Validation of a qualitative R5 dip-stick for gluten detection with a new mathematical-statistical approach

Specific and sensitive analytical methods are needed to guarantee the safety of gluten-free products for coeliac disease patients at the threshold of 20 mg gluten/kg. Immuno-chromatographic assays, usually available in dip-stick or lateral-flow format, provide rapid qualitative results indicating the presence or absence of gluten. In September 2015, an AOAC expert review panel accepted the qualitative immuno-chromatographic dip-stick system RIDA®QUICK Gliadin as an AOAC Official Method^SM of Analysis First Action 2015.16 for detection of gluten in corn-based food products. One part of the required validation was an international collaborative study with 18 participants who analysed 8 different processed and non-processed samples with different gluten concentrations. Since the official statistical evaluation at AOAC does not provide a calculation of: (1) the concentration and its prediction interval from which a user expects a positive result; and (2) what a negative result means, a new statistical approach was used. The method for modelling the probability of detection (POD) curve and confidence intervals of the LOD95 used a generalised linear mixed effects model together with a 4-parameter sigmoid curve to describe the variability of the POD curves for randomly chosen laboratories beyond those that participated in the study. The results confirmed that, in case of a positive result, a non-processed sample contains a minimum of 4.54 mg gluten/kg (prediction range 3.59-5.74 mg/kg) while a processed sample contains more than 6.29 mg gluten/kg (prediction range 3.88-10.20 mg/kg). Therefore, the method is able to identify negative samples clearly below the threshold of 20 mg gluten/kg which is a prerequisite for gluten testing in food.