Presentation (CTI)

OBJECTIVES: To clinically characterize glucagon-like peptide-1 receptor agonist (GLP-1RA) associated thrombocytopenia using clinician-confirmed cases and to evaluate the efficacy, completeness, cost, and timeliness of AI-assisted pharmacovigilance compared to previous surveillance methodologies.
METHODS: Southern Network on Adverse Reactions-Artificial Intelligence (SONAR-AI) and its predecessors, Research on Adverse Drug events and Reports (RADAR) and SONAR, have evaluated thrombocytopenia-associated adverse drug reactions. SONAR-AI investigated GLP-1RA-associated thrombocytopenia. Initial signals included three patients at 1 center with acute thrombocytopenia shortly after GLP-1RA initiation, with a fourth case subsequently reported by another hematologist. AI identified two cases from hospital research day and endocrinology conference presentations. A seventh case was reported in a review paper. AI facilitated the review of information from a public database (eHealthMe), social media postings (keywords: thrombocytopenia + GLP-1RA name), and international product labels. Data collection metrics were compared across analogous drug-associated thrombocytopenia evaluations from RADAR, SONAR, and SONAR-AI.
RESULTS: Seven patients developed severe thrombocytopenia following GLP-1RA exposure. Median platelet nadir during thrombocytopenia was 2,000/μL (range <2,000-45,000). Four patients required thrombopoietin receptor agonists, corticosteroids, plasmapheresis, and/or splenectomy. Rechallenge led to recurrent thrombocytopenia in one patient. In another case, a drug-dependent platelet-reactive IgG antibody specific for exenatide (GLP-1RA) was demonstrated. Additionally, incompletely described cases were identified in FDA Adverse Event Reporting System (54 cases, 1 death) and eHealthMe (175 cases) but not mentioned in product labels. Compared to prior investigations of thrombocytopenia associated with COVID-19 vaccines and clopidogrel, SONAR-AI evaluations were completed faster (median of 6 weeks vs 3 months and 1 year), at a lower cost ($500 vs $20,000 and $500,000), with more complete case information (95% vs 90% and 90%) and faster dissemination (1 month vs 3 months and 1 year).
CONCLUSIONS: Thrombocytopenia may complicate GLP-1RA therapy. This comprehensive evaluation of cases of thrombocytopenia associated with GLP-1RAs was cheaper, quicker, and more efficient than prior RADAR/SONAR evaluations.