Introduction to the 1995 Precollaborative Study on ColiChrome 2® Redigel®

During 1994, a precollaborative study titled “The Enumeration of Total Coliforms and Escherichia coli in Food and Dairy Products: A Comparative Study of The ColiChrome 2 Redigel Method, MPN, Petrifilm EC and VRB Redigel Methods” was completed by J.N. Roth and G.L. Bontrager of RCR Scientific. The study was submitted to the AOAC for the purpose of receiving permission to do an official collaborative study with the goal of having ColiChrome 2 Redigel being approved as an official method of the AOAC. The AOAC responded favorably with permission to proceed with a collaborative study, but at that point in time (1996), the patents and technology for the ColiChrome 2 Redigel method were purchased by 3M Company, and it became their option whether to do the study. They chose not to do the study for unexplained reasons, although it is noted that their Petrifilm EC competed for the exact same niche in the food and dairy market. The precollaborative study therefore has lain dormant since then until the reacquisition of the ColiChrome technology by RCR Scientific who sold it to Micrology Laboratories (2002-2003).

Micrology Laboratories currently serves the food and dairy industry with its line of Easygel and Coliscan media. It should be noted that the names ColiChrome 2 Redigel and Coliscan Easygel are synonymous for the same original product, so that whatever pertains to ColiChrome 2 Redigel is equally true for Coliscan Easygel.

This technology represents the first use of more than one chromogenic enzyme substrate being used in a diagnostic medium and the first patent was issued in 1993. Following the introduction of the technology by RCR Scientific, the idea was copied by most of the major media makers worldwide since the RCR patent only covered the U.S. Other companies in the U.S. developed media which superficially mimicked it (so that E. coli grew blue and other coliforms supposedly grew red), but none of those in use in the U.S. can legally duplicate the results of the original formulation.

The Enumeration of Total Coliforms and Escherichia coli in Food and Dairy Products: A Comparative Study of The ColiChrome 2 ® Redigel ® Method to MPN, Petrifilm ® EC and VRB Redigel ® Methods

J. N. Roth and G. L. Bontrager, RCR Scientific, Goshen, IN

ABSTRACT

A precollaborative study was conducted to compare a pectin gel (ColiChrome 2 Redigel) plating method to the MPN fermentation tube method, 966.23-955.24, the dry rehydratable film method (Petrifilm EC), 991.14, and the pectin gel violet red bile method (VRB Redigel), 989.11 for quantifying and differentiating total coliforms and fecal coliforms ( E. coli). The Petrifilm EC and VRB Redigel were included in the study due to the inherent lack of precision in the MPN fermentation tube method. The results of 300 samples of 20 foods were collected and statistically analyzed. In addition to the statistical analyses, the value for each sample run by the three rapid methods was compared to the MPN 95% confidence intervals for the sample. 92.7% of the total coliform samples and 94.7% of the E. coli samples for the ColiChrome 2 Redigel method fell within the MPN intervals. For the Petrifilm EC, the respective percentages were 89.7% and 93.4%, and for the VRB Redigel, the percentage was 88% for coliforms. The Duncan Multiple Range test indicated 90% and 100% of the coliform and E. coli samples respectively were not significantly different between the MPN and the ColiChrome 2 Redigel, while 75% and 90% respectively were not significantly different between the MPN and the Petrifilm EC, and 70% of coliforms were not significantly different between the MPN and VRB Redigel. The potential lack of precision in the MPN is demonstrated when comparing ColiChrome 2 Redigel to the other 2 rapid methods by the Duncan Multiple Range test. 100% and 95% of the coliforms and E. coli respectively were not significantly different between ColiChrome 2 Redigel and Petrifilm EC, while 95% of coliforms were not significantly different between ColiChrome 2 Redigel and VRB Redigel. Furthermore, when mean comparisons of each method and inoculum level were made, 93% of coliforms and 98.3% of E. coli results were not significantly different between the MPN method and ColiChrome 2 Redigel, while 90% and 93.3% respectively were not significantly different between the MPN and Petrifilm EC, and 88.3% of coliforms were not significantly different between the MPN and VRB Redigel. However, comparison of ColiChrome 2 Redigel with the other 2 rapid methods showed 98.3% of coliform results and 96.6% of E. coli results between ColiChrome 2 Redigel and Petrifilm EC were not significantly different, while 96.6% of coliform results between ColiChrome 2 Redigel and VRB Redigel were not significantly different.

INTRODUCTION

Redigel is an agar plate substitute whose only basic difference to standard agar plate media is the substitution of a low methoxyl pectin for agar as the gelling agent in the media. Therefore, various media formulations can be easily made according to the standard accepted ingredients for each individual medium. Two of the Redigel formulas (Total Count (SPC) Redigel and VRB Redigel) have undergone AOAC approved collaborative studies and are adopted official final action for use in food and dairy products respectively (1,2,3). These methods are also described in Edition 16 of Standard Methods for the Examination of Dairy Products (4). The Redigel (Pectin Gel) Method consists of two parts, the first being the liquid portion containing the ingredients plus the gelling agent, and the second part consisting of a petri dish coated with a thin gel layer containing calcium ions. When the liquid portion is poured into the pretreated petri dish, the calcium ions diffuse upward through the medium and complex with the pectin to form an agar-like calcium pectate gel.

The ColiChrome 2 Redigel method consists of a medium incorporating growth nutrients, selective factors and two substrates containing chromogenic indicators, one to detect activity and one to detect glucuronidase activity. Total coliforms produce the enzyme galactosidase which cleaves the first substrate to form an insoluble red dye. Therefore, total coliform colonies appear as red colonies in the clear medium. E. coli, in addition to producing galactosidase also produces glucuronidase which cleaves the second indicator to form an insoluble blue dye. Therefore, E. coli colonies appear purple in the clear medium due to the combination of the red and blue dyes in those colonies. Patent #5210022 for this unique method is assigned to RCR Scientific, Inc., Goshen, IN.

This study was designed in consultation with Dr. Wallace Andrews (AOAC General Referee for Food Microbiology) and the decision to compare the ColiChrome 2 Redigel Method to the MPN was based partly upon the fact that the MPN is a long used and accepted method and partly because other rapid methods had been evaluated similarly (5,6). Due to our awareness that MPN methods have been reported to exhibit significant variability (7), the authors decided to also include as part of the study a comparison to the VRB Redigel method for total coliforms and a comparison to the Petrifilm EC method for total coliforms and E. coli. Since both of these methods are also AOAC approved, we felt that such additional comparisons would be most informative and helpful.

METHODS AND MATERIALS

The study included 20 different food and dairy products representing a broad variety of the major groups. Five separate lots of each product were tested and each lot was represented by low, medium and high total coliform and E. coli levels. Therefore, 15 samples of each of the 20 products were tested by each method. Each test was done with the three tube MPN procedure and with duplicate plates in each of the other methods.

The food groups tested were as follows:

Dairy – Seafoods – Meats – Vegetables
cheddar cheese fish (frozen) beef (raw ground) broccoli (frozen)
cottage cheese oysters (fresh) meat pot pie (frozen) corn (frozen)
milk (pasteurized) shrimp (fresh) turkey (raw ground) mushrooms (fresh)
milk (raw)
yogurt

Spices – Grain Products – Nutmeat
pepper (black ground) flour peanuts (raw)
thyme rice walnuts

Preparation of Test Samples
Food samples were prepared according to instructions in AOAC 966.23B and 989.11C. The foods were spiked with a mixture of the following coliforms. (ATCC strains: E. coli 10586, 11303, 11775, 26922; Enterobacter aerogenes 13048; Enterobacter cloacae 23355; Klebsiella pneumoniae 13883; Citrobacter freundii 8090.) Cultures were grown 24 hours in tryptone broth at 35 C and then washed 3 times with Butterfield’s Phosphate Buffer. The total coliforms and E. coli were mixed to give about a 3:1 ratio of total coliforms to E. coli and sufficient of this inoculum was mixed with the sample food products to give low, ,medium and high inoculum levels. The inoculum and product were mixed in appropriate diluent (50 g product with 450 mL diluent) by blending for 2 minutes or according to other specific AOAC protocols. The actual counts are recorded in Table 5.

Microbiological Analysis of Prepared Samples
Each of the prepared samples was analyzed by the following methods: ColiChrome 2 Redigel method for total coliforms and E. coli; 3-Tube MPN method for total coliforms and E. coli; Petrifilm EC method for total coliforms and E. coli; VRB Redigel method for total coliforms.

The same sample dilutions were used as inoculum for all methods, with the goal of not exceeding 300 colonies(CFU)/plate for the methods other than MPN.

The ColiChrome 2 Redigel plates were incubated for 48 h at 35 C. Both red and purple (blue) colonies were counted at both 24 h and 48 h., but the data in this paper are all from the 24 h. counts. Red colonies were recorded as total coliforms and purple colonies were recorded as E. coli. Confirmation of total coliforms was done by picking 2 red colonies from each of the duplicate plates at 24 h. and inoculating BGLB tubes which were incubated at 35 C for 48 h. Gas production was considered positive. Confirmation of E. coli was done by picking 2 purple colonies from each duplicate plate at 24 h. and streaking onto Levine’s EMB agar plates. At 24 h, two typical isolated colonies were picked from each EMB plate and transferred to individual plate count agar slants. Following incubation, the IMViC series was run on each culture and each was tested for gas production in LST broth and gram stained. Gram negative, non-spore-forming rods producing gas in lactose broths and producing ++– or -+– IMViC patterns were counted as E. coli.

The 3 tube MPN method was done according to the described AOAC 966.24 procedure. Total coliforms were calculated from the gassing BGLB tubes and E. coli were calculated from the gassing EC broth tubes, both of which had been inoculated from originally gassing LST tubes. Further confirmation of the E. coli was done by streaking from gassing EC tubes onto Levine’s EMB agar, incubating 24 h at 35 C, and picking 2 typical colonies from each plate which were transferred to plate count slants. Following incubation, the IMViC series was run on each culture, each was tested for gas production in LST broth and gram stained. Gram negative, nonsporeforming rods producing gas in lactose broths and producing ++– or -+– IMViC patterns were counted as E. coli.

Petrifilm plates were incubated at 35 C for 48 h and colonies were counted at 24 and 48 h. Red and blue gassing colonies were counted at 24 h and the sum of the two was recorded as total coliforms. Blue gassing colonies were counted at 48 h and recorded as E. coli. Totally accurate differentiation of total coliforms and E. coli was not always obtained at 24 h. due to a small percentage of red colonies becoming blue by 48 h. The procedures were followed according to instructions provided with the product and the counts were made in accordance with the protocol established in the collaborative study of the product (6).

VRB Redigel plates were incubated at 35 C for 48 h and colonies were counted at 24 and 48 h. Red or pink colonies were counted and recorded as total coliforms. The data presented in this paper were from the 24 h. counts. Procedures were followed in accordance with the protocol established in the collaborative study of the medium (2).

All plates and tubes were incubated together in the same incubator when possible to eliminate any possible variation in results from incubation differences.

ANALYSIS OF DATA

The base 10 logarithms of the colony counts for the ColiChrome 2 Redigel, Petrifilm EC and VRB Redigel methods, and the base 10 logarithms of the MPN indices were used for statistical analyses. The complete set of data for all samples is given in Table 5. Statistical analyses were done by Mr. Foster McClure of the FDA in Washington D.C. Mean comparisons were done for each level of each of the methods (table 3) and the Duncan procedure for pairwise comparisons of the methods was also done (table 4). The analyses were based upon the 5% level of significance. This approach has been used and accepted as valid in past studies (5,6). Brief comments on each of the food products follows.

RESULTS

BROCCOLI (FROZEN)
The Duncan mean comparison for coliforms indicated no significant difference between the MPN and all the other methods. The Duncan mean comparisons for E. coli indicated no significant difference among all the methods.

The mean comparisons for each method and level indicated no significant differences among any of the methods or levels.

58 of 58 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 49 of 49 purple colonies picked as E. coli from ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

Percentage of samples falling within MPN confidence intervals for each method were as follows: ColiChrome 2 Redigel (coliforms-87%, E. coli-80%); Petrifilm EC (coliforms-87%, E. coli-73%); VRB Redigel (coliforms 93%).

CHEDDAR CHEESE
The Duncan mean comparison for coliforms indicated a significant difference between the MPN and all the other methods. However, there was no significant difference among the ColiChrome 2 Redigel, Petrifilm EC and VRB Redigel methods with the Duncan mean comparison for coliforms. The Duncan mean comparisons for E. coli indicated no significant difference among all the methods.

The mean comparisons for each method and level indicated no significant differences among any of the methods except for MPN and Petrifilm EC at the low level for E. coli.

60 of 60 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 55 of 55 purple colonies picked as E. coli from ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

Percentage of samples falling within MPN confidence intervals for each method were as follows: ColiChrome 2 Redigel (coliforms-93%, E. coli 86.7%); Petrifilm EC (coliforms-93%, E. coli-80%); VRB Redigel (coliforms-93%).

CORN (FROZEN)
The Duncan mean comparison for coliforms indicated no significant difference between the MPN and all other methods. Likewise, the Duncan mean comparisons for E. coli indicated no significant difference among all the methods.

The mean comparisons for each method and level indicated no significant differences among any of the methods except for MPN and ColiChrome 2 Redigel at low levels of coliforms and E. coli, and VRB Redigel at the low level for coliforms.

60 of 60 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 57 of 57 purple colonies picked as E. coli from ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

Percentage of samples falling within MPN confidence intervals for each method were as follows: ColiChrome 2 Redigel (coliforms-93%, E. coli-100%); Petrifilm EC (coliforms-93%, E. coli-100%); VRB Redigel (coliforms 93%).

COTTAGE CHEESE
The Duncan mean comparison for coliforms and E. coli indicated no significant difference between the MPN and all the other methods.

The mean comparisons for each method and level indicated no significant differences among the methods.

56 of 56 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 50 of 50 purple colonies picked as E. coli from ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

Percentage of samples falling within MPN confidence intervals for each method were as follows: ColiChrome 2 Redigel (coliforms-100%, E. coli-93%); Petrifilm EC (coliforms-89%, E. coli-100%); VRB Redigel (coliforms-93%).

FISH (FROZEN)
The Duncan mean comparison for coliforms indicated no significant difference between the MPN and all the other methods. The Duncan mean comparisons for E. coli indicated no significant difference among all the methods.

The mean comparisons for each method and level indicated no significant differences among any of the methods.

59 of 60 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 56 of 56 purple colonies picked as E. coli from ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

Percentage of samples falling within MPN confidence intervals for each method were as follows: ColiChrome 2 Redigel (coliforms-100%, E. coli-100%); Petrifilm EC (coliforms-93%, E. coli-100%); VRB Redigel (coliforms 93%).

FLOUR
The Duncan mean comparison for coliforms indicated no significant difference between the MPN and all other methods. Likewise, the Duncan mean comparisons for E. coli indicated no significant difference among all the methods.

The mean comparisons for each method and level indicated no significant differences among any of the methods and levels except for MPN with all the others at the low levels for coliforms.

56 of 60 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 51 of 51 purple colonies picked as E. coli from ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

Percentage of samples falling within MPN confidence intervals for each method were as follows: ColiChrome 2 Redigel (coliforms-93%, E. coli-100%); Petrifilm EC (coliforms-87%, E. coli-100%); VRB Redigel (coliforms 80%).

GROUND BEEF (RAW)
The Duncan mean comparison for coliforms indicated no significant difference between the MPN and all other methods. The Duncan mean comparisons for E. coli indicated no significant difference among all the methods with the exception of significant differences between the MPN and Petrifilm EC, and ColiChrome 2 Redigel and Petrifilm EC.

The mean comparisons for each method and level indicated no significant differences among any of the methods except for MPN and Petrifilm EC, and ColiChrome 2 Redigel and Petrifilm EC at the low level for E. coli.

60 of 60 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 56 of 56 purple colonies picked as E. coli from ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

Percentage of samples falling within MPN confidence intervals for each method were as follows: ColiChrome 2 Redigel (coliforms-93%, E. coli-100%); Petrifilm EC (coliforms-93%, E. coli-93%); VRB Redigel (coliforms 93%).

GROUND POULTRY (RAW)
The Duncan mean comparison for coliforms indicated no significant difference between the MPN and all other methods. Also, the Duncan mean comparisons for E. coli indicated no significant difference among all the methods.

The mean comparisons for each method and level indicated no significant differences among any of the methods.

60 of 60 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 57 of 57 purple colonies picked as E. coli from ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

Percentage of samples falling within MPN confidence intervals for each method were as follows: ColiChrome 2 Redigel (coliforms-93%, E. coli-100%); Petrifilm EC (coliforms-93%, E. coli-100%); VRB Redigel (coliforms 87%).

MEAT POT PIE (FROZEN)
The Duncan mean comparison for coliforms indicated no significant difference between the MPN and all other methods. Likewise, the Duncan mean comparisons for E. coli indicated no significant difference among all the methods.

The mean comparisons for each method and level indicated no significant differences among any of the methods except for the MPN and VRB Redigel at the low level for coliforms.

58 of 60 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 57 of 57 purple colonies picked as E. coli from ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

Percentage of samples falling within MPN confidence intervals for each method were as follows: ColiChrome 2 Redigel (coliforms-100%, E. coli-100%); Petrifilm EC (coliforms-93%, E. coli-100%); VRB Redigel (coliforms 80%).

MUSHROOMS (FRESH)
The Duncan mean comparison for coliforms indicated no significant difference between the MPN and ColiChrome 2 Redigel, or between the MPN and VRB Redigel methods. However, there was a significant difference between the MPN and Petrifilm EC methods with the Duncan mean comparison for coliforms. The Duncan mean comparisons for E. coli indicated no significant difference among all the methods.

The mean comparisons for each method and level indicated no significant differences among the MPN and any of the other methods except for MPN and Petrifilm EC at the medium level for coliforms. There was also a significant difference between VRB Redigel and the ColiChrome 2 Redigel and Petrifilm EC methods at the low levels.

60 of 60 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 56 of 56 purple colonies picked as E. colifrom ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

Percentage of samples falling within MPN confidence intervals for each method were as follows: ColiChrome 2 Redigel (coliforms-93%, E. coli-87%); Petrifilm EC (coliforms-93%, E. coli-93%); VRB Redigel (coliforms 93%).

OYSTERS (FRESH)
The Duncan mean comparison for coliforms indicated a significant difference between the MPN and all the other methods. However, there was no significant difference among all the other methods for both coliforms and E. coli.

The mean comparisons for each method and level indicated a significant difference between the MPN and all other methods at the low levels for coliforms, with no significant difference at high and medium levels. For E. coli, there was no significant difference among any of the methods except for MPN and Petrifilm EC at the low level.

59 of 60 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 40 of 40 purple colonies picked as E. coli from ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

Percentage of samples falling within MPN confidence intervals for each method were as follows: ColiChrome 2 Redigel (coliforms-53%, E. coli-93%); Petrifilm EC (coliforms-73%, E. coli-87%); VRB Redigel (coliforms 60%).

MILK (PASTEURIZED)
The Duncan mean comparison for coliforms indicated a significant difference between the MPN and both Petrifilm EC and VRB Redigel). However, there was no significant difference among the ColiChrome 2 Redigel, Petrifilm EC and VRB Redigel methods with the Duncan mean comparison for coliforms. The Duncan mean comparisons for E. coli indicated no significant difference among all the methods.

The mean comparisons for each method and level indicated no significant differences at the medium level, but significant differences at the high levels of coliforms for all methods compared to the MPN and at the low levels of coliforms for Petrifilm EC and VRB Redigel.

60 of 60 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 59 of 59 purple colonies picked as E. coli from ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

Percentage of samples falling within MPN confidence intervals for each method were as follows: ColiChrome 2 Redigel (coliforms-93.3%, E. coli-93%); Petrifilm EC (coliforms-93.3%, E. coli-93.3%); VRB Redigel (coliforms-93.3%).

PEANUTS (RAW)
The Duncan mean comparison for coliforms indicated no significant difference between the MPN and both the ColiChrome 2 Redigel and Petrifilm EC methods. However, there was a significant difference between the VRB Redigel and all the other methods with the Duncan mean comparison for coliforms. The Duncan mean comparisons for E. coli indicated no significant difference among all the methods.

The mean comparisons for each method and level indicated no significant differences among any of the methods except for ColiChrome 2 Redigel and Petrifilm EC with VRB Redigel at the low level for coliforms.

55 of 60 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 51 of 52 purple colonies picked as E. coli from ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

Percentage of samples falling within MPN confidence intervals for each method were as follows: ColiChrome 2 Redigel (coliforms-100%, E. coli-93%); Petrifilm EC (coliforms-100%, E. coli-100%); VRB Redigel (coliforms 80%).

PEPPER (BLACK GROUND)
The Duncan mean comparison for coliforms indicated a significant difference between the MPN and both the Petrifilm EC and VRB Redigel methods. However, there was no significant difference between the MPN and ColiChrome 2 Redigel methods or among the ColiChrome 2 Redigel, Petrifilm EC and VRB Redigel methods with the Duncan mean comparison for coliforms. The Duncan mean comparisons for E. coli indicated no significant difference among all the methods.

The mean comparisons for coliforms with each method and level indicated significant differences between the MPN and Petrifilm and VRB Redigel at the medium level, but no significant differences at the high and low levels for all methods compared to the MPN. There were no significant differences between the MPN and the other methods for E. coli.

60 of 60 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 58 of 58 purple colonies picked as E. coli from ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

Percentage of samples falling within MPN confidence intervals for each method were as follows: ColiChrome 2 Redigel (coliforms-100%, E. coli-100%); Petrifilm EC (coliforms-100%, E. coli-93%); VRB Redigel (coliforms-100%).

MILK (RAW)
None of the Duncan mean comparisons differed significantly.

Also, none of the mean comparisons for each method and level differed significantly.

59 of 59 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 53 of 53 purple colonies picked as E. coli from ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

One hundred percent of the samples for each of the methods (ColiChrome 2 Redigel, Petrifilm EC, VRB Redigel) fell within the MPN confidence intervals.

RICE
The Duncan mean comparison for coliforms and E. coli indicated no significant difference between the MPN and all other methods.

The mean comparisons for each method and level indicated no significant differences among any of the methods except for Petrifilm EC with ColiChrome 2 Redigel and VRB Redigel at the low levels for coliforms.

49 of 58 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 60 of 60 purple colonies picked as E. coli from ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

Percentage of samples falling within MPN confidence intervals for each method were as follows: ColiChrome 2 Redigel (coliforms-87%, E. coli-100%); Petrifilm EC (coliforms-100%, E. coli-93%); VRB Redigel (coliforms 80%).

SHRIMP (FRESH)
The Duncan mean comparison for coliforms indicated no significant difference between the MPN and all other methods. Likewise, the Duncan mean comparisons for E. coli indicated no significant difference among all the methods.

The mean comparisons for each method and level indicated no significant differences among any of the methods except for the MPN and Petrifilm EC at the high level of coliforms.

57 of 60 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 55 of 56 purple colonies picked as E. coli from ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

Percentage of samples falling within MPN confidence intervals for each method were as follows: ColiChrome 2 Redigel (coliforms-100%, E. coli-100%); Petrifilm EC (coliforms-93%, E. coli-100%); VRB Redigel (coliforms 93%).

THYME
The Duncan mean comparison for coliforms indicated no significant difference between the MPN and all other methods. The Duncan mean comparisons for E. coli indicated no significant difference among all the methods.

The mean comparisons for coliforms with each method and level indicated no significant differences between the MPN and all other methods. There were no significant differences between the MPN and the other methods for E. coli.

57 of 59 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 53 of 53 purple colonies picked as E. coli from ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

Percentage of samples falling within MPN confidence intervals for each method were as follows: ColiChrome 2 Redigel (coliforms-100%, E. coli-93%); Petrifilm EC (coliforms-91%, E. coli-100%); VRB Redigel (coliforms-93%).

WALNUTS
The Duncan mean comparison for coliforms indicated a significant difference between the MPN and VRB Redigel. However, there was no significant difference between the MPN and other methods for coliforms. The Duncan mean comparisons for E. coli indicated no significant differences among all the other methods.

The mean comparisons for each method and level of coliforms indicated no significant differences among any of the methods except for Petrifilm EC and VRB Redigel and at the low level. For E. coli, there were no significant differences among any of the methods.

51 of 60 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 57 of 58 purple colonies picked as E. coli from ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

Percentage of samples falling within MPN confidence intervals for each method were as follows: ColiChrome 2 Redigel (coliforms-93%, E. coli-87%); Petrifilm EC (coliforms-93%, E. coli-87%); VRB Redigel (coliforms 73%).

YOGURT
The Duncan mean comparison for coliforms indicated no significant difference between the MPN and all the other methods. The Duncan mean comparisons for E. coli indicated a significant difference between the MPN and Petrifilm EC methods.

The mean comparisons for each method and level indicated no significant differences among any of the methods except for MPN and Petrifilm EC at the low and high levels for E. coli.

58 of 58 red colonies picked as coliforms from ColiChrome 2 Redigel were confirmed and 54 of 54 purple colonies picked as E. coli from ColiChrome 2 Redigel were confirmed by EMB and IMViC reactions.

Percentage of samples falling within MPN confidence intervals for each method were as follows: ColiChrome 2 Redigel (coliforms-80%, E. coli-87%); Petrifilm EC (coliforms-73%, E. coli-80%); VRB Redigel (coliforms-87%).

DISCUSSION

This study was originally designed to satisfy the requirements of a precollaborative study leading to the approval and implementation of a collaborative study on the use of ColiChrome 2 Redigel for the differentiation of total coliforms and E. coli. The decision to base this study on a comparison of ColiChrome 2 Redigel to the standard MPN method was influenced by the fact that other recent collaborative studies have been done and approved on the basis of this format (5,6). Because it is generally accepted that the MPN method is somewhat imprecise (7), the authors decided to include two additional AOAC approved methods (Petrifilm EC and VRB Redigel) to this study to strengthen the overall results. As will be noted in the following discussion, on some occasions, the MPN results were significantly different from not only ColiChrome 2 Redigel, but also the other 2 methods used, and the ColiChrome 2 Redigel, Petrifilm EC and VRB Redigel methods were not significantly different. This kind of result raises the question of the accuracy of the MPN method for those cases.

Following are some observations and comments for Duncan’s Multiple Range Test for coliforms. In two of the foods (cheddar cheese and oysters), the MPN results were significantly different than the other three methods. However, in those cases, there was no significant difference between ColiChrome 2 Redigel and the other two methods (Petrifilm EC and VRB Redigel). The ColiChrome 2 Redigel did not differ significantly from the MPN in any other foods. There were some instances where the other 2 methods (Petrifilm EC and VRB Redigel) did differ from the MPN. Both differed from the MPN in the pasteurized milk and pepper, and the Petrifilm EC differed from the MPN in the mushrooms. The VRB Redigel also differed from the MPN in the raw peanuts and the walnuts.

For Duncan’s Multiple Range Test for E. coli, there were no significant differences between the MPN and ColiChrome 2 Redigel, and there were only 2 foods (ground beef and yogurt) where there was a significant difference between the MPN and Petrifilm.

In the general linear models procedure for coliforms, the MPN differed from all 3 other methods in two foods, but the differences were not for all levels of bacteria in each food (Flour, at low level only; oysters, at low level only). The MPN and ColiChrome 2 Redigel differed significantly in two other instances, those being in the corn at the low level and the pasteurized milk at the high level of bacteria. The MPN and Petrifilm EC differed significantly in five other instances, those being in the pasteurized milk at the low and high levels, the black pepper and mushrooms at the medium level and the shrimp at the high level. The MPN and VRB Redigel differed in corn, meat pot pie, and pasteurized milk at the low levels, the pepper at medium level and the pasteurized milk at the high level.

In the general linear model for E. coli, the MPN and ColiChrome 2 Redigel differed significantly in the corn at the low level. The MPN and Petrifilm EC differed significantly in the cheddar cheese, the ground beef and the yogurt at the low levels.

The number of samples falling within the MPN method confidence intervals was calculated for each of the methods tested and an overall high degree of agreement occurred for each method. For the ColiChrome 2 Redigel method, of 300 samples of 20 foods, 92.7% fell within MPN confidence intervals for coliforms and 94.7% fell within MPN confidence intervals for E. coli. For the Petrifilm EC method, of 294 samples of 20 foods, 89.7% fell within MPN confidence intervals for coliforms and 93.4% fell within MPN confidence intervals for E. coli. For the VRB Redigel method, of 300 samples of 20 foods, 88% fell within MPN confidence intervals for coliforms.

This study resulted in considerably more insights and information than originally expected. However, we believe that the direct comparisons of the four methods contribute significantly to questions about the methods.

It should be noted that the data presented for all three media (ColiChrome 2 Redigel, VRB Redigel, and Petrifilm EC) were taken at 24 h. incubation for the coliform counts, but that the E. coli data were taken at 24 h. for ColiChrome 2 Redigel and at 48 h. for Petrifilm EC (in accordance with protocol established in the collaborative study) (6).

Also, it should be noted that the results for the oysters were , in general, excellent for the E. coli, but that the results for the coliforms did not conform well to the MPN for any of the other methods. Even though there was no significant difference among ColiChrome 2 Redigel, VRB Redigel and Petrifilm EC, we wish to note the low counts recorded for the ColiChrome 2 Redigel and give some explanation of the results obtained. It is our experience that if low populations of coliforms occur in oysters so that 1mL of a 1:10 slurry of the oysters is plated out, there may be something present from the oyster tissue which affects the substrate/coliform interaction. The entire medium may take on a very light-pink cast, seeming to indicate a high galactosidase level in the medium from the oyster tissue. This oyster tissue activity apparently interferes with the obvious expression of red coliform CFUs. In-house studies with fresh clams and mussels have indicated that this activity is either lacking in their tissues or is at such low levels as to not interfere with excellent recovery of total coliforms with ColiChrome 2 Redigel.

It is particularly interesting to compare the three methods with the MPN method. The MPN method is outdated, cumbersome and less than precise, representing a range of estimates rather than actual counts of viable cells resulting in CFU in/on appropriate media. It is not only bulky and time consuming to set up, but confirmed results may not be available for 5-7 days.

Based upon the results of this study, we believe that any of the three rapid methods may be used effectively to differentiate and confirm coliforms or E. coli in food and dairy products. Results for coliforms can be obtained in 24 h. for all 3 methods. Results for E. coli can be obtained at 24 h. with ColiChrome 2 Redigel and at 48 h. with Petrifilm EC.

Acknowledgments
The authors thank AOAC General Referee W. H. Andrews, Food and Drug Administration, Division of Microbiology, Washington, D.C. for his direction and constructive criticisms throughout the duration of this study. Also, thanks to F.D. McClure of the Food and Drug Administration, Division of Mathematics, Washington, D.C. for performing the statistical analyses.

References
(1) Roth, J. (1988) J. Assoc. Off. Anal. Chem. 71, 343-349
(2) Roth, J. & Bontrager, G. (1989) J. Assoc. Off. Anal. Chem. 72, 298-302
(3) Official Methods of Analysis (1990) 15th Ed., AOAC., Arlington, VA, secs 988.18 – 989.11
(4) Standard Methods for the Examination of Dairy Products (1992) American Public Health Association, Washington, .DC, secs 6.3D – 7.14, 231 & 263
(5) Entis, P. (1989) J. Assoc. Off. Anal. Chem. 72, 936-949
(6) Curiale, M.S., Sons, T., McIver, D., McAllister, J.S., Halsey, B., Roblee, D., and Fox, T.L. (1991) J. Assoc. Off. Anal. Chem. 74 635-648
(7) Russek, E., and Colwell, R.R. (1983) Appl. Environ. Microbiol. 45 1646-1650