Prospective evaluation of testing with baked milk to predict safe ingestion of baked milk in unheated milk-allergic children

Allison Kwan, Maria Asper, Sasson Lavi, Elana Lavine, David Hummel, Julia E. Upton

Cow's milk allergy is one of the most common food allergies affecting young children. A subset of milk-allergic individuals can eat baked milk without allergic symptoms which is beneficial in terms of prognostication and liberalization of the diet. A retrospective study suggested that skin prick testing (SPT) with a baked milk (muffin) slurry may provide a sensitive means of predicting the outcome of a medically supervised baked milk oral food challenge. We evaluated the predictive value of SPT with baked milk to identify unheated milk-allergic children who are able to safely eat baked milk. Children aged 2-16 years with a prior history of reaction to milk and a milk extract SPT of 8-14 mm were recruited. Investigator-blinded SPT to muffin slurry and powdered milk in triplicate and specific IgE (sIgE) to casein and milk were performed. Graded oral challenge to egg-free baked milk muffins (total 2.6 gm milk protein) was performed in the hospital. Reliability of tests was analyzed for intraclass correlation. Statistical significance for clinical characteristics of population and muffin testing versus baked milk reactivity was calculated with Fisher exact test for dichotomous and t-test for continuous variables. Wilcoxon rank sum test was used to compare immunological characteristics between individuals who tolerated or reacted to baked milk. Fitted predicted probability curves and ROC curves were generated. Thirty-eight children were consented and 30 met study criteria. The muffin SPT and casein sIgE were significantly different in those who passed versus failed baked milk challenge. Negative (<3 mm) baked milk tests were found in 8/30 children (27 %) and were associated with non-reactivity to baked milk (p = 0.01) with a sensitivity of 1 (0.70-1.00). All children with negative SPT for baked milk passed the oral challenge. Specificity was 0.41 (0.19-0.67). The optimal decision point for the muffin SPT was 4 mm and the casein sIgE was 6 kU/L. The powdered milk test was not helpful. Skin prick testing with a baked milk (muffin) slurry may have a role in clinical practice to identify baked milk tolerance in milk-allergic patients.