Gastrointestinal toxicity of antibody–drug conjugates: a pharmacovigilance study using the FAERS database

Abstract

Background

Antibody–drug conjugate (ADC) product specifications identify gastrointestinal adverse reactions. Nevertheless, there is a scarcity of comparative studies pertaining to these side effects of similar medications. Special attention is warranted for adverse drug reactions (ADRs) affecting the gastrointestinal system that are inadequately documented in the drug literature.

Aims

Utilizing the U.S. Food and Drug Administration (FDA) Adverse Event Reporting System (FAERS), data mining was conducted to analyze gastrointestinal adverse reactions associated with ADCs. This analysis aimed to provide evidence supporting the safe use of ADCs in medical institutions.

Methods

We utilized the Openvigil 2.1 platform to extract adverse event data reported for each ADC from the FAERS database, covering the period from the drug’s launch until the second quarter of 2024. For data analysis, we employed the reporting odds ratio (ROR) and proportional reporting ratio (PRR) methods.

Results

A total of 23,886 adverse event reports were retrieved, with nine ADCs identified as the primary suspected drugs, including 1,517 reports of gastrointestinal adverse events linked to ADCs. The average patient age was 59.69 years, with a higher prevalence of female patients (919 patients, 60.58%) than male patients (319 patients, 24.32%). The gastrointestinal toxicity intensity, ranked from highest to lowest, was as follows: inotuzumab ozogamicin (IO) with ROR = 11.12 and PRR = 10.68, gemtuzumab ozogamicin (GO) with ROR = 7.87 and PRR = 7.54, polatuzumab vedotin (PV) with ROR = 6.47 and PRR = 6.20, brentuximab vedotin (BV) with ROR = 5.79 and PRR = 5.61, sacituzumab govitecan (SG) with ROR = 5.19 and PRR = 4.61, mirvetuximab soravtansine (MS) with ROR = 4.37 and PRR = 3.80, trastuzumab deruxtecan (TD) with ROR = 4.22 and PRR = 3.63, trastuzumab emtansine (TE) with ROR = 3.93 and PRR = 3.85, and enfortumab vedotin (EV) with ROR = 3.26 and PRR = 3.02. Adverse events resulted in 237 deaths, 43 life-threatening cases, and 439 initial or prolonged hospitalizations, with TD being the top ranking for deaths and hospitalizations, followed by TE, which presented the highest mortality rate due to adverse events. The most frequent adverse events were nausea (506 cases), diarrhea (262 cases), vomiting (216 cases), ascites (112 cases), colitis (90 cases), pancreatitis (52 cases), and intestinal obstruction (37 cases).

Conclusions

ADCs may increase the risk of gastrointestinal adverse events and thus require vigilant monitoring in clinical practice.

Introduction

Antibody–drug conjugates (ADCs) are a category of therapeutics created through the chemical linkage of monoclonal antibodies, linkers, and cytotoxic payloads [1]. ADCs facilitate the targeted and effective delivery of cytotoxic agents directly to cancer cells by selectively attaching to surface antigens on tumor cells with monoclonal antibodies. Compared with traditional anticancer therapies, ADCs demonstrate a significantly heightened level of selectivity in their targeting, which can enhance the effectiveness of cancer treatments while reducing the side effects associated with cytotoxic medications. This characteristic has contributed to their growing popularity in the realm of cancer treatment in recent years [2]. The U.S. Food and Drug Administration (FDA) has granted approval for 15 such drugs since the introduction of the first ADC in 2000. At this stage, more than 100 ADCs are undergoing various phases of clinical trials, suggesting strong market potential. The incidence of adverse events associated with ADCs is increasing annually with the expansion of their clinical applications globally. The “Expert Consensus on the Clinical Application of ADCs for Malignant Tumor Treatment (2020 Edition)” specifically emphasizes that ADCs can provoke a variety of systemic adverse effects during treatment, primarily impacting the hematologic, gastrointestinal, nervous, and cardiovascular systems, as well as leading to infections [3–4].

Adverse gastrointestinal reactions triggered by ADCs can negatively impact both treatment outcomes and overall quality of life for individuals with cancer, primarily manifesting as symptoms such as abdominal distension, digestive tract ulcers, bleeding, and infections [5]. Although some ADC product specifications recognize gastrointestinal adverse reactions, comparative studies on similar medications in terms of these side effects are lacking. Therefore, we utilized the Openvigil 2.1 tool to detect risk signals related to gastrointestinal adverse events within the ADC category sourced from the FAERS database. Adverse drug reactions (ADRs) affecting the gastrointestinal system that are not documented in the drug literature require special attention. This approach aims to create a comprehensive data reference to support the rational and safe clinical use of these medications.

Methods

Data source

The FAERS database serves as an accessible repository for pharmacovigilance, featuring millions of reports on adverse events contributed by healthcare professionals, staff from pharmaceutical companies, and various stakeholders [6]. This study utilized the OpenVigil 2.1 platform to analyze gastrointestinal adverse events (AEs) associated with each ADC in the FAERS database, covering the period from the drug’s launch until the second quarter of 2024. The pharmacovigilance tool OpenVigil 2.1 enables researchers to retrieve, process, and analyze structured information regarding adverse events from the FAERS database. This tool converts the large raw dataset from FAERS into a format that is better suited for further extraction and analysis [7]. The refined data conform to the definitions of adverse event codes outlined in the MedDRA version 27.0 FAERS database. Throughout this study, standard MedDRA version 27.0 analysis queries were employed for all searches conducted via the OpenVigil 2.1 tool. In this study, we examined various parameters, such as the sex and age of patients, the country of reporting, indications, events, and outcomes. Nevertheless, the differences in patient ethnicity and regional representation could influence the generalizability of the findings [8].

Data filtering

Some pharmaceutical drugs have been available on the market for a relatively short period, which has resulted in a reduced volume of reported data within the FDA Adverse Event Reporting System (FAERS) database. For the purpose of this study, nine FDA-approved ADCs were specifically chosen for comprehensive evaluation through meticulous data screening. The selected ADCs include brentuximab vedotin (BV), enfortumab vedotin (EV), gemtuzumab ozogamicin (GO), inotuzumab ozogamicin (IO), mirvetuximab soravtansine (MS), polatuzumab vedotin (PV), sacituzumab govitecan (SG), trastuzumab deruxtecan (TD), and trastuzumab emtansine (TE). The Medical Dictionary for Regulatory Activities (MedDRA) provides a structured framework for classifying medical terms, organized into five hierarchical levels. These levels range from the most detailed category, known as the lowest-level term (LLT), to broader categories: preferred terms (PTs), high-level terms (HLTs), high-level group terms (HLGTs), and finally, system organ classification (SOC) [9]. Following a thorough data- cleaning process, a significant total of 12,864,087 adverse event reports were collected. Among these reports, 23,886 specifically identified an ADC as the primary suspect for the reported adverse events. In the next phase of the analysis, all preferred terms associated with the System Organ Classification titled “Gastrointestinal Systemic Disorders” underwent a rigorous screening process. During this stage, duplicates containing identical information were systematically excluded. The identification of duplicates was based on various criteria, including adverse event name, International Statistical Reporting Standard (ISR) number, registration date, medication, indication, sex, reporting country, and patient age, among other factors. Additionally, gastrointestinal adverse events that could arise from concomitant medications or drug‒drug interactions were also filtered out. As a result, the final count of unique reports that were deemed relevant for the study was 1,517. These deduplicated reports formed the basis for the subsequent analysis, and the entire screening process is illustrated in detail in Fig. 1.

Fig. 1

Flowchart of the FAERS database’s pipeline for screening ADC-associated gastrointestinal adverse events

Data mining

Data analysis was conducted via the reporting odds ratio (ROR) and proportional reporting ratio (PRR) techniques to assess the potential link between ADRs and ADCs, employing the specific algorithms detailed in Table 1. The ROR technique signals a risk when the number of reports reaches a minimum of three and the lower bound of the 95% confidence interval (CI) surpasses one. In a similar vein, the PRR technique indicates a significant signal when there are at least three reports, the PRR is two or higher, and the rank sum is four or more [10]. Both the ROR and PRR approaches successfully identified credible signals, implying a statistical relationship between the drug and gastrointestinal adverse events. To reduce bias and enhance the reliability of the AE-positive signals included in the study, signals not pertinent to safety, such as those related to FDA-approved uses or product concerns, were omitted.

Table 1 Two algorithms used for signal detection

Statistical analysis

In our study, we employ Microsoft Excel 2021 as our primary tool for conducting the statistical analyses, while R language is utilized for the visualization of the results. This combination of software allows for a comprehensive approach to data handling and presentation, enhancing the clarity and effectiveness of our findings. When discussing serious adverse events, we categorized them into several critical classifications. These include instances that result in death, those that are life-threatening, occurrences that lead to initial or prolonged hospitalization, and other significant medical events. Such classifications are essential for a thorough evaluation of patient safety and treatment efficacy. Additionally, we identify several exposure factors that may influence treatment outcomes, specifically focusing on patient demographics such as age and sex, along with the specific drug treatment regimen being administered. Understanding these factors is crucial, as they can significantly impact both the patient experience and the overall effectiveness of the treatment protocols being studied.

Results

Adverse drug event reporting

A total of 1,517 cases involving AEs were documented between the first quarter of 2011 and the second quarter of 2024, as illustrated in Table 2. The distribution of these identified cases reveals a notable prevalence among various categories: 137 cases of BV, representing 9.04% of the total; 209 cases of EV, accounting for 13.80%; 22 cases of GO, accounting for 1.45%; 30 cases of IO, accounting for 1.98%; 54 cases of MS, accounting for 3.56%; 80 cases of PV, accounting for 5.27%; 183 cases of SG, accounting for 12.06%; a significant majority of 764 cases of TD, accounting for 50.36%; and 38 cases of TE, accounting for 2.50%. The data indicate that TD cases were the most common, followed by EV, SG, BV, PV, MS, TE, IO, and GO cases, in terms of overall case numbers. The demographic distribution of these cases revealed that females accounted for 60.58% of the cases, whereas males constituted 24.32%. Additionally, 15.10% of the cases lacked any recorded information regarding sex. Our analysis of the sex characteristics of AEs in each ADC gastrointestinal system revealed several notable findings. Notably, EV reports predominantly featured male patients, whereas both the SG and TD reports presented a greater proportion of female patients. Interestingly, there are no male reports of MS and TE; this may be attributable to the treatment scope, which specifically includes conditions impacting female patients with breast cancer, ovarian cancer, and endometrial cancer. The gender ratios of other drug-related cases appeared to be relatively balanced across the board. When the age distribution of the patients involved was analyzed, the average age was determined to be 59.69 years. The findings indicated a significant representation of middle-aged patients in both the SG and TD reports, whereas EV reports revealed a larger proportion of older individuals. Other drug reports tended to exhibit less variability in age. Furthermore, it is important to note that the data contributing to this study predominantly originated from the United States, Japan, France, and Canada, suggesting a significant regional influence on the adverse event reports collected during this timeframe.

The indications of the nine ADCs are noteworthy. For the BV report, a total of 134 cases were documented, with 97.81% of these cases providing clear indications. Among these patients, 80, accounting for 59.70%, were diagnosed specifically with Hodgkin lymphoma. In addition, 34 cases, representing 25.37%, were associated with other types of lymphomas, underscoring the prevalence of these conditions within the report. The EV report presented a total of 209 cases, with impressive 100% clarity in the indications provided. Among these, 127 cases, accounting for 60.77%, were identified as cases of transitional cell carcinoma, whereas 43 cases, accounting for 20.57%, were diagnosed with bladder cancer. This report highlights the range of urological malignancies that are being monitored in relation to drug consequences. In the case of the GO report, 18 instances were observed, with 81.82% identified as acute myeloid leukemia. This relatively high percentage indicates a concerning trend in the association of acute myeloid leukemia with the agents being studied. The IO report consisted of 17 total cases, with 56.67% providing clear indications. Within this subset, the majority were lymphomas, with 15 cases, reflecting a significant 88.24%. The remaining 2 cases, or 11.76%, were classified as acute myeloid leukemia, indicating an important divergence in the types of cancers associated with the drugs in question. The MS report revealed that 18 cases were identified, with a noTable 33.33% of the patients diagnosed with ovarian cancer. This information adds another dimension to understanding the impact of the drugs under review. In the PV report, 78 cases were clearly indicated, with an impressive 97.50% identified as cases of lymphoma. The concentration of lymphoma cases within this report signifies a potential area for further investigation regarding drug safety. The SG report revealed clear indications for 156 patients, with 85.25% clarity. Here, 131 patients, accounting for 83.97%, were diagnosed with breast cancer, highlighting the significant impact of this type of cancer in the context of adverse drug consequences. The TD report included a substantial total of 665 cases, with 87.04% clarity among indications. The report revealed that 531 cases, accounting for 79.85%, were associated with breast cancer, whereas 82 cases, accounting for 12.33%, were related to gastric cancer. This distribution of cases further emphasizes the importance of monitoring these specific malignancies in relation to drug exposure. Finally, the TE report indicated 22 cases, with a clarity rate of 57.89%. Notably, all of these patients were diagnosed with breast cancer, reiterating the critical need to explore the connections between drug use and this prevalent form of cancer. In summary, the characteristics of gastrointestinal AEs are correlated with reported adverse drug consequences, as outlined in Table 2 of the study.

Table 2 Characteristics of Gastrointestinal AE correlated with ADCs

Adverse events related to the gastrointestinal system associated with ADCs

Table 3 provides a detailed overview of the frequency of gastrointestinal AEs associated with each ADC, accompanied by their respective RORs, proportional reporting ratios (PRRs), and 95% confidence intervals (CIs). Among the various ADCs assessed in this analysis, the IO was found to have the fewest reported adverse events, with a sample size of 30 and a reporting odds ratio of 11.12, indicating a notable level of gastrointestinal safety despite the low occurrence of adverse events. The accompanying 95% confidence interval ranges from 4.86 to 6.90, which adds robustness to the findings. The PRR for IO was recorded at 10.68, with a 95% CI of 7.50–15.16 and a chi-square value of 254.45, further reaffirming the data. In comparison, EV had a larger sample size of 209, alongside a reporting odds ratio of 3.26, indicating a higher frequency of adverse events related to gastrointestinal safety. Despite this increased occurrence, the gastrointestinal safety risk associated with EV was determined to be lower. The 95% CI for EV ranged from 2.82 to 3.75, suggesting that even with a greater number of reported adverse events, the overall risk remained manageable. Its PRR was recorded at 3.02, with a confidence interval of 2.66–3.43 and a chi-square value of 295.56. When ranked according to gastrointestinal safety risk, the ADCs are prioritized from highest to lowest as follows: EV, TE, TD, MS, SG, BV, PV, GO, and IO. This ranking provides valuable insights into the safety profiles of these ADCs, guiding future assessments and clinical decisions.

Table 3 Safety-related adverse drug events among different ADCs

PT signal detection results

Within the context of the PT hierarchy utilized for detecting signals of adverse events, the ROR method was applied to analyze gastrointestinal adverse event signals stemming from nine ADCs. This rigorous analysis led to the identification of a total of 38 distinct signals. These signals cover a wide array of gastrointestinal complications, including but not limited to colitis, intestinal obstruction, pancreatitis, gastrointestinal toxicity, gastrointestinal perforation, gastrointestinal bleeding, and gastrointestinal ulcers. Additional conditions identified included ascites, diarrhea, abdominal pain, gum bleeding, stomatitis, and esophageal varices, as well as symptoms such as nausea and vomiting, jaundice, and dry mouth. A visual representation of these findings can be seen in Fig. 2. Breaking down the associations of these signals with individual ADCs revealed that the drug entity enfortumab vedotin (EV) was related to 14 different PT signals. In comparison, polatuzumab vedotin (PV) was linked to 12 signals, whereas Belantamab mafodotin (BV) was associated with 10 signals. Other ADCs, such as trastuzumab deruxtecan (TD) and tucatinib (TE), were associated with 9 and 8 signals, respectively. Furthermore, Sangamo (SG) was connected to 7 signals, Ibusilab (IO) to 4, Macrogenics (MS) to 4, and Gilead (GO) to 3 signals. This detailed analysis highlights the varying degrees of gastrointestinal adverse events linked to each ADC, as effectively illustrated in Fig. 2.

Figure 3 provides a comprehensive overview of the signal strength for various PTs associated with AEs across different ADCs. The detailed analysis revealed that the seven most frequently reported PTs are nausea, diarrhea, vomiting, ascites, colitis, pancreatitis, and intestinal obstruction, with each of these events occurring at varying incidence rates. Specifically, nausea was recorded in 506 instances, diarrhea in 262 instances, vomiting in 216 instances, and even less frequently occurring events such as ascites (112 cases), colitis (90 cases), pancreatitis (52 cases), and intestinal obstruction (37 cases). While these findings highlight certain prevalent trends within the data, it is essential to note the significant individual variations in drug-related AEs across different ADCs. Despite the overall strength of PT signals, specific drugs have unique adverse effects. For example, the most frequently reported AE for the drug Belantamab Mafodotin (BV) is pancreatitis, indicating a notable concern for patients receiving this treatment. In contrast, drugs such as endothelin receptor antagonists (EVs), mycophenolate (MS), and St. John’s Wort (SG) primarily cause diarrhea as the main AE. Other noteworthy observations include that gastrointestinal oncology (GO) agents and parenteral valganciclovir (PV) are linked with gastrointestinal bleeding, while intravenous oxaliplatin (IO) tends to be associated with ascites, whereas thalidomide (TD) is connected to nausea and vomiting. Finally, TE appears to commonly result in dry mouth and gum bleeding, underscoring the importance of tailoring patient care on the basis of these adverse effects.

Fig. 2

ROR heatmap of ADC-related gastrointestinal AEs

Fig. 3

Signal strength of ADC gastrointestinal AEs at the PT level in the FAERS database

Analysis of adverse drug event results

An investigation was carried out to evaluate the results of nine categories of AEs that occur within the gastrointestinal system and are associated with positive detection signals for ADCs. This thorough analysis encompassed a total of 1,517 reported AEs and resulted in significant findings: 237 cases resulted in fatalities, 43 incidents were deemed life-threatening, 439 instances required either initial or extended hospitalization, 548 other adverse events of various types, and 250 cases remained unreported. As detailed in Table 2, a horizontal comparison of the various ADCs highlighted notable disparities in outcomes. Specifically, among the different ADCs evaluated, TD had the greatest number of deaths and instances of hospitalization, thereby raising concerns regarding its risk profile. In contrast, the type identified as TE revealed the highest rate of mortality linked to adverse events, indicating a critical need for attention and further investigation into the safety implications associated with its use. These findings underscore the importance of continuous monitoring and assessment of AEs in the context of ADC therapies.

Discussion

Since the FDA approved the first ADC in 2011, this category of medication has swiftly developed into a model for precision therapy in the field of oncology [11]. Following a favorable regulatory position, the China Food and Drug Administration (CFDA) has approved multiple ADCs, such as BV, TD, SG, and IO, demonstrating China’s commitment to the effectiveness and safety of ADCs in treating tumors [12]. However, differences in the composition of monoclonal antibodies and cytotoxic agents among various ADCs can lead to significant variations in their therapeutic outcomes and side effect profiles. Examining a wider range of adverse responses, as well as the particular side effects associated with this drug class at the global level, is essential for assessing the safety of ADCs. This research, which leverages the FAERS database and the OpenVigil 2.1 mining tool, aims to analyze the gastrointestinal safety of ADCs after they have entered the market, thus providing crucial safety information for clinical application.

Case report analysis

TD and EV emerged as the leading contributors to adverse reactions, with TD being responsible for a total of 764 cases and EV being responsible for 209 cases. Notably, although TD was introduced to the market in 2019, it accounted for a substantial 50.36% of all gastrointestinal adverse reactions related to ADCs. This statistic is consistent with the results observed in TD clinical trials [13], highlighting the importance for healthcare providers to remain alert and actively monitor patients undergoing TD for various solid tumors. Clinicians should be prepared to provide appropriate symptomatic relief to manage any gastrointestinal adverse reactions that may arise during the course of treatment.

In the studied cohort, the majority of adverse events linked to EV gastrointestinal problems were reported by males (77.51%), associated with the payload MMAE (monomethyl auristatin E), which is primarily used for managing bladder cancer and other urological issues in males. Gastrointestinal toxicity represents a notable adverse effect associated with this medication [14]. TD and SG are mainly prescribed for treating female patients with metastatic breast cancer, with reports of adverse events primarily involving female patients. Interestingly, MS and TE, which are available for a shorter time and have fewer users, currently feature all reported adverse events in women. This sex distribution indicates a significant correlation between reports of adverse events for ADCs and the therapeutic purposes of the drugs, as well as their time on the market. Physicians are encouraged to remain attentive to potential adverse events related to these agents while addressing specific medical conditions.

ADCs and adverse events in the gastrointestinal system

In our study, we identified nine distinct types of gastrointestinal adverse event signals associated with ADCs. Through this analysis, we were able to extract a comprehensive total of 1,517 cases related to these events. To quantify the severity of each drug’s associations with these adverse events, we ranked them according to their signal strength, as measured by the ratio of occurrence (ROR value). The ranking revealed the following order of ADCs: EV, TE, TD, MS, SG, BV, PV, GO, and IO. Among these nine ADCs, 38 potential treatment-related signals were identified. The prevalence of these signals varied significantly among the drugs. Specifically, 14 signals were linked to EV, while PV had 12, BV had 10, TD had 9, TE had 8, SG had 7, MS had 4, IO had 4, and GO had 3 signals. Notably, the ADCs exhibiting the strongest signals for adverse events were BV, which was prominently linked to cases of paralytic ileus; EV, associated with enteritis, mechanical ileus, and small bowel perforation; GO, which indicated neutropenic colitis; IO, connected to ascites; PV, tied to gastrointestinal and small bowel perforation; SG, associated with gastrointestinal toxicity and neutropenic colitis; and TE, linked to the occurrence of esophageal varices. In summary, GO and IO exhibited the weakest signal intensities at the SOC level. They also registered fewer effective signals at the PT level compared to other ADCs, indicating a superior safety profile during therapeutic administration. Conversely, EV demonstrated stronger signal intensities and a higher incidence of PT signal detection at the SOC level, suggesting a lower safety margin. Therefore, increased vigilance regarding potential adverse effects is warranted during the clinical deployment of this drug.

ADC drug guidelines [15], clinical trials [16], and abundant post - marketing clinical data [17] have demonstrated a prevalent occurrence of gastrointestinal adverse events across all ADC variants currently on the market. A vast body of research and analysis on global ADCs indicates that intestinal obstruction [18, 19] is a significant gastrointestinal complication associated with BV, and neutropenic colitis [20, 21] is a common adverse reaction in GO and SG. Additionally, all nine ADC types are associated with a high risk of various colitis-related adverse effects. These findings are in line with the results of the present study. All the ADC types, apart from the IO, provided effective signals for different types of colitis. We recommend that clinicians implement a standardized pharmacological prophylaxis regimen to prevent colitis associated with antibody-drug conjugate (ADC) therapy. This regimen should include strategies such as bowel rest, electrolyte replenishment, and the consideration of parenteral nutrition. In addition, both IO and TE effectively detect signals of gingival bleeding, which may be associated with the underlying causes of thrombocytopenia. During clinical practice, greater attention should be given to platelet-related indicators, and patients should be reminded to be vigilant about bleeding-related risks [22]. BV and IO detected effective signals of melaena, which may be attributed to gastrointestinal ulcers or perforated bleeding. It is essential to carefully consider the risk of bleeding when administering these two medications in a clinical setting.

Notably, an adverse reaction that was not mentioned in the package insert has been identified: ascites. Possible mechanisms include the high expression of certain ADC targets in peritoneal metastasis, which may lead to local inflammation or tumor lysis, thereby increasing exudation. Additionally, ADCs could activate the immune system, potentially resulting in autoimmune peritonitis. Other contributing factors to the development of ascites include tumor progression, portal vein thrombosis, and heart failure. Despite these considerations, ascites is recognized as a significant adverse event associated with various ADCs. Therefore, clinicians must exercise caution and regard ascites as a potential risk linked to ADCs [23]. Injections of IO and EI can readily precipitate adverse events that are not listed in the instructions, such as gum bleeding; therefore, heightened caution is warranted when employing these agents. TE has been associated with notable esophageal varices that are not described in the product instructions, and this signal may contribute to gastrointestinal bleeding events in the adverse effects profile of TE. This association may be attributed to the relative liver toxicity of TE, which induces nodular regenerative hyperplasia (NRH) in the liver, leading to non - cirrhotic portal hypertension and, ultimately, the occurrence of esophageal varices [24]. Consequently, patients experiencing esophageal varices due to TE use without other signs of liver cirrhosis should be considered for drug-induced NRH.

Conclusion

In this study, we employed the Openvigil 2.1 tool, which is based on the FAERS database, to examine AEs following the market introduction of ADCs by utilizing the ROR and PRR methodologies. Through data mining in FAERS, we discovered a link between gastrointestinal toxicity and several ADCs, such as brentuximab vedotin, enfortumab vedotin, gemtuzumab ozogamicin, inotuzumab ozogamicin, mirvetuximab soravtansine, polatuzumab vedotin, sacituzumab govitecan, trastuzumab deruxtecan, and trastuzumab emtansine. Importantly, the adverse effects noted in this study, which were not indicated in the ADC product labeling, may provide additional support for improving drug labeling. We recommend that clinicians remain vigilant regarding gastrointestinal symptoms in patients receiving antibody-drug conjugate (ADC) medications. They should adopt targeted preventive measures for the specific gastrointestinal adverse reactions that may arise with different ADCs to ensure the safety of medication use during ADC therapy.

However, this study presents certain limitations: ① The FAERS database, being a spontaneous reporting system, is subject to variability in data quality due to differences among reporters and across countries, which could influence the outcomes; ② Despite the concurrent use of the ROR method and the PRR method to mitigate the appearance of false positive signals, the complete elimination of false positives remains unattainable; ③ The study’s reporting countries are predominantly the United States and Japan, which may introduce a regional bias, and the findings should be interpreted with caution, serving only as a reference; ④ This study is unable to ascertain the total number of patients treated with antibody-drug conjugates (ADCs), thus precluding the calculation of the incidence of ADC-related gastrointestinal toxicity. Consequently, only qualitative signals of ADC-related gastrointestinal adverse events can be derived; ⑤ The FAERS database lacks comprehensive patient-level data, including information on comorbidities, concurrent medications, and dosage. This limitation may impact the results. Nevertheless, the findings, derived from a large sample size, still possess significant clinical relevance for the application of ADC drugs in clinical practice.