Despite no global consensus on a definition of anaphylaxis, there is increasing recognition that just as allergic reactions lie on a spectrum of severity, the same is for anaphylaxis. A variety of severity scores exist in the literature. We review the approaches taken to develop these scores, and their relative advantages and disadvantages.

There have been four recent comparisons of published severity scores. All have highlighted the heterogeneity between scoring systems, and the lack of transferability from one approach to another. Notably, only one score has been developed using a data-driven approach, and none has undergone formal and comprehensive validation.

It is unclear whether a single severity score is achievable, or indeed desirable. If the aim is to guide management of acute reactions, then assignment of severity is not only unnecessary but might delay treatment and cause harm. Severity scores are needed in the research setting, but require an approach which can discriminate between reactions of similar but nonidentical severity (particularly, nonanaphylaxis reactions). Any approach should be fit for purpose, informed by patient and clinician experience, and ideally be data-driven to minimize subjective bias and facilitate objective validation.