Long COVID refers to persistent sequelae following SARS-CoV-2 infection, and sex-specific cardiovascular outcomes among hospitalized patients remain incompletely characterized. We queried the 2022 National Inpatient Sample (NIS) to identify adult hospitalizations with a diagnosis of long COVID (ICD-10 U09.9), excluding patients younger than 18 years or with missing outcome data. Analyses incorporated NIS discharge weights to generate national estimates. Multivariable (survey-weighted) logistic regression was used to estimate adjusted odds ratios (aORs) for cardiovascular diagnoses and in-hospital mortality. Among 87,415 weighted hospitalizations for long COVID, 49.4% were male, and 50.6% were female. Males had more complicated hypertension, coagulopathy, and alcohol use disorder, whereas females had higher rates of obesity, depression, and hypothyroidism (all p <0.05). Males had a higher in-hospital mortality rate (5.9% vs 4.7%, p <0.001). In adjusted analyses, females had lower odds of in-hospital mortality (aOR: 0.874 [95% CI 0.819–0.932]), cardiac arrhythmias (aOR: 0.652 [0.629–0.677]), venous thromboembolism (aOR: 0.846 [0.807–0.887]), and myocardial infarction (aOR: 0.767 [0.720–0.817]). Adjusted odds of ischemic cerebrovascular accident were not significantly different (aOR: 0.955 [0.795–1.146]). Females had higher odds of transient ischemic attack (aOR: 1.441 [1.067–1.945]). Median length of stay (5 vs 4 days) and total hospital charges ($50,447 vs $43,839) were lower in females (all p <0.001). In conclusion, in this nationally representative analysis of long COVID hospitalizations, sex-based differences were observed in cardiovascular diagnoses, mortality, and healthcare utilization, and these findings support sex-sensitive risk stratification and hypothesis generation for post-COVID care.
Graphical Abstract
Long COVID, broadly defined as persistent symptoms following SARS-CoV-2 infection, is increasingly recognized as a driver of cardiovascular morbidity, with manifestations including cardiac arrhythmias, thromboembolic events, and myocardial injury. While multiple studies have characterized the spectrum of postacute sequelae, , the influence of biological sex on the cardiovascular profile of long COVID hospitalizations remains incompletely characterized. Existing literature suggests that males and females may experience different trajectories of acute COVID-19 illness, with reports indicating that male patients often exhibit more severe acute disease and higher mortality rates, whereas female patients more frequently endure persistent, postacute sequelae. , However, whether these sex-based disparities extend to the context of long COVID hospitalizations, particularly in relation to cardiovascular complications and healthcare utilization, remains inadequately defined. To address this gap, we utilized the 2022 National Inpatient Sample (NIS), a nationally representative administrative dataset, to examine sex-based differences in cardiovascular diagnoses, in-hospital mortality, and resource utilization among adult patients hospitalized with long COVID. We aimed to describe demographic profiles, comorbidity burdens, and hospital-level outcomes stratified by sex, providing a foundation for hypothesis generation and future prospective investigation.
Methods
Data source and ethics statement
Hospitalization information was extracted from the NIS database, a component of the Healthcare Cost and Utilization Project (HCUP) managed by the Agency for Healthcare Research and Quality. The NIS stands as the most extensive publicly accessible, fully anonymized, all-payer inpatient database in the United States. It is based on billing records submitted by hospitals to state-level agencies nationwide and includes hospitalization-level data from a broad, representative sample of U.S. hospitals, but does not contain patient-level identifiers for tracking across multiple admissions. The database encompasses both patient-specific and hospital-specific data from approximately 4,000 hospitals, accounting for around 20% of all hospital admissions in the U.S., which translates to over 7 million individual hospitalizations annually. By applying the discharge-level sampling weights provided by HCUP, the database scales up to represent over 35 million inpatient stays annually, allowing for weighted national estimates of inpatient hospitalization trends. For every patient, the database records up to 40 discharge diagnoses and 25 procedure codes using the International Classification of Diseases, Tenth Revision (ICD-10) coding system.
The NIS database contains deidentified patient information and is publicly available for research purposes. As a result, this study was exempt from Institutional Review Board approval and informed consent requirements. The research was conducted in accordance with ethical standards in the 2013 Declaration of Helsinki. We adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines in reporting our findings. The corresponding STROBE checklist is available in Supplementary Table 2 .
Study population and patient selection
We identified all inpatient admissions in the NIS for the year 2022 with a documented diagnosis of long COVID using the ICD-10 code U09.9 (Post-COVID-19 condition, unspecified). After excluding patients younger than 18 years and those with missing outcome data, NIS sampling weights were applied to the remaining cohort. The resulting weighted cohort was then stratified by biological sex (male vs female) for comparative analyses ( Figure 1 ). Given that ICD-10 code U09.9 was newly introduced in October 2021, we recognized the potential for considerable variability in early coding practices across hospitals and states. Institutional differences in coding infrastructure, provider familiarity with the new code, and the absence of a standardized clinical definition for long COVID during this period may have contributed to heterogeneous case identification. Hospitals with higher volumes of post-COVID follow-up care may have adopted the code earlier and more consistently, potentially introducing selection bias. This concern was directly addressed through a temporal sensitivity analysis restricting the cohort to admissions from May through December 2022, when coding practices were expected to be more stable, as detailed below.
Flowchart of patient selection. Flowchart illustrating how 18,139 patients with long COVID were identified from 6,578,372 NIS admissions in 2022. After excluding those younger than 18 and missing outcome data, 17,483 remained. Weighting yielded 87,415 hospitalizations (43,155 males and 44,260 females) for final analysis. Abbreviation: NIS = National Inpatient Sample.
Data collection
Baseline demographic (e.g., age, sex, primary payer, race/ethnicity, median household income) and hospital-level characteristics (e.g., census division, location/teaching status, bed size) were extracted. Additionally, we utilized Elixhauser comorbidities which are shown in previous studies to affect outcomes in our study population ( Table 1 ). , Outcomes included resource utilization (hospital length of stay (LOS), hospital charges) and patient disposition (e.g., transfer to skilled nursing facility, home health care). ICD-10-coded cardiovascular events included in-hospital mortality; acute cardiovascular events: myocardial infarction (MI), heart failure (HF), cardiogenic shock, and cardiac arrest; cerebrovascular complications: ischemic cerebrovascular accident (CVA) and transient ischemic attack (TIA); cardiac arrhythmias: ventricular tachycardia/fibrillation and atrial fibrillation; venous thromboembolism: pulmonary embolism (PE) and deep vein thrombosis (DVT). The ICD-10 diagnosis codes used to define exclusion criteria, comorbidities, and outcomes are provided in Supplementary Table 1 .
Table 1
Baseline demographics and comorbidities
| Gender | |||
|---|---|---|---|
| Male ( n = 43,155) | Female ( n = 44,260) | p value | |
| Demographics characteristics | |||
| Age (years) | 67 (56–77) | 66 (52–77) | <0.001 |
| Race | <0.001 | ||
| White | 30,450 (72.7%) | 29,960 (69.4%) | |
| Black | 4,340 (10.4%) | 6,510 (15.1%) | |
| Hispanic | 4,990 (11.9%) | 4,375 (10.1%) | |
| Asian or Pacific Islander | 910 (2.2%) | 860 (2.0%) | |
| Native American | 255 (0.6%) | 490 (1.1%) | |
| Other | 930 (2.2%) | 945 (2.2%) | |
| Primary expected payer | <0.001 | ||
| Medicare | 23,245 (53.9%) | 24,865 (56.2%) | |
| Medicaid | 4,950 (11.5%) | 6,220 (14.1%) | |
| Private insurance | 11,785 (27.3%) | 11,520 (26.1%) | |
| Self-pay | 1,250 (2.9%) | 930 (2.1%) | |
| No charge | 70 (0.2%) | 15 (0.0%) | |
| Other | 1,795 (4.2%) | 670 (1.5%) | |
| Median household income (ZIP Code-demographic data) | <0.001 | ||
| 1–55,999 | 12,020 (28.2%) | 13,660 (31.2%) | |
| 56,000–70,999 | 11,610 (27.3%) | 11,600 (26.5%) | |
| 71,000–93,999 | 10,620 (25.0%) | 10,425 (23.8%) | |
| 94,000+ | 8,315 (19.5%) | 8,060 (18.4%) | |
| Patient location | <0.001 | ||
| Large central metro | 10,260 (23.8%) | 10,815 (24.5%) | |
| Large fringe metro | 10,755 (25.0%) | 10,955 (24.8%) | |
| Medium metro | 8,910 (20.7%) | 9,130 (20.7%) | |
| Small metro | 4,555 (10.6%) | 4,680 (10.6%) | |
| Micropolitan | 4,880 (11.3%) | 4,860 (11.0%) | |
| Non-core | 3,675 (8.5%) | 3,745 (8.5%) | |
| Clinical characteristics | |||
| Infectious/Immune | |||
| AIDS | 335 (0.8%) | 170 (0.4%) | <0.001 |
| Autoimmune disease | 1,780 (4.1%) | 4,235 (9.6%) | <0.001 |
| Anemias | |||
| Deficiency Anemia | 11,605 (26.9%) | 12,710 (28.7%) | <0.001 |
| Blood Loss (Anemia) | 360 (0.8%) | 365 (0.8%) | 0.877 |
| Cancers | |||
| Leukemia | 990 (2.3%) | 655 (1.5%) | <0.001 |
| Lymphoma | 1,090 (2.5%) | 835 (1.9%) | <0.001 |
| Metastatic cancer | 890 (2.1%) | 775 (1.8%) | <0.001 |
| Solid cancer | 1,935 (4.5%) | 1,540 (3.5%) | <0.001 |
| Cardiovascular/Hematologic | |||
| Coagulopathy | 5,485 (12.7%) | 3,960 (8.9%) | <0.001 |
| Hypertension, Complicated | 16,520 (38.3%) | 14,815 (33.5%) | <0.001 |
| Hypertension, Uncomplicated | 14,730 (34.1%) | 15,730 (35.5%) | <0.001 |
| Valvular disease | 4,005 (9.3%) | 3,835 (8.7%) | 0.001 |
| Peripheral vascular Disease | 3,410 (7.9%) | 2,605 (5.9%) | <0.001 |
| Neurological | |||
| Cerebrovascular disease | 1,580 (3.7%) | 1,260 (2.8%) | <0.001 |
| Sequelae of cerebrovascular disease | 1,270 (2.9%) | 1,040 (2.3%) | <0.001 |
| Neuro. movement disorder | 1,465 (3.4%) | 1,590 (3.6%) | 0.112 |
| Neuro. Other | 5,275 (12.2%) | 5,240 (11.8%) | 0.081 |
| Neuro. Seizure | 1,555 (3.6%) | 1,865 (4.2%) | <0.001 |
| Paralysis | 2,170 (5.0%) | 1,775 (4.0%) | <0.001 |
| Dementia | 2,795 (6.5%) | 3,000 (6.8%) | 0.073 |
| Psych/Behavior | |||
| Alcohol use | 1,775 (4.1%) | 725 (1.6%) | <0.001 |
| Drug abuse | 1,220 (2.8%) | 970 (2.2%) | <0.001 |
| Depression | 5,280 (12.2%) | 9,230 (20.9%) | <0.001 |
| Psychoses | 1,645 (3.8%) | 2,585 (5.8%) | <0.001 |
| Renal | |||
| Moderate renal failure | 7,190 (16.7%) | 5,875 (13.3%) | <0.001 |
| Severe renal failure | 3,460 (8.0%) | 3,400 (7.7%) | 0.065 |
| Liver | |||
| Mild liver disease | 3,050 (7.1%) | 2,825 (6.4%) | <0.001 |
| Severe liver disease | 750 (1.7%) | 500 (1.1%) | <0.001 |
| Respiratory | |||
| Chronic lung disease | 13,260 (30.7%) | 16,370 (37.0%) | <0.001 |
| Pulmonary circulation disorder | 3,640 (8.4%) | 4,055 (9.2%) | <0.001 |
| Metabolic/Endocrine | |||
| Obesity | 9,975 (23.1%) | 13,685 (30.9%) | <0.001 |
| Diabetes, complicated | 11,215 (26.0%) | 9,750 (22.0%) | <0.001 |
| Diabetes, uncomplicated | 5,415 (12.5%) | 5,305 (12.0%) | 0.011 |
| Hypothyroidism | 4,465 (10.3%) | 9,830 (22.2%) | <0.001 |
| Other/GI/Nutrition | |||
| Ulcer peptic disease | 585 (1.4%) | 485 (1.1%) | <0.001 |
| Weight loss (Cachexia) | 6,225 (14.4%) | 4,830 (10.9%) | <0.001 |
| Hospital characteristics and admission profiles | |||
| Census division of hospital | <0.001 | ||
| New England | 2,130 (4.9%) | 1,960 (4.4%) | |
| Middle Atlantic | 5,475 (12.7%) | 6,025 (13.6%) | |
| East North Central | 8,085 (18.7%) | 8,655 (19.6%) | |
| West North Central | 3,790 (8.8%) | 3,750 (8.5%) | |
| South Atlantic | 8,195 (19.0%) | 8,430 (19.0%) | |
| East South Central | 3,320 (7.7%) | 3,375 (7.6%) | |
| West South Central | 4,455 (10.3%) | 4,825 (10.9%) | |
| Mountain | 2,990 (6.9%) | 2,990 (6.8%) | |
| Pacific | 4,715 (10.9%) | 4,250 (9.6%) | |
| Location/teaching status of hospital | <0.001 | ||
| Rural | 5,150 (11.9%) | 5,585 (12.6%) | |
| Urban nonteaching | 7,300 (16.9%) | 7,225 (16.3%) | |
| Urban teaching | 30,705 (71.2%) | 31,450 (71.1%) | |
| Relative bed size category | <0.001 | ||
| Small | 10,815 (25.1%) | 11,395 (25.7%) | |
| Medium | 11,490 (26.6%) | 11,960 (27.0%) | |
| Large | 20,850 (48.3%) | 20,905 (47.2%) | |
| Elective vs non-elective admission | <0.001 | ||
| Non-elective | 38,790 (89.9%) | 39,730 (89.9%) | |
| Elective | 4,340 (10.1%) | 4,485 (10.1%) | |
| Admission day | <0.001 | ||
| Weekday | 33,815 (78.4%) | 33,965 (76.7%) | |
| Weekend | 9,340 (21.6%) | 10,295 (23.3%) | |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
