutritional risk.

The presence and severity of paravalvular aortic regurgitation was assessed using transthoracic echocardiography. Transesophageal echocardiography was performed in those patients with a suboptimal acoustic window.

The statistical analysis was performed with SPSS version 22.0 and Stata version 13. Baseline characteristics according to the development of post-TAVI heart failure (HF) during follow-up are described using number and percentage for categorical data and mean ± standard deviation for continuous data, respectively. Differences in characteristics were assessed by using chi-square tests and 2-sample Student t tests.

The association between post-TAVI HF and mortality was evaluated by Cox proportional hazards regression analyses with “post-TAVI HF” as a time-varying covariate. Results were graphically shown with Kaplan-Meier curves. Because HF hospitalization and death are semicompeting risks in which death precludes a subsequent HF hospitalization but death can still occur after a HF hospitalization, an illness-death, acyclic, multistate model was used.17 In this model, all participants were in the initial state of “discharge after TAVI” and were at risk of a HF hospitalization (transition 1) or death without a preceding HF hospitalization (transition 2). In addition, those who were hospitalized for HF were also at risk for death after a HF hospitalization (transition 3) (Figure 1). The illness-death regression model using Weibull parametrization was developed to model the effect of covariates on the cause-specific hazards of the 3-state transitions with separate (stratified) nonparametric baseline hazards for transitions into the “post-TAVI HF” state and into the “death” state. All variables associated with post-TAVI HF based on P < .05 in the univariate analyses were included in a multivariate model, together with those with clinical relevance. Hazard ratios (HR) were calculated with 95% confidence intervals (95%CI).

Figure 1.

Heart failure-death multistate model transitions. TAVI, transcatheter aortic valve implantation.

(0.1MB).

A total of 399 patients with severe AS underwent TAVI and were included in our registry between 2008 and 2017. The mean age of the cohort was 82.4 ± 5.8 years, and 53.4% (n = 213) were women.

Table 1 shows the baseline characteristics of the total population and of each group, including medical history, echocardiographic features, procedural details and in-hospital outcomes.

After a mean follow-up period of 27 ± 24.1 months and median of 21 months [interquartile range 6.5-40.7], 119 patients (29.82%) were admitted with a final diagnosis of AHF (cumulative incidence function 13.2%; 95%CI, 11.1%-15.8%) (Figure 2A). The average time until presentation of HF after the procedure was 20.9 ± 21.3 months with a median of 16.1 months [interquartile range 3.2-32.2]; 39.5% of AHF episodes (n = 47) occurred during the first 6 months after valve implantation.

Figure 2.

Cumulative incidence of HF after TAVI (A) and mortality according to HF (B). HF, heart failure; TAVI, transcatheter aortic valve implantation.

(0.22MB).

The results of the multivariate analysis are shown in Table 2. Prior to the procedure, AHF episodes and high STS score were the only independent predictors of postprocedure AHF. There was no difference between groups in left ventricular ejection fraction (not even when we stratified according to the latest HF guidelines classification)18.

During follow-up, 150 deaths occurred in our cohort: 31 patients (20% of total deaths) during the first 30 days and 119 during the remaining follow-up (Figure 2B). The factors associated with higher mortality are summarized in Table 3 (univariate) and Table 2 (multivariate; Table 2 and Table 3 for with and without AHF during follow-up, respectively). After adjustment for these factors, AHF was a strong independent predictor of mortality (HR, 1.84; 95%CI, 1.14-2.97; P < .12), with almost twice the mortality rate in comparison with those without follow-up AHF (Figure 3).

Figure 3.

Adjusted survival Kaplan-Meier curves of TAVI patients based on follow-up 95%CI, 95% confidence interval; HF, heart failure; HR, hazard ratio; TAVI, transcatheter aortic valve implantation.

(0.2MB).

Among the 119 patients who developed HF during follow-up, 62 (52.1%) died, with an average time between the first HF event and death of 25.7 ± 16.7 months, and a median of 16.8 months [interquartile rage 7.0-35.5]. Table 4 shows the predictive factors for mortality in this patient subgroup. Univariate analysis of patients readmitted with AHF who died showed that these patients were older, with a higher rate of AF and kidney failure, were at higher risk of malnutrition (assessed by NRI), and had increased values of N-terminal pro-B-type natriuretic peptide. Statistically significant differences in echocardiographic measurements were found only for mitral regurgitation greater than moderate and aortic regurgitation grade III or higher. In the multivariate analysis (Table 2 and Table 3), we only identified reduced NRI (HR, 0.93; 95%CI, 0.89-0.97; P = .002) and chronic obstructive pulmonary disease (HR, 2.35; 95%CI, 1.15-4.80; P = .018) as variables significantly associated with higher mortality (Figure 4).

Figure 4.

Adjusted survival Kaplan-Meier curves for independent predictors of mortality in patients with heart failure after transcatheter aortic valve implantation. 95%CI, 95% confidence interval; COPD, chronic obstructive pulmonary disease; HR, hazard ratio.

(0.24MB).

Several studies and registries have shown a high readmission rate in patients after TAVI.19 The rate of early (within 30 days) readmission ranged from 4.0% to 17.9%, and was higher in those patients who underwent TAVI through a transapical approach. The readmission rate during the first year post-TAVI was as high as 50% of patients, mostly for noncardiovascular causes. A study by Nombela-Franco et al.20 reported an incidence of 43.9% for all-cause readmissions up to 1 year after TAVI. Among them, 58.9% were admitted for noncardiovascular causes (mainly due to pre-existing comorbidities) and 41.1% for cardiac causes, mainly AHF (23.3%). Similar results were reported by Durand et al., 21 with 1-year total readmission rate and for AHF of 52.2% and 24.1%, respectively. Of note, in both studies the valve used was the SAPIEN device. Our registry, which mainly examined patients receiving the CoreValve device, shows similar results: 30% of our population was admitted due to AHF during follow-up, less than a half of these admissions being during the first year after the procedure (46.2%).

At the end of follow-up, 150 patients had died (37.59%). Our results are similar to previous published data by Avanzas et al.22: that group also remarked on the relevance of admission due to HF in TAVI patients, highlighting 92.6% of deaths after an admission. There are scarce reports on the determinants and prognosis of AHF after TAVI. As far as we know, the study by Durand et al.21 was pioneer in reporting the impact of AHF on mortality after TAVI. Among patients discharged from hospital, the rate of all-cause mortality was 13.7% at 1-year, and was 31.4% after a mean follow-up period of 27.2 + 0.7 months. Readmission due to AHF after TAVI was strongly associated with higher mortality at 1 year (24.2% vs 10.4%, P < .0001) and at the end of follow-up (50.0% vs 25.6%, P < .0001). Our results are fairly similar. After a mean follow-up period of 27 ± 24.1 months, mortality in patients admitted due to AHF was 52.1% vs 31.4% (HR, 1.84; 95%CI, 1.14-2.97; P < 0012). Nombela-Franco et al.20 also assessed the impact of early hospital admission on mortality, with a mean follow-up similar to that one in our study. Their reported mortality rate at 2 years was significantly increased in those patients who were admitted within 30 days after TAVI compared with those who were not readmitted (30.2% vs 19.2%; P = .002). Approximately 30% to 50% of the readmissions were related to cardiovascular causes, mostly AHF, and had a major prognostic impact on mortality.

Two independent factors were identified as predictors for AHF after the index discharge: an episode of hospital admission for AHF before TAVI implantation and high risk measured by STS score. Durand et al.21 found 4 independent predictors for AHF: low aortic mean gradient before TAVI, postprocedural blood transfusion, severe persistent postprocedural pulmonary hypertension, and left atrial dilatation, of which only 2 were procedure-related. In this study, a previous episode of AHF before TAVI failed to achieve statistical significance in the multivariate analysis. Baron et al.23 reported the impact of aortic valve gradient on the outcomes of TAVI in a large series of patients (n = 11 292); low aortic valve gradient (< 40mmHg) was associated with higher mortality (HR, 1.21; 95%CI, 1.11 - 1.32; P < .001) and higher rates of AHF (HR, 1.52; 95%CI, 1.36-1.69; P < .001) with no effect of LVEF. We observed no significant impact for pre-TAVI aortic valve gradient or for LVEF. These discrepancies may be explained by demographic and clinical differences between the populations.

In our registry, despite pulmonary arterial pressure being higher in the group that developed AHF after TAVI, it did not have an impact on prognosis in our cohort. Several publications showed that pulmonary hypertension before TAVI is frequent and increases mortality after TAVI.24,25 Patients with persistent severe pulmonary hypertension after TAVI have worse prognosis than those with a decrease in pulmonary artery systolic pressure below 60mmHg (2-year mortality rate 50.0% vs 18.6%, P = .001). It was postulated that right heart catheterization could therefore aid in Heart Team decision-making.

Major bleeding and transfusion is frequent following TAVI and was associated with an increased risk of early and late mortality.26 Durand et al.21 observed that severe bleeding and, particularly, the need for transfusions were independent predictors of admission due to AHF after TAVI. In our cohort the need for transfusions was not associated with increased AHF, but it was a marker of higher mortality after implantation.

NRI and chronic obstructive pulmonary disease after valve implantation were found to be key elements influencing the prognosis of this group of patients. Our results show for the first time how nutritional status assessed by using NRI is a powerful independent prognostic factor. NRI is a validated tool for estimating the risk of undernutrition in various populations. The NRI shows a strong correlation with mortality, adverse events and deterioration of functional capacity, which is superior to that achieved using body mass index and albumin separately.27–30 Malnutrition is highly prevalent and has been reported to be an independent risk factor for clinical events in HF. A large proportion of patients hospitalized for HF have moderate to severe malnutrition, and low NRI is associated with more readmissions and higher mortality in patients with AHF, as well as with higher mortality in patients with chronic HF.27–30 A relationship between classic nutritional status markers, such as body mass index and hypoalbuminemia, and prognosis after TAVI (with a J-shape curve) has been established in previous studies.31

Assessment of malnutrition risk must be part of the geriatric assessment of patients who undergo TAVI and plays a determining role in frailty status. The potential use of the NRI for early identification of patients at risk of malnutrition who are under assessment for TAVI could be highly relevant in daily clinical practice. Such patients could potentially benefit from interventions to improve their nutritional status prior to undergoing the procedure. Larger studies are needed to validate NRI within a geriatric and frailty score, alone or as a part of other predictive scores for events after TAVI.

Chronic obstructive pulmonary disease is common in HF patients, and its presence in those with systolic dysfunction is associated with an increased burden of comorbidities, lower use of evidence-based HF medications, longer hospital stays, and increased in-hospital noncardiovascular mortality.32