Collecting and analysing all available literature before starting an animal experiment is important and it is indispensable when writing a systematic review (SR) of animal research. Writing such review prevents unnecessary duplication of animal studies and thus unnecessary animal use (Reduction). One of the factors currently impeding the production of ‘high-quality’ SRs in laboratory animal science is the fact that searching for all available literature concerning animal experimentation is rather difficult. In order to diminish these difficulties, we developed a search filter for PubMed to detect all publications concerning animal studies. This filter was compared with the method most frequently used, the PubMed Limit: Animals, and validated further by performing two PubMed topic searches. Our filter performs much better than the PubMed limit: it retrieves, on average, 7% more records. Other important advantages of our filter are that it also finds the most recent records and that it is easy to use. All in all, by using our search filter in PubMed, all available literature concerning animal studies on a specific topic can easily be found and assessed, which will help in increasing the scientific quality and thereby the ethical validity of animal experiments.

Before starting a new animal experiment, thorough analysis of previously performed experiments is essential from a scientific as well as from an ethical point of view. The method that is most suitable to carry out such a thorough analysis of the literature is a systematic review (SR). An essential first step in an SR is to search and find all potentially relevant studies. It is important to include all available evidence in an SR to minimize bias and reduce hampered interpretation of experimental outcomes. Despite the recent development of search filters to find animal studies in PubMed and EMBASE, searching for all available animal studies remains a challenge. Available guidelines from the clinical field cannot be copied directly to the situation within animal research, and although there are plenty of books and courses on searching the literature, there is no compact guide available to search and find relevant animal studies. Therefore, in order to facilitate a structured, thorough and transparent search for animal studies (in both preclinical and fundamental science), an easy-to-use, step-by-step guide was prepared and optimized using feedback from scientists in the field of animal experimentation. The step-by-step guide will assist scientists in performing a comprehensive literature search and, consequently, improve the scientific quality of the resulting review and prevent unnecessary animal use in the future.

Collecting and analysing all available literature before starting a new animal experiment is important and it is indispensable when writing systematic reviews of animal research. In practice, finding all animal studies relevant to a specific research question turns out to be anything but simple. In order to facilitate this search process, we previously developed a search filter for retrieving animal studies in the most often used biomedical database, PubMed. It is a general requirement for systematic reviews, however, that at least two databases are searched. In this report, we therefore present a similar search filter for a second important database, namely Embase. We show that our filter retrieves more animal studies than (a combination of) the options currently available in Embase. Our search filters for PubMed and Embase therefore represent valuable tools for improving the quality of (systematic) reviews and thereby of new animal experiments.

Despite the routine application of septoplasty in clinical practice, the current body of evidence does not support firm conclusions on its effectiveness.

In this letter to the editor, an updated version of the Embase search filter is presented. The former version of the filter does not work adequately anymore, because of alterations in 'Emtree', the Elsevier life science thesaurus. A thesaurus is a hierarchically structured, controlled vocabulary for efficiently searching the database in question. The functioning of the search filter critically depends on the way in which the thesaurus terms are related to one another.As of January 2013, the Emtree term human/ is included in the Emtree term Chordata/, which in itself is part of the Emtree term animal/ (http://www.elsevier. com/online-tools/embase/emtree).Previously, the Emtree terms human/ and animal/ were in separate branches of the tree. As a consequence of the altered relation between human/ and animal/, the former version of the filter started to include all studies that used humans (and no other animals) as subjects, whereas the filter was intended to exclude exactly this type of studies.We have adapted the 'mode' in which the Emtree terms animal/ and Chordata/ are present in the search filter, so that the filter no longer automatically includes the purely human studies. The total number of records the updated search filter retrieves on its own (5,884,276) and in combination with a search strategy specific for the topic pancreatitis and probiotics (83) are similar to those in the original version of the filter (search date 3 May 2013). The new version of the search filter (version 2; 31 January 2013) can be found in Supplement 1 (see