Summary Report TEMPO EC

Transcription

1 ACCREDITATION N -044 PORTEE DISPONIBLE SUR NF VALIDATION Validation of alternative analytical methods Application in food microbiology Summary Report Validation study according to the NF EN ISO 640-:06 TEMPO EC (Certificate reference No: BIO /3-0/05) for the enumeration of glucuronidase-positive Escherichia coli in all human food products and pet foods Expert Laboratory: For: ADRIA Développement ZA Creac h 996 QUIMPER Cedex (France) biomérieux Chemin de l Orme F-6980 MARCY L ETOILE (France) This report consists of 69 pages, including 8 appendices. Only copies including the totality of this report are authorised. Competencies of the laboratory are certified by COFRAC accreditation for the analyses marked with the symbol. Version 0 April 5, 07 ADRIA DEVELOPPEMENT Creac h Gwen - F. 996 QUIMPER Cedex - Tél. (33) Fax (33) adria.developpement@adria.tm.fr - Site web : ASSOCIATION LOI DE 90 - N SIRET N EXISTENCE N TVA FR

2 INTRODUCTION... 4 METHOD PROTOCOLS Alternative method Reference method INITIAL VALIDATION STUDY AND EXTENSION/RENEWAL STUDIES: RESULTS Method comparison study Relative trueness study Accuracy profile study Inclusivity Exclusivity Limit of quantification (LOQ) Practicability Inter-laboratory study Study organisation Experimental parameters control Analysis results Calculations and interpretation Conclusion...30 Appendix - Artificial contamination 3 Appendix Trueness study: raw data (Initial validation study and extension studies) 37 Appendix 3 - Relative trueness study: statistical calculations 46 Appendix 4 - Accuracy study: raw data 50 Appendix 5 - Accuracy study: summarized results 56 Appendix 6 - Inclusivity / Exclusivity: raw data (initial validation, and extension study, 06) 58 Appendix 7 - Limit of quantification (LOQ): raw data 6 Appendix 8 - Inter-laboratory study results (initial validation, 005) (in French) 63 ADRIA Développement /69 April 5, 07

3 Quality Assurance documents related to this study can be consulted upon request from biomérieux. The technical protocol and the result interpretation were realised according to the EN ISO 640-:06 and the AFNOR technical rules (revision 5). Validation protocols NF EN ISO 640- (June 06) : Microbiology of the food chain - Method validation Part : Vocabulary Part : Protocol for the validation of alternative (proprietary) methods against a reference method AFNOR Technical Rules (Revision n 5) Reference method NF EN ISO (July 00): Horizontal method for the enumeration of glucuronidase positive Escherichia coli Part : colony-count technique at 44 C using 5-bromo-4-chloro-3 indoly -D-glucuronate Alternative method TEMPO EC Scope All human food and pet food Certification organism AFNOR Certification ( Analyses performed according to the COFRAC accreditation ADRIA Développement 3/69 April 5, 07

4 INTRODUCTION The TEMPO EC method was validated in February 005 for human food and pet food (Certificate number BIO /3-0/05) according to the ISO 640 (003) standard. In May 007: Extension validation for a specific protocol for milk s including the secondary diluents MOPS buffer instead of distilled water. In January 009: Renewal of the validation. In February 0: Extension Study to extend the scope of the validated TEMPO EC method to raw milks. An accuracy study was run. In November 0: Renewal of the validation. In January 07: Renewal of the validation according to the NF EN ISO 640-:06. METHOD PROTOCOLS. Alternative method Principle TEMPO system is an automated method associating an innovative card with an adapted medium to ensure rapid enumeration of several quality indicators. The method is based on the MPN principle (Most Probable Number), with the TEMPO card. The TEMPO EC test consists in a culture medium associated to the TEMPO card. The culture medium is inoculated with the sample to be tested and transferred by the TEMPO Filter into the card. ADRIA Développement 4/69 April 5, 07

5 Depending on the number and size of the positive wells, the TEMPO Reader calculates the number of E. coli. Card reading, interpretation and reporting are directly managed by the TEMPO system. Protocol The protocol is presented below (See figure ). Figure - Flow diagram of the alternative method Sample /0 dilution in a TEMPO stomacher bag Diluent: peptone-salt /40 dilution /400 dilution: 3 ml sterile distilled water () 3.9 ml sterile distilled water () 3 ml MOPS 0.4 M ph 7.5 for milk s 3.9 ml MOPS 0.4 M ph 7.5 for milk s in a TEMPO EC vial in a TEMPO EC vial ml suspension /0 in vial 0. ml dilution /0 in vial (one card) (one card) Incubation 4h - 7h at 37 C C Reading The /40 dilution allows 0 to cfu/g enumeration The /400 dilution allows 00 to cfu/g enumeration () Dilution in distilled water for all the products Reference of the kit insert since the last validation study The reference of the current kit insert is Ref L. No modification has been done since the last renewal study. The version changes have been done to follow new regulatory requirements (ex. REACH), and to harmonize with others package inserts. ADRIA Développement 5/69 April 5, 07

6 Restrictions for use Using the TEMPO EC method is not recommended to analyse the products with a high enzymatic activity. This restriction concerns particularly the raw mollusc and red offals, which needs a dilution rate >/40.. Reference method The standard method is the ISO (July 00): Horizontal method for the enumeration of glucuronidase positive Escherichia coli Part : colonycount technique at 44 C using 5-bromo-4-chloro-3 indoly -D-glucuronate. The protocol of the reference method is given below (See figure ). Figure - Flow diagram of the ISO (July 00) reference method Primary dilution and dilutions in peptone-salt ml per Petri dish Pour plating with TBX medium Injured micro-organisms Non injured micro-organisms Incubation 4 h at 37 C C Incubation 8-4 h at 44 C C Incubation 8-4 h at 44 C C Enumeration of characteristic colonies Analysis performed according to the COFRAC accreditation Note that for the Validation Study, Petri dishes were used per dilution for the analyses run in 004 and 006 ADRIA Développement 6/69 April 5, 07

7 3 INITIAL VALIDATION STUDY AND EXTENSION/RENEWAL STUDIES: RESULTS 3. Method comparison study The method comparison study is the part of the validation process that is performed in the organizing laboratory. It consists of four parts: - A comparative study of the results of the reference method to the results of the alternative method in a variety of different items (naturally and/or artificially) contaminated samples (so-called relative trueness study); - A comparative study of the results of the reference method to the results of the alternative method in artificially contaminated samples using replicates of a single item per category. The data are analysed using the accuracy profile (AP) approach (so-called AP study); - A limit of quantification (LOQ) study of the results of the alternative method in artificially contaminated samples using replicates of a single item per category. The data are used to calculate the LOQ of the alternative method. This study is only done for instrumentallybased methods (i.e. methods that are not based on the counting of individual colonies); - An inclusivity/exclusivity study of the alternative method. 3.. Relative trueness study Degree of correspondence between the response obtained by the reference method and the response obtained by the alternative method on identical samples 3... Artificial and natural contamination of the samples Naturally contaminated samples were analysed preferably. Artificially contaminated samples, using spiking or seeding protocols, were also analysed. The inoculated strains and the contamination protocols are provided in Appendix. Among the interpretable results, 64 samples were naturally contaminated and 77 samples were artificially contaminated. Note that all the artificially contaminated samples are indicated in bold characters all among the report. ADRIA Développement 7/69 April 5, 07

8 3... Number and nature of the samples The following samples were analysed: - 73 in 004: initial validation study, - 30 in 006: extension study for a specific protocol dedicated to milk s; - 9 in 00: extension study for raw milks, - 5 in 0: data for biomérieux (complementary data: meat products and composite foods) - 6 in 06: renewal study. In 004, 006 and 00, the analyses were realised in duplicate according to the ISO 640 (003). Raw data are provided in Appendix. A total of 90 samples were analysed providing 4 interpretable results. The number of samples tested per category and type is presented in table. Table - Number and nature of the samples Category Type Products Number of samples Number of samples with analysed interpretable results a Ready to eat 0 9 : RTE & b Ready to reheat 9 7 RTRH c Cooked delicatessen 5 5 Total 4 a Raw meat, except poultry meat 7 7 : Meat b Raw poultry meat c Delicatessen (to cook) 0 Total 9 8 a Cheeses and ice creams 8 6 3: Dairy b Milk s 30 3 c Raw milk 9 0 Total a Raw 3 5 4: Seafood b Reday to reheat 5 5 c Smoked 5 5 Total 3 5 a Raw, non processed 5 5:Fruits and b Ready to eat 6 0 Vegetables c Ready to reheat 8 Total 40 3 a Raw 5 5 6:Pet food b Sausages, pâtés 6 5 c Pellets 6 5 Total 7 5 TOTAL 90 4 ADRIA Développement 8/69 April 5, 07

9 3...3 Raw data The samples were analysed by the reference and the alternative methods in order to have 5 interpretable results per category, and 5 interpretable results per tested type. The data are classified in three categories (See Table ): - Interpretable results by the two methods; - Results with less than 4 colonies per plate with one of the two methods (indicated with * in the data) in order to have the more precise result. These results were not kept for calculation. - Results below or above the quantification limit. Category Type Products : RTE & RTRH : Meat 3: Dairy 4: Seafood 5:Fruits and Vegetables 6:Pet food Table Number of samples analysed Number of samples with <4 CFU/plate Number of samples below or above the detection limit Number of samples with no result Number of samples with interpretable results a Ready to eat b Ready to reheat c Cooked delicatessen Total a Raw meat, except poultry meat b Raw poultry meat c Delicatessen (to cook) Total a Cheeses and ice creams b Milk s c Raw milk Total a Raw b Ready to reheat c Smoked Total a Raw, non processed b Ready to eat c Ready to reheat Total a Raw b Sausages, pâtés c Pellets Total TOTAL The samples, which were not used in the calculations, are provided in table 3. ADRIA Développement 9/69 April 5, 07

10 Table 3 - Samples, which were not used in the calculations Sample Reference method: Alternative method: Product N ISO TEMPO EC Category Type Smoked salmon <.00 <.00 a 60 Ready to reheat food <.00 <.00 b 06 Ready to eat Chinese food <.00 <.00 b 865 Frozen sausage.00* <.00 c 93 Raw cow milk cheese.39*.88 3 a 4 Ice cream (red fruits) < a 509 Infant formula milk <.00 <.00 3 b 50 Infant formula milk 0.85* <.00 3 b 5 Infant formula milk.39 <.00 3 b 5 Infant formula milk 3.45 No result 3 b 504 Infant formula milk.05*.00 3 b 546 Infant formula milk.39*.00 3 b 547 Infant formula milk.53*.6 3 b 477 Raw cow milk 0.70* 0.70* 3 c 5055 Raw cow milk 0.70* 0.70* 3 c 476 Raw cow milk <.00 <.00 3 c 493 Raw cow milk < c 494 Raw cow milk < c 5056 Raw cow milk <.00 <.00 3 c 5 Raw goat milk <.00 <.00 3 c 53 Raw goat milk <.00 <.00 3 c 54 Raw goat milk <.00 <.00 3 c 00 Raw Clams <.00 No result 4 a 0 Shells <.00 <.00 4 a Raw Oysters <.00 <.00 4 a Sardines <.00 <.00 4 a 3 Trout <.00 <.00 4 a 56 Raw sea almonds <.00 <.00 4 a 57 Raw clams <.00 <.00 4 a 58 Raw oysters <.00 <.00 4 a 8 Cabbage 0.85* <.00 5 a 49 Buckwheat seed <.00 <.00 5 a 59 Carrots <.00 <.00 5 a 6 Green cabbage <.00 <.00 5 a 6 Figs <.00 <.00 5 a 69 Figs <.00 <.00 5 a 8 Cabbage 0.85* <.00 5 a 83 Mixed salad.4*.00 5 b 63 Dried pears <.00 <.00 5 b 67 Dried apricots <.00 <.00 5 b 68 Pistachios <.00 <.00 5 b 8 Grated carrots <.00 <.00 5 b 93 Dried apple < b 96 Ready to reheat vegetables <.00 <.00 5 c 97 Ready to reheat vegetables <.00 <.00 5 c 0 Ready to reheat vegetables <.00 <.00 5 c 70 Dried pears <.00 <.00 5 c 705 Sausage for dog.00 <.00 6 b 706 Pellets for cats.96 <.00 6 c * : < 4 colonies/plate Analyses performed according to the COFRAC accreditation ADRIA Développement 0/69 April 5, 07

11 Alternative method (log CFU/g) biomérieux Statistical interpretation The calculations are provided in Appendix 3. The obtained data were analysed using the scatter plot. The graphs are provided with the line of identity (y = x): The Figure 3 shows the data plotted for Ready to eat and ready to reheat; - The Figure 4 shows the data plotted for Meat products; - The Figure 5 shows the data plotted for Dairy products; - The Figure 6 shows the data plotted for Seafood products; - The Figure 7 shows the data plotted for Fruits and vegetables; - The Figure 8 shows the data plotted for Pet food; - The Figure 9 shows the data plotted for all the products. Figure 3 - Data plotted for Ready to eat and ready to reheat 8,00 Cat : Ready to eat, ready to reheat 7,00 6,00 5,00 4,00 3,00,00,00 0,00 0,00,00,00 3,00 4,00 5,00 6,00 7,00 8,00 Ref)erence method (log CFU/g) Type a Type b Type c corrected values <4 colonies/plate ADRIA Développement /69 April 5, 07

12 Alternative method (log CFU/g) Alternative method (log CFU/g) biomérieux Figure 4- Data plotted for Meat products Cat : Meat products 8,00 6,00 4,00,00 Type a Type b Type c corrected values <4 colonies/plate 0,00 0,00,00 4,00 6,00 8,00 Ref)erence method (log CFU/g) Figure 5- Data plotted for Dairy products 8 Cat 3 : Dairy products 6 4 Type a Type b Type c corrected values <4 colonies/plate 0 0,00,00 4,00 6,00 8,00 Ref)erence method (log CFU/g) ADRIA Développement /69 April 5, 07

13 Alternative method (log CFU/g) Alternative method (log CFU/g) biomérieux Figure 6- Data plotted for Seafood products Cat 4 : Seafood products 8,00 6,00 4,00,00 Type a Type b Type c corrected values <4 colonies/plate 0,00 0,00,00 4,00 6,00 8,00 Ref)erence method (log CFU/g) Figure 7- Data plotted for Fruits and vegetables 8,00 Cat 5 : Fruits and vegetables 6,00 4,00,00 Type a Type b Type c corrected values <4 colonies/plate 0,00 0,00,00 4,00 6,00 8,00 Ref)erence method (log CFU/g) ADRIA Développement 3/69 April 5, 07

14 Alternative method (log cfu/g)) Alternative method (log CFU/g) biomérieux Figure 8- Data plotted for Pet food Cat 6 : Pet food 8,00 6,00 4,00,00 Type a Type b Type c corrected values <4 colonies/plate 0,00 0,00,00 4,00 6,00 8,00 Ref)erence method (log CFU/g) Figure 9 - Data plotted for all the products All categories 0,00 8,00 6,00 4,00,00 0,00 0,00,00 4,00 6,00 8,00 0,00 reference method (log cfu/g) Cat Cat Cat 3 Cat 4 Cat 5 Corrected values < 4 colonies /plate y=x Cat 6 ADRIA Développement 4/69 April 5, 07

15 Difference biomérieux The calculated values are provided in Table 4 per category. Table 4 - Calculated values per category Category n SD 95% lower limit 95% upper limit / / / / / / / / / / / / All categories : Average difference SD: Standard deviation of differences The Bland-Altman difference plot for all the samples is given Figure 0. Figure 0 Bland-Altman difference plot for all the samples Bland-Altman -All categories 3,00,00,00 0,00 -,00 -,00-3,00 0,00,00 4,00 6,00 8,00 Mean log CFU/g Cat Cat Cat 3 Cat 4 Cat 5 Corrected values <4 colonies/plate Linear (95% lower limit) Linear (95 upper limit) Linear (bias) Cat 6 Samples for which the difference between the result observed with the reference and the alternative methods is above or lower than the limits are listed in the Table 5. ADRIA Développement 5/69 April 5, 07

16 Table 5 - Disagreements observed between the reference and the alternative method Category Type Reference method Alternative method N Product Sample Corrected True Corrected True values values values values Mean Difference c 369 Cooked ham / 3.6 / c 707 Pellets for cats /.75 / a 94 Raw goat milk cheese /. / a 6 Ground beef for animals /.65 / b 7 Sausage for dog /.3 / b 73 Pâté for dog / 3.63 / b 74 Pâté for dog / 3.9 / c 865 Frozen sausages / < b 50 Infant formula / < b 5 Infant formula / < a 8 Cabbage / < b 705 Sausage for dog / < c 706 Pellets for cats / < a 4 Ice cream (red fruits) 0.00 <.00 / a 0 Raw shells 0.00 <.0.00 < a Raw oysters 0.00 < < a 3 Trout 0.00 < < a 56 Raw sea almonds 0.00 < < a 57 Clams 0.00 < < b 93 Dried apples 0.00 <.00 / For a majority of samples (3), the differences observed between the reference and the alternative methods are explained by the fact that one of the two methods is below the detection limit. In the other cases (7 samples concerned), the difference was in favour of the alternative method (5 times) and in favour of the reference method ( times). The differences varied from -.4 to.38 log Conclusion The relative trueness of the alternative method is satisfied. ADRIA Développement 6/69 April 5, 07

17 3.. Accuracy profile study The accuracy profile is a graphical representation of the capacity of measurement of the quantitative method, obtained by combining acceptability intervals and β-expectation tolerance intervals, both reported to different levels of the reference value Matrices Six matrix/strain pairs were tested. A minimum of one type per category, and therefore different batches, were selected, using 6 samples per type. samples were contaminated at a low level, at intermediate level, at a high level. For each sample, 5 replicates (5 different test portions) were tested. In the end, 30 samples were tested per matrix type. The following matrix/strain pairs were studied (See Table 6). Table 6 - Matrix/strain pairs Category Matrix Inoculated Strain Origin. Ready to eat and ready to reheat Pâté E.coli 44 Ready to reheat meal. Meat products Ground beef E.coli 3 Ground beef 3. Dairy products Pasteurized milk E.coli 94 Cheese 4. Fruits and vegetables Grated carrots E.coli 9 Grated carrots 5. Seafood Raw fish filet E.coli Ad8 Fish 6. Pet food Dog pâté E.coli 3 Veal liver Inoculation level CFU/g Calculation and interpretation The raw data are provided in Appendix 4 and the summary tables (in log CFU/g) in Appendix 5. The statistical results and the accuracy profiles are provided in Figure. The calculations were done using the AP Calculation Tool MCS (Clause calculation and interpretation of accuracy profile study) ver available on The following interpretations were done: - Level : /40 Level : /400 when available Level 3: /400 - Level : /40 Level : /4000 Level 3: /4000 ADRIA Développement 7/69 April 5, 07

18 Bias Bias Bias Bias Bias Bias biomérieux Figure Accuracy profile Dilution /40 / /400 (Food) Category RTE (Food) Category Meat products (Food) Type Cooked delicatessen: liver pâté (Food) Type Raw meat: ground beef Cooked delicatessen: liver pâté Raw meat: ground beef 0,80 0,80 0,60 0,60 0,40 0,40 0,0 Bias 0,0 Bias 0,00 0,00,00,00 3,00 4,00 5,00 6,00 β-eti AL = +/ ,00 0,00,00,00 3,00 4,00 5,00 6,00 β-eti AL = +/ ,0-0,0-0,40-0,40-0,60 Reference Median -0,60 Reference Median β-eti β-eti β-eti β-eti Reference compared to compared to Reference compared to compared to Sample Name Bias Lower β-eti Upper β-eti Sample Name Bias Lower β-eti Upper β-eti central value AL=±0.5 final AL Central value AL=±0.5 final AL Acceptable Acceptable Acceptable Acceptable ,48 0,04-0,38 0,0 YES YES ,56 0,36 0,043 0,430 YES YES ,34 0,35-0,044 0,34 YES YES ,54 0, 0,08 0,405 YES YES ,7 0,334 0,55 0,53 NO NO ,77-0,030-0,4 0,63 YES YES ,77 0, -0,058 0,300 YES YES ,76 0,9-0,065 0,3 YES YES ,98 0,46 0,37 0,595 NO NO ,98 0,344 0,5 0,538 NO NO ,94 0,383 0,04 0,56 NO NO ,04 0,357 0,63 0,550 NO NO Reference Alternative SD repeatability of reference Reference Alternative SD repeatability of reference Final AL Final AL method method method <= 0.5 method method method <= 0.5 SD Repeatability 0,05 0,4 YES +/- 0,500 SD Repeatability 0,054 0,34 YES +/- 0,500 (Food) Category Dairy products (Food) Category Seafood (Food) Type Milk: pasteurized milk (Food) Type Raw fish: fish fillet Milk: pasteurized milk Raw fish: fish fillet 0,80 0,80 0,60 0,60 0,40 0,40 0,0 Bias 0,0 Bias 0,00 0,00,00,00 3,00 4,00 5,00 6,00-0,0 β-eti AL = +/ ,00 0,00,00,00 3,00 4,00 5,00 6,00-0,0 β-eti AL = +/ ,40-0,40-0,60 Reference Median -0,60 Reference Median β-eti β-eti β-eti β-eti Reference compared to compared to Reference compared to compared to Sample Name Bias Lower β-eti Upper β-eti Sample Name Bias Lower β-eti Upper β-eti Central value AL=±0.5 final AL Central value AL=±0.5 final AL Acceptable Acceptable Acceptable Acceptable ,3 0,55-0,06 0,335 YES YES ,4 0,06-0,4 0,38 YES YES ,5 0,038-0,43 0,8 YES YES ,45 0,76 0,000 0,35 YES YES ,58 0,53 0,07 0,433 YES YES ,63 0,45-0,03 0,3 YES YES ,70 0,93 0,03 0,374 YES YES ,65-0,00-0,86 0,66 YES YES ,83 0,496 0,36 0,677 NO NO ,76 0,467 0,9 0,643 NO NO ,99 0,44 0,063 0,44 YES YES ,87 0,0 0,034 0,386 YES YES Reference Alternative SD repeatability of reference Reference Alternative SD repeatability of reference Final AL Final AL method method method <= 0.5 method method method <= 0.5 SD Repeatability 0,055 0,5 YES +/- 0,500 SD Repeatability 0,064 0, YES +/- 0,500 (Food) Category Vegetables (Food) Category Pet food (Food) Type RTE: grated carrots (Food) Type Pâtés: dog pâté RTE: grated carrots Pâtés: dog pâté 0,80 0,60 0,60 0,40 0,40 0,0 Bias 0,0 Bias 0,00 0,00,00,00 3,00 4,00 5,00 6,00 β-eti AL = +/ ,00 0,00,00,00 3,00 4,00 5,00 6,00 β-eti AL = +/ ,0-0,0-0,40 é -0,40-0,60 Reference Median -0,60 Reference Median β-eti β-eti β-eti β-eti Reference compared to compared to Reference compared to compared to Sample Name Bias Lower β-eti Upper β-eti Sample Name Bias Lower β-eti Upper β-eti Central value AL=±0.5 final AL Central value AL=±0.5 final AL Acceptable Acceptable Acceptable Acceptable ,34 0,4 0,000 0,48 YES YES ,88 0,6-0,03 0,334 YES YES ,30 0,04-0,7 0,55 YES YES ,90 0,76 0,003 0,349 YES YES ,45 0,93 0,080 0,507 NO NO , 0,08 0,035 0,38 YES YES ,5 0,94-0,00 0,407 YES YES , 0,08 0,035 0,38 YES YES ,69 0,69 0,055 0,483 YES YES ,4 0,53-0,00 0,36 YES YES ,65 0,46 0, 0,640 NO NO ,34 0,348 0,75 0,5 NO NO Reference Alternative SD repeatability of reference Reference Alternative SD repeatability of reference Final AL Final AL method method method <= 0.5 method method method <= 0.5 SD Repeatability 0,076 0,48 YES +/- 0,500 SD Repeatability 0,048 0,0 YES +/- 0,500 ADRIA Développement 8/69 April 5, 07

19 Bias Bias Bias Bias Bias Bias biomérieux Accuracy profile Dilution /40 / /4000 (Food) Category RTE (Food) Category Meat products (Food) Type Cooked delicatessen: liver pâté (Food) Type Raw meat: ground beef Cooked delicatessen: liver pâté Raw meat: ground beef Bias β-eti Bias β-eti -0.0 AL = +/ AL = +/ Reference Median Reference Median β-eti β-eti β-eti β-eti Reference compared to compared to Reference compared to compared to Sample Name Bias Lower β-eti Upper β-eti Sample Name Bias Lower β-eti Upper β-eti central value AL=±0.5 final AL Central value AL=±0.5 final AL Acceptable Acceptable Acceptable Acceptable YES YES YES YES YES YES YES YES NO NO YES YES YES YES YES YES YES YES YES YES YES YES YES YES Reference Alternative SD repeatability of reference Reference Alternative SD repeatability of reference Final AL Final AL method method method <= 0.5 method method method <= 0.5 SD Repeatability YES +/ SD Repeatability YES +/ (Food) Category Dairy products (Food) Category Seafood (Food) Type Milk: pasteurized milk (Food) Type Raw fish: fish fillet Milk: pasteurized milk Raw fish: fish fillet Bias 0.0 Bias β-eti AL = +/ β-eti AL = +/ Reference Median Reference Median β-eti β-eti β-eti β-eti Reference compared to compared to Reference compared to compared to Sample Name Bias Lower β-eti Upper β-eti Sample Name Bias Lower β-eti Upper β-eti Central value AL=±0.5 final AL Central value AL=±0.5 final AL Acceptable Acceptable Acceptable Acceptable YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES NO NO YES YES NO NO YES YES Reference Alternative SD repeatability of reference Reference Alternative SD repeatability of reference Final AL Final AL method method method <= 0.5 method method method <= 0.5 SD Repeatability YES +/ SD Repeatability YES +/ (Food) Category Vegetables (Food) Category Pet food (Food) Type RTE: grated carrots (Food) Type Pâtés: dog pâté RTE: grated carrots Pâtés: dog pâté Bias 0.0 Bias β-eti AL = +/ β-eti AL = +/ é Reference Median Reference Median β-eti β-eti β-eti β-eti Reference compared to compared to Reference compared to compared to Sample Name Bias Lower β-eti Upper β-eti Sample Name Bias Lower β-eti Upper β-eti Central value AL=±0.5 final AL Central value AL=±0.5 final AL Acceptable Acceptable Acceptable Acceptable YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES YES NO NO NO NO Reference Alternative SD repeatability of reference Reference Alternative SD repeatability of reference Final AL Final AL method method method <= 0.5 method method method <= 0.5 SD Repeatability YES +/ SD Repeatability YES +/ ADRIA Développement 9/69 April 5, 07

20 When the /40 and /400 dilutions are used for interpretation, the upper limit is above the acceptability limit for the high inoculation level for all the tested matrices. For four matrices this concerns only one matrix batch among the two tested. For the lower and the medium inoculation levels, the limits are within the acceptability limits for the six matrix strain pairs tested. Contamination levels observed on naturally contaminated samples are generally very low. The criteria for E. coli are comprised between CFU to < 500 CFU/g (FCD, Commission Regulation N 03/05 on microbial criteria for foodstuffs). According the AFNOR Technical Rules (Version 5) inoculation levels between 300 and 0 5 CFU were tested. This resulted for using the /400 or /4000 dilutions. When /40 and /4000 dilutions are used, the upper limit is above the acceptability limit for: - Liver pâté, Level (Upper ETI = 0.56); - Pasteurized milk, Level 3 (Upper ETI = 0.53 and 0.657). A positive bias due to a better recovery of E. coli with TEMPO EC method compared to ISO was previously observed and demonstrated. If the average bias is subtracted from the values obtained for ETI for these two matrices, the upper limits (D = 0. log) are within the acceptability limits for all the tested matrices. For the renewal study run in 0, additional testing run on a whipped cream from pastry was tested. Using the two reference methods (ISO 640 Part (membrane enumeration) and Part (pour plate enumeration) and the TEMPO EC method, the following results were obtained: Product ISO ISO TEMPO EC Whipped cream from Baba au rhum The ISO is recommended for analysis of samples supposed to contain injured bacteria. The enumeration observed with TEMPO EC is close to the enumeration obtained with the ISO indicating that the TEMPO EC method has better performances than the ISO for enumeration of injured bacteria. This could explain the bias observed for the accuracy profiles. ADRIA Développement 0/69 April 5, 07

21 An article from Baylis, Green, Betts (006) confirms this hypothesis: 57 samples (fresh or frozen) were tested using the ISO 6649-, ISO and the TEMPO EC method. The ISO showed a better recovery of the injured cells, the enumeration obtained with the TEMPO method were similar to those of the ISO standard and the lowest enumerations were obtained using the ISO standard Conclusion Considering the positive bias of TEMPO EC method, the accuracy profiles are comprised within the acceptability limits when the /40 and /4000 dilutions are used, in agreement with the level of contamination of samples Inclusivity Exclusivity Inclusivity: study involving pure target strains to be detected or enumerated by the alternative method Exclusivity: study involving pure non-target strains, which can be potentially cross-reactive, but are not expected to be detected or enumerated by the alternative method Protocol During the initial validation, 30 target strains and 0 non target strains were tested in duplicata with the compared methods. The pure cultures were as well enumerated on non selective agar, i.e. PCA. 0 additional target strains and 0 non target strains were tested for the renewal study Results The raw data are given in Appendix 6. The E. coli strains which shown characteristic colonies on TBX ( -glucuronidase positive) were detected by the TEMPO EC method. For most of the tested target strains, the enumeration results are a little bit higher with the TEMPO EC method in comparison to the ISO No cross reaction was observed with the non target strains for both methods, i.e. TEMPO EC and ISO TEMPO EC method shows satisfying specificity and selectivity. ADRIA Développement /69 April 5, 07

22 3..4 Limit of quantification (LOQ) The limit of Quantification (LOQ) is the lowest analyte concentration that can be quantified with an acceptable level of precision and trueness under the conditions of the test. The LOQ was determined as is it needed for the instrumental methods which are related to the growth of the microorganism Experimental design Blank samples were tested for each category. These blank samples were used to verify the limit of quantification of the alternative method. 0 test portions from the same sample were analysed. The same food type matrices than those tested in the accuracy profile were used Calculation and interpretation The threshold standard deviation S 0 was calculated as followed: s n 0 ( y j y ) n j where: n = the total number of test portions used yj = the log transformed result of test portion j y = the average log transformed result of all test portions Results The limit of quantification is calculated as LOQ = 0 s o. Raw data and calculation are provided in Appendix 7. The results are summarized in table 7. Table 7 - Quantification limits per tested matrix Matrix S 0 LOQ Liver pâté 0 0 Pâté for dog 0 0 Grated carrots 0 0 Pasteurized milk 0 0 Fish fillet 0 0 Ground beef 0 0 ADRIA Développement /69 April 5, 07

23 3..5 Practicability The practicability was performed during the initial validation. Work flow study Table 8 Work flow study (time in minutes) Step Reference method: ISO Alternative method: TEMPO EC sample 0 samples sample 0 samples Sampling Dilution / stomach 5 5 Petri dish traceability () 0,3 3 / / Dilutions / Inoculation Reading,5 90 0,45 Time (total),8 7 8,45 8 Time per sample,8,4 8,45 4, The indicated times do not take into account medium preparation which is only necessary for the reference method. TEMPO method uses ready to use TEMPO EC medium. The use of TEMPO EC method reduces - handling time, in particular at the reading step o for instance, when analysing 0 samples, twice less time is needed for TEMPO EC than for ISO method, - waste management o Consumable volume is significantly reduced since only 90 ml diluent, TEMPO EC vial and card are needed per analysed sample. - required incubation space o 0 TEMPO cards holder (0 tests TC) measured.5 cm x 0.5 cm. () 3 dilutions x plates ADRIA Développement 3/69 April 5, 07

24 Traceability among the analyse TEMPO Filler and TEMPO Reader ensure a complete traceability of the following information: - sample identification, - test realisation time, - operator who realised the test, - material lot references, - incubation time and time remaining before card reading, - time card reading, - number of readings for each card - operator who realised the reading, - Results edition in CFU/g, - possible transfer towards a LIMS (Laboratory Information Management System). ADRIA Développement 4/69 April 5, 07

25 3. Inter-laboratory study The inter-laboratory study is a study performed by multiple laboratories testing identical samples at the same time, the results of which are used to estimate alternative-method performance parameters. The results of the inter-laboratory Study run in 005 were interpreted according to the EN ISO 640-:06 standard using Excel spread sheet available at (AP Calculation tool ILS (clause 6..3 Calculation summary and interpretations of data) ver ). 3.. Study organisation The used matrix was pasteurized half-skimmed milk. Twelve laboratories participated to this study 3.. Experimental parameters control 3... Strain stability In order to verify the stability of the strain in the matrix, two samples per inoculation level were enumerated the day of inoculation and after one day storage at - 8 C (See table 9). Table 9 - E. coli 94 enumeration with the reference method (CFU/ml) Day 0 Day Level Level Level No evolution was observed Homogeneity of inoculation This test was not run for the inter-laboratory study as it was not required in 005 (inter-laboratory study run according to the ISO 640:003). ADRIA Développement 5/69 April 5, 07

26 3..3 Analysis results The raw data are given in Appendix Results obtained by the expert Lab. The results obtained by the expert Lab. are the following (See table 0). Table 0 Results obtained by the expert Lab. Inoculation Reference method Alternative method level Duplicate Duplicate Duplicate Duplicate 0 < 0 < 0 < 0 < Targeted contamination levels were reached Results obtained by the collaborators A summary of the test results is given in Table (CFU/g) and Table (log CFU/g). Lab E made an inversion between one sample from Level 0 (E5) and one sample from Level (E). This lab was not kept for data interpretation at Level. ADRIA Développement 6/69 April 5, 07

27 Labor a- Table - Summary of data (CFU/g) Reference method Alternative method Reference method Alternative method Reference method Alternative method Reference method Alternative method Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate tories Blank level Low level Medium level High level Lab A <0 <0 <0 < Lab B <0 <0 <0 < Lab C <0 <0 <0 < Lab D <0 <0 <0 < Lab E 65 <0 <0 <0 < Lab F <0 <0 <0 < Lab G <0 <0 <0 < Lab H <0 <0 <0 < Lab I <0 <0 <0 < Lab J <0 <0 <0 < Lab K <0 <0 <0 < Lab L <0 <0 <0 < Table - Summary of data (log CFU/g) Labor Reference method Alternative method Reference method Alternative method Reference method Alternative method Reference method Alternative method a- Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate Duplicate toires Blank level Low level Medium level High level Lab A <.00 <.00 <.00 < Lab B <.00 <.00 <.00 < Lab C.85 < < Lab D <.00 <.00 <.00 < Lab E <.00 <.00 <.00 <.00 < Lab F.83 <.00 <.00 < Lab G <.00 <.00 <.00 < Lab H <.00 <.00 <.00 < Lab I <.00 <.00 <.00 < Lab J <.00 <.00 <.00 < Lab K <.00 <.00 <.00 < Lab L <.00 <.00 <.00 < ADRIA Développement 7/69 April 5, 07

28 Log (Alternative) biomérieux 3..4 Calculations and interpretation Visual linearity checking The Figure shows the data points after log 0 transformation. The visual inspection shows that the alternative method gives results, which are proportional to those of the reference method. The data are distributed closely to the first bisecting line. Figure - Visual linearity checking 6,00 5,00 Visual linearity checking 4,00 3,00,00,00 0,00 0,00,00,00 3,00 4,00 5,00 6,00 Log (Reference) Accuracy profile calculation Statistical calculations were done according to the Excel spreadsheet available on (ver ). A summary of the statistical test is provided in Table 3. ADRIA Développement 8/69 April 5, 07

29 Accuracy (difference of Log) biomérieux Table 3 - Summary of the statistical test Accuracy profile Study Name Tempo EC Date 005 Coordinator ADRIA Tolerance probability (beta) 80% 80% 80% Acceptability limit in log (lambda) Alternative method Reference method Levels Low Medium High Low Medium High Target value Number of participants (K) Average for alternative method Repeatability standard deviation (sr) Between-labs standard deviation (sl) Reproducibility standard deviation (sr) Corrected number of dof Coverage factor Interpolated Student t Tolerance interval standard deviation Lower TI limit Upper TI limit Bias Pooled repro standard dev = 0.45 Upper TI limit (beta = 80%) New acceptability limit (ALs) = Lower TI limit (beta = 80%) Lower Acceptability limit Upper Acceptability limit These values are collected in a graphical representation together with the acceptability limits (AL). This representation is given Figure 3. Figure 3 0,6 Bias Upper TI limit (beta = 80%) Lower TI limit (beta = 80%) Lower Acceptability limit Upper Acceptability limit 0,4 0, 0,0,5,0,5 3,0 3,5 4,0-0, -0,4-0,6 Levels Log(cfu/g) ADRIA Développement 9/69 April 5, 07

30 It is observed that for all the levels, the tolerance interval limits of the alternative method are within the acceptable limits of 0.5 log. The results obtained with the alternative method are not statistically different than those obtained with the reference method. 3.3 Conclusion For the method comparison study The relative trueness of the alternative method is satisfactory. The accuracy profiles are within the acceptability limits for the 6 matrix/strain pairs in the range of contamination usually observed for Escherichia coli enumerations. The specificity of the alternative method is satisfactory. For the inter-laboratory study The results observed fulfil the EN ISO 640-:06 requirements. For the 3 inoculation levels, the alternative method is accepted as equivalent to the reference method. ADRIA Développement 30/69 April 5, 07

31 Appendix - Artificial contamination N Sample Analysis date Product Artificial contamination Strain Origin Injury protocol Injury measurement Raw cow milk Escherichia coli 4 Raw milk 4 C / 7 days Raw cow milk Escherichia coli 94 Cheese 4 C / 7 days Raw cow milk Escherichia coli E7 Raw milk 4 C / 7 days Raw cow milk Escherichia coli 4 Raw milk 4 C / 7 days Raw cow milk Escherichia coli 9 Raw milk 4 C / 7 days Raw cow milk Escherichia coli 94 Cheese 4 C / 7 days /06/004 Ice cream (chocolate/pistachio) Cross contamination with raw milk Freezing / 4 08/06/004 Ice cream (red fruits) Cross contamination with raw milk Freezing / 4 08/06/004 Ice cream (vanilla) Cross contamination with raw milk Freezing / 43 08/06/004 Ice cream Cross contamination with raw milk Freezing / 44 09/06/004 Ready to reheat food Escherichia coli 08 Ready to reheat (bouchée à la reine) HT 56 C/5 min /06/004 Ready to reheat food Escherichia coli 08 Ready to reheat (bouchée à la reine) HT 56 C/5 min /06/004 Pellets for kitten Escherichia coli 44 Paella HT 56 C/5 min /06/004 Pellets Escherichia coli 44 Paella HT 56 C/5 min /06/004 Pellets Escherichia coli 44 Paella HT 56 C/5 min /06/004 Frozen ready to reheat vegetables Escherichia coli 9 Sliced carrots -0 C / month /06/004 Frozen ready to reheat vegetables Escherichia coli 9 Sliced carrots -0 C / month /06/004 Frozen ready to reheat vegetables Escherichia coli 9 Sliced carrots -0 C / month /07/004 Raw sea almonds Contamination by soaking 57 05/07/004 Raw clams Contamination by soaking 58 05/07/004 Raw oysters Contamination by soaking 59 05/07/004 Carrots Contamination by soaking 60 05/07/004 Turnips Contamination by soaking ADRIA Développement 3/69 April 5, 07

32 N Sample Analysis date Product Artificial contamination Strain Origin Injury protocol 6 05/07/004 Green cabbage Contamination by soaking 64 06/07/004 Ice cream (vanilla/pistachio) Cross contamination with raw milk Freezing / Injury measurement 7 /07/004 Sausage for dog Escherichia coli 08 Ready to reheat (bouchée à la reine) HT 56 C/5 min.09 7 /07/004 Sausage for dog Escherichia coli 3 Veal liver HT 56 C/5 min /07/004 Pâté for dog Escherichia coli Cured meat HT 56 C/5 min /07/004 Pâté for dog Escherichia coli 3 Ground beef HT 56 C/5 min /07/004 Frozen ground beef Escherichia coli 3 Ground beef -0 C / months /07/004 Frozen poultry meat Escherichia coli Cured meat -0 C / months /07/004 Frozen pork meat Escherichia coli 3 Veal liver -0 C / months /07/004 Cooked mussels Escherichia coli Ad8 Fish Osmotic stress (35 / NaCl) /07/004 Tuna Escherichia coli 93 Ready to reheat fish Osmotic stress (35 / NaCl) > /07/004 Fish fillet Escherichia coli 93 Ready to reheat fish Osmotic stress (35 / NaCl) >3.5 8 /07/004 Grated carrots Cross contamination with water 8 /07/004 Cabbage Cross contamination with water 83 /07/004 Mixed salad Cross contamination with water 93 0/07/004 Dried apple Contact with desiccated bacteria suspension and storage 4h at ambient temperature 94 0/07/004 Dried apple Contact with desiccated bacteria suspension and storage 4h at ambient temperature 504 6/0/006 Infant formula milk Escherichia coli 6 Raw milk HT 55 C/30 min /0/006 Infant formula milk Escherichia coli 6 Raw milk HT 55 C/30 min /0/006 Infant formula milk Escherichia coli 6 Raw milk HT 55 C/30 min /0/006 Infant formula milk Escherichia coli 6 Raw milk HT 55 C/30 min /0/006 Infant formula milk Escherichia coli 6 Raw milk HT 55 C/30 min /0/006 Infant formula milk Escherichia coli 4 Raw milk HT 55 C/30 min 0.6 ADRIA Développement 3/69 April 5, 07

33 N Sample Analysis date Product Artificial contamination Strain Origin Injury protocol Injury measurement 50 30/0/006 Infant formula milk Escherichia coli 4 Raw milk HT 55 C/30 min /0/006 Infant formula milk Escherichia coli 4 Raw milk HT 55 C/30 min /0/006 Infant formula milk Escherichia coli 4 Raw milk HT 55 C/30 min /0/006 Milk Escherichia coli 4 Raw milk HT 55 C/30 min //006 Infant formula milk Escherichia coli 5 Raw milk HT 55 C/30 min //006 Infant formula milk Escherichia coli 5 Raw milk HT 55 C/30 min //006 Infant formula milk Escherichia coli 5 Raw milk HT 55 C/30 min //006 Infant formula milk Escherichia coli 5 Raw milk HT 55 C/30 min //006 Infant formula milk Escherichia coli 5 Raw milk HT 55 C/30 min //006 Infant formula milk Escherichia coli 9 Raw milk HT 55 C/30 min //006 Infant formula milk Escherichia coli 9 Raw milk HT 55 C/30 min.6 5 0//006 Infant formula milk Escherichia coli 9 Raw milk HT 55 C/30 min.6 5 0//006 Infant formula milk Escherichia coli 9 Raw milk HT 55 C/30 min //006 Infant formula milk Escherichia coli 9 Raw milk HT 55 C/30 min //006 Infant formula milk Escherichia coli 97 Raw milk HT 55 C/30 min 7.7 ADRIA Développement 33/69 April 5, 07

34 N Sample Analysis date Product Artificial contamination Strain Origin Injury protocol Injury measurement 55 0//006 Infant formula milk Escherichia coli 97 Raw milk HT 55 C/30 min //006 Infant formula milk Escherichia coli 94 Raw milk HT 55 C/30 min //006 Infant formula milk Escherichia coli 94 Raw milk HT 55 C/30 min //006 Infant formula milk Escherichia coli E7 Raw milk HT 55 C/30 min //006 Infant formula milk Escherichia coli E7 Raw milk HT 55 C/30 min //006 Infant formula milk Escherichia coli Raw milk HT 55 C/30 min //006 Infant formula milk Escherichia coli Raw milk HT 55 C/30 min //006 Infant formula milk Escherichia coli Raw milk HT 55 C/30 min //006 Infant formula milk Escherichia coli Raw milk HT 55 C/30 min Ratatouille Escherichia coli 9 Sliced carrots Seeding HT 56 C/5 min / Chilli con carne Escherichia coli 44 Paella HT 56 C/5 min / Ready to cook vegetables Escherichia coli 44 Paella Seeding HT 56 C/5 min / 06 0 Ready to eat Chinese food Escherichia coli 93 Ready to reheat fish Seeding HT 56 C/5 min / 06 0 Pastry Escherichia coli Ad Egg product Seeding HT 56 C/5 min / Pastry Escherichia coli 4 Egg product Seeding HT 56 C/5 min / Pastry Escherichia coli Ad Egg product Seeding HT 56 C/5 min / Pork nem Escherichia coli 08 Ready to reheat meal Seeding HT 56 C/5 min / Chinese ravioli with pork Escherichia coli 08 Ready to reheat meal Seeding HT 56 C/5 min / ADRIA Développement 34/69 April 5, 07

35 N Sample Analysis date Product Artificial contamination Strain Origin Injury protocol Injury measurement 0 0 Frozen ratatouille Escherichia coli 9 Sliced carrots Seeding -0 C / 8 Days / 0 Ready to eat salad Escherichia coli 9 Sliced carrots Seeding -0 C / 8 Days / 0 Smoked salmon sandwich Escherichia coli 93 Ready to reheat fish Seeding -0 C / 8 Days / 3 0 Tabbouleh Escherichia coli 08 Ready to reheat meal Seeding -0 C / 8 Days / 50 0 Ready to eat pasta salad Escherichia coli Ad Egg product Seeding HT 0min 56 C / 5 0 Sandwich ham egg crudités Escherichia coli Ad Egg product 4 C / Day / 5 0 Smoked sausages Escherichia coli Raw delicatessen Seeding HT 0min 56 C / 53 0 Smoked sausages Escherichia coli Raw delicatessen Seeding HT 0min 56 C / 83 0 Frozen leeks quiche Escherichia coli 4 Egg product Seeding -0 C / 6 days / 84 0 Sandwich chicken and egg Escherichia coli 4 Egg product Seeding 4 C / 6 days / 85 0 Sausages Escherichia coli 6 Sausage Seeding 4 C / 6 days / 86 0 Poultry meat Escherichia coli 96 Turkey meat Seeding 4 C / 6 days / Salad carrots celery Escherichia coli 9 Sliced carrots Seeding 4 C / 7 days / Celery remoulade Escherichia coli 44 Paella Seeding 4 C / 7 days / Mixed vegetables Escherichia coli 44 Paella Seeding HT 56 C/5 min / Mixed vegetables Escherichia coli 4 Egg products Seeding HT 56 C/0 min / Sandwich ham egg tomatoes Escherichia coli 0 Pork Seeding HT 56 C/0 min / Cooked ham Escherichia coli 0 Pork Seeding HT 56 C/0 min / Garlic sausage Escherichia coli Pork Seeding HT 56 C/0 min / 47 0 Chorizo Escherichia coli Pork Seeding HT 56 C/0 min / Garlic sausage Escherichia coli Pork Seeding HT 56 C/0 min / Dry ham Escherichia coli Pork Seeding 4 C / 7 days / Bacon Escherichia coli 3A Pork Seeding 4 C / 7 days / Sliced bacon Escherichia coli Pork Seeding 4 C / 7 days / Ham Escherichia coli Pork Seeding 4 C / 7 days / ADRIA Développement 35/69 April 5, 07