2012-05-04 Evaluation Immunoassay Kits Aflatoxin PDF

CODA-CERVA
Evaluation of immunoassay kits for quantitative determination of ochratoxin in cereals

Veterinary and Agrochemical Research Centre
_________________________________________________________________________________________________________________
________________________________________
Evaluation of immunoassay
kits for aflatoxin
determination in corn & rice
AFSCA-FAVV
Federal
Error! Reference source Agency
not found. for the Safety of the Food Chain
1-1
CODA-CERVA
VETERINARY AND AGROCHEMICAL RESEARCH CENTRE
CHEMICAL SAFETY OF THE FOOD CHAIN
UNIT OF TOXINS AND NATURAL COMPONENTS
NATIONAL REFERENCE LABORATORY ON MYCOTOXINS
Evaluation of immunoassay
kits for aflatoxin
determination in corn & rice
December 2011
Bart Huybrechts
Correspondance to: Dr. Ir. Bart Huybrechts

Leuvensesteenweg 17
B-3080 Tervuren
Belgium
Tel: +32 02 769 22 76

Fax: + 32 02 769 23 05
[email protected]
We would like to thank the kit manufacturers;
Charm, Neogen, Tecna, R-Biopharm and
Romerlabs for providing us with all the materials,
tools, customer support and background
information necessary to bring this assessment to a
successful end.
Table of Contents
1. Foreword & scope & background ............................................................................... 1-1

1.1 Foreword ................................................................................................................................................................................ 1-1
1.2 Scope....................................................................................................................................................................................... 1-2
1.2.1 Objective and strategy of the inventory task .............................................................................................................................................. 1-2
1.2.2 Objective and strategy of the administrative evaluation............................................................................................................................ 1-2
1.2.3 Objective and strategy of the experimental evaluation.............................................................................................................................. 1-4
1.3 Background ............................................................................................................................................................................ 1-5
2. Inventory .................................................................................................................... 2-11
2.1 Inventory results of providers of aflatoxin kits ............................................................................................................. 2-13
3. Administrative evaluation .......................................................................................... 3-17
3.1 Analytical performance criteria ........................................................................................................................................ 3-19
3.2 Intra-run precision, between-run precision & reproducibility.................................................................................... 3-19
3.3 Accuracy & Recovery......................................................................................................................................................... 3-21
3.4 Cross-reactivity.................................................................................................................................................................... 3-23
3.5 Matrix applicability ............................................................................................................................................................. 3-23
3.6 Conclusions of the administrative evaluation ................................................................................................................ 3-25
4. Experimental evaluation ........................................................................................... 4-27
4.1 Materials and methods ....................................................................................................................................................... 4-29
4.1.1 Equipment and materials ............................................................................................................................................................................ 4-29
4.1.1.1 Kits used ................................................................................................................................................................................................. 4-29
4.1.1.2 Standards and reagents ............................................................................................................................................................................ 4-29
4.1.1.3 Read-out ................................................................................................................................................................................................. 4-30
4.1.1.4 Blank matrix.......................................................................................................................................................................................... 4-30
4.1.1.5 Accuracy ................................................................................................................................................................................................. 4-30
4.2 Work protocols ................................................................................................................................................................... 4-30
4.2.1 Cross-reactivity............................................................................................................................................................................................. 4-30
4.2.1.1 Study design ............................................................................................................................................................................................ 4-30
4.2.1.2 Data handling and statistical analyses ...................................................................................................................................................... 4-31
4.2.1.2.1 ELISA kits..................................................................................................................................................................................... 4-31
4.2.1.2.2 LFD ................................................................................................................................................................................................ 4-31
4.2.2 Accuracy ....................................................................................................................................................................................................... 4-31
4.2.2.1 Study design ............................................................................................................................................................................................ 4-31
4.2.3 Precision ....................................................................................................................................................................................................... 4-32
4.2.3.1 Study design ............................................................................................................................................................................................ 4-32
4.2.4 Detection capability (CCβ) ......................................................................................................................................................................... 4-32
4.2.4.1 Study design ............................................................................................................................................................................................ 4-32
4.2.5 Recovery ....................................................................................................................................................................................................... 4-33
4.2.5.1 Study design ............................................................................................................................................................................................ 4-33
4.2.6 Kit protocol .................................................................................................................................................................................................. 4-34
4.3 Results................................................................................................................................................................................... 4-35
4.3.1 Accuracy ....................................................................................................................................................................................................... 4-35
4.3.2 Precision ....................................................................................................................................................................................................... 4-37
4.3.3 Recovery ....................................................................................................................................................................................................... 4-40
4.3.4 Detection capability (CCβ) ......................................................................................................................................................................... 4-43
4.3.5 Cross-reactivity............................................................................................................................................................................................. 4-48
4.4 Conclusions of the experimental evaluation .................................................................................................................. 4-51
5. Conclusion and future prospects............................................................................... 5-53
List of Figures
Figure 1: Partnership ................................................................................................................................................................................................................. 1-1

Figure 2: Structure of AFB1, AFB2, AFG1 & AFG2 ........................................................................................................................................................... 1-6
Figure 3: Distribution of kits in function of: type test, B1 vs Total & Qual vs Quan ..................................................................................................... 2-13
Figure 4: Principle of recovery ............................................................................................................................................................................................... 4-34
Figure 5: Principle of accuracy representation ..................................................................................................................................................................... 4-35
Figure 6: Accuracy results for corn (T04138), results in ng/g............................................................................................................................................ 4-36
Figure 7: Accuracy results for rice (T04151), results in ng/g ............................................................................................................................................. 4-36
Figure 8: Precision results for Charm, results in ng/g ........................................................................................................................................................ 4-38
Figure 9: Precision results for Neogen, results in ng/g....................................................................................................................................................... 4-38
Figure 10: Precision results for R-Biopharm, results in ng/g ............................................................................................................................................. 4-39
Figure 11: Precision results for Romerlabs, results in ng/g................................................................................................................................................ 4-39
Figure 12: Precision results for Tecna, results in ng/g........................................................................................................................................................ 4-40
Figure 13: Intercept for corn .................................................................................................................................................................................................. 4-41
Figure 14: Intercept for rice ................................................................................................................................................................................................... 4-41
Figure 15: Slope for corn ........................................................................................................................................................................................................ 4-42
Figure 16: Slope for rice.......................................................................................................................................................................................................... 4-42
Figure 17: CCβ for Charm in corn ........................................................................................................................................................................................ 4-43
Figure 18: CCβ for Charm in rice .......................................................................................................................................................................................... 4-44
Figure 19: CCβ for R-Biopharm in corn ............................................................................................................................................................................... 4-44
Figure 20: CCβ for R-Biopharm in rice ................................................................................................................................................................................ 4-45
Figure 21: CCβ for Neogen in corn ...................................................................................................................................................................................... 4-45
Figure 22: CCβ for Neogen in rice ........................................................................................................................................................................................ 4-46
Figure 23: CCβ for Romerlabs in corn.................................................................................................................................................................................. 4-46
Figure 24: CCβ for Romerlabs in rice ................................................................................................................................................................................... 4-47
Figure 25: CCβ for Tecna in corn.......................................................................................................................................................................................... 4-47
Figure 26: Cross-reactivity for Charm ................................................................................................................................................................................... 4-48
Figure 27: Cross-reactivity for R-Biopharm ......................................................................................................................................................................... 4-49
Figure 28: Cross-reactivity for Neogen ................................................................................................................................................................................. 4-49
Figure 29: Cross-reactivity for Romerlabs ............................................................................................................................................................................ 4-50
Figure 30: Cross-reactivity for Tecna .................................................................................................................................................................................... 4-50
List of Tables
Table 1: Parameters for quantitative and qualitative kits ...................................................................................................................................................... 1-2

Table 2: Regulatory limits EU (ref. 3-4-6) .............................................................................................................................................................................. 1-4
Table 3: Criteria to be verified according to 657/20028........................................................................................................................................................ 1-5
Table 4: Advantages and disadvantages of traditional and emerging methods for mycotoxin analysis .......................................................................... 1-7
Table 5: Overview of contacted manufacturers and their kits ........................................................................................................................................... 2-14
Table 6: Overview of kits involved in the experimental evaluation................................................................................................................................... 2-15
Table 7: Calculation of parameters (for symbols, see list abbreviations) .......................................................................................................................... 3-19
Table 8: Intra-run precision.................................................................................................................................................................................................... 3-20
Table 9: Between-run precision ............................................................................................................................................................................................. 3-20
Table 10: Reproducibility ........................................................................................................................................................................................................ 3-21
Table 11: Reported accuracy .................................................................................................................................................................................................. 3-22
Table 12: Reported average recovery for reference matrix ................................................................................................................................................. 3-22
Table 13: Reported cross-reactivity ....................................................................................................................................................................................... 3-23
Table 14: Reference matrices ................................................................................................................................................................................................. 3-23
Table 15: Other possible matrices ......................................................................................................................................................................................... 3-24
Table 16: Standard solutions .................................................................................................................................................................................................. 4-29
Table 17: FAPAS quality control test materials ................................................................................................................................................................... 4-30
Table 18: Recovery according to 401/2006 .......................................................................................................................................................................... 4-33
Table 19: Extraction protocols .............................................................................................................................................................................................. 4-34
Table 20: Results (p-values) of ANOVA test ....................................................................................................................................................................... 4-37
Table 21: Detection capability (CCβ) .................................................................................................................................................................................... 4-43
Table 22: Cross-reactivity for 4 aflatoxins ............................................................................................................................................................................ 4-48
List of abbreviations and symbols
ACN : Acetonitril LFD : Lateral Flow Device

AFB1 : Aflatoxin B1 MeOH : Methanol
AFB2 : Aflatoxin B2 MS : Mass spectrometry
AFG1 : Aflatoxin G1 N : Number
AFG2 : Aflatoxin G2 NS : Not specified
ANOVA : Analysis of Variance NR : Not relevant
AOAC : Association of Analytical Communities (formerly “Association of NRL : National Reference Laboratory
International Official Analytical Chemists”) OTA : Ochratoxin A
Av : Average Q1 : First quartile
B : Absorbance Q2 : Second quartile (median)
Bo : Absorbance blank Q3 : Third quartile
CCα : Decision limit Rec% : Recovery in %
CCβ : Detection capability SCF Scientific Committee on Food
CEN : European Committee for Standardization SD : Standard deviation
CL : Confidence limit STER : Sterigmatocystin
CR% : Cross Reactivity % t-test : Student t-test
CRM : Certified Reference Material u : uncertainty
CV : Coefficient of Variation (in %) U(Ext) : Extended uncertainty
CA : Cyclopiazonic acid UK : United Kingdom
DON : Deoxynivalenol US : United States of America
EC : European Commission xe : Experimental mean (ng/g)
EFSA : European Food Safety Authority xt : True referenced mean (ng/g)
ELISA : Enzyme Linked Immuno Sorbent Assay ZEN : Zearalenon
EMAN : European Mycotoxins Awareness Network
EU : European Union
FAVV- AFSCA : Federal Agency for the Safety of the Food Chain
FAO : Food and Agriculture Organisation
FAPAS : Food Analysis Performance Assessment Scheme
FB1 : Fumonisin B1
FGIS : Federal Grain Inspection Service
(HP)LC-FL : Liquid Chromatography Fluorescence detection
FPIA : Fluorescence Polarization Immunoassay
FVO : Food and Veterinary Office
GIPSA : Grain Inspection, Packers and Stockyards Administration
HPLC : High performance liquid chromatography
IAC : Immunoaffinity clean-up
IC50 : Half maximal inhibitory concentration
ICH : International Conference on Harmonization of Technical
Requirements for Registration of Pharmaceuticals for Human Use
ISO : International Organization for Standardization
IQR : Inter Quartile Range
k : Coverage factor (usual = 2)
LC-MS/MS : Liquid Chromatography with tandem in space mass spectrometry
detection
1. Foreword & scope &
background
Evaluation of immunoassay kits for aflatoxin determination in corn & rice
1.1 Foreword
In Belgium, the Federal Agency for the Safety of the Food Chain (AFSCA-FAVV), aiming to preserve the safety of
the food chain and the quality of food, is responsible for the official control of all feed and food products, either
produced in Belgium or imported. As such, it is also the authority competent for monitoring mycotoxin
contamination throughout the food chain. Each year, a monitoring program is established by the food agency
specifying the number of samples, food and feed commodities, schedule, sampling method and the contaminants to
be analyzed. From a practical point of view, samples are collected all over Belgium throughout the year. Analyses are
carried out in approved and accredited laboratories for the purposes of the official control. A key requirement of any
food safety management system is the process of demonstrating its continuous effectiveness by verification of the
analytical methods. Given the potential application of the commercial rapid test kits as screening tools, it is
important that the performance of these immunochemical techniques should, when possible, be evaluated against
established reference methods to assess their reliability. In Belgium, AFSCA-FAVV has appointed CODA-CERVA,
as national reference laboratory (NRL), to evaluate the fast mycotoxin test kits that are currently available on the
market, based on currently available knowledge regarding applicability, reproducibility, precision, recovery, accuracy
and cost level. This evaluation aimed to promote reliable fast mycotoxin measurement systems in Belgium. In 2008-
2009 kits for deoxynivalenol were assessed1, in 2009-2010 kits for Ochratoxin A, this year kits for aflatoxin (B1
or/and total aflatoxin).
For the set-up of the evaluation protocol, a strategic group of experts was formed. This group included members of
the AFSCA-FAVV, CODA-CERVA and several kit users. This network was extended to the manufacturers and
allowed us to identify possible drawbacks of these kits. Figure 1 shows the Partnership developed for performing the
evaluation work.
The evaluation was performed in three main steps: i) inventory of available rapid fast AFLATOXIN testing kits (ii)
administrative evaluation based on questionnaires to be filled in by the kit producers and (iii) experimental evaluation
of quantitative kits by checking some critical parameters (drawbacks).
AFSCA - FAVV
(National Food Agency)
- Information exchange
- Workshop
- Group communication
- Feedback of reporting
- Scientific public
CODA - CERVA Guidance
National Reference Laboratory
(Evaluation works)
Kit providers
Kit Users' expert committee
(Manufacturers, Distributors) (Public & private Laboratories, Universities,
Evaluators…)
- Information
- Informationexchange
exchange - Information exchange for experimental evaluation
- Questionnaire
- Kits & devicesfilling
providing - Feedback of experimental protocol
- Feedback
- Demonstration
of experimental
trials protocol - Network data providing
- Kits
- Feedback
& devices
of reporting
providing - Participation of workshops
- Demonstration
- Scientific public
trials - Evaluation of intermediary report
- Feedback of reporting - Scientific public
- Scientific public
Figure 1: Partnership
1 http://www.var.fgov.be/
CODA-CERVA | Foreword & scope & background 1-1

1.2 Scope
1.2.1 Objective and strategy of the inventory task

At the present immunoassay kits are available in three different formats; FPIA (fluorescence polarization
immunoassay), ELISA (enzyme-linked immunosorbent assay) and LFD (lateral flow devices). Commercialized by
several providers they are available for all legally regulated mycotoxins.
The inventory of fast test kits attempted to
 Identify all the existing kit providers active on the Belgian market
 Identify the available kits for aflatoxin (either Aflatoxin B1 or total aflatoxin) determination in corn &
rice
 Establish collaboration with the identified kit providers
The inventory was based on the list of rapid mycotoxin testing kits available at the website “mycotoxins.org” of the
European Mycotoxins Awareness Network (EMAN) project. Since the last version of the EMAN website was
updated in 2003, the inventory was completed using the individual websites of the kit providers and personal contact
(telephone calls, mailings, meetings with kit producers or distributors).
1.2.2 Objective and strategy of the administrative evaluation

Information was gathered on the available aflatoxin kits from manufacturers through two questionnaires, one for
quantitative and one for qualitative aflatoxin kits. These were developed in close cooperation with the
aforementioned expert group and were addressed by mailing to the kit manufacturers. Table 1 summarizes the most
common parameters checked for quantitative or qualitative aflatoxin kits.
Table 1: Parameters for quantitative and qualitative kits2
Quantitative kits Qualitative kits
Detection limit Cut-off limit
Range and linearity dichotomic value
Sensitivity and specificity False positive and false negative rates
Measurement Uncertainty Unreliability region
Accuracy: trueness, precision Sensitivity and specificity
Selectivity: interferences Selectivity: interferences
Ruggedness or robustness Ruggedness or robustness
Formats and cost per kit Formats and cost per kit
Quantity of materials and reagents needed Quantity of materials and reagents needed
Protocol, throughput and time required for performing analysis Protocol, throughput and time required for performing analysis
Scientific supports and quality assurance in manufacturing Scientific supports and quality assurance in manufacturing
Easy-to-use’ information according to the manufacturers Easy-to-use’ information according to the manufacturers
Note: for a quantitative method, sensitivity should be a numerical value that indicates how the response changes whenever there is a variation in
the concentration of the analyte. However, this parameter will be evaluated in a different way if a qualitative method is used.
3
Regulation 1881/2006/EC (as amended by Regulation Nr. 165/20104) lays down maximal limits for certain
contaminants in specified foods, and as such determines the requirements with respect sensitivity of the analytical
2 E. Trullols Trends in Anal. Chem. 23, 2004, 137

3 http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:L:2006:364:0005:0024:EN:PDF
1-2 Foreword & scope & background | CODA-CERVA

methods. Furthermore, Regulation 401/2006EC 5 lays down methods for sampling and analysis, thereby defining
some method performance criteria (i.e. minimum recovery rate, maximum reproducibility, restrictions with respect to
the expanded analytical uncertainty). This is amended by regulation 178/2010 6 regarding groundnuts (peanuts), other
oilseeds, tree nuts, apricot kernels, liquorice and vegetable oil. Note that CEN-CR 13505:19997 and Decision
2002/657/EC8 indicate the parameters that are relevant for screening and confirmation methods. Decision
2002/657/EC8 states that screening methods should be validated and that the 'false compliant rate (false negatives)
should be <5% at the level of interest without providing much detail on how to establish this. Fortunately, a
complementary guideline has been published9. Essentially, both documents prescribe that in an initial validation, at
least 20 samples spiked at the anticipated screening reporting level need to be analyzed and the target analyte(s) need
to be detectable in at least 19 out of 20 samples.
For the approval of pre-export control on aflatoxins in peanuts from the U.S. Article 23 of Regulation (EC) No
882/200410 provides that pre-export checks carried out by a third country on feed and food immediately prior to
export to the Community with a view of verifying that the exported products comply with the Community
requirements may be approved. Such an EC approval of pre-export controls has as consequence that these pre-
export controls replace or reduce the documentary, identity and physical controls at import into the EC. An FVO
inspection carried out in September 2006 concluded that the US have a well-defined control system for aflatoxin
levels in peanuts and well performing approved laboratories. Therefore approval of pre-export control was granted
by Commission Decision 2008/47/EC11 in December 2007. As these pre-export controls replace effectively and
reliably the controls at import this should result in a significant decrease of controls at import. Special conditions
governing certain foodstuffs imported from certain third countries due to contamination risks of these products by
aflatoxins are laid down in Commission Regulation (EC) No 1152/2009 12. In order to assist the competent
authorities on the official control of aflatoxin contamination in food products which are subject to Commission
Regulation (EC) 1152/2009, a guidance document "Guidance document for competent authorities for the control of
compliance with EU legislation on aflatoxins”13 has been elaborated. This document is also applicable for the control
of aflatoxins in food products not subject to the safeguard Regulation. A number of scientific opinions were also
adopted: The Scientific Committee on Food (SCF) adopted on 23 September 1994 an opinion14 on toxicological
safety of aflatoxins B1, B2, G1, G2 and M1. In addition, the European Food Safety Authority (EFSA) has adopted
on 3 February 2004 an opinion15 related to aflatoxin B1 as undesirable substance in animal feed. EFSA adopted on
29 January 2007 an opinion16 related to the potential increase of consumer health risk by a possible increase of the
existing maximum levels for aflatoxins in almonds, hazelnuts and pistachios and derived products. EFSA adopted on
16 June 2009 a statement17 related to the effects on public health of an increase of the levels for aflatoxin total from
4 µg/kg to 10 µg/kg for tree nuts other than almonds, hazelnuts and pistachios.
5 http://eurlex.europa.eu/LexUriServ/LexUriServ.do?uri=CONSLEG:2006R0401:20100313:EN:PDF
7 http://www.cen.eu/CEN/Pages/default.aspx
9 http://ec.europa.eu/food/food/chemicalsafety/residues/Guideline_Validation_Screening_en.pdf
12 http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:32009R1152:EN:NOT
13 http://ec.europa.eu/food/food/chemicalsafety/contaminants/guidance-2010.pdf
14 http://ec.europa.eu/food/food/chemicalsafety/contaminants/scf_reports_35.pdf
15 http://www.efsa.europa.eu/EFSA/efsa_locale-1178620753812_1178620763196.htm
16 http://www.efsa.europa.eu/EFSA/efsa_locale-1178620753812_1178620761977.htm
17 http://www.efsa.europa.eu/en/scdocs/scdoc/1168.htm

Table 2: Regulatory limits EU (ref. 3-4-6)
Food product Mycotoxin Max Level

(µg/kg)
Groundnuts to be subjected to sorting, or other physical treatment, before human consumption or Aflatoxin B1 8
use as an ingredient in foodstuffs
Total 15
B1+B2+G1+G2
Nuts to be subjected to sorting, or other physical treatment, before human consumption or use as Aflatoxin B1 5
an ingredient in foodstuffs
Total 10
B1+B2+G1+G2
Groundnuts, nuts and processed products thereof, intended for direct human consumption or use Aflatoxin B1 2
as Aflatoxin B1 an ingredient in foodstuffs
Total 4
B1+B2+G1+G2
Dried fruit to be subjected to sorting, or other physical treatment, before human consumption or Aflatoxin B1 5
Total 10
B1+B2+G1+G2
Dried fruit and processed products thereof, intended for direct human consumption or use as an Aflatoxin B1 2
ingredient in foodstuffs
Total 4
B1+B2+G1+G2
All cereals and all products derived from cereals, including processed cereal products Aflatoxin B1 2
Total 4
B1+B2+G1+G2
Maize or rice to be subjected to sorting, or other physical treatment, before human consumption or Aflatoxin B1 5
Total 10
B1+B2+G1+G2
Raw milk, heat-treated milk and milk for the manufacture of milk-based products and as defined Aflatoxin M1 0.05
by Council Directive 92/46/EEC)
Aflatoxin B1 5
Following species of spices: Capsicum spp. (dried fruits thereof, whole or ground, including
Total 10
chillies, chilli powder, cayenne, and paprika), Piper spp. (fruits thereof, including white and black
B1+B2+G1+G2
pepper) Myristica fragrans (nutmeg) Zingiber officinale (ginger) Aflatoxin B1 Curcuma longa
(turmeric)
Processed cereal-based foods and baby foods for infants and young children Aflatoxin B1 0.1
Total -
B1+B2+G1+G2
Infant formulae and follow-on formulae, including infant milk and follow-on milk Aflatoxin M1 0.025
Dietary foods for special medical purposes intended specifically for infants Aflatoxin B1 0.1
Total -
B1+B2+G1+G2
Aflatoxin M1 0.025
1.2.3 Objective and strategy of the experimental evaluation

A research protocol has then been elaborated and discussed with the users’ committee. Several contacts have been
made with the kits providers to obtain their formal consent in participating to this research. Analyses were
performed by one analyst using the same lot of test kits in the same laboratory environment. The overall "ease of
use" of the kit was also appreciated by laboratory workers during the experimental evaluation.

According to Commission Decision 2002/657/EC the following parameters (Table 3) have to be validated for
screening methods:
Table 3: Criteria to be verified according to 657/20028
CCβ CCα Accuracy Recovery Precision Specificity Robustness

Qualitative + - - - - + +
Quantitative + - - - + + +
The main drawbacks identified, based on the literature, feedback from kit producers, experience of the members of
the expert group and the experiences gained in the evaluation of the DON & OTA kits the past two years year, the
following problems were indicated:
1. Cross-reactivity
2. Accuracy
3. Recovery
4. Detection capability (CCβ)
5. Precision
The kits of the following manufacturers were assessed: Romerlabs, R-Biopharm, Tecna, Charm & Neogen.
1.3 Background
Mycotoxins are secondary metabolites of a range of filamentous fungi with deleterious effects on humans and
animals, which can be found in agricultural commodities and animal foodstuffs 18. They are present in trace amounts
so there is a need for highly sensitive and selective analytical methods for these natural toxins19
Among various mycotoxins, aflatoxins have assumed significance due to their deleterious effects on human beings,
poultry and livestock. The aflatoxin problem was first recognized in 1960, when there was severe outbreak of a
disease referred as "Turkey 'X' Disease" in UK, in which over 100,000 turkey poults died20. The cause of the disease
was shown due to toxins in peanut meal infected with Aspergillus flavus and the toxins were named “aflatoxins”.
Food products contaminated with aflatoxins include cereal (maize, sorghum, pearl millet, rice, wheat), oilseeds
(groundnut, soybean, sunflower, cotton), spices (chillies, black pepper, coriander, turmeric, zinger), tree nuts
(almonds, pistachio, walnuts, coconut) and milk. Aflatoxins are potent toxic, carcinogenic, mutagenic,
immunosuppressive agents, produced as secondary metabolites by the fungi Aspergillus flavus and Aspergillus.
parasiticus on a variety of food products21 22. There are 18 different types of aflatoxins identified, major members are
aflatoxin B1, B2, G1 and G2. Aflatoxin B1 (AFB1) is normally predominant in cultures as well as in food products.
Pure AFB1 is pale-white to yellow crystalline, odorless solid. Aflatoxins are soluble in methanol, chloroform, actone,
acetonitrile. A. flavus typically produces AFB1 and AFB2, whereas Aspergillus. parasiticus produce AFG1 and
AFG2 as well as AFB1 and AFB2 23. Aflatoxin M1and M2 are major metabolites of aflatoxin B1 and B2 respectively,
found in milk of animals that have consumed feed contaminated with aflatoxins 24. The aflatoxins display potency of
toxicity, carcinogenicity, mutagenicity in the order of AFB1 > AFG1 > AFB2 > AFG2 as illustrated by their LD50
18 J. W. Bennett and M. Klich, Clin. Microbiolog. Rev., 2003, 16, 497.

19 I. Kralj Cigic and H. Prosen, Int. J. Mol. Sci., 2009, 10, 62.
20 T. Asao, G. Buchi, M.M. Abdel-Kader, S.B. Chang, E.L. Wick, G.N. Wogan, J. Am. Chem. Soc. 85 (1963) 1706.
21 IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, vol. 56, Some natural occurring Substances: Food
Items and Constituents, Heterocyclic aromatic Amines and Mycotoxins, International Agency for Research on Cancer, Lyon,
1993, p. 489.
22 Commission Regulation (EC) No. 1525/98, amending Regulation (EC) No 184/97 setting maximum residue levels for certain
contaminants in foodstuffs, Official J. Eur. Commun. L 201/43 (1998) 43.

23 D’Mello, J.P.F., MacDonald, A.M.C., 1997. Mycotoxins. Anim. Feed Sci. Technol. 69, 155–166.
24 Carnaghan, R.B.A., Hartley, R.D., O’Kelly, J., 1963. Toxicity and fluorescence properties of the aflatoxins. Nature (London)
200, 1101–1102.

values for day-old ducklings25 26. The aflatoxins fluoresce strongly in ultraviolet light (ca. 365 nm); B1 and B2
produce a blue fluorescence where as G1 and G2 produce green fluorescence27. According to FAO estimates, 25%
of the world food crops are affected by mycotoxins each year. And also crop loss due to aflatoxins contamination
costs US producers more than $100 million per year on average including $ 26 million for peanuts ($69.34/ha)28.
Figure 2: Structure of AFB1, AFB2, AFG1 & AFG2
Analytical methods for rapid, sensitive, and accurate determination of these mycotoxins in unprocessed cereals and
cereal-based products are highly needed in order to properly assess both the relevant risk of exposure and the
relevant toxicological risk for humans and animals, as well as to ensure that regulatory levels (see table 1) fixed by the
EU or other international organizations are met.
Testing for mycotoxins is conducted under many different circumstances and for a variety of reasons, which has led
to a proliferation in the number of test methods. Selecting the appropriate method depends upon the intended use
of the method. Factors such as the speed of the method, its accuracy, the skill level required for performing the
assay, and the cost will all impact method selection. The methods basically fall into two major categories: those that
can be conducted with minimal training in laboratories or in the field (screening assays), and those that must be
conducted by more fully trained personnel in analytical laboratories. In general, fast and easy-to-use ELISA based
aflatoxin screening kits are commercially available for all major types of aflatoxins. Quantification is predominantly
done with LC-FL 29. Detection limits in the low ppt range can easily be achieved when iodine is added post-column
to enhance method sensitivity. In addition, immunoaffinity sample clean-up has been shown to have a great potential
to increase method specificity and sensitivity by selective enrichment and isolation of the target aflatoxins29. A
number of quantitative methods have been published to determine major aflatoxins and the structurally related
25 Terao, K., Ueno, Y., 1978. Morphological and functional damage to cells and tissues. In: Uraguchi, K., Yamazaki, M. (Eds.),
Toxicology, Biochemistry and Pathology of Mycotoxins. Wiley, New York, pp. 189–210.
26 Cole, R.J., Cox, R.H., 1981. Handbook of Toxic and Fungal Metabolites. Academic Press, New York.
27 Sargeant, K.C.R.B.A.A.R., 1963. Chemistry and origin of aflatoxins, Chemical Ind. London, 53–55.
28 Council for Agriculture Science and Technology. “Mycotoxins:Economic and Health Risks”. Report 116, Council for
Agriculture Science and Technology. Ames, IA. 1989.

29 E. Papp, K. H-Otta, G. Zaray, E. Mincsovics, Microchem. J. 73 (2002) 39.

sterigmatocystin in food 30 31 32 33 34 35 36 37, milk 38 39, herbs 40, urine 41 42 43 44 and cigarette smoke 45 . In this field,
LC/MS seems to be just a minor alternative or confirmation technique for the already well established reliable and
robust LC-FL methodology 46 47 though it should be useful to confirm positive results and ELISA based screening
analysis.
Table 4: Advantages and disadvantages of traditional and emerging methods for mycotoxin analysis
Method Advantages Disadvantages
Expensive equipment, specialist expertise required,
Simultaneous analysis of mycotoxins, good derivatization required, matrix interference problems, non-
Gas Chromatography sensitivity, may be automated (autosampler), linear calibration curve, drifting response, carry-over effects
provides confirmation (MS detector). from previous sample, variation in reproducibility and
precision
Good sensitivity, good selectivity, good

High Performance Liquid precision, may be automated (autosampler), Expensive equipment, specialist expertise required, may
Chromatography short analysis times, official methods require derivatization
available.
Liquid Simultaneous analysis of mycotoxins, good Very expensive, specialist expertise required, sensitivity relies
Chromatography/Mass sensitivity (LC/MS/MS), provides on ionization technique, matrix assisted calibration curve (for
Spectrometry confirmation, no derivatization required. quantitative analysis)
Simple sample preparation, inexpensive
Cross-reactivity with related mycotoxins, matrix interference
Enzyme-Linked equipment, high sensitivity, simultaneous
problems, possible false positive/negative results,
Immunosorbent Assay analysis of multiple samples, suitable for
confirmatory LC analysis required.
screening, limited use of organic solvents.
Lateral Flow Rapid, no clean-up, no expensive equipment, Cross-reactivity with related mycotoxins, validation required
Device/Dipstick easy to use, no specific training required. for additional matrices.
Fluorescence Polarization Cross-reactivity with related mycotoxins, matrix interference
Rapid, no clean-up required
Immunoassay problems.
Rapid, non-destructive measurement, no Expensive equipment, calibration model must be validated,
IR (Infrared;) spectroscopy
extraction or clean-up, easy operation. knowledge of statistical methods, poor sensitivity
Cross-reactivity with related mycotoxins, extract clean-up

Biosensors Rapid, no clean-up procedure needed to improve sensitivity, variation in reproducibility and
precision.
In all cases, obtaining a representative test sample of the overall lot is essential in order to ensure that the results of
the test sample can be correctly ascribed to the lot. Sampling, subsampling, grinding, and extraction of commodities
30 S. Biselli, L. Hartig, H. Wegener, C. Hummert, LC–GC Eur. 17 (2004) 25.

31 S. Biselli, L. Hartig, H. Wegener, C. Hummert, Spectroscopy 20 (2005) 20.
32 M. Kokkonen, M. Jestoi, A. Rizzo, Food Addit. Contam. 22 (2005) 449.
33 M. Vahl, K. Jorgensen, Z. Lebensm. Unters. Forsch. A 206 (1998) 243.
34 K.A. Scudamore, M.T. Hetmanski, P.A. Clarke, K.A. Barnes, J.R. Startin, Food Addit. Contam. 13 (1996) 343.
35 J. Blesa, J.M. Soriano, J.C. Molto, R. Marin, J. Manes, J. Chromatogr. A 1011 (2003) 49.
36 T.F. Schatzki, W.F. Haddon, J. Agric. Food Chem. 50 (2002) 3062.
37 M. Takino, T. Tanaka, K. Yamaguchi, T. Nakahara, Food Addit. Contam. 21 (2004) 76.
38 C. Cavaliere, P. Foglia, E. Pastorini, R. Samperi, A. Lagana, J. Chromatogr. A 1101 (2006) 69.
39 L.K. Sorensen, T.H. Elbaek, J. Chromatogr. B 820 (2005) 183.
40 M. Ventura, A. Gomez, I. Anaya, J. Diaz, F. Broto, M. Agut, L. Comellas, J. Chromatogr. A 1048 (2004) 25.
41 M.Walton, P. Egner, P.F. Scholl, J.Walker, T.W. Kensler, J.D. Groopman, Chem. Res. Toxicol. 14 (2001) 919.
42 P.A. Egner, X. Yu, J.K. Johnson, C.K. Nathasingh, J.D. Groopman, T.W. Kensler, B.D. Roebuck, Chem. Res. Toxicol. 16
(2003) 1174.
43 P.F. Scholl, S.M. Musser, J.D. Groopman, Chem. Res. Toxicol. 10 (1997) 1144.
44 A. Kussak, C.-A. Nilsson, B. Andersson, J. Langridge, Rapid Commun. Mass Spectrom. 9 (1995) 1234.
45 L.E. Edinboro, H.T. Karnes, J. Chromatogr. A 1083 (2005) 127.
46 H.K. Abbas, W.P. Williams, G.L. Windham, H.C. Pringle, W. Xie, W.T. Shier, J. Agric. Food Chem. 50 (2002) 5246.
47 H.K. Abbas, R.D. Cartwright, W. Xie, W.T. Shier, Crop Prot. 25 (2006) 1.

take considerable attention and time. In fact, these steps often require more time than some of the rapid assays for
detecting the toxins. Therefore, where possible, it is preferable to combine a rapid extraction technique with rapid
assays in order to minimize the overall assay time. Extraction usually involves conventional procedures using
acetone, chloroform and methanol or mixtures thereof. Adding water to the extraction solvent generally gives higher
extraction efficiencies.
In spite of the striking importance of the quality of the commercial fast test kits for their effectiveness of mycotoxin
control, transversal evaluation or comparison of several kits is also needed in order to check how far they could help
to meet the ISO 17025 requirements. This report provides noticeable features of the inquiry as well as the overall
outcome of the evaluation process as performed for AFLATOXIN commercial test kits. The kits evaluated were
dedicated for cereal and cereal products because of its relevance to the official residue control in Belgium.
ELISA
Enzyme-Linked ImmunoSorbent Assay (ELISA) is an antibody-based assay that is commonly used to detect
mycotoxins48. A number of commercial ELISA kits are available for aflatoxins, deoxynivalenol, fumonisins,
ochratoxin, HT2 & T2 and zearalenone. This is usually a competitive assay in which the mycotoxin of interest from a
sample competes with a labeled mycotoxin for a limited number of specific antibody-binding sites. The greater the
amount of toxin present in the sample, the lower the binding of the labeled toxin and the lower the signal generated
by the assay49 50 51. ELISA is one of the more affordable methods for detecting mycotoxins, and has the advantage
that a large number of samples can be measured at once. Commercial 96-well assays and strip-tests are available for
many mycotoxins.
ELISA techniques have been shown to be less accurate and less sensitive than conventional chromatographic assays.
Very few correlations were found between the two types of techniques. In addition, false positive or false negative
results often occurred with ELISA because of cross-reactions between molecules or interferences with the antibody
reagents. They are thus considered to be suitable for qualitative assessment or for sample pre-screening but not for
quantitative determination. It is also recommended to use the ELISA techniques only for the foods they were
developed for.
Lateral flow devices

Immunochromatographic assays, also called “Lateral Flow Devices” are user-friendly formats requiring a very short
time for results52. Basically, a ligand that can be bound to a visually detectable solid support, such as dyed
microspheres, can be qualitatively tested and in many cases even quantitatively. The most popular label for the lateral
flow test are particles consisting of colloidal gold coated with the antibody which provide red-colored binding zones.
Liquid sample is added to the sample pad, the liquid components of the assay move along the membrane by capillary
flow to the absorbent pad53 54 55 at the end of the strip. When migrating, the sample suspends the gold particles and
the mycotoxin, if present, binds to the particles. When the test line is reached, which is coated with the analyte of
interest, the gold particles will bind with their antibody if they are still unbounded and color the line red. If the
sample was contaminated, the gold particles will be already bounded and flush over the test line. Therefore, absence
48 Michael Z. Zheng, John L. Richard, Johann Binder, Mycopathologia 2006, 161: 261–273
49 CAST. Management and detection of mycotoxins. In Mycotoxins: Risks in Plant, Animal, and Human Systems. Council for
Agricultural Science and Technology: Council for Agricultural Science and Technology, Ames, IA, USA, 2003: 86–103.
50 Anklam E, Stroka J, Boenke A. Food Control 2002; 13: 173–183
51 Seefelder W, Gossman M, Humpf HU. J. Agric. Food Chem. 2002; 50, 2778–2781.
52 Steven J. Saul and , Mark Tess, Lateral Flow Quantitative Method for the Detection of Mycotoxins, Food Contaminants,
Chapter 19, pp 314–319

53 United States Patent 6.475.805

of analyte results in red color for the test line. Additional chemicals or handling steps are, in contrast to ELISA tests,
not required.
Lateral flow devices with colloidal gold labels have been studied for the most important mycotoxins, such as
aflatoxins, DON, T-2, fumonisins, OTA and ZEN. They have the large advantage that no labor intensive and time
consuming washing steps are necessary making it possible to use them in the field.

2. Inventory
2.1 Inventory results of providers of aflatoxin kits
Table 5 gives an overview of the kit manufactures that offer kits for aflatoxin (B1 or total aflatoxin, aflatoxin M1 is
not included) determination. Of these manufacturers, 5 wished to participate (Table 6) in the evaluation.
Figure 3 gives on overview in function of: (1) type test, (2) B1 versus Total & (3) Qualitative versus Quantitative. It
can be seen clearly that the market is dominated by ELISA kits with strip based tests (Lateral Flow Devices or
dipsticks) following closely. Most of these kits are quantitative ones focusing on total aflatoxin. As quantitative kits
dominate the market it was decided to focus on them, participants however were free to decide whether they send in
a kit for AFB1 or for total aflatoxin. There was also no restriction set on the format that could be submitted for the
experimental evaluation.
Qualitative - Quantitative FLUO/COLUMN, Type test

4
NS, 1 NS, 1
FPIA, 3
Qualit, 16
ELISA, 25
Quant, 31 STRIP, 15
B1-Total aflatoxin
NS, 1
B1, 13
TOTAL
(B1,B2,G1,G2),
34
Figure 3: Distribution of kits in function of: type test, B1 vs Total & Qual vs Quan
CODA-CERVA | Inventory 2-13

Evaluation of immunoassay kits for determination of ochratoxin A in cereals
Table 5: Overview of contacted manufacturers and their kits

Producer Kit Afla Range Matrix Format Quant/Qual Remark
Abraxis Aflatoxin, ELISA Kit Total 2-100 ng/g Grain/grain products/nuts ELISA Quant
Cut-Off 5
Aflatoxin, ELISA Kit Dipstick Total Grain/corn/rice Strip Qual
ng/g
Aokin Mycontrol NS NS NS NS NS In development
Beacon Aflatoxin Plate kit Total 2-100 ng/g Grain ELISA Quant
Aflatoxin, Tube kit Total 2-100 ng/g Grain ELISA Quant
ROSA® Fast AFQ
Charm Sciences B1 0-150 ng/g Corn LFD Quant
(Quantitative)
ROSA® AFQ (Quantitative) B1 0-150 ng/g Grains LFD Quant methanol/ethanol
Cut-Off 10 /
ROSA® aflatoxin P/N B1 Grains LFD Qual methanol/ethanol
20ng/g
Cut-Off 10 /
ROSA® Best aflatoxin P/N B1 Grain LFD Qual water extraction
20ng/g
Cut-off: 20
Diachemix Aflatoxin FPA Qualitative Total Corn FPIA Qual
ng/g
Aflatoxin FPA 10 ng/g Cut-off: 10
Total Corn FPIA Qual
Qualitative Test Kit ng/g
Aflatoxin FPA Quantitative Total NS Grains/nuts FPIA Quant In development
Diagnostix EZ-Quant Aflatoxin Plate Kit Total 1-25 ng/g grain ELISA Quant
QuickTox™ Kit for Aflatoxin - Cut-off: 20
Envirologix Total Corn Strip Qual
20 ng/g - AS 101 ng/g
QuickTox Kit for QuickScan -
Total 2.5-30 ng/g Corn LFD Quant
Aflatoxin - AQ 109 BG
Europroxima Aflatoxin B1 ELISA B1 >0.5 ng/g Cereals/soy/nuts ELISA Quant
Aflatoxin Total ELISA Total >0.5 ng/g Wheat, corn, peanut, feed ELISA Quant
Aflatoxin B1 FTRT B1 > 2 ng/g Wheat, corn, peanut, feed LFD Qual
Aflatoxin Total FTRT Total > 4 ng/g Wheat, corn, peanut, feed LFD Qual
MycoMonitor™ Total Grains, cereals, nuts,
Helica Total 1 -20 ng/g ELISA Quant
Aflatoxin Assay animal feed
Total Aflatoxins (LOW grains and cereals, silage,
Total Jan-20 ELISA Quant
MATRIX) nuts, spices, animal feed
LCTech Aflaread B1 2-100ng/g Food/feed ELISA Quant
Cut-off 20
Neogen Agri-Screen for Aflatoxin Total corn, /soy ELISA Qual
ng/g
5 ng/g - 50
Veratox for Aflatoxin Total corn, /soy ELISA Quant
ng/g
Veratox for Aflatoxin AST 5 ng/g - 150
Total corn, /soy ELISA Quant
(Aflatoxin Single Test) ng/g
Veratox for Aflatoxin HS (High 1 ng/g - 8
Total corn, /soy ELISA Quant
Sensitivity) ng/g
NEOCOLUMN Total <10 ng/g, Penuts/grain Fluo Quant
>10 to <20
Reveal for Aflatoxin SQ Peanut Total ng/g, >20 peanuts LFD Qual
ng/g
R-Biopharm AFLAPLATE B1 1.5-50 Cereal ELISA Quant
RIDASCREEN® Aflatoxin B1
B1 0-50 ng/g Cereals ELISA Quant
30/15
RIDASCREEN® Aflatoxin
Total 0-40 Cereals ELISA Quant 96 wells
Total
RIDASCREEN®FAST
Total 1.7-45 Cereals ELISA Quant 48 wells
Aflatoxin
RIDASCREEN®FAST (calibration through
Total >2 ng/g Cereals ELISA Quant
Aflatoxin SC external provided curve
Aflascan Total 1-100 ng/g Nuts, Cereals Fluo Quant
Romerlabs FluoroQuant® Chemical Kits Total 1-100 ng/g Grain,carn… Fluo Quant solsep or IAC column
AgraQuant® Mycotoxin
Total 1-40 ng/g Grain,carn… ELISA Quant
ELISA Kits
corn, grain by-products,
Cut-off,
feeds, raw and roasted
AflaCup™ Kit Total 10ng/g, ELISA Qual
peanuts, peanut butter and
20ng/g
cottonseed
corn, DDGS, soybeans,
Cut-off, 4
cottonseed, peanuts, rice,
AgraStrip® Aflatoxin Test Kit Total ng/g, 10 ng/g LFD Qual
almonds, pistachio and
20 ng/g
many other commodities
Cut-off:,
SDI Myco aflatoxin test strips Total corn strip Qual
20ng/g
Cereals, nuts, dried fruit,
Celer AFLA, quantitative,
Tecna Total Feb-80 paprika and chilli, maize, ELISA Quant 15 min time
ELISA
germs and feeds.
Celer AFLA B1, quantitative, Cereals, maize germ, feeds,
B1 1-40 ng/g ELISA Quant 15 min time
ELISA dried fruits and nuts.
I'screen AFLA, quantitative, Cereals, paprika, chilli,
Total 0.5-20 ng/g ELISA Quant 50 min time
ELISA dried fruits and nuts
Transia Aflatoxin B1 B1 >0.5 ng/g spices, food and feed ELISA Quant
Aflatoxin B1 "sensitv" B1 0.03 ng/g baby food, food and feed ELISA Quant
Cut-off:, 2 or
Aflatoxin B1 membrane test B1 cereals and spices Strip Qual
4 ng/g
Aflatoxin Total membrane test Total Cut-off 4ng/g cereals and spices Strip Qual
Cut-off
Vicam Aflacheck Total 10ng/g, GRAINS/nuts Strip Qual
20ng/g
Aflatest Total 0.1-300 GRAINS/nuts/milk Fluo Quant
2-14 Inventory | CODA-CERVA

Table 6: Overview of kits involved in the experimental evaluation

Kit Manufacturer Total Afla / B1 ELISA LFD
AgraQuant® Total Aflatoxin Assay 1/20 Romer Labs Singapore Pte. Ltd. Total ν
Charm Sciences Inc
659 Andover Street
ROSA Aflatoxin Quantitative B1 ν
Lawrence, MA USA
01843-1032
R-Biopharm AG, An der neuen Bergstr. 17, 64297
RIDASCREEN® Aflatoxin total Total ν
Darmstadt, Germany
Neogen Corporation
620 Lesher Place
Veratox® HS Total Aflatoxin Total ν
Lansing, MI 48912
(800) 234-5333 ext 4450
Tecna S.r.l.
Area Science park
Celer AFLA B1 Padriciano Nr. 99 B1 ν
34149 Trieste
Italy
CODA-CERVA | Inventory 2-15

3. Administrative evaluation
3.1 Analytical performance criteria
All participants were asked to fill in a questionnaire in which data for recovery, accuracy, precision (intra-day & inter-
day), reproducibility and some general information had to be reported.
Table 7 indicates how recovery, standard combined uncertainty and Extended Uncertainty are calculated.
Table 7: Calculation of parameters (for symbols, see list abbreviations)
Parameter Equation
Recovery Equation 1 Rec% = (xe/xt)*100
Uncertainty Equation 2 u= √(CV2+bias2)
Extended Uncertainty Equation 3 U(Ext)= k (=2) * u
3.2 Intra-run precision, between-run precision & reproducibility
The data reported by the kit manufacturers for intra-run precision (same operator/same day/same batch, Table 8),
between-run precision (different operator and /or different day, same batch, Table 9) and reproducibility (different
operators and different days and different batches, Table 10) are reported.
According to document 401/20065 the recommended value for reproducibility should be calculated as by Horwitz
with the maximum allowable value for reproducibility 2 times the value from Horwitz. The precision can be
calculated as 0.66 * Horwitz. For 5ng/g this would generate an expected maximum reproducibility of 40%, a
maximum allowable reproducibility of 80% and an expected precision of 25%.
CODA-CERVA | Administrative evaluation 3-19

Table 8: Intra-run precision

Neogen R-Biopharm Romerlabs Tecna
Level (ng/g) 5.00 5.00 10.00 2.00
Run Run 1 Run 2 Run 3 Run 1 Run 2 Run 3 Run 1 Run 2 Run 3 Run 1 Run 2 Run 3
Uncertainty (%) 21 32 26 28 9 14 19 11 15
Extended Uncertainty (%, k=2) 42 65 52 57 17 27 39 23 29
Level (ng/g) 10.00 15.00 40.00 5.00
Uncertainty (%) 22 22 14 27 23 17 35 17 13
Level (ng/g) 20.00 50.00 160.00 10.00
Uncertainty (%) 13 18 15 19 31 22 32 22 16
Table 9: Between-run precision

Charm Neogen R-Biopharm
Level (ng/g) 5.70 5.00 5.00
Run Scenario 1 Scenario 2 Scenario 3 Scenario 1 Scenario 2 Scenario 3 Scenario 1 Scenario 2 Scenario 3
Uncertainty (%) 28 27 28 26 26 28 13
Extended Uncertainty
57 53 56 52 51 56 26
(%, k=2)
Level (ng/g) 9.20 10.00 15.00
Uncertainty (%) 16 25 23 26 25 19 20
33 50 46 52 51 38 40
(%, k=2)
Level (ng/g) 21.00 20.00 50.00
Uncertainty (%) 28 25 24 15 18 20 31
55 50 48 31 36 41 63
(%, k=2)
3-20 Administrative evaluation | CODA-CERVA

Table 9: Between-run precision (continued)

Romerlabs Tecna
Level (ng/g) 10.00 2.00
Run Scenario 1 Scenario 2 Scenario 3 Scenario 1 Scenario 2 Scenario 3
Uncertainty (%) 16 18 12 15
Extended Uncertainty (%, k=2) 31 37 25 31
Level (ng/g) 40.00 5.00
Level (ng/g) 160.00 10.00
Table 10: Reproducibility

Neogen Romerlabs
Level (ng/g) 0.00 10.00
Level (ng/g) 15.00 40.00
Level (ng/g) 25.00 160.00
3.3 Accuracy & Recovery
According to document 401/2006EC5 recoveries should lay within 70 and 110% of the referenced values. From
Table 11 & Table 12 it can be seen that in most cases the kits fulfill this requirement. Please note that Table 12
mentions the average recovery for all matrices and might disguise large differences in recovery efficiency for different
matrices.

Table 11: Reported accuracy

Conc 1 Conc 2 Conc 3 Conc 4 Conc 5
Reference value (ng/g) 5 20
Corn Meal Mean (ng/g) 6 25
Bias (%) 20% 25%
Corn Flour Mean (ng/g) 5 22
Bias (%) 0% 10%
Charm
Corn Soy Blend Mean (ng/g) 6 23
Bias (%) 20% 15%
Milled Rice Mean (ng/g) 6 21
Bias (%) 20% 5%
Milled Rice Mean (ng/g) 2.9 18.7
Bias (%) -42 -6.5
Corn Gluten Meal Mean (ng/g) 3.3 15.8
Bias (%) -34 -21
Neogen
Corn Meal Mean (ng/g) 5.7 23.4
Bias (%) 14 17
Corn/Soy Blend Mean (ng/g) 4.2 18.7
Bias (%) -16 -6.5
Reference value (ng/g) 4.5 14.3 31.6
Corn Mean (ng/g) 4.4 13.6 26.5
Bias (%) -2% -5% -16%
R-Biopharm
Reference value (ng/g) 2.2 11.5
Rice Mean (ng/g) 1.97 8.52
Bias (%) -10% -26%
Reference value (ng/g) 38.8 142
Rice Mean (ng/g) 35.6 168.6
Bias (%) 8.20% 18.70%
Reference value (ng/g) 10 20 30
Romerlabs Corn Mean (ng/g) 11.6 21.1 31.8
Bias (%) 16% 5.50% 6%
Reference value (ng/g) 20.15 101.9
Corn Mean (ng/g) 22.6 92.5
Bias (%) 12% 9%
Reference value (ng/g) 1.3 3.9 4.7 4.7 5.2
Corn (FAPAS) Mean (ng/g) 1.5 4.1 5 5 5.4
Bias (%) 13% 5% 6% 6% 4%
Reference value (ng/g) 7.3 11.3 18.1 23
Animal Feed (FAPAS) Mean (ng/g) 8.8 11.1 16.1 22.2
Bias (%) 17% -2% -12% -4%
Reference value (ng/g) 1.5
Tecna Nuts (FAPAS) Mean (ng/g) 1.5
Bias (%) 0%
Reference value (ng/g) 4.4
Figs (FAPAS) Mean (ng/g) 5
Bias (%) 12%
Reference value (ng/g) 9.9 23
Corn germ naturally incurred Mean (ng/g) 10.2 22.6
Bias (%) 3%
Table 12: Reported average recovery for reference matrix

Recovery (%)
Charm 110
Neogen 99
R-Biopharm 78
Romerlabs 107
Tecna 109

3.4 Cross-reactivity
To achieve accurate results it is imperative that immune-assay kits do not suffer from a matrix effect caused by other
components than the intended analyte reacting with the anti-bodies as this could result in false positives. All
participants reported no cross-reactivity for other components than for the 4 aflatoxins (Table 13). The kits of
Charm and Tecna which are intended for aflatoxin B1 only show some cross-reactivity for the other three aflatoxins.
Table 13: Reported cross-reactivity

Matrix AFB1 AFB2 AFG1 AFG2
Charm NS NS NS NS NS
Neogen Corn 100% 31% 28% 3.8%
R-Biopharm buffer 100% 48% 75% 18%
Romerlabs Corn 100% 65% 70% 42%
Tecna Methanol 100% 5% 19% <1
3.5 Matrix applicability
In Table 14 the matrices indicated by the kit manufacturers as the reference matrices are summarised, Table 15
summarises other possible matrices for which the kit can be used.
Table 14: Reference matrices
Target main matrix Charm Neogen R-Biopharm Romerlabs Tecna

Wheat X
Barley X
Oats X
Corn X X X X X
Rice X X

Table 15: Other possible matrices
Supplier Matrix
Wheat, barley, oats, corn, rice, Basmati rice, broken/brewer’s rice, corn flour, corn germ meal, corn gluten
Charm meal, corn meal, corn screenings, corn/soy blend, cracked corn, Distillers grains, Dried Distillers Grains with
Solubles, flaking corn grits, milled rice, millet, popcorn, rough rice, rye, sorghum, soybeans, wheat flour
Corn, Ammoniated corn, ammoniated cottonseed meal, barley, beet pulp*, coconut, copra, corn, corn bran,
corn grits, corn meal, corn germ meal, corn gluten meal, corn/soy blend, corn starch, cottonseed, cottonseed
Neogen meal, figs, flaxseed meal, hominy, kamut, lentils, milo, oat fiber, peanut hulls, peanuts (raw), peanuts (roasted),
peanut butter, peanut meal, pet food, popcorn, pumpkin seeds, quinoa, rice, rice bran, rice hulls, soy flour, soy
germ meal, soy meal, sunflower meal, wheat and wheat midds.*
R-Biopharm* Corn, rice
Corn, Almond, barley, beer, canola oil, chickpea, chili, corn bran, corn gluten meal, corn meal, corn/soy
Romerlabs blend, cotton seed, DDGs, feed, hazelnut, peanuts, petfood, pistachio, popcorn, raisin, rice, sesame, sorghum,
soybean, walnuts, wheat
Tecna Corn, Animal feed, nuts, dried fruits, corn germ
*) More matrices are available on costumer demand

3.6 Conclusions of the administrative evaluation
From the data provided by the manufacturers it can be seen that all the kits have the necessary sensitivity and
selectivity to adequately detect positive samples; the relatively high variability however might hamper their use as a
reference method. The most common used reference method, HPLC-FL with IAC achieves very good precision
(reproducibility of <20%) but this method is fully standardised, uses labour intensive clean-up and expensive lab
equipment that cannot be used in the field. LC-MS/MS makes the use of sample clean-up obsolete and, if an
isotopically labelled standard is used, achieves repeatabilities rivalling those of HPLC-FL methods but suffers from
the same main drawback; it cannot be used as a field screening method.

4. Experimental evaluation
4.1 Materials and methods
Kit manufactures were given the choice to send in a kit for total aflatoxin determination or one dedicated solely for
aflatoxin B1. As aflatoxin B1 is the most prevalent detected mycotoxin in Europe and, more important, is the most
toxic one, aflatoxin B1 will be used to screen the total aflatoxin kits. Corn and rice were chosen as the model
matrices; aflatoxins are not expected to be found on cereals grown in Belgium and are mainly detected on imported
cereals as corn and rice. Tecna asked not to screen rice with their kit as at the time the evaluation started it was not
yet validated for this matrix. As the MRL of focus 5 ng/g was chosen (the MRL in the EU for corn & rice (Table
2)). The final concentration used was corrected to 3.1 ng/g as the purchased reference solution was lower in
concentration than expected.
Compared to the evaluation of last year two points were no longer evaluated namely the concentration accuracy of
the standards and the cross-reactivity in matrix. The concentration accuracy was no longer verified for two reasons
(1) it is the final result in quality control materials that is important (2) some manufacturers don’t use real mycotoxins
as standard but an analogue which makes it very difficult to verify the concentration by fluorescence or LC-MS.
Cross-reactivity was not repeated as it yielded no extra information.
4.1.1 Equipment and materials
4.1.1.1 Kits used
The kits assessed are summarised in Table 6.
4.1.1.2 Standards and reagents
Acetonitril and methanol (HPLC-MS grade) were purchased from Biosolve (Valkenswaard, The Netherlands).
Deionised water was delivered by a Milli-Q system (Millipore, MA, USA). Certified reference solutions (Table 16)
AFB2, AFG1, AFG2, OTA & FB1 were purchased from Biopure (Tulln, Austria). Crystalline AFB1 and
sterigmatocystin were purchased from Biopure and dissolved in ACN, pure crystalline forms of cyclopiazonic acid
were also purchased from Sigma-Aldrich (Bornem, Belgium).
Table 16: Standard solutions
AFB2 AFG1 AFG2 OTA Sterigmatocystin Fumonisin B1
Provider Biopure Biopure Biopure Biopure Biopure Biopure
μg/ml 0.5 2 0.5 10 50 50
Quant 5 ml 5 ml 5 ml 1 ml 5 1
Solvent ACN ACN ACN ACN ACN 50 % ACN
A stock solution of 10 mg of cyclopiazonic acid in 100 ml ACN in a volumetric flask was prepared and stored in the
freezer at <-15°C. Daily working dilutions were prepared by diluting the stock solutions in 70% MeOH to a
concentration of 100 ng/ml. This solution was used to spike the blank matrix for the recovery experiments.
Successive dilutions for direct application on the kit were made in the appropriate extraction solvent.
CODA-CERVA | Experimental evaluation 4-29

4.1.1.3 Read-out
For the read-out of the ELISA kits a microplate reader model 550 of BIO-RAD (Hercules, US) was used. The read-
out of Charms LFD was done with the proprietary LFD reader of Charm.
4.1.1.4 Blank matrix
Two blank matrices, one corn and one rice, were used to assess the recovery. Both were purchased in a grocery store
and were determined to be blank using an LC-MS/MS based method (publication pending). For spike experiments
samples were weighted in appropriate HDPE extraction tubes and spiked with the 100 ng/ml working solution and
left overnight to allow the solvent to evaporate.
4.1.1.5 Accuracy
Contaminated matrices (quality control test materials) namely maize (TO4138) and rice (T04151) from FAPAS were
used to assess the accuracy (Table 17).
Table 17: FAPAS quality control test materials

Component Assigned value (µg/kg) Satisfactory range (µg/kg)
Afla B1 1.93 1.08-2.78
Afla B2 0.99 0.55-1.42

Rice (T04151) Afla G1 1.56 0.87-2.24
Afla G2 0.69 0.39-1.00
Aflatoxin (Total) 5.09 2.85-7.32
Afla B1 1.87 1.05-2.70
Afla B2 0.51 0.28-0.73
Corn (T04138) Afla G1 0.96 0.54-1.38
Afla G2 0.52 0.29-0.75
Aflatoxin (Total) 3.79 2.12-5.46
4.2 Work protocols
4.2.1 Cross-reactivity
4.2.1.1 Study design
The stock solutions of AFB1, AFB2, AFG1, AFG2, OTA, Fumonisin B1, sterigmatocystin and cyclopiazonic acid
were diluted using the appropriate extraction solvent (buffer) of the kit to obtain a calibration curve. This calibration
curve has an equal number of points with comparable concentrations as the kit calibrants. Dose-response curves
were constructed by plotting the theoretical concentration against the relative absorbance (B/Bo).
4-30 Experimental evaluation | CODA-CERVA

Concentrations on the abscissa are given in ng/g of components in the cereal, not the concentration in the solvent.
This permits a comparison between different kits as they do not always use the same standard concentration.
4.2.1.2 Data handling and statistical analyses
4.2.1.2.1 ELISA kits

Dose-response curves were constructed by plotting the theoretical concentration against the relative absorbance
(B/Bo). From this calibration curve the concentration at which 50% reduction in signal intensity occurs is calculated.
In a typical log versus logit curve of an ELISA kit this corresponds to the intercept. The relative cross-reactivity in %
(CR%) can be calculated as follows:
Equation 4 CR% = IC50%[ AFB1] / IC50%[component] * 100
With:
IC50% [AFB1]: concentration of AFB1 required for a 50% signal reduction of the zero standard
IC50%[COMPONENT]: concentration of specified component required for a 50% signal reduction of the zero standard
4.2.1.2.2 LFD
As the Charm LFD reader generates directly readings in concentration (ng/g), a different approach was used.
Calibration curves were constructed, the cross-reactivity was calculated by comparing the slopes.
Equation 5 CR% = Slope [COMPONENT] / Slope [AFB1] * 100
4.2.2 Accuracy
Although the determination of accuracy and precision is not demanded by 657/2002 it is our understanding from
end-users there is an interest in comparing the performance of immuno-assays and other tests.
Contaminated matrices maize (TO4138) and rice (T04151) from FAPAS as summarized in Table 17 were used to
assess the accuracy by analyzing 6 samples.
The following formula was used56.
Equation 6 xe-xt= (t*sd) / √N
With:
xe: experimental mean (ng/g)
56 Skoog, West & Holler, Fundamentals of Analytical chemistry, 7th Edition

xt: true value (ng/g)

t: statistic value at 95% confidence level for levels of probability
N: number of replicate measurements
4.2.3 Precision
On three different days 6 aliquots of corn and rice were spiked at the MRL (e.g. 3.1 ng/g) and analyzed. So in total
18 samples were analyzed per kit over the course of three days.
Data were compared using ANOVA with the data analysis function of Excel. In its simplest form ANOVA provides
a statistical test of whether or not the means of several groups are all equal, and therefore generalizes t-test to more
than two groups. Doing multiple two-sample t-tests would result in an increased chance of committing a type I error.
Data are represented using box plots. For this reason, ANOVAs are useful in comparing two, three or more means.
More information on ANOVA and box plots can be found in ref. 57.
4.2.4 Detection capability (CCβ)
For screening tests the β error (i.e. false compliant rate) should be < 5% 58. In the case of substances with an
established regulatory limit, CCβ is the concentration at which the method is able to detect permitted limit
concentrations with a statistical certainty of 1–β. In this case, CCβ must be less than or equal to the Regulatory Limit.
The protocol followed is described in: “Guidelines for the validation of screening methods for residues of veterinary
medicines”58.
In short the following procedure is used: a blank matrix (corn or rice) was spiked 20 times at 0 * MRL, 20 times at
0.5 * MRL, 20 times at 1 * MRL and 2 times 1.5 * MRL and 2 times 2* MRL (the latter two were included to obtain
an indication of the cut-off level if the kit failed for the 3 former levels . Then the lowest response in the spiked
57J. Miller, J. Miller, Statistics and Chemometrics for Analytical Chemistry, Pearson Education, 6th edition, 2010
58Guidelines for the validation of screening methods for residues of veterinary medicines (Initial validation and transfer), Community Reference
Laboratories Residues (CRLs), 2010.

samples was selected. This is defined as the Cut-Off Level, provided that the lowest response for the spiked samples
does not overlap with the highest response for the blank samples.
4.2.5 Recovery
For biochemical tests (e.g. ELISA), which can bind several analytes with varying cross-reactivity’s, initial validation
must be sufficient to demonstrate that all of the analytes in question (included in the scope of the method) will be
reliably extracted (if necessary) and detected.
Samples spiked at different levels (0 MRL, 0.5 MRL, 1 MRL, 1.5 MRL, 2 MRL), also used in the above described
experiment), covering the expected working range of the kit, will be extracted (amount as indicated within the kit
protocol), one in corn and one in rice.
Table 18: Recovery according to 401/2006
Criterion Concentration Recommended value
< 1,0 μg/kg 50 t/m 120 %

Recovery aflatoxins
1-10 μg/kg 70 t/m 110 %
(AFB1, AFB2, AFG1, AFG2)
> 10 μg/kg 80 t/m 110 %
The recovery is calculated is graphically depicted in Figure 4. In short, recovery was calculated as the slope of the
spiked concentration to the detected concentration, described by Vogelgesang et al.59 and should preferably between
70 and 110% for all components as set by Commission Regulation EC/401/20065. If the method does not suffer
from interference the intercept should not be distinguishable from 0. To keep the uncertainty on both intercept and
slope at 5% the bonferroni joint confidence interval was set at 1-α/2 with α = 5%60.
59J. Vogelgesang, J. Hädrich, Accredit. Qual. Assur., 3, (1998) 242-255

60 J. Neter, M. Kutner, W. Wasserman, C. Nachtsheim, J. Neter, Applied Linear Statistical Models, McGraw-Hill Irwin, 4th
edition, 1996

Figure 4: Principle of recovery
4.2.6 Kit protocol
In Table 19, the extraction protocols of the four kits are schematised.
Table 19: Extraction protocols

Charm Neogen R-Biopharm Romerlabs Tecna
Sample (g) 10 10 5 20 5
Extractant 70% MeOH 70% MeOH 70% MeOH 70% MeOH 70% MeOH
Volume extractant 20 50 25 100 25

(ml)
Other NR NR NR NR NaCl
consumables
Clean-up Filtration (0.45µm) centrifugation centrifugation centrifugation centrifugation
Heater block specific to strip, 4 strips

Incubator NR NR NR NR
size, timer, 10 min, 45°C
450nm with 630nm
Reader ROSA-M Reader 650 nm 450 nm 450 nm
differential filter
Extraction shaker shaker shaker shaker shaker

4.3 Results
4.3.1 Accuracy
The graphical representation of the accuracy results is given in Figure 5. The results for the accuracy are represented
in Figure 6 & Figure 7. The results from the kits are represented by the blue spots; the middle blue spot represents
the average value obtained (for 9 points at three different days). The lower and upper blue spot represent the
studentized confidence limits as described in 4.2.2.2. The solid red line represents the indicated reference value of the
FAPAS quality control test material, the dotted red lines represent the satisfactory range as indicated by FAPAS (see
Table 17). Both figures are divided in a left (for AFB1 dedicated kits, the red lines are for aflatoxin B1 solely in the
quality control test material) and a right part (total aflatoxin kits, the red lines represent the certified values for total
aflatoxin).
Figure 5: Principle of accuracy representation
The kits dedicated to AFB1 generated results that lay within the acceptable range indicated by FAPAS There seems
to be a tendency to slightly overestimate the AFB1 content; this could be caused by some cross-reactivity with the
other three aflatoxins present in the FAPAS quality control material . Only the kit of Neogen underestimates slightly
the referenced value.

Figure 6: Accuracy results for corn (T04138), results in ng/g
Figure 7: Accuracy results for rice (T04151), results in ng/g

4.3.2 Precision
Samples were analyzed as described in 4.2.3, the results are summarized in Figure 8 to Figure 12. ANOVA was used
to verify whether the means generated in each day can be considered as equal e.g. given the experimental variation
there is no statistical sound reason at a preset confidence level (5%) to assume the means are not equal (p<0.05).
Table 20 summarizes the results of this exercise and confirms what a visual inspection of the boxplots indicates
namely that, except for the Tecna kit, the inter-day variability is large enough to generate statistically significant day
to day variations. The LFD from Charm also generated p-values > 0.05 (= means are statistically indistinguishable).
When looking at the boxplots of the LFD (Figure 8) it is clear that the intra-day variability is relatively high, making
it impossible for the ANOVA test to detect inter-day differences. It is possible that this variability of the Charm kit is
due to the fact that the dedicated LFD reader only generates integer digits and at the low levels used here this will
easily result in a high apparent variability (see also discussion 4.3.4). Note also that in Figure 8 the kit generated
almost consistently 4 as result for rice on day 3 hence there seems to be no boxplot.
Table 20: Results (p-values) of ANOVA test
Corn Rice
Charm > 0.05 > 0.05
R-Biopharm < 0.05 < 0.05
Neogen < 0.05 < 0.05
Romerlabs < 0.05 < 0.05
Tecna > 0.05

n 4
g
/
g 3
0
Corn 1 Corn 2 Corn 3 Rice 1 Rice 2 Rice 3
Min Outlier Max Outlier
Labels Corn 1 Corn 2 Corn 3 Rice 1 Rice 2 Rice 3

Min 2 2 2 2 2 2
Q1 3 3 2 4 3 4
Median 3.5 3 2.5 4.5 3 4
Q3 4.75 3.75 3.75 5.75 3.75 4
Max 5 5 5 6 5 4
IQR 1.75 0.75 1.75 1.75 0.75 0
Upper Outliers 0 1 0 0 1 0
Lower Outliers 0 0 0 0 0 1
Figure 8: Precision results for Charm, results in ng/g
3.50
3.00
2.50
n 2.00
g
/
g 1.50
1.00
0.50
0.00

Min 2.31 2.79 2.19 1.44 2.08 1.81
Q1 2.57 3.11 2.55 1.93 2.36 2.01
Median 2.72 3.12 2.74 2.05 2.46 2.05
Q3 2.82 3.17 2.98 2.13 2.70 2.06
Max 2.91 3.19 3.05 2.24 2.92 2.22
IQR 0.25 0.06 0.43 0.20 0.33 0.04
Upper Outliers 0.00 0.00 0.00 0.00 0.00 1.00
Lower Outliers 0.00 1.00 0.00 1.00 0.00 1.00
Figure 9: Precision results for Neogen, results in ng/g

6.00
5.00
4.00
n
g
3.00
/
g
2.00
1.00
0.00

Min 2.18 2.94 2.70 2.80 3.26 2.93
Q1 2.23 2.99 2.93 3.22 3.63 3.97
Median 2.26 3.15 3.18 3.40 3.64 4.64
Q3 2.39 3.33 3.30 3.73 3.67 4.89
Max 2.57 3.57 3.42 4.00 3.87 5.46
IQR 0.17 0.33 0.37 0.51 0.03 0.91
Upper Outliers 0.00 0.00 0.00 0.00 1.00 0.00
Lower Outliers 0.00 0.00 0.00 0.00 1.00 0.00
Figure 10: Precision results for R-Biopharm, results in ng/g
7.00
6.00
5.00
n 4.00
g
/
g 3.00
2.00
1.00
0.00

Min 2.97 2.60 4.41 3.22 2.91 2.93
Q1 3.42 3.01 4.77 3.71 3.04 3.23
Median 3.50 3.12 4.97 4.30 3.10 3.33
Q3 3.58 3.31 5.19 4.49 3.30 3.43
Max 3.64 3.54 6.44 4.60 4.17 3.47
IQR 0.16 0.30 0.42 0.79 0.26 0.20
Upper Outliers 0.00 0.00 1.00 0.00 1.00 0.00
Lower Outliers 1.00 0.00 0.00 0.00 0.00 1.00
Figure 11: Precision results for Romerlabs, results in ng/g

6.00
5.00
4.00
n
g
3.00
/
g
2.00
1.00
0.00
Corn 1 Corn 2 Corn 3
Labels Corn 1 Corn 2 Corn 3

Min 3.61 3.76 3.25
Q1 3.64 3.82 4.08
Median 3.79 3.91 4.34
Q3 4.00 3.95 4.44
Max 4.06 4.62 4.86
IQR 0.36 0.13 0.36
Upper Outliers 0.00 1.00 0.00
Lower Outliers 0.00 0.00 1.00
Figure 12: Precision results for Tecna, results in ng/g
4.3.3 Recovery
The results of the recovery experiments are summarized in Figure 13 to Figure 16. The x-axis represents the kit
identity, the y-axis the value of the intercept (Figure 13 & Figure 14) or slope (Figure 15 & Figure 16). The vertical
lines belonging to each kit represents the 97.5% bonferroni confidence limit. For the intercept this should enclose 0,
the slope should lie between 70% and 110%.
From Figure 13 & Figure 14 it can be seen that most kits don’t seem to suffer from any noticeable matrix effect, only
the kit of Romerlabs gave a slight increased response for a blank. When looking at the slope the kit of Tecna gave a
higher recovery for corn while the kit of R-Biopharm gave an increased recovery for rice. The higher recovery of
Tecna might be explained by the fact that we centrifuged the samples before applying them to the kit instead of
filtering them, which often gives cleaner extracts, as the manufacturer recommends in the manual. This decision was
made as we had to process 64 samples on one day do determine both the recovery and the CCβ and this turned out
to be not feasible with filtration.
From a food safety point of view, false negative results due to a low recovery are not acceptable. Nearly all kits fulfill
this central requirement and are therefore fit-for-purpose.

Figure 13: Intercept for corn
Figure 14: Intercept for rice

Figure 15: Slope for corn
Figure 16: Slope for rice

4.3.4 Detection capability (CCβ)
For an immuno-assay to be used routinely as a high throughput screening method it is imperative that it can
differentiate in a sufficiently reliable (e.g. β<0.05) manner between blank samples and samples that are considered to
be positive. No specific protocol exists for mycotoxin dedicated immuno-assays, ref 58 gives a Guideline for the
determination of CCβ of screening methods for residues of veterinary medicines and was deemed a good starting
point to determine the CCβ for mycotoxin tests. In short a blank matrix has to be spiked at 0.5 * MRL (or if the kit
doesn’t achieve this level, at 1 * MRL) and analysed in 20-fold together with a 20-fold analysis of the blank matrix. If
the lowest signal of the spiked sample is lower than the highest level of the blank it has to be concluded that the kits
is unable to differentiate between a negative and a positive sample. Figure 17 to Figure 25 represent the obtained
data and are summarised in Table 21. The kit of R-Biopharm & Neogen differentiates 0 and 1.6 ng/g in both rice
and corn, Tecna detects reliably 1.6 ng/g in corn, Romerlabs 1.6 in corn and 3.1 in rice. The kit of Charm works well
in rice but struggles with corn. As stated earlier it is our opinion that this is due to some part to the fact this reader
only generates integer digits with possible increased errors due to rounding.
Table 21: Detection capability (CCβ)

Corn (ng/g) Rice (ng/g)
Charm 4.7 3.1
Neogen 1.6 1.6
R-Biopharm 1.6 1.6
Romerlabs 1.6 3.1
Tecna 1.6 NA
Figure 17: CCβ for Charm in corn

Figure 18: CCβ for Charm in rice
Figure 19: CCβ for R-Biopharm in corn

Figure 20: CCβ for R-Biopharm in rice
Figure 21: CCβ for Neogen in corn

Figure 22: CCβ for Neogen in rice
Figure 23: CCβ for Romerlabs in corn

Figure 24: CCβ for Romerlabs in rice
Figure 25: CCβ for Tecna in corn

4.3.5 Cross-reactivity
The cross-reactivity (specificity) of the anti-bodies for other possible interferents that are often found in samples
contaminated with aflatoxins was checked using solvent based calibration curves. The results are represented in
Figure 26 to Figure 30 and are summarized in Table 22. The x-axis represents the amount of analyte in equivalents
ng/g solid sample while the y-axis represents the results indicated by the FLD reader (Figure 26) or the standardized
dose response for the ELISA kits. The result for FB1 is not indicated on the graphs as this component was spiked at
a level 50 times higher than that of AFB1 (which one could expect in real life), for all kits no cross-reactivity was
detected.
None of the kits reacted to any of the components other than the aflatoxins, besides AFB1 all kits reacted to the
other 3 aflatoxins which given their structural similarity might be expected. The discriminating capability for the
different aflatoxins between AFB1 dedicated and total aflatoxin kits is not that large.
Table 22: Cross-reactivity for 4 aflatoxins

R-Biopharm Neogen Romerlabs Tecna Charm
AFB1 100% 100% 100% 100% 100%
AFB2 55% 64% 74% 37% 50%
AFG1 81% 76% 64% 51% 33%
AFG2 35% 68% 40% 30% 17%
Figure 26: Cross-reactivity for Charm

Figure 27: Cross-reactivity for R-Biopharm
Figure 28: Cross-reactivity for Neogen

Figure 29: Cross-reactivity for Romerlabs
Figure 30: Cross-reactivity for Tecna

4.4 Conclusions of the experimental evaluation
The results we obtained concord well with the data kindly provided by the kit manufacturers. The cross-reactivity,
recovery & detection capability experiments clearly show that these tests possess the selectivity and sensitivity to
adequately screen samples and differentiate positive from negative samples. Reassuringly, all extraction protocols
(often a significant source of error) are clearly adequate for the matrices investigated here. The day to day variability
on the other hand is somewhat troublesome, a point that was also discussed in the administrative evaluation.

5. Conclusion and future
prospects
If a general conclusion has to be drawn from this exercise it essentially doesn’t differ from the conclusions drawn the
past two years for the evaluation of DON and OTA kits: immuno-assays are very powerful screening tools to assess
a large number of samples in a very short timescale with a minimum of effort, costly labour hours or expensive lab
equipment. These kits, especially the strip type kits, are the preferred tool for in-field screening.
It is advisable to analyse samples flagged by immune-assays as positive by a more reliable (but much more expensive)
reference method.
The 4 ELISA kits are very competitive in terms of detection capability and sensitivity. The LFD of Charm struggled
a little bit more with corn but its strong point lies in its user friendliness: no calibration curve is needed which makes
it a true “field” test.
CODA-CERVA | Conclusion and future prospects 5-55

2012-05-04 Evaluation Immunoassay Kits Aflatoxin PDF

Uploaded by

Copyright:

Available Formats

2012-05-04 Evaluation Immunoassay Kits Aflatoxin PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2012-05-04 Evaluation Immunoassay Kits Aflatoxin PDF

Uploaded by

Copyright:

Available Formats

CODA-CERVA

Evaluation of immunoassay kits for quantitative determination of ochratoxin in cereals

CHEMICAL SAFETY OF THE FOOD CHAIN

UNIT OF TOXINS AND NATURAL COMPONENTS

NATIONAL REFERENCE LABORATORY ON MYCOTOXINS

Correspondance to: Dr. Ir. Bart Huybrechts

Tel: +32 02 769 22 76

1. Foreword & scope & background ............................................................................... 1-1

Figure 1: Partnership ................................................................................................................................................................................................................. 1-1

Table 1: Parameters for quantitative and qualitative kits ...................................................................................................................................................... 1-2

ACN : Acetonitril LFD : Lateral Flow Device

CODA-CERVA | Foreword & scope & background 1-1

1.2.1 Objective and strategy of the inventory task

1.2.2 Objective and strategy of the administrative evaluation

2 E. Trullols Trends in Anal. Chem. 23, 2004, 137

1-2 Foreword & scope & background | CODA-CERVA

CODA-CERVA | Foreword & scope & background 1-3

Table 2: Regulatory limits EU (ref. 3-4-6)

Food product Mycotoxin Max Level

1.2.3 Objective and strategy of the experimental evaluation

1-4 Foreword & scope & background | CODA-CERVA

Table 3: Criteria to be verified according to 657/20028

CCβ CCα Accuracy Recovery Precision Specificity Robustness

18 J. W. Bennett and M. Klich, Clin. Microbiolog. Rev., 2003, 16, 497.

contaminants in foodstuffs, Official J. Eur. Commun. L 201/43 (1998) 43.

CODA-CERVA | Foreword & scope & background 1-5

Figure 2: Structure of AFB1, AFB2, AFG1 & AFG2

Agriculture Science and Technology. Ames, IA. 1989.

1-6 Foreword & scope & background | CODA-CERVA

Good sensitivity, good selectivity, good

Cross-reactivity with related mycotoxins, extract clean-up

30 S. Biselli, L. Hartig, H. Wegener, C. Hummert, LC–GC Eur. 17 (2004) 25.

CODA-CERVA | Foreword & scope & background 1-7

Lateral flow devices

Chapter 19, pp 314–319

1-8 Foreword & scope & background | CODA-CERVA

CODA-CERVA | Foreword & scope & background 1-9

2.1 Inventory results of providers of aflatoxin kits

Qualitative - Quantitative FLUO/COLUMN, Type test

CODA-CERVA | Inventory 2-13

Table 5: Overview of contacted manufacturers and their kits

2-14 Inventory | CODA-CERVA

Table 6: Overview of kits involved in the experimental evaluation

CODA-CERVA | Inventory 2-15

3.1 Analytical performance criteria

Table 7: Calculation of parameters (for symbols, see list abbreviations)

Recovery Equation 1 Rec% = (xe/xt)*100

Uncertainty Equation 2 u= √(CV2+bias2)

Extended Uncertainty Equation 3 U(Ext)= k (=2) * u

3.2 Intra-run precision, between-run precision & reproducibility

CODA-CERVA | Administrative evaluation 3-19

Table 8: Intra-run precision

Level (ng/g) 5.00 5.00 10.00 2.00

Extended Uncertainty (%, k=2) 42 65 52 57 17 27 39 23 29

Level (ng/g) 10.00 15.00 40.00 5.00

Extended Uncertainty (%, k=2) 45 44 28 54 46 34 70 35 27

Level (ng/g) 20.00 50.00 160.00 10.00

Extended Uncertainty (%, k=2) 27 37 31 38 62 44 63 43 33

Table 9: Between-run precision