Control of Escherichia coli O157:H7 with Sodium Metasilicate

Transcription

1 151 Journal of Food Protection, Vol. 67, No. 7, 24, Pages Copyright, International Association for Food Protection esearch Note Control of Escherichia coli O157:H7 with Sodium Metasilicate EOE H. WEBE, 1 * JUDY K. O BIEN, 2 AND FEDIC. BENDE 3 1 eorgetown Technology roup, 1124 S.W. Lynnvale Drive, Portland, Oregon 97225; 2 NP Analytical Laboratories, Checkerboard Square, St. Louis, Missouri 63164; and 3 hodia Foods, CN75 Prospect Plains oad, Cranbury, New Jersey 8512, USA MS 3-314: eceived 1 July 23/Accepted 4 February 24 ABSTACT Three intervention strategiestrisodium phosphate, lactic acid, and sodium metasilicatewere examined for their in vitro antimicrobial activities in water at room temperature against a three-strain cocktail of Escherichia coli O157:H7 and a threestrain cocktail of generic E. coli. Both initial inhibition and recovery of injured cells were monitored. When 3.% (wt/wt) lactic acid, ph 2.4, was inoculated with E. coli O157:H7 (approximately 6 log CFU/ml), viable microorganisms were recovered after a 2-min exposure to the acid. After 2 min in 1.% (wt/wt) trisodium phosphate, ph 12., no viable E. coli O157:H7 microorganisms were detected. Exposure of E. coli O157:H7 to sodium metasilicate (5 to 1 s) at concentrations as low as.6%, ph 12.1, resulted in 1% inhibition with no recoverable E. coli O157:H7. No difference in inhibition profiles was detected between the E. coli O157:H7 and generic strains, suggesting that nonpathogenic strains may be used for in-plant sodium metasilicate studies. Strategies are currently in place that involve the use of chemicals as major impediments to the outgrowth of a variety of gram-negative microorganisms, mainly Salmonella species and Escherichia coli on meat and poultry products (11, 17, 22, 23, 25). In addition to processing protocols such as carcass pasteurization, common in beef processing, both acidic and alkaline chemical treatments such as lactic acid (LA), trisodium phosphate (TSP), peroxyacetic acid, and acidified sodium chlorite are used. Of those, treatments of meat and poultry with acidic (LA) or alkaline (TSP) solutions have had the longest history of commercial use (2 7, 1, 16, 2, 26). The specific mode of action of LA and its salts on various microorganisms is unclear but appears to be the result of the free diffusion of a protonized (or undissociated) form of the molecule into the cell, where the intracellular ph is consequently lowered. To offset this ph disruption, cellular energy is expended and cell growth is decreased. In addition, it is postulated that there could also be deleterious changes in the active transport of nutrient molecules, nucleic acid replication, and enzyme system integrity (9). Some LA studies involving beef have focused on E. coli O157:H7. ansom et al. (23) evaluated the decontamination potential of water, acetic acid, acidified chlorine, LA, acidified sodium chlorite, cetylpyridinum chloride, lactoferrin B, and peroxyacetic acid on fresh beef cuts inoculated with E. coli O157:H7. Their results indicate that, after immersion for 3 s in the test solutions, the most effective treatment was 2% LA with a greater than 3-log re- * Author for correspondence. Tel: ; Fax: ; gweber@easystreet.com. duction of E. coli O157:H7. However, less than ideal results were obtained by Bacon et al. (1). They rinsed chilled beef carcasses and cuts with 1.5 to 2.5% LA and determined that LA had little effect in reducing contamination of beef carcasses, subprimal cuts, and table surfaces during fabrication. It was postulated that the efficacy of organic acid reduction of microbial populations was time dependent, and the results they observed might have been a reflection of the ability of the microorganisms to attach, penetrate, and be protected in active biofilms. Data supporting the role of biofilms and a possible role of surfactants in antimicrobial efficacy are present in the work of Calicioglu et al. (11). Fresh beef carcasses were examined for both E. coli O157: H7 and indigenous E. coli survival after spraying with either 2% LA or combined LA application after treatment with Tween 2. Tween 2 pretreatment significantly enhanced the efficacy of LA inactivation through probable removal or partial detachment of susceptible microorganisms from the meat surface. Other surface-related LA combination intervention methods have also been examined. Castelo et al. (12) studied sequential processing treatments of water, hot air, and LA on both beef and pork products. esults showed a reduction in initial inoculated population levels and subsequent growth suppression of aerobic bacteria, coliforms, and E. coli after successive spray treatments with water and 2% (vol/vol) LA. Similar treatments of prechilled beef carcasses with 2% LA reduced the inoculated levels of E. coli O157:H7 greater than 5 log units (14), and an additional rinse with 4% LA was effective in reducing E. coli an additional 2. to 2.4 log units. Total bacterial counts were reduced by over 3 log units (13). The salts of organic acids such as sodium and potassium lactate have also been postulated to have an indirect

2 152 WEBE ET AL. J. Food Prot., Vol. 67, No. 7 effect on microbial growth by altering water activity and possibly extending the lag time of susceptible microorganisms (19). Antimicrobial studies with these compounds have been less successful than with LA. Huang and Juneja (18) studied the effect of temperature and the addition of up to 4.5% sodium lactate on E. coli O157:H7 in lean ground beef. They found that temperature alone was responsible for any change in D-values. As an alternative to, and predating the widespread commercial use of LA on beef, many poultry plants have adopted the use of TSP as an intervention strategy to control gram-negative pathogens, specifically Salmonella Typhimurium. In vitro studies, however, indicate that TSP is equally effective against other gram-negative pathogens such as E. coli O157:H7 and Campylobacter jejuni (2, 26). As with LA, the specific mechanisms of action of TSP are unclear, yet TSP washes appear to be effective in physically removing and inactivating Salmonella Typhimurium from poultry (3). Antimicrobial activity of TSP has been attributed to a number of factors. For example, the significant increase in ph and the strong ability of various forms of phosphate to chelate metal ions (21) are suspected to play important roles. Scanning electron micrographs of phosphate-treated, gram-negative microorganisms demonstrate cell wall and membrane disruption (data not shown). Overall, the only consensus is that phosphate antimicrobial effects, like those seen with LA salts, are multifaceted, and gram-negative microorganisms seem to be more affected than gram-positive microorganisms. However, attaining significant antimicrobial effectiveness on poultry and beef necessitates the use of high concentrations of TSP (e.g., 1%, wt/wt), and this high usage carries with its long-term use the prospect of a negative effect on wastewater treatment. Sodium metasilicate (SMS) is similar to TSP in that it too is a strong alkali. A.1% (wt/wt) solution of SMS has a ph of approximately 11.3, a 1.% (wt/wt) solution a ph of 12.3, and a 5% (wt/wt) solution a ph of It is currently accepted as a generally recognized as safe (AS) direct human food ingredient when used as a processing aid to wash fruits, vegetables, and nuts (28). Likewise, the U.S. Department of Agriculture lists SMS (CAS reg. no ) as an acceptable ingredient processing aid in select meat and poultry applications (29). The in vitro assay design in this study allowed comparison of LA, TSP, and SMS with respect to their antimicrobial effects on a three-strain cocktail of E. coli O157: H7 and a generic nonpathogenic E. coli cocktail. These experiments were also designed to attempt to mimic times and temperatures that might be experienced in meat processing operations. MATEIALS AND METHODS Bacterial strains. Three strains of E. coli O157:H7 were used in this study: ATCC 3515, a human feces isolate with cytotoxic activity from a case of hemorrhagic colitis; ATCC 43888, an enterotoxigenic human feces isolate that does not produce Shiga-like toxin I or II or possess the genes for these toxins; and ATCC 43895, a raw hamburger meat isolate that produces Shigalike toxin I and II and was implicated in a hemorrhagic colitis outbreak. Three additional strains of E. coli (termed generic) were also used: ATCC 11775, an E. coli urine isolate; ATCC 435, deposited as E. coli var. communior, isolated from the feces of a cow with diarrhea; and ATCC 35339, E. coli, isolated from Bali steer. Preparation of the inocula. Each strain of E. coli was individually grown in Trypticase soy broth (BBL, Cockeysville, Md.) at 35 C for 18 to 24 h. The cultures were grown individually overnight, harvested by centrifugation, and resuspended back to the original volume in sterile saline or Butterfield s phosphate buffer. For the screening experiment, cultures were combined and diluted 1:1 in saline to obtain an approximate 6-log CFU/ml working inoculum. Chemicals and treatments. Food-grade anhydrous SMS and TSP dodecahydrate were provided by hodia Foods (Cranbury, N.J.). For screening experiment A, LA was provided by Sigma (St. Louis, Mo.). In MIC experiment B, LA was obtained from a meat processing plant and supplied as an 88% solution (Purac of America, Lincolnshire, Ill.). Stock treatment solutions were made in sterile, deionized water. Where indicated, sterile Butterfield s phosphate buffer was used as a positive control. Chemical treatment studies of bacterial suspensions with SMS, LA, and TSP were done at room temperature (2 to 21 C) and were typically 1 ml in volume. Chemical treatment studies of E. coli O157:H7 with TSP, LA, or SMS. A three-strain cocktail of E. coli O157:H7 was treated with TSP, LA, or SMS at multiple concentrations during two time periods: a brief exposure of 5 to 1 s and 2 min. For each test condition, duplicate 9-ml samples were prepared, inoculated with 1. ml of the stock E. coli O157:H7 cocktail, and immediately vortexed. Within 5 to 1 s after the inoculation (brief exposure), 1.-ml samples were removed, diluted in 9. ml of Butterfield s phosphate buffer, and enumerated on plate count agar (PCA; BBL). After a total of 2 min had elapsed, a subsequent 1.-ml sample was removed, diluted in Butterfield s phosphate buffer, and enumerated on PCA. ecovery of sublethally injured microorganisms was monitored at each exposure time by enrichment of 1.-ml samples for up to 72 h in brain heart infusion (BHI; BBL) broth (see below). ecovery of injured survivors. In order to test for the recovery of viable but injured microorganisms, a 1-ml sample from each of the intervention test conditions was transferred to 9 ml of BHI broth acting as resuscitation medium and incubated at 35 C for up to 72 h. At 24-h intervals, each BHI recovery tube was scored ( / ) for turbidity. Those showing visible growth were enumerated on PCA after incubation at 35 C for 48 h. To confirm the presence of E. coli O157:H7, bacterial isolates from the PCA plates were transferred to sorbitol MacConkey agar (BBL). Protocols for analyses of E. coli O157:H7 concentration and SMS MIC. A2 (double-strength) SMS microtiter plate was prepared in the following manner. Each well of a microtiter plate was initially filled with 12 l of distilled water. One hundred twenty microliters of 2% (wt/wt) SMS was added to the far left column of wells. With the use of a multichannel pipetter, serial doubling dilutions were performed from left to right in the microtiter plate, leaving the far right column unmodified as the control column. One hundred twenty microliters was discarded from the last dilution column before the control column. Total volume per well in this 2 SMS plate was 12 l. An overnight cocktail of E. coli O157:H7 was prepared as above, and seven 1-fold serial dilutions were made in Butter-

3 J. Food Prot., Vol. 67, No. 7 CONTOL OF E. COLI O157:H7 WITH SODIUM METASILICATE 153 field s phosphate buffer. Each dilution of the E. coli O157:H7 cocktail was used to inoculate a separate row of a 2 SMS microtiter plate. After 2 min of incubation at room temperature, the contents of each well were removed and diluted 1:1 in BHI resuscitation medium. rowth was monitored as described above. SMS MIC protocols for E. coli generic and O157:H7 strains on microtiter plates. A2 SMS microtiter plate was prepared as previously described, except that 16% SMS was used as the stock solution, resulting in a concentration of SMS in the far left column of 8% prior to inoculation. Each well in the first, third, and fifth rows of the 2 SMS microtiter plate was inoculated with an equal volume (12 l) of an E. coli O157:H7 threestrain cocktail. Similarly, the second, fourth, and sixth rows were inoculated with an equal volume of the generic E. coli three-strain cocktail. The E. coli SMS solutions were incubated at room temperature for 2 min, after which the content of each well was removed and diluted 1:1 into BHI broth. The BHI broth tubes were incubated at 35 C for 72 h and checked for growth (turbidity) at 24-h intervals as above. Statistical analysis. Each set of microbiological data was analyzed with SAS software (27). SAS was used to invoke logistic regression, generate model coefficients, and provide a model to predict the performance of SMS on microbial survival. For the experimental conditions presented, the probability of growth or survival for a given concentration of SMS was defined as P 1/ (1 e t ), where t SMS dilution. ESULTS AND DISCUSSION Experiment A: preliminary in vitro analyses of the effectiveness of TSP, LA, and SMS on the viability of E. coli O157:H7. Preliminary in vitro experimental protocols were initiated in order to approximate the time conditions that might resemble conditions in a meat processing environment and to compare the effects of three individual intervention strategies on their inability to inactivate a three-strain cocktail of E. coli O157:H7 and prevent the recovery of viable microorganisms. Two separate times were chosen. The first, 5 to 1 s followed quickly by a water rinse, was an exposure meant to mimic the immediate effect on the test organisms of exposure to TSP, LA, or SMS. A second exposure time (2 min) was chosen to simulate a time comparable to that in beef processing, wherein a final treatment would remain on the carcass up to and including several minutes spent in the first cooler (or hot box ). Table 1 shows the inhibition and recovery results of duplicate in vitro assays of approximately 5 log CFU/ml E. coli O157:H7 exposure to TSP concentrations of.5 and 1.%, LA concentrations of 1.5 and 3.%, and eight SMS concentrations ranging from.1 to 1.%. Of the TSP treatments, only a 2-min exposure to 1.% TSP resulted in more than 99.99% growth inhibition and no recovery of injured cells. Brief exposure (5 to 1 s) to 1.% TSP resulted in more than 99% growth inhibition but subsequent recovery of injured cells. Similarly,.5% TSP exposure for 2 min reduced the inoculum by greater than 4 log CFU/ ml, but again injured cells were recovered. The results with LA indicated that brief exposure of approximately 5 log CFU/ml E. coli O157:H7 to 1.5% LA had virtually no inhibitory effect on the microorganisms. After a 2-min exposure to 1.5% LA, a 9% inhibition was observed. There was no apparent inhibition of E. coli O157: H7 after 5 to 1 s exposure to 3% LA. A 2-min exposure to 3% LA resulted in an initial inhibition of more than 99%, but as was seen with the 2-min exposure to.5% TSP, injured cells were recovered. Concentrations of SMS from.1 to 1.% were tested for their ability to inhibit E. coli O157:H7. Even though exposure (5 to 1 s) of E. coli O157:H7 to.3% or greater SMS resulted in an initial 5-log growth inhibition, injured cells were recovered. Brief exposure (5 to 1 s) to a minimum concentration of.6% SMS was needed to destroy 5 log units of E. coli O157:H7 with no recovery of viable cells. For exposure times of 2 min, a minimum effective SMS concentration of.4% was necessary to result in nonrecoverable E. coli O157:H7. Although precautions should be heeded in predictive modeling that uses data such as these, because of the small data set and short exposure times, a statistical analysis of the data was performed, and a model was developed that could be used to predict whether microbial survival would occur at various levels of SMS. For the experimental conditions presented, the probability of survival was defined as P 1/(1 e t ), where t SMS dilution. It was determined that if the SMS concentration is.8%, the estimated probability of E. coli O157:H7 survival is 4.37 log units. Experiment B: determining SMS MICs for varying levels of E. coli O157:H7 inoculation. A two-phase microtiter plate was prepared containing SMS dilutions from.5 to.3% (wt/wt) and inoculations of E. coli O157:H7 from approximately 6 to 1 log CFU/ml. Table 2 shows the results of E. coli O157:H7 recovery after a 2-min exposure to SMS. A threshold concentration of.13% SMS was needed to inhibit an E. coli O157:H7 level of 2.84 log CFU/ ml or less. E. coli O157:H7 levels in excess of 2.84 log CFU/ml required.25% SMS. Experiment C: comparison of SMS inhibition between generic E. coli and E. coli O157:H7 after a 2- min exposure. Table 3 presents inhibition and recovery profiles of duplicate experiments containing alternate inoculations of a three-strain generic E. coli cocktail and a cocktail of E. coli O157:H7, each at 7 log CFU/ml after a 2-min treatment with SMS. All E. coli strains were inhibited by the same concentration of SMS (.25%), suggesting that indicator E. coli could act as markers for further studies utilizing nonpathogenic E. coli. The MIC data in Tables 2 and 3 showed a MIC concentration of.25% SMS with no apparent recoverable organisms, whereas the data in Table 1 demonstrate recoverable E. coli O157:H7 after a 2-min exposure to.3% SMS. This apparent discrepancy can be explained by the respective reaction volumes and viable cell numbers tested in each experiment. For there to be no recorded viability in the 2- l microtiter wells in Tables 2 and 3, the viable (or recoverable) cell concentration could be less than 5 CFU/ml (e.g., 4 CFU/ml, or.8 CFU/2 l). At a comparable level of 4 CFU/ml, growth would be demonstrable in the 1-ml reaction vessels of Table 1.

4 154 WEBE ET AL. J. Food Prot., Vol. 67, No. 7 TABLE 1. Inhibition and recovery of E. coli O157:H7 after brief (5 to 1 s) and 2-min exposure at room temperature to trisodium phosphate, sodium metasilicate, and lactic acid Treatment Exposure (min) a Aerobic plate count (log CFU/ml) b % inhibition c ecovery in BHI b Control TSP (%) N SMS (%) N N N N N N N N Lactic acid (%) a Zero () minutes indicates 5 to 1 s of exposure. b Means of replicate results, n 2;, recovery; N, no recovery. c % inhibition [(control treatment)/control] 1. TABLE 2. Determination of sodium metasilicate MIC (%) for increasing levels of E. coli O157:H7 after a 2-min exposure E. coli O157:H7 (log CFU/ml) E. coli O157:H7 growth a.5%.25%.13%.6%.3% % a, no growth in either BHI broth or on tryptic soy agar (TSA) plates;, growth (as observed by turbidity in BHI broth and confirmation on TSA or sorbitol MacConkey agar). ecovery data are based on replicate results (n 2).

5 J. Food Prot., Vol. 67, No. 7 CONTOL OF E. COLI O157:H7 WITH SODIUM METASILICATE 155 TABLE 3. Determination of sodium metasilicate MICs (%) for approximately 7 log CFU/ml generic E. coli and E. coli O157:H7 after a 2-min exposure E. coli growth b Inoculum a 4.% 2.% 1.%.5%.25%.13% % E157 E. coli E157 E. coli E157 E. coli /NC /NC a E175, E. coli O157:H7 three-strain cocktail; E. coli, generic E. coli three-strain cocktail. b, no growth in either BHI broth or on tryptic soy agar (TSA) plates;, growth (as observed by turbidity in BHI broth and confirmation on TSA or sorbitol MacConkey agar); NC, negative control. ecovery data are based on replicate results. Even though LA treatment in beef and pork processing operations has been promoted as a viable method to minimize the growth of selective pathogens, there might be some concern as to acid adaptation of E. coli O157:H7 (8), its influence on subsequent tolerance to additional antimicrobials (15), and the microbial integrity of the plant environment (24). On the basis of the possibility that SMS will be used in commercial operations also employing LA as a microbial hurdle, the risk that brief exposure of E. coli O157:H7 to LA might affect the microorganisms ability to survive subsequent SMS treatment will be addressed in later studies. Our data indicate that SMS appears to be a viable and efficient alternative to LA and TSP treatment with respect to control of E. coli O157:H7. In addition, we have been able to demonstrate that indicator E. coli might be a substitute for E. coli O157:H7 for SMS in-plant studies in facilities that are restricted from using or are unable to use pathogens. ACKNOWLEDMENTS The authors thank Dr. John Norback of the University of Wisconsin Madison for consultation and assistance in the statistical analysis of the data and Walt Jones of NP Analytical for his excellent technical assistance. EFEENCES 1. Bacon,. T., J. N. Sofos, K. E. Belk, and. C. Smith. 22. Commercial application of lactic acid to reduce bacterial populations on chilled beef carcasses, subprimal cuts and table surfaces during fabrication. Dairy Food Environ. Sanit. 22: Bender, F.., and E. Brotsky Process for treating poultry carcasses to control salmonellae growth. U.S. patent no. 5,143, Bender, F.., and E. Brotsky Process for treating red meat to control bacterial contamination and/or growth. U.S. patent no. 5,192, Bender, F.., and E. Brotsky Process for treating red meat to control bacterial contamination and/or growth. U.S. patent no. 5,268, Bender, F.., and E. Brotsky Process for treating poultry carcasses to control bacterial contamination and/or growth. U.S. patent no. 5,283, Bender, F.., J. T. Elfstrum, and W. E. Swartz Process for treating poultry carcasses to increase shelf life. U.S. patent no. 5,512, Bender, F.., and W. Frees Process for treating red meat, poultry or seafood to control bacterial contamination and/or growth. U.S. patent no. 5,635, Berry, E. D., and C. Cutter. 2. Effects of acid adaptation of Eschericia coli O157:H7 on efficacy of acetic acid spray washes to decontaminate beef carcasses. Appl. Environ. Microbiol. 66: Bogaert, J.-C., and A. S. Naidu. 2. Lactic acid, p In A. S. Naidu (ed.), Natural food antimicrobial systems. CC Press, Boca aton, Fla. 1. Brotsky, E., and F.. Bender Process for treating poultry carcasses to control salmonellae growth. U.S. patent no. 5,69, Calicioglu, M., C. W. Kaspar, D.. Buege, and J. B. Luchansky. 22. Effectiveness of spraying with Tween 2 and lactic acid in decontaminating inoculated Escherichia coli O157:H7 and indigenous Escherichia coli biotype I on beef. J. Food Prot. 65: Castelo, M., M. Koohmaraie, and E. D. Berry. 21. Microbial and quality attributes of ground pork prepared from commercial pork trim treated with combination intervention processes. J. Food Prot. 64: Castillo, A., L. M. Lucia, I. Mercado, and.. Acuff. 21. Inplant evaluation of a lactic acid treatment for reduction of bacteria on chilled beef carcasses. J. Food Prot. 64: Castillo, A., L. M. Lucia, D. B. oberson, T. H. Stevenson, I. Mercado, and.. Acuff. 21. Lactic acid sprays reduce bacterial pathogens on cold beef carcass surfaces and in subsequent produced ground beef. J. Food Prot. 64: Cheng, H.-Y., H.-Y. Yang, and C.-C. Chou. 22. Influence of acid adaptation on the tolerance of Escherichia coli O157:H7 to some subsequent stresses. J. Food Prot. 65: Dickson, J. S., C.. Nettles-Cutter, and.. Siragusa Antimicrobial effects of trisodium phosphate against bacteria attached to beef tissue. J. Food Prot. 57: Dorsa, W. J., C. N. Cutter, and.. Siragusa Effects of acetic acid, lactic acid and trisodium phosphate on the microflora of refrigerated beef carcass surface tissues inoculated with Escherichia coli O157:H7, Listeria innocua, and Clostridium sporogenes. J. Food Prot. 6: Huang, L., and V. K. Juneja. 23. Thermal inactivation of Escherichia coli O157:H7 in ground beef supplemented with sodium lactate. J. Food Prot. 66: Jordan, K. N., and K. W. Davies. 21. Sodium chloride enhances recovery and growth of acid-stressed E. coli O157:H7. Lett. Appl. Microbiol. 32: Kim, J.-W., and M. F. Slavik Trisodium phosphate (TSP) treatment of beef surfaces to reduce Escherichia coli O157:H7 and Salmonella typhimurium. J. Food Sci. 59: Molins,. A Phosphates in food. CC Press, Boca aton, Fla. 22. Mustapha, A., T. Ariyapitipun, and A. D. Clarke. 22. Survival of Escherichia coli O157:H7 on vacuum-packaged raw beef treated with polylactic acid, lactic acid, and nisin. J. Food Sci. 67: ansom, J.., K. E. Belk, J. N. Sofos, J. D. Stoopforth, J. A. Scanga, and. C. Smith. 23. Comparison of intervention technologies

6 156 WEBE ET AL. J. Food Prot., Vol. 67, No. 7 for reducing Escherichia coli O157:H7 on beef cuts and trimmings. Food Prot. Trends 23: Samelis, J., J. N. Sofos, P. A. Kendall, and. C. Smith. 22. Effect of acid adaptation on survival of Escherichia coli O157:H7 in meat decontamination washing fluids and potential effects of organic acid interventions on the microbial ecology of the meat plant environment. J. Food Prot. 65: Smulders, F. J. M., and.. reer Integrating microbial decontamination with organic acids in HACCP programmes for muscle foods: prospects and controversies. Int. J. Food Microbiol. 44: Somers, E. B., J. L. Schoeni, and A. C. Wong Effect of trisodium phosphate on biofilm and planktonic cells of Campylobacter jejuni, Escherichia coli O157:H7, Listeria monocytogenes and Salmonella typhimurium. Int. J. Food Microbiol. 22: Statistical Analysis Systems. 21. Statistical analysis software, release 8.2. SAS Institute Inc., Cary, N.C. 28. U.S. Food and Drug Administration, Code of Federal egulations. 23. Sodium metasilicate. 21 CF a. 29. U.S. Food and Drug Administration, Code of Federal egulations. 23. Use of food ingredients and sources of radiation. 9 CF Yoon, K. S., and T. P. Oscar. 22. Survival of Salmonella typhimurium on sterile ground chicken breast patties after washing with salt and phosphates and during refrigerated and frozen storage. J. Food Sci. 67: