Statistical data of the AllergyScreen® / AlleisaScreen® test
Statistical data:
variation, sensitivity, specificity, accuracy, linearity, LoB and LoD (based on decimal classes) - most tested allergen groups are considered
Variation
| Sample level (classes) | Coefficients of variation (%) | |||
|---|---|---|---|---|
| AllergyScreen® | AlleisaScreen® | |||
| Intra-assay variation | Inter-assay variation | Intra-assay variation | Inter-assay variation | |
| Class 1 (cut-off) | 27.0 | 26.4 | 25.8 | 18.1 |
| Class 2-6 | 3.1 | 4.5 | 2.5 | 2.5 |
Inter-batch variation: mean value 8.2 % (AllergyScreen®/AlleisaScreen®)
Analytical performance of most tested allergens with an ELISA reference test (4820 determinations)
The sensitivity, specificity, concordance with the respective confidence intervals of most tested allergens over seven studies with a pairwise ELISA reference test in general (4820 determinations) are given in table below:
| Sensitivity | 95% confidence interval | Specificity | 95% confidence interval | Concordance | 95% confidence interval |
|---|---|---|---|---|---|
| 85.6% | 84.3% to 86.8% | 92.4% | 91.1% to 93.5% | 88.3% | 87.4% to 89.2% |
Linearity
Linearity between the measurement signal and the concentration of the analyte is, if r is > 0.97 and R² is > 0.95.
As seen in the table below, all calculated regression coefficients (r and R²) are close to 1 thus demonstrating a clear positive linear relationship between the measuring signal of the AllergyScreen®/AlleisaScreen® and the allergen concentration in the sample. Although the AllergyScreen®/AlleisaScreen® is determined as a semi-quantitative assay, the linearity is very good over a broad range of dilutions as it is recommended in quantitative assays.
| AllergyScreen®/AlleisaScreen® | Regression coefficient r | Coefficient of determination R² |
|---|---|---|
| Mean AllergyScreen® | 0.994 | 0.987 |
| Mean AlleisaScreen® | 0.993 | 0.987 |
LoD and LoB
The limit of detection (LoD) was determined by the measurement of samples known to contain a low concentration of analyte and the limit of blank (LoB) by measuring a high number (>100) of blank samples. The following tables show that the LoD and the LoB of the AllergyScreen®/AlleisaScreen® are in the range of equivalent devices and the mathematical cut-off of ≥ 0.35 IU/ml has compiled with.
LoB: Meanblank + 1.645 (SDblank)
| Parameter | AllergyScreen® | AlleisaScreen® |
|---|---|---|
| Blanks | 402 | 626 |
| Meanblank | 0.01 IU/ml | 0.01 IU/ml |
| SDblank | 0.03 | 0.03 |
| LoB | 0.06 IU/ml | 0.05 IU/ml |
LoD: LoB + 1.645 (SD low concentration sample)
| Parameter | AllergyScreen® | AlleisaScreen® |
|---|---|---|
| Low concentration | 83 | 123 |
| Meanblank | 0.39 IU/ml | 0.35 IU/ml |
| SDblank | 0.17 | 0.16 |
| LoD | 0.34 IU/ml | 0.31 IU/ml |
Furthermore, an interference study showed that there is no indication of a significant influence of different blood components (bilirubin, haemoglobin, lipids) on the results of the AllergyScreen®/AlleisaScreen® test. Only a very high IgE-level in patient's serum/plasma might lead to a minor to moderate reduction of the measured values in some cases due to the hook effect. But however, this does not lead to false-negative results. Patients treated with Omalizumab and similar drugs cannot be analysed with this system (Mizuma et al. 2015).