Background Threshold levels for peanut allergy determined by using oral challenges are important for the food industry with regard to allergen labeling. Moreover, the utility of biological markers in predicting threshold levels is uncertain. Objective We sought to use a modified oral food challenge regimen that might determine threshold levels for peanut allergy mimicking a more real-life exposure and to correlate the eliciting dose (ED) and severity of clinical reaction in children with peanut allergy with B-cell, T-cell, and effector cell markers. Methods A modified food challenge procedure with doses scheduled 2 hours apart was used in 63 children with peanut allergy. All children received a maximum of 8 semi-log increasing titration steps of roasted peanuts ranging from 3 to 4500 mg of peanut protein until objective allergic reactions occurred. Severity of symptoms was graded from I to V. Biological markers were measured before challenge. Results Forty-five of 63 patients showed objective symptoms after greater than 30 minutes, with a median latency of clinical reaction of 55 minutes. By using a log-normal dose-distribution model, the ED5 was calculated to be 1.95 mg of peanut protein. The ED was significantly and inversely correlated with peanut- and Ara h 2–specific IgE levels, skin prick test responses, basophil activation, and TH 2 cytokine production by PBMCs. Symptom severity did not correlate with any of the markers or the ED. Conclusion This modified food challenge procedure might better reflect threshold levels for peanut allergy than the standard procedure because most of the patients reacted at a time interval of greater than 30 minutes. By using this model, threshold levels, but not severity, could be correlated with biological markers.