To meet good pharmacovigilance practices, regulatory authorities expect Marketing Authorisation Holders of medicinal products to review the scientific literature for their product portfolio on a weekly basis.1 The purpose of this review is to identify individual case safety reports (ICSRs) relevant for 7/15-day reporting to Regulatory Authorities and non-ICSR publications relevant for the purpose of signal detection and for the production of aggregate reports – such as Periodic Safety Update Reports (PSURs) and Periodic Benefit Risk Evaluation Reports (PBRERs). Even for well-established medicinal products, literature database searches – such as PubMed, Embase, etc. – can easily yield several thousand hits per year. Therefore, strategies have to be in place to ensure efficient triaging and review of pharmacovigilance literature. Dr. Joerg Seebeck will present such a strategy and discuss, based on examples, why certain publications might be relevant and why others are not.
Due to the complexity and variety of the data sources available in scientific literature for the purpose of pharmacovigilance (as further detailed below), triage and review are difficult tasks. In addition, large product portfolios and limited resources require these analyses to be performed with high efficiency. Consequently, staff involved in these activities must have a thorough medical and scientific knowledge base, in addition to being adequately trained and experienced. The aim of this presentation is to contribute to that training effort.
Databases to be considered: Wide ranging databases such as Embase, PubMed, and Medline which are considered sufficient; and other databases which may be used when justified.
Search strategy: There is no ‘gold standard’ for search strings and search strategies may be database-specific. Usually it is a compromise between precision and recall (sensitivity). Search strings must include the name of the active ingredient / INN and standard search strings should be validated. The GVP module on PSURs states:
“Literature searches should be wider than those for individual adverse reaction cases as this should also include studies reporting safety outcomes in groups of subjects and other products containing the same active substance. The special types of safety information that should be included, but which may not be found by a search constructed specifically to identify individual cases, include:
- Pregnancy outcomes (including termination) with no adverse outcomes;
- use in paediatric populations;
- compassionate supply, named patient use;
- lack of efficacy;
- asymptomatic overdose, abuse or misuse;
- medication error where no adverse events occurred;
- important non-clinical safety results.
If relevant and applicable, information on other active substances of the same class should be considered.” The GVP module further states that literature data are relevant to identify new efficacy, effectiveness and safety information.2
Literature searches should be performed on a weekly basis and not only to identify ICSRs but also to identify relevant non-ISCR information which may lead to the identification of an Emerging Safety Issue (ESI) which has to be reported without delay.
Triaging results (hits): In order to be able to report expeditable ICSRs (7/15 day reporting time line) and ESIs in time, literature search results should be triaged and reviewed within 5 days. The initial triage should aim to identify all cases and case series using the standard 4 minimum criteria that define a valid case (identifiable reporter, drug, event and patient) and other cases that need to be databased for presentation in the PSUR (e.g. asymptomatic overdose, abuse, off-label use, misuse or medication error).
Identification of non-ICSR but relevant literature data: A more complex task. Data sources that are typically of high relevance are prospective large randomized controlled clinical trials (e.g. for demonstration of lack of efficacy and the relative risk for certain adverse events), state of the art drug-drug interaction studies if they demonstrate an impairment in the systemic availability of the drug of interest, thorough QT studies, meta-analyses investigating efficacy and safety issues (e.g. NSAID and cardiovascular outcomes and GI bleeding events) and other signal-generating analyses of existing data sets (e.g. data mining in large multidrug databases or analysis of registry studies for incidence rates of certain adverse events, any study indicating teratogenic or carcinogenic effects in animals or humans. In vitro data only, with exceptions, are often difficult to interpret due to uncertainty in terms of the relevance of the model chosen and of the drug concentration used.
Keywords: Antiviral Drugs, Scientific Literature