AOAC ISPAM Stakeholder Panel Meeting Book 9-24-17

Estimating Recovery of a Gluten ELISA Kit Method

121 122 123 124

Paul Wehling, Mark Arlinghaus

General Mills, Inc.

May 11, 2017 125 126 Introduction 127 128 As part of the AOAC International ISPAM initiative to draft an SMPR for Gluten in Oats, we 129 would like to outline a proposed method for estimating accuracy of response of an ELISA kit 130 method. The SMPR will describe the basic requirements for validation of a method, and set firm 131 criteria for method parameter estimates which must be satisfied for the method to proceed to 132 First Action status. The SMPR is used as a guide for the Expert Review Panel (ERP) to determine 133 which methods may pass to First Action. 134 135 In the past, gluten methods were evaluated for accuracy based on spiking wheat gluten into 136 various gluten-free matrices, and calculating “recovery” by the usual means of percentage of 137 analyte recovered from the collaboratively studied samples. Recent quantitative methods, such 138 as AOAC OMA methods 2012.01 and 2014.03 have used this method. In the case of validating 139 an ELISA method for gluten in oats, it will be essential to evaluate the kit responses to not only 140 wheat, but also barley and rye. This paper will outline a proposed procedure to evaluate kit 141 response to all 3 contaminant grains independently. 142 143 In 2010 and 2013, the AOAC Allergen Community published two papers in the JAOAC outlining 144 best practices for validating Allergen ELISA methods, and gluten ELISA methods (1, 2). Abbott 145 et. al.discussed spiking allergens into matrices to evaluate recovery. In the Koerner paper, we 146 briefly touched on recovery studies, but did not mention the differences between wheat, barley 147 and rye. In the case of determining gluten levels in oats, we think that barley will be a 148 significant source of contamination, so it will be critical to understand the kit response to barley 149 as well as wheat. Rye can be found in oat streams in North America, but not nearly to the 150 extent as wheat and barley. We propose here a method for estimating response of three 151 contaminant grains independently. 152 153 Concept 154 The concept here will be to obtain gluten free oat flour, and contaminant grains. The 155 contaminant grains will be ground to particle size similar to flour, and then analyzed by some 156 “reference” method for gluten content. The contaminant flours will then be spiked into the oat 157 flour quantitatively. In order to estimate the responses of the three contaminant grains 158 individually, there will necessarily need to be three series of spiked flour samples, one each for 159 wheat, rye and barley. Kit developers will analyze the series of samples (with replication) and 160 recovery estimates for the three grains can be calculated. 161 162 Our concept is that this experiment will be performed by the kit developer as part of the SLV, 163 and tested independently by a 3 rd party lab. This would be performed on oat flour samples only 164 as a test of kit/antibody response. We anticipate method precision will need to be 165 demonstrated on each claimed matrix by SLV and multi-lab experiments as required by the 166 SMPR. 167