Sociodemographic and Hospital-Level Disparities in Cardiovascular Hospitalizations Among Adults With Systemic Lupus Erythematosus: A Retrospective Cohort Study

Survival in systemic lupus erythematosus (SLE) has improved, yet cardiovascular morbidity and mortality remain disproportionately high [1-3]. Chronic inflammation, endothelial dysfunction, and traditional risk factors synergistically accelerate atherosclerosis in SLE, with systemic inflammation independently predicting adverse cardiovascular events [4]. Although hospitalization rates for cardiovascular complications are rising, the drivers of these admissions and the role of social determinants of health are incompletely understood [5]. We therefore examined sociodemographic and hospital-level predictors of cardiovascular-related hospitalizations in a contemporary, national SLE cohort.

This retrospective cohort study utilized hospitalization records from the Healthcare Cost and Utilization Project National Inpatient Sample (HCUP-NIS) from January 1, 2016, to December 31, 2021 [6]. This database is the largest publicly available all-payer inpatient database in the United States, which, when weighted, provides national estimates representing more than 35 million hospitalizations annually. As publicly available, de-identified data were used, informed consent or local institutional review board approval was waived. Hospitalizations were identified for adults aged ≥ 18 years with a diagnosis of SLE using the International Classification of Diseases, Tenth Revision (ICD-10) codes—M32.1x, M32.8, and M32.9. Cardiovascular hospitalization was determined by the principal diagnosis (ICD-10-CM codes I00–I99) recorded for each admission. The primary outcome of interest was cardiovascular-related hospitalization, categorized into specific events, including heart failure (HF), coronary artery disease (CAD), atrial fibrillation (AF), and stroke. National-level estimates were generated from discharge weights assigned to each hospitalization. Descriptive statistics were used to characterize the cohort, with categorical variables reported as counts and percentages. Multivariable modified Poisson regression with robust variance estimates was used to assess associations between sociodemographic and hospital-level factors and cardiovascular hospitalizations. We constructed a model that adjusted for sociodemographic factors and comorbid conditions. Generalized estimating equations (GEEs) with a compound symmetry correlation matrix were used to account for clustering at the hospital level. Effect sizes were expressed as relative risk (RRs) and their corresponding 95% confidence intervals (CIs). Statistical significance was set at a two-sided p < 0.05.

A total of 165 290 cardiovascular hospitalizations from 2016 to 2021 for patients with SLE were identified (Table 1). Most patients were women (84.8%) and either 45–64 years (38.4%) or ≥ 65 years (37.2%). Non-Hispanic White (NHW) patients accounted for 48.2% of admissions, non-Hispanic Black (NHB) 34.7%, and Hispanics 12.0%. Over one-third (36.5%) came from the lowest income quartile, and 58.2% were Medicare-insured. Admissions clustered in large (55.3%), urban teaching hospitals (77.4%), predominantly in the Southern region of the US (44.8%). Hypertension (82.1%) and chronic kidney disease (43.3%) were the leading comorbidities.

HF was the leading cause (24.8%), followed by CAD (14.2%), hypertension (10.5%), stroke (10.2%), and AF (6.7%) (Figure 1). Across strata, cardiovascular hospitalization rates rose steeply with age; hypertension-related admissions, however, remained relatively constant (≈16–18 per 1000 all-cause SLE admissions). NHB patients had more hypertension admissions than NHW peers (23.6 vs. 19.0 per 1000). Most events occurred in large, urban teaching hospitals in the South. Low-income individuals (quartile 1) experienced the greatest absolute burden of cardiovascular admissions (Figures S1–S5).

Compared with 18–44 years, patients ≥ 65 years faced markedly higher risks of HF (RR 2.46, 95% CI 2.32–2.60), CAD (RR 2.18, 1.99–2.39), AF (RR 3.34, 2.99–3.74), and stroke (RR 2.27, 2.01–2.56) (Table S1). Hypertension admissions were more common in younger patients (RR 0.47, 0.41–0.52 for ≥ 65 years vs. 18–44 years). Female sex conferred lower risks for CAD (RR 0.87, 0.81–0.93) and stroke (RR 0.88, 0.80–0.98) but did not significantly affect AF (RR 1.11, 0.94–1.32). Furthermore, NHB patients were more likely to be hospitalized for hypertension (RR 1.47, 1.34–1.61) and stroke (RR 1.17, 1.08–1.27) but less likely for CAD (RR 0.88, 0.82–0.94) than NHW patients. Hispanics had higher hypertension risk (RR 1.29, 1.16–1.45) yet lower stroke risk (RR 0.88, 0.78–0.95). Compared with the lowest income quartile, the highest quartile had lower risks of any cardiovascular admission (RR 0.95, 0.92–0.98), hypertension (RR 0.91, 0.84–0.99), and CAD (RR 0.88, 0.78–0.98). Lastly, large hospitals and urban teaching centers were independently associated with more overall cardiovascular and stroke admissions.

In a study of 165 290 cardiovascular hospitalizations among SLE patients, HF emerged as the leading cause of hospitalizations, followed by CAD and hypertension. Older age was significantly associated with increased risks of HF, CAD, AF, and stroke hospitalizations, while younger patients were more likely to be hospitalized for hypertension. Females exhibited lower risks for CAD and stroke but slightly higher AF hospitalization rates compared to males. NHB and Hispanic patients faced higher risks for hypertension hospitalizations but lower risks for CAD compared to NHW.

Sex differences were notable, with female sex being linked to a reduced risk of cardiovascular hospitalizations after multivariable adjustment. Supporting this, Mihailovic et al. and Pons-Estel et al. identified male sex as a significant predictor of CAD and myocardial infarction (MI) [7, 8]. Similarly, Urowitz et al. reported that male sex was independently associated with atherosclerotic events [9]. Racial disparities in cardiovascular outcomes were prominent. Joyce et al. found higher cardiovascular event prevalence among younger males, Hispanics, and NHB individuals [10]. Garg et al. reported a sevenfold increase in incident cardiovascular diseases among Black individuals compared to non-Blacks in the Georgia Lupus Registry [11]. Barbhaiya et al. noted elevated cardiovascular disease risks among Blacks, with Hispanics and Asians showing lower MI risks but higher stroke risks, particularly hemorrhagic stroke [12, 13]. In contrast, our study found lower total stroke hospitalization risks among Hispanics, possibly due to the inclusion of varied stroke types. Socioeconomic factors also played a role. Bolla et al. observed lower cardiovascular risk factor prevalence in middle-income countries but poorer risk factor control compared to high-income countries [14]. Maynard et al. linked low income to increased MI and stroke risks in White SLE patients, though not in African Americans [15]. Our findings confirmed higher cardiovascular hospitalization rates in lower-income groups, persisting after demographic adjustments. The limitations of our study include the absence of data on SLE disease duration, disease activity indices, specific medication use (including glucocorticoids, immunosuppressants, and biologics), and relevant laboratory markers such as complement levels or inflammatory biomarkers. These factors are known to influence cardiovascular risk and hospitalization patterns in SLE, and their omission may lead to residual confounding. Consequently, our findings should be interpreted as descriptive of population-level patterns rather than definitive causal associations, emphasizing the need for future studies with clinically granular datasets to validate and extend these observations.

In summary, HF drives cardiovascular hospitalizations in SLE, with Black race, older age, and lower socioeconomic status as key risk factors. These findings highlight the need for targeted interventions to address differences and improve cardiovascular outcomes in SLE patients.

Song Peng Ang: conception, methodology; Song Peng Ang and Jia Ee Chia: data acquisition, formal analysis, Investigation; Song Peng Ang, Jia Ee Chia, and Kanan Jahangirli: writing – original draft; Jose Iglesias and Debabrata Mukherjee: writing – review and editing; Debabrata Mukherjee: supervision.

The authors have nothing to report.

This study involved the utilization of publicly available databases with de-identified patient data. Hence, ethical approval was not required.

The authors declare no conflicts of interest.