Discussion
The EAACI Food Allergy Guidelines need to be updated, in particular the section on diagnostic tests for food allergy. The systematic review of the literature on diagnostic tests that informed the current EAACI Food Allergy Guidelines was conducted prior to 2014, thus a new systematic review is needed, especially since emerging tests such as CRD, MAT and BAT were not included in the previous review. We report herein the protocol for a systematic review of the literature, as registered at PROSPERO (CRD42021259186), and provide a description of the rationale and methods chosen for the review.
OFC is the reference-standard diagnostic test for food allergies; however, there are considerable risks involved. OFC involve ingesting the suspected allergen, which is associated with a risk of allergic reaction and may result in a life-threatening event. As a result, OFC must be conducted in a controlled environment with intensive care facilities in easy reach, and not all clinical settings can offer this service. Additionally, this risk of reaction may lead to significant anxiety in patients and their families, therefore a reliable and cost-effective alternative to OFC is needed. Since the previous guideline was published, newer tests such as CRD have entered mainstream use in Europe and other regions. Also, studies assessing the efficacy of novel diagnostic tests such as BAT and MAT have been published. All of these new tests must be reviewed to assess whether they are acceptable diagnostic tools for use in routine clinical practice. For more established allergy tests, such as SPT and sIgE, new studies have been performed since the previous review which may require updating of the previous recommendation. We aim to evaluate the diagnostic test accuracy of any index test, from SPT, sIgE to extracts, sIgE to individual allergens, sIgE to allergen peptides, BAT and MAT. Thus, we do not include search terms for the index test. We apply a specific filter for diagnostic test accuracy studies, though, which implicitly captures studies on diagnostics and/or tests. We only include search terms for “challenge” as part of the common terminology for the comparator, i.e. oral food challenge, to have a sensitive search regarding the comparator test.
The population of interest will include all ages and diverse clinical settings and geographical locations. We deliberately do not include search terms to specify the target population and studies that do not have evidence from all populations will be included in the review and judged for risk of bias. We expect to encounter very few if any diagnostic test accuracy studies conducted in animals; however, we will refrain from specifying search terms that will exclude animal studies because of the risk of falsely excluding studies on animal allergen sources. Certain food allergies are more prevalent in different areas. As such, we will include studies focussing on any food, to be inclusive to all populations, although we anticipate that there will be more studies on milk, egg and peanut allergies.
Although the gold standard is DBPCFC, we will also include studies where open oral food challenges are performed and allow for the inclusion of studies where a small proportion of patients were diagnosed without an oral food challenge due to previous severe or anaphylactic reactions. Studies focusing only on IgE sensitisation will be excluded, as some patients who experience sensitisation to foods are not necessarily allergic. It is important to note that the comparator reference standard test is oral food challenge and as such, the systematic review will inform on the diagnostic accuracy of potential alternative tests. However, the systematic review will not clarify which of these tests is best used in subjects in whom the oral food challenge is contraindicated.
As outcomes we are assessing sensitivity and specificity as measures of diagnostic test accuracy. The positive and negative predictive value of the test, which depends on the prevalence of the given food allergy in the population, is not within the scope of this systematic review. The details collected as part of the quality assessment and risk of bias will allow us to select high-quality studies at low risk of bias. The planned sensitivity analyses on these features may help to identify factors that influence the performance of the test and the modulation of the identified diagnostic cut-offs. If sufficient high quality studies are identified, the planned meta-analyses will provide evidence which may help to reduce the number of patients that have to undergo an OFC procedure.