Correlation of serum allergy (IgE) tests performed by different assay systems

Background: In vitro testing is commonly used to diagnose and manage allergies. Clinical reactivity has been correlated with food-specific IgE levels by using the ImmunoCAP (Phadia, Uppsala, Sweden). Objective: To determine whether IgE levels derived from different assays are equivalent to those measured by ImmunoCAP. Methods: Fifty patients from the Mount Sinai Pediatric Allergy practice were prospectively enrolled. For each deidentified sample, specific IgE levels were measured to egg, milk, peanut, cat, birch, and Dermatophagoides farinae at different laboratories, each using a different assay system (Phadia ImmunoCAP, Agilent Turbo-MP, and Siemens Immulite 2000). Results were analyzed to determine whether IgE measurements were equivalent. Food allergen-specific IgE levels were correlated with clinical data and around empirically determined thresholds that predict probability of clinical disease in 50% or 95% of subjects. Results: Variable degrees of agreement existed among the 3 assays. Immulite 2000 overestimated all specific IgE levels compared with ImmunoCAP. Turbo-MP overestimated for egg but underestimated for birch and D farinae. Differences for milk, peanut, and cat were observed, without a trend toward overestimation or underestimation. Furthermore, several values for the food allergens were discrepant around the 50% and 95% positive predictive values for clinical reactivity. Conclusion: Discrepancies in specific IgE values from 3 different assays can potentially lead to altered management and treatment. The predictive values for clinical reactivity associated with food-specific IgE levels determined by ImmunoCAP should not be applied to results from other assays.