Primary diagnostic cohort studies, nested case-control studies and diagnostic case-control studies in which controls were diagnosed with a specific gastrointestinal disease were eligible for inclusion. Included studies needed to evaluate accuracy of externally validated symptom-based criteria or history taking (symptoms) and/or physical examination (signs) for diagnosis of IBS in adults (18 years or more) with non-acute abdominal symptoms (duration two weeks or more) who presented to primary care or outpatient gastrointestinal clinics.
Studies had to report sufficient data to construct a 2x2 table of test performance. For exclusion of IBS, the reference standard had to consist of diagnostic work-up or clinical follow-up of at least one year. For diagnosis of IBS, studies were required to use externally evaluated symptom-based IBS criteria as the reference standard.
Included studies defined IBS as non-organic/functional disease without differentiating between upper and lower disease, lower functional gastrointestinal disease or a diagnosis using symptom-based IBS criteria. In the first two definitions, IBS was a diagnosis by exclusion and in the third, IBS was a positive diagnosis. The following symptom-based IBS criteria were evaluated: Manning, Kruis, Rome I, Rome II, Bellentani, Mazumdar, Talley and Wasson. Only two studies evaluated signs and symptoms as index tests; these used Rome I and Rome II as the reference standard. One study excluded patients with alarm features, two studies explored the influence of alarm features on accuracy. Three studies explicitly excluded patients with upper gastrointestinal disease. Six studies examined the effects on diagnostic performance of including upper functional disease.
Two reviewers independently assessed studies for inclusion. Disagreements were resolved through consultation with a third reviewer.