Hen’s egg allergy
We included 35 studies [33, 34, 41, 43, 49, 59, 60, 68, 72, 74, 78, 82, 84, 88, 94, 97, 105, 118, 126, 128, 129, 131, 133, 137, 142, 146, 152, 156-158, 161, 166, 168, 172, 180] on the accuracy of diagnostic tests for HE allergy. For meta-analyses, tests were divided into raw, cooked and baked HE.
For raw HE, SPT to egg white [33, 34, 137, 172] had a specificity of 80% compared to specificities of 96 % for SPT to egg yolk [33, 34, 49] and 91% for SPT to ovalbumin [34, 49, 172]. Nine studies on sIgE to egg white (EW) [33, 34, 43, 60, 74, 78, 137, 142, 172], 4 on sIgE-egg yolk [33, 34, 78, 142] and 6 on CRD [34, 43, 60, 74, 78, 142, 172] met the inclusion criteria. sIgE-EW showed pooled sensitivity of 73% with a specificity of 88%. (Table 2 ) This increased to 95% for subjects ≤2 years of age (Table 6 ). sIgE-egg yolk had low sensitivity and specificity. Ovomucoid [34, 43, 60, 74, 78, 172] showed high specificity of 91% with low sensitivity of 74% at a median cut-off 0.8 kUA/L and ovalbumin [34, 43, 60, 78, 142, 172] did not reach appropriate accuracy with sensitivity of 78% and specificity of 79%.
For cooked HE, 6 studies on SPT-EW [33, 82, 84, 118, 157, 168], 4 on SPP with raw EW [33, 82, 126, 168], 14 on sIgE-EW [33, 43, 74, 82, 88, 97, 118, 128, 131, 152, 157, 168, 172] , 7 on ovomucoid-sIgE [33, 34, 43, 74, 118, 142, 172] and 3 studies on ovalbumin [33, 43, 172] met the inclusion criteria. SPP [33, 82, 126, 168] was highly sensitive for cooked HE allergy diagnosis with pooled sensitivity of 94% and specificity of 66% at the 6 mm cut-off. SPT to EW [33, 82, 84, 118, 157, 168] showed pooled sensitivity of 68% with a specificity of 77%. The sensitivity increased to 79% for subjects ≤2 years. sIgE to EW showed pooled sensitivity of 85% with a specificity of 73%, respectively. Ovomucoid-sIgE had a sensitivity of 74% with high specificity of 91% at the 0.8 kUA/L cut-off. sIgG4 to HE had low sensitivity and low specificity in highly heterogenous studies. Studies for cooked HE allergy were less heterogeneous than those for raw or baked HE allergies.
For baked HE allergy, sIgE to EW [41, 137, 146] showed high specificity (94%) but very low sensitivity of 40% at the 8 kUA/L cut-off with values ranging widely from 6 to 50 kUA/L. The accuracy for SPT to EW [41, 137, 146, 166] was low for baked HE allergy. There was insufficient data for meta-analyses on accuracy of CRD in baked HE allergy. (Figure 2B)
For raw HE allergy, maximum sensitivity of ≥ 90% was not reached by analysed diagnostic tests. Maximum specificity was ≥ 90% for sIgE to EW, ovalbumin and ovomucoid. For cooked HE allergy, maximum sensitivity of ≥ 90% for SPP and maximum specificity was ≥ 90% for SPP and ovomucoid. For baked HE, maximum specificity was ≥ 90% for SPT and sIgE EW (Tables S5 and S6 ).