Anaphylaxis is a clinical diagnosis with no gold-standard test. Recent case definitions have attempted to provide objective criteria for diagnosis. The aim of this study was to compare the diagnostic concordance of the Brighton Collaboration case definition (the 'Brighton' case definition) to the consensus case definition from the Second Symposium on the Definition and Management of Anaphylaxis (the 'Symposium' definition). The study setting was a hospital-based emergency department in the UK. We identified cases of anaphylaxis by physicians' discharge diagnoses over a 2-year period from 2005 to 2006, and used randomly selected cases of allergic reaction, asthma and urticaria as a control group. Data was extracted by clinicians (who were unaware of the content of either case definition), and the two case definitions were applied by Boolean operators in a Microsoft Excel spreadsheet. Concordance between the case definitions was measured using Cohen's kappa (kappa) statistic. We reviewed 128 sets of notes, with 47 cases of anaphylaxis. Brighton and Symposium definitions had sensitivities of 0.681 and 0.671, respectively, and specificities of 0.790 and 0.704, respectively. A discordant result was found in 36/128 cases (28.1%; kappa = 0.414 [95% CI 0.253, 0.574]), which represents a moderate level of agreement between case definitions. The Brighton case definition has a similar diagnostic concordance to the Symposium case definition. It does not seem to over- or underestimate cases and is sufficiently unique that the identification of an allergic trigger does not have to form part of the case definition. This will be important in the recognition of anaphylaxis resulting from the administration of drug and vaccines, where causality should be examined separately from case ascertainment.