A study of the variability of the in vitro component-based microarray ISAC CDR 103 technique.

Specifi c immunoglobulin (Ig) E determination against allergens using the in vitro component-resolved diagnostic microarray technique (ImmunoCap ISAC CRD 103; Phadia, Uppsala, Sweden), has improved diagnostic accuracy [1-4], but few studies have analyzed the reproducibility of this semiquantitative technique [5]. Reproducibility is analyzed in successive determinations carried out using the KS11 control serum provided with the test kit. This serum contains different concentrations of specifi c IgE against 10 allergens: rApi g 1, rBet v 2, nBos d 4, nGal d 1, nGal d 2, nGal d 3, rHev b 8, rPhl p 5, rPhl p 6, and rPhl p 7. The resulting data are then used to generate a standard curve that relates the fl uorescence signal of the ISAC CRD103 microarray acquired by a laser scanner (LuxScan 10K/A, CapitalBio, Beijing, China) to known concentrations of specifi c IgE measured in ISAC standardized units (ISUs). The microarray assay was performed according to the manufacturer ́s instructions. The KS11 serum was analyzed for intraslide variability (4 times), intra-assay variability (8 times), and interassay variability (12 times). The reproducibility of the technique was analyzed by calculating the intraclass correlation coeffi cient (ICC) for overall variability using the software package SPSS 15.0 and the coeffi cient of variation (CV) for the variability of each of the 10 allergens using Microsoft Excel 97. According to the classifi cation of Fleiss [6], the ICCs were almost perfect for all 3 tests, with a score of 0.998 for intraslide variability (P<.0001), of 0.997 for intra-assay variability (P<.0001), and of 0.989 for interassay variability (P<.0001). For the intraslide analysis, 7 of the 10 allergens detected by KS11 had CV values of 10% or less, and for the intra-assay analysis, 5 allergens had CV values of 15% or less. In the interassay analysis, all of the allergens had CV values of over 20%, three allergens (nBos d 4, rPhl p 5, and rPhl p 6) had values of between 20% and 30%, while 5 (rBet v 2, nGal d 1, nGal d 2, rHev b 8, and rPhl p 6) had values of under 40%. The remaining 2 allergens, rApi g 1 and nGal d 3, had values of 44% and 51%, respectively (Table). While excellent results were observed for overall intraslide, intra-assay, and interassay variability, rApi g 1, nGal d 3, and rPhl p 6 all showed high variability in the individual analyses. Jahn-Schmid et al [5] reported similar results for rPhi p6. Moreover, in our study, nGal d3 had the highest CV. This suggests the existence of a technical problem related to the adhesion of the allergen to the slide, highlighting the need to validate each allergen individually. This requirement should be even stricter for allergens used to establish the standard curve, and to defi ne specifi c IgE levels for all the allergens in the ISAC CRD 103 microarray. Moreover, the fact that the variability of the data in the interassay analysis can be improved suggests that this technique can be used to assess sensitization profi les but is not appropriate for monitoring sensitization. On the basis of our results, it can be concluded that, overall, the semi-quantitative ISAC CDR 103 method is a reproducible technique. However, the high variability detected for certain allergens suggests that this in vitro tool is valid for an initial study but probably not for follow-up or monitoring studies, or for establishing therapeutic decisions. In such cases, we recommend the use of quantitative tests such as specifi c IgE determination.