arXiv:2511.00093v1 Announce Type: new
Abstract: This study aimed to systematically identify and quantify risks for drug-induced rhabdomyolysis (DIR) using real-world data and to propose an evidence-based risk mitigation framework. We conducted a retrospective pharmacovigilance study using the FDA Adverse Event Reporting System (FAERS) database from Q1 2005 to Q1 2025. A two-stage analysis involved initial signal detection using the Reporting Odds Ratio (ROR), followed by a LASSO-optimized multivariate logistic regression to calculate adjusted odds ratios (aORs) for 54 target drugs while controlling for confounders. Our analysis confirmed potent DIR risks for known agents, such as gemfibrozil (aOR 173.67) and statins (lovastatin aOR 97.20, simvastatin aOR 85.12). Crucially, we identified strong, novel risk signals for drugs currently lacking warnings, most notably levetiracetam (aOR 11.02) and donepezil (aOR 8.90). A significant “labeling gap” was quantified: 61.1% of drugs with a statistically significant DIR risk lack a corresponding warning in U.S. drug labels. We subsequently developed a three-tiered risk stratification model. The proposed framework provides a data-driven foundation for developing tiered clinical decision support systems, enhancing prescribing safety, and guiding future regulatory action to bridge the identified evidence-to-labeling gap.
The Hidden Power of Normalization: Exponential Capacity Control in Deep Neural Networks
arXiv:2511.00958v1 Announce Type: cross Abstract: Normalization methods are fundamental components of modern deep neural networks (DNNs). Empirically, they are known to stabilize optimization dynamics and