Background: The prognosis for many children with cow's milk allergy (CMA) is remission within 3 years, and the clinical parameters that predict duration of disease have not been measured incrementally.
Objective: To prospectively determine prognostic predictors of tolerance in a random cohort of referrals using CMA workup outcomes as covariates and tolerance as the status variable in a duration model of CMA.
Methods: The 2001-2006 Milan Cow's Milk Allergy Cohort (MiCMAC) enrolled children referrals using double-blind, placebo-controlled food challenges (DBPCFCs) as study end points (confirmation of CMA; onset of tolerance). The Cox regression model was used to analyze all clinical factors that contributed to tolerance. Covariates analyzed were skin, gastrointestinal, and respiratory symptoms; history and demographics at presentation; age at diagnosis and DBPCFC outcomes; sensitization (skin and serum) by cow's milk protein fractions; sensitization to other food and inhalant allergens; total IgE levels; specific IgE concentrations for cow's milk protein fractions, other ingestants, and aeroallergens; and threshold doses at DBPCFC. Sensitization and DBPCFC were performed at 6-month intervals.
Results: A total of 112 infants were enrolled (mean [SD] age, 13.85 [9.84] months), and 59 achieved tolerance (mean [SD] age when tolerance was achieved, 27.58 [11.81] months). On univariate analysis, asthma and/or rhinitis at presentation was an independent predictor of persistence (hazard ratio [HR], 2.19; 95% confidence interval [CI], 1.26-3.82). On multivariate analysis, predictors of persistence were a fresh milk wheal diameter increment of 1 mm (HR, 1.18; 95% CI, 1.07-1.31) and a positive skin prick test result with soy (HR, 6.99; 95% CI, 1.56-31.25).
Conclusions: This is the first study, to our knowledge, to identify incremental biological predictors of delayed tolerance to cow's milk in children that should be integrated into DBPCFC schedules for CMA in infants.