Adverse reactions of immune checkpoint inhibitors combined with acetaminophen/non-steroidal anti-inflammatory drugs (NSAIDs): a pharmacovigilance analysis of drug–drug interactions

1 Introduction

The advent of immune checkpoint inhibitors (ICIs), including antibodies targeting cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), programmed cell death protein 1 (PD-1), and its ligand (PD-L1), has fundamentally transformed the therapeutic landscape for a broad spectrum of malignancies (Michielin et al., 2022). By reinvigorating exhausted T cells and overcoming tumor-mediated immune suppression, ICIs have exhibited unprecedented and durable clinical responses across diverse cancer types, facilitating their rapid integration into standard-of-care regimens—both as monotherapy and in combination with other treatment modalities (Ribas and Wolchok, 2018). This widespread andoften prolonged use underscores the critical need for a comprehensive understanding of their safety profile and potential interactions with concomitant medications.

Concurrently, acetaminophen (APAP) and non-steroidal anti-inflammatory drugs (NSAIDs) remain among the most frequently prescribed medications globally for the management of cancer-related pain, inflammation, and fever. Their analgesic and anti-inflammatory effects, primarily mediated through the inhibition of cyclooxygenase (COX) enzymes, render them a cornerstone of supportive care in oncology (Khalil et al., 2024). As a result, the co-administration of APAP/NSAIDs with ICI therapy is a commonplace clinical scenario, fueled by the need to alleviate symptoms and enhance patients’ quality of life during cancer treatment (Sohi et al., 2021).

However, this frequent co-administration raises important pharmacological concerns. Both ICIs and APAP/NSAIDs profoundly impact the immune system, albeit through distinct mechanisms. ICIs induce a systemic, pro-inflammatory state essential for antitumor immunity but also predispose patients to immune-related adverse events (irAEs). In contrast, the cyclooxygenase pathways, inhibited by NSAIDs, have been implicated in promoting tumor growth, progression, and immunosuppression (Zelenay et al., 2015; Bonavita et al., 2020). The potential pharmacokinetic and, more critically, pharmacodynamic interactions between these two drug classes could theoretically alter the efficacy or toxicity profiles of either agent (Sharma and Allison, 2020).

The concomitant use of multiple drugs can lead to drug–drug interactions (DDIs), which can alter the efficacy and safety profiles of medications, and may consequently result in toxicities and treatment failure (Blower et al., 2005). However, there is a paucity of research investigating DDIs focused specifically on the co-administration of ICIs and APAP/NSAIDs. Existing research on the co-administration of ICIs and APAP/NSAIDs has primarily focused on the clinical outcomes, yielding inconsistent and complex findings (Stone et al., 2026; Zhang et al., 2021; Mao et al., 2022; Li et al., 2026). However, research addressing their safety remains limited, with only three relevant studies reported to date. One study, which performed a multivariate regression analysis from the FAERS database, found that aspirin use was associated with an increased risk of irAEs across a pan-cancer setting (Yang et al., 2023). Notably, low-dose aspirin is currently primarily used for cardiovascular protection. Unlike other NSAIDs, it is not clinically employed for antipyretic or analgesic purposes. Additionally, two studies have confirmed an association between kidney injury and NSAID use in the context of ICI therapy. One meta-analysis investigating the risk factors of acute kidney injury (AKI) associated with ICIs identified NSAID use as one of the significant contributing factors (Liu et al., 2023). Another pharmacovigilance study, which analyzed the World Health Organization (WHO) international pharmacovigilance database Vigibase, demonstrated an increased reporting of nephritis in patients receiving both NSAIDs and ICIs (Laurent et al., 2025). However, neither of these studies employed methodologies specifically designed to evaluate DDIs. In general, the aforementioned studies only described simple correlation phenomena without rigorous verification of interaction effects, resulting in a weak clinical base for the safe concurrent medication of ICIs and anti-inflammatory analgesics. Currently, large-scale, standardized evidence regarding DDIs involving the co-administration of ICIs and APAP/NSAIDs remains insufficient and has not been comprehensively summarized in the existing literature. Against this background, our study further supplements real-world pharmacovigilance evidence in this field and strengthens the research positioning in the existing ICI-DDI evidence system.

To better understand the potential DDIs between ICIs and APAP/NSAIDs, it is essential to gather real-world data. We conducted descriptive and disproportionality analyses using the United States Food and Drug Administration (FDA) Adverse Event Reporting System (FAERS) database to investigate the potential harmful effects of these co-administrations. Based on real-world spontaneous reporting big data, our study characterized disproportional signals of ICI-associated analgesic interactions, described the clinical features and distribution patterns of adverse events, and provided preliminary reference evidence for rational clinical medication and risk awareness.

2 Methods

2.1 Data source

The FAERS database was used for this retrospective pharmacovigilance study (U.S. Food and Drug Administration, 2021). The raw data can be obtained from the FAERS database at the following link: FAERS Quarterly Data Extract Files (fda.gov). The FAERS database includes adverse event (AE) reports from healthcare professionals, patients, and manufacturers worldwide, and the data are publicly accessible. Structurally, FAERS consists of seven standardized core data tables, including patient demographic information (DEMO), drug administration records (DRUG), adverse event terms (REAC), patient outcomes (OUTC), report sources (RPSR), therapy duration information (THER) and indicate information (INDI), which enables integrated retrieval of medication exposure, combined medication information and corresponding adverse reactions. To eliminate duplicate reports, the most recent FDA_DT was selected when the CASEID was identical. If both CASEID and FDA_DT were identical, the PRIMARY_ID with the higher value was selected, following the FAERS user instructions (U.S. Food and Drug Administration, 2021). Data were extracted for the period from the first quarter (Q1) of 2014 to the second quarter (Q2) of 2025.

2.2 Procedures

Data were obtained from the REAC files according to the Medical Dictionary for Regulatory Activities (MedDRA, version 23.0) at the preferred term (PT) level, based on all 27 Standardized MedDRA Query System Organ Class (SOC) (Maintenance and support service organization). The 27 SOCs are listed in Supplementary Table S1. The DRUG file contained records of generic drug names, with the role_code selected as either ‘PS’ (primary suspect) or ‘SS’ (secondary suspect). Combined immunotherapy drugs included anti-PD-1 agents (pembrolizumab, nivolumab, cemiplimab, dostarlimab, prolongolimab, tislelizumab, toripalimab, sintilimab, camrelizumab, penpulimab, zimberelimab, and serplulimab), anti-PD-L1 agents (atezolizumab, durvalumab, avelumab, envafolimab, and sugemalimab), and anti-CTLA4 agents (ipilimumab and tremelimumab). APAP/NSAIDs included acetaminophen, non-selective COX inhibitors, and selective COX-2 inhibitors, comprising a total of 33 medications (Supplementary Table S2). The inclusion criteria were that the role_code (PS) included both ICIs and either APAP or NSAIDs, with clear records of adverse events. The exclusion criteria were reports with incomplete key information and duplicate reports. Of note, three-step data cleaning was conducted before analysis (Figure 1). First, duplicates and multiple cases or reports were removed, and only the most recent version of each report without duplication was extracted. Second, a drug event combination of ICIs and APAP/NSAIDs was established, and only those AE reports with reported roles of drugs as “suspect” were included. Finally, based on different therapeutic regimens, total reports were divided into four subgroups: (1). No ICIs, no APAP/NSAIDS, (2). APAP/NSAIDS only, no ICIs, (3). ICIs only, no APAP/NSAIDS, (4). ICIs combined with APAP/NSAIDS. Descriptive analyses were conducted to synthesize the clinical features of patients receiving ICIs in combination with APAP/NSAIDs using data from the FAERS database.

2.3 Data analysis

In spontaneous reporting databases, the detection of potential drug interactions relies on the observation that a suspected adverse event (AE) occurs more frequently when two drugs are used in combination than when they are used individually (van Puijenbroek et al., 2000; Yue et al., 2014; Capogrosso Sansone et al., 2017). To analyze this phenomenon, the reports were categorized into three index groups: (i) patients exposed to ICIs alone, without APAP or NSAIDs; (ii) patients exposed to APAP/NSAIDs alone, without ICIs; and (iii) patients exposed to both ICIs and APAP/NSAIDs at the time of the event. The reference group comprised patients who were not exposed to either ICIs or APAP/NSAIDs.

Disproportionality was assessed using the reporting odds ratio (ROR), along with the corresponding 95% confidence interval (95% CI). A result was considered statistically significant when the lower bound of the 95% CI exceeded 1, provided that at least three cases of interest were reported (van Puijenbroek et al., 2002). To control for potential confounders, crude RORs were recalculated using unconditional logistic regression, adjusting for age, sex, and reporting year (Noguchi et al., 2019). The logistic regression model used was as follows: Log(risk of the event) = β0 + β1 (ICIs) + β2 (APAP/NSAIDs) + β3 (ICIs × APAP/NSAIDs) + β4 (age) + β5 (gender) + β6 (reporting year). An interaction was considered significant if the combination of drugs was associated with an enhanced ROR compared to the other index groups (Thakrar et al., 2007).

Various methods are available for detecting DDI signals. The logistic regression model is a statistical method used to identify DDI signals from a spontaneous reporting system. The ROR, similar to the odds ratio, is a statistical model that can be adjusted for age and gender using logistic regression. The additive and multiplicative models are also used to detect potential DDI signals. In the additive model, drug-related risks increase additively, whereas in the multiplicative model, they increase synergistically (Noguchi et al., 2019). Given that no universally accepted statistical method exists for detecting DDI signals from spontaneous reporting data, we employed the adjusted ROR and additive/multiplicative models for signal detection.

To evaluate the consistency and reliability of the detected DDIs, both the multiplicative and additive models were used (Noguchi et al., 2019). The analysis provided a threshold for detecting DDI signals. In the multiplicative model, the risk associated with a drug is multiplied by the background risk, while in the additive model, the risk associated with the drug is added to the background risk. In the multiplicative model, a positive interaction is indicated when: The risk (drug 1 × drug 2)/((risk (drug 1) × risk (drug 2)) >1. In the additive model, a positive interaction is indicated when: risk (drug 1 × drug 2) – (risk (drug 1) + risk (drug 2)) >0. If the interaction term exceeds 0 in the additive model or one in the multiplicative model, it suggests a positive drug interaction (Noguchi et al., 2020). Data management and analyses were conducted using SPSS (version 22.0; IBM Corp., Armonk, NY, United States of America).

3 Results

3.1 Descriptive analysis

From January 2014 to June 2025, a total of 58, 442, 079 reports were extracted from the FAERS database, of which 47, 592, 909 reports were included in the final analysis. Among these, 4,613 reports involved the co-administration of ICIs and APAP/NSAIDs. The clinical characteristics of the patients are summarized in Table 1. The median age of patients was 64 years (IQR: 53–72 years). The proportion of male patients (47.56%) was slightly higher than that of female patients (44.00%). Healthcare professionals submitted a significant proportion of the reports (43.67%). Europe accounted for the highest rate of adverse drug reactions (ADRs) (49.00%), followed by Asia (23.78%), North America (23.00%), Oceania (1.44%), and Africa (0.11%). The number of reports steadily increased over time, from 0.56% in 2014 to 19.56% in 2024.

TABLE 1. Characteristics of patients treated with immune checkpoint inhibitors combined with acetaminophen/non-steroidal anti-inflammatory drugs (NSAIDs) in the FAERS database (1 January 2014, to 30 June 2025).

3.2 Disproportionality analysis

The signal detection results for DDIs associated with theco-administration of ICIs and APAP/NSAIDs across all 27 SOCs are presented in Table 2. The crude and adjusted RORs and 95% CIs for all comparisons are provided using a disproportionality method. DDIs were estimated using both additive and multiplicative models. Positive signals were identified in four SOCs: gastrointestinal disorders, investigations, renal and urinary disorders, and skin and subcutaneous tissue disorders (Table 3). For gastrointestinal disorders, the adjusted ROR for APAP/NSAIDs alone was 1.28 (95% CI: 1.28–1.29), and for ICIs alone, it was 1.24 (95% CI: 1.23–1.25). A signal increase was observed with the concomitant use of ICIs and APAP/NSAIDs (adjusted ROR: 1.57, 95% CI: 1.41–1.70). This positive interaction signal was further supported by the additive model (0.0549). For investigations, the adjusted ROR for APAP/NSAIDs alone was 1.05 (95% CI: 1.04–1.06), and for ICIs alone, it was 1.07 (95% CI: 1.06–1.08). An increase in the signal was detected with the concomitant use of ICIs and APAP/NSAIDs (adjusted ROR: 1.27, 95% CI: 1.13–1.43). The positive interaction signal was supported by both the additive (0.3847) and multiplicative (1.3135) models. For renal and urinary disorders, the adjusted ROR for APAP/NSAIDs alone was 1.07 (95% CI: 1.06–1.09), and for ICIs alone, it was 1.50 (95% CI: 1.48–1.53). The signal for concomitant use of ICIs and APAP/NSAIDs showed a notable increase (adjusted ROR: 4.68, 95% CI: 4.19–5.24). The positive interaction was supported by both the additive (3.5771) and multiplicative (3.6312) models. For skin and subcutaneous tissue disorders, the adjusted ROR for APAP/NSAIDs alone was 0.99 (95% CI: 0.99–1.00), and for ICIs alone, it was 1.02 (95% CI: 1.01–1.04). An increase in the signal was observed for the concomitant use of ICIs and APAP/NSAIDs (adjusted ROR: 1.64, 95% CI: 1.47–1.83). This positive interaction was supported by both the additive (0.6818) and multiplicative (1.5996) models.

TABLE 2. Reporting odds ratios (RORs) and drug interaction approaches for immune checkpoint inhibitors (ICIs) and acetaminophen/non-steroidal anti-inflammatory drugs (NSAIDs).

Bold values indicate positive signals.

TABLE 3. Summary of methods used to analyze drug-drug interactions for all SOCs.

SOC, with positive signal.

3.3 Subset data analyses

For the four positive SOCs, we conducted a subset analysis of the co-administration of ICI with acetaminophen, non-selective COX inhibitors, and selective COX-2 inhibitors (Tables 4, 5). For gastrointestinal disorders and investigations, the co-administration of ICIs and APAP revealed positive DDI signals, which were verified using additive and multiplicative models. For renal and urinary disorders, all combinations of APAP/NSAIDs with ICIs exhibited positive DDI signals, which were confirmed via additive and multiplicative models. In skin and subcutaneous tissue disorders, the combination of NSAIDs with ICI revealed positive DDI signals, which were verified using additive and multiplicative models.

TABLE 4. Disproportionality analyses and drug interaction approaches for the various drug combinations.

Bold values indicate positive signals.

TABLE 5. Summary of methods used to analyze drug–drug interactions in four SOCs.

PD-1, programmed death-1 inhibitors; PDL-1, programmed death ligand-1 inhibitors; CTLA-4, cytotoxic T-lymphocyte antigen-4, inhibitors.

Drug combination with positive signal.

4 Discussion

To the best of our knowledge, this is the first pharmacovigilance study to systematically assess DDIs between ICIs and APAP/NSAIDs. Our study provides the most comprehensive real-world data regarding the clinical characteristics and DDIs associated with theco-administration of ICIs and APAP/NSAIDs. Several key findings emerged: (1) DDIs involving the co-administration of ICIs and APAP/NSAIDs occurred in four systems: gastrointestinal disorders, investigations, renal and urinary disorders, and skin and subcutaneous tissue disorders. (2) The co-administration of ICIs and APAP was associated with gastrointestinal disorders and investigations. (3) All co-administrations of APAP/NSAIDs with ICIs were associated with renal and urinary disorders. (4) The co-administrations of NSAIDs with ICIs was associated with skin and subcutaneous tissue disorders.

ICIs, such as anti-PD-1/PD-L1 and anti-CTLA-4 antibodies, act by blocking inhibitory signaling pathways, releasing the “brakes” on T cells, thereby activating adaptive immune responses and inducing anti-tumor effects. However, this process can also trigger immune overactivation and irAEs (Sharma and Allison, 2020). As the number of cancer patients receiving long-term ICI treatments increases, the interaction between these immunomodulatory agents and commonly prescribed medications become increasingly clinically relevant. Drug interactions can be categorized into pharmacodynamic and pharmacokinetic interactions. Pharmacodynamic interactions occur when different drugs affect the same physiological processes, whereas pharmacokinetic interactions involve effects on the absorption, distribution, metabolism, or excretion of a drug (Sharma and Allison, 2020). DDIs can result in either enhanced or reduced therapeutic effects, AEs, or a unique response not caused by either agent alone (Blower et al., 2005). This study focuses on DDIs that lead to increased toxicity associated with the co-administration of ICIs and APAP/NSAIDs.

NSAIDs exert their anti-inflammatory, analgesic, and antipyretic effects by inhibiting cyclooxygenase (COX) enzymes, thereby reducing the synthesis of prostaglandins (e.g., PGE₂) (Khalil et al., 2024). COX enzymes, which are part of the myeloperoxidase family, metabolize arachidonic acid into prostaglandins and thromboxane (Pajiep et al., 2023). Prostaglandins regulate a variety of physiological processes, including tumor development, migration, differentiation, inflammation, and cell adhesion (Chandrasekharan and Simmons, 2004). COX2-mediated PGE₂ synthesis suppresses anti-tumor immunity through multiple mechanisms, such as inhibition of myeloid cell activation, reduced recruitment of natural killer (NK) and dendritic cells (DC), induction of regulatory T-cell (Tregs), and decreased proliferation and effector function of CD4⁺/CD8⁺ T-cell (Figure 2). In animal models, COX inhibition has been shown to overcome tumor immune evasion and synergize with checkpoint blockade (Zelenay et al., 2015; Wang and Dubois, 2006).

APAP, widely regarded as safe and one of the most frequently used drugs in oncology (Zhang et al., 2021), possesses analgesic and antipyretic properties similar to NSAIDs, but unlike NSAIDs, it lacks anti-inflammatory activity (Figure 2) (Zelenay et al., 2015). However, like conventional NSAIDs, it suppresses prostaglandin synthesis (Böttcher et al., 2018). The mechanism by which acetaminophen exerts its analgesic effects is not fully elucidated. Although acetaminophen lacks significant anti-inflammatory activity, it is thought to act as a selective COX-2 inhibitor (Wiffen et al., 2017). Studies have demonstrated that acetaminophen inhibits prostaglandin production in the brain and spleen, which is consistent with COX-2 inhibition (Wiffen et al., 2017). Additionally, acetaminophen may modulate the endocannabinoid system, thereby contributing to its analgesic effects (Jóźwiak-Bebenista and Nowak, 2014). Previous studies have shown that prophylactic administration of acetaminophen can limit vaccine responses in children (Hinz et al., 2008; Högestätt et al., 2005), leading the World Health Organization to recommend against its use before or during vaccination (Prymula et al., 2009).

Emerging evidence suggests that APAP may exert unintended immunomodulatory effects, raising concerns about its potential to compromise anti-tumor immune responses (Falup-Pecurariu et al., 2017). Mechanistic studies have demonstrated that APAP can inhibit T-cell proliferation, suppress interferon-gamma (IFN-γ) production, and promote regulatory T-cell (Treg) expansion—actions that collectively impair cytotoxic immunity (Falup-Pecurariu et al., 2017). A meta-analysis identified a consistent association between APAP administration and worsened clinical outcomes in cancer patients receiving ICIs therapy, with nearly a 30% increased risk of death or disease progression in APAP-exposed individuals (ACIP Vaccine Administration Guidelines for Immunization, 2022).

Studies have also explored the impact of NSAID use on ICI treatment outcomes, yielding mixed results -ranging from improved efficacy and survival in specific clinical contexts to neutral or even detrimental effects (Falup-Pecurariu et al., 2017; Bessede et al., 2022; Lin et al., 2025; Wang et al., 2020). Two meta-analyses confirmed that the concurrent administration of NSAIDs was not significantly associated with prognosis in patients receiving ICI therapy (Zhang et al., 2021; Mao et al., 2022). However, these studies failed to distinguish between APAP and NSAIDs, despite the fact that both are often erroneously sclassified as NSAIDs. From a clinical perspective, there is little practical difference between the two, yet from an immunotherapy standpoint, this failure to differentiate could introduce significant bias.

ICIs stimulate the immune system, thereby enhancing anti-tumor responses (Cortellini et al., 2020). While effective in combating malignancy, this immune stimulation can trigger immune-mediated inflammatory responses, known as irAEs, which can occur in any organ system of the body (Bessede et al., 2022). Some researchers propose that activated effector T-cell attack normal, non-tumor tissues while exerting their anti-tumor activity (Kanai et al., 2021; Wei et al., 2018). T-cell receptor (TCR) sequencing studies have provided supportfor this theory (Pauken et al., 2019; Khan and Gerber, 2020). In patients treated with Ipilimumab, researchers observed greater diversity of CD4⁺ and CD8⁺ T-cell in those with irAE compared to those without significant adverse reactions (Ronen et al., 2022), offering mechanistic insight into organ-specific toxicities.

ICIs can induce a range of kidney-related irAEs, including AKI, which often presents as acute interstitial nephritis (AIN), though glomerular disease and electrolyte disturbances have also been reported (Porciello et al., 2022). The incidence of AKI in patients receiving ICI therapy can be as high as 17%, but ICI-induced AKI is estimated to account for 3%–5% of these cases (Porciello et al., 2022). Risk factors for ICI-AKI include baseline chronic kidney disease (CKD), combination treatment with anti-CTLA-4 and anti-PD-1/PD-L1 agents, prior or concurrent extrarenal irAEs, and the use of drugs associated with AIN (Sanromán et al., 2023; Oh et al., 2017; Herrmann et al., 2025). NSAIDs can bind to renal tubules and trigger immune responses that are normally suppressed, thereby resulting in kidney damage. This synergistic effect, when combined with ICIs, can lead to AKI (Gupta et al., 2021). A meta-analysis identified NSAID administration as a risk factor for AKI in patients receiving ICI (Zhang et al., 2021). A recent study using the WHO’s Vigibase database found that NSAIDs were associated with increased reports of nephritis in the context of ICI therapy, suggesting the need for careful consideration in clinical prescription (Laurent et al., 2025). Our study further confirms that renal toxicity arises from synergistic effects and reveals for the first time the correlation between the co-administration of NSAIDs and ICIs and toxicity in the gastrointestinal, skin, and subcutaneous tissue systems.

Cutaneous irAEs are a common consequence of ICI treatment, as immune system overactivation leadsto dermatologic toxicities (Seethapathy et al., 2019). These cutaneous irAEs are among the most frequent complications of ICI therapy (Cortazar et al., 2020; Perazella and Shirali, 2018). Many present similarly to primary dermatoses and may share features with autoimmune skin disorders (Cortazar et al., 2020; Laurent et al., 2025; Tattersall et al., 2020). Cutaneous irAEs occur in up to 30%–50% of patients receiving ICIs therapy (Perazella and Shirali, 2018; Laurent et al., 2025), though most are mild and can be managed without interrupting immunotherapy. Interestingly, these reactions can mimic a wide range of primary dermatoses, suggesting a T-cell-mediated mechanism common to various skin disorders.

Gastrointestinal irAEs are the second most common following cutaneous involvement and are often the primary reason for discontinuation of immunotherapy (Haanen et al., 2018; Curry et al., 2017; Villadolid and Amin, 2015). Diarrhea is the most frequently reported symptom, often accompanied by abdominal pain, bloating, rectal bleeding, or fever (Lacouture and Sibaud, 2018; Sibaud, 2018). APAP-induced gastrointestinal toxicity is a significant aspect of its safety profile, with reported rates of gastrointestinal AEs ranging from 5.3% to 9.0% (Schneider et al., 2021). While less severe than NSAID-related gastrointestinal damage, APAP can cause dyspepsia, nausea, abdominal pain, and in rare cases, gastric erosions, ulcers, and bleeding (García Rodríguez and Hernández-Díaz, 2001). Risk factors for more severe outcomes include age, a history of peptic ulcers, alcohol consumption, and prolonged fasting (Prescott, 2000).

Our study has certain limitations. First, the FAERS database is a spontaneous reporting system (SRS) with inherent limitations, including underreporting and inconsistent reporting quality. It was not feasible to quantify adverse reaction signals based on the total number of AEs using the collected data; instead, the study aimed to provide a qualitative indicator solely based on the signal intensity between the drugs and the corresponding reactions. Second, compared with clinical trials and cohort studies, SRS data are generally less reliable due to incomplete clinical details, residual confounding factors, and missing denominator data. The identification and reporting of AEs in SRS are subject to less stringent quality control. Third, it is challenging to identify significant risk factors between disorders and drugs due to deficiencies in the documentation of preexisting disorders and comorbidities, which may influence the assessment of disease causality. Furthermore, this study was not restricted to a specific disease, which differs significantly from clinical trials. Fourth, this study used SOCs as the target reactions to encompass all AEs; however, the scope was overly broad, and subgroup analyses were not performed. Fifth, the medications listed in the same report are not entirely equivalent to those actually taken by the patient concurrently, and there may be reporting bias associated with combined medication use. For DDIs, the relative reporting rate of drug-induced AEs may be overestimated if there is a lack of information on either of the two drugs involved (Riveiro-Barciela et al., 2024; Moore and Scheiman, 2018; Cheung and Brain, 2020; Noguchi et al., 2021; Sakaeda et al., 2013). Sixth, calculations, justifications, and power analyses for the selected sample size were not conducted, as the study intended to include all eligible ADRs. Seventh, the FAERS database has inherent reporting biases, including the Weber effect (reporting rates increase in the first 2 years after a drug’s launch and then decline), the ripple effect (reporting of other drugs in the same class is expedited due to the notoriety of one drug), and the masking or cloaking effect (reports of the same adverse event linked to other drugs can suppress signal scores). Finally, data mining revealed imperfect reporting, with inaccuracies and incomplete entries (e.g., drug dosage and duration of use), which may introduce analytical bias. Nevertheless, our findings, despite being limited by the inherent constraints of the FAERS database, shed light on the DDIs between ICIs and APAP/NSAIDs and provide a framework for more rigorous follow-up research to validate these findings.

5 Conclusion

Our study identified four types of DDIs associated with the co-administration of ICIs and APAP/NSAIDs using real-world data from the FAERS. Therefore, close attention should be paid when these two classes of drugs are co-administered. Preclinical trials are required to explore the underlying mechanisms of these interactions, and further well-designed, robust clinical studies are necessary to clarify these relationships, elucidate the risk of DDIs associated with this combination therapy, and provide more granular details.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fphar.2026.1833275/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Olivier Feron, Université catholique de Louvain, Belgium

Reviewed by: Manish Kumar Sharma, MPI Research, United States

Raffaele Giusti, Sapienza University of Rome, Italy