ABSTRACT
-
Background/Aims
Ursodeoxycholic acid (UDCA) downregulates angiotensin-converting enzyme 2, the cellular entry receptor for SARS-CoV-2, and may reduce acute coronavirus disease 2019 (COVID-19) severity. However, it remains unknown whether UDCA prevents long COVID outcomes in patients with steatotic liver disease (SLD)—a population vulnerable to adverse outcomes. We investigated the association between pre-infection UDCA use and risk of long COVID outcomes in patients with SLD.
-
Methods
We conducted a nationwide retrospective cohort study using the Korean COVID-19 registry linked to National Health Insurance Service claims data (2019–2022). Patients with SLD (fatty liver index ≥30) who experienced COVID-19 were included. Exposure was defined as at least one UDCA prescription within the 90 days preceding infection. Outcomes of interest were 20 incident long COVID conditions across eight organ systems assessed ≥84 days post-infection. After propensity score (PS) fine stratification, hazard ratios (HR) with 95% confidence intervals (CI) were estimated using Cox regression.
-
Results
Among 469,108 patients with SLD and COVID-19, 23,560 (5.0%) were UDCA users. After PS fine stratification weighting, UDCA use was not associated with reduced risk for most long COVID outcomes. A protective association was observed for atrial fibrillation (HR 0.51, 95% CI 0.30–0.88), whereas increased risks were found for type 2 diabetes mellitus (HR 1.24, 95% CI 1.09–1.42) and epilepsy (HR 1.69, 95% CI 1.09–2.61).
-
Conclusions
Pre-infection UDCA use was not associated with reduced risk for most long COVID outcomes in patients with SLD. The observed associations warrant cautious interpretation given potential residual confounding and surveillance bias.
-
Keywords: Ursodeoxycholic acid; Long COVID; Fatty liver disease; Angiotensin-converting enzyme 2; Pharmacoepidemiology
Study Highlights
• In a nationwide cohort of patients with steatotic liver disease and COVID-19, pre-infection ursodeoxycholic acid use was not associated with reduced risk of most long COVID outcomes across multiple organ systems.
• However, ursodeoxycholic acid use showed a protective association with atrial fibrillation alongside increased risks of type 2 diabetes and epilepsy; these findings should be interpreted cautiously.
Graphical Abstract
INTRODUCTION
Since the coronavirus disease 2019 (COVID-19) pandemic onset, more than 778 million confirmed cases have been reported worldwide as of December 2025 [
1]. Although advances in acute management, vaccination, and antiviral therapy have substantially reduced mortality, a considerable proportion of patients experience long COVID—persistent or newly emerging symptoms lasting at least three months after acute infection [
2,
3]. Long COVID affects multiple organ systems contributing to functional impairment, increased healthcare utilization, and substantial public health burden.
Patients with steatotic liver disease (SLD) are particularly vulnerable to adverse COVID-19 outcomes due to metabolic dysfunction, chronic inflammation, and immune dysregulation [
4,
5]. Individuals with cirrhosis demonstrate elevated morbidity and mortality following COVID-19 infection [
6,
7]. Moreover, COVID-19 has been linked to persistent hepatobiliary and gastrointestinal disorders extending beyond the acute phase [
2]. However, no studies have specifically investigated the risk of long COVID outcomes across multiple organ systems in patients with SLD, representing a critical knowledge gap.
Ursodeoxycholic acid (UDCA), a hydrophilic bile acid widely used in cholestatic and metabolic liver diseases, has emerged as a potential protective agent against COVID-19. Experimental studies suggest UDCA may attenuate viral infection by downregulating angiotensin-converting enzyme 2 (ACE2) expression through inhibition of farnesoid X receptor (FXR) signaling [
8–
10]. However, real-world evidence on the association between UDCA use and COVID-19 outcomes remains inconsistent, with some studies reporting protective effects [
11–
13] while others show no association [
14]. Furthermore, the potential impact of UDCA on long COVID outcomes, particularly in patients with SLD, has not been evaluated.
We therefore conducted a nationwide cohort study using linked healthcare data and the national COVID-19 registry to examine the association between pre-infection UDCA use and the risk of 20 long COVID outcomes in patients with SLD. This study addresses limited evidence on pharmacologic factors influencing long COVID and aims to inform future research on preventive strategies in vulnerable populations.
MATERIALS AND METHODS
Ethics approval
Ethical approval was obtained from the Institutional Review Board of Sungkyunkwan University (IRB No. SKKU 2023-04-012), where the requirement of informed consent was waived as this study used anonymized administrative data. This study adhered to ethical principles outlined in the Declaration of Helsinki.
Data source
We utilized the Korea Disease Control and Prevention Agency-COVID-19-National Health Insurance Service (K-COV-N) cohort [
15,
16], which integrates the nationwide COVID-19 registry with healthcare data from the National Health Insurance Service (NHIS) covering January 2019 through December 2022. The NHIS database encompasses individual-level healthcare utilization records, health examination data, and death records for more than 99% of the Korean population [
17]. It includes sociodemographic characteristics, diagnoses coded using the International Classification of Diseases, 10th Revision (ICD-10), detailed prescription records, and procedure codes from inpatient, outpatient, and emergency department visits. Regular health examination data include smoking status, alcohol consumption, anthropometric measurements, and laboratory values. The COVID-19 registry from the Korea Disease Control and Prevention Agency (KDCA) contains comprehensive information on confirmed COVID-19 cases, including dates of confirmation, routes of transmission, and COVID-19 vaccination status (number of doses and vaccine type).
We conducted a retrospective cohort study to estimate the association between UDCA use and risk of long COVID outcomes among patients with SLD. The base cohort comprised patients aged ≥20 years with SLD who had confirmed COVID-19 infection between January 1, 2020, and December 31, 2022. SLD was defined using fatty liver index (FLI) ≥30, a validated noninvasive surrogate marker for hepatic steatosis that has demonstrated excellent diagnostic accuracy in large population-based epidemiological studies [
18,
19]. The calculation method for FLI is presented in
Supplementary Table 1. COVID-19 infection was confirmed by positive real-time reverse transcription polymerase chain reaction assay or approved antigen test using nasal or pharyngeal swabs, as validated by KDCA [
15,
16,
20]. The cohort entry date was defined as the date of first confirmed COVID-19 infection during the study period.
Patients were classified as UDCA users if they received at least one UDCA prescription (Anatomical Therapeutic Chemical code: A05AA02) within the 90-day period preceding cohort entry. This exposure window was chosen to capture biologically relevant drug exposure during the critical periods of viral incubation and early viral replication, when ACE2 downregulation by UDCA might theoretically exert protective effects. Patients without a documented UDCA prescription during this 90-day window were classified as non-users and served as the reference group.
Outcomes and follow-up
The outcomes of interest comprised incident diagnoses of 20 prespecified long COVID–related conditions across eight organ systems: cardiovascular (hypertension, atrial fibrillation), pulmonary (interstitial lung disease, chronic obstructive pulmonary disease, asthma), neurological (neuropathy, epilepsy, migraine), psychiatric (insomnia, depression, anxiety, psychosis), gastrointestinal (gastroesophageal reflux disease [GERD], irritable bowel syndrome), endocrine (type 2 diabetes mellitus [T2DM], hypothyroidism, Hashimoto’s thyroiditis, hyperthyroidism), dermatological (atopic dermatitis), and musculoskeletal (rheumatoid arthritis) (
Supplementary Table 2) [
3]. In accordance with the World Health Organization definition of long COVID, the index date was defined as 84 days (12 weeks) after initial confirmed COVID-19 infection [
21].
Patients were followed from the index date until the earliest occurrence of: (1) the outcome of interest, (2) death, (3) COVID-19 reinfection, or (4) end of study period (December 31, 2022). Patients who died, had COVID-19 reinfection, reached the study end, and were diagnosed with viral hepatitis (to exclude competing liver disease) before the index date were excluded. Patients receiving COVID-19 vaccination between cohort entry date and index date were also excluded. To ensure ascertainment of incident outcomes, individuals with a documented history of each outcome prior to the index date were excluded from the respective outcome-specific analysis.
Covariates
Age, sex, insurance type, and dominant COVID-19 variant were assessed at cohort entry. Comorbidities, concomitant medications, the Charlson Comorbidity Index, and COVID-19 vaccination history were assessed during the 365 days preceding cohort entry. Lifestyle factors were assessed prior to cohort entry based on the most recent health examination, and SLD subtype was defined according to the SLD diagnosis recorded before cohort entry. Detailed definitions are provided in
Supplementary Table 3. The detailed study diagram is shown in
Figure 1.
For each outcome-specific analysis, propensity score (PS) fine stratification weighting was applied to adjust for confounding [
22,
23]. The PS—the conditional probability of UDCA use versus non-use given prespecified baseline covariates— was estimated using multivariable logistic regression. After trimming individuals in non-overlapping regions of PS distributions, we constructed 50 fine strata based on the PS distribution among UDCA users. UDCA users were assigned a weight of 1, while non-users were reweighted proportionally to the number of UDCA users within each PS stratum. Covariate balance was assessed using absolute standardized mean differences (ASD), with values <0.1 indicating adequate balance. Weighted incidence rates with 95% confidence intervals (CI) were calculated based on the Poisson distribution. Weighted Cox proportional hazards models estimated hazard ratios (HR) and 95% CI for each outcome. Given the use of NHIS- and registry-based administrative data, the extent of missing data was minimal.
To assess whether the association between UDCA use and long COVID risk differed across clinically relevant subgroups, we conducted stratified analyses by age group (20–40, 41–65, >65 years), sex, prior use of renin-angiotensin system (RAS) inhibitors (ACE inhibitors or angiotensin II receptor blockers), presence of cirrhosis, dominant SARS-CoV-2 variant period (pre-Delta, Delta, Omicron), and COVID-19 vaccination status. To assess the robustness of findings, sensitivity analyses were performed using a more stringent SLD definition (FLI ≥60). Additionally, post-hoc E-value analysis was conducted to assess the minimum strength of unmeasured confounding required to nullify observed associations.
All statistical analyses were performed using SAS software, version 7.1 (SAS Institute, Cary, NC, USA). Statistical significance was defined as two-tailed P-value <0.05.
RESULTS
Baseline characteristics
We identified a base cohort of 469,108 patients with SLD infected with COVID-19 between January 1, 2020 and December 31, 2022, comprising 23,560 (5.0%) UDCA users and 445,548 (95.0%) non-users (
Fig. 2). For outcome-specific analyses, we created 20 distinct cohorts, each excluding individuals with a prior history of the outcomes of interest. Sample sizes for UDCA users and non-users in each outcome-specific analysis are shown in
Supplementary Tables 4–
23.
The overall study population was predominantly male (73.4%) with a mean age of 51.5 years (standard deviation 13.4). The most prevalent type of SLD was metabolic dysfunction-associated steatotic liver disease (MASLD; 89.2%), followed by metabolic dysfunction and alcohol-related liver disease (8.1%) and alcohol-related liver disease (2.4%). Most COVID-19 infections (98.6%) occurred during the Omicron-dominant phase, and 97.1% of patients had received at least one COVID-19 vaccine dose prior to infection.
Compared with non-users, UDCA users were older (mean age, 54.5 vs. 51.4 years) and exhibited greater comorbidity burden, including higher prevalence of T2DM (33.1% vs. 14.9%), dyslipidemia (52.4% vs. 28.2%), cirrhosis (31.3% vs. 7.0%), thyroid disease (7.5% vs. 5.0%), cancer (7.1% vs. 3.4%), and osteoporosis (21.8% vs. 16.4%). Concomitant medication use was more prevalent among UDCA users across most therapeutic categories (
Table 1). Baseline characteristics for each outcome-specific cohort are presented in
Supplementary Tables 4–
23. After PS fine stratification, all covariates achieved adequate balance (ASD <0.1).
During the median follow-up ranging from 134.3 to 141.6 days, UDCA use was not associated with reduced risk for most long COVID outcomes (
Table 2). For cardiovascular outcomes, no association was observed for hypertension (HR 1.03; 95% CI 0.74–1.44). However, atrial fibrillation occurred in 14 UDCA users and 315 non-users, corresponding to weighted incidence rates of 0.14 and 0.17 per 100 person-years, respectively. UDCA use was associated with lower risk of atrial fibrillation (HR 0.51; 95% CI 0.30–0.88). For pulmonary outcomes, no associations were observed for interstitial lung disease (HR 0.82; 95% CI 0.36–1.86), chronic obstructive pulmonary disease (HR 1.14; 95% CI 0.74–1.75), or asthma (HR 0.96; 95% CI 0.73–1.27). For neurological outcomes, UDCA use was associated with increased risk of epilepsy (HR 1.69; 95% CI 1.09–2.61), whereas no associations were observed for neuropathy (HR 0.94; 95% CI 0.79–1.13) or migraine (HR 0.73; 95% CI 0.51–1.03). No significant associations were observed for psychiatric outcomes, including insomnia (HR 0.76; 95% CI 0.53–1.09), depression (HR 1.01; 95% CI 0.82–1.25), anxiety (HR 1.15; 95% CI 0.94–1.42), and psychosis (HR 1.38; 95% CI 0.70–2.72), or for gastrointestinal outcomes, including GERD (HR 1.02; 95% CI 0.92–1.14) and irritable bowel syndrome (HR 1.02; 95% CI 0.90–1.15). For endocrine outcomes, UDCA use was associated with a higher risk of T2DM (HR 1.24; 95% CI 1.09–1.42). Hazard ratios were also elevated, though not statistically significant, for hypothyroidism (HR 1.35; 95% CI 0.99–1.84), Hashimoto’s thyroiditis (HR 1.90; 95% CI 0.87–4.13), and hyperthyroidism (HR 1.26; 95% CI 0.74–2.17). Finally, no associations were observed for dermatologic or musculoskeletal outcomes, including atopic dermatitis (HR 1.07; 95% CI 0.78–1.45) and rheumatoid arthritis (HR 2.05; 95% CI 0.88–4.76).
Figure 3 presents subgroup analyses stratified by RAS inhibitor use. Among RAS inhibitor users, increased risks of epilepsy (HR 2.05; 95% CI 1.13–3.70) and T2DM (HR 1.35; 95% CI 1.13–1.62) were consistent with main findings. A more pronounced protective association with atrial fibrillation was observed among non-RAS inhibitor users (HR 0.26; 95% CI 0.08–0.80). No significant effect modification by RAS inhibitor use was detected (
P for interaction > 0.05). Additional subgroup analyses stratified by age group, sex, presence of cirrhosis, SARS-CoV-2 variant period, and COVID-19 vaccination status are presented in
Supplementary Tables 24–
28.
Sensitivity analyses using a more stringent SLD definition (FLI ≥60) yielded results consistent with main findings (
Supplementary Table 29). E-values for point estimates were 3.33 for atrial fibrillation, 2.77 for epilepsy, and 1.79 for T2DM, suggesting that moderate to strong unmeasured confounding would be required to explain away the observed associations.
DISCUSSION
This nationwide cohort study represents the first comprehensive evaluation of the long-term impact of UDCA on post-COVID-19 sequelae in patients with SLD. Contrary to the biological hypothesis that UDCA-mediated ACE2 downregulation might mitigate viral pathogenicity and downstream complications, pre-infection UDCA use was not associated with reduced risk for the vast majority of long COVID outcomes across cardiovascular, pulmonary, neurological, psychiatric, gastrointestinal, endocrine, dermatological, and musculoskeletal systems. A protective association with atrial fibrillation was observed, whereas increased risks were found for T2DM and epilepsy.
Prior studies suggested potential benefits of UDCA during the acute phase of COVID-19 [
24]. Brevini et al. [
10] demonstrated that FXR inhibition by UDCA reduces ACE2 expression in human lung and liver organoids, thereby limiting SARS-CoV-2 infection
in vitro. However, mechanisms influencing early viral entry may not directly affect post-infectious sequelae characterizing long COVID. The pathophysiology of long COVID is multifactorial, involving persistent immune dysregulation, T-cell exhaustion, latent viral reactivation, tissue viral persistence, autoimmunity, and endothelial dysfunction [
25–
28]. Accordingly, ACE2 downregulation by UDCA during the early phase of infection may be insufficient to influence downstream chronic inflammatory and immunological, and thrombotic pathways that characterize long COVID. A recent randomized controlled trial found a 2-week UDCA course did not improve long COVID recovery at 8 weeks compared with placebo [
29]. Our observed lack of effectiveness of UDCA across multiple organ systems might reflect discrepancies between experimental findings and real-world observations, including insufficient dosing and distinct immune and metabolic alterations in SLD patients, along with suboptimal timing of UDCA exposure relative to SARS-CoV-2 infection.
Among the 20 outcomes examined, the only protective association observed was for new-onset atrial fibrillation, and this finding was consistent across several subgroups, particularly among patients not receiving RAS inhibitors. Arrhythmias are well-documented complications of long COVID, potentially driven by autonomic dysfunction, persistent myocardial inflammation, direct viral myocardial injury, and sustained cardiac metabolic stress [
30–
32]. UDCA possesses multiple pleiotropic properties relevant to cardiovascular protection, including anti-inflammatory effects, cytoprotection through stabilization of mitochondrial function, reduction of oxidative stress, and modulation of endoplasmic reticulum stress [
33–
35]. These effects may attenuate systemic inflammation and myocardial stress, which are key contributors to atrial remodeling, fibrosis, and arrhythmogenesis, in the setting of long COVID. In addition, ACE2 is expressed in multiple cardiac cell types, including cardiomyocytes and cardiac pericytes, and altered cardiac ACE2 signaling has been implicated in SARS-CoV-2–related myocardial injury and cardiovascular complications [
36]. Therefore, modulation of the FXR–ACE2 axis by UDCA may represent one biologically plausible mechanism potentially contributing to the observed association with atrial fibrillation. Additionally, UDCA may reduce susceptibility to atrial fibrillation by modulating cardiomyocyte ion channel activity and calcium handling, as bile acid signaling has been implicated in cardiac electrophysiology and rhythm disorders [
37]. Given that SLD patients frequently manifest coexisting metabolic dysfunction, insulin resistance and cardiovascular risk [
38], the observed protective association with atrial fibrillation may reflect a combination of cardiac electrophysiological effects, modulation of ACE2-related pathways, and indirect improvements in systemic metabolic and inflammatory profiles. This finding aligns with UDCA’s demonstrated anti-arrhythmic potential in other contexts, such as cholestatic fetal cardiac arrhythmia [
39], though further mechanistic validation through experimental and clinical research is warranted.
The observed increased risks of T2DM and epilepsy among UDCA users warrant cautious interpretation. MASLD is associated with an increased risk of gallstones [
40], and the nonlinear relationship between liver steatosis severity and gallstone risk [
41] suggests a higher likelihood of UDCA use among patients with more severe liver disease. Furthermore, COVID-19 infection itself has been associated with increased risk of new-onset T2DM through multiple mechanisms including direct pancreatic beta-cell injury, immune-mediated destruction, and metabolic decompensation [
42,
43], and with neurological complications, including seizures and epilepsy, potentially due to neuroinflammation, direct neuronal injury, or unmasking of latent seizure susceptibility [
44]. Patients prescribed UDCA likely differ from non-users in underlying clinical characteristics, including more advanced hepatobiliary disease, greater metabolic dysfunction, and higher systemic inflammation burden, all of which may predispose individuals to post-COVID metabolic and neurological complications. Therefore, the observed associations may reflect differences in underlying metabolic and inflammatory phenotypes, infection-related vulnerability, or disease severity rather than direct adverse pharmacologic effects of UDCA. Although we rigorously adjusted for baseline confounders using PS fine stratification, residual confounding and unmeasured differences in disease severity may persist. The E-value analyses suggest that moderate to strong unmeasured confounding (Evalue of 1.79 for T2DM and 2.77 for epilepsy) would be required to fully explain the observed associations, yet residual confounding cannot be entirely excluded. Increased healthcare utilization among patients receiving chronic UDCA therapy may also have contributed to greater detection of clinically subtle conditions such as T2DM. Therefore, these positive associations should be interpreted with appropriate caution.
Associations between UDCA use and long COVID outcomes were generally consistent across subgroups stratified by age, sex, RAS inhibitor use, and cirrhosis presence, with no clear evidence of effect modification. Although RAS inhibitors directly modulate RAS and ACE2-related pathways—theoretically creating a mechanistic synergy with UDCA’s ACE2 downregulation—the associations did not differ meaningfully according to RAS inhibitor use. This finding suggests that the relationship between UDCA and long COVID outcomes may be largely independent of RAS pathway modulation, or that the effects are too subtle to detect meaningful interaction in this population.
This study represents the largest population-based investigation of UDCA and long COVID outcomes to date, utilizing high-quality nationwide COVID-19 registry data with comprehensive capture of medical history and pharmaceutical records. Its focus on SLD patients is particularly relevant given their elevated risk of severe COVID-19 and metabolic complications. Nevertheless, several limitations warrant consideration. First, the observational design precludes causal inference, and residual confounding cannot be completely excluded despite comprehensive PS adjustment. Acute COVID-19 severity and related treatment patterns were not directly incorporated into the adjustment model and may have influenced subsequent long COVID risk. Future studies evaluating the interaction between acute disease severity, COVID-19 therapeutics, and long COVID outcomes in UDCA users would provide additional clinical insight. In addition, claims data may not have been sufficient in capturing the greater likelihood of liver disease severity among those prescribed UDCA, which might have served as an unmeasured confounder. Other unmeasured confounders such as genetic susceptibility, dietary patterns, and psychosocial stressors may also have influenced results. Second, SLD was defined using FLI, a surrogate marker, rather than radiologic or histological confirmation, which may have led to misclassification of SLD. Nevertheless, FLI has demonstrated positive predictive value of 89% in the Korean population and is widely used in large-scale epidemiological research [
45]. In addition, the sensitivity analysis using a more stringent definition (FLI >60) yielded results consistent with the primary findings. Third, exposure to UDCA was ascertained from prescription records, which may not perfectly reflect actual medication consumption due to non-adherence. Additionally, over-the-counter UDCA use, though uncommon in South Korea, was not captured in the database, potentially resulting in exposure misclassification biasing results toward the null. Fourth, outcome misclassification is plausible and results should be interpreted cautiously. Defining long COVID based on ICD-10 codes might have introduced variability in results due to varying diagnostic criteria used by physicians. However, we adopted outcome definitions from prior studies and restricted cases to ≥1 inpatient or ≥2 outpatient diagnoses to improve outcome validity. Fifth, we cannot exclude the possibility that UDCA prescribing patterns reflect unmeasured markers of disease severity or healthcare-seeking behavior. Finally, the study was conducted predominantly during the Omicron variant era (98.6% of infections), characterized by lower acute severity than earlier variants but potentially different long COVID profiles. Generalizability to other viral variants or populations with different demographic and metabolic characteristics may be limited.
In this nationwide cohort study of patients with SLD and COVID-19, pre-infection UDCA use was not associated with reduced risk for most long COVID outcomes across multiple organ systems. A protective association was observed for atrial fibrillation, whereas increased risks were found for T2DM and epilepsy. These findings should be interpreted cautiously, as observed associations may reflect underlying clinical differences in disease severity and metabolic phenotype, post-COVID vulnerability, surveillance bias or residual confounding rather than direct pharmacologic effects of UDCA. The protective association with atrial fibrillation warrants further investigation through mechanistic studies and prospective clinical trials. This investigation builds on recent drug repurposing research of existing medications [
29,
46,
47] examining the association between UDCA use and long COVID outcomes. Our findings suggest that UDCA’s potential protective effects, if any, may be limited to specific outcomes rather than providing broad protection against long COVID. Further mechanistic and clinical research is warranted to elucidate the potential role of UDCA in long COVID prevention and identify patient subgroups with SLD who might benefit from targeted interventions. In addition, further studies comparing the findings of this study with those in the general population with FLI <30 or in the populations with FLI >60 would be informative.
FOOTNOTES
-
Authors’ contribution
J-YS had full access to all data and takes responsibility for data integrity and accuracy of data analysis. Concept and design; acquisition, analysis, or interpretation of data; administrative, technical, or material support; critical revision of the manuscript for important intellectual content: all authors. Drafting of the manuscript: KJ, G-AK, JW, JY, WK. Statistical analysis: KJ, JW. Supervision: WK, J-YS.
-
Acknowledgements
The authors thank the Korea Disease Control and Prevention Agency, the National Academy of Medicine of Korea, and the National Health Insurance Service (NHIS) for collaborative efforts in making nationwide data available for analysis and providing technical assistance. This study used database from KDCA and NHIS for policy and academic research (research number: KDCA-NHIS-2023-1-508).
This research was supported by grants from the Ministry of Food and Drug Safety, South Korea (22183MFDS431, 2022-2025; 21153MFDS607, 2021-2025) and the National Research Foundation (NRF) of Korea (RS-2021-NR056442, RS-2022-NR067269, RS-2023-00223831, RS-2024-00440883, and RS-2025-25458964). This work was supported by KBSMC-SKKU Future Clinical Convergence Academic Research Program, Kangbuk Samsung Hospital & Sungkyunkwan University, 2025. This research was supported by a grant (RS-2026-25511535) from Ministry of Food and Drug Safety in 2026. This research was supported by a grant of Patient-Centered Clinical Research Coordinating Center (PACEN) funded by the Ministry of Health & Welfare, Republic of Korea (RS-2025-02215225).
Data may be obtained from a third party and are not publicly available due to legal data sharing restrictions. Data are provided by the Korea Disease Control and Prevention Agency (KDCA) and the National Health Insurance Service (NHIS). Although legal agreements prohibit direct disclosure, researchers can request database access by submitting well-defined data access requests including explicit details on required data elements, analysis methodology, and planned dissemination to KDCA and NHIS through their official data request portals.
-
Conflicts of Interest
JYS received grants from the Ministry of Food and Drug Safety, the Ministry of Health and Welfare, the National Research Foundation of Republic of Korea, and pharmaceutical companies, including SK Bioscience, Yuhan, UCB, and Pfizer, outside the submitted work. WK received grants from GSK, Gilead, Novartis, Pfizer, Roche, Boehringer-Ingelheim, 89BIO, Aligos, Novo Nordisk, Lilly, Hanmi, Daewoong, and KOBIOLABS; consulting fees from Gilead, Boehringer-Ingelheim, GSK, Novo Nordisk, Lilly, YUHAN, Hanmi, KOBIOLABS, Olix Pharma, TSD Life Sciences, Daewoong, QUEST, Therasid Bioscience, and Korea United Pharm; honoraria for lectures from Ildong, Samil, Bukwang, and Novo Nordisk, owns stocks in KOBIOLABS and Lepidyne, and is the founder of Remedygen. JY is employee of Gilead Sciences. The other authors declare no conflicts of interest.
SUPPLEMENTARY MATERIAL
Supplementary material is available at Clinical and Molecular Hepatology website (
http://www.e-cmh.org).
Supplementary Table 4.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for hypertension)
cmh-2026-0283-Supplementary-Table-4.pdf
Supplementary Table 5.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for atrial fibrillation)
cmh-2026-0283-Supplementary-Table-5.pdf
Supplementary Table 6.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for interstitial lung diseases)
cmh-2026-0283-Supplementary-Table-6.pdf
Supplementary Table 7.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for COPD)
cmh-2026-0283-Supplementary-Table-7.pdf
Supplementary Table 8.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for asthma)
cmh-2026-0283-Supplementary-Table-8.pdf
Supplementary Table 9.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for neuropathy)
cmh-2026-0283-Supplementary-Table-9.pdf
Supplementary Table 10.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for epilepsy)
cmh-2026-0283-Supplementary-Table-10.pdf
Supplementary Table 11.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for migraine)
cmh-2026-0283-Supplementary-Table-11.pdf
Supplementary Table 12.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for insomnia)
cmh-2026-0283-Supplementary-Table-12.pdf
Supplementary Table 13.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for depression)
cmh-2026-0283-Supplementary-Table-13.pdf
Supplementary Table 14.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for anxiety)
cmh-2026-0283-Supplementary-Table-14.pdf
Supplementary Table 15.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for psychosis)
cmh-2026-0283-Supplementary-Table-15.pdf
Supplementary Table 16.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for GERD)
cmh-2026-0283-Supplementary-Table-16.pdf
Supplementary Table 17.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for irritable bowel syndrome)
cmh-2026-0283-Supplementary-Table-17.pdf
Supplementary Table 18.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for type 2 diabetes mellitus)
cmh-2026-0283-Supplementary-Table-18.pdf
Supplementary Table 19.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for hypothyroidism)
cmh-2026-0283-Supplementary-Table-19.pdf
Supplementary Table 20.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for Hashimoto’s thyroiditis)
cmh-2026-0283-Supplementary-Table-20.pdf
Supplementary Table 21.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for hyperthyroidism)
cmh-2026-0283-Supplementary-Table-21.pdf
Supplementary Table 22.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for atopic dermatitis)
cmh-2026-0283-Supplementary-Table-22.pdf
Supplementary Table 23.
Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use, before and after propensity score fine stratification (analysis for rheumatoid arthritis)
cmh-2026-0283-Supplementary-Table-23.pdf
Supplementary Table 29.
Sensitivity analysis for association between UDCA use and long COVID risk, redefining steatotic liver disease by fatty liver index greater than 60
cmh-2026-0283-Supplementary-Table-29.pdf
Figure 1Study design flow diagram. Figure illustrates the study design and follow-up structure. The timeline shows: (a) baseline period for exposure ascertainment (UDCA prescription within 90 days prior to COVID-19 infection); (b) index date defined as day 84 post-infection (consistent with WHO long COVID definition); (c) follow-up period until occurrence of outcome, death, COVID-19 reinfection, or end of study (December 31, 2022). COVID-19, coronavirus disease 2019; SLD, steatotic liver disease; UDCA, ursodeoxycholic acid. Censored until the occurrence of a study outcome, date of death, COVID-10 reinfection, or end of the study period (December 31, 2022), whichever occurred earlier.
Figure 2Study population flow chart. Figure presents the flow of study participants through inclusion/exclusion criteria. Starting with the K-COV-N cohort identified through COVID-19 registry, the diagram shows: (a) number of patients with SLD (fatty liver index ≥30) with confirmed COVID-19 infection (Jan 2020–Dec 2022); (b) exclusion criteria applied; (c) classification into UDCA users (prescription within 90 days pre-infection) and non-users (23,560 UDCA users and 445,548 non-users). COVID-19, coronavirus disease 2019; SLD, steatotic liver disease; UDCA, ursodeoxycholic acid.
Figure 3Association between UDCA use and risk of long COVID outcomes, stratified by concomitant RAS inhibitor use. Figure presents a forest plot comparing adjusted hazard ratios (95% CI) for selected long COVID outcomes between UDCA users and non-users, separately for patients taking RAS inhibitors (ACE inhibitors or ARBs) versus non-users. This figure highlights the potential effect modification by RAS inhibitor use. Outcomes shown include cardiovascular (hypertension, atrial fibrillation), pulmonary (interstitial lung disease, chronic obstructive pulmonary disease, asthma), neurological (neuropathy, epilepsy, migraine), psychiatric (insomnia, depression, anxiety, psychosis), gastrointestinal (gastroesophageal reflux disease, irritable bowel syndrome), endocrine (type 2 diabetes mellitus, hypothyroidism, Hashimoto’s thyroiditis, hyperthyroidism), dermatological (atopic dermatitis), and musculoskeletal (rheumatoid arthritis) domains. The horizontal dashed line at HR=1 represents no association. ACE, angiotensin-converting enzyme; ARBs, angiotensin II receptor blockers; CI, confidence interval; COPD, chronic obstructive pulmonary disease; COVID, coronavirus disease; GERD, gastroesophageal reflux disease; HR, hazard ratio; IR, incidence rate; RAS, renin-angiotensin system; T2DM, type 2 diabetes mellitus; UDCA, ursodeoxycholic acid.
Table 1Baseline characteristics of patients with SLD infected with COVID-19, stratified by UDCA use
Table 1
|
All patients (n=469,108) |
UDCA user (n=23,560) |
UDCA non-user (n=445,548) |
ASD |
|
Age |
51.5 (13.4) |
54.5 (12.3) |
51.4 (13.4) |
0.242 |
|
Sex |
|
|
|
0.007 |
|
Male |
344,236 (73.4) |
17,360 (73.7) |
326,876 (73.4) |
|
|
Female |
124,872 (26.6) |
6,200 (26.3) |
118,672 (26.6) |
|
|
Insurance type |
|
|
|
0.121 |
|
Health insurance |
463,205 (98.7) |
22,883 (97.1) |
440,322 (98.8) |
|
|
Medical aid |
5,903 (1.3) |
677 (2.9) |
5,226 (1.2) |
|
|
Smoking |
232,108 (49.5) |
12,203 (51.8) |
219,905 (49.4) |
0.060 |
|
Alcohol |
338,803 (72.2) |
16,417 (69.7) |
322,386 (72.4) |
0.044 |
|
SLD type |
|
|
|
0.097 |
|
MASLD |
418,583 (89.2) |
20,343 (86.4) |
398,240 (89.4) |
|
|
MetALD |
38,043 (8.1) |
2,373 (10.1) |
35,670 (8.0) |
|
|
ALD |
11,066 (2.4) |
780 (3.3) |
10,286 (2.3) |
|
|
Others |
1,416 (0.3) |
64 (0.3) |
1,352 (0.3) |
|
|
Comorbidities |
|
Cirrhosis |
38,655 (8.2) |
7,370 (31.3) |
31,285 (7.0) |
0.648 |
|
COPD |
6,829 (1.5) |
562 (2.4) |
6,267 (1.4) |
0.072 |
|
Diabetes |
74,116 (15.8) |
7,791 (33.1) |
66,325 (14.9) |
0.436 |
|
Dyslipidemia |
137,994 (29.4) |
12,353 (52.4) |
125,641 (28.2) |
0.510 |
|
Hypertension |
8,549 (1.8) |
694 (3.0) |
7,855 (1.8) |
0.078 |
|
Atrial fibrillation |
4,574 (1.0) |
344 (1.5) |
4,230 (1.0) |
0.047 |
|
Heart failure |
11,606 (2.5) |
926 (3.9) |
10,680 (2.4) |
0.066 |
|
Stroke |
7,571 (1.6) |
609 (2.6) |
6,962 (1.6) |
0.072 |
|
Ischemic heart disease |
19,627 (4.2) |
1,442 (6.1) |
18,185 (4.1) |
0.093 |
|
Chronic kidney disease |
4,018 (0.9) |
304 (1.3) |
3,714 (0.8) |
0.044 |
|
Thyroid disease |
24,089 (5.1) |
1,774 (7.5) |
22,315 (5.0) |
0.104 |
|
Cancer |
16,999 (3.6) |
1,681 (7.1) |
15,318 (3.4) |
0.166 |
|
Obstructive sleep apnea |
13,187 (2.8) |
1,055 (4.5) |
12,132 (2.7) |
0.043 |
|
Osteoporosis |
78,076 (16.6) |
5,139 (21.8) |
72,937 (16.4) |
0.139 |
|
Psoriasis |
3,014 (0.6) |
209 (0.9) |
2,805 (0.6) |
0.028 |
|
Depression |
14,788 (3.2) |
1,144 (4.9) |
13,644 (3.1) |
0.092 |
|
Epilepsy |
3,177 (0.7) |
247 (1.1) |
2,930 (0.7) |
0.042 |
|
Anxiety |
17,866 (3.8) |
1,361 (5.8) |
16,505 (3.7) |
0.098 |
|
Comedication |
|
RAS inhibitors |
136,974 (29.2) |
11,523 (48.9) |
125,451 (28.2) |
0.436 |
|
Anticoagulants |
11,714 (2.5) |
1,367 (5.8) |
10,347 (2.3) |
0.177 |
|
Beta blockers |
14,783 (3.2) |
1,239 (5.3) |
13,544 (3.0) |
0.111 |
|
Bisphosphonates |
6,585 (1.4) |
509 (2.2) |
6,076 (1.4) |
0.061 |
|
Calcium channel blockers |
109,024 (23.2) |
9,292 (39.4) |
99,732 (22.4) |
0.376 |
|
Corticosteroids |
228,622 (48.7) |
13,167 (55.9) |
215,455 (48.4) |
0.151 |
|
Diuretics |
40,564 (8.7) |
3,614 (15.3) |
36,950 (8.3) |
0.220 |
|
Nitrates |
1,984 (0.4) |
199 (0.8) |
1,785 (0.4) |
0.056 |
|
Platelet inhibitors |
168,716 (36.0) |
9,993 (42.4) |
158,723 (35.6) |
0.140 |
|
Proton pump inhibitors |
176,808 (37.7) |
12,713 (54.0) |
164,095 (36.8) |
0.349 |
|
Histamine H2-receptor antagonists |
195,133 (41.6) |
12,270 (52.1) |
182,863 (41.0) |
0.223 |
|
Fibrates |
25,694 (5.5) |
3,275 (13.9) |
22,419 (5.0) |
0.306 |
|
Statins |
150,577 (32.1) |
14,071 (59.7) |
136,506 (30.6) |
0.611 |
|
Antidiabetics |
67,540 (14.4) |
7,567 (32.1) |
59,973 (13.5) |
0.456 |
|
Anticonvulsants |
38,878 (8.3) |
3,025 (12.8) |
35,853 (8.1) |
0.157 |
|
Antidepressants |
19,749 (4.2) |
1,626 (6.9) |
18,123 (4.1) |
0.125 |
|
Antipsychotics |
61,480 (13.1) |
4,233 (18.0) |
57,247 (12.9) |
0.142 |
|
Benzodiazepines |
108,552 (23.1) |
8,689 (36.9) |
99,863 (22.4) |
0.321 |
|
CCI score |
0.5 (0.9) |
1.2 (1.3) |
0.4 (0.9) |
0.719 |
|
CCI group |
|
|
|
0.817 |
|
0 |
345,133 (73.6) |
9,810 (41.6) |
335,323 (75.3) |
|
|
1 |
54,555 (11.6) |
2,571 (10.9) |
51,984 (11.7) |
|
|
2 |
52,817 (11.3) |
8,457 (35.9) |
44,360 (10.0) |
|
|
3+ |
16,603 (3.5) |
2,722 (11.6) |
13,881 (3.1) |
|
|
Variants of COVID-19 |
|
|
|
0.165 |
|
Before Delta |
1,786 (0.4) |
162 (0.7) |
1,624 (0.4) |
|
|
Delta |
4,922 (1.1) |
363 (1.5) |
4,559 (1.0) |
|
|
Omicron |
462,400 (98.6) |
23,035 (97.8) |
439,365 (98.6) |
|
|
COVID-19 vaccination |
455,551 (97.1) |
22,697 (96.3) |
432,854 (97.2) |
0.046 |
Table 2Association between use of ursodeoxycholic acid and long COVID outcomes in patients with steatotic liver disease
Table 2
|
UDCA user |
UDCA non-user |
Weighted HR (95% CI) |
|
No. of events |
No. of patients |
Weighted IR per 100 person-years |
No. of events |
No. of patients |
Weighted IR per 100 person-years |
|
Cardiovascular |
|
Hypertension |
36 |
22,765 |
0.36 (0.25–0.50) |
443 |
436,165 |
0.24 (0.21–0.26) |
1.03 (0.74–1.44) |
|
Atrial fibrillation |
14 |
23,166 |
0.14 (0.08–0.23) |
315 |
440,464 |
0.17 (0.15–0.19) |
0.51 (0.30–0.88) |
|
Pulmonary |
|
Interstitial lung diseases |
6 |
23,491 |
0.06 (0.02–0.13) |
83 |
444,888 |
0.04 (0.03–0.05) |
0.82 (0.36–1.86) |
|
COPD |
22 |
23,139 |
0.22 (0.14–0.33) |
242 |
441,098 |
0.13 (0.11–0.14) |
1.14 (0.74–1.75) |
|
Asthma |
54 |
22,391 |
0.55 (0.42–0.72) |
748 |
427,688 |
0.41 (0.38–0.44) |
0.96 (0.73–1.27) |
|
Neurological |
|
Neuropathy |
121 |
22,112 |
1.26 (1.04–1.50) |
1,843 |
426,567 |
1.01 (0.96–1.05) |
0.94 (0.79–1.13) |
|
Epilepsy |
22 |
23,328 |
0.22 (0.14–0.33) |
140 |
442,856 |
0.07 (0.06–0.09) |
1.69 (1.09–2.61) |
|
Migraine |
33 |
22,875 |
0.33 (0.23–0.46) |
687 |
435,952 |
0.37 (0.34–0.40) |
0.73 (0.51–1.03) |
|
Psychiatric |
|
Insomnia |
31 |
22,894 |
0.31 (0.21–0.44) |
582 |
438,225 |
0.31 (0.28–0.34) |
0.76 (0.53–1.09) |
|
Depression |
88 |
22,242 |
0.91 (0.73–1.12) |
1,329 |
429,350 |
0.72 (0.68–0.76) |
1.01 (0.82–1.25) |
|
Anxiety |
93 |
22,365 |
0.95 (0.77–1.17) |
1,199 |
430,577 |
0.65 (0.61–0.69) |
1.15 (0.94–1.42) |
|
Psychosis |
9 |
23,347 |
0.09 (0.04–0.17) |
90 |
443,096 |
0.05 (0.04–0.06) |
1.38 (0.70–2.72) |
|
Gastrointestinal |
|
GERD |
325 |
14,680 |
5.05 (4.51–5.62) |
5,576 |
320,784 |
4.04 (3.94–4.15) |
1.02 (0.92–1.14) |
|
Irritable bowel syndrome |
252 |
18,283 |
3.16 (2.78–3.58) |
4,101 |
378,170 |
2.53 (2.45–2.61) |
1.02 (0.90–1.15) |
|
Endocrinal |
|
T2DM |
225 |
15,908 |
3.23 (2.82–3.68) |
2,755 |
382,520 |
1.68 (1.61–1.74) |
1.24 (1.09–1.42) |
|
Hypothyroidism |
43 |
22,771 |
0.43 (0.31–0.58) |
462 |
435,590 |
0.25 (0.22–0.27) |
1.35 (0.99–1.84) |
|
Hashimoto’s thyroiditis |
7 |
23,481 |
0.07 (0.03–0.14) |
83 |
444,219 |
0.04 (0.03–0.05) |
1.90 (0.87–4.13) |
|
Hyperthyroidism |
14 |
23,339 |
0.14 (0.08–0.23) |
209 |
442,804 |
0.11 (0.10–0.13) |
1.26 (0.74–2.17) |
|
Dermatological |
|
Atopic dermatitis |
43 |
22,991 |
0.43 (0.31–0.58) |
675 |
435,826 |
0.36 (0.33–0.39) |
1.07 (0.78–1.45) |
|
Musculoskeletal |
|
Rheumatoid arthritis |
6 |
23,492 |
0.06 (0.02–0.13) |
40 |
444,710 |
0.02 (0.01–0.03) |
2.05 (0.88–4.76) |
Abbreviations
angiotensin-converting enzyme 2
alcohol-related liver disease
absolute standardized mean difference
gastroesophageal reflux disease
International Classification of Diseases, 10th Revision
Korea Disease Control and Prevention Agency-COVID-19-National Health Insurance Service
Korea Disease Control and Prevention Agency
metabolic dysfunction-associated steatotic liver disease
metabolic dysfunction and alcohol-related liver disease
National Health Insurance Service
REFERENCES
- 1. World Health Organization. WHO COVID-19 dashboard. World Health Organization web site. <https://data.who.int/dashboards/covid19/summary>. Accessed 23 Mar 2026.
- 2. Lee K, Park J, Lee J, Lee M, Kim HJ, Son Y, et al. Long-term gastrointestinal and hepatobiliary outcomes of COVID-19: a multinational population-based cohort study from South Korea, Japan, and the UK. Clin Mol Hepatol 2024;30:943-958.
- 3. Ely EW, Brown LM, Fineberg HV. Long COVID defined. N Engl J Med 2024;391:1746-1753.
- 4. Yan M, Man S, Ma L, Guo L, Huang L, Gao W. Immunological mechanisms in steatotic liver diseases: an overview and clinical perspectives. Clin Mol Hepatol 2024;30:620-648.
- 5. Woo C, Jeong WI. Immunopathogenesis of liver fibrosis in steatotic liver disease. Clin Mol Hepatol 2024;30:299-302.
- 6. Ioannou GN, Liang PS, Locke E, Green P, Berry K, O’Hare AM, et al. Cirrhosis and severe acute respiratory syndrome coronavirus 2 infection in US veterans: risk of infection, hospitalization, ventilation, and mortality. Hepatology 2021;74:322-335.
- 7. Baldelli L, Marjot T, Barnes E, Barritt AS, Webb GJ, Moon AM. SARS-CoV-2 infection and liver disease: a review of pathogenesis and outcomes. Gut Liver 2023;17:12-23.
- 8. Hoffmann M, Kleine-Weber H, Schroeder S, Krüger N, Herrler T, Erichsen S, et al. SARS-CoV-2 cell entry depends on ACE2 and TMPRSS2 and is blocked by a clinically proven protease inhibitor. Cell 2020;181:271-280e8.
- 9. Datta PK, Liu F, Fischer T, Rappaport J, Qin X. SARS-CoV-2 pandemic and research gaps: understanding SARS-CoV-2 interaction with the ACE2 receptor and implications for therapy. Theranostics 2020;10:7448-7464.
- 10. Brevini T, Maes M, Webb GJ, John BV, Fuchs CD, Buescher G, et al. FXR inhibition may protect from SARS-CoV-2 infection by reducing ACE2. Nature 2023;615:134-142.
- 11. Costello RE, Waller KMJ, Smith R, Mells GF, Wong AYS, Schultze A, et al. Ursodeoxycholic acid and severe COVID-19 outcomes in a cohort study using the OpenSAFELY platform. Commun Med (Lond) 2024;4:238.
- 12. Moon SY, Son M, Kang YW, Koh M, Lee JY, Baek YH. Association between ursodeoxycholic acid use and COVID-19 in individuals with chronic liver disease: a nationwide case-control study in South Korea. Virol J 2024;21:202.
- 13. John BV, Bastaich D, Webb G, Brevini T, Moon A, Ferreira RD, et al. Ursodeoxycholic acid is associated with a reduction in SARS-CoV-2 infection and reduced severity of COVID-19 in patients with cirrhosis. J Intern Med 2023;293:636-647.
- 14. Corpechot C, Verdoux M, Frank-Soltysiak M, Duclos-Vallée JC, Grimaldi L. Exploring the impact of ursodeoxycholic acid therapy on COVID-19 in a real-world setting. J Med Virol 2024;96:e29418.
- 15. Choi Y, Kim HJ, Park J, Lee M, Kim S, Koyanagi A, et al. Acute and post-acute respiratory complications of SARS-CoV-2 infection: population-based cohort study in South Korea and Japan. Nat Commun 2024;15:4499.
- 16. Heo YW, Jeon JJ, Ha MC, Kim YH, Lee S. Long-term risk of autoimmune and autoinflammatory connective tissue disorders following COVID-19. JAMA Dermatol 2024;160:1278-1287.
- 17. Lim SJ, Jang SI. Leveraging national health insurance service data for public health research in Korea: structure, applications, and future directions. J Korean Med Sci 2025;40:e111.
- 18. Bedogni G, Bellentani S, Miglioli L, Masutti F, Passalacqua M, Castiglione A, et al. The fatty liver index: a simple and accurate predictor of hepatic steatosis in the general population. BMC Gastroenterol 2006;6:33.
- 19. Cho EJ, Jung GC, Kwak MS, Yang JI, Yim JY, Yu SJ, et al. Fatty liver index for predicting nonalcoholic fatty liver disease in an asymptomatic Korean population. Diagnostics (Basel) 2021;11:2233.
- 20. Lee SW, Lee J, Moon SY, Jin HY, Yang JM, Ogino S, et al. Physical activity and the risk of SARS-CoV-2 infection, severe COVID-19 illness and COVID-19 related mortality in South Korea: a nationwide cohort study. Br J Sports Med 2022;56:901-912.
- 21. World Health Organization. Post COVID-19 condition (long COVID). World Health Organization web site. <https://www.who.int/news-room/fact-sheets/detail/post-covid-19-condition-(long-covid)>. Accessed 23 Mar 2026.
- 22. Desai RJ, Franklin JM. Alternative approaches for confounding adjustment in observational studies using weighting based on the propensity score: a primer for practitioners. BMJ 2019;367:l5657.
- 23. Desai RJ, Rothman KJ, Bateman BT, Hernandez-Diaz S, Huybrechts KF. A propensity-score-based fine stratification approach for confounding adjustment when exposure is infrequent. Epidemiology 2017;28:249-257.
- 24. Song JH, Shim SR, Shin J, Choe WH, Park J, Lee TH, et al. Clinical effects of ursodeoxycholic acid in COVID-19 infection: a systematic review and dose-response meta-analysis. Front Pharmacol 2026;17:1719144.
- 25. Iwasaki A, Putrino D. Why we need a deeper understanding of the pathophysiology of long COVID. Lancet Infect Dis 2023;23:393-395.
- 26. Gupta G, Buonsenso D, Wood J, Mohandas S, Warburton D. Mechanistic insights into long COVID: viral persistence, immune dysregulation, and multi-organ dysfunction. Compr Physiol 2025;15:e70019.
- 27. Davis HE, McCorkell L, Vogel JM, Topol EJ. Long COVID: major findings, mechanisms and recommendations. Nat Rev Microbiol 2023;21:133-146.
- 28. Altmann DM, Whettlock EM, Liu S, Arachchillage DJ, Boyton RJ. The immunology of long COVID. Nat Rev Immunol 2023;23:618-634.
- 29. Lim SY, Lee J, Chang E, Kwon JS, Jang CY, Seo Y, et al. Neither metformin nor ursodeoxycholic acid effectively treats postacute sequelae of COVID-19 : a randomized clinical trial. Ann Intern Med 2026;179:478-485.
- 30. Huseynov A, Akin I, Duerschmied D, Scharf RE. Cardiac arrhythmias in post-COVID syndrome: prevalence, pathology, diagnosis, and treatment. Viruses 2023;15:389.
- 31. Zhang T, Li Z, Mei Q, Walline JH, Zhang Z, Liu Y, et al. Cardiovascular outcomes in long COVID-19: a systematic review and meta-analysis. Front Cardiovasc Med 2025;12:1450470.
- 32. Davis HE, Assaf GS, McCorkell L, Wei H, Low RJ, Re’em Y, et al. Characterizing long COVID in an international cohort: 7 months of symptoms and their impact. EClinicalMedicine 2021;38:101019.
- 33. Li X, Hu Y, He B, Li L, Tian Y, Xiao Y, et al. Design, synthesis and evaluation of ursodeoxycholic acid-cinnamic acid hybrids as potential anti-inflammatory agents by inhibiting Akt/NF-κB and MAPK signaling pathways. Eur J Med Chem 2023;260:115785.
- 34. Lapenna D, Ciofani G, Festi D, Neri M, Pierdomenico SD, Giamberardino MA, et al. Antioxidant properties of ursodeoxycholic acid. Biochem Pharmacol 2002;64:1661-1667.
- 35. Rodrigues CM, Ma X, Linehan-Stieers C, Fan G, Kren BT, Steer CJ. Ursodeoxycholic acid prevents cytochrome c release in apoptosis by inhibiting mitochondrial membrane depolarization and channel formation. Cell Death Differ 1999;6:842- 854.
- 36. Chen L, Li X, Chen M, Feng Y, Xiong C. The ACE2 expression in human heart indicates new potential mechanism of heart injury among patients infected with SARS-CoV-2. Cardiovasc Res 2020;116:1097-1100.
- 37. Rainer PP, Primessnig U, Harenkamp S, Doleschal B, Wallner M, Fauler G, et al. Bile acids induce arrhythmias in human atrial myocardium--implications for altered serum bile acid composition in patients with atrial fibrillation. Heart 2013;99:1685-1692.
- 38. Moon JH, Jeong S, Jang H, Koo BK, Kim W. Metabolic dysfunction-associated steatotic liver disease increases the risk of incident cardiovascular disease: a nationwide cohort study. EClinicalMedicine 2023;65:102292.
- 39. Miragoli M, Kadir SH, Sheppard MN, Salvarani N, Virta M, Wells S, et al. A protective antiarrhythmic role of ursodeoxycholic acid in an in vitro rat model of the cholestatic fetal heart. Hepatology 2011;54:1282-1292.
- 40. Cho T, Fukunaga S, Ohzono D, Tanaka H, Minami S, Nakane T, et al. Metabolic dysfunction-associated steatotic liver disease is a risk factor for gallstones: a multicenter cohort study. Hepatol Res 2025;55:663-674.
- 41. Chen M, Wang L, Song G. The nonlinear relationship between the hepatic steatosis index and the risk of gallstones: A threshold analysis based on the National Health and Nutrition Examination Survey data. J Int Med Res 2025;53:03000605251384456.
- 42. Kim SH, Arora I, Hsia DS, Knowler WC, LeBlanc E, Mylonakis E, et al. New-onset diabetes after COVID-19. J Clin Endocrinol Metab 2023;108:e1164-e1174.
- 43. Lakić B, Škrbić R, Uletilović S, Mandić-Kovačević N, Grabež M, Šarić MP, et al. Beneficial effects of ursodeoxycholic acid on metabolic parameters and oxidative stress in patients with type 2 diabetes mellitus: a randomized double-blind, placebo-controlled clinical study. J Diabetes Res 2024;2024:4187796.
- 44. You M, Yuan P, Li L, Li B, Peng Z, Xu H. The association between epilepsy and COVID-19: analysis based on Mendelian randomization and FUMA. Front Neurosci 2023;17:1235822.
- 45. Lee YH, Bang H, Park YM, Bae JC, Lee BW, Kang ES, et al. Non-laboratory-based self-assessment screening score for non-alcoholic fatty liver disease: development, validation and comparison with other scores. PLoS One 2014;9:e107584.
- 46. Tsutsumi T, Charlton MR, Oddo VM, Vilt S, Kim HY, Micic D, et al. Aspirin use and Mediterranean diet adherence as factors associated with overall mortality in patients with MASLD in NHANES. Hepatol Res 2026;56:330-338.
- 47. Reis G, Dos Santos Moreira Silva EA, Medeiros Silva DC, Thabane L, Ferreira TS, Reis LLF, et al. The effect of fluvoxamine and metformin for fatigue in patients with long COVID : an adaptive randomized trial. Ann Intern Med 2026;179:621-628.
, Won Kim6,7
