Reducing variability in the measurement of gluten contamination in oats, oilseeds, and pulses by improving sample preparation

Abstract

Background and objectives:The variability in gluten in non-gluten-containing grains (NGCG) processed using two preparation schemes was evaluated with the aim of minimizing effects of sample heterogeneity on gluten determined by enzyme linked immunosorbent assay (ELISA). The relationship between gluten concentration as determined by ELISA and visually assessed contamination of NGCG with wheat, durum, barley, and rye was investigated. Findings: Low variability between duplicate aliquots taken from test portions (0-30.6% relative standard deviation [RSD]) demonstrated the ELISA itself was precise. In the first scheme, variability among test portions ranged from 1% to 143% RSD, with only half in the range of 1-50%. Using scheme 2, variability in gluten among test portions ranged from 1% to 85% RSD, with more than three quarters in the range of 1-50%. High lipid content hemp seed was a particular challenge to grind, and this was reflected in higher variability in gluten results between test portions (mean RSD = 61%). Conclusions: Subsampling ground samples using rotary sample division and the use of a 1-g test portion in scheme 2 decreased the variability of gluten results for most samples. At concentrations relevant to existing thresholds of gluten contamination (e.g. 20 mg/kg), there was no relationship between gluten concentration in NGCG and cereal contamination as determined by visual inspection. Significance and novelty: This study provides guidance on how to improve the analysis of gluten contamination in NGCG by ELISA and describes the absence of a relationship between ELISA-determined gluten and the visual assessment of contamination in NGCG.