Drug-induced impulse control disorders: integrating pharmacovigilance findings into the cumulative evidence synthesis process.

Impulse control disorders (ICoDs), including pathological gambling, compulsive shopping, and hypersexuality, significantly impair quality of life. Traditionally regarded as always idiopathic, ICoDs have also been recognized as adverse drug reactions (ADRs) over the past 25 years, with individual case safety reports–ICSRs providing unique insights. This PhD project aims to expand our understanding of drug-induced ICoDs by challenging the notion that ICSRs and disproportionality analysis (DPA) are only for preliminary signal detection. The simplicity of DPA has led to misuse and misinterpretation, leading several pharmacovigilance experts to advocate for its preclusion from the scientific literature. Nonetheless, integrative approaches to DPA show it can offer more profound insights. This project proposes a methodology that enhances the utility of ICSRs and DPA by incorporating established knowledge and complementary tools. The project begins with phenotyping ICoDs in ICSRs, validating an operational definition through a scoping review and DPA and analysing free-text narratives to identify irrelevant reports. I use this enhanced understanding of ICoDs phenotype to integrate disproportionality analysis with Bradford Hill’s criteria, organizing intrinsic and extrinsic evidence for a class effect by third-generation antipsychotics. In the absence of a comprehensive mechanistic explanation for drug-induced ICoDs, DPA combined with pharmacometrics suggested a potential pathogenetic mechanism involving 5-HT1a receptor agonism. Additionally, network analysis combined with an event-event disproportionality analysis characterized ICoDs as a dynamic syndrome, identifying behavior-specific sub-syndromes and potential exacerbating events, and pointing to possible therapeutic targets. Beyond advancing knowledge on drug-induced ICoDs, this PhD work contributes to methodological advances in DPA and ICSRs, advocating for a Bayesian epistemological framework in pharmacovigilance. This framework integrates intrinsic and extrinsic evidence, empowering pharmacovigilance to provide an answer to causal questions and profile ADRs beyond drug-event combination, to retrieve actionable findings to support enhanced regulatory and clinical management strategies.