Answers are grounded in this paper and cite their sources.
Background Immune checkpoint inhibitors (ICIs) are used with different chemotherapy backbones in small-cell lung cancer (SCLC) and non-small-cell lung cancer (NSCLC). We examined whether adverse-event reporting-category composition differed across histology–drug co-reporting contexts in two spontaneous-reporting databases. Methods We analyzed deduplicated reports from the U.S. Food and Drug Administration Adverse Event Reporting System (FAERS; through 2026 Q2) and Japanese Adverse Drug Event Report database (JADER; July 2026 release). Eligibility required lung-cancer indications, compatible histology, and specified suspect-drug roles for an ICI and chemotherapy components. Co-reported drugs defined SCLC platinum–etoposide (EP), NSCLC pemetrexed–platinum (PP), and NSCLC paclitaxel–platinum (PC) contexts without confirming concurrent administration. MedDRA version 29.0 Preferred Term dictionaries defined bone-marrow (BM), infection (INF), and lung-injury (LUNG) modules and five mutually exclusive reporting categories. Multinomial relative risk ratios (RRRs) compared each category with unmapped reports (UNC), using EP as the exposure reference. FAERS maximum-likelihood estimates and exploratory JADER Firth estimates were analyzed separately; Benjamini–Hochberg correction covered six primary comparisons per database. A separate 1,111-report subset underwent human review. Results The full analytic cohort contained 11,111 reports: 6,445 from FAERS and 4,666 from JADER. BM-only represented 36.3%, 22.3%, and 27.7% of FAERS EP, PP, and PC reports, respectively, and 46.5%, 10.9%, and 24.2% of JADER reports. PP-versus-EP BM-only RRRs were 0.46 (95% confidence interval 0.38–0.56) in FAERS and 0.29 (0.22–0.38) in JADER; PC-versus-EP estimates were 0.59 (0.50–0.70) and 0.50 (0.40–0.61), respectively (all q<0.001). Within NSCLC, PP-versus-PC RRRs were 0.77 (0.65–0.91) and 0.65 (0.51–0.82), but stricter nonsquamous common-support estimates were imprecise and did not pass multiplicity correction. Infection and lung-only results lacked consistent adjusted concordance. Human review assessed report coding and supplied adjudicated phenotype labels without an independent clinical reference standard. Conclusions BM-only reporting composition differed across the three co-reporting contexts, with directional concordance between databases. Structural coupling of histology and chemotherapy, uncertain concurrent exposure, and reporting bias prevent attribution to individual drugs. These findings generate hypotheses about reporting patterns; they do not estimate clinical incidence, absolute toxicity risk, or causal treatment effects.
No takes yet. Share an insight, caveat, or question.
Wang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: