Soleris®Enterobacteriaceae for the Detection of Enterobacteriaceae in Select Foods: AOAC Performance Tested MethodSM 121901

Abstract Background Soleris®Enterobacteriaceae is a growth-based, automated method for detection of Enterobacteriaceae in food. Objective A study was conducted to validate the Soleris method for detection of Enterobacteriaceae in select foods (pasteurized milk, yogurt, mozzarella cheese, ice cream, dried milk, pasteurized liquid egg, frozen cooked chicken, deli ham, lettuce, and dry dog food) at a threshold of ≥ 10 CFU/g of product. Methods Inclusivity and exclusivity of the Soleris method were assessed by testing 55 and 38 target and non-target bacterial strains, respectively. Matrix testing was performed with one naturally contaminated and nine inoculated foods. Efficacy of the Soleris method was compared to that of the ISO 21528-2:2017 direct plating reference method using probability of detection analysis. Independent laboratory testing was conducted to verify method performance in two matrixes (yogurt and deli ham). Method robustness, stability, and lot-to-lot consistency of the Soleris reagents were also assessed. Results Inclusivity of the Soleris test was 91% and exclusivity was 100%. In matrix testing, there were no significant differences in the number of positive results obtained with the Soleris and reference methods for any of the matrixes examined. Overall, of 370 test portions, there were 176 positive results by the Soleris method and 177 positive results by the reference procedure. Conclusions Soleris Enterobacteriaceae is an effective method for detection of Enterobacteriaceae in the foods evaluated, with performance equivalent to that of the ISO 21528-2:2017 reference method. Highlights The Soleris method offers the advantages of labor savings and results within 18 h.

Soleris Enterobacteriaceae is a member of a large family of tests in the Soleris platform. Several Soleris methods have received AOAC Performance Tested Method certification, including methods for total viable count (2,3), coliforms (4), Escherichia coli (5), and yeast and mold (6,7). Here we report results of a study designed to validate the performance of the Soleris method for detection of EBAC in select foods at levels ! 10 CFU/g. Soleris method performance was compared to that of the ISO 21528-2:2017 reference method (1), which is based on a conventional colony count technique. The study was conducted in accordance with the current AOAC International Methods Committee Guidelines for Validation of Microbiological Methods for Food and Environmental Surfaces (8). Definitions (a) Probability of Detection (POD).-The proportion of positive analytical outcomes for a qualitative method for a given matrix at a given analyte level or concentration. POD is concentration dependent. Several POD measures can be calculated: POD R (reference method POD), POD C (confirmed candidate method POD), POD CP (candidate method presumptive result POD), and POD CC (candidate method confirmation result POD). (b) Difference of Probabilities of Detection (dPOD).-Difference of probabilities of detection is the difference between any two POD values. If the confidence interval of a dPOD does not contain zero, then the difference is statistically significant at the 5% level.

Principle
The Soleris vial is comprised of an upper portion containing a selective growth medium and a pH indicator, and a lower detection portion containing a matrix which excludes particulates but allows diffusion of gasses and small molecules. The vial contains a peptone yeast extract base with glucose as the carbon source. The selective agents include bile salts, sodium lauryl sulfate, and other Gram-positive inhibitors. The Soleris instrument is comprised of temperature-controlled chambers and optical sensors which monitor the color in the detection portion of the vial over time. An aliquot of a test sample homogenate or further dilution is introduced into the Soleris vial. The vial is capped and placed into the Soleris instrument programmed with specific test parameters including temperature and test duration. As EBAC grow and ferment glucose in the vial, the pH is reduced and the indicator color changes from purple to yellow. This change occurs in both the growth and detection portions of the vial. When a color change of a specific magnitude is detected, the instrument signals the test result as positive. If no change is detected within 18 h, the test result is reported as negative. Culture confirmation of Soleris results may be conducted by sampling from the upper chamber of the vial when the test is complete.

Materials and Methods
Test Kit Information

Sample Preparation
(a) Combine 10 g sample and 90 mL sterile buffered peptone water in a stomacher-type bag, homogenize thoroughly. (b) Check pH and adjust if necessary, to pH 7.0 6 1.0. (c) For testing at a threshold level of ! 10 CFU/g, the sample homogenate is used without further dilution. For testing at higher threshold levels, prepare the appropriate dilution in buffered peptone water.

Recommended Confirmation Procedure
Positive results may be confirmed by streaking the vial contents to violet red bile glucose agar and continuing with identification of presumptive Enterobacteriaceae colonies using standard methods (1).

Internal Validation Studies
Inclusivity Testing (a) Methodology.-Inclusivity testing was conducted using 55 bacterial species of the family Enterobacteriaceae. Strains were grown in nutrient broth overnight at 37 6 1 C and then diluted to approximately 100 CFU/mL (100 times the limit of detection of the Soleris method). One mL was introduced to the Soleris vial and the test run on the Soleris instrument for 18 h at 36 6 1 C. Strains were randomized, blind coded, and intermixed with exclusivity strains. (b) Results.-Results are shown in Table 1. Fifty of the 55 strains (91%) produced a positive result within 18 h. The five organisms that showed no detection within 18 h were Buttiauxella warmboldiae, one of two strains of Pantoea agglomerans, Serratia grimesii, Serratia protemaculans, and Yersinia enterocolitica. Three of the five strains were detected outside of the 18 h test duration (see Table 1).  Table 2. Of the 38 strains tested, all produced no detection within 18 h for exclusivity of 100%.

Matrix Testing
(a) Methodology.-Performance of the Soleris EBAC method at a threshold level of ! 10 CFU/g was compared to that of the ISO 21528-2:2017 reference colony count method in testing of 9 food matrixes. A tenth matrix was tested at a higher threshold level. The same amount from each test portion (1 mL of a 1:10 food sample homogenate, or 0.1 g) was used for both the Soleris and reference methods, therefore the two methods have the same theoretical detection limit. For the reference method, plate counts were scored for each test portion. For comparison to Soleris results at the ! 10 CFU/g threshold, plate counts ! 10 CFU/g were scored as positive and those < 10 CFU/g were scored as negative. The number of positive results obtained by the two methods was compared using POD analysis.
(1) Sample preparation.-Food matrixes and inoculation organisms are shown in Table 3. Levels shown in CFU/ g reflect mean results of the reference method plate counts. Lettuce with naturally occurring EBAC was available, but all other matrixes required inoculation. As the lettuce contained EBAC at a high level (approximately 5Â10 5 cfu/g), the test threshold for this matrix was set at ! 100 000 CFU/g by making further dilutions of the sample homogenate. A liquid inoculum was used for all foods except dried milk which was   Table 3. Results for the Soleris confirmed and reference methods are shown in Table 4. At the fractional level, inoculation levels determined from the mean reference method plate counts ranged from 3 to 22 CFU/g. These levels are consistent with the fractional positive data sets obtained at the ! 10 CFU/g test threshold level. Inoculation levels for the high-level test portions ranged from 12 to 218 CFU/g. The mean reference method plate count for naturally occurring EBAC in lettuce was 4.7 Â 10 5 CFU/g.
Soleris presumptive and Soleris confirmed results were identical; there were no unconfirmed positive results by the Soleris test (Table 3). Comparing the Soleris and reference methods, out of 220 fractional-level results for the 10 matrixes combined, there were 98 positive results by the Soleris method and 100 positive results by the reference plating method (Table 4). Using an unpaired POD test at P < 0.05, at the fractional level there were no significant differences in the number of positive results obtained by the Soleris and reference methods for any of the 10 matrixes examined. At the high level, of 45 test portions (there were no high-level test portions for lettuce), there were 43 positives by each method, with no significant differences between methods for any matrix. There were no positive results on uninoculated test portions by either method.

Robustness Testing
(a) Methodology.-The effect of modest perturbations introduced to Soleris operating parameters was studied in a robustness experiment. Variations were introduced simultaneously to three operating parameters (sample volume, temperature, and test duration) in a matrix of nine test conditions ( Table 5). The ninth condition represents the standard conditions for the Soleris EBAC test. Test samples included an E. coli culture dilution at 1-5 CFU/vial (positive) and a Pseudomonas aeruginosa culture dilution at approximately 1 Â 10 5 CFU/vial (negative). Ten replicate tests were performed for each sample type under each of the nine conditions. The number of positive results at each of the eight conditions containing variations to normal operating parameters were compared to the number of positive results at the standard condition by unpaired POD analysis at P < 0.05. (b) Results.-Results are shown in Table 5. For the negative sample, all Soleris tests were negative for all conditions. For the positive sample, the standard condition produced Stability and Lot-to-Lot Consistency Testing

Independent Laboratory Study
(a) Methodology.-Performance of the Soleris EBAC method was verified in testing of two matrixes by the independent laboratory. Yogurt and deli ham were tested using procedures consistent with those employed in in-house testing. (b) Results.-Soleris presumptive and confirmed results are shown in Table 3, while Soleris and reference method results are shown in Table 4. For yogurt, at the fractional level, there were 10 positive Soleris results, and all were confirmed by oxidase and glucose fermentation tests. There were 9 positive results by the reference method. This difference is not significant by unpaired POD analysis at P < 0.05. All high-level test portions were positive and all uninoculated control portions were negative by both methods. For deli ham, there were 15 Soleris positive results at the fractional level, and all were confirmed. There were also 15 positive results by the reference method. All highlevel test portions were positive and all uninoculated control portions were negative by both methods. These results confirm the efficacy of the Soleris EBAC method for these two matrixes.

Discussion
Results of this validation study demonstrate that the Soleris EBAC method is an accurate and effective procedure for detection of EBAC in a variety of foods. Inclusivity was 91% for target bacteria tested and exclusivity was 100%. Strains of five organisms (Buttiauxella warmboldiae, one of two strains of Pantoea agglomerans, Serratia grimesii, Serratia protemaculans, and Yersinia enterocolitica) were not detected within 18 h by the Soleris test. In repeat testing, these strains were again not detected. An additional strain of Yersinia enterocolitica (ATCC 27729) was tested and produced a positive result, with a detection time of 17.4 h (data not shown). Eleven additional ATCC strains of Pantoea agglomerans were tested; nine were positive with detection times ranging from 8.6 to 16.3 h (data not shown). Results of the additional testing indicate that the original results were strain-specific and not necessarily indicative of the response of these organisms in the Soleris test. An additional strain of Serratia grimesii (ATCC 14460) was tested and again produced no detection within 18 h using the standard test parameters. This strain was also tested with the Soleris method using a temperature of 30 C rather than the normal 36 C. A positive result was obtained with a detection time of 15.0 h (data not shown). Temperature sensitivity may also explain the negative results obtained with Serratia proteamaculans and Buttiauxella warmboldiae; both of these organisms have been described as having optimal growth temperatures of 30 C or below in liquid media (9-11).
Considering the in-house and independent laboratory matrix testing data combined, there were 176 positive results by the Soleris method and 177 positive results by the ISO 21528-2:2017 reference plating method. In 12 matrix trials, there were no significant differences in results between the Soleris and reference methods as determined by POD analysis at P < 0.05.
Robustness testing established that the Soleris method can withstand modest variation to three critical test parameters simultaneously. Real-time stability testing results support expiration dating for the Soleris EBAC vials of 6 months from date of manufacture.
In this study, all matrixes except lettuce required inoculation with EBAC and all were tested at a positive/negative test threshold of ! 10 CFU/g. Lettuce contained naturally occurring EBAC at a high level and was tested at a threshold of ! 1 Â 10 5 CFU/g. Test thresholds for the Soleris method can be adjusted to any level to match product specifications for EBAC. In addition to this flexibility, the Soleris method offers labor savings and decreased analysis time in comparison to the reference plating method. Soleris results are available within 18 h, while the reference method requires 22 h to produce negative results, and a minimum of an additional 44 h to produce a confirmed positive result.

Conclusions
Based on results of the validation study reported herein, it is recommended that the Soleris Enterobacteriaceae test be granted AOAC Performance Tested Method status for detection of Enterobacteriaceae in pasteurized milk, yogurt, mozzarella cheese, ice cream, dried milk, pasteurized liquid egg, frozen cooked chicken, deli ham, lettuce, and dry dog food.