Repositorio Institucional

CCQM-K138 Final Report

CCQM-K138 Determination of aflatoxins (AFB1, AFB2, AFG1, AFG2 and Total AFs) in

Dried Fig

Key Comparison Track C

Final Report November 2018

Ahmet C Goren, Taner Gokcen, Simay Gunduz, Mine Bilsel TUBITAK UME, National Metrology Institute, Gebze/Kocaeli, 41470 Turkey

With contributions from: Mathias Koch Bundesanstalt fuer Materialforschung und –pruefung (BAM), Berlin, Germany Elias Kakoulides, Panagiota Giannikopoulou EXHM/GCSL-EIM, Chemical Metrology Laboratory (General Chemical State Laboratory Hellenic Metrology Institute), Athens, Greece Gary Tang Wai-tong, Andy Chan Government Laboratory, Hong Kong (GLHK), HKSAR, China Estela Kneeteman National Institute of Industrial Technology, Toxicology and Nutrition Laboratory (INTI) Buenos Aires, Argentina Isaac Mugenya, Geoffrey Murııra Kenya Bureau of Standards, Food and Agriculture (KEBS), Nairobi, Kenya Cheerapa Boonyakong National Institute of Metrology of Thailand (NIMT), Pathum Thani, Thailand Maria Fernandes-Whaley National Metrology Institute of South Africa (NMISA), Brummeria, South Africa Anatoliy Krylov, Alena Mikheeva D.I. Mendeleyev Institute for Metrology (VNIIM), St. Petersburg, Russian Federation

i



CCQM-K138 Final Report

SUMMARY

The presence of any aflatoxin contamination in exported figs needs to be monitored and measured through reliable and traceable methods, which require pure and matrix certified reference materials. On the other hand, certified reference materials (CRM) for determination of aflatoxins in dried fig are not yet available. Moreover, there is a lack of CRMs to be used in routine testing laboratories for method validation and quality control. The routine testing laboratories, participating in commercial proficiency testing (PT) programs, use the results available from consensus values to evaluate the performance of the participating laboratories, rather than metrologically traceable assigned values. This study initially proposed as a key comparison and presented at the EURAMET TC-MC SCOA meeting in Malta in 2015 and subsequently at the CCQM OAWG meeting in April 2015, proposes a CRM candidate for determination of levels of aflatoxins B1, B2, G1, G2 and their total in dried fig [1-4]. Evidence of successful participation in formal, relevant international comparisons are needed to document measurement capability claims (CMCs) made by national metrology institutes (NMIs) and designated institutes (DIs).

In total nine NMI/DI participated in the Track C Key Comparison CCQM-K138 Determination of aflatoxins (AFB1, AFB2, AFG1, AFG2 and Total AFs) in Dried Fig. Participants were requested to evaluate the mass fractions expressed in ng/g units, of aflatoxins B1, B2, G1, G2 and total aflatoxin in a dried food matrix, dried fig. The CCQMK138 results for the determination of aflatoxins (AFB1, AFB2, AFG1, AFG2 and total AFs) are ranging from 5.17 to 7.27 ng/g with an %RSD of 10.47 for AFB1, ranging from 0.60 to 0.871 ng/g with an %RSD of 11.69 for AFB2, ranging from 1.98 to 2.6 ng/g with an %RSD of 10.36 for AFG1, ranging from 0.06 to 0.32 ng/g with an %RSD of 35.6 for AFG2, and ranging from 8.29 to 10.31 ng/g with an %RSD of 7.69 for Total AFs. All participants based their analyses on LC-MS/MS, HPLC-FLD, HR-LC/MS, and IDMS. Brief descriptions of the analytical methods used by the participants, including sample preparation, analytical technique, calibrants, and quantification approach are summarized in Appendix F. Linear Pool was used to assign the Key Comparison Reference Values (KCRVs) for B1, B2, G1, G2 and total aflatoxins. Due to the traceability requirements for the calibrants not being met, results of KEBS, INTI, VNIIM and BAM were excluded from KCRV determination.

Successful participation in CCQM-K138 demonstrates the following measurement capabilities in determining mass fraction of organic compounds, with molecular mass of 100 g/mol to 500 g/mol, having high polarity (pKow > -2), in mass fraction range from 0.05 ng/g to 500 ng/g in dried food matrices.

ii



CCQM-K138 Final Report

TABLE OF CONTENTS

INTRODUCTION .......................................................................................................................... 1 TIMELINE...................................................................................................................................... 2 MEASURANDS ............................................................................................................................. 2 STUDY MATERIALS ................................................................................................................... 3 PARTICIPANTS, INSTRUCTIONS AND SAMPLE DISTRIBUTION .................................... 10 RESULTS ..................................................................................................................................... 11 DEGREES OF EQUIVALENCE (DoE) ...................................................................................... 27 USE OF CCQM-K138 IN SUPPORT OF CALIBRATION AND MEASUREMENT CAPABILITY (CMC) CLAIMS .................................................................................................. 35 CONCLUSIONS........................................................................................................................... 35 ACKNOWLEDGEMENTS .......................................................................................................... 36 REFERENCES ............................................................................................................................. 36

LIST OF TABLES

Table 1: Timeline for CCQM-K138 ............................................................................................... 2 Table 2: Statistical Evaluation Result of Homogeneity of the Study Material............................... 5 Table 3: Homogeneity Results of the Study Material..................................................................... 6 Table 4: Results of the homogeneity assessment for AFB1, AFB2, AFG1, AFG2 in dried fig. ...... 7 Table 5: Short Term Stability Test Results..................................................................................... 8 Table 6: Long Term Stability Test Results of the Study Material ............................................... 10 Table 7: Institutions Registered for CCQM-K138........................................................................ 10 Table 8: Metrological Traceability of Participants’ Results ......................................................... 14 Table 9: Reported Results for AFB1, ng/g .................................................................................... 18 Table 10: Reported Results for AFB2, ng/g .................................................................................. 19 Table 11: Reported Results for AFG1, ng/g.................................................................................. 19 Table 12: Reported Results for AFG2, ng/g.................................................................................. 20 Table 13: Reported Results for Total AF, ng/g............................................................................. 20 Table 14: Reported Results for AFG2, ng/g.................................................................................. 23 Table 15. Candidate Key Comparison Reference Values............................................................. 24 Table 16: Degrees of Equivalence for AFB1 ................................................................................ 27 Table 17: Degrees of Equivalence for AFB2 ................................................................................ 28 Table 18: Degrees of Equivalence for AFG1 ................................................................................ 28 Table 19: Degrees of Equivalence for AFG2 ................................................................................ 28 Table 20: Degrees of Equivalence for Total AF ........................................................................... 29

LIST OF FIGURES

Figure 1: Structure of AFB1 ............................................................................................................ 3 Figure 2: Structure of AFB2 ............................................................................................................ 3 Figure 3: Structure of AFG1............................................................................................................ 3 Figure 4: Structure of AFG2............................................................................................................ 3 Figure 5: The flowchart of the production process of dried fig ...................................................... 4 Figure 6: Reported Results for AFB1, ng/g................................................................................... 21 Figure 7: Reported Results for AFB2, ng/g................................................................................... 21 Figure 8: Reported Results for AFG1, ng/g .................................................................................. 22

iii



CCQM-K138 Final Report

Figure 9: Reported Results for AFG2, ng/g .................................................................................. 22 Figure 10: Reported Results for Total AF, ng/g ........................................................................... 23 Figure 11: Linear pool KCRV relative to the reported results for AFB1, ng/g............................ 25 Figure 12: Linear pool KCRV relative to the reported results for AFB2, ng/g............................. 25 Figure 13: Linear pool KCRV relative to the reported results for AFG1, ng/g ............................ 26 Figure 14: Linear pool KCRV relative to the reported results for AFG2, ng/g ............................ 26 Figure 15: Linear pool KCRV relative to the reported results for Total AF, ng/g ....................... 27 Figure 16: Absolute degrees of equivalence for AFB1 in CCQM-K138. ..................................... 29 Figure 17: Relative degrees of equivalence for AFB1 in CCQM-K138. ...................................... 30 Figure 18: Absolute degrees of equivalence for AFB2 in CCQM-K138. ..................................... 30 Figure 19: Relative degrees of equivalence for AFB2 in CCQM-K138. ...................................... 31 Figure 20: Absolute degrees of equivalence for AFG1 in CCQM-K138...................................... 31 Figure 21: Relative degrees of equivalence for AFG1 in CCQM-K138....................................... 32 Figure 22: Absolute degrees of equivalence for AFG2 in CCQM-K138...................................... 32 Figure 23: Relative degrees of equivalence for AFG2 in CCQM-K138....................................... 33 Figure 24: Absolute degrees of equivalence for Total AF in CCQM-K138................................. 33 Figure 25: Relative degrees of equivalence for Total AF in CCQM-K138.................................. 34

LIST OF APPENDICES

Appendix A: Call for Participation ............................................................................................. A1 Appendix B: Protocol ..................................................................................................................B1 Appendix C: Registration Form...................................................................................................C1 Appendix D: Reporting Form ..................................................................................................... D1 Appendix E: Core Competency Form.......................................................................................... E1 Appendix F: Summary of Participants’ Analytical Information.................................................. F1

Table F-1: Summary of Sample Size, Extraction, and Cleanup for CCQM-K138................ F2 Table F-2: Summary of Analytical Techniques for CCQM-K138 ........................................ F6 Table F-3: Summary of Summary of Calibrants and Standards for CCQM-K138 ............. F10 Table F-4: Summary of Assessment and Verification Methods for CCQM-K138 ............. F11 Table F-5: Additional Comments for CCQM-K138............................................................ F13 Appendix G: Summary of Participants’ Uncertainty Estimation Approaches ........................... G1 Appendix H: Participants’ Results as Reported.......................................................................... H1

iv



CCQM-K138 Final Report



ACRONYMS



BAM CCQM

CMC CRM CV

DI DoE EXHM GLHK HPLC-DAD LC-HRMS LC-MS LC-MS/MS ID IDMS INTI INRAP KC KCRV KEBS LC MADe

MRM NICOB NIMT NMISA NMI NMR OAWG pKow PSE qNMR QuEChERS RMP SIM SPE SRM UME VNIIM



Bundesanstalt fuer Materialforschung und –pruefung, DI: Germany Consultative Committee for Amount of Substance: Metrology in Chemistry and Biology Calibration and Measurement Capability certified reference material coefficient of variation, expressed in %: CV = 100·s/

designated institute degrees of equivalence Chemical Metrology Laboratory, DI: Greece Government Laboratory, Hong Kong, DI: Hong Kong high pressure liquid chromatography with diode array detection liquid chromatography with high-resolution mass spectrometry detection liquid chromatography with mass spectrometry detection liquid chromatography with tandem mass spectrometry detection isotope dilution isotope dilution mass spectrometry National Institute of Industrial Technology, Buenos Aires, Argentina National Institute of Research and Physical and Chemical Analysis, Tunisia Key Comparison Key Comparison Reference Value Kenya Bureau of Standards, NMI: Kenya liquid chromatography median absolute deviation from the median (MAD)-based estimate of s: MADe = 1.4826·MAD, where MAD = median(|xi-median(xi)|) multiple reaction monitoring NIST Consensus Builder National Institute of Metrology of Thailand, Thailand National Measurement Institute South Africa, NMI: South Africa national metrology institute nuclear magnetic resonance spectroscopy Organic Analysis Working Group logarithm of the octanol-water partition coefficient pressurized solvent extraction quantitative nuclear magnetic resonance spectroscopy “Quick, Easy, Cheap, Effective, Rugged, Safe” liquid/solid extraction Reference Measurement Procedure selected ion monitoring solid phase extraction Selected reaction monitoring National Metrology Institute of Turkey, NMI: Turkey D.I. Mendeleyev Institute for Metrology, DI: Russia



v



di %di k n s

ts u(xi)

(x) U(x) U95(x)

Uk=2(x) x xi

zi



CCQM-K138 Final Report

SYMBOLS

degree of equivalence: xi - KCRV percent relative degree of equivalence: 100·di/KCRV coverage factor: U(x) = k·u(x) number of quantity values in a series of quantity values standard deviation of a series of quantity values:

Student’s t-distribution expansion factor standard uncertainty of quantity value xi pooled uncertainty: expanded uncertainty expanded uncertainty defined such that x ±U95(x) is asserted to include the true value of the quantity with an approximate 95 % level of confidence expanded uncertainty defined as Uk=2(x) = 2·u(x) a quantity value the ith member of a series of quantity values mean of a series of quantity values: z-score, a standardized quantity value:



vi



CCQM-K138 Final Report

INTRODUCTION

Dried fig, which is known to be a healthy food with its high nutritional value, could either be consumed directly or can be made into a paste/slurry to be used in desserts and candies [5]. The agricultural practices in the production of dried fig such as ripening, harvesting and sun-drying, present significant risk of fungal infection and subsequent mycotoxin contamination. In many products, severe limitations have been introduced by the European Union (EU) (Commission Regulation No 1058/2012 amending Regulation No 1881/2006) and the maximum limits have been established in the European legislation for various mycotoxins, which are extremely toxic, carcinogenic, tetratogenic and hepatotoxic such as aflatoxins B1, B2, G1 and G2. Due to the high level of aflatoxins, some products exported to the EU have been rejected and withdrawn. Weekly alert notifications are released on the internet for the member states through a Rapid Alert System, which is considered very important to protect both consumers and producers prior to consuming/processing [6-11]. The major producers of dried fig are Turkey, USA, Iran and Mediterranean countries, among which Turkey, as the producer of 60 % of the total worldwide supply, is involved in half of the international trade in dried figs. This makes Turkey a major exporter of dried fig, which requires it to comply with the internationally accepted sanitation and hygiene standards during production, storage and delivery. Thus, presence of any aflatoxin contamination in exported figs needs to be monitored according to Commission Regulation (EU) No 1058/2012 of 12 November 2012 amending Regulation (EC) No 1881/2006 as regards maximum levels for aflatoxins in dried figs and measured through reliable and traceable methods, which require pure and matrix certified reference materials.

Extraction, chromatographic separation, and quantification of low-concentration organic compounds in complex matrices are core challenges for reference material producers and providers of calibration services. Evidence of successful participation in formal, relevant international comparisons are needed to document measurement capability claims (CMCs) made by national metrology institutes (NMIs) and designated institutes (DIs).

In April 2015, the Consultative Committee for Amount of Substance: Metrology in Chemistry and Biology (CCQM) approved the Key Comparison (KC) CCQM-K138 Determination of aflatoxins (AFB1, AFB2, AFG1, AFG2 and Total AFs) in Dried Fig. CCQM-K138 was designed to assess participant capabilities for determination of mid-polarity contaminants in a food matrix. AFB1, AFB2, AFG1, AFG2 and Total AFs can be successfully evaluated using either Liquid chromatography (LC) Mass spectrometry, high performance liquid chromatography (HPLC) with different detection methods. Aflatoxins must be removed by extraction, following cleanup.

The following sections of this report document the timeline of CCQM-K138, the measurands, study material, participants, results and the measurement capability claims that participation in CCQM-K138 can support. The Appendices reproduce the official communication materials and summaries of information about the results provided by the participants.

1 of 36



CCQM-K138 Final Report



TIMELINE



Table 1 lists the timeline for CCQM-K138.



Table 1: Timeline for CCQM-K138



Date



Action



Apr 2015 Proposed to CCQM



October 2015



Draft protocol presented to OAWG as potential Track A or C Key Comparison



November 2015



OAWG authorized CCQM-K138 as a Track C Key Comparison; protocol approved



November 2015 Call for participation to OAWG members



March 2016 to

June 2016



Study samples shipped to participants. The range in shipping times reflects delays from shipping and customs.



September 2016 Results due to coordinating laboratory



October 2016 Draft A report distributed to OAWG



Apr 2018 Draft B report distributed to OAWG



TBD



Final report approved by OAWG



MEASURANDS

The measurands to be determined are the mass fractions of Aflatoxin (B1, B2, G1, G2 and total) in dried fig. The structures of Aflatoxins (AFB1, AFB2, AFG1 and AFG2) are given in Figure 1. The nominal values of Aflatoxins B1, B2, G1, G2 and total Aflatoxin are between mass fractions of 3 7 ng/g, 0.3 - 1 ng/g, 1 - 3 ng/g, 0.08 - 0.3 ng/g and 6 - 9.5 ng/g, respectively.

Figures 1- 4 below display the molecular structure of B1, B2, G1 and G2.



2 of 36



CCQM-K138 Final Report



Figure 1: Structure of AFB1

Aflatoxin B1 AFB1

pKOW 1.23



Figure 2: Structure of AFB2

Aflatoxin B2 AFB2

pKOW 1.45



Figure 3: Structure of AFG1

Aflatoxin G1 AFG1

pKOW 0.50



Figure 4: Structure of AFG2

Aflatoxin G2 AFG2

pKOW 0.71



STUDY MATERIALS

The test material is a candidate material for a dried fig certified reference material (CRM 1302).

Raw materials used in the production of dried figs were obtained from the Aydın province that meets about 70-75% of the production in Turkey. 300 kg of uncontaminated dried fig and 25 kg of dried fig contaminated by aflatoxin as Sarı Lop (Calimyrna) type were supplied from an exporting company in Aydın province as a starting material for the production of certified reference materials of aflatoxins in dried fig. The starting material was examined considering the visual UV findings before beginning of the process. All raw materials were subjected to gamma irradiation at around 5.3 kGy to prevent any microbiological activity. Since the aflatoxin content of the starting material was known to be stable under dry, dark and cold conditions, raw material was then kept in cold storage rooms at -18°C until the processing.

One of the most important and critical steps in the processing was the lyophilization, which is necessary to reduce moisture content of the material to minimize biological activity and improve long term storage stability. Lyophilization process was optimized for powder fig material which was obtained with the use of a blender homogenizer (Robot Coupe, Blixer 23, USA) with the

3 of 36



CCQM-K138 Final Report

addition of 13 % moisture-retaining material. The flowchart of the production process is given in Figure 5.

After lyophilization (loss of mass was 14%) and blending processes, all powder material was sieved with 500 μm sieve. After homogenization with 3-D mixer (3-D MegaMix, HKTM, Turkey), material was bottled (as 160 g to each bottle) using a semi-automatic filling machine (Augapack, Vectofill, Belgium) and capped materials were subjected to second gamma irradiation (5.3 kGy) before storing at -80°C.



Gamma irradiated contaminated dried fig Chopping with robot coupe homogenizer Lyophilization, sieving and vacuum packing in

aluminum sachets, storage at -18 °C Homogenisation with 3D mixer

Determination of aflatoxin content



Gamma irradiated uncontaminated (blank) dried fig Chopping with robot coupe homogenizer

Lyophilization, sieving and vacuum packing in aluminum sachets, storage at -18 °C Homogenisation with 3D mixer

Determination of aflatoxin content



Mixing blank and contaminated samples with 3D mixer, vacuum packing in HDPE containers, storage at -20 °C

Bottling, capping, labeling, gamma irradiation and storage at -80 oC



Homogeneity Test



Short Term Stability Test



Long Term Stability Test



Characterisation Study



Figure 5: The flowchart of the production process of dried fig

The powder product obtained from the production was filled into light-impermeable airtight brown bottles as 160 g. Totally 511 units were produced. Samples were randomly selected with TRaNS and subjected to homogeneity, stability and characterisation tests. The results obtained by the analysis of selected units were evaluated statistically.

Each participant received 2 units of candidate reference material: HDPE bottles into aluminum sachet, containing about 160 g of powder dried fig. The recommended minimum sample amount for analysis was at least 6 g. Measurement results were to be reported on as received basis.



4 of 36



CCQM-K138 Final Report



Homogeneity Assessment of Study Material

Homogeneity study between the units was performed to show that the assigned value was valid for all units within the stated uncertainty. In this study, 10 units were selected by using random stratified sampling software (TRaNS) and were reserved for the study of homogeneity between units. Homogeneity tests were carried out for all analytes of candidate CRM by measuring 3 subsamples (6 g sample size) under repeatability conditions. The method used for these measurements was validated and the samples to be analysed were introduced to the instrument by random order to find out any trend arising from analytical and/or filling sequences. All homogeneity measurements were carried out using HPLC-FLD method.



The data for all analytes were evaluated statistically by regression analysis for the presence of any trend in analytical and filling sequence. After evaluation of data, no trend was found for any analyte in CRM candidate at 95% confidence level.



Grubbs test was applied to all data for the presence of outlier at 95% confidence level. According to data obtained for each analyte, it was found that the distribution was found to be normal and no outliers were found (Table 2).



Table 2: Statistical Evaluation Result of Homogeneity of the Study Material



Analyte

AFB1 AFB2 AFG1 AFG2 Total AF



Is there aTrend?



Analytical sequence



Filling sequence



No



No



No



No



No



No



No



No



No



No



Is there an Outlier? All data Unit averages



No



No



No



No



No



No



No



No



No



No



Distribution All data

Normal/unimodal Normal/unimodal Normal/unimodal Normal/unimodal Normal/unimodal



Analysis of Variance (ANOVA) is a statistical tool used to estimate the uncertainty contribution from homogeneity of the materials. All data were examined for normal data distribution using Shapiro-Wilk test and histograms before applying one way ANOVA test. All analytes (AFB1, AFB2, AFG1, AFG2 and total AF) showed normal distribution on Shapiro-Wilk test and histogram diagrams. The uncertainties of homogeneity between units were evaluated with one way ANOVA for all analytes. The equation (1) was used for the calculation of the repeatability of the method (swb) and equation (2) was used for the calculation of standard deviation between units (sbb).



5 of 36



CCQM-K138 Final Report



       =            ℎ    

(1) where MSwithin : mean of square of variance within the unit

swb equals to “s” of the method as long as sub samples represent the whole unit.



where,



      



=



                  



−            ℎ       



(2)



MSbetween : mean of square of variance between units



n : number of replicates per unit



MSbetween is found to be smaller than MSwithin in conditions for which the heterogeneity of the material was smaller than heterogeneity that can be determined by the applied analytical method or measurement fluctuations that may have occurred randomly. In these cases, since sbb cannot be calculated, u*bb was calculated as heterogeneity contributing to uncertainty including method repeatability using equation (3).



   ∗   



=



         



4



2



             ℎ    



(3)



where, νMSwithin : degree of freedom of MSwithin

The uncertainty values obtained from the homogeneity study are given in Table 3.



Table 3: Homogeneity Results of the Study Material



Analyte AFB1 AFB2



Average value (ng/g) 5.38



sbb,rel 2(%.27)



u*bb,rel 2(%.28)



ubb,rel 2(%.28)



0.60



MSbetween<MSwithin



4.07



4.07



6 of 36



AFG1 AFG2 Total AF



CCQM-K138 Final Report



2.21



7.46



0.18



4.00



8.37



1.52



4.31 7.46 4.68 4.68 2.05 2.05



The values of MSbetween were found to be smaller than the values of MSwithin for analyte AFB2. So, u*bb was calculated and used as the uncertainty contribution due to homogeneity. For the cases where both sbb and u*bb can be calculated, the bigger one was taken as uncertainty contribution due to between bottle homogeneity (ubb).

The Results of the homogeneity assessment for AFB1, AFB2, AFG1, AFG2 in dried fig are given at Table 4.



Table 4: Results of the homogeneity assessment for AFB1, AFB2, AFG1, AFG2 in dried fig.



ANOVA Estimate Within-packet, CVwth: Between-packet, CVbtw: Total analytical variability, CV: Probability of falsely rejecting the

hypothesis that all samples have the same

concentration:



AFB1 7.34% 2.27% 2.28%



AFB2 13.1% MSbetween<MSwithin 4.07%



AFG1 13.9% 7.46% 4.31%



AFG2 15.1% 4.00% 4.68%



Total AFs 6.60% 1.52% 2.05%



76%



39%



83% 65% 56%



Stability Assessment of Study Material

Stability studies were performed with an isochronous design which is cited in ISO Guide 35. For the Short Term Stability (STS) test, two different temperatures (-20˚C and 4˚C) and 4 time points (1, 2, 3 and 4 weeks) were tested. 10 samples were selected by TRaNS. 8 samples were subjected to the test temperatures for the specified time intervals.

Samples were moved to -80˚C (reference temperature) after completion of the test time. All samples were analysed at the same time. Two replicate samples were prepared from each unit (6 g sample size) and were analyzed by HPLC-FLD method under the repeatability conditions for determining the mass fractions of AFB1, AFB2, AFG1, AFG2 and total AF.

The data for each temperature were first examined by single Grubbs test for both 95% and 99% confidence intervals to find out outliers. The number of detected outliers is given in the Table 5.

Since no technical reason can be found to reject these data, all outliers were included in the STS calculations.

7 of 36



CCQM-K138 Final Report



Values calculated for each time point were plotted against the time for the assessment of short term stability. The relationship between variables were analyzed in order to determine if any significant change exists in mass fraction values with the testing time (regression analysis). It was found that the slopes were not significantly different than zero for all in the 95% confidence interval.



Uncertainty calculations were done using equation (4). The maximum time for transfer was chosen as 2 weeks.



      



         = 



      (     −   )2



(4)



where,



RSD :



relative standard deviation obtained from all data in STS



ti



:



:



time point for each replicate mean of all time points



t



:



maximum time suggested for transfer: 2 weeks



Results obtained from short term stability are given in Table 5.



Table 5: Short Term Stability Test Results



Analyte



-20 °C usts,rel (%) for 2 weeks



AFB1



2.6



AFB2



2.1



AFG1



6.5



AFG2



6.1



Total AF



3.3



* One-sided Grubbs Test



4 °C usts,rel (%) for 2 weeks

2.9

2.8

4.9

4.6

4.0



Number of outliers in 95%

confidence interval*



Number of outliers in 99%

confidence interval*



Is there a significant trend

in 95% confidence

interval?



Is there a significant trend

in 99% confidence

interval?



-20 °C 4 °C -20 °C 4 °C -20 °C 4 °C -20 °C 4 °C



1



-



1



-



No



No



No



No



-



-



-



-



No



No



No



No



-



-



-



-



No



No



No



No



-



-



-



-



No



No



No



No



1



-



-



-



No



No



No



No



Result of this study showed that the sample could be transferred to the end users within a two week time interval ensuring the temperature not to exceed + 4C with cooling elements.

8 of 36



CCQM-K138 Final Report



Stability Assessment of Study Material (Long Term)

Shelf life of the produced CRM was determined by the long-term stability (LTS) studies. +4 oC was chosen as the test temperature for long term stability tests and in total 10 units were reserved for this study. Samples were selected by TRaNS software and kept at +4 oC for 9 months. Two units for each time point (0, 2, 4, 6, and 9 months) were stored at +4 oC and transferred to -80˚C (reference temperature) after completion of the test time. Two replicate samples (6 g sample size) were prepared from each unit and analyzed by HPLC-FLD under the repeatability conditions for determining the mass fractions of AFB1, AFB2, AFG1, AFG2 and total AF.



The data was first examined by one-sided Grubbs test for both 95% and 99% confidence intervals to find out outliers. The numbers of detected outliers are given in the Table 5. Since no technical reason was present to reject these data, all outliers were included in the LTS calculations.



Values calculated for each time point were plotted against the time for the assessment of LTS.



The relationship between variables were analyzed in order to determine if any significant change



exists in mass fraction values with the testing time (regression analysis). It was found that the



slopes were not significantly different than zero for all analytes in the 95% confidence interval.



The potential uncertainty contribution of long term stability, ults, was calculated using equation



(5) for 1 year of shelf life at +4 oC.



      



         = 



      (     −   )2



(5)



where,

RSD : relative standard deviation obtained from all data in LTS ti : time point for each replicate t : mean of all time points t : shelf life suggested at +4 oC: 1 year



9 of 36



CCQM-K138 Final Report



Table 6: Long Term Stability Test Results of the Study Material



Analyte



ults,rel (%) at +4 oC for 1

year



AFB1



11.1



AFB2



13.5



AFG1



12.8



AFG2



13.5



Total AF



10.1



* Single Grubbs Test



Number of outliers in

95% confidence

interval*



Number of outliers in 99%

confidence interval*



Is there a significant trend

in 95 % confidence

interval?



Is there a significant trend

in 99% confidence

interval?



-



-



No



No



1



-



No



No



1



1



No



No



-



-



No



No



-



-



No



No



PARTICIPANTS, INSTRUCTIONS AND SAMPLE DISTRIBUTION

The call for participation was distributed in November 2015 with the intent to distribute samples in February 2016, receive results in July 2016, and discuss results at the CCQM OAWG meeting, October 2016. See Table 1 for study timeline. Appendix A reproduces the Call for Participation; Appendix B reproduces the study Protocol.



Table 7 lists the institutions that registered for CCQM-K138



Table 7: Institutions Registered for CCQM-K138



NMI or DI



Code



Country



Contact



Bundesanstalt fuer



BAM



Germany Matthias Koch



Materialforschung und –pruefung



Matthias.Koch@bam.de



Chemical Metology Laboratory EXHM/GCSL- Greece



Elias Kakoulides



(General Chemical State Laboratory EIM



metrology@gcsl.gr



- Hellenic Metrology Institute)



Government Laboratory, Hong GLHK



Hong Kong Andy Chan



Kong



cmchan@govtlab.gov.hk



National Institute of Industrial



INTI



Argentina Estela Kneeteman



Technology, Toxicology and



estelak@inti.gob.ar



Nutrition Laboratory



Kenya Bureau of Standards, Food KEBS



Kenya



Mr. Isaac Mugenya



and Agriculture



mugenya@kebs.org



National Institute of Metrology of NIMT



Thailand Cheerapa Boonyakong



Thailand



cheerapa@nimt.or.th



National Metrology Institute of NMISA



South Africa Maria Fernandes-Whaley



South Africa



MFWhaley@nmisa.org



10 of 36



CCQM-K138 Final Report



D.I. Mendeleyev Institute for Metrology TUBITAK UME, National Metrology Institute



VNIIM

TUBITAK UME



Russia Turkey



Anatoliy Krylov ak@vniimex.ru Ahmet Ceyhan Gören Taner Gokcen



The participants were informed of the date of dispatching of samples. Each participant received 2 units of candidate reference material (HDPE bottles into aluminium sachet containing about 165 g of powder dried fig).

Due to delays in sample shipping and customs issues, the last set of material was delivered in June 2016. Because of these delays, the deadline for submission of results was postponed to 30 Sep 2016.

The participants were requested to report results from the mean of two samples, with corresponding standard and expanded uncertainty. The value of the results and their associated standard uncertainties must be expressed in ng/g. If the final result has been calculated from more than one method, the individual results from the contributing methods must also be reported. Participants were asked to provide information about the applied analytical procedure including the sample preparation and calibration methods and their metrological traceability. Each participant was asked to make an assessment of the measurement uncertainty. Each variable contributing to the uncertainty of the result was to be identified and quantified in order to be included in the combined standard uncertainty of the results. A full uncertainty budget was to be reported, as part of the results. All cells in all sheets (Result Reporting Form, Method Information, Comparison Results and Moisture Content Method) in Annex 2 “Report Form” was requested to be filled out in the Excel file provided in electronic form by TUBITAK UME.

RESULTS

Participants were requested to report a single estimate of the mass fraction ng/g for AFB1, AFB2, AFG1, AFG2 and total AF of independent measurements of two bottles. Results ranged from 5.17 to 7.27 ng/g with an %RSD of 10.47 for AFB1, ranged from 0.60 to 0.871 ng/g with an %RSD of 11.69 for AFB2, ranged from 1.98 to 2.6 ng/g with an %RSD of 10.36 for AFG1, ranged from 0.06 to 0.32 ng/g with an %RSD of 35.6 for AFG2, and ranged from 8.29 to 10.31 ng/g with an %RSD of 7.69 for Total AF.

In addition to the quantitative results, participants were instructed to describe their analytical methods, approach to uncertainty estimation, and the Core Competencies they felt were demonstrated in this study. Appendices C, D, and E reproduce the relevant report forms.

CCQM-K138 results were received from 9 of the 9 institutions that received samples.



11 of 36



CCQM-K138 Final Report

Calibration Materials Used by Participants

Participants used a range of different calibration materials, in several cases from commercial providers. Table 8 lists the calibrants used by each institute and how participants attempted to establish the traceability of the calibrants, where this was carried out. If this was via their own measurements, its assigned purity, the method used, and how the participant had demonstrated their competence in the use of the method(s) were also given in Table 8.

The issue of calibrant traceability was discussed at the OAWG meeting in September 2017. At that meeting it was flagged that many of the calibration materials employed did not meet the CIPM traceability requirements from CIPM 2009-24. This document allows two pathways: in house assessment using capabilities whose effectiveness has been demonstrated or the use of another NMI/DI’s capabilities where they have also been demonstrated.

The commercial materials used did not meet these CIPM criteria and thus where institutes did not carry out an independent in-house assessment then results using these calibrants could not be included in the KCRV. One instance that caused particular issue was the use of the IRMM ERMs. Several institutes used these materials assuming they would meet the CIPM traceability requirements, however these are certified by consensus from a range of different laboratories and hence they were not deemed to be acceptable.

Two institutes carried out in house assessment of the commercial calibrants in a way that was deemed sufficient to provide traceability.

EXHM purity assigned pure materials by mass balance and qNMR and then made up gravimetric solutions and measured them via IDMS versus the IRMM solutions. The purities of the AFs were given as below:

AFB1=96.13 ± 3.18%, AFB2=93.32 ± 3.13%, AFG1=98.60 ± 3.35%, AFG2=94.02 ± 3.12% (k=3 due to limited material)

The values on the certificate of the IRMM-ERM materials were:

AFB1=3.79 µg/g ± 2.90 %, AFB2=3.80 µg/g ± 2.11%, AFG1=3.78 µg/g ± 3.44%, AFG2=3.80 µg/g ± 1.84%, (k=2). The determined values by EXHM agreed with these values within their uncertainties.

TUBITAK UME purity assigned commercially available highly-pure substances by in-house qNMR purity assignment traceable to UME CRM 130. The purities of the AFs were given as below:

AFB1=85.47 ± 0.94%, AFB2=83.35 ± 1.25%, AFG1=77.13 ± 4.77%, AFG2=70.18 ± 0.46%.

12 of 36



CCQM-K138 Final Report The values of the Sigma standards were: AFB1=99.64% AFB2=98.50%, AFG1=100%, AFG2=100%.The values used by UME for these materials were those assigned in-house. Some other institutes did carry out assessment of materials. BAM used checks versus different lot numbers and different suppliers but as they were all commercial materials this was not deemed sufficient, BAM did use LC-MS for identity confirmation. INTI and KEBS used spectrophotometric analysis of their commercial calibration solutions however this was also deemed inappropriate. NIMT, NMISA and GLHK used the IRMM calibrants with no assessment and VNIIM used the Biopure materials with no assessment. As a result of the full analysis of the approaches used by all participants, due to the traceability requirements for the calibrants not being met, the results of KEBS, INTI, VNIIM and BAM were excluded from KCRV determination. If the institutes that had employed the IRMM materials were also excluded this would have left two institutes valid for the KCRV calculation. In this case a compromise was agreed to whereby it was deemed that the work done by EXHM had demonstrated the IRMM materials had valid assigned values and in this case the institutes that utilised those materials would have their results included. It is noted that all institutes except EXHM and UME would need to use different approaches to their calibration if they wished to have a CMC considered associated with this comparison considered.

13 of 36



NMI/DI BAM

EXHM / GCSL-EIM



Analyte

AFB1 AFB2 AFG1 AFG2

AFB1 AFB2 AFG1 AFG2



Table 8: Metrological Traceability of Participants’ Results



Source of Traceability Gravimetric sample preparation Aflatoxin B1:

16192B Aflatoxin B2:

15483A Aflatoxin G1:

15331C Aflatoxin G2:

15391A

IRMM-ERMAC057

IRMM-ERMAC058

IRMM-ERMAC059

IRMM-ERMAC060



Material Biopure

IRMM



Mass Fractiona Purity, %



Purity Techniquesb



Evidence of Competence



Certified standard solutions used.



Purities of calibration standards were



independently confirmed by LC-MS



measurements (scan mode; ESI+/-).



The specified aflatoxin contents of



-



the used certified standard solutions



N/A



were cross checked by certified



standards of different lot numbers



(same provider) and certified



standard solutions of a second



provider



Commercial



The solid aflatoxins were



solid aflatoxins: characterized for their purity using



AFB1=96.13±3. the Mass Balance approach and l Participation



18%,



qNMR. The concentration of the in CCQM-



AFB2=93.32±3. solutions prepared gravimetrically K104, P117.c,



13%,



were assigned against the IRMM CCQM-K131,



AFG1=98.60±3. CRM solutions (ERM AC 057, 058, CCQM-K78



35%,



059, 060) using IDMS experiments underpins



AFG2=94.02±3. The values assigned were found to claimed



12%.



agree with the gravimetric



uncertainties



preparations within the stated



uncertainties



14 of 36



NMI/DI GLHK

INTI



Analyte

AFB1 AFB2 AFG1 AFG2

AFB1 AFB2 AFG1 AFG2



Source of Traceability IRMM-ERM-

AC057 IRMM-ERM-

AC058 IRMM-ERM-

AC059 IRMM-ERM-

AC060

Aflatoxin B1 Cat. Code: 5032

Aflatoxin B2 Cat. Code: 5033

Aflatoxin G1 Cat. Code:5035

Aflatoxin G2 Cat. Code: 5036



Material IRMM

Fluka AG



Mass Fractiona Purity, %



Purity Techniquesb



N/A



Evidence of Competence



One solution of each aflatoxin was



prepared to obtain 4 stock solutions



of 8-10 ug/ml in acetonitrile. These



solutions were verified using an



Spectrophotometric method (AOAC



Manufacturer declaration using TLC and HPLC (more than 98%)



971.22). After the measurement of the stock solution at 350nm, it was

adjusted the purity of each calibration solution.

The assignment of purity was determined following the next



N/A



equation:



% purity = ccstandard stock x 10ml x 5000ul x100

50 ul x 1000 ug/mg x 10 mg



15 of 36



NMI/DI Analyte



Source of Traceability



Material



Mass Fractiona Purity, %



AFB1 Lot# AF017 Fermentek



AFB2 Lot#



Trilogy



AFB1



141104-070



Analytical



KEBS



AFB2 AFG1



AFG1 Lot#



Trilogy



-



AFG2



150305-070



Analytical



AFG2 Lot# 150309-070



Trilogy Analytical



NIMT



AFB1 AFB2 AFG1 AFG2



NMISA



AFB1 AFB2 AFG1 AFG2



IRMM-ERMAC057

IRMM-ERMAC058

IRMM-ERMAC059

IRMM-ERMAC060

IRMM-ERMAC057

IRMM-ERMAC058

IRMM-ERMAC059

IRMM-ERMAC060



IRMM IRMM

16 of 36



Purity Techniquesb HPLC/FLD

N/A N/A



Evidence of Competence

N/A



NMI/DI VNIIM



Analyte

AFB1 AFB2 AFG1 AFG2



Source of Traceability

Aflatoxin B1 in acetonitrile

Aflatoxin B2 in acetonitrile

Aflatoxin G1 in acetonitrile

Aflatoxin G2 in acetonitrile



Material



Mass Fractiona Purity, %



Biopure



Purity Techniquesb N/A



Evidence of Competence



TUBITAK UME



AFB1 AFB2 AFG1 AFG2



AFB1 A6636 AFB2 A9887 AFG1 A0138 AFG2 A0263



Sigma



AFB1=85.471 ±



0.943%, AFB2=83.351 ±

1.253%, AFG1=77.131 ±

4.767%, AFG2=70.178 ±



Purity of commercially available highly-pure substances were

determined by in-house qNMR purity assignment traceable to UME

CRM 1301



Participation in CCQM-K55b-d

underpins claimed uncertainties



0.455%



a Stated as Value ± U95(Value) b DSC: Differential scanning calorimetry

GC-FID: Gas chromatography with flame ionization detection HPLC-DAD: High pressure liquid chromatograph with diode-array detection MB: Mass balance qNMR: Quantitative nuclear magnetic resonance



17 of 36



Methods Used by Participants

Each laboratory was requested to use a properly validated method, calibration standards with a metrologically traceable assigned value (an appropriate CRM or material where its purity has been suitably assessed by the participant) according to criteria established by the CCQM OAWG for the inclusion of results in the calculation of the KCRV.



All participants based their analyses on LC-MS/MS, HR-LC-MS and HPLC-FLD. Brief descriptions of the analytical methods used by the participants, including sample preparation, analytical technique, calibrants and quantification approach is summarized in Appendix F Tables F1-5. The participants’ approaches to estimating uncertainty are provided in Appendix G.



The spread of results for each analyte was reasonably broad but there was no trend observed from the techniques used. INTI and KEBS used fluorescence detection whereas all other participants used IDMS. Significant effort was put into sample clean up by most participants, with immunoaffinity clean up being the most common. Only BAM used a simple centrifugation step which may have provided less selectivity.

Participant Results for AFB1, AFB2, AFG1, AFG2 and total AFs

The results for CCQM-K138 for the determination of aflatoxins (AFB1, AFB2, AFG1, AFG2 and total AFs) are detailed in Table 9 - 13 and presented graphically in Figure 6 -10 respectively. Results are ranging from 5.17 to 7.27 ng/g with an %RSD of 10.47 for AFB1, ranging from 0.60 to 0.871 ng/g with an %RSD of 11.69 for AFB2, ranging from 1.98 to 2.6 ng/g with an %RSD of 10.36 for AFG1, ranging from 0.06 to 0.32 ng/g with an %RSD of 35.64 for AFG2, and ranging from 8.29 to 10.31 ng/g with an %RSD of 7.69 for Total AF.



Table 9: Reported Results for AFB1, ng/g



NMI BAM

EXHM/GCSL-EIM GLHK INTI KEBS NIMT NMISA VNIIM

TUBITAK UME n



x 5.41 5.994 5.8 5.17 7.27 6.6 6.20 6.22 5.72 9.00



u(x) 0.15 0.123 0.5 0.33 0.8 0.40 0.28 0.23 0.33



AFB1, ng/g u(x) % k U(x) 2.77 2.571 0.40 2.05 2.03 0.249 8.62 2 1.1 6.38 2 0.66 11.00 2 1.6 6.06 2.03 0.9 4.52 2 0.56 3.70 2 0.46 5.77 2 0.66



U(x) % 9.06 4.15 18.97 12.77 22.83 13.64 9.03 7.40 11.54



6.04 s 0.63 CV 10.47



n = number of results included in summary statistics; = mean; s = standard deviation;

CV = 100·



18 of 36



Table 10: Reported Results for AFB2, ng/g



AFB2, ng/g



NMI



x u(x) u(x) % k U(x) U(x) %



BAM EXHM/GCSL-EIM

GLHK INTI

KEBS NIMT NMISA VNIIM TUBITAK UME



0.66 0.871 0.74 0.69 0.6 0.8 0.755 0.81 0.67



0.03 0.022 0.07 0.13 0.1 0.05 0.04 0.06 0.05



4.55 2.53 9.46 18.84 16.67 6.25 5.30 7.41 7.46



2.571 0.08 2.11 0.047 2 0.14 2 0.26 2 0.2 2.04 0.1 2 0.08 2 0.12 2 0.09



12.12 5.40 18.92 37.68 33.33 12.50 10.60 14.81 13.43



n 9.00



0.73 s 0.09 CV 11.69

n = number of results included in summary statistics;

CV = 100·



= mean; s = standard deviation;



Table 11: Reported Results for AFG1, ng/g



AFG1. ng/g



NMI



x u(x) u(x) % k U(x) U(x) %



BAM EXHM/GCSL-EIM

GLHK INTI

KEBS NIMT NMISA VNIIM TUBITAK UME



2.01 2.093 2 2.5 2.39 2.6 2.24 1.98 2.16



0.11 0.061 0.2 0.07 0.4 0.18 0.2 0.11 0.15



5.47 2.91 10.00 2.80 16.74 6.92 8.93 5.56 6.94



2.571 0.27 2.07 0.125 2 0.5 2 0.14 2 0.8 2.1 0.4 2 0.4 2 0.22 2 0.3



13.43 5.97 25.00 5.60 33.47 15.38 17.86 11.11 13.89



n 9.00



2.22 s 0.23 CV 10.36

n = number of results included in summary statistics;

CV = 100·



= mean; s = standard deviation;



19 of 36



Table 12: Reported Results for AFG2, ng/g



AFG2. ng/g



NMI



x u(x) u(x) % k U(x) U(x) %



BAM EXHM/GCSL-EIM

GLHK INTI

KEBS NIMT NMISA VNIIM TUBITAK UME



0.22 0.264 0.22 0.32 0.06 0.3 0.214 0.15 0.23



0.01 0.01 0.04 0.04 0.01 0.03 0.025 0.04 0.02



4.55 3.79 18.18 12.50 16.67 10.00 11.68 26.67 8.70



2.571 0.03 2.2 0.022 2 0.07 2 0.08 2 0.02 2 0.1 2 0.049 2 0.08 2 0.04



13.64 8.33 31.82 25.00 33.33 33.33 22.90 53.33 17.39



n 9.00



0.22 s 0.08 CV 35.60

n = number of results included in summary statistics;

CV = 100·



= mean; s = standard deviation;



Table 13: Reported Results for Total AF, ng/g



Total AFs ng/g



NMI



x u(x) u(x) % k U(x) U(x) %



BAM 8.29 0.19 2.29 2.571 0.49 5.91 EXHM/GCSL-EIM 9.223 0.141 1.53 2 0.282 3.06



GLHK 8.7 0.6 INTI 8.68 0.57

KEBS 10.31 1.34 NIMT 10.3 0.44 NMISA 9.4 0.65



6.90 2 1.2 6.57 2 1.14 13.00 2 2.68 4.27 2.57 1.2 6.91 2 1.3



13.79 13.13 25.99 11.65 13.83



VNIIM 9.16 0.27 2.95 2 TUBITAK UME 8.78 0.35 3.99 2



0.54 5.90 0.7 7.97



n 9.00



9.20 s 0.71



CV 7.69

n = number of results included in summary statistics; = mean; s = standard deviation;

CV = 100·



20 of 36



Figure 6: Reported Results for AFB1, ng/g

Panels A and B display the reported results for AFB1; panel A displays the results sorted alphabetically by NMI Acronym, panel B displays results sorted by increasing reported value. Dots represent the reported mean values, x; bars their 95 % expanded uncertainties, U(x). The thin horizontal gridlines are provided for visual guidance.

Figure 7: Reported Results for AFB2, ng/g

Panels A and B display the reported results for AFB2; panel A displays the results sorted alphabetically by NMI Acronym, panel B displays results sorted by increasing reported value. Dots represent the reported mean values, x; bars their 95 % expanded uncertainties, U(x). The thin horizontal gridlines are provided for visual guidance.

21 of 36



Figure 8: Reported Results for AFG1, ng/g

Panels A and B display the reported results for AFG1; panel A displays the results sorted alphabetically by NMI Acronym, panel B displays results sorted by increasing reported value. Dots represent the reported mean values, x; bars their 95 % expanded uncertainties, U(x). The thin horizontal gridlines are provided for visual guidance.

Figure 9: Reported Results for AFG2, ng/g

Panels A and B display the reported results for AFG2; panel A displays the results sorted alphabetically by NMI Acronym, panel B displays results sorted by increasing reported value. Dots represent the reported mean values, x; bars their 95 % expanded uncertainties, U(x). The thin horizontal gridlines are provided for visual guidance.

22 of 36



Figure 10: Reported Results for Total AF, ng/g

Panels A and B display the reported results for Total AF; panel A displays the results sorted alphabetically by NMI Acronym, panel B displays results sorted by increasing reported value. Dots represent the reported mean values, x; bars their 95 % expanded uncertainties, U(x). The thin horizontal gridlines are provided for visual guidance.



Discussion of Results



The Draft A Report was sent to the participants to review in March 2017. The examination of the data revealed that NMISA correctly reported two individual results for AFG2 however the mean was incorrectly calculated, they reported a corrected mean result for AFG2 before the April 2017 OAWG meeting. The NMISA results for AFG2 are given in table 14.

Table 14: Reported Results for AFG2, ng/g



Participating



Overall Mean



u



k



Institutes



(ng/g)



(ng/g)



NMISA (First result)



0.214



0.025



2



U (ng/g)

0.049



NMISA (Corrected



0.229



0.026



2



result)



0.053



VNIIM had followed-up on their results following the October 2016 OAWG meeting, and in the April 2017 meeting they confirmed that their AFG2 result remained unchanged at 0.15 ng/g.



23 of 36



KEY COMPARISON REFERENCE VALUE (KCRV)



Selecting an appropriate KCRV estimator for these small and reasonably variable datasets was carefully considered. It was decided at the OAWG meeting in September 2017 in Ottawa, for UME to consider the suitability of using a Linear Pool as a potential KCRV estimator. The linear pool estimator is suitable as it reflects the overall diversity amongst the individual results and calculates the KCRV as the average expected value that would be reported by any participant. It is considered a good estimator where there are small datasets with variability.



The results of Linear pool KCRV estimator are given in Table 15, in conjunction with other estimators that were considered. The Linear pool KCRV relative to the reported results for AFB1, AFB2, AFG1, AFG2 are presented graphically in Figure 11 -15.

Table 15. Candidate Key Comparison Reference Values



Estimator u?a



Median



No



DL-Mean 1 No



DL-Mean 2 No



Bayesian



No



Linear Pool Yes



X 5.99 6.022 6.02 6.04 6.06



AFB1, ng/g u(X) U95(X)b



0.17



0.34



0.066 0.183



0.09



0.26



0.16 -0.30/+0.34



0.47 -0.93/+1.01



X 0.76 0.774 0.774 0.777 0.766



AFB2, ng/g



u(X)



U95(X)b



0.037



0.075



0.037



0.103



0.035



0.096



0.044 -0.092/+0.084



0.083 -0.126/+0.134



Estimator u?a



AFG1, ng/g



X



u(X) U95(X)b



AFG2, ng/g



X u(X)



U95(X)b



Median



No 2.16 0.066 0.133



0.23 0.013



0.027



DL-Mean 1 No 2.195 0.090 0.251 0.248 0.014



0.038



DL-Mean 2 No 2.20 0.095 -0.27/+0.26 0.248 0.015



0.040



Bayesian



No 2.18 0.098 -0.18/+0.21 0.250 0.016 -0.034/+0.029



Linear Pool Yes 2.22 0.265 -0.46/+0.59 0.246 0.042 -0.079/+0.089



Estimator u?a Median No DL-Mean 1 No DL-Mean 2 No Bayesian No Linear Pool Yes



Total AF, ng/g



X



u(X)



U95(X)b



9.22 0.367



0.735



9.272 0.252



0.699



9.27 0.27 -0.74/+0.73



9.27 0.33 -0.64/+0.68



9.28 0.742 -1.35/+1.52



a) Does the estimator utilize the information in the reported uncertainties? b) U95(X) = ts·u(X), where ts is the appropriate two-tailed Student’s t critical value for 95 % coverage.



24 of 36



Figure 11: Linear pool KCRV relative to the reported results for AFB1, ng/g

The results are sorted by increasing reported value. Dots represent the reported mean values, x; bars their standard uncertainties, u(x). The blue horizontal line denotes the candidate KCRV. The bracketing dashed lines denote the standard uncertainty of the candidate KCRV. The red data points were not included in the KCRV calculation.

Figure 12: Linear pool KCRV relative to the reported results for AFB2, ng/g

The results are sorted by increasing reported value. Dots represent the reported mean values, x; bars their standard uncertainties, u(x). The blue horizontal line denotes the candidate KCRV. The bracketing dashed lines denote the standard uncertainty of the candidate KCRV. The red data points were not included in the KCRV calculation.

25 of 36



Figure 13: Linear pool KCRV relative to the reported results for AFG1, ng/g

The results are sorted by increasing reported value. Dots represent the reported mean values, x; bars their standard uncertainties, u(x). The blue horizontal line denotes the candidate KCRV. The bracketingdashed lines denote the standard uncertainty of the candidate KCRV. The red data points were not included in the KCRV calculation.

Figure 14: Linear pool KCRV relative to the reported results for AFG2, ng/g

The results are sorted by increasing reported value. Dots represent the reported mean values, x; bars their standard uncertainties, u(x). The blue horizontal line denotes the candidate KCRV. The bracketing dashed lines denote the standard uncertainty of the candidate KCRV. The red data points were not included in the KCRV calculation.

26 of 36



Figure 15: Linear pool KCRV relative to the reported results for Total AF, ng/g

The results are sorted by increasing reported value. Dots represent the reported mean values, x; bars their standard uncertainties, u(x). The blue horizontal line denotes the candidate KCRV. The bracketing dashed lines denote the standard uncertainty of the candidate KCRV. The red data points were not included in the KCRV calculation.



DEGREES OF EQUIVALENCE (DoE)

The absolute degrees of equivalence for the participants in CCQM-K138 are estimated as the signed difference between the combined value and the KCRV: di = xi – KCRV. KCRV is estimated from Linear Pool Procedure of 5 participants’ results. Since only 5 participants’ results are entered to NICOB database to estimate KCRV and their DoE.U95 values, in order to calculate DoE.U95 values for other participants NICOB program ran a second time with all values and their values derived from this outcome. Table 16-20 below lists the numeric values of di, U95(di), di, and U95(di) for all participants in CCQM-K138 for AFB1, AFB2, AFG1, AFG2 and Total AFs. DOE and DOE% graphs are given in figure 16-25.



Table 16: Degrees of Equivalence for AFB1



NMI INTI BAM UME GLHK EXHM NMISA VNIIM NIMT KEBS



d

-0.90 -0.66 -0.34 -0.27 -0.07 0.14 0.15 0.53 1.20



AFB1, ng/g Uk=2(d) %d



1.16



-14.78



1.01



-10.82



1.15



-5.68



1.36



-4.40



1.00



-1.17



1.11



2.23



1.07



2.55



1.23



8.81



1.82



19.87



Uk=2(%d)

19.07 16.70 18.98 22.40 16.56 18.26 17.58 20.28 30.02



27 of 36



The entries in italic are results not included in the KCRV calculation



Table 17: Degrees of Equivalence for AFB2



AFB2, ng/g



NMI



d



Uk=2(d) %d Uk=2(%d)



KEBS



-0.167 0.252 -21.79 32.80



BAM



-0.107 0.162 -13.98 21.15



UME



-0.097 0.183 -12.67 23.88



INTI



-0.078 0.302 -10.12 39.37



GLHK



-0.027 0.209



-3.57



27.23



NMISA -0.012 0.172



-1.57



22.40



NIMT



0.033 0.184



4.30



23.96



VNIIM



0.043 0.195



5.58



25.46



EXHM 0.103 0.156 13.40 20.28



The entries in italic are results not included in the KCRV calculation



Table 18: Degrees of Equivalence for AFG1



AFG1, ng/g



NMI



d



Uk=2(d) %d Uk=2(%d)



VNIIM -0.239 0.564 -10.77 25.40



GLHK



-0.219 0.655



-9.86



29.51



BAM



-0.209 0.567



-9.42



25.54



EXHM -0.126 0.538



-5.68



24.24



UME



-0.059 0.600



-2.65



27.02



NMISA 0.022 0.655



0.97



29.49



KEBS



0.170 0.943



7.64



42.48



INTI



0.281 0.543 12.65 24.47



NIMT



0.381 0.629 17.16 28.35



The entries in italic are results not included in the KCRV calculation



Table 19: Degrees of Equivalence for AFG2



AFG2, ng/g



NMI



d



Uk=2(d) %d Uk=2(%d)



KEBS



-0.186 0.086 -75.49 34.94



VNIIM -0.096 0.114 -38.93 46.38



NMISA -0.032 0.097 -12.88 39.24



GLHK -0.026 0.114 -10.50 46.31



BAM



-0.026 0.086 -10.46 34.92



UME



-0.016 0.092



-6.36



37.56



EXHM 0.018 0.086



7.46



35.04



NIMT



0.054 0.102 22.07 41.39



INTI



0.074 0.115 30.19 46.57



The entries in italic are results not included in the KCRV calculation



28 of 36



Table 20: Degrees of Equivalence for Total AF



Total AF, ng/g



NMI



di



U(di) % di % U(di)



BAM



-0.992 1.494 -10.68 16.09



INTI



-0.603 1.837



-6.49



19.80



GLHK



-0.582 1.875



-6.27



20.20



UME



-0.503 1.612



-5.42



17.37



VNIIM



-0.123 1.546



-1.32



16.66



EXHM -0.061 1.475



-0.66



15.89



NMISA 0.120 1.941



1.30



20.91



NIMT



1.018 1.689 10.97 18.20



KEBS



1.027 3.002 11.07 32.35



The entries in italic are results not included in the KCRV calculation



Figures 16-25 below graphically presents both the DOE and DOE% for AFB1, AFB2, AFG1, AFG2 and Total AFs.



Figure 16: Absolute degrees of equivalence for AFB1 in CCQM-K138.

All results are sorted by increasing value. The axis to the left edge displays the absolute DoE, d, in units [ng/g]. The vertical bars correspond to ±U(di). The horizontal blue line marks the zero deviation from the KCRV.

29 of 36



Figure 17: Relative degrees of equivalence for AFB1 in CCQM-K138.

All results are sorted by increasing value. The axis to the left edge displays the relative DoE, 100•d/KCRV, as percent. The vertical bars correspond to ± U(%di). The horizontal blue line marks the zero deviation from the KCRV.

Figure 18: Absolute degrees of equivalence for AFB2 in CCQM-K138.

All results are sorted by increasing value. The axis to the left edge displays the absolute DoE, d, in units [ng/g]. The vertical bars correspond to ±U(di). The horizontal blue line marks the zero deviation from the KCRV.

30 of 36



Figure 19: Relative degrees of equivalence for AFB2 in CCQM-K138.

All results are sorted by increasing value. The axis to the left edge displays the relative DoE, 100•d/KCRV, as percent. The vertical bars correspond to ± U(%di). The horizontal blue line marks the zero deviation from the KCRV.

Figure 20: Absolute degrees of equivalence for AFG1 in CCQM-K138.

All results are sorted by increasing value. The axis to the left edge displays the absolute DoE, d, in units [ng/g]. The vertical bars correspond to ±U(di). The horizontal blue line marks the zero deviation from the KCRV.

31 of 36



Figure 21: Relative degrees of equivalence for AFG1 in CCQM-K138.

All results are sorted by increasing value. The axis to the left edge displays the relative DoE, 100•d/KCRV, as percent. The vertical bars correspond to ± U(%di). The horizontal blue line marks the zero deviation from the KCRV.

Figure 22: Absolute degrees of equivalence for AFG2 in CCQM-K138.

All results are sorted by increasing value. The axis to the left edge displays the absolute DoE, d, in units [ng/g]. The vertical bars correspond to ±U(di). The horizontal blue line marks the zero deviation from the KCRV.

32 of 36



Figure 23: Relative degrees of equivalence for AFG2 in CCQM-K138.

All results are sorted by increasing value. The axis to the left edge displays the relative DoE, 100•d/KCRV, as percent. The vertical bars correspond to ± U(%di). The horizontal blue line marks the zero deviation from the KCRV.

Figure 24: Absolute degrees of equivalence for Total AF in CCQM-K138.

All results are sorted by increasing value. The axis to the left edge displays the absolute DoE, d, in units [ng/g]. The vertical bars correspond to ±U(di). The horizontal blue line marks the zero deviation from the KCRV.

33 of 36



Figure 25: Relative degrees of equivalence for Total AF in CCQM-K138.

All results are sorted by increasing value. The axis to the left edge displays the relative DoE, 100•d/KCRV, as percent. The vertical bars correspond to ± U(%di). The horizontal blue line marks the zero deviation from the KCRV.

34 of 36



USE OF CCQM-K138 IN SUPPORT OF CALIBRATION AND MEASUREMENT CAPABILITY (CMC) CLAIMS

How Far the Light Shines

Successful participation in CCQM-K138 demonstrates the following measurement capabilities in determining mass fraction of organic compounds, with molecular mass of 100 g/mol to 500 g/mol, having high polarity pKow > -2, in mass fraction range from 0.05 ng/g to 500 ng/g in dried food matrices.

It is noted that figs are a high carbohydrate form of dried foods and thus extrapolation to other types of dried food matrices should take this into account.

Core Competency Statements and CMC support

Tables E1 to E9 list the Core Competencies claimed by the participants in CCQM-K138. The information in these Tables is as provided by the participants; however, the presentation of many entries has been condensed and standardized. Details of the analytical methods used by each participant in this study are provided in Appendix F. The core competency tables are annotated to reflect the actual performance of the participants.

CONCLUSIONS

The results for CCQM-K138 represent a highly challenging set of measurands and involve very low level measurement of complex analytes in a situation where there is very limited availability of appropriate calibration materials. Participants have demonstrated capabilities to measure these analytes at levels of ranging from 5.41 ng/g to 7.27 ng/g with uncertainties ranging from 0.12 ng/g to 0.80 ng/g for AFB1; levels from 0.60 ng/g to 0.871 ng/g with uncertainties ranging from 0.022 ng/g to 0.13 ng/g for AFB2; levels from 1.98 ng/g to 2.6 ng/g with uncertainties ranging from 0.061 ng/g to 0.4 ng/g for AFG1; levels from 0.06 ng/g to 0.32 ng/g with uncertainties ranging from 0.01 ng/g to 0.04 ng/g for AFG2; levels from 8.29 ng/g to 10.31 ng/g with uncertainties ranging from 0.141 ng/g to 1.34 ng/g for Total AF.

In terms of analytical methods, most participants used immunoaffinity column cleanup and only one used SPE cleanup. All participants used liquid chromatography technique. 2 participants used florescence detector and 7 used MS detector.

Areas for improvement largely involve appropriate assessment of the traceability of the calibrants used for these measurements.

Due to the variability in results the degrees of equivalence for these analytes were reasonably large and this will need to be taken into consideration in the assessment of proposed CMCs.

35 of 36



ACKNOWLEDGEMENTS

The study coordinators thank all of the participating laboratories for providing the requested information during the course of these studies. We would like to thank to Lindsey Mackay, Michael Nelson and Katrice Lippa for their invaluable contributions to the report.

REFERENCES

[1] Steiner. W.E.. Rieker. R.H.. Battaglia. R.. 1988. “Aflatoxin contamination in dried figs: distribution and association with fluorescence”. Journal of Agricultural and Food Chemistry. 36. 88-91.

[2] Anklam. E.. Gilbert. J.. 2002. “Validation of analytical methods for determining mycotoxins in foodstuffs”. Trends in Analytical Chemistry. 21. 468-486.

[3] Gilbert. J.. Şenyuva. H.. 2008. “Fungal and mycotoxin contamination of dried figs-a review”. Mycotoxins. 58 (2). 73-82.

[4] Imperato. R.. Campone. L.. Piccinelli. A.L.. Veneziano. A.. Rastrelli. L.. 2011. “Survey of aflatoxins and ochratoxin a contamination in food products imported in Italy”. Food Control. 22. 1905-1910.

[5] Bircan. C.. Barringer. S.A.. Ulken. U.. Pehlivan. R.. 2008. ”Aflatoxin levels in dried figs. nuts and paprika powder for export from Turkey”. International Journal of Food Science and Technology. 43. 1492-1498.

[6] htts://ec.europa.eu/food/food/rapidalert/report2007-en.pdf. [7] http://ec.europa.eu/food/food/rapidalert/docs/report2009\_en.pdf. [8] http://ec.europa.eu/food/safety/rasff/docs/rasff\_annual\_report\_2010\_en.pdf. [9] http://ec.europa.eu/food/safety/rasff/docs/rasff\_annual\_report\_2011\_en.pdf. [10] http://ec.europa.eu/food/safety/rasff/docs/rasff\_annual\_report\_2012\_en.pdf. [11] http://ec.europa.eu/food/safety/rasff/docs/rasff\_annual\_report\_2013.pdf

36 of 36



APPENDIX A: Call for Participation



From: Lindsey.Mackay@measurement.gov.au Date: 25.11.2015 21:50



Dear OAWG colleagues

Attached please find all of the documentation for our next Track C key comparison for aflatoxins in fig. Please return registration forms to UME by 4 December and contact me if you have any questions about the comparison.

Many thanks

Lindsey



Attachments:



CCQM K138/P174\_ Registration form.docx CCQM K138/P174 Technical protocol.docx CCQM K138 Core Competency Table .doc CCQM K138/P174 Report Form.xlsx CCQM K138/P174 Sample Receipt Confirmation form



A 1 of 1



APPENDIX B: Protocol

CCQM-K138 and P174

Mass fractions of aflatoxins (AFB1, AFB2, AFG1, AFG2 and total AFs) in dried fig

TECHNICAL PROTOCOL

CCQM-K138 and P174

Key and Pilot Comparisons on

“Determination of aflatoxins (AFB1, AFB2, AFG1, AFG2 and Total AFs) in Dried Fig”

Call for Participants and Technical Protocol

(February 24, 2016)

1. Introduction Dried fig can be consumed directly or as fig paste/slurry in desserts and candies [1]. It is considered a healthy food as its nutritional value is high. It has highly alkaline property, which makes it useful in balancing the pH of fibre. It is a rich source of potassium and calcium, which is important in helping to regulate blood pressure and as an alternative to dairy products for the people who have allergies. Calcium and potassium are also important in preventing osteoclasis. Dried fig contains good level of magnesium, iron, copper and manganese. Tryptophan in fig induces good sleep and helps in preventing sleeping disorders like insomnia. It helps to reduce the risk of breast cancer and blood cholesterol level [2].

The production of dried fig involves some unique agricultural practices such as ripening, harvesting and sun-drying. These practices present significant risk of fungal infection and subsequent mycotoxin contamination. National and international institutions and organization such as the European Commission (EC), the US Food and Drug Administration (FDA), the World Health Organization (WHO) and the Food and Agricultural Organization (FAO) have recognized that mycotoxins have potential risk to human and animal health. Regulations have been established in many countries to protect consumers from their harmful effects. The European Union (EU) has introduced severe limits in many products for major mycotoxin classes as high risk of contamination (Commission Regulation No.1881/2006). The European legislation has set maximum limits for various mycotoxins in food and feed, including Aflatoxins (B1, B2, G1 and G2), which are extremely toxic, carcinogenic, tetratogenic and hepatotoxic.

B-1 of 11



Exported products to the EU are sometimes rejected and withdrawn because of high levels of aflatoxins. Alert notifications are published weekly on the internet to inform the member states by a Rapid Alert System as the monitoring is very important for consumer protection and producers of raw products prior to transport or processing [3-8].

Turkey, USA, Iran and Mediterranean countries are the major producers of dried fig. Half of the international trade in dried figs is conducted by Turkey, which produces 60 % of the total worldwide supply. Therefore, the sustainable export of dried figs has great significance for the Turkish agricultural economy. To ensure its sustainability, it is necessary to satisfy internationally accepted sanitation and hygiene standards during production, storage and delivery to consumers. To this end. aflatoxin contamination in exported figs should be monitored through reliable and traceable measurement methods. The traceability of aflatoxin measurement results can be achieved through the use of pure and matrix certified reference material. However, for the determination of aflatoxins in dried fig, such certified reference materials are not yet available. There is a lack of certified reference materials (CRMs) for use by routine testing laboratories in method validation and as quality controls. In addition, commercial proficiency testing (PT) programmes, commonly participated in by routine testing laboratories, make use of consensus results instead of metrologically traceable assigned values to evaluate the performance of the participating laboratories. The proposed study material is a candidate certified reference material for the determination of aflatoxin (B1, B2, G1, G2 and total) levels in dried fig [9-12].

The study was first proposed as a key comparison and presented at the EURAMET TC-MC SCOA meeting in Malta in 2015. During the meeting, three NMIs / DIs expressed interest to participate in the study. Hence, the meeting recommended that the study should proceed and be presented during the CCQM OAWG meeting by EURAMET. The study was subsequently presented at the CCQM OAWG meeting in April 2015. CCQM OAWG members from other RMOs would also be invited to participate in the study. During the meeting, five NMIs/DIs expressed interest to participate in the study. An approval was subsequently obtained from the CCQM OAWG Chair to organize this study as a Track C key comparison and as a pilot study.

2. Test material

The test material is a candidate material for a dried fig certified reference material (CRM). The mass fraction of aflatoxin (B1, B2, G1, G2 and total) in dried fig will be certified in the near future. The results of this comparison will be mentioned in the certification document.

2.1. Preparation of Study Samples by TUBITAK UME

The raw material to be used in this study was obtained from the province of Aydın which supplies 70-75 % of all dried figs in Turkey. Dried fig material was stored at -18 ºC until processing. Raw material was blended by a blade mixer, dehydrated using a freeze dryer, grounded, sieved and packed as 165 g in bottles. The bottles were packed with foil-laminate

B-2 of 11



sachets under vacuum. All the sample bottles were stored at room temperature (-80 ± 3) C inside prior to distribution or use. Steps in the preparation of study samples are given in Scheme 1 below.

Scheme 1. Preparation steps of study samples 2.2. Homogeneity and Stability Testing of samples The homogeneity of the material was investigated by analyzing 12 bottles selected from 500 bottles. The bottles were randomly and stratified selected. Three subsamples (6 g) were taken from each bottle for homogeneity. The data were treated with ANOVA. The samples were measured in a random order under repeatability conditions. The data were technically scrutinised and statistically evaluated according to ISO Guide 30 to 35. The preservatives were found to be sufficiently homogeneous in the study material. The relative standard uncertainties due to between-bottle inhomogeneity for AFB1, AFB2, AFG1, AFG2 and total AFs were found to be

B-3 of 11



2.28 %, 4.61 %, 4.31 %, 4.68 % and 2.05 %, respectively. The results of the homogeneity study are summarized in Table 1.



Table 1. Results of the homogeneity assessment for target aflatoxin (B1, B2, G1, G2 and total) in dried fig



S (%) between

u*bb (%) ubb (%)

RSD F

F-critic P-value



AFB1 2.27 2.28 2.28 7.67 1.29

0.29



AFB2 MSbetween<

MSwithin

4.07



AFG1 7.46 4.31



4.07 8.40 0.63

0.79



7.46 15.67 1.86 2.22 0.10



AFG2 4.00 4.68 4.68 15.57 1.21

0.33



Total AFs 1.52 2.05 2.05 6.76 1.16

0.36



When MSbetween is smaller than MSwithin, Sbetween cannot be calculated. This does not prove that the material is perfectly homogeneous, but only indicates that study set-up was not good enough to quantify heterogeneity. Instead of Sbb, u*bb, the heterogeneity that can be hidden by the method repeatability, is calculated.

A four week isochronous study was performed to evaluate stability of candidate reference material during transport. The bottles were selected using a random stratified sample picking computer programme. Two subsamples (50 g) were taken from each bottle for stability tests. For a short-term stability study, -20 ºC and 4 ºC were selected as test temperatures. The selected test periods were 0, 1, 2, 3 and 4 weeks. After the indicated storage periods. the samples were stored at -80 °C until analysis. For each test temperature and test periods, 2 bottles were analyzed. Each sample bottle was analyzed in duplicate. Two replicates of each bottle were analyzed randomly under repeatability conditions. All data were evaluated for short-term stability test according to ISO Guides 30 to 35. Regression lines were calculated to detect possible degradation. Although the slope was found to be indistinguishable from zero for storage temperatures of -20 ºC and 4 ºC, a significant slope was found when the samples were stored at -80 ºC. The uncertainty of the short-term stability (usts) can be assumed to be negligible if the sample shipment is carried out with cooling elements or on dry ice. Results of the short-term stability study are summarized in Table 2.

Table 2. Results of the short-term stability assessment for target Aflatoxin (B1, B2, G1, G2 and total) in dried fig



B-4 of 11



AFB1 AFB2 AFG1 AFG2 Total AFs



o

u (%) (-20 C)

sts

1.20 1.22 3.23 2.88 1.38



o

u (%) (+4 C)

sts

1.20 1.36 2.46 2.31 1.22



The same method (HPLC-FLD) was used for the homogeneity and short-term stability measurements.



For the long-term stability study, -20 ºC and 4 ºC remained as the selected test temperatures, while the test periods of 2, 4, 6 and 9 months were used. All data will be evaluated for long-term stability test according to ISO Guides 30 to 35 until the deadline for submission of results.



Different amount of subsamples was used for the minimum sample intake study.



Table 3. Results of the minimum sample intake study for target Aflatoxin (B1, B2, G1, G2 and total) in dried fig



Analyte



RSD



2g



4g



6g



50 g



AFB1



20



19



9



9



AFB2



13



13



7



4



AFG1



16



21



15



15



AFG2



22



22



18



14



Results of the minimum sample intake study are summarized in Table 3. According to results, minimum sample intake is recommended as at least 6 g.

In the same day (within day repeatability), the relative standard deviation of measurement results of AFB1, AFB2, AFG1, AFG2 and total AFs were found 7.75 %, 8.74 %, 16.2 %, 16.7 % and 7.09 %, respectively for 6 g sample intake.



B-5 of 11



3. Measurands

The measurands to be determined are the mass fractions of aflatoxin (B1, B2, G1, G2 and total) in dried fig. The structures of Aflatoxins (AFB1, AFB2, AFG1 and AFG2) are given in Figure 1.



Figure 1. Structures of Aflatoxins (AFB1, AFB2, AFG1 and AFG2)



The nominal values of Aflatoxin B1 are between mass fractions of 3 ng/g to 7 ng/g. Aflatoxin B2 between mass fractions of 0.3 ng/g to 1 ng/g, Aflatoxin G1 between mass fractions of 1 ng/g to 3 ng/g, Aflatoxin G2 between mass fractions of 0.08 ng/g to 0.3 ng/g, total Aflatoxin between mass fractions of 6 ng/g to 9.5 ng/g.



Analytes and those nominal values in the candidate reference material are also given in Table 3.



Table 3. AFB1, AFB2, AFG1, AFG2 and Total AFs Expected Mass Fractions



Analytes



Mass Fraction (ng/g)



AFB1 AFB2 AFG1



3-7 0.3-1 1-3



B-6 of 11



AFG2 Total AFs



0.08-0.3 6-9.5



4. Handling and storage

To avoid any decomposition, the samples should be kept sealed until they are used. They should be stored at the temperature from -20 oC to +4 oC in its original bottle, tightly capped and not exposed to intense direct light and ultraviolet radiation. The samples should be opened carefully and the measurement should be carried out immediately after the samples are opened.

5. Distribution

The participants will be informed of the date of dispatching of samples. Each participant will receive 2 units candidate reference material (HDPE bottles into aluminium sachet containing about 165 g of powder dried fig).

Participants are required to acknowledge the receipt of the sample. and return the receipt to TUBITAK UME by e-mail. If there is any damage on the sample, TUBITAK UME will send a substitute sample on request. A Sample Receipt Confirmation Form as a receipt form will be distributed to the participants. After receiving the sample, it should be kept at a temperature between -20 and +4 °C.

6. Methods/procedures

Each participant is encouraged to use their typical analytical method. Please include a full description of your method of analysis when reporting the results. For this purpose, a “Report Form” will be sent to the participants. NMIs or officially designated institutes are welcome to participate in this comparison. If ID-MS methods are used, the source of isotopically labeled spike material used should be reported.

7. Analysis and Uncertainty Evaluation

The units should be stored between -20 to +4 °C and should be equilibrated to room temperature before analysis for 2 hours.

Before opening the sample, the material must be homogenised by shaking the container for 2 min to prevent possible clumping. The analysis should be conducted with a recommended sample size of at least 6 g.

The report should comprise a brief description of the measurement method (including sample preparation) as well as a brief description of quality assurance measures. The calibration solutions and the individual results (for each parameter analyzed) should be reported in ng/g. All



B-7 of 11



results must be linked to the TUBITAK UME sample identification number (unit number) and to the date of the analyses.

Each participant laboratory should use an appropriate approach following the ISO/GUM and the approach used to derive the uncertainty budget must be briefly described in the report. Each variable contributing to the uncertainty of the result should be identified and quantified in order to be included in the combined standard uncertainty of the result. A full uncertainty budget must be included in the report.

Every participant laboratory should use its usual aflatoxin calibrants and establish their traceability.

8. Reporting and submission of results and core capability assessment

The result should be reported as the mass fraction of each measurand, mean of from two sample, to TUBITAK UME, accompanied by a full uncertainty budget. The result should be submitted using the attached Report Form.

Furthermore, all participants in this comparison are required to complete a Core Capability Table for the measurement technique they used. Templates for the appropriate techniques will be sent to the registered participants when the sample is distributed. The filled-out table should be submitted together with the measurement result.

Please complete and submit the attached Report Form and the Core Capability Table to TUBITAK UME (E-mail: ahmetceyhan.goren@tubitak.gov.tr) by e-mail before the scheduled deadline.

The report must include:

 Result should be reported as a value of independent measurements of two bottles of comparison sample with corresponding standard and expanded uncertainty.

 The value of the results and their associated standard uncertainties must be expressed in ng/g.

 If the final result has been calculated from more than one method, the individual results from the contributing methods must also be reported.

 A detailed description of the applied analytical procedure including the sample preparation and calibration methods.

 Participants are asked to provide information about their metrological traceability.  Each participant should make an assessment of the measurement uncertainty. Each

variable contributing to the uncertainty of the result should be identified and quantified in order to be included in the combined standard uncertainty of the results. A full uncertainty budget must be reported, as part of the results.  All cells in all sheets (Result Reporting Form and Method Information) in Annex 2 “Report Form” should be filled out in the excel file that will be provided in electronic form by TUBITAK UME.

B-8 of 11



9. KCRV

Each laboratory should use a properly validated method, calibration standards with a metrologically traceable assigned purity value (CRM or material where its purity has been suitably assessed by the participant) according to criteria established by the CCQM OAWG for the inclusion of results in the calculation of the KCRV. Exclusion of data points in the KCRV calculation will be using a sound of metrological basis. On the basis of this information, appropriate estimators and uncertainty evaluation for the KCRV will be proposed (see for reference the "OAWG Practices and Guidelines" document). It is expected that it is most likely that each reference value will be the median of the submitted data from NMIs and officially designated institutes, though it will be decided after discussion in CCQM OAWG meeting. If any participant submitted individual results by multiple methods, their best result (i.e., with the smallest uncertainty) will be selected to calculate the reference value. The final decision regarding the assignment of a KCRV and its uncertainty for CCQM-K138 and P174, will be taken after discussion in the November 2016 CCQM OAWG meeting.

10. How Far Does the Light Shine?

This Key Comparison will demonstrate capabilities for low molecular mass (100 g/mol to 500 g/mol) analytes of high polarity (pKow > -2) at the 0.05 ng/g to 500 ng/g mass fraction range in dried food matrices.

11. Program schedule

 Draft protocol and conformation: October 2015  Call for participation: November 2015  Deadline for registration: December 2015  Distribution of study sample: February 2016  Deadline for submission of results: 30th September 2016  Presentation/initial discussion of results: November 2016 CCQM OAWG  Draft A report: December 2016

12. Participants

Participation is open to all interested NMIs or officially designated institutes that can perform the determination.

13. Coordinating laboratory

The CCQM-K138 and P174 are coordinated by TUBITAK UME. TUBITAK UME takes all responsibilities for the development and operation of the key comparison, including preparation and distribution of samples, initial data analysis and evaluation of results to facilitate OAWG discussions, draft reports, and communications with participants.

B-9 of 11



14. Registration

Please complete and return the attached registration forms to TUBITAK UME (E-mail: ahmetceyhan.goren@tubitak.gov.tr) for the participation. Successful registration will be notified by e-mail. Please register no later than 04 December 2015.

15. Confidentiality

The participating laboratories will receive the reports giving all results for assessment/comments. The participating laboratories will be identified in the reports. The key comparison is conducted in the belief that participants will perform the analysis and report results with scientific rigor. Collusion between participants or falsification of results is clearly against the spirit of this study. Once approved by the OAWG, this report will be available on the open access section of the BIPM website. Participants may not publish any such data until the key comparison report has been published on the KCDB.

16. Contact

For any enquiries, participants may wish to contact the persons from coordinating laboratory are as follows:

TUBITAK Ulusal Metroloji Enstitusu (UME)

Dr. Ahmet Ceyhan GOREN

E-mail: ahmetceyhan.goren@tubitak.gov.tr

Phone: 00 90 262 679 50 00 (6102)

Fax: 00 90 262 679 50 01

17. References

[1] Bircan. C.. Barringer. S.A.. Ulken. U.. Pehlivan. R.. 2008. ”Aflatoxin levels in dried figs. nuts and paprika powder for export from Turkey”. International Journal of Food Science and Technology. 43. 1492-1498.

[2] http://www.figyork.com/production.html. [3] htts://ec.europa.eu/food/food/rapidalert/report2007-en.pdf. [4] http://ec.europa.eu/food/food/rapidalert/docs/report2009\_en.pdf. [5] http://ec.europa.eu/food/safety/rasff/docs/rasff\_annual\_report\_2010\_en.pdf. [6] http://ec.europa.eu/food/safety/rasff/docs/rasff\_annual\_report\_2011\_en.pdf. [7] http://ec.europa.eu/food/safety/rasff/docs/rasff\_annual\_report\_2012\_en.pdf.

B-10 of 11



[8] http://ec.europa.eu/food/safety/rasff/docs/rasff\_annual\_report\_2013.pdf [9] Steiner. W.E.. Rieker. R.H.. Battaglia. R.. 1988. “Aflatoxin contamination in dried figs:

distribution and association with fluorescence”. Journal of Agricultural and Food Chemistry. 36. 88-91. [10] Anklam. E.. Gilbert. J.. 2002. “Validation of analytical methods for determining mycotoxins in foodstuffs”. Trends in Analytical Chemistry. 21. 468-486. [11] Gilbert. J.. Şenyuva. H.. 2008. “Fungal and mycotoxin contamination of dried figs-a review”. Mycotoxins. 58 (2). 73-82. [12] Imperato. R.. Campone. L.. Piccinelli. A.L.. Veneziano. A.. Rastrelli. L.. 2011. “Survey of aflatoxins and ochratoxin a contamination in food products imported in Italy”. Food Control. 22. 1905-1910.

B-11 of 11



APPENDIX C: Registration Form

CCQM-K138 and P174 Mass fractions of aflatoxins (AFB1, AFB2, AFG1, AFG2 and total AFs) in

dried fig

REGISTRATION FORM



Please complete the following:



Name of Institute



:



Acronym of Institute (if available)



:



Name of Laboratory/Department



Name of Contact Person



:



Designation



:



E-mail Address



:



Telephone Number



:



Fax Number



:



Postal Address



:



Postal Code



:



Country



:



Date



:



Please tick the appropriate boxes.



C-1 of 2



1) We would like to register for the following measurements; Analytes



Aflatoxin B1 (AFB1)



Aflatoxin B2 (AFB2)



Aflatoxin G1 (AFG1)



Aflatoxin G2 (AFG2)



Total Aflatoxin (Total AFs)



K138 P174



K138 P174



K138 P174



K138 P174



K138 P174



2) Do you require a special custom permit for the samples to be sent to your laboratory?

Yes No (If yes, please give the details in a separate paper.)

Please note that any import taxes or charges, imposed on the material during transportation, shall be met by the participating laboratory.

Kindly complete and return this form by e-mail or fax no later than 04 December 2015 to:



Dr. Nilgun TOKMAN TUBITAK UME Gebze Yerleskesi P.K. 54 41470 Gebze-Kocaeli/Turkey E-mail: nilgun.tokman@tubitak.gov.tr Phone: 00 90 262 679 50 00 (6203) Fax: 00 90 262 679 50 01



Dr. Ahmet Ceyhan GOREN TUBITAK UME Gebze Yerleşkesi P.K. 54 41470 Gebze-Kocaeli/Turkey E-mail: ahmetceyhan.goren@tubitak.gov.tr Phone: 00 90 262 679 50 00 (6201) Fax: 00 90 262 679 50



If you do not receive an acknowledgement for your registration from us within 4 working days, please send us an email.



C-2 of 2



APPENDIX D: Reporting Form

The original form was distributed as an Excel workbook. The following are pictures of the relevant portions of the workbook’s two worksheets. “Result Reporting Form” worksheet

D-1 of 4



“Result Reporting Form” worksheet (continued)

“Method Information” Worksheet D-2 of 4



“Method Information” Worksheet (continued) D-3 of 4



D-4 of 4



APPENDIX E: Core Competency Tables CCQM OAWG: Competency Template for Analyte(s) in Matrix



Instructions:

 In the middle column place a tick, cross or say the entry is not applicable for each of the competencies listed (the first row does not require a response)

 Fill in the right hand column with the information requested in blue in each row  Enter the details of the calibrant in the top row, then for materials which would not meet the CIPM traceability

requirements the three rows with a # require entries. 

Table E.1. Core Competencies Demonstrated in CCQM-K138 by BAM



CCQM-K138 Scope of Measurement:



Mass fractions of aflatoxins (AFB1, AFB2, BAM AFG1, AFG2 and total AFs) in dried fig



This Track C Key Comparison will demonstrate capabilities for low molecular mass (100 g/mol



to 500 g/mol) analytes of high polarity (pKow > -2) at the 0.05 ng/g to 500 ng/g mass fraction range in dried food matrices.



Tick.



Competency



cross. or “N/A”



Specific Information as Provided by NMI/DI



Competencies for Value-Assignment of Calibrant



Calibrant: Did you use a “highly-pure



 calibration solutions (Biopure. RomerLabs):



substance” or calibration solution?



B1: 16192B, B2: L15483A., G1: L15331C, G2:



Identity verification of analyte(s) in



L15391A, 13C17-Afla-Mix: I15383M







Mass spectrometric investigations (MRM.



calibration material.#



fragmentation pattern)



For calibrants which are a highly-pure



N/A Commercial certified standard solutions used, not



substance: Value-Assignment / Purity



suited for e.g. qNMR (purity assessment).



Assessment method(s).# For calibrants which are a calibration







Certified standard solutions used. Purities of



solution: Value-assignment method(s).#



calibration standards were independently confirmed



by LC-MS measurements (scan mode; ESI+/-). The



specified aflatoxin contents of the used certified



standard solutions were cross checked by certified



standards of different lot numbers (same provider)



and certified standard solutions of a second provider.



Sample Analysis Competencies



Identification of analyte(s) in sample







Retention time, internal standard, mass spec ion



Extraction of analyte(s) of interest from



ratios (quantifier/qualifier)







shaking extraction



matrix Cleanup - separation of analyte(s) of interest







IAC (AflastarTM)



from other interfering matrix components (if



used)



Transformation - conversion of analyte(s) of



N/A -



interest to detectable/measurable form (if



used) Analytical system







HPLC-MS/MS



E-1 of 10



Calibration approach for value-assignment of analyte(s) in matrix

Verification method(s) for value-assignment of analyte(s) in sample (if used) Other







IDMS; six-point calibration; linear regression



N/A N/A -



E-2 of 10



Table E.2. Core Competencies Demonstrated in CCQM-K138 by EXHM



Mass fractions of aflatoxins (AFB1, AFB2,



CCQM-K138



EXHM AFG1, AFG2 and total AFs) in dried fig



Scope of Measurement:



This Track C Key Comparison will demonstrate capabilities for low molecular mass (100 g/mol



to 500 g/mol) analytes of high polarity (pKow > -2) at the 0.05 ng/g to 500 ng/g mass fraction



range in dried food matrices.



Tick,



Competency



cross, or “N/A”



Specific Information as Provided by NMI/DI



Competencies for Value-Assignment of Calibrant



Calibrant: Did you use a “highly-pure



ERM AC 057, 058, 059, 060 solutions



substance” or calibration solution?



in-house Aflatoxin solutions



Identity verification of analyte(s) in calibration material.#



✔



LC-MS/MS



For calibrants which are a highly-pure substance: Value-Assignment / Purity



✔



mass balance (LC-UV,KF titration, ICPMS)



qNMR



Assessment method(s).#



For calibrants which are a calibration



✔



solution: Value-assignment method(s).#



IRMM CRMs: used certified values checked versus qNMR analysis of



in house pure materials: UV-Vis (according to EN



14123 )



Sample Analysis Competencies



Identification of analyte(s) in sample



✔



Retention time, mass spec ion ratios



Extraction of analyte(s) of interest from



✔



matrix



Cleanup - separation of analyte(s) of interest ✔ from other interfering matrix components (if



used)



Transformation - conversion of analyte(s) of N/A



interest to detectable/measurable form (if



used)



Analytical system



✔



Liquid/liquid, ASE immunoaffinity column -LC-MS/MS



Calibration approach for value-assignment of ✔ analyte(s) in matrix



IDMS, exact matching matrix matched, single-point calibration



Verification method(s) for value-assignment N/A



--



of analyte(s) in sample (if used)



Other



N/A



--



E-3 of 10



Table E.3. Core Competencies Demonstrated in CCQM-K138 by GLHK



CCQM-K138



Mass fractions of aflatoxins (AFB1, AFB2, GLHK AFG1, AFG2 and total AFs) in dried fig



Scope of Measurement:



This Track C Key Comparison will demonstrate capabilities for low molecular mass (100 g/mol



to 500 g/mol) analytes of high polarity (pKow > -2) at the 0.05 ng/g to 500 ng/g mass fraction range in dried food matrices.



Tick,



Competency



cross, or “N/A”



Specific Information as Provided by NMI/DI



Competencies for Value-Assignment of Calibrant



Calibrant: Did you use a “highly-pure



Calibration Solutions Used:



substance” or calibration solution?



Aflatoxin B1 : IRMM ERM – AC057



Aflatoxin B2 : IRMM ERM – AC058



Aflatoxin G1 : IRMM ERM – AC059



Aflatoxin G2 : IRMM ERM – AC060



Identity verification of analyte(s) in



N/A



--



calibration material.#



For calibrants which are a highly-pure



N/A



--



substance: Value-Assignment / Purity



Assessment method(s).#



For calibrants which are a calibration



N/A



--



solution: Value-assignment method(s).#



Sample Analysis Competencies



Identification of analyte(s) in sample



✔



Retention time and ion ratio of mass spectrometric



analysis



Extraction of analyte(s) of interest from



✔



matrix



Liquid/solid extraction by high speed homogenizer – 2 times extraction by water and followed by 3 times extraction by 80% methanol



Cleanup - separation of analyte(s) of interest ✔ from other interfering matrix components (if used) Transformation - conversion of analyte(s) of N/A interest to detectable/measurable form (if used)



Analytical system



✔



Calibration approach for value-assignment of ✔ analyte(s) in matrix



Verification method(s) for value-assignment N/A



of analyte(s) in sample (if used)



Other



N/A



Cleanup - by immunoaffinity column

--

LC-MS/MS Quantification mode used - Isotope Dilution Mass Spectrometry Calibration mode used – Standard addition ---



E-4 of 10



Table E.4. Core Competencies Demonstrated in CCQM-K138 by INTI



CCQM-K138



INTI



Mass fractions of aflatoxins (AFB1, AFB2, AFG1, AFG2 and total AFs) in dried fig



Scope of Measurement:



This Track C Key Comparison will demonstrate capabilities for low molecular mass (100 g/mol



to 500 g/mol) analytes of high polarity (pKow > -2) at the 0.05 ng/g to 500 ng/g mass fraction range in dried food matrices.



Tick,



Competency



cross, or “N/A”



Specific Information as Provided by NMI/DI



Competencies for Value-Assignment of Calibrant



Calibrant: Did you use a “highly-pure



Pure material. Fluka AG. Aflatoxins B1, B2, G1 and



substance” or calibration solution?



G2 from Aspergillus Flavus. Aflatoxin B1 Cat. Code:



5032 Batch 2216541280. Aflatoxin B2 Cat. Code:



5033 Batch 202621578. Aflatoxin G1 Cat.



Code:5035 Batch 219939181. Aflatoxin G2 Cat.



Code: 5036 Batch 219940181.



Identity verification of analyte(s) in calibration material.#



✓



Spectrophotometric method (AOAC 971.22).



For calibrants which are a highly-pure substance: Value-Assignment / Purity



✓



One solution of each aflatoxin was prepared to obtain



4 stock solutions of 8-10 ug/ml in acetonitrile. These



Assessment method(s).#



solutions were verified using an Spectrophotometric



method (AOAC 971.22). After the measurement of



the stock solution at 350nm. it was adjusted the



purity of each calibration solution.



The assignment of purity was determinated following



the next equation:



For calibrants which are a calibration solution: Value-assignment method(s).#

Sample Analysis Competencies

Identification of analyte(s) in sample

Extraction of analyte(s) of interest from matrix Cleanup - separation of analyte(s) of interest from other interfering matrix components (if used) Transformation - conversion of analyte(s) of interest to detectable/measurable form (if used) Analytical system

Calibration approach for value-assignment of analyte(s) in matrix Verification method(s) for value-assignment of analyte(s) in sample (if used)

Other



% purity = ccstandard stock x 10ml x 5000ul x100 50 ul x 1000 ug/mg x 10 mg

N/A Indicate how you established analyte mass fraction in calibration solution



✓



Retention time with external standard



✓



The analyte is extracted using solvent extraction



(MeOH+H20) (8+2 v/v)



✓



Cleanup with immunoaffinity column.



Chromatographic Separation with LC.



✓



Post-column derivatization involving bromination.



(Kobra Cell)



✓



LC-FD (Liquid Chromatography with fluorescence



detector.



✓



a) external standard



b) 5 points calibration curve



✓



We do not use any verification. Verification is not



necessary due specificity of cleanup separation used



in the method.



N/A Indicate any other competencies demonstrated.



E-5 of 10



Table E.5. Core Competencies Demonstrated in CCQM-K138 by KEBS



CCQM-K138



KEBS



Mass fractions of aflatoxins (AFB1, AFB2, AFG1, AFG2 and total AFs) in dried fig



Scope of Measurement: This Track C Key Comparison will demonstrate capabilities for low molecular mass (100 g/mol to 500 g/mol) analytes of high polarity (pKow > -2) at the 0.05 ng/g to 500 ng/g mass fraction range in dried food matrices.



Competency



Tick, cross, or “N/A”



Specific Information as Provided by NMI/DI



Competencies for Value-Assignment of Calibrant

Calibrant: Did you use a “highly-pure substance” or calibration solution?



Calibration solution used AFB1 Source FERMENTEK Lot# AF017 AFB2 Source TRILOGY ANALYTICAL LABORATORY Lot# 141104-070 AFG1 Source TRILOGY ANALYTICAL LABORATORY Lot# 150305-070 AFG2 Source TRILOGY ANALYTICAL LABORATORY Lot# 150309-070



Identity verification of analyte(s) in calibration material.#



√



UV/VIS with Acetonitrile as solvent



For calibrants which are a highly-pure substance: Value-Assignment / Purity Assessment method(s).# For calibrants which are a calibration solution: Value-assignment method(s).#



N/A



--



√



UV/VIS with Acetonitrile as solvent and



application of Beer’s Law



Sample Analysis Competencies

Identification of analyte(s) in sample



√



Retention time



Extraction of analyte(s) of interest from



√



Extraction by using a shaker and 80% Methanol



matrix



Cleanup - separation of analyte(s) of interest



√



Immunoaffinity columns used



from other interfering matrix components (if



used)



Transformation - conversion of analyte(s) of



√



Electro-chemical derivertization



interest to detectable/measurable form (if



used)



Analytical system



√



HPLC with FL detection



Calibration approach for value-assignment of analyte(s) in matrix



a)External standard



√



b) 5- point calibration curve



Verification method(s) for value-assignment



N/A



--



of analyte(s) in sample (if used)



Other



N/A



--



NOTE: KEBS results for AFG2 was not consistent with the KCRV and had a DoE that did not cross



zero. The specific reason for this deviation was not identified although KEBs did not use appropriate



traceable calibrants for all of the analytes.



E-6 of 10



Table E.6. Core Competencies Demonstrated in CCQM-K138 by NIMT



CCQM-K138



Mass fractions of aflatoxins (AFB1, AFB2, NIMT AFG1, AFG2 and total AFs) in dried fig



Scope of Measurement:



This Track C Key Comparison will demonstrate capabilities for low molecular mass (100 g/mol



to 500 g/mol) analytes of high polarity (pKow > -2) at the 0.05 ng/g to 500 ng/g mass fraction range in dried food matrices.



Tick,



Competency



cross, or “N/A”



Specific Information as Provided by NMI/DI



Competencies for Value-Assignment of Calibrant



Calibrant: Did you use a “highly-pure



ERM-C057, ERM-C058, ERM-C059, ERM-C060



substance” or calibration solution?



Identity verification of analyte(s) in



N/A --



calibration material.#



For calibrants which are a highly-pure



N/A --



substance: Value-Assignment / Purity



Assessment method(s).#



For calibrants which are a calibration



N/A Gravimetric



solution: Value-assignment method(s).#



Sample Analysis Competencies



Identification of analyte(s) in sample







The analytes in the samples were identified against



ERM-CO57, ERM-CO58, ERM-CO59 and ERM-



CO60 standards by comparing their retention times



and m/z of LC-MS/MS.



Extraction of analyte(s) of interest from







Liquid-liquid extraction using 70:30 MeCN: water



matrix



with 20 mL of extraction solvent: 10 grams sample



Cleanup - separation of analyte(s) of interest







Immunoaffinity column (IAC)



from other interfering matrix components (if



used)



Transformation - conversion of analyte(s) of



N/A Indicate chemical transformation method(s), if any,



interest to detectable/measurable form (if



(i.e., hydrolysis, derivatization. other)



used)



Analytical system







LC-MS/MS



Calibration approach for value-assignment of







a) IDMS.



analyte(s) in matrix



b) 6-point calibration



Verification method(s) for value-assignment of analyte(s) in sample (if used) Other



N/A -N/A --



E-7 of 10



Table E.7. Core Competencies Demonstrated in CCQM-K138 by NMISA



CCQM-K138 Scope of Measurement:



Mass fractions of aflatoxins (AFB1, AFB2, NMISA AFG1, AFG2 and total AFs) in dried fig



This Track C Key Comparison will demonstrate capabilities for low molecular mass (100 g/mol



to 500 g/mol) analytes of high polarity (pKow > -2) at the 0.05 ng/g to 500 ng/g mass fraction range in dried food matrices.



Tick,



Competency



cross, or “N/A”



Specific Information as Provided by NMI/DI



Competencies for Value-Assignment of Calibrant



Calibrant: Did you use a “highly-pure



IRMM ERM individual aflatoxin B1, B2, G1 and G2



substance” or calibration solution?



ERM solutions respectively:



ERM\_AC057AFB1 ILM010 Lot I15231A



ERM\_AC058AFB2 ILM011 Lot I15345B



ERM\_AC059AFG1 ILM012 Lot I15345A



Identity verification of analyte(s) in



ERM\_AC060 AFG2 ILM013 Lot I15232G







Verification by comparison of HPLC-FLD. UPLC-



calibration material.#



MS/MS Retention time, FLD excitation-emission wavelength, multi-reaction monitoring (MRM) ion



ratio transitions unique to the toxins using IRMM



ERMs and other commercial standards of the



mycotoxins (Biopure™ and Trilogy™)



For calibrants which are a highly-pure



N/A -



substance: Value-Assignment / Purity



Assessment method(s).#



For calibrants which are a calibration



N/A -



solution: Value-assignment method(s).#



Sample Analysis Competencies



Identification of analyte(s) in sample







Identification by comparison of HPLC-FLD, UPLC-



MS/MS Retention time. FLD excitation-emission



wavelength., multi-reaction monitoring (MRM) ion



ratio transitions unique to the toxins using IRMM



ERMs and other commercial standards of the



Extraction of analyte(s) of interest from



mycotoxins (Biopure™ and Trilogy™)







Methanol: Water (80:20)saline solid-liquid extraction



matrix Cleanup - separation of analyte(s) of interest



of the dried fig powder with shaking 60 min.







Immunoaffinity clean-up (VICAM Aflatest)



from other interfering matrix components (if



used)



Transformation - conversion of analyte(s) of



N/A -



interest to detectable/measurable form (if



used) Analytical system







UPLC-ESI-MS/MS, HPLC-FLD independent check.



Calibration approach for value-assignment of







a) double IDMS, standard addition, external standard



analyte(s) in matrix



b) 9-point std addition; 6-point external calibration; 3



brackets dIDMS.



Verification method(s) for value-assignment



N/A -



of analyte(s) in sample (if used)



Other



N/A -



E-8 of 10



Table E.8. Core Competencies Demonstrated in CCQM-K138 by VNIIM



CCQM-K138



Mass fractions of aflatoxins (AFB1, AFB2, VNIIM AFG1, AFG2 and total AFs) in dried fig



Scope of Measurement:



This Track C Key Comparison will demonstrate capabilities for low molecular mass (100 g/mol



to 500 g/mol) analytes of high polarity (pKow > -2) at the 0.05 ng/g to 500 ng/g mass fraction range in dried food matrices.



Tick,



Competency



cross, or “N/A”



Specific Information as Provided by NMI/DI



Competencies for Value-Assignment of Calibrant



Calibrant: Did you use a “highly-pure



Calibration solution. RM from Biopure



substance” or calibration solution?



Identity verification of analyte(s) in







LCMS



calibration material.#



For calibrants which are a highly-pure



N/A -



substance: Value-Assignment / Purity



Assessment method(s).#



For calibrants which are a calibration



N/A From certificate of analysis



solution: Value-assignment method(s).#



Sample Analysis Competencies



Identification of analyte(s) in sample







Retention time, mass spec ion ratios



Extraction of analyte(s) of interest from matrix Cleanup - separation of analyte(s) of interest from other interfering matrix components (if used) Transformation - conversion of analyte(s) of interest to detectable/measurable form (if used) Analytical system

Calibration approach for value-assignment of analyte(s) in matrix







Sonication







SPE



N/A -







LC-MS/MS







IDMS, single point calibration



Verification method(s) for value-assignment of analyte(s) in sample (if used) Other



N/A N/A -



E-9 of 10



Table E.9. Core Competencies Demonstrated in CCQM-K138 by TUBITAK UME



CCQM-K138



Mass fractions of aflatoxins (AFB1, AFB2, UME AFG1, AFG2 and total AFs) in dried fig



Scope of Measurement:



This Track C Key Comparison will demonstrate capabilities for low molecular mass (100 g/mol



to 500 g/mol) analytes of high polarity (pKow > -2) at the 0.05 ng/g to 500 ng/g mass fraction range in dried food matrices.



Tick,



Competency



cross, or “N/A”



Specific Information as Provided by NMI/DI



Competencies for Value-Assignment of Calibrant



Calibrant: Did you use a “highly-pure



 Highly pure substance. commercially available from



substance” or calibration solution?



SIGMA: Aflatoxin B1 from aspergillus flavus



A6636; Aflatoxin B2 A9887; Aflatoxin G1 A0138;



Aflatoxin G2 A0263



Identity verification of analyte(s) in







High Resolution LC-MS



calibration material.# For calibrants which are a highly-pure







Purity of commercially available highly-pure



substance: Value-Assignment / Purity



substances were determined by in-house qNMR



Assessment method(s).#



purity assignment traceable to UME CRM 1301



For calibrants which are a calibration



N/A



solution: Value-assignment method(s).#



Sample Analysis Competencies



Identification of analyte(s) in sample







Retention time, MS ion



Extraction of analyte(s) of interest from matrix Cleanup - separation of analyte(s) of interest from other interfering matrix components (if used) Transformation - conversion of analyte(s) of interest to detectable/measurable form (if used) Analytical system

Calibration approach for value-assignment of analyte(s) in matrix







Solid-liquid extraction







Immuno Affinity Column (R-BIOPHARM . EASI



EXTRACT AFLATOXIN RP70N)



N/A -







High Resolution LC-MS







Isotope Dilution Mass Spectrometry (IDMS), five-



point calibration



Verification method(s) for value-assignment of analyte(s) in sample (if used) Other



N/A N/A -



E-10 of 10



APPENDIX F: Summary of Participants’ Analytical Information

The following Tables summarize the detailed information about the analytical procedures each participant provided in their “Analytical Information” worksheets. The presentation of the information in many entries has been consolidated and standardized. The participant’s measurement uncertainty statements are provided verbatim in Appendix G.

F-1 of 20



Table F.1. : Summary of Sample Size, Extraction, and Cleanup for CCQM-K138



Institute



Pre-treatment



Extraction Method



Sample Size (units)



Clean-up



BAM



Fifteen grams of the homogenised sample were weighed into a 120 mL polypropylene (PP) centrifugation tube



Fifteen grams of the homogenised



sample were weighed into a 120 mL



polypropylene (PP) centrifugation tube,



followed by the addition of 1.5 g



sodium chloride and 90 mL of a



mixture of methanol/water (80:20, v/v),



followed by the addition of 1.5 g



15 g



sodium chloride and 90 mL of a



mixture of methanol/water (80:20, v/v).



The tube was closed and the mixture



was shaken for 30 min at ambient



temperature in a mechanical shaker



(300 strokes/min).



The aqueous-methanolic layer was separated by centrifugation (ambient temperature, 10 min, 3000 rpm (1942 g)).



EXHM



11,4 g of the test material are mixed with water at 1:2 ratio to produce a slurry and is then spiked with labelled aflatoxins (13C17 B1, B2, G1 and G2) and left for 1 hour to equilibrate. 10 g of the slurry is mixed with 1 g NaCl and is then extracted with 60 mL MeOH:H2O 80:20 in a high sheer mixer for 3 min.



11.4 g (slurried with 23,6 g Η2Ο)



Immunoaffinity column The extract is filtered and 16,5 g are mixed with 60 mL PBS buffer and pass through an IAC column. The column is washed with water and flushed with MeOH to collect the aflatoxins. The resultant solution is evaporated to dryness and redisolved in MeOH:H2O 2:1 and analysed in an LCMS/MS system

IA columns: Aflastar R (Rohmer Labs)



F-2 of 20



Institute GLHK

INTI KEBS



Pre-treatment



Sample



Extraction Method



Size (units)



Clean-up



Liquid/solid extraction by high speed



homogenizer – 2 times extraction by



water and followed by 3 times extraction by 80% methanol Volume of water used : 72 mL



6 g per analysis



Clean-up by immunoaffinity column



Volume of 80% methanol used 108 mL



A test portion of 25g is extracted with



MeOH-H2O (8+2). Extract is filtered,



diluted with PBS and applied to an



affinity column. Aflatoxins are



removed from the affinity column with MeOH and are quantified by reversed-



25 g



phase liquid chromatography with post-



column derivatization involving



bromination (Kobra cell) and



determined by fluorescence detection.



Extraction by using a shaker and 80%



Methanol



SHAKING WITH 80% METHANOL Shake sample with extraction solvent



0.0



for 40min, Filter using filter paper,



15mL filtrate mixed with 85mL PBS,



Cleanup with immunoaffinity column. Chromatographic Separation with LC. Immunoaffinity Column Liquid Chromatography with Post-Column Derivatization (AOAC 999.07) Clean up method: Immunoaffinity column Aflatest WB VICAM - Elution solvent: MeOH

Immunoaffinity columns used 10mL (filtrate with PBS) loaded to IAC then washed with PBS. Elution with 2mL Methanol for HPLC



F-3 of 20



Institute NIMT

NMISA



Pre-treatment



Sample



Extraction Method



Size (units)



Clean-up



Liquid-liquid extraction using 70:30



MeCN: water with 20 mL of extraction



solvent: 10 grams sample 1. Weigh out 10 g of dried fig sample. An appropriate amount of each aflatoxin labeled solution and 1 g of NaCl were then added to the sample. 2. Add 20 mL of 70:30 MeOH:water to 10 g the mixture and mix vigorously for 60 mins 3. Centrifuge at 3000 rpm for 10 min, collect the supernatant and filter through 0.45 micron GMF.



Immunoaffinity column (IAC) 5. Add 5 mL of 1XPBS (pH 7.4) and pass the mixture thorugh IAC column. 6. Wash the IAC with 5 mL water and elute with 2 portions of 2.5 mL acetonitrile. 7. Evaporate the eluate under N2 stream at 45 oC to dryness. 8. Reconsitute with 200 mL acetonitrile and filter with 0.2 micron PVDF before injecting onto HPLC



4. Evaporate out organic consituent



under N2 stream at 45 oC for 30 mins.



Methanol: Water (80:20)saline solid-



liquid extraction of the dried fig powder



with shaking 60 min. Modified from the AOAC 999.07 method. In short, approximately 6-10 g of sample was weighed and extracted with 36 mL of extraction solvent and 10% NaCl (m/m sample). The samples were extracted for 1 hour by orbital shaking at approximately 200 rpm. A



6-10 g



Immunoaffinity clean-up (VICAM Aflatest) The full extract 72 mL was loaded onto a VICAM AFLAtest immunoaffinity clean up cartridge. Samples were eluted after washing, with 3 mL methanol. The eluate was dried down and resuspended in 300 µL of LC solvent.



12 mL aliquot of the extract was diluted



into 60 mL of PBS,



F-4 of 20



Institute VNIIM

TUBITAK UME



Pre-treatment

10 g of sample was put into a 100-ml Erlenmeyer flask. the internal standards (13C17aflatoxines B1. B2. G1. G2) and 40 mL of acetonitrile-water (84:16. v/v) were added.



Sample



Extraction Method



Size (units)



Clean-up



10 g of sample was put into a 100-ml



SPE.



Erlenmeyer flask. the internal standards



After sonication for 30 minutes. the



(13C17-aflatoxines B1. B2. G1. G2) and 40



supernatant was filtered through a glass



mL of acetonitrile-water (84:16. v/v) were



microfiber filter. Filtrate was purified



added. After sonication for 30 minutes. the



by passing through the MycoSep 228



supernatant was filtered through a glass



AflaPat cartridge at flow rate of 1 mL/min. The cleaned filtrate was



10 g



microfiber filter. Filtrate was purified by passing through the MycoSep 228 AflaPat



evaporated to dryness at 40 °C under a



cartridge at flow rate of 1 mL/min. The



gentle stream of nitrogen. The residue



cleaned filtrate was evaporated to dryness at



was reconstituted in 1 mL of methanol-



40 °C under a gentle stream of nitrogen. The



water (55:45 v/v). containing 10 mM



residue was reconstituted in 1 mL of



ammonium acetate.



methanol-water (55:45 v/v). containing 10



mM ammonium acetate.



Solid-liquid extraction



6 grams of sample were weighed into a



Immuno affinity cleanup



50 mL polypropylene centrifugation



Diluted extract was transferred to reservoir



tube, and 100 uL IS stock solution



on immuno affinity column (R-BIOPHARM



added and weighed. Then, 0.6 g sodium



EASI EXTRACT AFLATOXIN RP70N)



chloride and 36 mL of extraction



and passed with application of vacuum, after



solvent ( methanol:water 80:20 v/v)



extract was passed, column washed twice



added. Tube wrapped with aluminum



with 10 mL ultrapure water. Column dried



foil and vortex for 20 min at room



6g



for 5 seconds under vacuum. Aflatoxins



temperature with Heidolph Multi Reax.



eluted with 2 mL methanol to 4 mL amber



Then centrifuge at 10000 rpm at 15 °C



vial by gravity. Concentrated under nitrogen



for 20 minutes. Extract was filtered



stream till 0.5 mL remains. 1.5 mL ultrapure



through Macharey Nagel (product #



water added and vortex for 1 minute, if clear



405012) glass fiber filter paper and 25



transfer to LC vial otherwise filter through



mL of filtered extract is diluted with



0.2 μm syringe filter. Analyze with Thermo



150 mL PBS buffer (pH 7.4, Sigma



Scientific Q Exactive Orbitrap HR-LC/MS.



P4417).



F-5 of 20



Table F.2. Summary of Analytical Techniques for CCQM-K138



Institute



Analytical Technique



Chromatographic Column



Chromatographic and Mass Spectrometry Conditions



ion/MRM monitored



BAM



HPLC-SIDAMS/MS



Parameter Table



CUR: 15.00



Scan Type: MRM (MRM). Scheduled MRM: No



IS: 4000.00



Polarity: Positive . Scan Mode: N/A. Ion Source: Turbo



TEM: 550.00



Spray. Resolution Q1: Unit. Resolution Q3: Unit.



GS1: 70.00



AFG1 quant: 329.000 → 243.100. DP 79.00. CE 39.00. CXP



GS2: 50.00



14.00 1



ihe: ON



AFG1 qual: 329.000 → 311.100. DP 79.00. CE 31.00. CXP



CAD: 4.00



21.00



EP 10.00



13C AFG1: 346.100 → 257.200. DP 94.00. CE 40.00. CXP



Dwell(msec): 50.00



15.00



AFG2 quant: 331.000 → 285.100. DP 46.00. CE 38.00.



Mobile phase: water and methanol CXP 18.00



HPLC column:



(each 0.1% formic acid and 5mM AFG2 qual: 331.000 → 313.100. DP 46.00. CE 32.00. CXP



Agilent Zorbax



ammonium formate; HPLC gradient). 10.00



Eclipse XDB C18. column temperature: 30°C.



13C AFG2: 348.200 → 330.300. DP 94.00. CE 35.00. CXP



2.1x100mm. 1.8µm. analysis time: 14.5 min.



23.00



flow rate: 300 µL/min.



AFB1 quant: 313.000 → 285.200. DP 86.00. CE 33.0. CXP



injection volume: 5 µL



18.00



AFB1 qual: 313.000 → 241.200. DP 86.00. CE 52.00. CXP



13.00



13C AFB1: 330.200 → 301.200. DP 91.00. CE 35.00. CXP



19.00



AFB2 quant: 315.000 → 287.100. DP 66.00. CE 37.00.



CXP 18.00



AFB2 qual: 315.000 → 259.100. DP 66.00. CE 43.00. CXP



15.00



13C AFB2: 332.000 → 332.100. DP 91.00. CE 39.00. CXP



18.00



F-6 of 20



Institute



Analytical Technique



Chromatographic Column



Chromatographic and Mass Spectrometry Conditions



ion/MRM monitored



HESI - multiple reaction monitoring



EXHM



ID-LC-MS/MS



Waters XTerra MS 15 mm. 2.1 mm. 3 μm



Capillary Temp: 270, Vaporizer Temp: 350, Sheath Gas Pressure: 40.0, Ion Sweep Gas Pressure: 0.0, Aux Gas Flow: 10.0, Spray Voltage: + 4000.0

Mobile phase: Water (A) - MeOH (B) gradient: 0 min - 90A/10B. 4 min90A/10B. 12 min 30A/70B. 16 min 10A/90B. 20 min 10A/90B. 21 min 90A/10B. 25 min 90A/10B flow rate: 150 mL. injection vol. 20 mL



AfB1 (313.1 to 285Q. 241q). AflaB2 (315.1 to 243. 259Q. 287q). AflaG1 (329.1 to 200q. 215. 243Q). AflaG2 (331.1 to 201. 217. 245Q. 257q. 275. 313). 13C-AfB1 (330.1 to 255. 301). 13CAflaB2 (332 to 259. 303). 13CAflaG1 (346 to 212. 317). 13CAflaG2 (348 to 259. 313)



F-7 of 20



Institute



Analytical Technique



Chromatographic Chromatographic and Mass



Column



Spectrometry Conditions



Operation mode : ESI positive



ionization



Source temperature : 450 °C



Ion spary voltage : 5500 V



ion/MRM monitored



GLHK LC-MS/MS



ACQUITY UPLC C18 (2.1 x 100 mm. 1.7 µm)



Mobile phase A : 10mM ammonium formate. 0.1% formic acid. 5% MeOH in water Mobile phase B : methanol Gradient program : t = 0min. 95%A;



t



MRM scanning Operation mode : ESI positive Source temperature : 450 °C Ion spray voltage : 5500 V



ionization



= 1min. 60%A; t =7. 55%A; t = 7.5-



10.5min. 5%A; t = 11-15min. 95%A



Flow rate : 0.25 mL/min



Analysis time : 15 min



Injection volume : 20 µL



Column temperature 35 °C



INTI KEBS



Inmunaffinity columns, electrochemical derivatization (Kobra Cell) and LC with fluorescence detection



Reversed Phase Column ODS 4.6 mm x 15 cm, 5 um.



Mobile Phase: H20+MeOH (6+4) + 216.4 mg KBr/L +159.1 ul (HNO3 4N)/L. Fluorescence detector wavelengths 360 nm excitation filter and 420 nm emission filter. Column Temperature: 40°C. Column Pressure 61 bar. Analysis time 20 min. Flow rate 1 ml/min. Injection Volume 100 ul.



HPLC- FL DETECTION



C18 150 mm 5u column



MP- Water:Methanol:CAN (5:4:1). Detector FL Ex365 Em435. Temp 30. Flow rate 1ml/min. Analysis time 9 min. 10uL injection



F-8 of 20



Institute



Analytical Technique



NIMT



LC-MS/MS IDMS



Chromatographic Chromatographic and Mass



ion/MRM



Column



Spectrometry Conditions



monitored



Detection by MS/MS:Positive ESI with SRM mode



AFB1: 313.1>241. 313.1>268 labeled AFB1: 330.1>301.1. 330.1>255.1



Luna C18 4.6x150 mm 5 mm 100 Å



Chromatographic conditions: MP: MeCN:H2O with 20 mM Formic acid (42:58) Flow rate: 0.5 mL/min Injection vol: 20 mL Column temp: 40 °C



AFB2: 315.1>259.05. 315.1>287.1 labeled AFB2: 332.2>303.2. 332.2>273.1 AFG1: 329.1>242.95. 329.1>200.1 labeled AFG1: 346.1>257.1. 346.1>212.1 AFG2; 331.15>245. 331.15>275 labeled AFG2: 348.1>330.1. 348.1>259.1



NIMSA UPLC-MS/MS



VNIIM



LC-MS/MS IDMS



The cone voltage was set at 2 V with



a collision energy for the various



transitions range from 24 to 40 eV.

The capillary voltage was set at 2.5

kV and the desolvation temperature at 550oC. The cone gas flow was set



The MS/MS analysis was performed on a WatersTQS triple quadrupole instrument The quantifier transitions for each of the toxins was:



Acquity UPLC BEH to 150 L/h whilst the desolvation gas



C18 1.7 µm. 2.1 x flow was 800 L/h.



100 mm column



(40oC)



Mobile phase 5 mM ammonium



formate aqueous and methanol



solvents at a flow rate of 0.35



mL/min and the total runtime was 7.5 . min. The maximum pressure reached



during a run is approximately 11500



psi.



ESI(+)



Mobile phase: A - ammonium acetete ESI(+); MRM: aflatoxin B1 (313 → 241). aflatoxin B2 (315



Hydroshere C18 10mmol/L (45%).



→ 259). aflatoxin G1 (329 → 243). aflatoxin G2 (331 →



100mm x 4,6 mm, 3 B - methanol (55%); isocratic



245). 13C17-aflatoxin B1 (330 → 255). 13C17-aflatoxin B2



µm;



eluation; flow rate 0.8 ml/min;



(332 → 273). 13C17-aflatoxin G1 (346 → 257).13C17-



column temperature 30°C; injection aflatoxin G2 (348 → 259).



volume 5 µl



F-9 of 20



Institute



Analytical Technique



TUBITAK HR-LC-MS UME IDMS



Chromatographic Chromatographic and Mass



Column



Spectrometry Conditions



MS Resolution: 70000



HESI Positive



Capillary Temperature 280 °C



Aux gas heater temp. 250 °C



Sheath gas flow rate: 45



Aux gas flow rate: 10



Spray voltage (kV): 3.60



Scan range: 100 - 1000 m/z



B1



: 313.0700



mobile phase:



B1-13C17 : 330.1270



A: 95 % water 5 % MeOH 5 mM B2



: 315.0860



Ammonium Acetate 0.1 %



B2-13C17 : 332.1430



B: MeOH



G1



: 329.0650



column temperature: 40°C.



G1-13C17 : 346.1220



Autosampler 4 °C



G2



: 331.0810



injection volume: 10 µL



G2-13C17 : 348.1370



Ret(min) Flow (μL/min) % A % B



00



0.3



95 5



06



0.3



50 50



10



0.3



5 95



15



0.3



5 95



15.1



0.3



95 5



18



0.3



95 5



ion/MRM monitored



F-10 of 20



Table F.3. Summary of Calibrants and Standards for CCQM-K138



Institute



Type of Calibration



Calibrants Commercial standards (Biopure, RomerLabs), gravimetric sample preparation



Internal Standards



BAM



Aflatoxin B1 in Acetonitril



(2.01µg/mL +/- 0.03 µg/mL;



Biopure) B1: 16192B



Six point



Aflatoxin B2 in Acetonitril



internl standard



(0.502µg/mL +/- 0.008 µg/mL;



calibration (SIDA),



Biopure) B2: L15483A



IDMS, linear regression



Aflatoxin G1 in Acetonitril



(2.01µg/mL +/- 0.03 µg/mL;



Biopure) G1: L15331C



Alfa -Mix 13C17-B1: I15383M 13C17-B2: I15383M 13C17-G1: I15383M 13C17-G2: I15383M



Aflatoxin G2 in Acetonitril (0.500µg/mL +/- 0.008 µg/mL; Biopure) G2: L15391A



EXHM



Exact matching matrix matched standards

ID-LC-MS/MS



IRMM ERM–AC057 IRMM ERM – AC058 IRMM ERM – AC059 IRMM ERM – AC060



C13 labelled aflatoxin solutions were purchased from LGC (B1) and Romer Labs (B2. G1. G2)



F-11 of 20



Institute GLHK



Type of Calibration



Calibrants



3 - 5 calibration points



Aflatoxin B1 : IRMM ERM – AC057



Quantification mode used - Isotope Dilution Mass Spectrometry Calibration mode used –



Aflatoxin B2 : IRMM ERM – AC058 Aflatoxin G1 : IRMM ERM – AC059 Aflatoxin G2 : IRMM ERM –



Standard addition



AC060



Internal Standards [13C17] Aflatoxin B1 from LGC [13C17] Aflatoxin B2 from LGC [13C17] Aflatoxin G1 from LGC [13C17] Aflatoxin G2 from LGC



INTI



Fluka AG, Aflatoxins B1, B2, G1 and G2 from Aspergillus Flavus



External standard, 5



Aflatoxin B1 Cat. Code: 5032



points calibration curve Batch 2216541280. Aflatoxin



Spectrophotometric



B2 Cat. Code: 5033 Batch



-



method (AOAC 971.22) 202621578. Aflatoxin G1 Cat.



Code:5035 Batch 219939181.



Aflatoxin G2 Cat. Code: 5036 Batch 219940181.



KEBS



External Calibration



AFB1 Source FERMENTEK Lot# AF017 AFB2 Source TRILOGY ANALYTICAL LABORATORY Lot# 141104- 070 AFG1 Source TRILOGY ANALYTICAL LABORATORY Lot# 150305070 AFG2 Source TRILOGY



F-12 of 20



Institute



Type of Calibration



Calibrants ANALYTICAL LABORATORY Lot# 150309070



Internal Standards



NIMT NIMSA VNIIM



6-pt calibration with labeled internal standard, IDMS



IRMM ERM–AC057 IRMM ERM – AC058 IRMM ERM – AC059 IRMM ERM – AC060



labeled AFB1 labeled AFB2 labeled AFG1 labeled AFG2



Double IDMS, standard addition. External



IRMM ERM–AC057, ILM010 Lot I15231A



BIOPURE 13C Afla B1



standard 9-point std addition; 6-



IRMM ERM – AC058, ILM011



Lot I15345B



BIOPURE 13C Afla B2



point external



IRMM ERM – AC059, ILM012 Lot I15345A



BIOPURE 13C Afla G1



calibration; 3 brackets dIDMS.



IRMM ERM – AC060, ILM013 BIOPURE 13C Afla G2 Lot I15232G



Single point, IDMS



Commercial standards

Aflatoxin B1 in acetonitrile. Biopure

Aflatoxin B2 in acetonitrile. Biopure

Aflatoxin G1 in acetonitrile. Biopure



13C17-aflatoxin B1 solution in acetonitrile (cat. № ILM010). 13C17-aflatoxin B2 solution in acetonitrile (cat. № ILM011). 13C17-aflatoxin G1 solution in acetonitrile (cat. № ILM012) 13C17-aflatoxin G2 solution in acetonitrile (cat. № ILM013) were obtained from Biopure.



Aflatoxin G2 in acetonitrile.



F-13 of 20



Institute



Type of Calibration



Calibrants



Biopure



Internal Standards



TUBITAK UME



Five point

internal standard calibration, IDMS



Commercial standards purity



determined by QNMR traceable AFB1-13C17 SIGMA 32764l to UME CRM 1301, gravimetric



sample preparation



AFB2-13C17 SIGMA 32771



AFB1 SIGMA A6636 AFB2 SIGMA A9887 AFG1 SIGMA A0138 AFG2 SIGMA A0263



AFG1-13C17 SIGMA 32772 AFG2-13C17 SIGMA 32777



F-14 of 20



Table F.4. Assessment and Verification Methods for CCQM-K138



Institute



Purity Assessment



The purity of the used certified calibration standards on three



ways:



Result Verification



BAM



- Purities of calibration standards were independently confirmed by LC-MS measurements (scan mode; ESI+/-).



The specified aflatoxin contents of the used certified



-



standard solutions were cross checked by certified standards



of different lot numbers (same provider)



Additional cross check using certified standard solutions of a second provider.



EXHM



The solid aflatoxins used by EXHM have been characterized



for their purity using the mass balance approach and qNMR.



The concentration of the solutions prepared has been



assigned against the IRMM CRM solutions (ERM AC 057,



058, 059, 060) using IDMS experiments, and this is the



reason why we attribute traceability to IRMM.



-



The actual values were:



AFB1=96.13±3.18%, AFB2=93.32±3.13%, AFG1=98.60±3.35%, AFG2=94.02±3.12%.



GLHK



-



-



F-15 of 20



Institute INTI



Purity Assessment



Result Verification



One solution of each aflatoxin was prepared to obtain 4



stock solutions of 8-10 ug/ml in acetonitrile. These solutions



were verified using an Spectrophotometric method (AOAC



971.22). After the measurement of the stock solution at



350nm, it was adjusted the purity of each calibration solution. The assignment of purity was determinated following the



We do not use any verification. Verification is not necessary due specificity of cleanup separation used in the method.



next equation:



% purity = ccstandard stock x 10ml x 5000ul x100 50 ul x 1000 ug/mg x 10 mg



KEBS



-



-



NIMT



-



-



NIMSA -



-



VNIIM -



-



F-16 of 20



Institute



Purity Assessment



TUBITAK Purity of commercially available highly-pure substances



were determined by in-house qNMR purity assignment



-



UME



traceable to UME CRM 1301



Result Verification



F-17 of 20



Table F.5. Additional Comments for CCQM-K138



Institute BAM



Additional Comments Remarks: LOD/LOQ (as mass fraction): 0.083 µg/kg / 0.328 µg/kg for AFB1, 0.033 µg/kg / 0.128 µg/kg for AFB2, 0.136 µg/kg / 0.539 µg/kg for AFG1, 0.016 µg/kg / 0.064 µg/kg for AFG2



EXHM



Information of quality control sample: None

Remarks: -C13 labelled aflatoxin solutions were purchased from LGC (B1) and Romer Labs (B2, G1,G2) -The product ions 259 and 245 were used to quantify Afla B2 and Afla G2 respectively, due to pronunced matrix interference for the more abundant ions. -IA columns: Aflastar R (Rohmer Labs)



LOD/LOQ (as mass fraction): 5/15 ng/g



Information of quality control sample: FAPAS T04280QC



Remarks : Concentration of calibrants in standard addition does not include the concentration of AFs from sample



GLHK



LOD/LOQ (as mass fraction): LOQ of analyte calibrated by standard addition is regarded as the sample concentration in mass fraction



Information of quality control sample: IRMM ERM – BE375 Compound Feedingstuff



F-18 of 20



INTI



Remarks: Preparation of standards: 1) Preparation of Calibrant: To container of 10 mg of each dry aflatoxin was added a volume of 5 ml of Toluene : Acetonitrile (9+1). Final concentration aprox. 2 mg/ml. 2) Preparation of stock solution: From calibrant solution (50ul) were prepared individuals stock solution in acetonitrile of each aflatoxin. Final concentration (10 ml) aprox. 8-10 ug/ml. 3) Working solution: The working solutions were prepared mixed the four toxins from stock solutions. The solutions were prepared in four levels: 0.075 ng/ml B1, B2, G1 and G2 0.375 ng/ml B1, B2, G1 and G2- 1.25 ng/ml B1, B2, G1and G2 - 2.5 ng/ml B1, B2, G1 and G2. The accuracy of the method was determinated making a recovery test in-house material. The

results obtained were the following: AfB1: 112%, AfB2 88% AfG1 110%, AfG2 93%. The

repetibility was determinated analyzing each sample four times in the same day as individual

replicates.



LOD/LOQ (as mass fraction): 0.1 ng/g / 0.3 ng/g



Information of quality control sample: Recovery test using in-house material (Recovery values 88%-112%)

Remarks: N/A



LOD/LOQ (as mass fraction): KEBS N/A



Information of quality control sample: CRM-ERMBE375 Remarks:-



NIMT



LOD/LOQ (as mass fraction): 0.3/0.8 ng/g for B1, 0.06/0.15 ng/g for B2, 0.14/0.4 ng/g for G1, 0.03/0.1 ng/g for G2

Information of quality control sample: Spiked blank



F-19 of 20



NIMSA VNIIM



Remarks: -The solvent proportions were maintained when extracting increased masses of sample. -Recovery on QC >90% -Matrix enhancement effects were observed (compensated for by the isotope) and there was limited stability of the low concentration calibrant solutions for G1 and G2. The homogeneity of the sample appears to be a significant contributor to the variability as multiple aliquots from a single extract yielded very similar results, suggesting that the large variability between repeat analyses is not as a result of the clean-up and analytical method. Initial tests were run using HPLC-FLD which confirmed data obtained using LC- MS/MS. FAPAS fig slurry was used as QC, recoveries >90% achieved.

LOD/LOQ (as mass fraction):

0.14/ 0.46 ng/g for AB1 0.027/ 0.090 ng/g for AB2 0.074/ 0.25 ng/g for AG10.022/ 0.074 ng/g for AG2

Information of quality control sample: FAPAS T04258 Fig Slurry 1.72 µg/kg (0.96 - 2.48) for AFB1 FAPAS T04258 Fig Slurry 1.30 µg/kg (0.73 - 1.87) for AFB2 FAPAS T04258 Fig Slurry 0.94 µg/kg (0.52 - 1.35) for AFG1 FAPAS T04258 Fig Slurry 0.88 µg/kg (0.49 - 1.27) for AFG2

Remarks: Internal standards: 13C17-aflatoxin B1 solution in acetonitrile (cat. № ILM010), 13C17-aflatoxin B2 solution in acetonitrile (cat. № ILM011), 13C17-aflatoxin G1 solution in acetonitrile (cat. № ILM012) and 13C17-aflatoxin G2 solution in acetonitrile (cat. № ILM013) were obtained from Biopure.

LOD/LOQ (as mass fraction): N/A Information of quality control sample: Sample of dried fig with addition of AFB1, AFB2, AFG1, AFG2



Remarks: -



TUBITAK UME



LOD/LOQ (as mass fraction): 0.029 µg/kg / 0.096 µg/kg for AB1 , 0.003 µg/kg / 0.009 µg/kg AB2 0.008 µg/kg / 0.023µg/kg for AG1 , 0.001 µg/kg / 0.002 µg/kg AG2



Information of quality control sample: None



F-20 of 20



APPENDIX G: Summary of Participants’ Uncertainty Estimation Approaches

The following are text excerpts and/or pictures of the uncertainty-related information provided by the participants in the reporting form. Information is grouped by participant and presented in alphabetized acronym order.

Uncertainty Information from BAM

w\_sample= ((r\_area - i\_cal)/sl\_cal) ∙ m\_is/m\_sample w\_sample: mass fraction of aflatoxin in sample r\_area: area ratio native compound/internal standard i\_cal: intercept of calibration line sl\_cal: slope of calibration line m\_is: mass of internal standard added to sample m\_sample: mass sample

Uncertainty estimation was performed. using the following equation:

U\_(95%)=k ∙ √((s/m)^2+ (u(c\_cal )/c\_cal )^2+ (u\_(x\_pred )/x\_pred )^2 )

U\_(95%): expanded uncertainty 95% confidence k: coverage factor m: mean s: standard deviation of the mean u\_(c\_cal ): uncertainty of the standard substances u\_(x\_pred ): uncertainty of the calibration

where u\_(x\_pred) was calculated according to EURACHEM CITAC Guide:

var(x\_pred )= S^2/(b\_1^2 ) ∙ (1/p+ 1/n + (x\_pred- x ̅ )^2/((∑(x\_i^2 )-(∑x\_i )^2/n))) ; S^2= (∑w\_i (y\_i-y\_fi)^2))/((n-2))

(y\_i-y\_fi): residual for the ith point n: number of data points in the calibration b\_1 : calculated best fit gradient p: number of measurements x\_i. y\_i: data points x\_pred: estimated concentration x ̅: mean

Estimation of standard measurement uncertainties:

G-1 of 12



method precision: standard deviation of the mean (n = 6) standard substances: based on given uncertainties of the standards calibration: uncertainties from linear least squares calibration according to EURACHEM CITAC Guide Uncertainty estimation for u\_c;sumAfla was performed. using the following equation: u\_c;sumAfla=√((u\_c;AFB1)^2 + (u\_c;AFB2)^2 + (u\_c;AFG1)^2 + (u\_c;AFG2)^2) u\_c;AF: combined uncertainty of the respective aflatoxin

G-2 of 12



Uncertainty Information from EXHM/GCSL-EIM The measurement equation is:



where wM.S = aflatoxin mass fraction in the sample. (μg/kg)



wM.C = aflatoxin mass fraction in the calibration solution. (μg/kg)



F



= sample fraction in slurry (g/g)



Rec = recovery (%). assessed against other independent methods



mis.S = mass of internal standard solution added to sample blend. (g)



mM.S = mass of slurry in sample blend. (g)



mM.C = mass of the calibration solution added to calibration blend. (g)



mis.C = mass of internal standard solution added to calibration blend. (g)



RS = measured peak area ratio of the selected ions in the sample blend



RC = measured peak area ratio of the selected ions in the calibration blend



The equation used to estimate standard uncertainty is:



where sR is the standard deviation under reproducibility conditions. n the number of determinations and Cj the sensitivity coefficients associated with each uncertainty component. The uncertainty of the peak area ratios was considered to have been included in the estimation of method precision.

Uncertainty estimation was carried out according to JCGM 100: 2008. The standard uncertainties were combined as the sum of the squares of the product of the sensitivity coefficient (obtained by partial differentiation of the measurement equation) and standard uncertainty to give the square of the combined uncertainty. The square root of this value was multiplied by a coverage factor (95% confidence interval) from the t-distribution at the total effective degrees of freedom obtained from the Welch-Satterthwaite equation to give the expanded uncertainty.

The uncertainty budgets for the four aflatoxins are shown in the pages that follow.



G-3 of 12



Aflatoxin B1 Aflatoxin B2

G-4 of 12



Aflatoxin G1 Aflatoxin G2

G-5 of 12



Uncertainty Information from GLHK The mass fraction of aflatoxins (AFB1.AFB2. AFG1. AFG2) were quantified by isotope dilution mass spectrometry (IDMS). Standard addition was employed for calibration in this comparison. Measurement equations used : 1) Linear equation for standard addition

where i - ith solution of standard addition Ri - peak area of analyte/ peak area of IS of the ith solution of standard addition mx.i - mass of sample x in the ith solution of standard addition my.i - mass of IS solution y added to the ith solution of standard addition mz.i - mass of added standard z in the ith solution of standard addition a0 - y-intercept of the linear fit function of standard addition calibration curve a1 - slope of the linear fit function of standard addition calibration curve 2) Equation for mass fraction calculation

where wx - mass fraction of the analyte in sample x wz - mass fraction of the analyte in standard z

Individual uncertainty contributions 1) Uncertainty in mass fraction. wx. calibrated by standard addition

where

G-6 of 12



2) Weighing - mass of sample. mass of standard added and mass of IS added



- estimated by combining uncertainty in weighing. including uncertainty from analytical



balances



3) Precision



- estimated by the variation in response factors from repeated measurement of samples



4) Recovery



- estimated by recovery of blank sample spikes calibrated by standard addition method



5) Uncertainty of mean value



- estimated from the deviation of mass fractions from 2 different sample units



Combining individiual uncertainties



Take AFB1 as an example. combining individual uncertainties for each of the sample unit:



Sample 145



C B1 (ng/g)



Weighing



Recovery



Precision



Sample 340



C B1 (ng/g)



Weighing



Recovery



Precision



Value (xi)



5.6345



1



1



1



Value (xi)



5.8681



1



1



1



u(x i )



2.1958E-01 3.2150E-04 2.8355E-03 6.5841E-02



u(x i )



2.8325E-01 3.1265E-04



2.8355E-03



7.8778E-02



C B1 (ng/g) Weighing Recovery Precision

C B1 (ng/g) u(y,x i ) u(y,x i ) 2

u(C B1 ) , (ng/g)



5.6345 1 1 1

5.6345

0.1861

0.4314



5.8541 1.0000 1.0000 1.0000

5.8541 2.1958E-01 4.8213E-02



5.6345 1.0003 1.0000 1.0000

5.6364 1.8115E-03 3.2816E-06



5.6345 1.0000 1.0028 1.0000

5.6505 1.5977E-02 2.5526E-04



5.6345 1.0000 1.0000 1.0658

6.0055 3.7099E-01 1.3763E-01



C B1 (ng/g) Weighing Recovery Precision

C B1 (ng/g) u(y,x i ) u(y,x i ) 2

u(C B1 ) , (ng/g)



5.8681 1 1 1

5.8681

0.2942

0.5424



6.1514 1.0000 1.0000 1.0000

6.1514 2.8325E-01 8.0233E-02



5.8681 1.0003 1.0000 1.0000

5.8700 1.8347E-03 3.3660E-06



5.8681 1.0000 1.0028 1.0000

5.8848 1.6639E-02 2.7686E-04



5.8681 1.0000 1.0000 1.0788

6.3304 4.6228E-01 2.1370E-01



Include also the standard uncertainty of the mean value and calculate the overall expanded uncertainty. k = 2:



AFB1 Mean Value

Mean AFB1 = u(between-bottle

deviation) = u (AFB1) = U (AFB1) =



5.7513 ng/g 0.1168 ng/g 0.5443 ng/g 1.0886 ng/g



G-7 of 12



Uncertainty Information from INTI

Measurement Equation of each analyte: CCinter = Interception of calibration curve from repetibility of each sample V1 = 5 ml with volumetric flask V2 = 20 ml with Volumetric pipette V3 = 30 ml with Graduated cylinder V4 = 9 ml con Adjustable pipette V5 = 100 ml with Graduated Cylinder V6 = 30 ml + 9 ml (Graduated Cylinder + Adjustable pipette) m = mass with laboratory balance

Measurement equation to calculate uncertainty: ΔCCinter afx= Measurement uncertainty ΔCCinter = Interception of calibration curve from repetibility of each sample - 0.0751ng/ml; 0.0284ng/ml; 0.0160ng/ml; 0.0094ng/ml (Afx B1; Afx B2; Afx G1; Afx G2) ΔV1 = 0.014 ml with volumetric flask (internal calibration) ΔV2 = 0.0176 ml with Volumetric pipette (internal calibration) ΔV3 = 0.151ml with Graduated cylinder (internal calibration) ΔV4 = 0.005 ml con Adjustable pipette (internal calibration) ΔV5 = 0.0939 ml with Graduated Cylinder (internal calibration) ΔV6 = 0.1515 ml (Graduated Cylinder + Adjustable pipette) (internal calibration) Δm = 0.01 g mass with laboratory balance (internal calibration)

G-8 of 12



Uncertainty Information from NIMT

wx= Mass fraction of aflatoxin (ng/g) in test sample w0 = Mass fraction ratio (between unlabeled/labeled) obtained from the calibration curve (ng/ng) wy(x) = Mass fraction of aflatoxin internal standard added to the sample. ng/g

my(x) = Mass of internal standard spiked into the sample (g)

mx= Mass of sample (g) R = Recovery factor



u(my). u(mx) = standard uncertainties due to weighing estimated from bias of balance u(w0)= standard uncertainty of the mass fraction ratio (between unlabeled/labeled) obtained from the calibration curve (ng/ng) estimated from the regression u(Fcal) = standard uncertainty of mid concentration calibration standard estimated from bias and random effects (type B and type A) u(FE) = standard uncertainty of extraction

u(FP) = standard uncertainty of method precision

u(R) = standard uncertainty of recovery



u(wx )  wx



 



u(my my



)



2 











u(mx mx



)



2











u(w0 w0



)



2











u(Fcal Fcal



)



2











u(FE FE



)



2

  











u(FP FP



)



2

  







 



u(R) R



2 



G-9 of 12



Uncertainty Information from NMISA



Measurement equation for determining (   ), the mass fraction of Aflatoxin in fig test portion:



Where:



   =



             +                 +          3



     (    )



            

                             ′     ′    



             =



     ×



           



×



         



×



  ′     ′    



Mass fraction of Aflatoxin in fig test portion (ng/g) obtained using bracketing double-isotope dilution



Mass fraction of aflatoxin in calibration CRM (g)



Mass of CRM added to calibration blend (g)



Mass of isotope added to calibration blend (g)



Mass of isotope added to sample blend (g)



Mass of test sample (g)



Peak area ration of analyte/isotope in sample blend



Peak area ration of analyte/isotope in calibration blend



Similarly: Where: Derived from:



                =                      ×    ×                 =                      ×    ×                                =       =      +   



Where through linear regression of the calibration data:



               

                             

                     



Mass fraction of Aflatoxin in fig test portion (ng/g) obtained using standard addition Mass fraction of Aflatoxin in fig test portion (ng/g) obtained using external calibration

Mass fraction of aflatoxin in calibration CRM (g)

Peak area of analyte

Slope

Mass fraction (ng/g) of calibration CRM added

y-intercept

Dilution factor

Recovery factor applied, determined from isotope recovery standard



G-10 of 12



Uncertainty Information from TUBITAK UME

w\_sample= ((r\_area - i\_cal)/sl\_cal) ∙ m\_is/m\_sample w\_sample: mass fraction of aflatoxin in sample r\_area: area ratio native compound/internal standard i\_cal: intercept of calibration line sl\_cal: slope of calibration line m\_is: mass of internal standard added to sample m\_sample: mass sample

Uncertainty estimation was performed, using the following equation calculated according to EURACHEM CITAC Guide “Quantifying Uncertainty in Analytical Chemistry”:



U k



u  u 2 precision



2 re co v ery



                     =    2  +    2 



     =             = (       −        )/          =        /(   − 1)        =        /(   −   ) SSb and SSw are obtained from one way ANOVA (p -1) and (N – p) are degrees of freedom obtained from one way ANOVA



G-11 of 12



Uncertainty Information from VNIIM



w- mass fraction of analyte in the sample, ng/g;

mis - mass of internal standard added to sample before sample preparation, ng; m - mass of the sample, g;

F - response factor.

F=(Sancal*mis)/(Siscal*man) man- mass of analyte in calibration solution; mis - mass of internal standard in calibration solution; Sancal - peak area for the analyte; Siscal - peak area for the internal standard



wан



=



Sан  mIS SIS  F  m



Source of uncertainty mass of sample (m) preparation of calibration solution concentration of reference standard solutions

RSD of F determination

mass of internal standard added to sample before extraction (mIS) (volume of IS solution added to sample) RSD of results. %

comb.std uncertainty expanded uncertainty (k=2)



AF B1 0.0006

1.29 0.87 2.65

0.58

2.15 3.8 7.6



u, % AF B2 0.0006 3.14

0.87 2.69

0.9

5.9 7.3 15



AF G1 0.0006

1.85 0.87 1.45

1.45

4.66 5.5 11



AF G2 0.0006 3.27 0.87 9.85

0.96

23 25 50



G-12 of 12



APPENDIX H: Participants’ Quantitative Results as Reported

The following are text excerpts and/or pictures of the quantitative results as provided by the participants in the reporting form. Information is grouped by participant and presented in alphabetized acronym order.

Quantitative Results from BAM



Measurand

AFB1 AFB2 AFG1 AFG2 Total AF



Mass Fraction (ng/g)

5.41 0.66 2.01 0.22 8.29



Combined Standard Uncertainty

(ng/g)

0.15 0.03 0.11 0.01 0.19



Coverage Factor (k)

2.571 2.571 2.571 2.571 2.571



Expanded Uncertainty

(ng/g)

0.40 0.08 0.27 0.03 0.49



Quantitative Results from EXHM/GCSL-EIM



Measurand

AFB1 AFB2 AFG1 AFG2 Total AF



Mass Fraction (ng/g)

5.994 0.871 2.093 0.264 9.223



Combined Standard Uncertainty

(ng/g)

0.123 0.022 0.061 0.01 0.141



Coverage Factor (k)

2.03 2.11 2.07 2.20 2.00



Expanded Uncertainty

(ng/g)

0.249 0.047 0.125 0.022 0.282



Quantitative Results from GLHK



H-1 of 4



Measurand

AFB1 AFB2 AFG1 AFG2 Total AF

Measurand

AFB1 AFB2 AFG1 AFG2 Total AF

Measurand

AFB1 AFB2 AFG1 AFG2



Mass Fraction (ng/g)

5.8 0.74 2.0 0.22 8.7



Combined Standard Uncertainty

(ng/g)

0.5 0.07 0.2 0.04 0.6



Coverage Factor (k)

2 2 2 2 2



Expanded Uncertainty

(ng/g)

1.1 0.14 0.5 0.07 1.2



Quantitative Results from INTI



Mass Fraction (ng/g)

5.17 0.69 2.5 0.32 8.68



Combined Standard Uncertainty

(ng/g)

0.33 0.13 0.07 0.04 0.57



Coverage Factor (k)

2 2 2 2 2



Expanded Uncertainty

(ng/g)

0.66 0.26 0.14 0.08 1.14



Quantitative Results from KEBS



Mass Fraction (ng/g)

7.27 0.60 2.39 0.06



Combined Standard Uncertainty

(ng/g)

0.8 0.1

0.4 0.01



Coverage Factor (k)

2 2 2 2



Expanded Uncertainty

(ng/g)

1.6 0.2 0.8 0.02



H-2 of 4



Total AF

Measurand

AFB1 AFB2 AFG1 AFG2 Total AF

Measurand

AFB1 AFB2 AFG1 AFG2 Total AF



10.31



1.34



2



2.68



Quantitative Results from NIMT



Mass Fraction (ng/g)

6.6 0.8 2.6 0.3 10.3



Combined Standard Uncertainty

(ng/g)

0.40 0.05 0.18 0.03 0.44



Coverage Factor (k)

2.03 2.04 2.10 2.00 2.57



Expanded Uncertainty

(ng/g)

0.9 0.1 0.4 0.1 1.2



Quantitative Results from NMISA



Mass Fraction (ng/g)

6.20 0.755 2.24 0.214 9.4



Combined Standard Uncertainty

(ng/g)

0.28 0.040 0.20 0.025 0.65



Coverage Factor (k)

2 2 2 2 2



Expanded Uncertainty

(ng/g)

0.56 0.080 0.40 0.049 1.3



H-3 of 4



Measurand

AFB1 AFB2 AFG1 AFG2 Total AF

Measurand

AFB1 AFB2 AFG1 AFG2 Total AF



Quantitative Results from TUBITAK UME



Mass Fraction (ng/g)

5.72 0.67 2.16 0.23 8.78



Combined Standard Uncertainty

(ng/g)

0.33 0.05 0.15 0.02 0.35



Coverage Factor (k)

2 2 2 2 2



Expanded Uncertainty

(ng/g)

0.66 0.09 0.30 0.04 0.70



Quantitative Results from VNIIM



Mass Fraction (ng/g)

6.22 0.81 1.98 0.15 9.16



Combined Standard Uncertainty

(ng/g)

0.23 0.06 0.11 0.04 0.27



Coverage Factor (k)

2 2 2 2 2



Expanded Uncertainty

(ng/g)

0.46 0.12 0.22 0.08 0.54



H-4 of 4