This retrospective pharmacovigilance study used the FDA Adverse Event Reporting System (FAERS) to compare the real-world adverse event (AE) profiles of enfortumab vedotin (EV) monotherapy versus EV combined with pembrolizumab (EV+P). The primary aim was to quantify differences in reporting frequency across System Organ Class (SOC), High-Level Group Term (HLGT), and Preferred Term (PT) hierarchies and to assess time-to-onset patterns and validate notable signals (notably pneumonitis) via narrative literature review.
FAERS data spanning 2020 Q4 through 2025 Q3 were downloaded and linked across seven FAERS file domains using PRIMARYID and CASEID. Initial import yielded 8,649,232 reports before deduplication. FAERS reaction terms were mapped to MedDRA v27 PTs and aggregated to HLGT and SOC levels. Institutional review board approval was waived because only de-identified public-domain data were used.
EV reports were retrieved by searching generic and trade names (for example, enfortumab and Padcev). EV+P exposure was defined operationally as a FAERS report listing both enfortumab vedotin and pembrolizumab as suspect drugs in the same report. A sensitivity analysis was later performed restricted to reports where the exposure-defining drugs were coded as Primary Suspect.
A case/non-case design compared AE reporting for EV+P versus EV monotherapy (reference). Logistic regression estimated crude and adjusted odds ratios (ORs); age and sex were included in multivariable models. The Benjamini–Hochberg False Discovery Rate (FDR) correction addressed multiple testing; a signal required the lower 95% CI bound to exceed 1 and FDR-adjusted P < 0.05. Additional shrinkage analyses assessed the influence of sparse reporting. Time-to-onset was calculated from START_DT to EVENT_DT and analyzed with Kaplan–Meier curves; FDR-adjusted log-rank P values were reported.
A total of 3,004 AE reports were identified for EV monotherapy and 2,265 for combination therapy. Reports for combination therapy rose steadily over the study period, while EV monotherapy reports peaked in 2023 and decreased in 2024. Most reports involved patients aged ≥65 years, with the 70–75-year subgroup most represented. Men predominated (~3:1). Top reporting countries for EV monotherapy included Japan, the United States, and France; for EV+P, Japan, the United States, and Canada led reporting. The most frequent serious outcome category was initial or prolonged hospitalization, followed by death.
Of 22 SOCs analyzed, nine showed statistically significant higher reporting with EV+P compared with EV monotherapy (FDR P < 0.05). The largest SOC-level signals were for endocrine disorders (crude OR reported ~5.88; adjusted OR ~5.47), immune system disorders (OR ~2.03), and respiratory, thoracic and mediastinal disorders (OR ~1.71). Other SOCs with significant increased reporting included vascular, hepatobiliary, cardiac, renal and urinary, gastrointestinal, and investigations. Sensitivity analyses adjusted for age and sex and ROR shrinkage confirmed the directional consistency of these SOC-level findings.
Fifteen HLGT signals remained robust after FDR correction. Key HLGT signals included adrenal gland disorders (OR 9.21), thyroid gland disorders (OR 6.59), salivary gland conditions (OR 3.59), gastrointestinal inflammatory conditions (OR 2.76), immune disorders NEC (OR 3.56), lower respiratory tract disorders (OR 2.56), and myocardial disorders (OR 5.55). Some signals with high point estimates—such as hearing disorders (OR 10.64) and dental/gingival conditions (OR 9.31)—did not retain statistical robustness after multiple-testing correction in the main HLGT table.
At the PT level, top significant signals included hepatitis, immune-mediated enterocolitis, adrenal insufficiency, and pneumonitis among other specific event terms. The authors reported that vascular-related PTs, particularly those reflecting blood pressure fluctuations, featured prominently among top-ranked disproportionality signals. Detailed PT rankings and ORs are provided in the original analysis tables.
A complete-case analysis of START_DT and EVENT_DT fields was used to compute time-to-onset, and cumulative incidence curves were generated via Kaplan–Meier methods. The combination therapy suggested potentially earlier onset for some toxicities; statistically significant earlier onset was reported for eye disorders (FDR P = 0.047) and skin disorders (FDR P = 0.011). Missing or inaccurate date fields in FAERS limited the time-to-onset sample to complete cases.
A sensitivity analysis restricted to Primary Suspect reports retained 4,187 of the original 5,269 reports (≈79.5%) and produced SOC-level results directionally consistent with the primary analysis. Multivariable adjustment for age and sex and a shrinkage ROR analysis also supported the stability and directional consistency of the principal signals identified.
To validate pneumonitis signals, the authors conducted a PubMed and conference-abstract review of clinical trials and recent meeting data within the prior five years. Literature-derived incidence estimates supported a higher rate of any-grade pneumonitis with EV+P (reported as 10%) versus EV monotherapy (reported as 3.5%) in the reviewed sources. Extraction prioritized larger or more recent reports when cohorts overlapped.
This FAERS-based comparative analysis indicates that EV combined with pembrolizumab has a broader and more intense AE reporting profile than EV monotherapy, driven largely by immune-related toxicities (endocrine, immune system, and respiratory disorders) and potential synergistic organ injury. Several HLGT and PT-level signals—including adrenal and thyroid disorders, hepatitis, immune-mediated enterocolitis, pneumonitis, and myocardial disorders—merit heightened clinical vigilance. The authors emphasize the need for proactive monitoring and early intervention when using the EV+P combination. Detailed numerical tables and supplementary materials in the original source present complete signal rankings and statistical metrics.