Tumor necrosis factor-α (TNF-α) inhibitors have become a cornerstone in the management of immune-mediated diseases. These agents exert their effects by antagonizing TNF-α, a pro-inflammatory cytokine that plays a central role in immune activation and inflammation. TNF-α inhibitors have been widely utilized in rheumatology, gastroenterology, and dermatology by alleviating inflammation and improving disease symptoms (1). To date, adalimumab, etanercept, infliximab, certolizumab pegol, and golimumab have been approved by the U.S. Food and Drug Administration (FDA) for use in humans (2). These biologics are indicated for a broad spectrum of immune-mediated conditions, including rheumatoid arthritis, Crohn’s disease, ulcerative colitis, and psoriasis (3–6). According to current guidelines from the European League Against Rheumatism (EULAR) and the American College of Rheumatology (ACR), TNF-α inhibitors are recommended as first-line biologic agents for patients inadequately controlled by conventional therapies (7, 8).

Despite their clinical efficacy, TNF-α inhibitors are associated with a range of adverse effects. These adverse events (AEs) span a wide spectrum, from mild injection-site reactions to serious infections and malignancies (9). Rare but severe AEs include cardiovascular events, intestinal obstruction, immune deficiencies, and hepatobiliary disorders (10).

Among these, hepatobiliary disorders constitute a rare but clinically important category of AEs. Several studies and case reports have described hepatobiliary AEs, including drug-induced hepatitis, autoimmune hepatitis, and cholestasis (11–14). For instance, infliximab has been linked to severe liver injury requiring liver transplantation, while adalimumab and etanercept have been associated with elevated liver enzymes. Case series and a pharmacovigilance database study have reported autoimmune hepatitis and other hepatobiliary events with anti-TNFα agents (14, 15); a systematic review has also summarized biologic-related liver injury across agents (13), but head-to-head comparisons among TNF inhibitors remain scarce. Moreover, whether hepatobiliary reporting patterns are consistent across different spontaneous reporting systems (e.g., FAERS vs JADER) remains unclear.

Therefore, this study aimed to systematically evaluate hepatobiliary AEs associated with TNF-α inhibitors using the FAERS (2004–2024) and JADER (2004–2025) databases, providing a cross-database hepatobiliary disproportionality comparison across TNF-α inhibitors, and to explore potential risk factors for the hepatobiliary AE signal-positive TNF-α inhibitor. The findings may help inform clinical decision-making and support risk–benefit assessments for TNF-α inhibitor therapy.

In our study, a pharmacovigilance analysis of hepatobiliary AEs associated with TNF-α inhibitors was performed using real-world data from FAERS and JADER. Disproportionality analysis revealed a significant signal between adalimumab and hepatobiliary AEs. We further characterized the clinical features of these events and developed a logistic regression model to detect risk factors for hepatobiliary disorders related to adalimumab. The disproportionality analysis showed notable variability in hepatobiliary risk among different TNF-α inhibitors. Adalimumab exhibited the highest risk in both two databases. Golimumab showed a positive signal only in FAERS. According to ROR values, the hepatobiliary AE risk ranking among TNF-α inhibitors was adalimumab, golimumab, certolizumab pegol, infliximab and etanercept in FAERS. In contrast, the rank in JADER was adalimumab, certolizumab pegol, etanercept, infliximab, golimumab. This difference may be related to drug selection preferences in different regions. The dual-database analysis using FAERS and JADER provides complementary validation. Similar dual-database approaches have been adopted in previous pharmacovigilance studies using FAERS and JADER. (24, 25).

The potential mechanisms underlying hepatobiliary AEs induced by TNF-α inhibitors have not yet been fully elucidated. Current research suggests that these events may be associated with immune dysregulation and enhanced autoimmune responses. Although speculative, potential mechanisms proposed in literature include disruption of Treg homeostasis. TNF-α is crucial for maintaining regulatory T cells (Tregs), a subset of CD4+ T lymphocytes. Tregs suppress excessive immune activity and preserve immune homeostasis (26). TNF-α promotes the survival and function of Tregs through TNFR2 receptor signaling (27). Inhibition of TNF-α may disrupt this pathway, resulting in decreased Treg numbers or impaired function, which in turn may trigger immune imbalance and contribute to hepatobiliary injury. Additionally, anti-TNF-α therapy may impair the clearance of autoreactive B cells and inhibit apoptosis of cytotoxic CD8+ T cells, thereby facilitating the production of autoantibodies such as ANA, ASMA, Anti-LKM, Anti-ds-DNA, etc. (28). Clinical case reports have documented the emergence of these autoantibodies following infliximab therapy, with some patients meeting diagnostic criteria for autoimmune hepatitis (AIH) (15, 29). These potential mechanisms are speculative and warrant further experimental validation.

To further investigate specific risk factors, a focused analysis was conducted on adalimumab-related hepatobiliary disorders. Logistic regression analysis identified several significant risk factors of hepatobiliary AEs, including patient sex, drug dosage, patients’ age, weight and six categories of comorbidities: infections and infestations, surgical and medical procedures, skin and subcutaneous tissue disorders, general disorders and administration site conditions, congenital, familial and genetic disorders, and renal and urinary disorders. Among these, skin and subcutaneous tissue disorders was identified as a common factor in the risk factor analysis of two datasets, indicating a strong correlation with hepatobiliary disorders. Notably, renal and urinary disorders presented an opposite signal (β<0), suggesting inverse association of hepatobiliary disorders related to adalimumab. This signal may be related to the order of variable inclusion, and its clinical significance remains to be verified. Some differences in factors may be attributed to varying baseline data of the populations represented by the two databases, for example, the population of greater weight cases in FAERS is higher than that in JADER.

In Table 3, analyses based on logistic regression stratified by database (FAERS vs JADER) showed that, although females accounted for the majority of adalimumab users in both the databases (66.27% in FAERS and 60.86% in JADER), males were found to have a higher association of developing hepatobiliary disorders (OR = 1.225 in FAERS; OR = 2.378 in JADER). This increased risk in males is speculated to be associated with differences in hormonal levels or immune responses. Estrogen is believed to exert hepatoprotective effects by regulating lipid metabolism, inhibiting inflammation (30, 31). During reproductive age, the prevalence and severity of non-alcoholic fatty liver disease (NAFLD) tend to be higher in men than in women. However, after menopause, women exhibit a higher incidence of NAFLD than age-matched men, suggesting that estrogen plays a protective role in liver health (32). Moreover, Kupffer cells in male livers exhibit a more pronounced response to hepatic injury signals. Men infected with hepatitis B virus (HBV) experience more frequent and severe liver damage during viral reactivation compared to women (33). The biological explanations discussed remain hypotheses rather than direct evidence.

Psoriasis and related Th17-driven dermatoses frequently co-occur with NAFLD and other hepatic conditions, reflecting shared systemic inflammation (IL-17/IL-23/TNF-α axes) (34). Against this background, TNF-α blockade with adalimumab may perturb immune homeostasis in the liver. In patients with latent or established liver disease, therapy may function as a second hit superimposed on underlying pathology, thereby bringing liver injury to clinical attention (35). While our pharmacovigilance design cannot infer incidence or causation, this pathobiological rationale is consistent with the hepatobiliary disproportionality signal observed and underscores the value of baseline and early monitoring in at-risk patients.

The FAERS and JADER databases are both spontaneous reporting systems. Their data are prone to selection bias from factors like race, geography, drug approval times, clinical usage rates, AE awareness, and lack of reporting of some AEs that actually occurred. In our collected cases, gender, age, and weight data have varying degrees of missingness, with weight data missing the most (65.62% in FAERS and 43.03% in JADER). Importantly, neither database provides incidence rates and is subject to reporting bias, indication bias, and underreporting. Despite these unavoidable limitations may affect disproportionality signals and risk-factor estimates, the consistent patterns observed may help generate more specific hypotheses, and inform the design of future prospective cohort studies or other real-world data analyses.

In summary, among the five TNF-α inhibitors evaluated in this study, adalimumab showed a disproportionality association of hepatobiliary AEs. The other TNF-α inhibitors did not show consistent disproportionality at the hepatobiliary SOC level. Male sex and skin and subcutaneous tissue disorders were regarded as consistent independent risk factors in both datasets. High drug dosage, high age, high weight, and the presence of an additional six comorbidities could also require attention in clinics because these factors were obviously observed in certain single databases. Our findings suggest that adalimumab therapy requires monitoring hepatobiliary function, assessment of individualized risks (particularly gender-related risks), and special attention to screening for skin and subcutaneous tissue disorders.