Comparison between two automated systems to determine specific IgE: CAP and ELItest

Summary Background and objective The ELItest® is a newly developed system to measure specific IgE based on allergen bound to paper rings and an alkaline phosphatase conjugated second antibody detection system. It was compared to the CAP® system, a method based on allergen conjugated to an encapsulated cellulose polymer and a β-galactosidase conjugated fluorescence detection system. Methods Sera of 300 patients with positive history and positive skin-prick tests to common allergens (birch, timothy-grass, cat dander, dermatophagoides pteronyssinus, wasp venom) and 30 negative controls were tested in both systems. Serial dilutions of high titre sera were measured; inter- and intraassay coefficients of variation (cv) were determined. Results The CAP system proved to be more sensitive (92.3%) compared to ELItest (84%) but marginally less specific (94.7% for CAP versus 96.7% for ELItest). Intraassay cv were slightly lower in the ELItest (7.2% CAP versus 6.4% ELItest), whereas the interassay cv was roughly twice as high for ELItest (20.1%) than for the CAP system (11.4%). Linearity over an 8-fold dilution was good in both tests (r2 0.979 ELItest versus 0.996 CAP), although ELItest levelled off at higher allergen concentrations. Similarly, correlation analysis between both systems revealed that ELItest consistently measured lower values, especially at higher concentrations of specific IgE. The slope of the linear regression line of a log/log plot of measured IgE concentrations was significantly lower than 1 in birch, cat and wasp; the y-intersect was significantly lower than 0 in all analysed allergens. Conclusion These results suggest that the ELItest system for the measurement of specific IgE is not quite as reproducible and sensitive as the CAP system but slightly more specific, and that higher concentrations of specific IgE are measured lower in the ELItest. One potential reason might be that the amount of allergen bound to a paper ring might be smaller than that bound to a cellulose polymer, but further experiments are necessary to prove this hypothesis.