Consequences of the exclusion of known adverse drug reactions in a spontaneous reporting system on the possibility to detect unknown drug/ADR combinations

Background: The Reporting Odds Ratio (ROR) is one of the expressions to analyse disproportionallity of adverse drug reactions (ADRs) in a spontaneous reporting system. The ROR is defined as the extent to which the association between a suspected drug and ADR stands out against the background frequency of the same ADR in all other drugs. The (de)nominator is influenced by known associations, unknown associations and 'background noice'. The question is if the sensitivity of the system would improve by omitting the 'known ADRs'. Objectives: This study investigates the increase of disproportional, unknown associations in a spontaneous reporting system that excludes known drug/ADR combinations for two ADRs: myocardial infarction and myalgia. Methods: All cases reported to the Netherlands Pharmacovigilance Centre were included in this analysis. A drug/ADR combination was disproportional if n>3, and ROR - 95%CI >1. Disproportionallity analysis was the ADRs 'myocardial infarction' (MI) and 'myalgia'. RORs were calculated using cases from the standard database, and on a database in which all cases with known drug/ADR combination were excluded. Results: For myalgia, 150 drug/ADR combinations were present in the Lareb database, of which 141 are known and 34 of these were disproportional. Of the nine unknown drug/ADR combinations all were disproportional when known drug/ADR combinations were excluded. For MI, 39 drug/ADR combinations were present, of which 35 were known and 27 of these were disproportional, including the four unknown drug/ADR combinations. Conclusions: Exclusion of known drug/ADR combinations increases the number of disproportional, unknown drug/ ADR combination. This may improve the sensitivity of signal detection and can be relevant when the denominator is influenced by relatively over reporting. An analysis using a 'corrected ROR' in which all cases of known drug/ADR combinations are excluded, might be helpful to study new drug/ADR combinations