The complementary use of fish extracts enhances the diagnostic performance of the molecular approach
Besides PVs, the presence of IgE to other fish allergens should be investigated before continuing with FC to confirm specific fish tolerance. Our analyses showed lower IgE levels to raw and heated fish extracts than to PVs for all bony fish (Figure 4A ). To determine the correlation between the IgE levels for PV and RE, PV and HE, or RE and HE, Spearman’s rank correlation coefficient was calculated. Data were interpreted (weak/moderate/strong/very strong correlation) according to Schober et al.17 A strong significant positive correlation (Rho>0.7) of all three pairs was observed for herring and salmon (Figure 4B ). In addition, a very strong significant positive correlation (Rho=0.98) between values for cod PV and RE was shown, indicating that for cod, a RE is sufficient to detect the PV sensitization. For other analyzed species, correlations of IgE levels to PVs and extracts were weak to moderate (Figure 4B ). Data for REs and HEs correlated strongly for herring (Rho=0.85) and salmon (Rho=0.75), while moderate to weak correlations were observed for other species. All correlations were significant, except for tuna PV vs RE and tuna PV vs HE (Figure S7 ). Individual scatterplots for each species and calculated P values for the correlation coefficients are shown in Figure S7 .
We next assembled data for patients with higher IgE values to extracts than to PV from one or more species (44 patients in total). The percentage of these patients varied across the countries (8% - 28%) (Figure 4C ). Data for each patient was then analyzed in detail (Figure 4D ). Up to 7 patients per fish species had no IgE for the respective PV, but were positive for RE and/or HE. These patients could be falsely diagnosed as negative to some species if only the PV-sIgE would be quantified.