Quantitative Signal Detection for COVID-19 Medicinal Products Based on Retrospective Analysis of Spontaneous Reports from the Russian Pharmacovigilance Database

Abstract

SCIENTIFIC RELEVANCE. The pandemic of novel coronavirus infection (COVID-19) led to a drastic increase in the use of medicinal products of various therapeutic groups and increased spontaneous reporting of adverse drug reactions (ADRs). Therefore, it is necessary to analyse the reported information to identify potential safety signals.AIM. This study aimed at systematisation and quantitative analysis of data on the safety of COVID-19 medicinal products from the Russian pharmacovigilance database.MATERIALS AND METHODS. This retrospective analysis included spontaneous ADR reports submitted to the Russian pharmacovigilance database from 1 January 2020 to 31 December 2022. The authors applied disproportionality analysis to generate safety signals.RESULTS. During the stated period, the database website published 873 spontaneous reports on 1,636 ADRs associated with COVID-19 treatment. Most ADRs were associated with favipiravir (493 reports), hydroxychloroquine (87 reports), and olokizumab (85 reports). The most common ADRs included 273 (16.7%) abnormal investigation results, 203 (12.4%) hepatobiliary disorders, and 191 (11.6%) gastrointestinal disorders. The majority of adverse events, 674 (77%) cases, had favourable outcomes. Using frequency-based disproportionality analysis, the authors identified 23 potential safety signals based on disproportionately reported ADRs for azithromycin, dexamethasone, levilimab, lopinavir+ritonavir, molnupiravir, olokizumab, tofacitinib, tocilizumab, umifenovir, and favipiravir.CONCLUSIONS. The new safety signals require additional assessment of ADR reporting forms for causal relationship analysis, validation, prioritisation, and clinical interpretation. The frequency-based method did not identify safety signals for hydroxychloroquine, remdesivir, baricitinib, sarilumab, and anti-COVID-19 human immunoglobulin, but this does not rule out the possibility of detecting new causal relationships.