MC-Media Pad ACplus for Enumeration of Aerobic Counts in a Variety of Foods

Transcription

1 TERAMURA & BETTS: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. nnn, FOOD BIOLOGICAL CONTAMINANTS MC-Media Pad ACplus for Enumeration of Aeroic Counts in a Variety of Foods AOAC Performance Tested Method SM Astract TheMC-MediaPadACplus is a dry, rehydratale film medium for the enumeration of aeroic acterial colonies. The performance of the method in a variety of foods was compared to that of U.S. reference methods: U.S. Department of Agriculture (USDA) Food Safety and Inspection Service (FSIS) Microiology Laoratory Guideook (MLG) Chapter 3.02 Quantitative Analysis of Bacteria in Foods as Sanitary Indicators (USDA/FSIS MLG 3.02); Standard Methods for the Examination of Dairy Products (SMEDP) Chapter 6 Microiological Count Methods, Standard Plate Count Method (SMEDP 6); AOAC Official Method SM Microiological Methods; and ISO :2013 Microiology of the food chain Horizontal method for the enumeration of microorganisms Part 1: Colony count at 30 degrees C y the pour plate technique. The validated matrixes included raw chicken reast and raw ground pork for USDA/FSIS MLG 3.02; cream cheese and yogurt drink for SMEDP 6; parsley, vegetale juice, prawns, tuna pate, sandwiches, and pasta salad for AOAC Method , and raw chicken reast, raw ground pork, cream cheese, yogurt drink, parsley, vegetale juice, prawns, tuna pate, sandwiches, and pasta salad for ISO :2013. In each matrix study, five replicates at each of three contamination levels were tested as paired test portions. All 10 matrixes were compared to the appropriate U.S. reference methods under MC-Media Pad ACplus standard-usage conditions (35 ± 1 C for 48 ± 2 h). Across all matrixes, the difference of mean log 10 valueanged from 0.43 to 0.44, within the acceptale range of 0.50 to The candidate method repeataility SD ( )variedfrom0.03to0.23log 10 CFU/g, comparing favoraly to the reference method SD, which ranged from 0.06 to 0.30 log 10 CFU/g. Seven matrixes were compared to the appropriate U.S. reference methods under MC- Media Pad ACpluapid-usage conditions (35 ± 1 C for 24 ± 2 h). Of the 21 matrix/concentration cominations, only three instances of difference of mean >0.5 log were oserved. The ranges of values of the rapid-usage candidate method ( ) and the reference method ( ) were similar for the seven matrixes tested. All 10 matrixes were compared to the International Organization for Standardization (ISO) reference method under MC-Media Pad ACplus alternate-method conditions (30 ± 1 C for 72 ± 3 h). All 10 matrixes yielded a mean difference etween methods of <0.5 log, and the ranges of values were similar etween the candidate alternate method ( ) and the ISO reference method Sumitted for pulication Septemer 14, The method was independently tested, evaluated, and certified y the AOAC Research Institute as a Performance Tested Method SM. See for information on certification. Corresponding author s h.teramura@jnc-corp.co.jp DOI: ( ). The product consistency study demonstrated no significant difference etween lots of product and supported the 2-year shelf life. Roustness testing yielded no significant differences when small variations were made in sample volume, incuation temperature, and incuation time. Thus, the data show equivalent or etter performance of the MC-Media Pad ACplus method compared to the relevant reference methods in support of AOAC Performance Tested Method SM certification. Participants METHOD AUTHORS HAJIME TERAMURA JNC Corp., Yokohama Research Center, 5-1, Ookawa, Kanazawa-ku, Yokohama, Kanagawa , Japan GAIL BETTS Campden BRI Ltd, Station Rd, Chipping Campden, Gloucestershire GL55 6LD, United Kingdom SUBMITTING COMPANY JNC Corp., Yokohama Research Center, 5-1, Ookawa, Kanazawa-ku, Yokohama, Kanagawa , Japan INDEPENDENT LABORATORY Campden BRI Ltd, Station Rd, Chipping Campden, Gloucestershire GL55 6LD, United Kingdom REVIEWERS YI CHEN U.S. Food and Drug Administration, Center for Food Safety and Nutrition, 5100 Paint Branch Pkwy, College Park, MD MICHAEL BRODSKY Brodsky Consultants, 73 Donnamora Crescent, Thornhill, ON, Canada L3T 4K6 YVONNE SALFINGER Consultant, Association of Pulic Health Laoratories, 1488 Madison St, No. 501, Denver, CO Scope of Method (a) Target organisms. Total aeroic count. () Matrixes. Cooked, peeled, chilled, cold-water prawns (1.3% salt); prewashed agged flat-leaf parsley; chilled tuna pate (23% fat, 0.9% salt); fermented strawerry yogurt drink (1.8% fat containing Lactoacillus acidophilus, Bifidoacterium, and L. casei); fresh raw chicken reast fillets (2% fat); fresh raw lean pork mince (5% fat); cold-press green vegetale juice (celery, cucumer, lemon, lettuce, spinach, pear, and pineapple); full-fat _Teramura 1

2 2 TERAMURA & BETTS: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. nnn, 2017 Tale 1. Parameters to e varied Principle of the Method Parameter Low value Nominal value High value Sample volume, ml Incuation temperature, C Incuation time, h soft cream cheese (24% fat); fresh chilled egg and mayonnaise sandwich on white sliced read (5% fat, 0.9% salt); and fresh feta cheese deli pasta salad (3.4% fat, 0.5% salt). (c) Summary of validated performance claims. Performance equivalent to that of the U.S. Department of Agriculture (USDA) Food Safety and Inspection Service (FSIS) Microiology Laoratory Guideook (MLG) Chapter 3.02 Quantitative Analysis of Bacteria in Foods as Sanitary Indicators (USDA/ FSIS MLG 3.02) reference method for 50 g test portions of raw ground pork and raw chicken reast under standard and rapidusage conditions; the Standard Methods for the Examination of Dairy Products (SMEDP) Chapter 6 Microiological Count Methods, Standard Plate Count Method (SMEDP 6) reference method for 11 ml test portions of yogurt drink under standardand rapid-usage conditions and for 11 g test portions of cream cheese under standard-usage conditions; AOAC Official Method SM Microiological Methods for 50 g test portions of parsley, vegetale juice, tuna pate, and pasta salad under standard- and rapid-usage conditions and for 50 g test portions of prawns and sandwiches under standard-usage conditions; and ISO :2013 Microiology of the food chain Horizontal method for the enumeration of microorganisms Part 1: Colony count at 30 degrees C y the pour plate technique for 10 g test portions of all claimed matrixes under alternate-method conditions. Definitions (a) Repeataility SD ( ). SD of replicates for each analyte at each concentration of each matrix for each method. () Mean difference etween candidate and reference methods. Mean difference etween candidate and reference method transformed results with 95% confidence interval for each analyte at each concentration of each matrix. The MC-Media Pad ACplus is a test pad coated with growth medium and three redox indicators for the detection of aeroic acterial colonies. After the liquid sample is inoculated onto the test pad, the sample diffuses through the whole pad y capillary action, reconstituting the medium. If aeroic acteria are present, they grow aed colonies on the test pad. For most foods, a 50 g or 50 ml test portion is homogenized with 450 ml Butterfield s Phosphate Buffer (BPB). Serial decimal dilutions are prepared as needed. For dairy products, an 11 g test portion of cream cheese is added to 99 ml 2% sodium citrate warmed to C and homogenized. An 11 ml test portion of yogurt drink is diluted into 99 ml BPB. Serial dilutions of dairy products are prepared with 1 or 11 ml homogenate diluted into 99 ml BPB. A 1 ml aliquot of diluted sample is applied to the test pad, the cover film ieplaced, and the plate is incuated at 35 ± 1 C for 24 ± 2 h (rapid usage) or 48 ± 2 h (standard usage). Even though the rapid usage can e used for processed food and other foods after eing verified y the user, it does not apply to foods containing large amounts of lactic acid acteria or psychrophilic acteria. Alternatively, a 10 g test portion of matrix is diluted into 90 ml Maximum Recovery Diluent (MRD). Serial dilutions are prepared with 10 ml homogenate diluted into 90 ml MRD. A 1 ml aliquot of diluted sample is applied to the test pad, the cover film ieplaced, and the plate is incuated at 30 ± 1 C for 72 ± 3 h (alternate method). General Information Aeroic acteria are routinely tested in food manufacture as a sanitary process indicator and as an assessment of the general microiological quality of foods. Total aeroic count methods are intended to enumerate all microorganisms that are ale to grow and form colonies in a solid medium after aeroic incuation under the stated temperature. In this study, this involved a temperature of 35 ± 1 C for the candidate method and 32 ± 1 C or 35 ± 1 C for the reference methods. The total aeroic plate count is not a selective test; therefore, it will detect a wide range of vegetative cells and acterial spores, including Gram-positive and Gram-negative microorganisms. The total aeroic count will vary dependent on the matrix analyzed and may include Pseudomonas, Enteroacteriaceae, lactic acid Tale 2. Roustness testing with E. coli NBRC Treatment comination Sample volume, ml Incuation temperature, C Incuation time, h n a High level: CFU Mean colony count Low level:1 10 CFU Mean colony count a n = Numer of replicates. = SD of replicates _Teramura 2

3 Tale 3. Roustness testing with P. aeruginosa NBRC TERAMURA & BETTS: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. nnn, Treatment comination Sample volume, ml Incuation temp, C Incuation time, h n a High level: CFU Mean colony count Low level:1 10 CFU Mean colony count a n = Numer of replicates. = SD of replicates. acteria, and Bacillus spp. Aeroic plate counts are used frequently y the food industry for monitoring the general microiological quality of foods, and an easy, ready-to-use alternative to standard methods is needed. The MC-Media Pad ACplus was developed to provide ease of use, convenience, and shorter time to results over traditional agar plates. This paper reports on the validation of the MC-Media Pad ACplus. Materials and Methods Test Kit Information (a) Kit name. MC-Media Pad ACplus. () Cat. Nos. SK01A25, SK01B25, SK01A10, and SK01B10. (c) Ordering information. JNC Corp., Life Chemical Launch Office, 2-2-1, Otemachi, Chiyoda-ku, Tokyo , Japan; mcmp@jnc-corp.co.jp; Tel: ; Fax: Test Kit Components MC-Media Pad ACplus medium sheets. Additional Supplies, Reagents, and Apparatus for MC-Media Pad ACplus Method (a) Buffered peptone water (BPW). Cat. No. CM0509 (Oxoid Ltd). () Sodium citrate, 2% solution. Prepare y adding 20 g trisodium citrate dihydrate (Na 3 C 6 H 5 O 7 2H 2 O) to 990 ml distilled water. Stir and, if necessary, heat to C to dissolve. Adjust ph, if necessary, to 7.5 and dispense into required volumes. Sterilize at 121 C for 15 min. (c) BPB. Prepare y adding 34 g KH 2 PO 4 to 500 ml distilled water. Adjust to ph 7.2 with 1 N NaOH and ring final volume to 1 L with distilled water. Dispense into 90 ml, 99 ml, or 1 L amounts aequired. Sterilize at 121 ± 1 C for 15 min. (d) MRD. Cat. No. CM0733 (Oxoid Ltd). (e) Stomacher (or equivalent) homogenizer. For homogenizing samples. (f) Filtered stomacher ags. (g) Blender. Waring type. (h) Sterile lender jars. (i) Pipets. 1 ml. (j) Incuator. Capale of maintaining 35 ± 1 C or 30 ± 1 C. (k) Colony counter. Additional Organisms, Media, and Apparatus for Validation Studies (a) Escherichia coli NBRC National Institute of Technology and Evaluation (NITE) Biological Resource Center (Chia, Japan). () Pseudomonas aeruginosa NBRC NITE Biological Resource Center (Chia, Japan). (c) Staphylococcus aureus NBRC NITE Biological Resource Center (Chia, Japan). (d) Incuator. Capale of maintaining 32 ± 1 C. Tale 4. Comined lot-to-lot and staility testing with E. coli NBRC h incuation 48 h incuation High level: CFU Low level:1 10 CFU High level: CFU Low level:1 10 CFU Lot No. n a Mean colony count Mean colony count Mean colony count Mean colony count a n = Numer of replicates. = SD of replicates _Teramura 3

4 4 TERAMURA & BETTS: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. nnn, 2017 Tale 5. Comined lot-to-lot and staility testing with S. aureus NBRC h incuation 48 h incuation High level: CFU Low level: 1 10 CFU High level: CFU Low level: 1 10 CFU Lot No. N a Mean colony count Mean colony count Mean colony count Mean colony count a n = Numer of replicates. = SD of replicates. (e) Plate Count Agar (PCA). Cat. No. CM0325 (Oxoid Ltd). (f) Standard methods agar. Standard Plate Count Agar, Cat. No. CM0463 (Oxoid Ltd). (g) Tryptic Soy Broth (TSB). Cat. No. BD Precautions (a) The test is designed for use y QC personnel and others familiar with testing samples potentially contaminated with aeroic microes. () Use personal protective equipment such as gloves, laoratory coat, and eye protection. Treat samples and inoculated devices as potentially iohazardous. All used test kits and devices must e sterilized y autoclaving or oiling. (c) Read the instruction manual carefully efore use. (d) After opening the aluminum ag, unused plates should e stored in the aluminum ag, sealed with tape, and kept in a cool (2 15 C) environment. After opening, use all plates within 1 month. (e) Do not expose unused plates to sunlight or UV light. (f) Do not use a discolored or damaged plate. Tale 6. Food matrix analyses Matrix Final contamination level Replicates Test portion Reference method Raw ground pork 10 2 CFU/g 5 50 g USDA/FSIS MLG CFU/g CFU/g 5 Raw chicken reast 10 2 CFU/g 5 50 g USDA/FSIS MLG CFU/g CFU/g 5 Fresh parsley 10 3 CFU/g 5 50 g AOAC Method CFU/g CFU/g 5 Vegetale juice 10 3 CFU/g 5 50 g AOAC Method CFU/g CFU/g 5 Cooked prawns 10 2 CFU/g 5 50 g AOAC Method CFU/g CFU/g 5 Tuna pate 10 2 CFU/g 5 50 g AOAC Method CFU/g CFU/g 5 Sandwiches 10 2 CFU/g 5 50 g AOAC Method CFU/g CFU/g 5 Pasta salad 10 2 CFU/g 5 50g AOAC Method CFU/g CFU/g 5 Yogurt drink 10 2 CFU/mL 5 11 ml SMEDP CFU/mL CFU/mL 5 Cream cheese 10 2 CFU/mL 5 11 g SMEDP CFU/mL CFU/mL _Teramura 4

5 TERAMURA & BETTS: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. nnn, Figure 1. Plots of candidate method versueference method for each food matrix under standard-usage conditions, with line of equivalence. (g) A wrinkle on the test pad should not affect detection. (h) Small fragments of faric on or around the test pad should not affect detection. (i) Do not use the plates after the expiration date. The quality of an expired plate is not warranted. (j) The measurement range is <300 CFU/plate. If >300 CFU/plate are counted, further dilution iecommended. (k) The rapid-mode test is not suitale for all foods. Therefore, suitaility should e verified using your own samples efore applying. (l) The nature of the food (e.g., foods that have high viscosity or include dye) may affect test usage or results. In this case, the cause needs to e eliminated y dilution or other means. (m) The used kit must e sterilized y autoclaving or oiling, and then disposed according to local regulations for waste. Sample Preparation (a) Raw meat and poultry. Weigh 50 g in a filtered stomacher ag, add 450 ml BPW, and stomach for 2 min. Make serial dilutions y adding 10 ml diluted matrix to 90 ml BPW and shaking 25 times to mix. () Cream cheese. Weigh 11 g into a filtered stomacher ag, add 99 ml 2% sodium citrate at C, and stomach for 2 min. Make serial dilutions y adding 11 ml diluted cream cheese to 99 ml BPB and shaking 25 times to mix _Teramura 5

6 6 TERAMURA & BETTS: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. nnn, 2017 Figure 1. Continued. (c) Yogurt drink. To 11 ml yogurt drink, add 99 ml BPB, and invert 25 times to mix. Make serial dilutions y adding 11 ml diluted yogurt drink to 99 ml BPB and shaking 25 times to mix. (d) All other foods. Weigh 50 g matrix into a lender jar, add 450 ml BPB, and lend for 2 min. Make serial dilutions y adding 10 ml diluted matrix to 90 ml BPB and shaking 25 times to mix. (e) Alternate method. Weigh 10 g matrix into a filtered stomacher ag, add 90 ml MRD, and stomach for 1 min. Make decimal serial dilutions in MRD y adding 10 ml diluted matrix to 90 ml MRD. Analysis (a) Open the aluminum ag and remove the MC-Media Pad. If necessary, write information on the cover film. () Lift the cover film and drop 1.0 ml sample solution onto the test pad. It iecommended to lift the cover film diagonally for easy and sure resealing. (c) Replace the cover film and lightly press the edges of film to seal. (d) For standard usage, incuate test plate at 35 ± 1 C for 48 ± 2 h. (e) For rapid usage, incuate test plate at 35 ± 1 C for 24 ± 2 h. Note: The standard usage is applicale for all foodstuffs. For foodstuffs that contain large amounts of lactic acid acteria (e.g., Lactoacillus spp.) or psychrophilic acteria (e.g., Pseudomonas spp.), the rapid usage may not e applicale. Large amounts of lactic acid acteria may affect color development of the redox indicator due to acid production. Further, psychrophilic acteria may not e detected within 24 h ecause they tend to grow slowly in nonoptimum temperature. (f) For the alternate method to ISO :2013, incuate test plate at 30 ± 1 C for 72 ± 3 h. Calculations, Interpretation, and Test Result Report Count all reddish-colored colonies. Certain acteria (in particular, Bacillus spp. strains) may form diffuse and fuzzy round shapes. In this case, dark-colored points should e counted as colonies. For large numers of colonies, colony counts can e Tale 7. Statistically significant results from the Grus tests for standard usage Method Matrix/contamination level Outlier, log CFU/g P Reference Cream cheese/low highest Reference Cream cheese/high highest Reference Vegetale juice/high highest Candidate Yogurt drink/medium lowest Candidate Yogurt drink/high highest Candidate Fresh parsley/medium highest Candidate Fresh parsley/high lowest Candidate Sandwich/low lowest _Teramura 6

7 Tale 8. Matrix study results for standard usage (35 C, 48 h) TERAMURA & BETTS: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. nnn, Candidate method Reference method 95% confidence interval Matrix Contamination level n a Mean, log CFU/g Mean, log CFU/g Mean difference, Dlog CFU/g LCL c UCL d Raw ground pork Low Medium High Raw chicken reast Low Medium High Fresh parsley Low Medium High Vegetale juice Low Medium High Cooked prawns Low Medium High Tuna pate Low Medium High Sandwiches Low Medium High Pasta salad Low Medium High Yogurt drink Low Medium High Cream cheese Low Medium High a c d n = Numer of replicates. = SD of replicates. LCL = Lower confidence limit. UCL = Upper confidence limit. estimated y counting colonies in one grid square and multiplying y 20. If more than 10 4 microes are grown, the entire test pad may appear stained and look as if no individual colonies were formed. If this is the case, dilute the sample further and retest. If necessary, a target colony can e picked up with a sterile needle from the test pad for further analysis. Validation Study This validation study was conducted under the AOAC Research Institute Performance Tested Method SM program and the AOAC INTERNATIONAL Methods Committee Guidelines for Validation of Microiological Methods for Food and Environmental Surfaces (1). Method developer studies were conducted in the laoratories of JNC Corp., and included the product consistency and staility studies and roustness testing. The independent laoratory study was conducted y Campden BRI Ltd, and included the matrix study for all the claimed food matrixes. Method Developer Studies (a) Roustness study. (1) Methodology. Three method parameters were varied (sample volume, incuation temperature, and incuation time), as shown in Tale 1. A factorial experimental design was followed, as shown in Tales 2 and 3. E. coli NBRC and P. aeruginosa NBRC were cultured in TSB for 24 h at 35 C and diluted to a high level and a low level within the quantitative range of the MC-Media Pad ACplus method. For each target concentration, 1 ml acterial dilution was inoculated onto each of five MC-Media Pad ACplus devices for each treatment comination shown in Tales 2 and 3. After counting colonies, the values of each treatment comination at each concentration were calculated and compared to determine differences among parameter settings _Teramura 7

8 8 TERAMURA & BETTS: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. nnn, 2017 Figure 2. Plots of candidate method versueference method for seven food matrixes under rapid-usage conditions, with line of equivalence. (2) Results. Roustness study results are shown in Tales 2 and 3 for E. coli and P. aeruginosa, respectively. For E. coli, the values of each treatment comination were less than 3-fold the of treatment comination 9 (normal value) at oth the high level and the low level. For P. aeruginosa, the values of each treatment comination were also less than 3-fold the of treatment comination 9 (normal value) at oth the high level and the low level. Therefore, the MC-Media Pad ACplus method ioust for sample volumes of ml, incuation temperatures of C, and incuation times of h. () Product consistency (lot-to-lot) and staility studies. (1) Methodology. Comined lot-to-lot and staility studies were conducted y testing three lots of kits: one newly manufactured lot (Lot No AC, 3 months postmanufacture), one lot at the middle of the expiration term (Lot No AC, 13 months postmanufacture), and one lot at or near expiration (Lot No AC, 24 months postmanufacture). All lots were stored etween 2 and 15 C and tested on the same day. E. coli NBRC and S. aureus NBRC were cultured in TSB for 24 h at 35 C and diluted to a high level and a low level within the quantitative range of the MC-Media Pad ACplus method. Each target concentration was tested in quintuplicate on each MC-Media Pad ACplus lot. After counting colonies, the values of each lot at each concentration were calculated and compared to determine differences among the three lots. (2) Results. The results for the comined lot-to-lot and staility studies are shown in Tales 4 and 5. For E. coli at oth the 24 and 48 h incuation times, the values of each lot were less than 3-fold the of the newly manufactured lot (lot 1) at oth the high level and the low level. For S. aureus at oth the 24 and 48 h incuation times, the values of each lot were also less than 3-fold the of the newly manufactured lot (lot 1) at oth the high level and the low level. These results show that MC- Media Pad ACplus devices are stale, even after 24 months from date of manufacture, supporting the claimed shelf life of 2 years. Furthermore, MC-Media Pad ACplus devices showed no significant difference among lots and no significant difference etween the 24 and 48 h incuation times. Independent Laoratory Studies (a) Matrix study: comparison to U.S. reference methods. (1) Methodology. Matrixes were otained from local markets. To otain the three different contamination levels for each matrix, samples of each food were stored at 8 ± 1 C for varying lengths of time. For low-level contamination samples, foods were purchased as close to the start of their shelf life as possile and analyzed immediately after purchase. For medium-level contamination, food samples were stored at 8 ± 1 C for the duration of the designated shelf life of the food and then tested. For high-level contamination, samples were stored at 8 ± 1 C for 1 3 weeks after the designated shelf life and then frozen. This approach was taken in order to allow the naturally present organisms to grow to a high enough concentration to achieve medium and high levels. For each food matrix at each contamination level, a ulk sample (300 g) was weighed into a sterile stomacher ag and _Teramura 8

9 TERAMURA & BETTS: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. nnn, Figure 2. Continued. thoroughly mixed to ensure homogeneity. Five replicate test portions (11 g, 11 ml, or 50 g, depending on the matrix) were taken, diluted, and then plated out using the appropriate U.S. reference method for the food type (see Tale 6). For raw ground pork and raw chicken reast, the MC-Media Pad ACplus method was compared to the USDA/FSIS MLG 3.02 reference method (2) in a paired study. Briefly, 50 g test portions taken from the homogenized ulk sample were placed in stomacher ags, and 450 ml BPW was added. Each diluted test portion was stomached for 2 min ± 15 s, and then serial dilutions were made y diluting 10 ml homogenate with 90 ml BPW. Each dilution was shaken 25 times in a 1 ft arc. From each dilution, 1 ml aliquots were plated in duplicate on PCA using the pour plate technique. Plates were incuated at 35 ± 1 C for 48 ± 2 h. Colonies were counted on plates containing CFU/plate, and the mean of duplicates was used to determine the CFU per gram of product for each replicate test portion at each contamination level, taking the dilution factor into account. Concurrently, 1 ml aliquots from each dilution were inoculated in singlet on MC-Media Pad devices, incuated at 35 ± 1 C, and then colonies were counted on devices containing CFU at 24 ± 2 h (rapid usage) and at 48 ± 2 h (standard usage). For yogurt drink and cream cheese, the MC-Media Pad ACplus method was compared to the SMEDP 6 reference method (3) in a paired study. Briefly, 11 g test portions of cream cheese taken from the homogenized ulk sample were placed in stomacher ags, and 99 ml 2% sodium citrate dilution uffer warmed to C was added. The diluted test portion was stomached for 2 min. Yogurt drink test portions (11 ml) taken from the homogenized ulk sample were placed in stomacher ags, and 99 ml BPB was added. The diluted test portions were mixed y inverting 25 times in 7 s in a 1 ft arc. Serial dilutions of each matrix were made y diluting 11 ml homogenate with 99 ml BPB. Each dilution was shaken 25 times in 7 s in a 1 ft arc. From each dilution, 1 ml aliquots were plated in singlet on Standard Plate Count Agar using the pour plate technique. Plates were incuated at 32 C for 48 ± 3 h. Colonies were counted on plates containing CFU/plate, and the dilution with the correct amount of colonies was used to determine the CFU per gram or CFU per milliliter of product for each replicate test portion at each contamination level, taking the dilution factor into account. Concurrently, 1 ml aliquots from each dilution were inoculated in singlet on MC-Media Pad devices, incuated at 35 ± 1 C, and colonies were counted on devices containing CFU at 24 ± 2 h (rapid usage) and at 48 ± 2 h (standard usage). All other matrixes were compared to AOAC Official Method SM (4) in a paired study. Briefly, 50 g or 50 ml test portions taken from the homogenized ulk sample were placed in lender jars, and 450 ml BPB was added to each Tale 9. Statistically significant results from the Grus tests for rapid usage Method Matrix/contamination level Outlier, log CFU/g P Reference Vegetale juice/high highest Candidate Yogurt drink/high highest < _Teramura 9

10 10 TERAMURA & BETTS: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. nnn, 2017 Tale 10. Matrix study results for rapid usage (35 C, 24 h) Candidate method Reference method 95% confidence interval Matrix Contamination level n a Mean, log CFU/g Mean, log CFU/g Mean difference, Dlog CFU/g LCL c UCL d Raw ground pork Low Medium High Raw chicken reast Low Medium High Fresh parsley Low Medium High Vegetale juice Low Medium High Tuna pate Low Medium High Pasta salad Low Medium High Yogurt drink Low Medium High a c d n = Numer of replicates. = SD of replicates. LCL = Lower confidence limit. UCL = Upper confidence limit. jar and lended for 2 min ± 15 s. Serial dilutions were made y diluting 10 ml homogenate with 90 ml BPB, and each dilution was shaken 25 times in a 30 cm arc. From each dilution, 1 ml aliquots were plated in duplicate on PCA using the pour plate technique. Plates were incuated at 35 ± 1 C for 48 ± 2 h. Colonies were counted on plates containing CFU/plate, and the mean of duplicates was used to determine the CFU per gram or CFU per milliliter of product for each replicate test portion at each contamination level, taking the dilution factor into account. Concurrently, 1 ml aliquots from each dilution were inoculated in singlet on MC-Media Pad devices, incuated at 35 ± 1 C, and colonies were counted on devices containing CFU at 24 ± 2 h (rapid usage) and at 48 ± 2 h (standard usage). (2) Results. Raw data were transformed using log 10 [CFU per unit + (0.1)f], where f = the CFU per unit (grams or milliliters) corresponding to the smallest reportale result. The smallest reportale result is 10 CFU/g or 10 CFU/mL, so the transformation was carried out with f = 10 for the whole data set. The candidate method claims to give a total aeroic plate count under standard conditions (48 h at 35 C) and under rapid conditions (24 h at 35 C). The manufacturer does note in the kit insert that rapid usage may not e applicale for foods that contain a large amount of lactic acid acteria (e.g., Lactoacillus spp.) or psychrophilic acteria (e.g., Pseudomonas spp.). From the matrixes tested in this study, there were seven products that gave sufficient data to compare the performance of the candidate versus the reference method under rapid incuation. These were vegetale juice, fresh parsley, fresh pork mince, fresh chicken reasts, deli pasta salad, tuna pate, and yogurt drink. The data analysis for rapid usage, therefore, iestricted to these seven matrixes. For the standard usage (48 h), all 10 matrixes were used in the data analysis. For each product at the various contamination levels, graphs were produced plotting the candidate method results versus the reference method results, showing the line of equivalence to look for any apparent discrepancies. The plots for standard usage (48 h) are presented in Figure 1. For sandwiches, the medium and high levels were very close to each other. Because these samples were not inoculated and contained only natural microorganisms, it was not easy to find very distinct levels. Otherwise, no major discrepancies were apparent. The data were examined statistically for outliers using the Cochran test and the Grus test from the R Package outliers test (5), with default values for all options. Statistically significant evidence was assessed using the commonly accepted criterion of P < The Cochran test was performed twice, once for the reference method and once for the candidate method, identifying the product/level cominations with the highest variance. There was statistically significant evidence of high variance for the candidate method for cream cheese at the high level (P = 0.008). The Grus test was performed 60 times, once for each method/product/level comination, identifying individual values far from the corresponding group mean. Eight data points were identified _Teramura 10

11 TERAMURA & BETTS: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. nnn, Figure 3. Plots of candidate method versueference method for each food matrix under alternate-usage conditions, with line of equivalence. as outliers y the Grus test, as shown in Tale 7; however, no justifiale cause was identified to remove these statistical outliers. In addition, the spread of the data was visualized using a stem-and-leaf display, a letter-and-value display, and a ox plot (figures not shown), ut revealed no ovious differences etween the two methods. The matrix study results and statistical analyses for the standard-usage conditions are summarized in Tale 8. For the reference method, the ranged from to 0.300, and for the candidate method, the range was This shows similar repeataility values for the candidate and reference methods. The largest mean differences etween methods were for yogurt drink at the high level ( 0.498), fresh parsley at the high level (0.439), and fresh parsley at the low level ( 0.432), with the candidate method yielding a higher result in one case and the reference method yielding a higher result in two cases. From a microiological viewpoint, a difference of 0.5 log is not considered to e practically significant. When this was taken into account, none of the mean differences from the 10 matrixes at three levels was considered to e of any practical significance. Furthermore, there was no trend or ias for one method to consistently yield greater results than the other. Raw ground pork, raw chicken reast, fresh parsley, vegetale juice, tuna pate, deli pasta salad, and yogurt drink were also analyzed under the rapid-usage conditions (24 h incuation). Plots of the candidate method results versueference method results are presented in Figure 2. The candidate method results are lower than the reference method results in some instances and higher in others; however, there are no major apparent discrepancies to suggest that any results are aerrant. Tuna pate is the only product showing similar negative ias at all inoculum levels, whereas the candidate method is on average more than 0.5 log lower than the reference method. Outlier analysis y the Cochran test identified statistically significant evidence of high variance for the candidate method for deli salad at the low level (P = 0.02). Only two of 42 Grus tests gave statistically significant evidence of outlying values, as shown in Tale 9. Lastly, the stem-and-leaf display, letter-and-value display, and the ox plot did not show any ovious differences etween the two methods (plots not shown). There was no valid explanation to support excluding any of the outlier results; therefore, all results were included in the statistical analysis. The matrix study results and statistical analyses for the rapidusage conditions are summarized in Tale 10. From the seven matrixes at three levels, there were three instances with a mean difference >0.5 log. These are tuna pate at the high and medium levels as mentioned previously, with the reference method yielding higher results than the candidate method; and parsley at the medium level, with the candidate method yielding higher results than the reference method. The of the reference method ranged from to 0.236, and for the candidate method, the range was This shows slightly higher values for the candidate method than for the reference method, although this is of little practical microiological significance. () Matrix study: comparison to ISO reference method. (1) Methodology. In a separate study, all matrixes were compared to the ISO :2013 (6) method using 10 g test portions in a _Teramura 11

12 12 TERAMURA & BETTS: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. nnn, 2017 Figure 3. Continued. paired study. Matrixes were otained from local markets and then temperature aused as previously descried to achieve three levels of contamination. For each food matrix, a ulk sample (100 g) was weighed into a sterile stomacher ag and thoroughly mixed to ensure homogeneity. Five replicate test portions (10 g) were weighed into a stomacher ag, and 90 ml MRD was added to give a 1:10 dilution. The test portion was then homogenized in a stomacher for 1 min ± 15 s, diluted decimally in MRD, and each relevant dilution was tested for oth the candidate method and the reference method. For the International Organization for Standardization (ISO) reference method, 1 ml aliquots of each relevant dilution were aseptically transferred to a 90 mm Petri dish and poured with ml PCA. For the lowest dilution only (i.e., 10 1 dilution), duplicate plates were prepared. The plates were allowed to set, inverted, and then incuated at Tale 11. Statistically significant results from the Grus tests for alternate usage Method Matrix/contamination level Outlier, log CFU/g P Reference Prawns/low lowest Reference Deli salad/low lowest Reference Pork mince/high lowest Candidate Cream cheese/medium lowest Candidate Cream cheese/high lowest Candidate Prawns/high highest <0.001 Candidate Tuna pate/high highest Candidate Pork mince/high highest _Teramura 12

13 Tale 12. Matrix study results for alternate usage (30 C, 72 h) TERAMURA & BETTS: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. nnn, Candidate method Reference method 95% confidence interval Matrix Level n a Mean, log CFU/g Mean, log CFU/g Mean difference, Dlog CFU/g LCL c UCL d Raw ground pork Low Medium High Raw chicken reast Low Medium High Fresh parsley Low Medium High Vegetale juice Low Medium High Cooked prawns Low Medium High Tuna pate Low Medium High Sandwiches Low Medium High Pasta salad Low Medium High Yogurt drink Low Medium High Cream cheese Low Medium High a c d n = Numer of replicates. = SD of replicates. LCL = Lower confidence limit. UCL = Upper confidence limit. 30 ± 1 C for 72 ± 3 h. After the incuation period, all colonies present were counted in the plates showing etween 15 and 300 colonies per plate. The MC-Media Pad ACplus alternate method was used: From each dilution, 1 ml aliquots in singlet were aseptically pipetted into the center of the film. The covering was replaced, and the film was incuated at 30 ± 1 C for 72 ± 3 h. Following incuation, red colonies were counted and recorded up to a maximum of 300 colonies/plate. (2) Results. Raw data were transformed using log 10 [CFU per unit + (0.1)f], where f = the CFU per unit corresponding to the smallest reportale result. The smallest reportale result is 10 CFU/g or 10 CFU/mL, so the transformation was carried out with f = 10 for the whole data set. For each product at the various contamination levels, graphs were produced plotting the candidate method results versus the reference method results to look for any apparent discrepancies (see Figure 3). There were no major apparent discrepancies to suggest that any results were aerrant. Statistical outliers were determined using the Cochran and Grus tests from the R Package outliers test (5), with default values for all options. The Cochran test was performed twice, once for the reference method and once for the candidate method, identifying the product/level cominations with the highest variance. The Grus test was performed 60 times, once for each method/product/level comination, identifying individual values far from the corresponding group mean. Statistically significant evidence was assessed using the commonly accepted criterion of P < The Cochran test identified two instances of high variance: the candidate method medium level for pasta salad (P = 0.01) and the reference method low level for cream cheese (P = 0.001). Statistically significant results from the Grus test are shown in Tale 11. No justifiale cause was identified to remove these statistical outliers, so all data points were included in the final statistical analyses. The spread of the data was visualized using a stem-and-leaf display, a letter-and-value display, and a ox plot, ut these did not show _Teramura 13

14 14 TERAMURA & BETTS: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. nnn, 2017 any ovious differences etween the two methods (plots not shown). The matrix study results and statistical analyses for the alternate-usage conditions are summarized in Tale 12. For the reference method, the ranged from to 0.437, and for the candidate method, the range was This shows similar repeataility for the candidate method and the reference method. The mean difference etween methods (candidate result minueference result) ranged from to With a mean difference of >0.5 log considered to e practically significant, none of the matrixes at any level was considered to e of any practical significance from a microiological point of view. Discussion The MC-Media Pad ACplus method performed as well or etter than the relevant reference methods tested in the matrix study. In all matrixes tested, there was less than a 0.5 log 10 CFU/g difference etween the candidate and reference methods, and the repeataility of the methods was very similar. The results showed that the alternative method was very versatile in its application. It showed comparale performance to the U.S. reference methods under standard usage (48 ± 2 h at 35 ± 1 C) for all 10 matrixes and under rapid usage (24 ± 2 h at 35 ± 1 C) for seven matrixes. In addition, it showed comparale results to the ISO reference method (72 ± 3 h at 30 ± 1 C) for all 10 matrixes. The results of this validation study indicate that the MC-Media Pad ACplus method can e used for rapid and accurate enumeration of total aeroic count in a variety of foods and under a variety of test conditions. Conclusions The MC-Media Pad ACplus was shown to e equivalent to the USDA/FSIS MLG 3.02 reference method for raw chicken reast and raw ground pork; to the SMEDP Chapter 6 reference method for cream cheese and yogurt drink; to AOAC Method for parsley, vegetale juice, prawns, tuna pate, sandwiches, and pasta salad; and to ISO :2013 for raw chicken reast, raw ground pork, cream cheese, yogurt drink, parsley, vegetale juice, prawns, tuna pate, sandwiches, and pasta salad. Acknowledgments We are grateful to Sharon Brunelle (AOAC Research Institute) for her support in editing the manuscript. References (1) Official Methods of Analysis (2016) 20th Ed., AOAC INTERNATIONAL, Gaithersurg, MD, Appendix J, eoma.aoac.org (accessed on July 20, 2017) (2) U.S. Department of Agriculture, Food Safety and Inspection Service (2015) Microiology Laoratory Guideook, Chapter 3.02, Quantitative Analysis of Bacteria in Foods as Sanitary Indicators, f af d6592cc9/mlg-3.pdf? MOD=AJPERES (accessed on July 20, 2017) (3) Standard Methods for the Examination of Dairy Products (2004) 17th Ed., Chapter 6, Microiological Count Methods, APHA Press, Washington, DC, ook/ / # (accessed on July 20, 2017) (4) Official Methods of Analysis (2016) 20th Edition, AOAC INTERNATIONAL, Gaithersurg, MD, Method , accessed on July 20, 2017 (5) Komsta, L. (2011) Tests for outliers, R Package version 0.14, (accessed on July 20, 2017) (6) ISO :2013 Microiology of the food chain Horizontal method for the enumeration of microorganisms Part 1: Colony count at 30 degrees C y the pour plate technique, iso.org/standard/53728.html (accessed on July 20, 2017) _Teramura 14