Introduction

Transcatheter aortic valve replacement (TAVR) has emerged as an effective minimally invasive treatment option for patients suffering from severe aortic valve stenosis (AS) and aortic regurgitation (AR). These patients are deemed high-risk for surgical aortic valve replacement1. Nowadays, TAVR has experienced rapid advancement, driven by innovative technological breakthroughs and strong clinical evidence emerging across the world. Quality of life (QoL) is a crucial factor in predicting treatment effectiveness and has significant prognostic value, making it extremely important in medical decision-making2. Changes in QoL may better reflect patients’ experiences during disease progression and treatment-related complications. This, in turn, may lead to refinements and enhancements in treatment approaches and patient care3. Although the Placement of Aortic Transcatheter Valves (PARTNER) trial4 and another study5 have demonstrated an improvement in QoL for patients after TAVR, there remains a dearth of systematic and comprehensive research exploring the trajectory of QoL over time for those who have undergone TAVR.

Frailty is one of the most prevalent factors that influence the QoL of patients after TAVR. It not only has a major effect on the recovery ability of these older patients6,7 but also acts as a risk factor leading to poor outcomes. Such outcomes include longer hospital stays, disabilities, and even death after the TAVR procedure8,9,10. Significantly, frailty has been identified as an independent factor that can predict a deteriorating QoL. It was found that the risk of frail patients not having an improvement in their QoL one year after TAVR was two times higher than that of non-frail patients11,12. As a general concept, frailty is a clinical condition characterized by reduced capacities in multiple physiological systems, determining a state of increased vulnerability to negative psychological conditions like anxiety or depression and susceptibility to adverse health-related QoL13.

As reported from a worldwide study involving 40 countries, poorer cardiac function characterized by the New York Heart Association (NYHA) class had a strong correlation to poorer health-related QoL of heart failure patients14. Heart failure patients with complex life demands would lead to adverse psychosocial stressors not merely anxiety and depression15. Meanwhile, anxiety and depression are established predictors of adverse clinical outcomes in elderly cardiovascular patients with multiple comorbidities, significantly compromising their QoL16,17,18. Currently, less is known about how mental health is related to TAVR. This includes the prevalence of depression and anxiety during the TAVR perioperative period in real-world situations, as well as the outcomes after TAVR that are associated with the patient’s baseline mental health status19. Based on findings from other settings, it can be inferred that depression in TAVR patients is connected with a variety of functional declines, particularly manifesting as frailty, reduced mobility, and impaired QoL19. Surman et al. outlined specific prospective outcomes related to depression, frailty, and QoL in TAVR patients. Nonetheless, this finding did not go deeper into exploring the inherent relationships and mechanisms among these three factors20.

To fill these knowledge gaps, we aimed to investigate the relationships and figure out the interaction mechanisms between frailty, mental health, cardiac function, and QoL after TAVR procedures. Our conceptual framework (Fig. 1) posits three core hypotheses for patients who had TAVR. First of all, we assumed that frailty had a direct negative influence on mental health and QoL, while cardiac function had a direct positive influence on mental health and QoL for these patients. Secondly, we supposed that the mental health of patients after TAVR had a direct negative influence on their QoL. Thirdly, we hypothesized that frailty indirectly influenced the QoL of patients after TAVR through the mediating role of their mental health. Additionally, we also hypothesized that patients’ cardiac function indirectly influenced their QoL by the effects of mental health.

Fig. 1
figure 1

The hypothesized model.

Full size image

Methods

Study design, setting, and participants

This study was a cross-sectional descriptive study designed to evaluate the interconnections among frailty, cardiac function, mental health, and QoL. It took place at a public hospital in Sichuan Province, China.

A total of 396 patients who had undergone TAVR were selected as convenience samples in this study from February 2022 to November 2023. Patients were eligible for inclusion if (1) they were elderly with severe calcified stenosis of the aortic valve; (2) they had symptoms like palpitation, exertional dyspnea, angina, chest pain, heart failure, syncope, a history of presyncope, or presyncope during exercise testing; (3) they were asymptomatic severe AS with a left ventricular ejection fraction (LVEF) < 55% (in the absence of left ventricular systolic dysfunction caused by other incentives); (4) they were AS patients<70 years old who were at high risk of surgical procedures or have other risk factors, such as those who have had chest radiotherapy, suffer from liver failure, have severe diffuse aortic calcification, or are extremely frail21. Patients with mental diseases who were unable to communicate were excluded from the study.

Measurement instruments and variables

Exogenous variable: frailty

Frailty was measured using the Frailty Phenotype (FP), the initial and most widely used assessment scale for providing an operational definition of physical frailty. This scale consists of 5 items, with at least three of the five criteria required to diagnose frailty: gait speed over 5 m, grip strength, weight loss, self-reported exhaustion, and inactivity22. If a patient meets only one or two of these criteria, they are classified as pre-frailty, while meeting none is categorized as non-frailty. Frailty is diagnosed when three or more criteria are met.

Endogenous variable: quality of life

The short-form health survey 12 (SF-12) is a shortened version of the SF-36 Health Survey from the United States. It provides a more concise alternative to assess the QoL23. The SF-12 is scored on a 0-100 scale containing 12 items that contribute to two components: the physical component summary (PCS-12) and the mental component summary (MCS-12). Raw scores from each item response have to be linearly transformed to a 0-100 scale to calculate PCS-12 and MCS-12 scores24.

Endogenous variable: mental health

Mental health was measured utilizing the Hospital Anxiety and Depression Scale (HADS), which is a commonly employed brief screening instrument in hospital settings to detect emotional disorders25,26. It comprises 14 items, of which seven items are used to assess the anxiety dimension (HADS-A), and the remaining seven are applied to assess the depression dimension (HADS-D). Each item is rated on a 4-point Likert scale ranging from 0 to 3. A raw score exceeding 8 suggests a mild disorder, while a score surpassing 10 indicates a moderate disorder.

Endogenous variable: heart functional status

The patients’ cardiac functional status was evaluated using the NYHA class27 and brain natriuretic peptide (BNP) levels28. The NYHA class is widely used in clinical practice to assess the degree of heart failure. It comprises four classifications ranging from class I, indicating no physical limitations during activity, to class IV, indicating symptoms present even when the patient is at rest. BNP and the aminoterminal part of its pro-hormone (NT-pro-BNP) are cardiac hormones mainly secreted by ventricular cardiomyocytes. They have the property of vasodilation and can help reduce preload and afterload. The levels of these hormones are correlated with myocardial wall stress and the severity of AS29.

Data collection and ethical considerations

Our study adhered to the Declaration of Helsinki guidelines30. Approval was granted by the Ethics Committee of the West China Hospital (2022 Review No. 909). We obtained the informed consent of every participant before the study initiation. On the second day after the TAVR operation, once the study’s purpose had been thoroughly explained, eligible participants independently completed the baseline demographic survey and self-report questionnaires (SF-12, FP, and HADS scale). They did this without any interruption or guidance. The NYHA class and the NT-pro-BNP values were extracted from the electronic medical records before the operation. Out of 423 potential patients, 396 consented to take part in the study, yielding a response rate of 93.6%.

Statistical analysis and model testing

We employed IBM SPSS Statistics (version 24.0; made by IBM, Chicago, IL, USA) and Amos (version 20.0; made by IBM, Chicago, Illinois, USA) to perform the statistical analysis. A p-value < 0.05 was regarded as statistically significant. Continuous variables were presented as mean ± standard deviation or were described in the median and interquartile range (IQR), while categorical variables were expressed as frequency and the percentage distribution within each group. We conducted a Spearman correlation analysis to examine the relationships between all the variables. Additionally, independent-sample t-test and one-way ANOVA were conducted to explore the differences in the main variables of interest among different sociodemographic groups.

A structural equation model (SEM) was utilized to validate the hypothesized model and yielded the path coefficients of the factors. The goodness-of-fit index of SEM and their acceptable threshold levels were reported to evaluate the appropriateness of the hypothesized model31. The χ2/degrees of freedom ratio (df) < 3.0, the goodness-of-fit index (GFI) > 0.9, the adjusted goodness-of-fit index (AGFI) > 0.9, the root mean square error of approximation (RMSEA) < 0.08, the nonstandard fit index (NFI) > 0.9, the relative fit index (RFI) > 0.9, the incremental fit index (IFI) > 0.9, the Tucker-Lewis index (TLI) > 0.9, and the comparative fit index (CFI) > 0.9 indicate a well-fitting structural model32,33,34,35. In addition, we adopted bootstrapping analysis with a sample size of 2000 to estimate indirect pathways36.

Results

Demographic characteristics

In our study, most respondents were male (60.1%) and resided in urban areas (65.4%). The mean age of all the patients, male patients, and female patients was 72.03 (SD 7.06), 72.59 (SD 7.33), and 71.19 (SD 6.55) years, respectively. The majority of patients who did not report experiencing a history of syncope (90.2%) were in a partnered relationship (88.4%) and exhibited abnormal NT-pro-BNP levels (90.9%). More than half of the patients (52.8%) had a normal BMI. Among all the patients with cardiovascular disease, only 32.8% had more than five comorbidities. Concerning the NYHA class, a large proportion of the patients (55.8%) were grouped into level II or lower, while the rest were at level III or higher. Nearly half of the patients (48.2%) were classified as having a pre-frailty phenotype. Table 1 shows the demographic profile of the participants.

Table 1 The demographic characteristics and the between-group differences of participants.
Full size table

The profiles for all the variables of interest

Specifically, the means and standard deviations for total SF-12, PCS-12, MCS-12, HADS-A, and HADS-D were 50.29 (SD 20.035), 33.32 (SD 23.931), 67.14 (SD 22.764), 4.51 (SD 2.469), 4.19 (SD 2.497), respectively. Details are presented in (Table 2). The median value of NT-pro-BNP was 1962.5 (interquartile range [IQR] 567 to 4859.75). The mean scores of HADS-A and HADS-D indicated that there were no significant levels of anxiety and depression in the patients. There were statistically significant differences in the scores of total SF-12, PCS-12, and MCS-12 when grouped by NYHA class and frailty phenotype. Regarding HADS-A and HADS-D, they had statistically significant differences when considering factors like frailty phenotype, age groups, and history of syncope.

Relationships among frailty, QoL, mental health, and cardiac function

The Spearman correlation analysis revealed statistically significant positive correlations between the NYHA class and the NT-pro-BNP value (r = 0.172, P < 0.01), the NYHA class and frailty (r = 0.225, P < 0.01), and between the NT-pro-BNP value and frailty (r = 0.110, P < 0.05). Additionally, positive correlations were observed between the NYHA class and the HADS-D scores (r = 0.106, P < 0.05), and between frailty and both the HADS-A scores (r = 0.221, P < 0.01) and HADS-D (r = 0.191, P < 0.01). In contrast, all the scores of the total SF-12, the PCS-12, and the MCS-12 were negatively correlated with the NYHA class and frailty. This implies that patients with poorer heart function, as indicated by a higher NYHA class, and those with a frailty phenotype generally had a lower QoL. These results are presented in (Table 2).

Table 2 Correlations, means and standard deviations of variables of interest.
Full size table

Testing of the hypothesized model

In summary, the fit index (χ2[114] = 452.55, p < 0.01, NFI = 0.989, RFI = 0.973, IFI = 1.005, TLI = 1.011, CFI = 1.00, RMSEA < 0.01) of the examined model yielded a good fit to our data. As illustrated in Fig. 2, among all the eight path coefficients within the structural model under examination, all but two were statistically significant. The exceptions were the path coefficient directly linking cardiac function to mental health and the indirect path coefficient from cardiac function to QoL.

Fig. 2
figure 2

Standardized coefficients of the structural equation model. **P < 0.05; HADS-D the hospital anxiety and depression scale-depression dimension, HADS-A the hospital anxiety and depression scale-anxiety dimension, NYHA the New York heart association class, BNP brain natriuretic peptide, NT-pro-BNP the aminoterminal part of BNP’s pro-hormone, SF-12 the short-form health survey 12, PCS-12 the physical component summary of SF-12, MCS-12 the mental component summary of SF-12.

Full size image

To be specific, frailty had a statistically direct effect on patients’ mental health (β = 0.237, 95% CI 0.068 to 0.365, p < 0.01) and QoL (β=-0.375, 95% CI -0.524 to 0.193, p < 0.01). Similarly, both patients’ mental health (β = 0.159, 95% CI 0.031 to 0.30, p < 0.05) and cardiac function (β=-0.356, 95% CI -0.599 to -0.168, p < 0.01) showed significant direct effects on QoL. When we modeled patients’ mental health as a mediator, the results of the bootstrapping analysis indicated that it significantly mediated the relationship between frailty and QoL (β = 0.038, 95% CI 0.006 to 0.072, p < 0.05). Conversely, it did not impact the relationship between cardiac function and QoL (β = 0.009, 95% CI -0.019 to 0.064, p > 0.05). Table 3 illustrates the standardized total, direct, and indirect effects of the structural equation model we examined.

Table 3 Parameter estimates of the standardized pathways in the structural equation model.
Full size table

Discussion

Given the increasing number of TAVR procedures being performed worldwide and the expanding availability of this treatment option, gaining insights into patients’ experiences and post-procedure outcomes is of great importance. To the best of our knowledge, our study is the first to examine the relationships between frailty, mental health, cardiac function, and QoL in patients with TAVR. We found that the relationship between frailty and QoL was mediated by the effect on patients’ mental health.

In our study, male patients outnumbered female patients. Most of the patients were older than 65 years, which is consistent with the findings from a nationwide study in the United States37. Our findings revealed that among females, the group with lower SF-12 scores and PCS-12 scores tended to be more common compared to males. This suggests that tailored interventions based on gender differences might be necessary38. There were significant between-group differences in the HADS-A and HADS-D scores among patients of different ages. This highlights the complex biological process of aging, which may make individuals more prone to depression or anxiety disorders39. Furthermore, robust evidence indicates that the occurrence of syncope, angina, or other symptoms, along with increased blood levels of BNP/NT-pro-BNP, can predict the rapid progression of AS process29,40,41. Although only a small proportion of our participants had experienced syncope, most of them reported abnormal NT-pro-BNP levels. It is well-established that unpredictable and sudden syncope episodes could trigger anxiety and depression42. However, caution is warranted when interpreting the differences in the HADS-A, HADS-D, and MCS-12 scores between the two groups due to the sample size disparity. Conflicting findings have been reported regarding the association between elevated BNP/NT-proBNP levels and adverse outcomes after TAVR, with some studies supporting the association while others do not43,44,45. It has been speculated that low BNP/NT-proBNP levels in patients with AS may indicate an underdeveloped compensatory mechanism, which is a substrate for poor outcomes after TAVR46. Previous research has developed biomarker interventions to improve clinical and QoL outcomes substantially47. Nevertheless, we did not find any statistically significant differences between groups based on NT-pro-BNP levels regarding outcomes related to anxiety, depression, and QoL. Maybe NT-pro-BNP by itself is not the only biomarker influencing the QoL of patients with valvular diseases. Furthermore, while NT-pro-BNP serves as an objective indicator partially reflecting patients’ cardiac function, it may not necessarily have a direct relationship with subjective indicators of individual mental health status.

Compared to the previous two studies involving high surgical-risk AS patients48,49, our participants demonstrated distinct clinical characteristics, including younger age and better NYHA levels. Our findings revealed higher QoL scores, PCS-12 scores, and MCS-12 scores than those reported in these two studies. Correspondingly, other findings also indicated significant differences in QoL across different NYHA class levels50. This discrepancy may reflect an enhancing patient-reported health status that is typically associated with younger age and a lower risk of heart failure51. Heart failure patients with lower NYHA class not only had less debilitating physical symptoms but also reported less anxiety and depression, resulting in a better QoL52.

The mean scores for anxiety and depression after TAVR observed in this study were lower than in the previous study17. Maybe there exists a certain number of undisclosed cases of anxiety or depression among TAVR patients. This could stem from the self-reported attributes of the data and unmeasured cultural disparities53. Besides, as seen from Table 1, most of the patients in our study have partners. Moreover, due to the relatively high cost of the TAVR surgery, the patients who underwent the TAVR procedure had a relatively high family income, and the overall family support is good, which reduces the likelihood of anxiety and depression. Additionally, most of the patients had less than 5 comorbidities (67.2%). There were few patients with severe symptoms such as syncope (9.8%) and a relatively small proportion of frail patients (20.7%). Therefore, the degree of anxiety and depression caused by disease symptoms among our participants is relatively low.

Our results also showed that only a small portion (20.7%) of our participants experienced frailty. Inconsistent with similar research findings, we observed a relatively lower proportion of frail participants8,9. This could be partially attributed to the younger mean age of our study cohort. On the other hand, this finding may be attributed to the high proportion of participants with normal or elevated BMI (90.4%) in our study cohort, which is associated with higher bone mineral density and hormonal osteo-protective effects that can help weight-bearing and mobilization54. Therefore, the impact would lead to a lower risk of frailty development.

Exhaustion symptoms are one of the major frailty phenotypes in patients with valvular disease and are commonly observed in heart failure that is objectively assessed by NYHA class and BNP/NT-pro-BNP levels55,56,57. It is identical that we found frailty was positively associated with significantly increased NT-pro-BNP levels and as well a higher NYHA class. BNP/NT-proBNP, secreted by the myocardium in response to increased mechanical wall stress, the exceptionally high levels (> 10000ng/L) and low levels (< 800ng/L) have been confirmed a strong correlation with poor functional improvement after the TAVR procedure assessing by NYHA class44. When we examined the association of QoL, mental health, and the frailty phenotype, we documented significant negative interactions between the frailty phenotype and the SF-12 scores, PCS-12 scores, and MCS-12 scores. In addition, positive interactions between the frailty phenotype and the HADS-A and the HADS-D scores were noted. Frail patients with diminished physiological reserve are impaired in physical activity, endurance, mobility, strength, and other aspects. Therefore, frailty is an independent predictor of QoL for patients undergoing TAVR11. However, a TAVR procedure that only treats anatomic abnormalities might not be sufficient to continuously improve overall QoL for patients whose frailty dynamically fluctuates among different states22,58,59. Some indicators of frailty, like exhaustion, have an effect on individuals’ emotions. Strong bidirectional correlations between frailty, anxiety, and depression have recently been proven in elderly individuals with heart failure60. Effective prevention and treatment of anxiety and depression may hold promising prospects for improving frailty by increasing physical and social activities61.

The SEM findings from our study further the understanding of mental health status, particularly anxiety and depression, as a potential mediator in the relationship between frailty and QoL for patients following TAVR. Boureau et al.62 indicated that the baseline depression level in patients may not, at the very least, necessarily correlate with a worsening of their QoL six months after TAVR. Conversely, other studies have shown a correlation between depression and QoL both before and after the TAVR procedure63. We also observed not unanimous results in this study. There was no significant correlation between the HADS-A scores or HADS-D scores and the SF-12 scores, the PCS-12 scores, or the MCS-12 scores in the correlation analysis. However, using SEM, we found a direct correlation between mental health status and QoL. It is not surprising, given that the proportion of participants with anxiety (14.2%) or depression (11.1%) was relatively low, and the incidence of post-TAVR complications that impair their QoL was also low. SEM can delve deeper into the underlying relationships among variables compared to the Spearman correlation analysis, and it can also uncover some mediating effects. Besides, the uncertain reliability of self-reported scales is one of the reasons contributing to the inconsistent results. There are studies have reported that the prevalence of anxiety and depression symptoms based on scales was roughly three times higher than other diagnostic means64.

According to the Frailty Phenotype criteria, it may seem natural that frail individuals’ physical component QoL would be worse than those classified as robust. The mental QoL was also much worse among frail patients than among robust individuals. It can be speculated that frailty, even if mainly defined in physical terms, can substantially impact the mental components of the QoL65. This insight may provide a mechanism suggesting that incorporating interventions involving both physical and psychological factors could potentially improve QoL for patients undergoing TAVR.

There are several limitations in this study. Firstly, the cross-sectional study design we adopted restricted our possibility to interpret trends and trajectories longitudinally about the factors of interest. Therefore, it warrants further longitudinal research to explore the dynamics concerning QoL, frailty, anxiety, and depression among perioperative TAVR patients. Secondly, the chosen measurements for assessing QoL, frailty, anxiety, and depression are not the sole choices available. We were forced to make a selection, and our SEM lacked external validation, which downgraded the comparability and generalizability of the results. Finally, although the concept of frailty is increasingly being used, its translation from research to clinical practice remains a challenge in the coming years.

In summary, this is the first study that adds theoretical insights into patients’ QoL and its determinants, including patients’ cardiac function, frailty phenotype, and mental health status. In the future, it would be beneficial to launch campaigns to raise awareness of frailty among patients and conduct targeted interventions before and after TAVR. Therefore, specificity and standardization of frailty measurements are essential for its generalization66. Additionally, efforts should take into account the potential effects of mental health on the overall QoL of patients after TAVR, requiring beyond just anxiety and depression. Our results could also serve as an impetus for further research and the conduct of more extensive prospective trials, including additional psychological interventions targeting patients following TAVR67.