Background: Oral food challenges (OFC) are required to diagnose food allergies but are resource-intensive. Objective: To reduce the need for OFCs, we sought to determine serum specific immunoglobulin E (sIgE) cutoff levels for cow's milk and its major allergens predicting oral milk challenge outcomes in children with suspected cow's milk allergy. Methods: A total of 135 Finnish children (median age, 1.8 years [range, 1.0‐14.1 years]) with suspected cow's milk allergy underwent open OFC with unheated cow's milk. The sIgE levels to milk (f2), casein (Bos d 8), alpha-lactalbumin (Bos d 4), beta-lactoglobulin (Bos d 5), and bovine serum albumin (BSA) (Bos d 6) were measured and compared with the challenge outcomes. Results: Of the 135 OFCs, 5 were excluded from the study due to purely subjective symptoms. Of the 130 remaining OFCs, 98 results (75%) were positive. In a receiver operating characteristic analysis with 1-2-year-old children, no individual allergen sIgE had a better area under the curve than milk sIgE (0.824). A milk sIgE level > 6.30 kU/L gave 94% specificity and 33% sensitivity for positive OFCs. In 3‐14-year-old children, a cutoff value >13.9 kU/L predicted a positive OFC result with 93% specificity and 25% sensitivity. Children with moderate-to-severe reactions had higher sIgE levels to milk, alpha-lactalbumin, and BSA than did children with mild reactions. Conclusion: Molecular allergy diagnostics did not improve the predictive performance compared with milk sIgE. The milk sIgE value that exceeds the cutoff for 95% specificity in combination with the clinical history may help to reduce the need for OFCs. The severity of an allergic reaction cannot reliably be predicted from sIgE measurements.
