Interleukin-23 (IL-23) plays a pivotal role in the development and function of T helper cells, with Th17 cells emerging as a significant T cell subset involved in the pathogenesis of immune-mediated diseases (Sharma et al., 2022). IL-23 is a heterodimeric cytokine composed of the IL-23-specific p19 subunit (IL-23p19) and the p40 subunit, which is shared with IL-12 (IL-23p40). Its primary sources include myeloid cells such as monocytes, macrophages, and dendritic cells (Krueger et al., 2024). Studies have shown that elevated levels of IL-23 are detectable in the affected skin of patients with psoriasis (PsO) (Kelemen et al., 2022), synovial tissue in patients with psoriatic arthritis (PsA) (Hum et al., 2025), and intestinal mucosa in patients with Crohn’s disease (CD) (Liu et al., 2011).

As a result, selective blockade of the IL-23 pathway has become an effective therapeutic strategy for these inflammatory diseases (Ghoreschi et al., 2021; García-Domínguez, 2025). Therapeutic monoclonal antibodies targeting the IL-23p19 subunit, such as guselkumab and risankizumab, have demonstrated efficacy in a variety of conditions including PsO (Yang et al., 2021), PsA (Mohanakrishnan et al., 2022), CD (Fanizza et al., 2023), and ulcerative colitis (UC) (Saab et al., 2025). Risankizumab (brand name SKYRIZI®) was approved by the U.S. FDA in April 2019 for the treatment of moderate-to-severe PsO and was later approved in January 2022 and June 2022 for active PsA and moderate-to-severe active CD (Pang et al., 2024), respectively. In June 2024, it was further approved for the treatment of moderate-to-severe active UC in adult patients (Johnson and Loftus, 2024). Guselkumab (brand name TREMFYA®) has also received similar indications (US Food and Drug Administration, 2025), and these two drugs have become representative therapies within the IL-23 inhibitor class.

Although risankizumab and guselkumab share the same core mechanism of action (Bourgonje et al., 2025), which involves binding specifically to the IL-23p19 subunit to block its interaction with the receptor, they possess distinct molecular structures. Guselkumab is a fully human IgG1λ monoclonal antibody with a natural Fc region, whereas risankizumab’s Fc region is engineered with L234A and L235A mutations (known as the “LALA” substitution) (Cross, 2025). This modification significantly weakens its ability to bind to Fcγ receptors (such as CD64). A series of in vitro studies have shown that guselkumab, with its natural Fc region, can bind to CD64⁺ myeloid cells (the primary source of IL-23 in PsA) and capture IL-23 secreted by these cells. This dual mechanism of action has been described as a potential “bifunctional” effect (Doggrell, 2025). In contrast, the “LALA” modification of risankizumab precludes this function. Further experiments have demonstrated that guselkumab exhibits enhanced Fc-dependent efficacy in inhibiting IL-23 signaling (Sachen et al., 2025). However, the long-term clinical implications of these in vitro findings remain uncertain, and further research is required to determine whether these molecular differences have a practical impact on clinical outcomes, including efficacy and safety.

To date, while the safety profiles of IL-23 inhibitors relative to older biologics (e.g., tumor necrosis factor-α inhibitors) have been studied, there remains a lack of direct comparative pharmacovigilance analyses specifically between risankizumab and guselkumab in real-world practice. Given their structural differences, distinguishing their specific safety signals is clinically relevant. Therefore, pharmacovigilance analysis using large-scale real-world data is crucial for identifying and comparing the safety profiles of these two drugs across diverse populations. This study aims to conduct a comprehensive comparative analysis of post-marketing adverse event (AE) reports for risankizumab and guselkumab using the FDA Adverse Event Reporting System (FAERS) database, providing valuable, data-driven evidence to guide therapeutic choices for patients with immune-mediated inflammatory diseases.

This study, based on real-world data from the FAERS database, presents the first comprehensive and systematic comparative analysis of the AE profiles of two important IL-23 inhibitors: guselkumab and risankizumab. Our findings reveal significant differences between the two drugs in terms of baseline characteristics, indication distribution, AE signal categories, key risk signals, and event timing patterns, providing new evidence for clinical safety monitoring and risk management.

A core finding of this study is that the primary reported events for guselkumab were not pharmacologically related AEs, but rather a series of drug administration and management-related issues, such as product dose omission and product storage error. This phenomenon is strongly linked to its primary indications (e.g., PsO), a chronic disease requiring long-term self-injection by patients. These management-related events coexist with high-frequency reports of therapeutic product effect decreased. This contrasts with controlled clinical settings like the VOYAGE one trial, where strict protocols minimized administration errors, though mild injection site reactions were still reported (Blauvelt et al., 2017). Therefore, our findings suggest that in a real-world setting, improper drug administration and poor adherence may be key factors contributing to suboptimal clinical outcomes, rather than a decrease in the drug’s efficacy. This contrasts with risankizumab, which primarily presented signals linked to events with clear pathophysiological significance. To address the interference from administration-related events for guselkumab, we stratified our analysis by specific SOCs (e.g., neoplasms and infections) to specifically compare the pharmacological safety profiles of the two agents.

Regarding cancer risk, the disproportionality analysis revealed notable signal differentiation. Risankizumab showed strong associations with non-melanoma skin cancers (NMSC) (Fahradyan et al., 2017), particularly basal cell carcinoma and squamous cell carcinoma of the skin, especially in patients aged 65 and older. This aligns with safety data from pivotal trials in PsO (UltIMMa-1/2) and UC (COMMAND), which also reported NMSC and other malignancies, though these events remained rare and comparable to placebo (Gordon et al., 2018; Louis et al., 2024). While PsO patients inherently have a higher baseline risk for NMSC (Wang et al., 2020) and existing clinical trial data have not confirmed an increased risk (Gordon et al., 2024), the signal in the FAERS database, notable for both its high reporting frequency and strong statistical disproportionality, suggests this potential association with NMSC warrants attention in long-term post-market surveillance. In contrast, guselkumab did not show significant skin cancer signals, but ROR analysis identified a rare signal for angioimmunoblastic T-cell lymphoma, a hematologic malignancy (ROR = 13.84, IC = 3.77). Notably, our FAERS findings of specific malignancies align with observations from the VOYAGE 1/2 (PsO) and QUASAR (UC) trials, which documented sporadic cases of basal cell carcinoma, prostate cancer, and breast cancer in treatment arms (Blauvelt et al., 2017; Reich et al., 2017b; Peyrin-Biroulet et al., 2023). This consistency between trial case reports and real-world signals suggests that, while the absolute risk remains low, vigilance regarding malignancy is warranted. The differential signals imply that the potential carcinogenic risks of the two drugs may involve different targets or mechanisms, which warranting further clarification through future epidemiological studies.

Regarding infection risk, both drugs exhibited signals consistent with the known class effects of biologics (Ouranos et al., 2025). Both drugs showed associations with tuberculosis. Notably, risankizumab displayed a particularly strong signal for mycobacterium tuberculosis complex test positive (ROR = 17.45, IC = 4.02), while guselkumab was associated with latent tuberculosis. However, triangulation with Phase 3 trials suggests that these signals may reflect screening bias rather than true toxicity. In particular, serious infection rates in the GALAXI (CD) and QUASAR (UC) studies were comparable to or even lower than placebo, likely due to better disease control (Peyrin-Biroulet et al., 2023; Panaccione et al., 2025). This highlights the importance of rigorous tuberculosis screening before initiating IL-23 inhibitor therapy. Although retrospective studies have shown that IL-23 inhibitors do not increase the risk of reactivation of asymptomatic latent tuberculosis infection (LTBI) (Torres et al., 2024), a recent population-based TriNetX analysis indicated a higher incidence of tuberculosis in patients treated with IL-23 inhibitors compared to the general population (Woodie et al., 2025). Additionally, COVID-19 was a high-frequency reported event for both drugs, but the signal strength (ROR) was not significant, suggesting that the signal likely reflects the global pandemic context during the analysis period rather than a significant drug-specific risk increase.

This study also identified and summarized potential new safety signals beyond the drug labels. For risankizumab, we found potential associations with various serious cardiovascular and cerebrovascular events (e.g., cerebrovascular accident, myocardial infarction), gastrointestinal events (e.g., intestinal obstruction), and malignancies (e.g., skin cancer, colon cancer). Although early Phase 2 trials did not report major cardiovascular events (Feagan et al., 2018), recent case reports (Polat Ekinci et al., 2025) have shown that two PsO patients developed cerebrovascular accidents following risankizumab treatment, and other FAERS analyses have suggested a potential increase in cardiovascular event risk (Egeberg and Thyssen, 2023; Woods, 2023). Given that such trials are often underpowered to detect rare events, further long-term observational data are needed. For guselkumab, newly identified safety signals mainly focused on a broader category of malignancies (e.g., Hodgkin’s disease, lung adenocarcinoma), severe skin reactions (e.g., pemphigoid, dermatitis exfoliative generalised), and rare autoimmune diseases (e.g., autoimmune thyroiditis, myelitis transverse). These real-world signals are consistent with rare observations from the VOYAGE one and two trials, which documented sporadic cases of malignancies and immune-mediated reactions in treatment arms (Blauvelt et al., 2017; Reich et al., 2017b). Future studies should further validate these new AE signals to elucidate their clinical significance and inform intervention strategies.

The differences in safety signal profiles suggest that the mechanisms of action of the two drugs may not be identical. A key molecular distinction lies in their Fc fragment function. Guselkumab is a fully human IgG1λ antibody with a natural Fc region. Studies have shown that its Fc region can bind to CD64 (FcγRI) on myeloid cells (such as macrophages) (Pugliano et al., 2025), potentially triggering broader immune regulatory effects beyond simple IL-23 blockade. This may explain the more diverse autoimmune and skin-related safety signals observed with guselkumab. In contrast, risankizumab is a humanized IgG1 antibody, with an engineered Fc region (e.g., L234A and L235A mutations) designed to significantly reduce binding to Fcγ receptors (including CD64) and complement C1q, minimizing unnecessary cytotoxic effects. This “silencing” difference in molecular design likely accounts for the distinct safety profiles of the two drugs in real-world clinical use. Our findings emphasize that even monoclonal antibodies targeting the same cytokine, subtle structural differences (particularly in the Fc region) can translate into clinically significant safety differences, which are crucial for guiding clinical decision-making and risk monitoring.

Finally, the time-to-onset analysis, reinforced by survival analysis models, revealed distinct temporal risk profiles for the two drugs. The Cox regression analysis (HR = 1.24) and cumulative incidence curves confirmed that AEs for guselkumab were more likely to occur early in treatment (especially within the first 30 days), suggesting that early risk management (such as patient education and post-injection observation) is particularly important. In contrast, risankizumab’s AE time-to-onset had a longer median, with a relatively gradual and delayed risk profile, indicating that more prolonged monitoring is needed for patients using risankizumab.

This study has several inherent limitations. First, as FAERS is a spontaneous reporting system, it is subject to reporting bias (e.g., selective reporting, underreporting, or duplicate reporting) and data gaps (e.g., incomplete information on clinical details or disease severity). Second, the disproportionality analysis used in this study can only identify signal associations and cannot establish causality. Third, the large disparity in report numbers (particularly the surge in risankizumab reports in 2022) may affect the stability of signal detection. Fourth, the data in this study were predominantly from the United States (risankizumab: 79.55%; guselkumab: 85.13%). As a result, the findings may reflect specific U.S. clinical practices and patient demographics, which could limit the generalizability of the safety profiles to populations outside the United States. Finally, the FAERS database lacks comprehensive clinical data, specifically detailed prior disease history. While our analysis of concomitant medications suggests a comparable baseline between the groups, we were unable to fully standardize patient characteristics due to these missing records. Therefore, although both drugs are prescribed for similar indications, we cannot fully exclude the possibility that the observed safety signals may be influenced by prior treatments or other unmeasured variables. Nevertheless, the strength of this study lies in its large sample size and real-world environment, capturing rare or delayed AE signals that may be difficult to detect in tightly controlled clinical trials, offering valuable perspectives on the safety comparison of the two drugs.

In conclusion, our study demonstrates that although guselkumab and risankizumab both belong to the IL-23 inhibitor class, their safety signal profiles differ significantly in the real world. These differences may stem from factors such as distinct indication populations, report sources, and potential drug-specific effects. The multiple signals of concern identified in this study, particularly risankizumab’s cancer signals and guselkumab’s efficacy-related and emerging autoimmune signals, provide specific directions for clinical risk management and underscore the need for continued post-market pharmacovigilance. Future research should combine multi-source data, such as electronic health records, to further validate and conduct causal inference on these signals.

Our study provides real-world confirmation of known AEs associated with IL-23 inhibitors, while also identifying distinct safety profiles that necessitate drug-specific risk management. Specifically, risankizumab exhibited stronger signals for severe organic pathologies, including skin cancer and gastrointestinal obstruction. Consequently, clinicians should exercise heightened vigilance when prescribing risankizumab, particularly for elderly patients (≥65 years), who demonstrated significantly elevated risks for these specific AEs.

In contrast, guselkumab presented a relatively more favorable safety profile regarding organic lesions but was predominantly associated with medication management challenges, such as dose omission and accidental exposure. Therefore, the clinical focus for guselkumab should prioritize patient education on proper administration to ensure adherence and avoid potential therapeutic compromises.

These findings underscore the importance of individualized therapy. Factors such as patient age, comorbidities, and the specific safety signals identified in this study should guide treatment selection. As real-world evidence continues to accumulate, these insights will aid in optimizing treatment strategies, maximizing patient benefits, and minimizing drug-specific AEs.