The impact of duration of treatment on reported time-to-onset in spontaneous reporting systems for pharmacovigilance.

Abstract

Within pharmacovigilance, knowledge of time-to-onset (time from start of drug administration to onset of reaction) is important in causality assessment of drugs and suspected adverse drug reactions (ADRs) and may indicate pharmacological mechanisms involved. It has been suggested that time-to-onset from individual case reports can be used for detection of safety signals. However, some ADRs only occur during treatment, while those that do occur later are less likely to be reported. The aim of this study was to investigate the impact of treatment duration on the reported time-to-onset. Case reports from the WHO Global ICSR database, VigiBase, up until February 5(th) 2010 were the basis of this study. To examine the effect of duration of treatment on reported time-to-onset, angioedema and hepatitis were selected to represent short and long latency ADRs, respectively. The reported time-to-onset for each of these ADRs was contrasted for a set of drugs expected to be used short- or long-term, respectively. The study included 2,980 unique reports for angioedema and 1,159 for hepatitis. Median reported time-to-onset for angioedema in short-term treatments ranged 0-1 days (median 0.5), for angioedema in long-term treatments 0-26 days (median 8), for hepatitis in short-term treatments 4-12 days (median 7.5) and for hepatitis in long term treatments 19-73 days (median 28). Short-term treatments presented significantly shorter reported time-to-onset than long-term treatments. Of note is that reported time-to-onset for angioedema for long-term treatments (median value of medians being 8 days) was very similar to that of hepatitis for short-term treatments (median value of medians equal 7.5 days). The expected duration of treatment needs to be considered in the interpretation of reported time-to-onset and should be accounted for in signal detection method development and case evaluation.