The European Commission s science and knowledge service Joint Research Centre Gavin O Connor
The Role of Reference Methods and Materials in Harmonising Allergen Measurement Results Gavin O'Connor JRC- Geel Frontiers in Food Allergy and Allergen Risk Assessment and Management April 2018
Outline What do we need to measure? Legislation Reference materials What are they? What use are they? Current status on the impact of a reference material
What do we need to measure? What does the legislation actually say? Regulation (EU) No 1169/2011 Food information to consumers (FIC) Article 36 9 1. 3., The Commission indication of shall the following adopt implementing particulars shall acts be on mandatory: the application of the requirements referred to in paragraph 2 of this Article to the following voluntary (c) food any information: ingredient or processing aid listed in Annex II or derived from a substance or product listed in Annex (a) information II causing on allergies the possible or intolerances and unintentional used the manufacture presence in food or preparation of substances of a or food products and still causing present allergies in or the intolerances finished product, even if in an altered form;
What do we need to measure? Regulation (EC) No 178/2002 laying down the general principles and requirements of food law. Article 50 14 establishes RASFF Food safety requirements 1. Food shall not be placed on the market if it is unsafe. 2. Food shall be deemed to be unsafe if it is considered to be: (a) injurious to health; (b) unfit for human consumption
What do we need to measure? Listed below are cases where MS have considered that the risk was such as to possibly require rapid reaction (in some cases following an ad hoc risk assessment) As regards food: j) food in which an unintentional presence of allergenic substances is detected, which is no mentioned on the label;
On-going activities Joint DG JRC and DG SANTE workshop. 16 th -17 th June 2016- Harmonization of approaches informing EU allergen labelling legislation 1 PAL should only be used when an associated risk assessment has been performed Guidance on a harmonised risk assessment procedure or approach for PAL is necessary Possible agreement on analytical marker(s) and their conversion to a common reporting unit should be encouraged. The most appropriate reporting unit for reporting analytical results is mg total allergenic ingredient protein per kg food. Establish EU Allergen laboratory Network. 1 Technical report: G. O'Connor, M. Haponiuk ( DG SANTÉ), F. Ulberth, Joint DG SANTÉ and DG JRC Workshop- Harmonisation of Approaches for informing EU allergen labelling legislation, JRC10825, https://ec.europa.eu/jrc
What is a Reference material? Reference material (RM) Material sufficiently homogeneous and stable with respect to one of more specified properties, which has been established to be fit for its intended use in a measurement process. Can be qualitative properties (nominal properties) Certified reference material (CRM) Reference material characterised by a metrologically valid procedure for one or more specified properties, accompanied by an RM certificate that provides the value of the specified property, its associated uncertainty, and a statement of metrological traceability. ISO Guide 33:2015(E) Reference materials Good practice in using reference materials
What is a Reference material used for? Used by measurement scientists (analysts) to assure the quality and traceability of measurement results. Calibration of equipment or measurement procedure Establishing metrological traceability Method validation Assigning values to other materials Quality control of a measurement or measurement procedure
What is a Reference material used for? Reference materials come in many forms but are used for many different reasons Pure substances (characterised for purity) Standard solutions (often gravimetric preparations) Matric reference materials Physical/chemical properties (melting point) Objects or artefacts (taste/smell)
What is a Reference material used for? Matrix CRM Calibration CRM Definition of the measurement Sampling sample preparation Instrument Calibration Calibration CRM Matrix CRM Evaluation of measurement uncertainty Calculation of results Data evaluation Measurement of sample(s) Calibration CRM Calibration CRM QC criteria evaluation Reporting
Achieving comparable measurement results (Approach 1- Harmonisation) The comparison of all results to a common food ingredient/method may possibly provide a solution. We will be able to assess methods relative to the common food ingredient. However, comparison of the results with others who do not have access to the same common standards/methods and with the clinical thresholds will be difficult.
Achieving comparable measurement results (Approach 1- Standardisation) The comparison of results via a well-defined unit enables the independent assessment of different approaches. Can link results to other groups working on the assessment of methods via a variety of different technologies. The linking of results via this unbroken chain of calibration is often referred to as "metrological traceability". True value (unit) Peptide standards of know purity Equimolar mixed standard of prototypic peptides Protein standards of know molar concentration Food ingredient characterised for amount of target proteins Preparation of food matrix with known amount of protein Assessment of methods relative to the concentration of peptides/protein it contains. Comparison of results to a quantity value that can be realised across space and time.
Why has comparability of measurement results been an issue? Same outcome no matter where, when or by whom it was measured! Is this achievable at a specified level/legislative level? Clinically relevant levels? Zero level- often based on instrument performance. (but how low is low enough?) drives Legislation Measurement performance Innovation Accepted Thresholds/ Levels Reference materials Food Markers
Why has comparability of measurement results been an issue?
Trueness Nitrogen content Conversion factor Interferences Food Ingredient Target degradation Target stability Target homogeneity Product Total protein content Analytical Target Variance in expression level
How comparable are food allergen results? Assessment against Peers. lysozyme in wine at current OIV LOD Study of lysozyme in wine by four different test kits Mean 0.22 µg g -1 Std. dev. 0.02 µg g -1 RSD% 12
How comparable are food allergen results? Assessment against "Truth" lysozyme in wine at current OIV LOD Study of lysozyme in wine by four different test kits Mean 0.22 µg g -1 Std. dev. 0.02 µg g -1 RSD% 12 Gravimetric value 0.25 µg g -1 Difference 13.8%
How comparable are food allergen results? Assessment against "fit for purpose" lysozyme in wine at current OIV LOD Gravimetric value 0.25 µg g -1 Difference 13.8% Acceptance range 25%
How comparable are food allergen results? Study 1 Study 2 Study 1 Study 2 zz = (xx XX) 25 xx. 100 Performance criteria of 25 % used to establish a Z score based on "fitness for purpose"
Assessment of "allergens" in wine lysozyme in wine at current OIV LOD & LOQ for fining agents Youden plot representing pairwise comparison of lysozyme in wine.
Assessment of "allergens" in wine Fining agent in wine at current OIV LOD Youden plot representing pairwise comparison of egg white fining agent in wine.
How comparable are food allergen results? Kit Manufacturer Kit Units (measurand) Antibody Range (ppm gluten) (gluten conc. determined directly from graph w/o sample dilution) Morinaga Institute of Biological Science, Inc. ELISA Technologies Inc. Wheat Protein Elisa kit (Gliadin) ng/ml Wheat protein polyclonal 0.3-10 ALLER-TEK ng/ml Gluten 401.21 5-80 NEOGEN Corporation Gluten Assay Kit ng/g Gluten 401.21 3-50 Romer Labs UK Ltd. AgraQuant ng/g Gluten G12 4-200 Biomedal Diagnostics GlutenTox Elisa ng/ml Gliadin G12 1.5-25 BIOCONTROL TRANSIA Plate ng/ml Gliadin R5 3-50 (prolamins) R-biopharm AG RIDASCREEN Gliadin ppb Gliadin R5 5-80
Comparability of gluten measurements? Four matrix materials were gravimetrically spiked with a gluten extract to yield a mass fraction of 20 and 100 mg/kg Gluten in each material. corn flour, rice flour, cooked corn cookie, cooked soy cookie
Comparability of gluten measurements? Four matrix materials were gravimetrically spiked with a gluten extract to yield a mass fraction of 20 and 100 mg/kg Gluten in each material. corn flour, rice flour, cooked corn cookie, cooked soy cookie
Comparability of gluten measurements? common calibration standards used across all kits To assess the effect of a purified common calibration standard, all kit calibrators were compared in terms of mass fraction using the pure gluten extract as a calibrator. The effect of this recalibration on the results of the spiked samples was then assessed.
Comparability of gluten measurements? Spiked materials Corrected for mean error of matrix results. i.e. global sample recovery correction. Caution! This is correcting the data set for the observed difference of the data!!!! But it may help indicate the best case scenario if we were able to remove systematic (calibration) bias from the kits.
Comparability of gluten measurements? Samples were sourced from routine test samples submitted to public analysist. (10-150 ppm gluten) Samples used in proficiency testing schemes. A candidate reference material. Samples extracted and analysed in triplicate.
Comparability of gluten measurements? Uncorrected results from real food samples. Sampled paired in to match closely related concentrations and matrices.
Comparability of gluten measurements How can we explain the differences? A pair wise comparison of kits revealed a lack of correlation between many of the kits. A CRM is unlikely to improve comparability of results where the kits target and report different quantities.
Will a Reference Material Help? BME-IGRM-0, BME-IGRM-10 and IBM-IGRM-50 Three candidate reference materials produced by the Budapest University of Technology and Economics, Hungary (Bugyi, Török, Hajas, Adonyi, Poms, Popping, et al., 2012) Cooked cookies that have been spiked with 10 and 50 mg kg -1 WGPAT gliadin extract. Assessment of material commutability
Comparability of gluten measurements? Corrected using the slope of the matrix reference material.
Impact on Legislation? 4 2 1 A "fit for purpose" assessment was made in terms of legislative impact of each kits determined concentration for each sample. Sample level (ppm gluten) Kits < 20 mg/kg 20 100 mg/kg > 100 mg/kg
Legislative impact Uncorrected data Probability of a false classification as gluten free ppm (mg kg-1) % 20 51 90 3 80 5
Legislative impact Corrected data Probability of a false classification as gluten free ppm (mg kg-1) % 20 63 30 9 40 2
LCMS analysis of Food Allergens
Preliminary assessment of method Determination of the milk protein content of a cookie using peptides alone! Baked cookie prepared with 10 mg milk protein/kg. Average of all results is 9.3 ± 3.0 mg total milk protein/kg
Conclusions The current allergen measurement infrastructure is evolving! The lack of agreement of targets and the gap between these targets and the intended measurand is not well understood. The impact of requested CRMs requires careful consideration. Multiple extraction Reference methods Understanding all the contributions to the overall measurement uncertainty of the final result is important when comparing results for different approaches.
JRC_Allergens@ec.europa.eu Jørgen Nørgaard Chiara Nitride Maria-Jose Martinez Esteso Malgorzata Rzychon Marcel Brohée Andrew Flanagan Regional public analyst s laboratory, Galway, Ireland. Adam Dabrowski Biotechnika Laboratoryjna, Nuscana, Poznan, Poland. Sandor Tomoskozi Budapest University of Technology and Economics
Stay in touch! ec.europa.eu/jrc @EU_ScienceHub EU Science Hub - Joint Research Centre Joint Research Centre EU Science Hub
Thanks Questions? You can find me at gavin.o connor@ec.europa.eu
The importance of measurement uncertainty Measurement result value ± uncertainty i) ii) iii) iv) i. Result minus uncertainty above limit. ii. Result above limit but limit within uncertainty. iii. Result below limit but limit within uncertainty. iv. Result plus uncertainty below limit.