Real-world pharmacovigilance of drug-related bone metabolism disorders: integrating FAERS and VigiAccess with a Bradford Hill-based causal plausibility assessment.

BACKGROUND

Metabolic bone diseases (MBDs) are a group of conditions characterized by imbalanced bone remodeling, including osteoporosis, osteomalacia, and rickets. In recent years, drug-induced secondary MBDs have emerged as a major clinical concern. However, systematic drug risk assessment studies based on large-scale real-world data remain scarce.

METHODS

Based on the U.S. Food and Drug Administration Adverse Event Reporting System (FAERS, 2004-2024) and VigiAccess, four complementary disproportionality algorithms (reporting odds ratio, ROR; proportional reporting ratio, PRR; Bayesian confidence propagation neural network, BCPNN; multi-item gamma Poisson shrinker, MGPS) were used to screen drug-event signals. The signals were stratified by gender and age type, and their temporal distribution characteristics were characterized using the Weibull model. Finally, the causal plausibility was assessed using the updated Bradford Hill framework.

RESULTS

A total of 55,783 adverse event reports were included, and 33 drugs showed consistent positive signals in both the FAERS and VigiAccess databases. Major signal drugs included bisphosphonates (e.g., alendronic acid, zoledronic acid), antiviral drugs (e.g., tenofovir disoproxil fumarate, emtricitabine/tenofovir combination), proton pump inhibitors (e.g., esomeprazole), and endocrine antineoplastic drugs (e.g., anastrozole). Within the Anatomical Therapeutic Chemical (ATC) classification, signals were primarily concentrated in anti-infectives (category J), antineoplastic and immunomodulatory drugs (category L), and musculoskeletal drugs (category M).

CONCLUSION

This study systematically identified robust pharmacovigilance signals and disproportionality-based associations between multiple drug classes and metabolic bone disorders through an integrated analysis of the FAERS and VigiAccess databases, providing high-confidence real-world evidence for pharmacovigilance signal detection.