Baseline population

The baseline study population comprised 264 patients hospitalized between 2020 and 2024 in 1 of 2 adjacent cardiology departments at a tertiary care center. Patients with any nondegenerative aortic pathology, past surgical interventions on the aortic valve, or congenital heart defects (except for bicuspid aortic valve) were excluded.

All patients included in the study were diagnosed with moderate or severe AS, and were referred for either conservative management or an invasive intervention during their index hospitalization. All invasive procedures, including pre- and perioperative care, were performed at an affiliated tertiary care center, following a joint decision of the local heart team and the patient. The qualification for TAVI, SAVR, or conservative treatment was based on the available guidelines for the management of valvular heart disease⁹ and the expertise of the local dedicated “valvular” heart team with over 15 years of experience. The criteria considered during the qualification comprised the severity and hemodynamic profile of AS, coronary artery anatomy, left ventricular function, along with general clinical characteristics, including age, comorbidities, and fragility status, as well as patient preferences.

The study was approved by the Jagiellonian University Medical College Ethics Committee (118.0043.1.218.2025). Informed consent was provided verbally during the telephone interview.

Follow‑up protocol

A practicing physician from the corresponding cardiology department initiated each follow‑up assessment via a telephone call. If the patient did not answer the first call, their proxy (an authorized person indicated by the patient) was contacted instead. The patient proxies were instructed to answer on the patient’s behalf. Up to 3 telephone calls, either to the patient or their proxy, were attempted before a given patient was deemed lost to follow‑up. Each interview followed a predefined protocol (described in detail in Supplementary material, Table S1), starting with obtaining oral consent for participation in the study. The patients who refused to participate were excluded from all subsequent aspects of the study.

During the interview, occurrence of dyspnea, angina, syncope, and vertigo during 2 distinct short‑term intervals was retrospectively assessed to limit the confounding effects of preprocedural AS progression and postprocedural recovery. The preprocedural interval was set between the first and second month before the procedure, and the postprocedural interval between the first and second month after the procedure. The pre- and postprocedural severity of dyspnea and angina were assessed using the New York Heart Association (NYHA) scale and the Canadian Cardiovascular Society (CCS) scale, respectively. The patients were asked to rate how much having undergone a TAVI or SAVR procedure subjectively affected their QoL. A QoL grading scale of 0–10 was used, with a score of 10 meaning the best possible outcome, a score of 5, no significant outcome, and a score of 0, the worst possible outcome. Postprocedural major adverse cardiovascular events (MACEs) were defined as the first instance of either nonfatal myocardial infarction (MI), nonfatal ischemic stroke, cardiovascular‑related hospitalization (due to heart failure decompensation, any thromboembolic events or any cardiac arrhythmias requiring hospital care), or cardiovascular‑related death (due to MI, ischemic stroke, or end‑stage heart failure, or sudden cardiac death).

A total of 32 patients (12.1%) from the baseline population were deemed lost to follow‑up, and further 14 (5.3%) refused to participate in the study. Of the remaining 218 patients, only 116 (53.2%) were deemed eligible for inclusion in the final study population based on the postinterview exclusion criteria, as listed in Supplementary material, Figure S1 and Table S2.

Clinical parameters

All clinical parameters and comorbidities analyzed in the final study population were those assessed during the index hospitalization. Basic echocardiographic parameters, namely the aortic valve area, peak velocity of the aortic valve (V_max), mean aortic valve pressure gradient (AVG), left ventricular ejection fraction (LVEF), stroke volume, left ventricular posterior wall thickness at end‑diastole, left atrial area (LAA), and cardiac output, were measured in each patient. All echocardiographic examinations were performed during the first week of the index hospitalization by an experienced echocardiographer as part of routine diagnostic workup before evaluation by the heart team, in accordance with the guidelines for diagnostic imaging of AS.⁹ A senior heart team consultant supervised the diagnostic imaging process and validated the results for all included patients. All echocardiographic measurements were taken using a GE Vivid E9 or GE Vivid E95 ultrasound machine (General Electric, Boston, Massachusetts, United States). The measured LAA was indexed to the body surface area (LAA index [LAAi]). Detailed descriptions of the echocardiographic measurements are presented in Supplementary material, Table S3.

The N‑terminal pro‑B–type natriuretic peptide levels were measured in 82 patients. A total of 105 patients underwent routine coronary angiography during their hospitalization. A reduction by at least 75% in the luminal diameter of any main coronary artery or its major branches (or a reduction by >50% in the case of the left main coronary artery) was considered as indicative of significant coronary stenosis.

Statistical analysis

Variables are expressed as mean with SD or median with interquartile range (IQR), depending on the data normality, as assessed by the Shapiro–Wilk test. Categorical variables were compared using the χ² test. The t test or Wilcoxon rank sum test was used for 2‑group comparisons. The relationship between the QoL score and selected clinical parameters was analyzed using β regression models with a logit link function. For regression analysis, the QoL score values were linearly rescaled to an open unit interval (0, 1) and treated as a quasi‑continuous variable. Logistic regression models were built to predict the postprocedural decrease (by at least 1 class category) in NYHA and CCS classes. Each regression model included an interaction term between the procedure type (SAVR vs TAVI) and a predictor value.

The threshold for statistical significance in all tests was set at a P value below 0.05. All analyses were performed using RStudio (version 4.3.2, R Foundation for Statistical Computing, Vienna, Austria).

Baseline characteristics

A total of 116 patients with AS treated with either TAVI or SAVR were included in the study. Of those, 43.1% underwent TAVI, while 56.9% underwent SAVR (Table 1). The mean (SD) patient age was 72 (9.3) years, and women constituted 42.2% of the population. Three individuals presented with moderate AS during the index hospitalization, while the remaining patients were initially diagnosed with severe AS. The most common comorbidities at the index hospitalization were hypertension (78.4%), diabetes mellitus (34.5%), and chronic kidney disease (32.8%). No TAVI or SAVR procedures were repeated in any of the patients.

The median (IQR) waiting time for the procedure (ie, the period between the end of the index hospitalization and the procedure) was 4 (1–8.3) months for all patients, and was longer in the TAVI than the SAVR group (6.5 [3–10.8] vs 3 [1–5] months, respectively; P = 0.001). The median follow‑up (ie, the period between the procedure and the follow‑up call) was 18.5 (10–26) months for all patients, 15 (7–23) months for the TAVI group, and 22 (12.75–28.75) months for the SAVR group, with a significant difference between the groups (P = 0.02).

Major adverse cardiovascular event occurrence

In the final study population, 20% of the patients experienced at least 1 postprocedural MACE (Table 1). Cardiovascular‑related hospitalization was the most common postprocedural MACE (18.1%), and was less prevalent in the TAVI than the SAVR patients (8% vs 25.8%, respectively; P = 0.03). In the SAVR group, there were 3 instances of stroke, 3 instances of MI, and 3 instances of cardiovascular‑related death. In the TAVI group, only 1 cardiovascular‑related death was observed, and there were no instances of stroke or MI.

Symptom severity

Up to 2 months before the procedure, 77.6% of all patients were affected by dyspnea (NYHA class II–IV), 47.4% by angina (CCS class I–IV), 53.4% by vertigo, and 25.9% by syncope (Table 1). Dyspnea persisted 2 months postprocedure in only 25.9% of the study population, angina in 15.5%, vertigo in 24.1%, and syncope in 3.4%. Postprocedural dyspnea was more prevalent in the TAVI than the SAVR group (40% vs 15.1%, respectively; P = 0.01), even though preprocedurally, the rate of dyspnea was comparable between the groups (82% for TAVI vs 74.2% SAVR; P = 0.53). Vertigo was the only symptom consistently less prevalent in the SAVR than the TAVI group both before (44.6% vs 66%; P = 0.04) and after the procedure (16.1% vs 36%; P = 0.03). Moreover, the rate of asymptomatic patients (ie, without dyspnea, angina, syncope, or vertigo) after the procedure was greater in the SAVR than the TAVI group (60.2% vs 42%; P = 0.01).

Logistic regression analysis for predicting a postprocedural decrease in the NYHA and CCS classes was performed using a separate model for each respective clinical parameter and symptom class decrease (Supplementary material, Figure S2). The patients with higher values of AVG and V_max were predisposed to a decrease in the NYHA class (odds ratio [OR], 1.21; 95% CI, 1.02–1.48 for AVG and OR, 1.12; 95% CI, 1.03–1.23 for V_max; Supplementary material, Figure S2A). A greater LAA was a factor predisposing to a decrease in the CCS class (OR, 1.55; 95% CI, 1.08–2.39), whereas a greater LAAi had a marginal but nonsignificant effect on a decrease in the CCS class (OR, 1.15; 95% CI, 1–1.36; Supplementary material, Figure S2B).

Differences between the SAVR and TAVI patients regarding the effects of echocardiographic parameters on symptom relief were analyzed using the interaction terms. The effect of a greater LAA on a decrease in the CCS class was marginally stronger in the SAVR group (OR, 0.63; 95% CI, 0.35–1.06 for the interaction term). Similarly, the effect of a greater LAAi on a decrease in the CCS class was marginally stronger in the SAVR group (OR, 0.85; 95% CI, 0.69–1.02 for the interaction term). The effects of other analyzed parameters on a decrease in the CCS and NYHA classes were comparable between the SAVR and TAVI patients (Supplementary material, Figures S3 and S4).

Quality of life score

The median QoL score assessed in the study population was 8 (6–10) points, with a noticeable tendency toward higher ratings overall (Table 1). A maximal score (10 points) was declared by 30.2% of the patients, and a positive score (QoL >5) was reported by 75%. Although the TAVI patients declared marginally lower scores than the individuals treated with SAVR, the difference was not significant (median QoL score, 7.5 [5–9] vs 9 [7–10], respectively; P = 0.06). The total number of patients reporting maximal QoL scores (TAVI vs SAVR, 24% vs 34.9%; P = 0.27) and positive QoL scores (TAVI vs SAVR, 70% vs 78.8%; P = 0.31) was comparable between the groups.

Postprocedural decrease in dyspnea severity had a nonconclusive effect on the QoL scores. A decrease in NYHA class was associated with higher QoL scores in the entire study population (P = 0.02) but not in the TAVI (P = 0.06) or SAVR (P = 0.1) groups when analyzed separately (Supplementary material, Table S4). In contrast, postprocedural reduction in angina severity was more strongly associated with higher QoL scores. The patients who experienced a CCS class decrease reported higher QoL scores than those without such a decrease, both in the whole study population (P = 0.01) and in the subgroup analysis (P = 0.03 for TAVI and P = 0.04 for SAVR). Among the comorbidities assessed during the index hospitalization, only a history of ischemic stroke was associated with lower QoL scores in the whole patient population (P = 0.01). MACE occurrence was associated with lower QoL scores in the SAVR (P = 0.004) but not in the TAVI group (P = 0.94).

The only significant predictor of the QoL score in the β regression analysis was LAA, which was greater in the patients reporting an increased QoL score (OR, 1.21; 95% CI, 1.01–1.44; Supplementary material, Figure S2C). A greater LAAi had a marginal effect on the QoL score (OR, 1.07; 95% CI, 1–1.15). The subsequent analysis of interaction terms showed no significant variation in the effect depending on the procedure type for any of the analyzed parameters (Supplementary material, Figure S5). However, the positive effect on the QoL score was marginally stronger in the SAVR group for LAA (OR, 0.78, 95% CI, 0.6–1.03 for the interaction term) and LAAi (OR, 0.91; 95% CI, 0.82–1 for the interaction term).

Limitations

The main limitation of this study is its small sample size, contributing to the marginal significance and uncertainty of the reported results. Data regarding periprocedural parameters or follow‑up echocardiographic assessments were not available. The qualification for TAVI and SAVR was performed by the same heart team, and only a single medical care center was included in the data collection process, which limits the generalizability of the acquired results. The interviews conducted during this study could yield imprecise measures of symptom severity due to potential cognitive impairment in elderly respondents. To address this issue, the decrease in symptom severity was summarized as a binary outcome, rather than being analyzed as the magnitude of the NYHA or CSS class change. As the timeframe allotted for each follow‑up call was limited, a simplified QoL grading scale was used in conjunction with the NYHA and CCS class assessment, rather than more robust, multidimensional questionnaires. Despite these limitations, we believe that this study provides valuable insights into the symptom‑specific determinants of postprocedural QoL improvement based on the in‑hospital evaluation before TAVI and SAVR.

Conclusions

Several echocardiographic parameters assessed in patients with AS before TAVI or SAVR procedures were associated with a postprocedural decrease in symptom severity and QoL outcomes. The patients with an elevated preprocedural AVG or V_max had greater odds of experiencing a postprocedural decrease in dyspnea severity. Similarly, an enlarged LAA was associated with a higher likelihood of decreased angina severity and better patient‑reported QoL. Patient age, sex, preprocedural LVEF, and aortic valve area were not associated with either a reduction in symptom severity or improvement in QoL.

The consideration of LA dimensions together with key hemodynamic parameters, including AVG and V_max, can guide the shared decision‑making process by helping clinicians counsel patients on the likelihood of improvement in both QoL and symptom status after TAVI and SAVR procedures. Integrating those parameters into clinical evaluation supports a patient‑centered care approach, aligning intervention strategies with individual patient profiles and anticipated benefits.