Key Elements in Adverse Drug Interaction Safety Signals
(English)Manuscript (preprint) (Other academic)
Background: Effective surveillance of adverse drug interactions (a problematic drug combination resulting in an adverse drug reaction (ADR)) in large collections of individual case safety reports (ICSRs) requires a combination of expert clinical assessments and efficient algorithms. To date, most methods proposed for adverse drug interaction surveillance focus on disproportionality analysis, although a recent study proposed that the reported clinical and pharmacological information is also useful for systematic screening of adverse drug interactions.
Objective: The purpose of this study is to identify and describe key elements in adverse drug interaction safety signals.
Methods: Altogether 137 case reports from three previously published safety signals of suspected adverse drug interaction were re-evaluated using an operational algorithm for causality analysis of drug interactions; the Drug Interaction Probability Scale (DIPS). Reports in the WHO Global ICSR Database, VigiBase, and their corresponding original files were analysed, examining whether the DIPS elements were registered. The retrieved case information was specified as being listed in the structured fields, free text and, in total. In addition, information not covered by DIPS, such as explicit notifications of a suspected drug interaction by the reporter or the pharmacovigilance centre was also registered.
Results: As expected from the data used in this analysis, the most frequently fulfilled DIPS elements were: objective evidence (such as ADR) of a drug interaction (137 cases; 100%). Other frequent elements were (ranked order) plausible time to onset (53 cases; 38%), and resolution of the ADR after terminating the drug inducing the interaction (10 cases; 7%). Ten cases (7%) fulfilled both a plausible time to onset and resolution of the ADR after stopping the drug. Positive rechallenge was only reported in 3 cases (2%). For 32 cases additional information was reported in free text in the original files that were not available in VigiBase. A suspected drug interaction was noted by the reporter in 47 original cases (35%) and more than 80% of these were assessed as a possible or probable drug interaction according to the DIPS classification. Among cases without notes of suspected interactions were 58 original reports (64%) assessed as possible (56 cases) or probable (2 cases).
Conclusions: A plausible time to onset pattern and resolution of the ADR after withdrawal of the drug inducing the interaction frequently strengthened the suspected causality of a drug interaction. Particularly strong cases were those containing both these key elements. Since this information is often available in structured format, it could potentially be used to automatically highlight strong cases in firstpass screening. Finally, this analysis also demonstrated the importance of free text where particularly relevant clinically details such as timeliness, severity, resolution of the reaction after withdrawal of the drug inducing the interaction, possible alternative causes, and dosage changes are available.
Adverse drug interactions, signals, systematic surveillance system, plausible time to onset
Medical and Health Sciences
IdentifiersURN: urn:nbn:se:liu:diva-70847OAI: oai:DiVA.org:liu-70847DiVA: diva2:442061