D3.87 - Assessing the agreement between skin prick test and specific-IgE for ten common food allergens in infants

Food allergy diagnostic guidelines recommend using either skin prick test (SPT) or specific-IgE (sIgE) to confirm IgE sensitisation despite some evidence of disagreement between these tests. Further, SPT and sIgE agreement has rarely been assessed in young children, particularly for foods other than egg white, cow’s milk, and peanut. This study therefore aimed to investigate the agreement between SPT and sIgE at 12 months old for ten common food allergens.

The study included infants from Vitality (NCT02112734), a double-blinded, randomised, placebo-controlled trial which sought to determine the role of daily oral vitamin D in infancy on food allergy prevalence (n = 2739). Infants who underwent both SPT and sIgE (peanut, egg white, cow’s milk, cashew, sesame, hazelnut, almond, soy, shellfish, wheat) at 12 months were eligible for the current analysis. Kendall's tau-b (τ_b) correlation coefficient was used to determine the relationship between SPT and sIgE. Gwet’s agreement coefficient-1 (AC₁) was used to assess binary test agreement between SPT and sIgE using different sensitisation cut-offs for SPT (1 mm, 2 mm, or 3 mm) and sIgE (0.10 kU_A/L or 0.35 kU_A/L).

In total, 571 infants had SPT and sIgE results to all ten foods at 12 months old. Figure 1 shows SPT and sIgE had positive associations ranging from weak for shellfish (τ_b = 0.17 [95% CI = -0.07, 0.32]) to strong for cashew (τ_b = 0.63 [95% CI = 0.53, 0.73]). More children had at least one positive sIgE result ≥ 0.10 kU_A/L (40.6% [95% CI = 36.6%, 44.8%], n = 232/571) than SPT ≥ 1 mm (16.3% [95% CI = 13.4%, 19.6%], n = 93/571) to any of the ten foods. Agreement was highest when using SPT ≥ 3 mm and sIgE ≥ 0.35 kU_A/L. At these cut-offs, individual food agreement ranged from 92.6% for egg white (AC₁ = 0.903 [95% CI = 0.873, 0.933], n = 529/571) to 99.8% for shellfish (AC₁ = 0.998 [95% CI = 0.995, 1.000], n = 570/571). Agreement was poorer between SPT and sIgE when the positive sensitisation cut-off for sIgE was lowered to ≥ 0.10 kU_A/L.

IgE test agreement at 12 months old was highest when using sensitisation cut-offs typically used clinically, i.e. SPT ≥ 3 mm and sIgE ≥ 0.35 kU_A/L, across ten common food allergens. Even though test agreement was high, a small number of children with negative tests of sensitisation by one measure may have a positive test by the other. The clinical implication of this finding will be explored in future analyses.