Adverse drug reaction reporting: how can drug consumption information add to analyses using spontaneous reports?

Kristian Svendsen, Kjell H. Halvorsen, Solveig Vorren, Hilde Samdal, Beate Garcia

Spontaneous reporting of adverse drug reactions (ADRs) is a cornerstone in pharmacovigilance. However, information about the underlying consumption of drugs is rarely used when analysing spontaneous reports. The purpose of this study was to combine ADR reports with drug consumption data to demonstrate the additional information this gives in various scenarios, comparing different drugs, gender-stratified sub-populations and changes in reporting over time. We combined all Norwegian ADR reports in 2004-2013 from the EudraVigilance database (n = 14.028) with dispensing data from the Norwegian Prescription Database (more than 800 million dispensed prescriptions during 2004-2013). This was done in order to calculate drug-specific consumption-adjusted adverse drug reaction reporting rates (CADRRs) by dividing the number of reports for each drug with the number of users of the drug during the same time period. Among the ten drugs with the highest number of ADR reports and the ten drugs with the highest CADRR, only four drugs were in both categories. This indicates that drugs with a high number of reports often also have a high number of users and that CADRR captures drugs with potentially relevant safety issues but a smaller number of users. Comparing reported ADRs in females and males using methylphenidate, we found that the two groups report different ADRs. Finally, we showed that changes in ADR reporting for simvastatin and atorvastatin during 2004-2013 were due to changes in consumption and that atorvastatin had a higher CADRR but fewer reports than simvastatin. CADRR provides additional information compared with number of reports alone in studies using spontaneous reports. It is important for researchers to adjust for consumption whenever possible in pharmacovigilance studies.