The data that support the findings of this study are available from the corresponding author upon reasonable request. The data for this study were obtained from the NIS database, spanning the years 2006 to 2018. The NIS is a comprehensive database that contains inpatient health diagnoses and procedures from a diverse set of sources, including ≈7 million hospital stays each year and representing around 20% of all hospitalizations in the United States. It is a publicly available database developed by the Agency for Healthcare Research and Quality for the Healthcare Cost and Utilization Project. The NIS collects all‐payer data from participating states, covering over 97% of the US population. Primary and secondary diagnoses, as well as procedural diagnoses, are coded using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD‐9‐CM) and International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Clinical Modification (ICD‐10‐CM) coding systems.11The study population consisted of 312 085 hospitalization of patients aged ≥18 years with a diagnosis of TC between 2006 and 2018. TC was identified using the ICD‐9‐CM code 429.83 and the ICD‐10‐CM code I5181 (refer to Table S1 for details). To ensure accurate identification of patients with TC, encounters with a primary admitting diagnosis of ACS were excluded using the corresponding ICD‐9‐CM and ICD‐10‐CM codes (refer to Table S2 for details). The primary focus was on 2 racial groups: self‐identified White and Black patients. Patients with missing race information, those belonging to small sample size racial or ethnic groups (such as American Indian, Asian, or Pacific Islander), or individuals identified as Hispanic ethnicity (as this diverse group is considered a single race category in the NIS database) were excluded from the analysis. A total of 206 617 TC hospitalizations from the White and Black racial groups were included in the analysis (Figure 1). Our primary outcomes of interest included death, while secondary outcomes included cardiac arrest, cardiogenic shock, and length of stay. The relevant ICD‐9‐CM and ICD‐10‐CM codes for these outcomes can be found in Table S1. To analyze the outcomes, we adjusted for various factors, including age; sex; Elixhauser comorbidity scores; and socioeconomic factors such as household income, primary insurance payer, and hospital teaching status. Elixhauser comorbidity variables were derived from the ICD‐9‐CM and ICD‐10‐CM codes provided by the Healthcare Cost and Utilization Project's NIS.12Flowchart of study population.Appropriate stratifying, clustering, and weighting of samples using standard NIS methods were applied for all analyses. Continuous variables were reported as median and interquartile range, while categorical variables were presented as frequencies and percentages. Group comparisons were performed using Wilcoxon ranked‐sum tests for continuous variables and χ2 tests for categorical variables. Univariate and multivariable logistic regression models were used to evaluate the racial disparities in the risk of death, cardiogenic shock, and cardiac arrest. The multivariable models included several predictors, such as race, age, Elixhauser comorbidity index, sex, hospital location, expected primary payer, and median household income. To investigate the racial difference in length of stay, a Poisson regression model was employed. In addition, Bayesian model averaging was applied to determine the most important predictors in each race. Five candidate models were defined, each with a different set of predictor variables, and equal prior probabilities were assigned to all models. Bayesian methods were used to fit each model to the data and compute posterior probabilities. A single predictive distribution was obtained by averaging over the posterior distributions of all models, with the final model being selected on the basis of the lowest Bayesian information criterion and highest posterior probability. Bayesian model averaging accounted for model uncertainty and provided more reliable predictions while following a statistically rigorous approach to model selection. Odds ratios and risk ratios with corresponding 95% CIs were used to express the outcomes and adjusted accordingly for length of stay. The analysis was conducted using standard NIS methods to obtain weighted national estimates. Data analysis was performed using SAS version 9.4 software, and a 2‐tailed P value of <0.05 was considered statistically significant. The plots were generated using R, specifically using the ggplot2 and PupillometryR package (R Foundation for Statistical Computing, Vienna, Austria).13 As the NIS contained publicly available deidentified patient data, institutional review board approval and informed consent were not required for this study. However, the main authors who accessed the data directly had signed the Healthcare Cost and Utilization Project data use agreement.
RESULTS
A total of 206 617 hospitalizations with TC from the NIS 2006 to 2018 met the inclusion criteria. Among them, 188 608 identified as White, while the 18 009 hospitalizations were part of the Black cohort. The differences in baseline demographics and clinical characteristics between the 2 racial groups are summarized in Table 1.Demographics and Baseline Clinical CharacteristicsQR indicates interquartile range.Pearson's χ2 test; Wilcoxon rank‐sum test; Fisher's exact test.Compared with White patients, Black patients tended to be younger, with a higher proportion of men and lower median household income. Furthermore, significant differences were observed in several clinical characteristics between the 2 racial groups. Among Black patients, a higher prevalence of comorbidity conditions, as reflected by a higher Elixhauser comorbidity score, hypertension, diabetes, obesity, congestive heart failure, pulmonary circulation disease, renal failure, paralysis, metastatic cancer, coagulopathy, fluid and electrolyte disorder, deficiency anemias, alcohol abuse, drug abuse, and liver disease were observed. While peripheral vascular disease, chronic pulmonary disease, and depression were more prevalent among White counterparts (all P <0.001; Table 1).Compared with White patients, Black patients experienced poorer outcomes, including a higher percentage of deaths (7% versus 6%), a higher likelihood of cardiac arrest (2% versus 1%), and longer hospital stays (median length of stay of 4.7 versus 3.8 days) (all P<0.001; Table 1). In univariate models, compared with White patients, Black patients had a higher likelihood of death (odds ratio, 1.28 [95% CI, 1.11–1.47]) and cardiac arrest (odds ratio, 1.45 [95% CI, 1.15–1.82]; Table 2). In multivariable models, after adjusting potential confounders, higher mortality rate and risk of cardiac arrest were observed among Black patients, although the difference did not reach statistical significance. In terms of the length of stay, Black patients had a significantly longer hospital stay in both the univariate model (risk ratio, 1.26 [95% CI, 1.22–1.31]) and the multivariable model (risk ratio, 1.06 [95% CI, 1.04–1.07]; Table 2, Figure 2).Predictors of the Outcome at Univariate and Multivariate AnalysisOR indicates odds ratio; and RR, risk ratio.*Multivariable logistic regression models adjusted for age, Elixhauser comorbidity index, sex, hospital location, expected primary payer, and median household income. +Poisson regression model adjusted for age, Elixhauser comorbidity index, sex, hospital location, expected primary payer, and median household income.In the Black cohort, Bayesian model averaging model revealed that sex, congestive heart failure, pulmonary circulation disorder, weight loss, and drug abuse were the most significant comorbidities associated with death (posterior probability=48%) (Table 3). In the White cohort, the posterior probability of mortality predictors was 46%. The significant predictors included age, sex, hospital type, and the following 12 comorbidities: renal failure, congestive heart failure, pulmonary circulation disorder, peripheral vascular disease, neurologic disorders, hypothyroidism, liver diseases, lymphoma, metastatic cancer, psychiatric disorders, depression, and hypertension (Table 4). Of particular importance, sex, congestive heart failure, and pulmonary circulation disease were consistently observed as mortality predictors in both racial groups.Predictors of Death in the Multivariate Logistic Model in Black Patients With TCOR indicates odds ratio; and TC, Takotsubo cardiomyopathy.Predictors of Death in the Multivariate Logistic Model in White Patients With TCOR indicates odds ratio; and TC, Takotsubo cardiomyopathy.
DISCUSSION
The objective of this study was to compare the clinical characteristics and in‐hospital outcomes of White and Black patients with TC. The study yielded several key findings: (1) notable differences in clinical characteristics were observed between the 2 racial groups; (2) Black patients exhibited a higher incidence of cardiac arrest and a longer length of hospital stay; and (3) differences in baseline characteristics between the 2 racial groups led to the development of a distinct predictive model for all‐cause death.Our study observed significant disparities in demographics and comorbidity burden between Black and White patients. Specifically, Black patients were found to be younger and more likely to be men and exhibited a higher prevalence of comorbidities such as hypertension, diabetes, and chronic kidney disease compared with their White counterparts. These findings align with previous research conducted on US‐wide analyses of TC hospitalizations using NIS 2006 to 2014 data14 or conducted on heart failure populations.15 The ARIC (Atherosclerosis Risk in Communities) study on heart failure with preserved ejection fraction revealed that Black patients had a younger age, higher rates of hypertension, diabetes, higher body mass index, and renal failure compared with White patients.15 These different comorbidity profiles play a crucial role in the development of distinct predictive models for death among different racial groups, as evidenced by the results presented in Tables 3 and 4. Importantly, a novel finding from this study, our analysis demonstrated that sex, congestive heart failure, and pulmonary circulation disease emerged as consistent predictors of death in both Black and White cohorts, underscoring their substantial impact on mortality outcomes irrespective of race.Black patients had a higher incidence of cardiac arrest and longer hospital stays than their White counterparts. These findings are in line with previous studies indicating that Black patients are at a greater risk of sudden cardiac death (SCD) than White patients.16 Zhao et al conducted a large community‐based cohort study and reported that Black subjects were significantly more likely to experience SCD than White subjects, with socioeconomic status and traditional risk factors accounting for around 65% of the racial disparity in SCD.16 Similarly, the The Losartan Intervention For Endpoint Reduction (LIFE) study followed 533 Black and 8660 non‐Black patients with hypertension for 4.8 years and found that Black patients had almost twice the risk of SCD than non‐Black patients, even after adjusting for other risk factors.17 Given the baseline clinical and socioeconomic differences observed in our study, it is likely that similar dynamics account for the observed differences in outcomes.Black patients experienced longer hospital stays than their White counterparts, consistent with previous research conducted in various populations. Zheng et al found that differences in hospital practices accounted for the length‐of‐stay disparities across race and ethnicity, with Black patients having longer stays than White patients.18 Similarly, in our study, significant differences in insurance types including Medicare, Medicaid, and private insurance were observed, which likely impacted discharge disposition and contributed to the longer hospitalization durations among Black patients. Furthermore, higher rates of SCD observed in Black patients is a known factor in longer hospital stays.These worse outcomes among Black patients may be explained partially by higher stressors from chronic medical conditions and socioeconomic status. In addition, higher prevalences of alcohol abuse and drug abuse found in Black patients compared with White patients could also result in worse TC outcomes. Substance abuse disorder, as well as alcohol withdrawals were reported to induce TC due to catecholamine surges.19, 20, 21 Further studies need to investigate the interaction effects of race and those confounders on TC clinical outcomes.Several limitations of our study exist. First, as it was conducted retrospectively, unmeasured confounders may have potentially affected our results. Second, since the NIS database records hospitalization encounters, rather than clinical course of individual patients, some recorded hospitalizations may have been recurrences in the same subject rather than separate patients. However, this likely did not affect our results since the recurrence rate of TC was reported to be <5%.22 In addition, we excluded patients who were diagnosed with TC at the same time as or after ACS due to concerns about diagnostic bias. Primary Takotsubo syndrome primarily affects women and typically arises following emotional stressors. It often presents without significant coronary artery disease, characterized by minor troponin release, a transient reduction in left ventricular function, and generally has a favorable prognosis. In contrast, secondary TC affects both men and women and is usually precipitated by physical stressors. It is commonly accompanied by coronary artery disease, significant troponin release, and more severe left ventricular dysfunction and is associated with a poorer prognosis.23, 24 Future prospective studies should investigate potential outcome disparities between Black and White populations in both primary and secondary TC. Finally, the long‐term outcomes have not been reported in our study.
CONCLUSIONS
There are notable disparities in baseline demographic and clinical characteristics between White and Black patients presenting with TC as well as discernible differences in clinical outcomes. Hence, it is crucial to emphasize the significance of individualized care that is tailored to address the specific needs of each racial group.
