EndocrineXiangling Duan, Wenhui Liu, Junlong Ma, Bao Sun
BACKGROUND: Although glucagon-like peptide-1 receptor (GLP-1R) is widely expressed in multiple tissues and organs including skin and subcutaneous tissue, with diverse physiological roles in metabolism and immunity, the potential causal association between GLP-1R expression and dermatological diseases remains unclear. METHODS: Available cis-expression quantitative trait loci (cis-eQTLs) were selected as genetic instruments for GLP-1R expression. A two-sample Mendelian randomization (MR) analysis was employed to assess the potential causal association between GLP-1R expression and dermatological diseases. Subsequently, a two-step mediation MR analysis was conducted to explore the mediators between GLP-1R expression and dermatological diseases. Finally, a real-world pharmacovigilance analysis using the Food and Drug Administration Adverse Event Reporting System (FAERS) database was conducted to provide supplementary real-world evidence. RESULTS: Our study revealed distinct associations between GLP-1R expression and dermatological diseases. Specifically, GLP-1R expression was significantly associated with a reduced risk of bullous pemphigoid (OR = 0.47, 95% CI = 0.31-0.72, P < 0.001), as well as an increased risk of psoriasis (OR = 1.19, 95% CI = 1.08-1.32, P < 0.001) and urticaria (OR = 1.41, 95% CI = 1.23-1.61, P < 0.001). Further mediation MR analysis suggested that the immune cell phenotype HLA-DR on B cells might mediate the causal association between GLP-1R expression and bullous pemphigoid, with an estimated mediation proportion of 35.7% (P = 0.003). Other immune cells including Monocytic Myeloid-Derived Suppressor Cells Absolute Count, HLA-DR on CD14 + CD16- monocytes, and HLA-DR on CD14 + monocytes were also identified as potential mediators, with estimated mediation proportions of 23.6% (P = 0.010), 13% (P = 0.024), and 12.7% (P = 0.036), respectively. In the complementary FAERS analysis, consistently, GLP-1RAs reports showed a lower reporting signal for bullous pemphigoid than sodium-glucose cotransporter 2 inhibitors (SGLT2is) reports. CONCLUSION: Our findings suggest that GLP-1R expression is associated with a reduced risk of bullous pemphigoid, which may be partly mediated by specific immune cell phenotypes.
European journal of clinical pharmacologySoo Hyeon Lee, Seojun Lee, Sangyoon Chris Lee, Yeo Jin Choi
PURPOSE: This study aims to evaluate the comparative adverse drug event (ADE) reporting patterns associated with antidiabetic drugs and develop machine learning (ML)-based classification models for the seriousness of reported ADEs. METHODS: We performed a retrospective analysis of 28,633 antidiabetic-related ADEs reported to the Korea Institute of Drug Safety and Management- Korea Adverse Event Reporting System database (KAERS DB 2505A0010) from 2015 to 2024. Disproportionality analyses identified safety signals using reporting odds ratios (RORs) with 95% confidence intervals (CIs). Factors associated with serious adverse event (SAE) classification were assessed using multivariate logistic regression, and three ML-based classification models were developed. RESULTS: Older adults accounted for 64.1% of the reported ADEs, with 2.35% classified as SAEs. Sulfonylureas demonstrated the highest SAE reporting signal (ROR 2.60, 95% CI 2.22-3.05). Male sex, older age, and sulfonylurea exposure were associated with higher odds of reports being classified as serious. Across the ML models, diabetic neuropathy treatment consistently emerged as the most influential feature contributing to SAE classification. CONCLUSION: ML-based classification complemented disproportionality analyses by identifying features associated with classification of reported ADEs as serious. Further validation using integrated longitudinal real-world data is warranted to confirm the robustness of these findings.
Journal of the American College of Clinical Pharmacy : JACCPBrian Murray, Jackie Rowe, Duncan X Dobbins, Phuong Duong, Megan Kunka Fritz, Madison Brooke Grizzle, Steven T Johnson, Abbie D Leino, Rajsumeet Macwan, Farah …
Artificial intelligence (AI) is increasingly shaping the pharmaceutical industry. This ACCP commentary examines the implications of AI for industry-based clinical pharmacists across the pharmaceutical life cycle, including drug development, regulatory affairs, medical affairs, health economics and outcomes research, and pharmacovigilance. Artificial intelligence-enabled tools may support target identification, clinical trial design, regulatory intelligence, evidence synthesis, medical content generation, real-world evidence analysis, economic modeling, adverse event processing, and safety signal detection. As these tools mature, the role of the clinical pharmacist is likely to shift from primarily task execution toward clinical interpretation, quality assurance, strategic decision-making, and governance of AI-supported outputs. However, AI implementation also introduces important risks and practical implementation challenges. These limitations reinforce the need for clinical pharmacists to remain actively engaged as human-in-the-loop experts who can assess whether AI-generated insights are scientifically valid, clinically relevant, ethically sound, and appropriate for decision-making. Ultimately, AI may expand the reach and efficiency of pharmaceutical industry functions, but successful integration will depend on pharmacist leadership in evaluation, oversight, and responsible implementation.
Pakistan journal of pharmaceutical sciencesJieyuan Chen, Zhaojun Wang, Songsong Mao
BACKGROUND: Sevoflurane, an ether-derived inhalational anesthetic widely used for general anesthesia, requires comprehensive safety evaluation. OBJECTIVES: This study aimed to identify sevoflurane-associated adverse events and detect unexpected safety signals using data from the FDA Adverse Event Reporting System (FAERS). METHODS: We analyzed FAERS data from the first quarter of 2004 through the first quarter of 2024, applying four disproportionality analysis algorithms (ROR, PRR, BCPNN, MGPS) to reports designating sevoflurane as the "primary suspect" drug. RESULTS: Among 1,649 reports, we identified 27 significant preferred terms affecting four major organ systems. Notable unexpected safety signals included cardiac arrest, anesthesia awareness, delayed recovery, pulmonary alveolar hemorrhage, anaphylactic shock and seizures. CONCLUSION: These findings provide critical insights into sevoflurane's real-world safety profile, particularly revealing severe yet underrecognized risks that can inform enhanced clinical monitoring and safer anesthetic practice. However, it should be acknowledged that disproportionality analysis is a hypothesis-generating or refinement approach, and further confirmatory studies are warranted to establish causal relationships.
Pharmacoepidemiology and drug safetyValcieny Sandes, Alice Ramos-Silva, Adriana Ivama-Brummell, Albert Figueras, Elisangela Costa Lima
PURPOSE: Appropriate antibiotic dosing is essential to minimise treatment failure and reduce selective pressure that may contribute to antimicrobial resistance. This study characterised antibiotic medication-error drug-event pairs (MEPs) reported to the medicines Individual Case Safety Reporting (ICSR) system, VigiMed and assessed the distribution of dosing-related error drug-event pairs (DEPs) across World Health Organization (WHO) AWaRe (Access, Watch, and Reserve) antibiotic classes. METHODS: We conducted an exploratory cross-sectional analysis of reports involving systemic antibiotics (Anatomical Therapeutic Chemical, ATC Classification System as J01 - antibacterials for systemic use) from Brazil's open-access pharmacovigilance database (VigiMed) from 2018 to 2025. MEPs were identified using the Standardised MedDRA Query for Medication Error; within this group, Preferred Terms (PTs) related to dose amount, timing, frequency, duration, or rate of administration were classified as DEPs (32 PTs). Descriptive analyses and modified Poisson regression were used to examine demographic and reporting characteristics associated with MEP classification and, among MEPs, the association between AWaRe class and DEP classification. RESULTS: The adjusted prevalence of MEP classification was higher among neonates (PR 3.90), infants (PR 1.78) and older adults (PR 1.70) than among adults. DEPs accounted for 31.7% of MEPs (n = 2587; 61 antibiotics), with dose omission (51.4%) and dosing regimen problems (16.7%) as the leading error subtypes. Among DEPs, 35.9% involved Access, 56.7% Watch and 7.1% Reserve antibiotics. In adjusted analyses, Watch (PR 1.14; 95% CI 1.07-1.22) and Reserve (PR 1.31; 95% CI 1.16-1.49) antibiotics showed a higher prevalence of DEPs than Access antibiotics. CONCLUSIONS: This study analyses pharmacovigilance data on antibiotic dosing-related medication errors, highlighting potential opportunities for antimicrobial stewardship.
BMJ openIoana Rada Popa Ilie, Steliana Ghibu, Anca Butuca, Carmen Maximiliana Dobrea, Adina Frum, Laurentiu Stoicescu, Cãlin Homorodean, Felicia Gabriela Gligor, Claud…
OBJECTIVE: As obesity and diabetes are on the verge of an alarming growth, so is the use of tirzepatide, a novel dual glucose-dependent insulinotropic peptide/glucagon-like peptide-1 (GLP-1) receptor agonist (RA) and currently the most effective weight loss-drug. This research aimed to identify reporting patterns related to suboptimal therapeutic outcomes and tirzepatide-related drug-use issues, mining the EudraVigilance (EV). DESIGN: Retrospective pharmacovigilance study using descriptive and disproportionality analyses of reports retrieved from the EV database. SETTINGS: Analysis of individual case safety reports involving tirzepatide in comparison with other GLP-1 RAs in the overall dataset (healthcare professionals (HP) and non-HP reports combined) and in the HP group. OUTCOME MEASURES: Reporting ORs (RORs) with 95% CIs for selected Preferred Terms (PT) related to drug-use issues. RESULTS: Among all analysed PTs, the most frequently reported were 'Off-label use' (n=1521), 'Drug ineffective' (n=425) and 'Off-label use device' (n=99). PTs in HP reports showed lower reporting odds compared with non-HP reports. In the overall dataset, reports involving tirzepatide showed lower reporting odds of the PT 'Drug ineffective' than those involving liraglutide (ROR 0.61, 95% CI 0.54 to 0.70), dulaglutide (ROR 0.72, 95% CI 0.63 to 0.82), exenatide (ROR 0.78, 95% CI 0.67 to 0.92) and semaglutide (ROR 0.81, 95% CI 0.72 to 0.91). However, in HP reports, no difference in reporting odds was observed between tirzepatide and semaglutide. A different pattern was observed for 'off-label use' in the full dataset, with higher reporting odds for tirzepatide vs lixisenatide, dulaglutide and liraglutide, but lower reporting odds vs semaglutide (ROR 0.52, 95% CI 0.49 to 0.55). In contrast, a higher reporting odd for the PT 'off-label use of device' was observed for tirzepatide versus semaglutide (ROR 43.47, 95% CI 16.00 to 118.13) in the overall reports; however, this finding should be interpreted with caution given the wide CI. CONCLUSIONS: This study complements existing evidence from clinical trials and current clinical practice.
Pharmacoepidemiology and drug safetyJin-Hwan Kim, Jonghun Kim, Saerom Kim
BACKGROUND: In South Korea, a substantial volume of prescription psychotropics is dispensed outside the national health insurance reimbursement system, creating a pharmacovigilance blind spot. We investigated nationwide utilization patterns of medical narcotics to identify potential safety signals using the Narcotics Information Management System (NIMS) database. METHODS: This population-based descriptive study analyzed NIMS data covering all controlled substances dispensed in South Korea from January 2019 to May 2023. We examined overall trends and stratified patterns for anorexiants and ADHD medications. Prescription volumes were expressed as estimated dispensing units (EDU) per 1000 persons per month using resident registration population data as denominators. RESULTS: Among 3.1 million dispensing cases, anorexiants exhibited a stark gender disparity. Women at primary care clinics received approximately 440 EDUs per 1000 women per month-approximately 15 times the volume dispensed to men-concentrated among women aged 30-49, suggesting widespread prescribing outside approved indications for aesthetic weight management. ADHD medications were predominantly dispensed to males, with the largest gender gap observed in adolescents aged 10-19 and a progressively narrowing disparity in older age groups. Dispensing rates for both drug classes increased steadily over the study period, with the most pronounced rise in ADHD medications observed after the resumption of in-person schooling following COVID-19 disruptions. CONCLUSION: This study provides the first population-level analysis of medical narcotics utilization in South Korea's non-reimbursed sector. The disproportionate exposure of young women to psychotropic anorexiants constitutes a compelling pharmacovigilance signal. Addressing this requires not only an integrated data infrastructure linking NIMS with national health insurance records, but also the restoration of meaningful professional oversight over prescribing outside the reimbursement system.
Human vaccines & immunotherapeuticsWenjun Fang, Ben Liu, Li Cai
Rituximab is widely used for B‑cell malignancies and autoimmune diseases, but its pediatric safety profile requires systematic real‑world evaluation. This research mined safety signals of adverse events (AEs) linked to rituximab in pediatric patients using U.S. Food and Drug Administration Adverse Event Reporting System (FAERS) data from 2004 to 2025. A total of 2,231 pediatric AE reports involving rituximab were identified and analyzed using four disproportionality methods (ROR, PRR, BCPNN, and MGPS) for signal detection and Weibull distribution analysis for time-to-onset patterns. Results showed a balanced sex distribution in AE reports (47.7% male, 48.5% female). Most reports involved adolescents aged 12-17, with an overall increasing trend. A total of 149 positive safety signals were detected via disproportionality analysis. Several of these signals pertained to immune system - and infection‑related adverse events. Notably, the elevated signal magnitudes for certain signals were substantially confounded by indication bias, pre‑existing comorbidities, and reports associated with therapeutic failure. Stratified analyses revealed differences in the signal distribution of adverse‑event reports across age and sex subgroups. Time‑to‑onset analysis of adverse‑event reports indicated that most events documented in these reports occurred within 30 d of therapy initiation, with a median of 7 d. This study characterizes the reported safety profile of rituximab in pediatric patients based on FAERS spontaneous reports. Our findings represent exploratory signal hypotheses only and can serve as a reference for subsequent relevant investigations.
Breast cancer research and treatmentIoannis-Alexios Koumprentziotis, Alexander Cawley, Elliot Ewig, Alexandre O Gérard, Maëlys Labat, Grażyna Kamińska-Winciorek, Marcin Kubeczko, Luca Rapparini, …
PURPOSE: Sacituzumab govitecan (SG) is an antibody-drug conjugate approved across multiple metastatic breast cancer settings. Although hematologic and gastrointestinal toxicities dominate its safety profile, dermatologic adverse events (dAEs) remain poorly characterized, despite the physiological expression of Trop-2 in epidermal keratinocytes. We aimed to comprehensively characterize the spectrum, severity, time to onset, and oncologic impact of SG-associated dAEs across real-world and pharmacovigilance data sources. METHODS: An international retrospective cohort study was conducted within the EADV Task Force "Dermatology for Cancer Patients" (n = 56 breast cancer patients). Findings were triangulated with disproportionality analyses of the WHO Global Pharmacovigilance database (VigiBase®; n = 696 cases) and the French Pharmacovigilance Database (FPVD; n = 46 cases), alongside a literature review of SG clinical trials reporting dAEs. RESULTS: Across all data sources, alopecia was the most frequently reported dAE (EADV cohort: 41.1%; FPVD: 52.2%; VigiBase®: 67.6%), followed by stomatitis/mucositis, pruritus, xerosis, and maculopapular rash. In the EADV cohort, 13 distinct dAE types were identified; median time to first dAE was 56 days, and 96.4% of events were grade 1-2. Treatment discontinuation attributable to dAEs occurred in only 3.6% of patients. VigiBase® disproportionality analysis identified significant reporting signals for alopecia (IC 3.3 [3.09; 3.53]), infusion-related reactions (IC 1.9 [1.07; 2.55]), and stomatitis (IC 1.6 [0.86; 2.30]). Preliminary signals of cutaneous infections were observed but did not consistently meet the threshold for statistical significance. CONCLUSION: SG-associated dAEs are predominantly low-grade and infrequently precipitate treatment discontinuation. Pharmacovigilance data raises a potential signal for cutaneous infections that warrants further investigation. As SG increasingly enters combination regimens, structured dermatologic monitoring and prospective characterization of dAEs are essential to guide oncodermatologic management and protect oncological outcomes. CLINICAL TRIAL NUMBER: Not applicable.
Journal of medical Internet researchGarang Majok Dut
Gaps in pharmaceutical governance could widen with the adoption of AI, even as AI promises better pharmacovigilance in low-income countries (LICs). While advanced regulatory systems like Australia's are integrating AI into pharmaceutical governance, LICs with underdeveloped regulatory capabilities, such as South Sudan, lag behind. The potential divergence disorients the World Health Organization's "Medicine Without Harm" agenda and effective global pharmacovigilance. Moreover, evolving global governance initiatives, including the newly established United Nations scientific panel on AI, may be hampered by this global divergence in capabilities. This makes 3 critical interrelated questions: what are the moral trade-offs in the introduction of AI in health care, what power dynamics impact the introduction of AI into health systems, and how could AI be used for pharmacovigilance in LICs? This viewpoint aims at informing global policies and regulations on AI in pharmacovigilance. It uses clinical, policy, and regulatory practitioner insights to synthesize evidence on the ethical, economic, and clinical contours of AI in pharmacovigilance. It contrasts the high-income context of Australia with the low-income context of South Sudan and shows that national capabilities are instrumental for institutionalizing global practice. It identifies current ethical challenges with applying AI and digital health, which straddle epistemic, normative, and metaethical domains, such as misguidance, cultural devaluation, and trust deficit. These filter into demerits observed with current applications of AI to pharmacovigilance, from the detection of adverse drug events and adverse drug reactions to the simulation of clinical trials. The merits of current applications are multiple and depend on data quality, ranging from the detection of adverse drug reactions to real-time surveillance of medical errors and predictive application to population risk quantification of adverse drug events. The widening gaps in global capabilities amid rapid evolution of AI suggest the need for inclusive global governance in the early stages, especially because AI may be deterministic and effects may not be retrospectively surmountable. The viewpoint also assesses the sufficiency of current evaluation frameworks, noting that health economic models currently lag in capturing gains and losses from the adoption of AI in health systems, digital health frameworks are largely retrospective and overlook sociopolitical and financial contexts, and influential service-oriented frameworks for health systems overlook outcomes. It observes that, although AI could be harnessed across the breadth of the pharmaceutical system, effective evaluation of potential risks is hampered by upstream decisions in software development and procurement, which preclude aspects of subsequent application. This introduces inscrutability and weakens clinicians' role in risk adjudication, which may worsen with nonrepresentative evolution of AI. Using these insights and a case study on the low-income context of South Sudan, the viewpoint commends an integrated health systems framework and country-level investments in infrastructure and regulatory capabilities as requisites for effective global governance and equitable use of AI in pharmacovigilance.
Selective β1-blockers are fundamental in managing hypertension, coronary artery disease, and heart failure. Yet, despite their widespread use, practical guidance on safe and individualized prescribing remains limited. Existing evidence is largely derived from controlled clinical trials, which may not fully capture drug-specific adverse events, sex-related susceptibility, or early-onset risks observed in routine clinical practice. To generate real-world, evidence-based recommendations for the safe use of selective β1-blockers, we compared the adverse event profiles of metoprolol, bisoprolol, and atenolol using 2 large pharmacovigilance systems. Adverse drug events reported for the 3 agents between 2004 and Q2 2025 were retrieved from the US Food and Drug Administration Adverse Event Reporting System, a spontaneous-reporting system, and compared with reports from the Canadian Vigilance Adverse Reaction Online Database. Disproportionality analyses, preferred term mapping, exploratory sex-stratified comparisons, and time-to-onset modeling were conducted to characterize shared reporting patterns, drug-specific signals, and potential patient-level modifiers. Across all 3 agents, disproportionate reporting was observed for a common cardiovascular spectrum that included bradycardia, conduction abnormalities, heart failure, and blood pressure instability. Metoprolol showed prominent reporting signals for BRASH syndrome and neuropsychiatric and suicide-related events. Bisoprolol showed signals for bradyarrhythmia, hyperkalemia, and acute kidney injury. Atenolol showed reporting patterns involving blood pressure perturbations, electrolyte imbalance, and interaction-related events. Exploratory sex-stratified analyses identified differences in the reporting distributions of several PT = preferred terms for metoprolol and atenolol, whereas bisoprolol showed a more balanced distribution. All 3 agents displayed early-failure time-to-onset patterns, supporting closer monitoring after treatment initiation. This real-world pharmacovigilance assessment identifies drug-specific patterns of disproportionate reporting that may inform hypothesis generation and individualized monitoring. Because spontaneous-reporting data cannot establish incidence or causality, the findings should complement, rather than replace, clinical judgment and confirmation in controlled or longitudinal data sources.
Ensifentrine is a newly approved inhaled dual phosphodiesterase 3 and 4 inhibitor for the maintenance treatment of chronic obstructive pulmonary disease. Because post-marketing safety experience is still limited, this study evaluated adverse event reporting patterns for ensifentrine. Quarterly US Food and Drug Administration (FDA) Adverse Event Reporting System/Adverse Event Monitoring System files from 2024Q3 to 2026Q1 were analyzed. Duplicate reports were removed using Case Identification Number (CASEID), FDA receipt date (FDA_DT), and Primary Identification Number (PRIMARYID), and the primary analysis was restricted to reports in which ensifentrine was recorded as the primary suspect drug. Preferred Terms (PTs) were mapped to Medical Dictionary for Regulatory Activities primary System Organ Class categories, and report-level disproportionality was assessed using reporting odds ratio, proportional reporting ratio, an information component approximation, and an observed-to-expected approximation. After deduplication, 2,831,030 records were reduced to 2,500,712 unique reports. Ensifentrine was identified in 1137 reports across all drug roles and in 823 primary-suspect reports, which included 1817 PT records. At the System Organ Class level, Respiratory, thoracic and mediastinal disorders was the only category meeting all predefined robust signal criteria. At the PT level, respiratory, cardiovascular, psychiatric/neuropsychiatric, and product quality- or medication-use-related terms were notable. Time-to-onset could be calculated for 139 reports, with a median of 12 days. Respiratory reporting patterns predominated in early post-marketing reports for ensifentrine. These findings should be interpreted as hypothesis-generating pharmacovigilance signals rather than estimates of incidence or causality.
OBJECTIVE: To identify disproportionality signals linking androgenetic alopecia (AGA) drugs (finasteride, dutasteride, and minoxidil) with male infertility-related adverse events and to explore infertility-related biological features using testicular transcriptomic datasets. METHODS: A dual-database replication design was employed using FAERS (2004-2025) and EudraVigilance (2002-2025). Disproportionality analysis with multiple signal detection metrics assessed drug-infertility associations. Multi-level bioinformatic analyses-including toxicity prediction, drug-associated target screening, testicular transcriptomic analysis, single-cell RNA sequencing, intercellular communication analysis, gene set enrichment analysis (GSEA), and immune infiltration analysis-were integrated to explore biological features potentially relevant to male infertility. RESULTS: Disproportionality analyses detected signals for all three drugs, with finasteride demonstrating the most prominent reporting signal, followed by dutasteride and minoxidil. Multi-level bioinformatic analysis identified HIF1A as an overlap-derived candidate under the specified datasets and screening criteria, and HIF1A expression was higher in testicular tissue from patients with male infertility. Single-cell analysis showed higher HIF1A expression in late spermatocytes, Leydig cells, and myoid cells from infertility samples, together with group-dependent inferred communication patterns for selected VEGF-, PDGF-, IGF-, and FGF-related signaling axes. GSEA associated higher HIF1A expression with immune-response-, wound-healing-, and cell-adhesion-related processes, whereas lower HIF1A expression was associated with spermatid-development- and cilium/flagellum-dependent motility-related processes. ssGSEA-based analysis showed positive correlations between HIF1A expression and several immune-cell signature scores. CONCLUSION: This study systematically evaluated associations between AGA drugs and male infertility using real-world pharmacovigilance data and integrated bioinformatic analyses. The HIF1A-related transcriptomic findings provide a hypothesis-generating biological context for male infertility but do not establish a shared or drug-specific mechanism linking the three medications to infertility. The pharmacovigilance findings indicate reporting signals rather than incidence or causality.
ObjectiveTo investigate whether genetically proxied antihypertensive drug targets are associated with pancreatic exocrine diseases and whether corresponding antihypertensive agents show pancreatitis-related pharmacovigilance signals.MethodsThis integrative genetic epidemiology and pharmacovigilance study combined linkage disequilibrium score regression, two-sample Mendelian randomization, summary-data-based Mendelian randomization, Bayesian colocalization, and FAERS disproportionality analysis. Publicly available GWAS summary statistics were derived predominantly from European-ancestry participants, and whole-blood cis-eQTLs were used to genetically proxy antihypertensive drug-target expression. FAERS reports submitted from 2004 to 2023 were analyzed for pancreatitis-related reporting signals.ResultsAmong 21 antihypertensive drug targets, ADRB2, CACNA1I, JUN, NR3C2, and SLC12A4 showed nominal associations with pancreatic outcomes; however, none remained significant after Bonferroni correction. Absolute PP.H4 values did not support robust colocalization. Although several target-outcome pairs had high conditional PP.H4 values, these findings provide only relative support for H4 over H3 when both traits are assumed to have regional association signals. Several antihypertensive agents also generated pancreatitis-related reporting signals in FAERS; these signals do not quantify incidence or establish causality.ConclusionsThe findings are exploratory and prioritize selected antihypertensive targets and drugs for replication and mechanistic study. They do not establish causal effects, robust colocalization, or an increased incidence of pancreatitis.
European journal of clinical pharmacologyNuria Sols Cueto, María Del Mar Gutiérrez-Lobón, Araceli Núñez Ventura, Cristina Fernández-Fernández
PURPOSE: This study aimed to analyse the potential masking effect of drug-event combinations (DECs) with extreme reporting rates and to evaluate the impact on disproportionality analysis and therefore in signal detection. METHODS: An algorithm is proposed, based on the approach established by Juhlin et al., that identifies influential outliers and excludes them through six unmasking strategies to recalculate the Reporting Odds Ratio (ROR). This study was performed in the Spanish spontaneous reporting database FEDRA. The dataset included reports from 1 January 1981 to 17 February 2025 excluding those in which the suspected drug was a vaccine (ATC group J07). RESULTS: A total of 287,145 DECs were analysed. Of these, 0.4% (1,211 out of 287,145) were considered influential outliers. Almost 21% (81,371 out of 389,262) of the FEDRA reports included an influential outlier. About 14% (494 out of 3,447) of the drugs and 5.9% (576 out of 9,745) of the adverse drug reactions in FEDRA were part of an influential outlier. Regarding the disproportionality analysis, the proportion of DECs whose lower limit of the 95% confidence interval of their ROR increased after the different strategies ranged from 14.6 to 48.1%. CONCLUSION: The study demonstrates the existence of a masking effect in FEDRA caused by certain highly reported DECs. Their exclusion according to the proposed unmasking strategies could reveal additional SDRs that may otherwise remain masked during routine signal detection activities. This study aimed to analyse how certain highly reported drug-event combinations (DECs) may obscure relevant safety signals. To address it, a method was developed to identify and remove these extreme DECs to get a clearer picture. The analysis was conducted using data from the Spanish spontaneous reporting database FEDRA, excluding vaccines (ATC group J07). A total of 287,145 DECs were assessed, of which 0.4% were identified as influential outliers. Approximately 14.3% of the drugs and 5.9% of the adverse drug reactions in FEDRA were involved in these extreme DECs. Between 14.6% and 48.1% of the DECs showed an important increase in their lower limit of the 95% confidence interval of their reporting odds ratios. In conclusion, the study demonstrates that certain highly reported drug-adverse drug reaction pairs can hide other important issues. By removing these extreme DECs, the potential to detect true safety signals increases.
European journal of pediatricsYun Lu, Yan Zhou, Yu Wu, Chun Liu, Lu Zhou, Ya Zou, Hua Wei, Fangqing Xie, Hongju Wang, Shihao Yan, Xirui Guo, Qinchuan Li, Jia Chen
UNLABELLED: This study aimed to assess infection-related adverse events (AEs) associated with calcineurin inhibitors (CNIs), including cyclosporine (CsA) and tacrolimus (TAC), in the pediatric population using combined postmarketing surveillance databases. Data from the FDA Adverse Event Reporting System (FAERS) and the Japanese Adverse Drug Event Report (JADER) were analyzed for pediatric patients. Disproportionality analysis was conducted to evaluate AEs. Additionally, subgroup analysis, fatal outcome assessment, and time-to-onset (TTO) analysis were also performed. A total of 426 CsA-related and 1284 TAC-related infection reports were identified in FAERS, with corresponding findings in JADER (CsA, n = 140; TAC, n = 323). Cross-database analysis identified positive overlapping signals for BK virus, cytomegalovirus, and Epstein-Barr virus infections for both drugs. Age- and sex-stratified analysis revealed differences in infection-related reporting patterns, with the percentage of fatal outcomes varying significantly across age groups and being highest in the youngest subgroups: 3-5 years for CsA (29.2%) and 0-2 years for TAC (18.4%) (P < 0.05). CONCLUSION: This study provides real-world evidence on infection-related safety signals associated with CNIs in pediatric patients, particularly for viral infections. The distinct signal profiles and age- and sex-specific patterns highlight the importance of individualized infection risk monitoring and immunosuppressive management strategies. WHAT IS KNOWN: • Calcineurin inhibitors (CNIs), including cyclosporine (CsA) and tacrolimus (TAC), are widely used in pediatric transplantation and immune-mediated diseases, but infection remains an important safety concern during CNI-based immunosuppression. • Evidence on the comparative infection-related safety profiles of these two agents in children remains limited. WHAT IS NEW: • This is the first dual-database pharmacovigilance analysis to characterize infection-related safety signals associated with CsA and TAC in pediatric patients. • Age- and sex-specific patterns were identified, with fatal outcomes significantly associated with younger age groups for both drugs.
The AAPS journalInês Lucas, João Sousa, Carla Vitorino
Digital transformation in pharmaceutical regulatory affairs is accelerating as global submissions grow in complexity and traditional document-based workflows reach their limits. Artificial intelligence (AI), particularly natural language processing (NLP), is increasingly being explored to support regulatory data management, document preparation, and decision support activities. This review examines AI adoption across pharmaceutical regulatory science, including initiatives from major regulatory agencies, AI-supported regulatory workflows, and emerging governance and interoperability frameworks. Current applications include document classification, data extraction, Common Technical Document (CTD) support, pharmacovigilance, and predictive analytics. Key implementation challenges, including explainability, traceability, validation, data quality, interoperability, cybersecurity, and Good Practice (GxP) compliance requirements, are critically discussed. The review further examines emerging regulatory data ecosystems and governance frameworks that may support the responsible integration of AI into regulatory processes. Collectively, these developments highlight the potential of AI to support more structured, interoperable, and efficient regulatory systems while maintaining regulatory oversight and accountability. Current evidence suggests that AI implementation has progressed from conceptual research toward early operational deployment. However, robust evidence demonstrating sustained improvements in regulatory performance and long-term operational impact remains limited.
Revista da Sociedade Brasileira de Medicina TropicalAna Luiza Silva Rezende, Janaína de Pina Carvalho, Beatriz Prado Noronha, Sarah Nascimento Silva
BACKGROUND: Cutaneous leishmaniasis treatment is frequently associated with adverse events (AEs). We investigated the profile of AEs and examined their associations. METHODS: This prospective observational study was conducted at a Brazilian reference center to describe AEs according to their occurrence rate, severity, and seriousness. RESULTS: Among 318 AEs recorded in 92 patients, the highest AE occurrence rate was among those treated with miltefosine. The highest rate of serious AEs was among patients treated with liposomal amphotericin B. Local therapy using meglumine antimoniate showed a more favorable safety profile. CONCLUSION: The different profiles highlight the importance of pharmacovigilance in real-world settings.
Frontiers in immunologyXinchi Luan, Bingcheng Fan, Xuezhe Wang, Xiaoxuan Li, Yuhui Song, Xiaolei Zhang, Huhu Zhang, Ruolan Chen, Yi Li, Zelin Yang, Ning Liu, Weiwei Qi, Wensheng Qiu,…
BACKGROUND: Glycogen synthase kinase 3 beta (GSK-3β) inhibitors have received substantial attention for their therapeutic potential; however, their systemic safety profile remains incompletely characterized. OBJECTIVE: This study characterized the research landscape and exploratory safety-reporting signals associated with agents with reported GSK-3β activity by integrating bibliometric analysis, pharmacovigilance, and preliminary experimental assessment. METHODS: A multi-layered framework included bibliometric analysis, disproportionality analyses of FAERS, JADER, and CVARD reports, and transcriptomic profiling. SH-SY5Y cells were treated with 9-ING-41 (1 μM, 24 h); cell viability, qRT-PCR, and DCFH-DA-based intracellular oxidative-stress measurements were assessed. RESULTS: Bibliometric analysis showed sustained growth in GSK-3β-related research. Pharmacovigilance identified neurological and hematologic disproportionality signals across 27 System Organ Classes. In SH-SY5Y cells, 9-ING-41 was associated with modest changes in neuronal-function, inflammatory-response, and GSK-3β/Wnt-pathway transcripts, increased DCFH-DA fluorescence, and high cell viability. These cell-based observations are preliminary and do not establish clinical causality. CONCLUSION: The literature-derived study-drug panel showed exploratory neurological and hematologic reporting signals. Cross-database recurrence can prioritize hypotheses, whereas pharmacological heterogeneity and the limitations of spontaneous reporting require cautious interpretation. The SH-SY5Y experiments provide preliminary biological plausibility only.
PloS oneYao Zhou, Jie Gong, Lele Shen, Jie Ling, Shiting Wu, Anqi Ge, Lin Qi, Lifang Liu
Eribulin is used to treat metastatic breast cancer, but its adverse event (AE) profile requires comprehensive analysis using FDA Adverse Event Reporting System (FAERS) data to guide clinical use. We analyzed eribulin-related AE reports in FAERS (Q1 2011 - Q2 2024). Disproportionality analyses (ROR, PRR, MGPS, BCPNN) identified significant AE signals. Our study finalized 3570 AE reports with eribulin as the primary suspect drug from the FAERS database, including 7455 coded AE records. The included patients were predominantly female (90.1%) and aged 35-59 years (42.61%). Eribulin-associated AEs spanned 26 organ systems. Only "Blood and lymphatic system disorders" (n = 1,788) met significance across all four algorithms. At the preferred term level, 95 positive disproportionality signals emerged. Common expected AEs included neutropenia, febrile neutropenia, and leukopenia. Critically, serious unexpected AEs not consistently referenced in drug inserts were identified, such as pleural effusion, respiratory failure, and cardiac failure. Clinical use of eribulin requires great attention to blood and lymphatic adverse events, as well as enhanced monitoring and reporting of adverse events not mentioned in the insert to prevent serious AEs.
Drug design, development and therapyShaopeng Ming, Zhouyan Wu, Jingjing Li, Yicheng Su, Jianyou Yu, Hongtao Liu, Yanzhuo Zhang
BACKGROUND: Dexmedetomidine is widely used for clinical sedation, while clinical data suggest a potential correlation between its administration and diabetes insipidus. This exploratory study aimed to characterize molecular correlates linking dexmedetomidine and diabetes insipidus via multiomics and pharmacovigilance analysis. METHODS: FAERS data (2004-2024) were mined for disproportionality analysis to screen suggestive association signals. Network toxicology predicted shared targets of dexmedetomidine and diabetes insipidus, followed by GO/KEGG enrichment, PPI network construction and molecular docking. Transcriptome sequencing of 10 treated patients preliminarily validated bioinformatic correlations. RESULTS: A strong suggestive association signal was detected (ROR=471.47). A total of 105 overlapping targets were screened, among which IL6, IL10, INS, IL1B, AKT1 and IFNG served as core correlated hub genes. Docking confirmed stable binding between dexmedetomidine and these proteins. Enrichment revealed enriched MAPK cascade, kinase activity and PI3K/AKT pathways, which may correlate with abnormal AQP2 function. Transcriptomics identified differential expression of inflammation-immune genes after infusion, consistent with predicted molecular correlations. CONCLUSION: This study combined FAERS pharmacovigilance analysis, network toxicology, molecular docking and transcriptome sequencing to explore potential mechanisms of dexmedetomidine-induced diabetes insipidus. FAERS data mining uncovered a robust adverse signal, indicating a strong correlation between dexmedetomidine and diabetes insipidus for clinical reference. Six hub genes and PI3K/AKT, MAPK pathways may be associated with this side effect. These findings facilitate high-risk population management and individualized sedation, and supply molecular candidates for future validation studies.
BACKGROUND: Antineoplastic drugs are effective for malignant tumours, but they frequently cause severe adverse drug reactions (ADRs), which seriously affect treatment continuity and patient prognosis. Post - marketing real - world safety monitoring is therefore critical. METHODS: We conducted a retrospective pharmacovigilance study using FAERS (FDA Adverse Event Reporting System) data covering the period from 2004 Q1 to 2025 Q4. Safety signals were detected via disproportionate analysis, and associations were identified through LASSO and multivariable logistic regression, followed by an analysis of time to onset (TTO) for haematotoxicity. RESULTS: A total of 347, 248 reports involving 367 antitumour drugs were analysed. Female patients (46.51%) outnumbered males (38.73%), with a median age of 59 years; patients aged 18 - 64.9 years represented 34.79% of the cohort. Fatal or life - threatening events were recorded in 78, 824 cases (22.70%). The leading drugs by report frequency were lenalidomide (ROR (Reporting Odds Ratio) = 1.51), methotrexate (ROR = 2.98), and rituximab (ROR = 3.69). Most events (61.65%) emerged within the first month, and median TTO differ significantly by sex. Furthermore, WSP analysis revealed that 26 of the top 30 drugs followed an early failure pattern. CONCLUSIONS: Our findings reveal a strong signal of haematological toxicity associated with antitumour agents, supporting enhanced routine monitoring to mitigate clinical risks. Nevertheless, confirmation through pharmacoepidemiological studies with rigorous causality assessment is required, owing to the inherent limitations of the FAERS spontaneous reporting system.
PloS oneJiaLe Yang, JiaWen Liu, GuanBo Zhao, HeChen Li, Ge Sun
BACKGROUND: Entrectinib is effective for ROS1-positive non-small cell lung cancer (NSCLC), but its postmarketing safety profile remains incompletely characterized outside clinical trials. We conducted a dual-database pharmacovigilance study to characterize real-world reporting patterns, identify cross-database replicated adverse event signals, explore age- and sex-related reporting heterogeneity, and evaluate time-to-onset patterns. METHODS: Reports were retrieved from the U.S. Food and Drug Administration Adverse Event Reporting System (FAERS; 2020Q1-2025Q4; 520 patients; 1,573 preferred-term [PT] events) and the Japanese Adverse Drug Event Report database (JADER; 2020Q1-2025Q3; 254 patients; 393 PT events). Molecular fusion status was not available for verification. Four disproportionality methods were applied, with reporting odds ratio (ROR) as the primary signal criterion. RESULTS: At the system organ class level, shared positive signals involved nervous system disorders (FAERS/JADER ROR: 4.48/4.76), cardiac disorders (3.20/5.43), and renal and urinary disorders (2.04/3.92). At the PT level, 23 signals were detected in both databases, whereas 58 PTs were unique to FAERS and 8 to JADER. Representative shared PT signals included dizziness (12.94/14.33), taste disorder (38.03/13.28), renal impairment (6.89/8.18), blood creatinine increased (8.31/24.99), cardiac failure (7.57/8.90), cognitive disorder (18.04/78.29), ataxia (51.73/351.45), syncope (8.87/89.45), myocarditis (3.47/5.91), electrocardiogram QT prolonged (4.01/5.35), and hyperuricaemia (7.36/59.63). Subgroup analyses suggested exploratory age- and sex-related reporting heterogeneity, including relatively more renal and mobility-related reports in older patients and female predominance for ataxia in FAERS. In the FAERS-based time-to-onset analysis, 203 of 520 reports (39.0%) had valid onset data. The median time to onset was 13 days, 70.94% of evaluable reports had onset dates within 30 days, and Weibull analysis suggested an early-failure pattern. CONCLUSIONS: Entrectinib-associated reports in NSCLC showed reproducible neurologic, cardiac, renal, and laboratory-related disproportional reporting signals across FAERS and JADER. These findings may help prioritize early safety monitoring but should be interpreted as hypothesis-generating signals rather than evidence of incidence or causality.
Frontiers in neurologyAdelė Antanaitytė, Mantas Jokubaitis, Jorinta Jokubaitė, Kristina Ryliškienė
INTRODUCTION: Calcitonin gene-related peptide (CGRP)-targeting therapies have transformed migraine management and are generally associated with a favorable safety profile. However, accumulating real-world evidence suggests a potential association with alopecia that was not systematically captured in pre-approval clinical trials. This review aimed to synthesize current evidence on alopecia associated with CGRP-targeting therapies, with a focus on clinical patterns, pharmacovigilance signals and underlying pathophysiological mechanisms. METHODS: A structured literature search of PubMed was conducted to identify relevant evidence on CGRP-targeting therapies and hair loss through August, 2026. Eligible studies included clinical trials, observational studies, pharmacovigilance analyses and case reports reporting alopecia outcomes. Reference lists were screened for additional relevant studies. Due to heterogeneity of study designs and predominance of low-level evidence, findings were synthesized narratively without formal quality assessment. RESULTS: Evidence from pharmacovigilance data and case-based observations suggests a reproducible signal of predominantly reversible, non-scarring alopecia occurring within weeks to months after initiation of CGRP-targeting therapies. Reports span multiple monoclonal antibodies and gepants, supporting a potential class effect. However, the available evidence is largely derived from spontaneous reporting systems and case series, limiting causal inference and precluding reliable estimation of incidence. Proposed mechanisms include neuroimmune dysregulation, impairment of hair follicle immune privilege, reduced microvascular perfusion, and disruption of the CGRP-insulin-like growth factor-1 axis. CONCLUSION: Alopecia has emerged as a potentially clinically relevant safety signal associated with CGRP-targeting therapies. Although the association is biologically plausible, the available evidence does not establish causality, frequency or the typical clinical course of this adverse event. Most cases appear to be mild, non-scarring and reversible. Given the expanding use of these therapies, increased clinical awareness and individualized management are essential. Future prospective studies with standardized dermatological assessment are needed to better define incidence, risk factors and underlying mechanisms.
BACKGROUND: Drug-associated esophageal ulceration and perforation are uncommon but clinically serious adverse events. This study characterized post-marketing reporting patterns for these events across three spontaneous reporting systems. METHODS: Reports from the U.S. Food and Drug Administration Adverse Event Reporting System (FAERS), Japanese Adverse Drug Event Report database (JADER), and Canada Vigilance from 2004 to 2025 were analyzed using predefined Medical Dictionary for Regulatory Activities (MedDRA) preferred terms. Disproportionality was assessed using the reporting odds ratio (ROR), proportional reporting ratio (PRR), Bayesian confidence propagation neural network (BCPNN), and multi-item gamma Poisson shrinker (MGPS). In FAERS, least absolute shrinkage and selection operator (LASSO) regression followed by multivariable logistic regression was used for exploratory signal prioritization. Reported time-to-onset was summarized for drug-report records with valid therapy start and event onset dates. RESULTS: FAERS contained 8,061 reports meeting the prespecified composite case definition. Among the 50 most frequently reported FAERS primary-suspect drugs, the largest disproportionality estimates were observed for doxycycline (ROR 41.15, 95% confidence interval [CI] 36.47-46.43), alendronic acid (ROR 21.45, 95% CI 19.14-24.04), and clindamycin (ROR 18.59, 95% CI 15.40-22.44). JADER and Canada Vigilance showed descriptively overlapping but heterogeneous reporting patterns. Reported time-to-onset was heterogeneous, with both short and long recorded intervals. CONCLUSIONS: This multidatabase analysis identified prominent reporting signals for established pill-injury drugs and exploratory reporting patterns involving other drug categories. These findings are hypothesis-generating and should not be interpreted as estimates of incidence, comparative clinical risk, or causal effects.
International journal of public healthGita Kusnadi, Grace Wangge, Arif Perdana
OBJECTIVE: To examine the application of artificial intelligence (AI) in pharmacovigilance across the Asia-Pacific and identify reported methodological implementation challenges. METHODS: MEDLINE, Scopus, and Google Scholar were searched using terms related to artificial intelligence, computational signal detection, pharmacovigilance, and Asia-Pacific countries. Peer-reviewed original studies published in English were included. PRISMA 2020 guideline was followed. RESULTS: We included 64 studies in 14 countries primarily focused on 1) Adverse Drug Reaction (ADR) identification, 2) ADR prediction and risk factor modelling, 3) Drug safety, monitoring, and evaluation, 4) Predictive modelling, and 5) Data information management. Machine Learning (ML) techniques were the most commonly applied AI methods in pharmacovigilance, followed by natural language processing, deep learning, neural networks, and symbolic and explainable AI. Disproportionality analysis methods were also commonly used across studies. Some challenges reported were relevant to data quality issues, generalizability, clinical workflow integration, implementation technicalities, and cultural barriers. CONCLUSION: To overcome the challenges of AI application in the Asia-Pacific, a tiered implementation strategy can be employed through establishing a regional collaboration framework and taking into account disparities in technological maturity across countries.
This study aimed to compare the safety profiles of intravenous and oral linezolid by analyzing adverse event (AE) reports from the US Food and Drug Administration Adverse Event Reporting System. We extracted linezolid-related AE reports from US Food and Drug Administration Adverse Event Reporting System (2004 Q1-2025 Q2). After de-duplication and exclusion of reports with unknown administration routes, 7591 reports (intravenous: 3445; oral: 4146) were included. Disproportionality analyses using reporting odds ratio, proportional reporting ratio, Bayesian confidence propagation neural network, and multi-item gamma Poisson shrinker were performed to detect AE signals. Time-to-onset analysis was conducted using Weibull distribution. Intravenous administration was associated with acute, severe AEs such as neonatal complications, acute kidney injury, and serotonin syndrome, predominantly occurring within 30 days (median onset: 6 days). Oral administration showed stronger signals for chronic toxicities, including peripheral neuropathy, optic neuropathy, anemia, and metabolic disturbances, with a later median onset of 11 days and prolonged event distribution. Hematologic toxicity patterns also differed: the oral route was linked to anemia and pancytopenia, while the intravenous route was associated with leukopenia and agranulocytosis. The safety profiles of linezolid differ significantly between intravenous and oral routes, largely driven by treatment duration, patient population, and clinical setting. These findings support route-specific monitoring strategies: acute toxicity surveillance for intravenous therapy and long-term neurotoxicity and hematologic monitoring for oral therapy.
Frontiers in public healthDorota Kopciuch, Agata Kapcińska, Krzysztof Kus, Emilio Russo
BACKGROUND: Medical cannabis is increasingly used in routine care, but its pharmacovigilance profile remains difficult to monitor because products, formulations, doses, routes of administration and clinical indications vary substantially. This study assessed pharmacovigilance knowledge, adverse drug reaction (ADR) reporting practice and attitudes toward medical cannabis among physicians in Poland, with particular attention to factors associated with ADR reporting. METHODS: We conducted a cross-sectional analytical survey among 253 clinically active physicians in Poland using an anonymous self-administered questionnaire. The survey covered demographic and professional characteristics, pharmacovigilance knowledge and awareness, general ADR reporting, clinical exposure to medical cannabis and cannabis-related safety perceptions. Crude associations were examined using chi-square tests, Fisher's exact tests for cannabis-specific sparse-data comparisons, Cramer's V, odds ratios (ORs) and 95% confidence intervals, with Benjamini-Hochberg false-discovery-rate adjustment. For the primary outcome of general ADR reporting, a forced-entry multivariable logistic regression included PV-purpose knowledge, any medical specialization, professional experience greater than 10 years, hospital/university workplace, medical cannabis exposure and sex. Cannabis-specific reporting was analyzed separately using exact methods because only 19 events occurred. RESULTS: Most respondents reported clinical exposure to medical cannabis (81.0%), whereas only 13.4% reported adequate formal education in this area. In crude analysis, knowledge of the purpose of pharmacovigilance showed the strongest association with ADR reporting (OR 46.58; 95% CI 21.20-102.32; p < 0.001; q < 0.001). In the adjusted model, PV-purpose knowledge (adjusted OR [aOR] 24.71; 95% CI 9.40-64.96), any medical specialization (aOR 9.97; 95% CI 3.66-27.20) and hospital/university workplace (aOR 5.72; 95% CI 2.10-15.55) retained positive associations with general ADR reporting. Professional experience greater than 10 years, medical cannabis exposure and sex were not statistically significant after adjustment. Cannabis-specific ADR reporting was uncommon. CONCLUSIONS: In this sample, PV-purpose knowledge, medical specialization and hospital/university workplace retained positive associations with general ADR reporting after multivariable adjustment. Crude associations for longer professional experience and medical cannabis exposure were attenuated after adjustment. The combination of frequent medical cannabis exposure and rare cannabis-specific ADR reporting suggests an exposure-reporting gap. The observational findings support practical, therapy-specific pharmacovigilance training but should not be interpreted causally.
BACKGROUND: Immune checkpoint inhibitors (ICIs) have exhibited remarkable clinical benefits in the treatment of lung cancer, yet a large-scale pharmacovigilance investigation on ICI-related thyroid dysfunction (ICI-TD) remained limited. This study aimed to comprehensively analyze the demographic characteristics of ICI-TD in lung cancer patients and the pharmacovigilance signal patterns of the adverse events (AEs) using data from the Food and Drug Administration Adverse Event Reporting System (FAERS) database. METHODS: All existing data ranging from January 1st 2014 to March 31st 2026 were retrieved from the FAERS database. In descriptive analysis, the demographic characteristics of ICI-TD in lung cancer patients were collected and categorized. Disproportionality analysis was conducted to evaluate the signal characteristics of ICI-TD by employing three disproportional algorithms, namely reporting odds ratio (ROR), proportional reporting ratio (PRR), and information component (IC). RESULTS: A total of 2,001 reports of ICI-TD during treatment of lung cancer were identified after data processing, involving 1,758 patients and 10 ICIs. In general, ICI-TD were mainly reported in male patients (n = 959, 54.55%) and patients aged ≥ 65 years (n = 856, 48.69%; median [lower quartile (Q1), upper quartile (Q3)] = 67 [60, 73]). After data cleaning, we found that reports from 752 patients had valid time-to-onset (TTO) values. The median TTO of ICI-TD was 42 days [Q1, Q3] = [14, 98], and ICI-TD was predominantly reported within 30 days after ICI administration (n = 323, 42.95% of all 752 patients). The vast majority of patients were reported as having serious reaction outcome (n = 1,690, 96.13%). Among all ICIs, nivolumab had the greatest number of ICI-TD reports (a = 655, 32.73% of all 2,001 reports, ROR025 = 2.97, IC025 = 1.32, PRR = 3.21, χ2 = 782.34). The most frequently reported ICI-TD was "hypothyroidism" (a = 901, 45.03% of all 2,001 reports, ROR025 = 4.83, IC025 = 1.22, PRR = 5.36, χ2 = 1,138.77). CONCLUSIONS: The present study provided a comprehensive overview of the demographic features of patients and pharmacovigilance signal patterns of ICI-TD during the treatment of lung cancer. Our findings would offer a referential perspective for the clinical understanding of the pharmacovigilance signals of ICI-TD.
BMC medical informatics and decision makingYuLong He, Yan Mao, XinYu Wang
BACKGROUND: The rapid deployment of mRNA vaccines during the COVID-19 pandemic exposed limitations in traditional pharmacovigilance systems, including delayed reporting, high underreporting rates, and inability to calculate true incidence. Machine learning (ML) offers new pathways to overcome these challenges by integrating multi-source real-world data. METHODS: We systematically reviewed English-language studies from database inception to June 2026. Searches were performed in PubMed, Embase, and Web of Science. Two reviewers independently screened records. Given substantial heterogeneity across ML tasks (signal detection, text extraction, risk prediction, prognosis stratification), algorithms, data sources, and metrics, we performed narrative synthesis. Risk of bias was assessed using adapted QUADAS-2. RESULTS: We identified 43 studies. For adverse-event prediction, tree-based models reported AUCs of 0.85-0.87, though estimates derive from heterogeneous settings. NLP reduced redundant signals by 17% in vaccine reporting systems. For myocarditis, ML models reached AUCs up to 0.899 in cardiovascular cohorts, but direct mRNA vaccine applications remain limited and retrospective. Emerging platforms (self-amplifying and tumor mRNA vaccines) lack post-marketing data, rendering ML applications largely conceptual. CONCLUSION: ML-assisted pharmacovigilance enables a shift from passive to active, intelligent monitoring. Despite challenges in data quality, model interpretability, and regulatory approval, intelligent pharmacovigilance systems will become essential infrastructure for safeguarding public health.
The concomitant use of ceftriaxone and proton pump inhibitors (PPIs) is common in hospital practice. However, it is unclear whether individual PPIs differ in their effects on QT interval prolongation, ventricular arrhythmia, or cardiac arrest, collectively termed as QVC events. We conducted a two-stage, real-world study. First, we screened the United States Food and Drug Administration Adverse Event Reporting System (FAERS) and the Canada Vigilance Adverse Reaction (CVAR) database with standard disproportionality measures (reporting odds ratio and proportional reporting ratio) and six drug-drug interaction (DDI) algorithms to identify combination signals that exceeded component signals. Second, we validated signal-positive combinations in the Medical Information Mart for Intensive Care IV (MIMIC-IV) intensive care unit (ICU) electronic health record (EHR) cohort by assembling adult inpatients with overlapping ceftriaxone-PPI exposures. The primary outcome was 28-day QVC events. Multivariable Cox proportional hazards models were the main analysis and complemented by propensity score matching, inverse probability of treatment weighting, and Fine-Gray competing-risk models. To address external generalisability, an additional validation was performed using ECG-ViEW II, an Asian electrocardiogram-linked real-world database. The combination of ceftriaxone and lansoprazole was significantly associated with QVC events, revealing notable DDIs (e.g., in FAERS, Ω025 = 0.54). To validate these findings, a cohort of 5,594 patients receiving ceftriaxone combined with PPIs from the MIMIC-IV database was analyzed using Cox proportional hazards models. The analyses corroborated the initial findings (lansoprazole vs. other PPIs, multivariate HR = 1.30; 95% CI: 1.10-1.54), with the risk associated with the three PPI combinations ranked as lansoprazole > pantoprazole > omeprazole. ECG-ViEW II provided supportive Asian external validation, showing a higher QVC risk for ceftriaxone plus lansoprazole than for ceftriaxone plus other PPIs. Evidence from two national pharmacovigilance systems and an ICU EHR cohort indicated that PPI choice modified cardiac safety during ceftriaxone therapy. Lansoprazole co-use confers a higher risk of QVC, whereas omeprazole appears relatively safer. Therefore, prospective confirmation is warranted.
Human vaccines & immunotherapeuticsXueqing Ma, Jiaojiao Fan, Ben Liu
Dupilumab is effective for the treatment of moderate‑to‑severe type 2 inflammatory diseases; however, its real‑world safety profile in pediatric patients remains to be further clarified. This retrospective pharmacovigilance study analyzed U.S. Food and Drug Administration Adverse Event Reporting System (FAERS) data from 2017 to 2025, including 44,593 reports where dupilumab was the primary suspect drug in children. Disproportionality analyses (ROR, PRR, BCPNN, MGPS), time-to-onset analysis, stratified analysis, and Weibull modeling assessed adverse event (AE) distribution and patterns. Results showed balanced sex distribution (male 50.1%, female 48.6%), with adolescents (12-17y) most common (45.3%), and report counts increasing over time. 93 positive safety signals were detected. The most common core AEs were rash (4,105 cases; ROR = 4.55, 95% CI: 4.38-4.72), injection site pain (3,582 cases; ROR = 8.59, 95% CI: 8.22-8.97), and skin exfoliation (1,834 cases; ROR = 3.41, 95% CI: 3.23-3.59). Potential unlisted signals included vitiligo (33 cases; ROR = 10.46, 95% CI: 6.52-16.78), skin depigmentation (32 cases; ROR = 13.04, 95% CI: 7.85-21.66), and eye color change (13 cases; ROR = 7.06, 95% CI: 3.54-14.11). Stratified analyses indicated distinct patterns of adverse‑event reporting across age and sex strata. Among cases with complete temporal information, most AEs occurred within 30d after dupilumab initiation, with a median time‑to‑onset of 14d. In conclusion, this large-scale spontaneous-reporting study detected potential safety signals in children. Given FAERS data limitations, the associations should be viewed as hypothesis-generating rather than confirmatory, and may inform future signal validation and prospective studies.
Current oncology (Toronto, Ont.)Zaid Ahmed, Rashid Sayyid, Omid Yazdanpanah, Ravand Samaeekia, Arash Rezazadeh Kalebasty, David I Lee, Mohammed Shahait
Therapeutic options for advanced prostate cancer have expanded in recent years, incorporating multiple-system treatment approaches with differing mechanisms of action. However, comparative real-world safety data following drug approval remain limited. As such, the aim of this study is to characterize adverse events and disproportionate safety signals among advanced prostate cancer therapies using the FDA Adverse Event Reporting System (FAERS). A retrospective pharmacovigilance study of FAERS reports evaluated enzalutamide, darolutamide, apalutamide, abiraterone acetate, relugolix, niraparib/abiraterone, talazoparib, rucaparib, cabazitaxel, sipuleucel-T, and lutetium-177 vipivotide. Adverse events were categorized by System Organ Class and Preferred Terms. Reporting odds ratios (RORs) with 95% confidence intervals identified safety signals. Among 172,440 reports, most involved patients aged 65-85 years. Cabazitaxel had the highest proportion of serious reports (86.6%) and deaths (22%), whereas relugolix had the lowest (23.3% and 4.8%). Nervous system disorders predominated with enzalutamide and darolutamide, gastrointestinal disorders with abiraterone, rucaparib, and niraparib/abiraterone, and hematologic toxicities with cabazitaxel, talazoparib, and lutetium-177 vipivotide. Significant safety signals were identified for abiraterone and cabazitaxel, but not other therapies. The absence of a detected signal should not be interpreted as evidence of safety or equivalence, as reporting volume, detection bias, and statistical power varied across therapies. Overall, the therapies demonstrated distinct toxicity profiles, which may inform treatment selection, toxicity monitoring, and patient counseling.
Indian journal of pharmacologyEmmanuel Delali Kofi Fiagbey, Linda Nyame, Zexiang Bao, Zikomo Gaudence Kipanga, Isaac Nii Lante Lamptey, Daniel Kofi Nyame, Helen Dekyem, Yusuf Nasir Maigari,…
OBJECTIVES: To assess the knowledge, awareness, and practices of adverse drug reaction (ADR) reporting among healthcare professionals (HCPs) in the Ashanti Region of Ghana and to further identify the predictors of HCPs' reporting behaviors. METHODS: A cross-sectional survey was conducted among 326 HCPs from various healthcare facilities across the Ashanti Region. Data were collected using structured, pretested questionnaires, and analyzed with R and STATA software, using descriptive statistics and univariate logistic regression to explore the factors associated with reporting behaviors. RESULTS: 93.6% of HCPs were aware of pharmacovigilance (PV), with 85.6% acknowledging ADR reporting as a professional mandate. However, only 49.4% were aware of Ghana's National Pharmacovigilance Center, with only 50% having submitted fully-filled ADR forms. Barriers to reporting included unavailability of forms (30.3%) and lack of training (23.3%). Age, experience, and previous training significantly influenced reporting (P < 0.05). The most ADR-implicated drugs were antibiotics (36.5%) and antihypertensives (17.8%), with gastrointestinal (34.4%), and central nervous system (32.4%) manifestations most often reported. CONCLUSIONS: This study reveals relevant insights into the PV landscape in Ghana's Ashanti Region. Despite high awareness among HCPs in the Ashanti Region, operational knowledge and reporting practice were suboptimal, with most cited barriers being unavailability of forms and lack of training. Our findings highlight the need for targeted interventions such as continuous PV training, improved access to reporting tools, and inclusion of underrepresented HCP cadres like Medical Herbalists to strengthen PV and improve ADR reporting in Ghana.
Human psychopharmacologySara Jiménez-Fernández, Carlos De Las Cuevas, Emilio J Sanz, Jose de Leon
OBJECTIVES: This study aimed to: (1) systematically review published cases of accidental clozapine poisoning in children and adolescents, and (2) study accidental pediatric clozapine intoxications reported to VigiBase, the global pharmacovigilance database. METHODS: A systematic review was conducted following PRISMA guidelines, and a retrospective analysis of accidental pediatric clozapine intoxication cases was performed using VigiBase. RESULTS: A systematic review identified 5 studies describing 8 cases of accidental clozapine ingestion in children aged 10 months to 10 years. The most common clinical manifestations were acute neuro-respiratory complications. One fatality was reported, in a two-year-old child. The VigiBase observational analysis included 32 accidental intoxication cases, 3 in adolescents and 29 in children under seven years of age. Among adolescents, all cases resulted from in-hospital medication administration errors; none were fatal. In the children, nearly half required prolonged hospitalization, and 2 resulted in death due to aspiration pneumonia or cardiopulmonary failure. CONCLUSIONS: Accidental clozapine poisoning in pediatric populations, although rare, can result in severe and potentially life-threatening outcomes, particularly in very young children. At home, preventive strategies are essential to reduce accidental overdose. The 3 cases of hospitalized adolescents are relatively recent, and many hospitals currently have systems in place to prevent medication errors.
Briefings in bioinformaticsOlivér M Balogh, Mátyás Pétervári, Áron M Csernák, Eszter Puhl, András Horváth, Péter Ferdinandy, Bence Ágg
Post-marketing surveillance is crucial for drug safety, yet the tools of pharmacovigilance rely solely on text-based data that may limit contemporary machine learning methodologies in the support of decision-making. With the recent surge of employing large language models (LLMs) for text-based tasks, there also arises an unmet need for a different approach which is not grounded in the linguistic patterns of unfiltered natural text, like LLMs, but rather based on real-world drug safety data. Here, we adapt contrastive learning algorithms to generate adverse event vector representations from spontaneous adverse event reports to serve as machine-readable (i.e. numerical) resources for downstream pharmacovigilance applications, such as drug-event association prediction for signal detection or causality assessment. We present comprehensive interpretability analyses of the resulting representations through density-based clustering, semantic evaluation, and comparison of multivariate dispersions, revealing patterns that reflect both functional and causal relations of the adverse events while also capturing drug-safety-related information better than existing medical terminologies and encoder-only LLMs. Furthermore, we demonstrate the applicability of our representations as input features in our downstream classifier model, outperforming the reporting odds ratio method, commonly used by regulatory agencies, and also LLM-generated representations (area under the receiver operating characteristic curve: 0.88 versus 0.76-0.83) on drug-event association prediction benchmarks. Therefore, we propose an interpretable adverse event vector representation, serving as a general resource that could enable the development of a wide array of machine learning applications to support decision-making in pharmacovigilance and facilitate patient safety.
Pharmacoepidemiology and drug safetyEmelith Florendo Cerbito, Lana Kattan, Hevna Dhulkifle, Hesham M Korashy, Zaid H Maayah
PURPOSE: Doxorubicin (DOX), an effective anticancer agent, is associated with dose-dependent cardiovascular toxicity. Understanding its mechanisms and risk factors will facilitate novel interventions to minimize DOX-induced cardiovascular toxicity. Thus, this study performed data mining of the FDA Adverse Event Reporting System (FAERS) to detect and analyze DOX-induced cardiac-related adverse events (AEs). METHODS: Data from 2004 to 2025, where DOX was the primary suspect, were extracted from FAERS via OpenVigil. Using the Medical Dictionary for Regulatory Activities, AEs were categorized into preferred terms (PTs) and system organ classes (SOCs). This study used descriptive analysis and signal detection algorithms including Proportional Reporting Ratio (PRR), Reporting Odds Ratio (ROR), Multi-item Gamma Poisson Shrinker (MGPS), and Bayesian Confidence Propagation Neural Network (BCPNN). RESULTS: A total of 2421 reports of DOX-associated AEs were extracted in FAERS. Reports primarily involved female cancer patients (55.02%) between 18and 65 years (48.12%) from the United States. Most signals were of moderate to strong signal intensity, with cardiotoxicity, cardiac failure, and cardiomyopathy being the most reported. Interestingly, while cardiac disorders were mainly found to be strongly linked to DOX, novel vascular AEs like endothelial dysfunction were also among the top signals strongly associated with DOX. Out of the six serious AEs associated with DOX-induced cardiotoxicities, "Death" and "Initial/prolonged hospitalization" were the most commonly reported. CONCLUSION: This study provides valuable insights into DOX-induced cardiovascular toxicity using real-world data from FAERS. Vascular events like endothelial dysfunction were shown to be significant novel AEs experienced by cancer patients undergoing treatment with DOX. Thus, careful monitoring of potential DOX-AEs associations will help improve the risk-benefit ratio of DOX-treated patients.
BACKGROUND: Bexagliflozin and velagliflozin are currently the only sodium-glucose cotransporter 2 inhibitors approved by the United States Food and Drug Administration (FDA) for treating diabetes in cats. There are limited real-world safety reports on their associated adverse events (AEs). HYPOTHESIS/OBJECTIVES: To analyze AEs associated with bexagliflozin and velagliflozin using real-world data from the FDA Animal Drug Adverse Events (ADAE) database. ANIMALS: None. METHODS: AE reports submitted for cats receiving bexagliflozin and velagliflozin. Data were obtained from the ADAE database up to the second quarter of 2025. Disproportionality analysis was conducted employing four algorithms: the reporting odds ratio (ROR), proportional reporting ratio, Bayesian confidence propagation neural network, and multi-item gamma Poisson shrinker. RESULTS: Of the 34 187 AE reports for cats, 2876 were related to bexagliflozin and 2776 to velagliflozin. AEs associated with bexagliflozin spanned 22 system organ classes (SOCs), with stronger signals observed for weight fluctuation, glucosuria, and diabetic ketoacidosis. Velagliflozin-related AEs occurred in 19 SOCs, with higher RORs for ketonuria, hypochloremia, and acid-base disorders. Bexagliflozin was associated with stronger signals for DKA and ketosis, whereas velagliflozin showed stronger signals for ketonuria and hypochloremia. Velagliflozin-related AEs occurred earlier (median onset: 5 vs. 9 days), resolved more quickly (median duration: 8 vs. 14 days). CONCLUSIONS AND CLINICAL IMPORTANCE: This study provides a comprehensive safety profile of bexagliflozin and velagliflozin for diabetes in cats. These findings support veterinarians in implementing differentiated risk monitoring and individualized therapeutic decisions in the management of diabetes in cats.
Cancer medicineSaikat Mandal, Manideepa Maji, Arkadeep Dhali
BACKGROUND: T-cell-engaging bispecific antibodies are increasingly utilised in haematological malignancies and are beginning to be adopted in solid-tumour treatment protocols. Cytokine release syndrome and neurotoxicity are well recognised, but target-specific dermatological adverse event reporting patterns remain poorly defined. METHODS: We analysed 12,333,305 deduplicated FAERS reports from 2016Q1 to 2026Q1, including 23,386 reports exposed to ten T-cell-engaging bispecific antibodies grouped by target: CD19, CD20, BCMA, GPRC5D and DLL3. A disease-matched haematologic-oncology comparator cohort contained 1,062,195 reports. Canada Vigilance provided a secondary directional comparison using 755,911 reports from 2016Q1 to 2025Q4 including 662 unique exposed reports. Outcomes were any cutaneous adverse event, broad severe cutaneous adverse reaction, and narrow Stevens-Johnson syndrome/toxic epidermal necrolysis. Multivariable models adjusted for age, sex, cancer indication, polypharmacy and classical culprit drugs. Cox models assessed target-specific timing. RESULTS: Among exposed FAERS reports, 1394 (5.96%) included any cutaneous adverse event, 61 (0.26%) broad severe cutaneous adverse reaction and 9 (0.04%) narrow Stevens-Johnson syndrome/toxic epidermal necrolysis. After disease matching, no clear class-level excess of severe cutaneous reaction reporting was observed. Talquetamab had the highest cutaneous reporting proportion (471/1822; 25.8%) and an elevated timing estimate (hazard ratio 1.34; 95% CI, 1.06-1.69), with enrichment for skin, hair, sweat-gland and rash phenotypes. Tarlatamab showed an exploratory early-onset pattern (hazard ratio 3.36; 95% CI, 1.68-6.72; median onset, 4 days), based on only eight reports with usable latency data. Canada Vigilance showed a similar direction of reporting for GPRC5D/any cutaneous adverse events. CONCLUSIONS: T-cell-engaging bispecific antibodies did not show clear class-level excess in severe cutaneous adverse-event reporting after disease matching. Talquetamab showed higher reporting of skin changes, hyperhidrosis, hair abnormalities and rash, whereas tarlatamab showed a possible early-onset cutaneous reporting pattern based on a small number of reports with usable timing data. These findings support target- and timing-aware dermatological monitoring and require prospective confirmation.
Frontiers in immunologyHan Chen, Yongqi Shan, Hanfang Xu, Keer Xuan, Tianshu Ren, Qingchun Zhao
BACKGROUND: Although immune checkpoint inhibitor (ICI) therapy has revolutionized cancer treatment, the incidence of cardiac toxicity associated with ICIs remains unclear. This study aimed to evaluate the cardiac toxicity risks associated with ICI and identify the most common types of cardiac toxicity caused by ICI. A further objective was to determine the types of ICI that require the most monitoring of cardiac toxicity through systematic review and network meta-analysis, assisted by pharmacovigilance studies. DESIGN: Systematic review and network meta-analysis, complemented by a pharmacovigilance study of the FAERS database. DATA SOURCE AND METHODS: A systematic search of electronic databases up to November 2025 was conducted. Randomized controlled trials (RCTs) were eligible if they compared ICIs with appropriate controls without restrictions. Data were analyzed using random-effects pairwise and network meta-analyses to evaluate cardiac toxicity. Disproportionality analysis was performed using FAERS data from 2011 to 2026 (Q1), employing PRR, ROR, IC, and EBGM to quantify the risk and incidence of cardiac toxicity associated with ICIs. RESULTS: In total, 135 RCTs were included in the network meta-analysis. Pairwise meta-analysis demonstrated that ICIs can induce cardiac toxicity in four key manifestations, which were selected through pairwise meta-analysis: acute myocardial infarction, cardiac arrest, myocarditis, and ventricular tachycardia. Network meta-analysis indicated that both PD-1/PD-L1 monotherapy and combination therapies present a higher risk of cardiac toxicity. The risks associated with PD-1 and PD-L1 agents were relatively elevated when used as monotherapies. Furthermore, combinations of PD-1 (nivolumab, pembrolizumab) and PD-L1 (avelumab) with CTLA-4 or other small molecule targeted inhibitors were associated with increased risk of the four aforementioned cardiac toxicities. Disproportionality analysis revealed that both PD-1 and PD-L1 inhibitors carry risks of myocarditis, and the combination of bevacizumab, relatlimab, and ipilimumab may cause increased cardiac toxicity. CONCLUSIONS: ICIs, particularly PD-1/PD-L1 inhibitors, increase the risk of cardiac toxicity. PD-1 (nivolumab, pembrolizumab) and PD-L1 (atezolizumab, avelumab) agents present a higher risk of myocarditis and acute myocardial infarction. Moreover, combination regimens involving PD-1/PD-L1 further elevate the risk of cardiac toxicity, underscoring the necessity for vigilant monitoring in patients with underlying heart disease. CLINICAL TRIAL REGISTRATION: https://www.crd.york.ac.uk/PROSPERO/, identifier CRD420261357478.