Introduction

Aortic stenosis (AS) is a prevalent condition, especially among the elderly. In patients aged over 75 years, the prevalence of AS is 12.4%, and severe AS requiring intervention stands at 3.4%, signifying a significant disease burden1. Aortic valve replacement (AVR) encompasses two treatment modalities for AS: surgical aortic valve replacement (SAVR) and transcatheter aortic valve replacement (TAVR). SAVR used to be the standard treatment modality for severe symptomatic AS2. In recent times, TAVR has emerged as a less invasive alternative, especially for patients deemed high-risk or unsuitable for traditional surgery3,4.

Disparities in healthcare outcomes are often associated with socioeconomic status (SES). A significant body of research indicates that people with lower SES tend to face increased post-operative mortality following various surgical intervention procedures5,6,7,8. When it comes to cardiovascular diseases, the population from lower SES typically shows a greater prevalence of risk factors. Such disparities often stem from restricted access to specialized medical care and lower adherence to prescribed treatments5,8,9.

In patients undergoing AVR, while prior research has found SES disparities in accessing TAVR10,11,12, there is limited data on the effect of SES on post-procedural outcomes in either TAVR or SAVR. Thus, this study aimed to conduct a population-based analysis to assess the influence of SES on AVR outcomes.

Methods

Data source

Using the National/Nationwide Inpatient Sample (NIS), patients diagnosed with aortic stenosis (AS) who underwent either TAVR or SAVR procedures were identified. This identification was based on the International Classification of Diseases, 10th Revision, Procedure Coding System (ICD-10-PCS) and the International Classification of Diseases, 10th Revision, Clinical Classification (ICD-10-CM) from the last quarter of 2015 to 2020. The specific codes used for AS included I06.0, I06.1, I06.2, I06.8, I06.9, Q23.1, I35.1, I35.0, and Q23.0. For TAVR, the codes were 02RF3JZ, 02RF38Z, 02RF3KZ, 02RF37Z, 02RF3JH, 02RF38H, 02RF3KH, and 02RF37H. For SAVR, the codes used were 02RF08Z, 02RF0JZ, 02RF0KZ, and 02RF07Z. The study excluded patients under the age of 18 years.

The NIS database categorizes patients into quartiles based on the estimated median household income of residents in the patient's ZIP Code. Accordingly, patients were grouped into these four quartiles reflecting their neighborhood's income. Table 1 provides a breakdown of patients from each of these income quartiles who were diagnosed with AS and underwent TAVR or SAVR. To assess the impact of SES on AVR, only isolated TAVR or SAVR were considered, excluding concomitant procedures including coronary artery bypass grafting (CABG; ICD-10-PCS: 0210xxx), percutaneous intervention (PCI; ICD-10-PCS: 02703xx, 02713xx, 02723xx, 02733xx), and mitral valve replacement (MVR; ICD-10-PCS: 02RGxxx). Patients from neighborhoods in the lowest and highest income quartiles were chosen as the cohorts for comparison in isolated TAVR and SAVR, respectively.

Table 1 A summary of patients from various SES backgrounds diagnosed with AS and who underwent TAVR/SAVR and isolated TAVR/SAVR. TAVR/SAVR includes isolated TAVR/SAVR and AVR with concomitant procedures.
Full size table

Pre-procedural variables

Tables 2, 3, 4 and 5 summarize the comparison of pre-procedural factors for isolated TAVR and SAVR between patients from low and high-income neighborhoods. These factors encompass age, sex, primary payor status, and various hospital characteristics. Hospital characteristics include bed size, geographical location, and teaching status of the hospital. The classification of hospital bed sizes is determined by both the hospital's location and its teaching status. The American Hospital Association's annual survey refines these categories into small, medium, and large. Patient comorbidities were ascertained using the Elixhauser measure13 and ICD-10-CM codes, as detailed in Supplementary Table S1. In addition, All Patients Refined Diagnosis Related Groups (APR-DRG) Risk of Mortality Subclass was calculated to characterize the risk of the patients14.

Table 2 Demographic, primary payer status, hospital characteristics, transfer status, and admission status comparing patients living in neighborhoods of income at the lowest and highest quartiles who underwent isolated TAVR.
Full size table
Table 3 APR-DRG score, comorbidities, and clinical conditions comparing patients living in neighborhoods of income at the lowest and highest quartiles who underwent isolated TAVR.
Full size table
Table 4 Demographic, primary payer status, hospital characteristics, transfer status, and admission status comparing patients living in neighborhoods of income at the lowest and highest quartiles who underwent isolated SAVR.
Full size table
Table 5 APR-DRG score, comorbidities, and clinical conditions comparing patients living in neighborhoods of income at the lowest and highest quartiles who underwent isolated SAVR.
Full size table

Outcome variables

Post-procedural outcomes encompassed in-hospital mortality, major adverse cardiovascular events (MACE), myocardial infarction (MI), stroke, transient ischemic attack (TIA), neurological and pericardial complications, pacemaker implantation, cardiogenic shock, respiratory complications, mechanical ventilation, acute kidney injury (AKI), post-procedural renal failure, venous thromboembolism (VTE), pulmonary embolism (PE), hemorrhage/hematoma, surgeries reopened for bleeding control, infections, sepsis, both deep and superficial wound complications, vascular complications, diaphragmatic paralysis, transfers to other hospital facilities, the wait time from admission to operation, hospital length of stay (LOS), and the total hospital charge. The specific ICD-10 codes used to define these outcomes are detailed in Supplementary Table S1. It is important to note that the NIS database does not capture post-discharge patient follow-up. As a result, all post-TAVR/SAVR outcomes presented in this study are derived exclusively from in-hospital data.

Statistical analysis

Binary pre-procedural variables for patients from low- and high-income neighborhoods undergoing isolated TAVR or SAVR were assessed using Fisher's exact test. To calculate the odds of patients from different quartiles of neighborhood household income undergoing SAVR and TAVR compared to those from the lowest quartile of neighborhood household income, as well as binary post-procedural outcomes after isolated TAVR or SAVR, multivariable logistic regression was used, adjusting for pre-procedural variables that exhibited substantial differences (p < 0.1) in the Fisher's exact test. This provided adjusted odds ratios (aORs) along with their 95% confidence intervals (CI). Continuous variables, including time from admission to operation, LOS, and total hospital charges, were analyzed using generalized linear models (GLM), adjusting for all pre-procedural factors.

All statistical analyses were conducted using SAS (version 9.4). A p-value below 0.05 was considered statistically significant. The authors had full access to the dataset and assumed full responsibility for the accuracy and integrity of the analyses. Given the de-identified nature of the NIS database and the study's retrospective design, it received an exemption from the Institutional Review Board (IRB) approval at The George Washington University.

Ethics approval

This study was exempt from the IRB approval by The George Washington University as it analyzed retrospective, deidentified NIS data.

Results

A summary of the distribution among patients of different SES who had diagnosis of AS and those who underwent TAVR or SAVR are shown in Table 1. From the last quarter of 2015 to 2020, there were 613,785 adult patients presented with AS diagnosis, where 59,976 (9.77%) and 62,171 (10.13%) of them underwent TAVR and \SAVR, respectively. The rates that AS patients underwent TAVR or SAVR gradually decreased with lower average income in the neighborhoods, even after adjusting for pre-procedural differences as shown by the significant aORs in Table 1. In addition, the ratio of TAVR/SAVR decreased with lower income in the community. After excluding concomitant procedures, 58,064 patients underwent isolated TAVR, and 43,694 underwent isolated SAVR. In isolated TAVR, 12,355 (21.28%) patients resided in neighborhoods with incomes in the lowest quartile, while 15,212 (26.2%) lived in neighborhoods with incomes in the highest quartile. For isolated SAVR, 10,029 (22.95%) patients were from neighborhoods in the lowest income quartile, and 10,811 (24.74%) were from those in the highest income quartile.

In isolated TAVR, compared to patients living in the higher-income neighborhoods, those from lower-income neighborhoods were more likely to be female (p < 0.01), aged less than 75 years (p < 0.01), be African American (p < 0.01), Hispanic (p < 0.01), Native Americans (p < 0.01), use Medicaid (p < 0.01) as the primary payer, stay in a rural hospital (p < 0.01), urban private practice (p < 0.01), or a hospital with small bed size (p < 0.01), transferred in from a different acute care hospital (p < 0.01), and under emergent admission (p < 0.01) (Table 2). In contrast, patients from lower-income neighborhoods were less likely to have age over 85 years (p < 0.01), be Caucasian (p < 0.01) or Asian (p < 0.01), use Medicare (p < 0.01) or private insurance (p = 0.04) as the primary payer, or stay in urban teaching hospitals (p < 0.01) (Table 2).

In isolated TAVR, patients from lower-income neighborhoods were more likely to have APR-DRG Class 3 (p < 0.01) and Class 4 (p = 0.01), diabetes (p < 0.01), drug abuse (p = 0.03), complicated hypertension (p < 0.01), chronic pulmonary disease (p < 0.01), obesity (p < 0.01), peripheral vascular disease (p = 0.03), advanced renal failure (p < 0.01), pulmonary hypertension (p < 0.01), anemia (p < 0.01), coronary artery disease (p < 0.01), previous myocardial infarction (MI; p < 0.01) and previous CABG (p < 0.01) (Table 3). In contrast, patients from lower-income neighborhoods were less likely to have autoimmune conditions (p = 0.01), lymphoma (p = 0.01), leukemia (p = 0.01), metastatic cancer (p = 0.01), solid tumor without metastasis, malignant (p < 0.01), uncomplicated hypertension (p < 0.01), hypothyroidism (p = 0.02), other thyroid disorders (p < 0.01), atrial fibrillation (p < 0.01), first-degree (p < 0.01), second-degree (p = 0.02), or complete atrioventricular block (p < 0.01), hyperlipidemia (p < 0.01), or thrombocytopenia (p = 0.01) (Table 3).

In isolated SAVR, patients from lower-income neighborhoods were more likely to be female (p < 0.01), aged less than 55 years (p < 0.01), be African American (p < 0.01), Hispanic (p < 0.01), Native American (p < 0.01), use Medicaid (p < 0.01) and self-pay (p < 0.01) as the primary payer, stay in rural hospital (p < 0.01), urban private practice (p < 0.01), hospitals with medium bed size (p < 0.01), transferred in from a different acute care hospital (p < 0.01), transferred in from another type of health facility (p < 0.01), and under emergent presentation (p < 0.01) (Table 4). Patients from lower-income neighborhoods were less likely to have ages over 65 years (p < 0.01), be Caucasian (p < 0.01) or Asian (p < 0.01), under private insurance (p < 0.01), stay in urban teaching hospital (p < 0.01) or a hospital with large bed size (p < 0.01) (Table 4).

In isolated SAVR, compared to patients from high-income neighborhoods, those living in lower-income neighborhoods were more likely to have APR-DRG class 4 (p < 0.01), acquired immune deficiency syndrome (p < 0.01), alcohol abuse (p < 0.01), cerebrovascular disease (p < 0.01), dementia (p = 0.01), depression (p < 0.01), diabetes (p < 0.01), drug abuse (p < 0.01), complicated hypertension (p < 0.01), chronic pulmonary disease (p < 0.01), obesity (p < 0.01), paralysis (p < 0.01), advanced renal failure (p < 0.01), pulmonary hypertension (p < 0.01), endocarditis (p < 0.01), ventricular fibrillation (p = 0.01), anemia (p < 0.01), tobacco use (p < 0.01), previous MI (p < 0.01), previous CABG (p < 0.01) (Table 5). On the other hand, patients from lower-income neighborhoods were less likely to have lymphoma (p = 0.04), uncomplicated hypertension (p < 0.01), peripheral vascular disease (p < 0.01), other thyroid disorders (p < 0.01), coronary artery disease (p < 0.01), atrial fibrillation (p < 0.01), atrial flutter (p < 0.01), first-degree (p < 0.01) or complete atrioventricular block (p < 0.01), hyperlipidemia (p < 0.01), thrombocytopenia (p < 0.01), or sleep apnea (p < 0.01) (Table 5).

In isolated TAVR, compared to patients living in neighborhoods of income at the highest quartiles, those living in lower-income neighborhoods had higher risks of mechanical ventilation (aOR 1.20, 95 CI 1.02–1.42, p = 0.03) but lower risks of superficial wound complication (aOR 0.74, 95 CI 0.59–0.92, p = 0.01) (Table 6 and Fig. 1). Patients from low-income neighborhoods had lower total hospital charges (207,314 ± 1106 vs 227,423 ± 1188 dollars, p < 0.01) (Table 6).

Table 6 In-hospital outcomes comparing patients living in neighborhoods of income at the lowest and highest quartiles who underwent isolated TAVR.
Full size table
Figure 1
figure 1

In-hospital outcomes comparing patients living in neighborhoods of income at the lowest and highest quartiles who underwent isolated TAVR. The aOR and 95% CI are shown. *p-value < 0.05. AKI acute kidney injury, aOR adjusted odds ratio, CI confidence interval, MACE major adverse cardiovascular event, MI myocardial infarction, PE pulmonary embolism, Ref reference, SES socioeconomic status, TAVR transcatheter aortic valve replacement, TIA transient ischemic attack, VTE venous thromboembolism.

Full size image

In isolated SAVR, compared to patients living in neighborhoods of income at the highest quartiles, those living in lower-income neighborhoods had higher risks of mortality (aOR 1.44, 95% CI 1.22–1.71, p < 0.01), respiratory complications (aOR 1.13, 95% CI 1.03–1.23, p = 0.01), mechanical ventilation (aOR 1.23, 95% CI 1.11–1.35, p < 0.01), AKI (aOR 1.14, 95% CI 1.06–1.23, p < 0.01), and transfer out to other facilities (aOR 1.19, 95% CI 1.09–1.29, p < 0.01) (Table 7 and Fig. 2). On the other hand, patients living in lower-income neighborhoods had lower risks of pericardial complications (aOR 0.80, 95% CI 0.68–0.94, p = 0.01), and hemorrhage/hematoma (aOR 0.85, 95% CI 0.80–0.90, p < 0.01) (Table 7 and Fig. 2). Patients from low-income neighborhoods had longer time from admission to operation (2.22 ± 0.05 vs 1.42 ± 0.05 days, p = 0.01), longer LOS (11.37 ± 0.12 vs 9.39 ± 0.09, p < 0.01), but lower total hospital charge (275,909 ± 2731 vs 276,058 ± 2942 dollars, p < 0.01) (Table 7).

Table 7 In-hospital outcomes comparing patients living in neighborhoods of income at the lowest and highest quartiles who underwent isolated SAVR.
Full size table
Figure 2
figure 2

In-hospital outcomes comparing patients living in neighborhoods of income at the lowest and highest quartiles who underwent isolated SAVR. The aOR and 95 CI are shown. *p-value < 0.05. AKI acute kidney injury, aOR adjusted odds ratio, CI confidence interval, MACE major adverse cardiovascular event, MI myocardial infarction, PE pulmonary embolism, Ref reference, SES socioeconomic status, SAVR surgical aortic valve replacement, TIA transient ischemic attack, VTE venous thromboembolism.

Full size image

Discussion

This study conducted a population-based analysis to examine the impact of SES on post-procedural outcomes following AVR. Patients with AS from lower-income communities had less access to both TAVR and SAVR, with TAVR accessibility being particularly limited. In isolated TAVR, outcomes were largely similar between patients from both lower- and higher-income neighborhoods. However, in isolated SAVR, patients from lower-income communities had a higher risk of in-hospital mortality, as well as increased pulmonary and renal complications; additionally, these patients experienced longer waits from admission to operation and had extended hospital stays.

Although the NIS categorizes patients' SES based on their neighborhood's household income rather than the income of the patients, evidence from the population differences between the low- and high-income patients in our study validates this stratification. For example, patients from lower-income neighborhoods were more likely to be under Medicaid, a program for those with low income. Additionally, racial minorities such as African Americans, Hispanics, and Native Americans, as well as comorbidities including obesity, diabetes, uncontrolled hypertension, peripheral vascular disease, chronic pulmonary disease, severe renal failure, tobacco use, and substance abuse are all strongly associated with low SES15,16,17,18,19,20,21,22,23. These factors were more prevalent in the low-income cohort of our study, which supports the NIS's approach to stratifying SES.

The prevalence of AS rises exponentially with age1. Despite being at a younger age, patients from lower-income neighborhoods were still prominently represented in AS diagnoses when compared to those from higher-income areas in both TAVR and SAVR. This discrepancy might be attributed to the higher prevalence of AS risk factors in lower-income communities. Established risk factors for AS, such as obesity, diabetes, uncontrolled hypertension, and chronic kidney disease24, were more common among patients from lower-income neighborhoods, which may lead to their earlier onset of AS.

The rates of patients with AS undergoing either TAVR or SAVR declined with decreasing income of the community, even when accounting for pre-procedural differences including age and comorbidities. Notably, the TAVR to SAVR ratio also declined with reduced income in the neighborhood. This indicates that patients with AS from lower-income communities face barriers to accessing both TAVR and SAVR, with TAVR being especially less accessible. This aligns with prior studies that highlighted lower SES patients encounter challenges in accessing TAVR10,11,12. Such disparities may stem from the high costs associated with TAVR valves combined with inadequate Medicare reimbursements for TAVR, discouraging some hospitals from offering the procedure, which disproportionately impacts socioeconomically disadvantaged patients12. Consequently, during TAVR's initial growth phase in the US, hospitals catering to more affluent patients were quicker to adopt the procedure, which resulted in unequal distribution of TAVR services10.

In isolated TAVR, post-procedural outcomes were largely comparable between patients from both lower- and higher-income neighborhoods, aligning with findings from Rogers et al.25. This suggests that when provided with access, patients from lower-income communities can benefit from the novel, lifesaving TAVR as much as their counterparts in wealthier communities. Therefore, efforts should be concentrated on ensuring equitable access to TAVR for patients in economically disadvantaged communities.

In isolated SAVR, patients from lower-income neighborhoods experienced a heightened risk of in-hospital mortality and elevated pulmonary and renal complications. This contrasts with the findings by Rogers et al., who reported no significant disparities in SAVR outcomes25. The discrepancy might arise from the fact that Rogers et al. sourced their data solely from Florida and Washington state registries, which may not be wholly representative of the broader US population25. In contrast, our study utilized a comprehensive national registry. The worse SAVR outcomes in lower-income communities could be attributed to delayed diagnosis and subsequent treatment. This is underscored by the extended wait time from admission to operation in these patients, averaging 2.22 days, compared to just 1.42 days in patients from wealthier neighborhoods especially considering about one-third of SAVR cases were emergent. This delay remained significant even after accounting for pre-operative differences, such as higher comorbidity burden in patients from lower-income communities, which often necessitates additional stabilization time prior to surgery. Additionally, patients from lower-income communities were more likely to get transferred in from other facilities, leading to potentially more delays in the transfer process prior to admission. These delays may arise from limited healthcare resources in lower-income communities. Efforts should be directed towards ensuring equitable access to specialized care in these areas. Another potential factor could be the restricted access to TAVR in the low SES communities. Younger patients with AS are often directed towards SAVR, primarily because of the possibility of implanting a mechanical valve, which can reduce the need for future interventions. In contrast, older patients were typically recommended for TAVR, given the potential risks associated with open surgery and diminished necessity for long-term re-interventions. Among patients who are aged 85 and above, patients from a low SES background were less likely to be treated with TAVR (25.93% vs 33.91%, p < 0.01) compared to those of high SES. As a result, patients who might have benefited more from TAVR were disproportionately directed, leading to poorer outcomes. In light of this finding, broadening TAVR accessibility in economically disadvantaged areas becomes imperative.

This study comes with several limitations. First, while the NIS database's income stratification finds support in our findings, using neighborhood income as a proxy for individual income introduces inherent variability. However, individual income data of the patients is rarely gathered during medical visits. This makes recording exact patient income, especially in expansive national registries, a challenge, rendering neighborhood income a practical alternative. Second, the NIS database is primarily structured based on diagnostic codes. This means it does not capture granular details such as medication, and laboratory results, or allows the calculation of surgical risk scores like the STS (Society of Thoracic Surgeons) or EuroSCORE (European System for Cardiac Operative Risk Evaluation) scores. Additionally, the database only has records of the hospital stay, lacking post-discharge follow-up data. This limitation restricts our ability to evaluate long-term prognoses post-TAVR and SAVR, where previous studies indicated disadvantaged SES may negatively influence long-term outcomes in AVR12. Nevertheless, NIS is the most extensive in-patient database in the US that records 20% of all discharges. This breadth provides one of the most representative samples of the US patient demographic, enabling a thorough population-level examination of the impact of SES on AVR outcomes.

In this population-based analysis, we explored the influence of SES on short-term outcomes post-TAVR and SAVR. It was found that lower-income communities had reduced access to both TAVR and SAVR, with the disparity in accessibility being more pronounced for TAVR. When given access to isolated TAVR, patients from lower-income neighborhoods had mostly comparable outcomes as those from higher-income communities. However, for isolated SAVR, patients from low-income communities faced a higher risk of in-hospital mortality and increased pulmonary and renal complications, most likely due to delays in treatment. Addressing these disparities is crucial, with a particular focus on ensuring equitable specialized healthcare resources including expanding TAVR access in economically disadvantaged communities.