Disproportionality analysis of drug-associated tinnitus safety signals: a 43-year pharmacovigilance study.

Drug-induced tinnitus is a prevalent yet under-characterised adverse drug reaction. No prior study has systematically screened the FDA Adverse Event Reporting System (FAERS) for drug-tinnitus disproportionality signals using a rigorous, deduplication-validated framework. A case/non-case pharmacovigilance study was conducted using 53,243 tinnitus-associated individual case safety reports from the FAERS Public Dashboard (1983-2025). Following four-tier deduplication, the reporting odds ratio (ROR) was computed for 900 eligible active ingredients (N ≥ 10 ICSRs; signal criterion: ROR > 1, lower 95% CI > 1, Yates-corrected p < 0.05). Bayesian sensitivity analyses applied. Seventy-four. positive disproportionality signals were identified (8.2%; ROR 1.29-10.71). The strongest signal was diphenhydramine citrate/ibuprofen (ROR 10.71), followed by chloroquine phosphate (9.28). Signals spanned seven ATC Level-1 classes; the nervous system class contributed most signal drugs (n = 12). BCPNN positive 48/74 signals (64.9%; Spearman ρ = 0.997 vs ROR). Median reporting delay was 169 days (IQR 56-446). BH-FDR sensitivity analysis (q ≤ 0.05, m = 900) retained 60 signals, all BCPNN-corroborated. This 43-year analysis identifies 74 positive drug-associated tinnitus signals, including novel immunobiologic (guselkumab, lifitegrast) and cardiovascular (lovastatin, terazosin) associations without current audiological labelling warnings. The delayed onset profile implicates cumulative cochlear injury mechanisms. These findings provide an actionable basis for regulatory prioritisation and prospective audiological safety research for drug classes where audiological monitoring protocols are currently absent.