Comparison of Post-marketing Surveillance Approaches Regarding Infections Related to Tumor Necrosis Factor Inhibitors (TNFIs) Used in Treatment of Autoimmune Diseases

Objectives: To examine agreement between the FDA Adverse Event Reporting System (FAERS) and observational studies in common infections for tumor necrosis factor inhibitors (TNFi’s). Methods: Using MedDRA® preferred terms, all infection cases in FAERS with each TNFi were retrieved using EvidexTM. Observational studies reporting TNFi-related infections were identified from PubMed (OS-PM) and ClinicalTrials.gov (OS-CT). Infections with a reporting rate of ≥2% (based on percentage of total number of infections) from each data source were compiled. Fleiss’s kappa and Cohen’s kappa (κ) were calculated to determine agreement across all three sources and between each two sources. Results: A total of 163,789 FAERS infection cases, 53 OS-PM studies and 52 OS-CT studies were identified. The Fleiss’ kappa that comparing all 3 data sources demonstrated lack of agreement. Significant moderate agreements were found between FAERS and OS-CT for etanercept and adalimumab, respectively (κ=0.53, p=0.02; κ=0.56, p=0.02), but no agreements (κ<0) when comparing FAERS vs. OS-PM or OS-CT vs. OS-PM. Conclusion: For common TNFi-related infections, passive (FAERS) and active (observational studies) pharmacovigilance results are similar between FAERS vs. OS-CT for etanercept and adalimumab but dissimilar across the 3 sources. Our findings suggest incorporating both active and passive pharmacovigilance methods in post-marketing drug safety assessment.