Interpretation of tests for nut allergy in one thousand patients, in relation to allergy or tolerance

Abstract

Summary: Background Peanut and tree nut allergy are common, increasing in prevalence and the commonest food cause of anaphylaxis. In the USA, 7.8% are sensitized (have nut‐specific IgE), but not all those sensitized are allergic. Lack of data makes interpretation of tests for nut‐specific IgE difficult. Objectives This is the first study to investigate the clinical significance of test results for peanut and tree nut allergy in allergic or tolerant patients. Findings are related to the severity of the allergy. Method An observational study of 1000 children and adults allergic to at least one nut. History of reactions (severity graded) or tolerance to up to five nuts was obtained and skin prick test (SPT)/serum‐specific IgE (CAP) performed. Results There was no correlation between SPT size and graded severity of worst reaction for all nuts combined or for peanut, hazelnut, almond and walnut. For CAP, there was no correlation for all nuts. Where patients tolerated a nut, 43% had positive SPT of 3–7 mm and 3% ≥ 8 mm. For CAP, 35% were positive (0.35–14.99 kU/L) and 5% ≥ 15 kU/L. In SPT range 3–7 mm, 54% were allergic and 46% were tolerant. There was poor concordance between SPT and CAP (66%). Of patients with a clear nut‐allergic history, only 0.5% had negative SPT, but 22% negative CAP. Conclusions Magnitude of SPT or CAP does not predict clinical severity, with no difference between minor urticaria and anaphylaxis. SPT is more reliable than CAP in confirming allergy. Forty‐six per cent of those tolerant to a nut have positive tests ≥ 3 mm (sensitized but not allergic). One cannot predict clinical reactivity from results in a wide ‘grey area’ of SPT 3–7 mm; 22% of negative CAPs are falsely reassuring and 40% of positive CAPs are misleading. This emphasizes the importance of the history. Understanding this is essential for accurate diagnosis. Patients with SPT ≥ 8 mm and CAP ≥ 15 kU/L were rarely tolerant so these levels are almost always (in ≥ 95%) diagnostic.