Evaluation of Commonly Used Antimicrobial Interventions forFresh Beef Inoculated with Shiga Toxinâ•fiProducing Escherichia coli Serotypes O26, O45, O103, O111, O121, O145, and O157:H7

Although numerous antimicrobial interventions targeting Escherichia coli O157:H7 have been developed and implemented to decontaminate meat and meat products during the harvesting process, the information on efficacy of these interventions against the so-called Big Six non-O157 Shiga toxin–producing E. coli (STEC) strains is limited. Prerigor beef flanks (160) were inoculated to determine if antimicrobial interventions currently used by the meat industry have a similar effect in reducing nonO157 STEC serogroups O26, O45, O103, O111, O121, and O145 compared with E. coli O157:H7. A high (10 CFU/cm) or a low (10 CFU/cm) inoculation of two cocktail mixtures was applied to surfaces of fresh beef. Cocktail mixture 1 was composed of O26, O103, O111, O145, and O157, while cocktail mixture 2 was composed of O45, O121, and O157. The inoculated fresh beef flanks were subjected to spray treatments by the following four antimicrobial compounds: acidified sodium chlorite, peroxyacetic acid, lactic acid, and hot water. High-level inoculation samples were enumerated for the remaining bacteria populations after each treatment and compared with the untreated controls, while low-level inoculation samples were chilled for 48 h at 4uC before enrichment, immunomagnetic separation, and isolation. Spray treatments with hot water were the most effective, resulting in mean pathogen reductions of 3.2 to 4.2 log CFU/cm, followed by lactic acid. Hot water and lactic acid also were the most effective interventions with the low-level inoculation on surfaces of fresh beef flanks after chilling. Peroxyacetic acid had an intermediate effect, while acidified sodium chlorite was the least effective in reducing STEC levels immediately after treatment. Results indicate that the reduction of non-O157 STEC by antimicrobial interventions on fresh beef surfaces were at least as great as for E. coli O157:H7. However, the recovery of these low inoculation levels of pathogens indicated that there is no single intervention to eliminate them. Foodborne diseases caused by microorganisms are the number one food safety concern among consumers and regulatory agencies. Illnesses attributed to foodborne microorganisms can cause severe debilitating symptoms and in some cases, these illnesses can result in death. Escherichia coli O157:H7 is a common human infectious agent globally (3), and an estimated 63,153 (24) cases of E. coli O157:H7 infection occur in the United States annually. There are other serotypes of Shiga toxin–producing E. coli (STEC) called non-O157 STEC, which cause human disease similar to that produced by E. coli O157:H7. The Centers for Disease Control and Prevention (CDC) estimated that non-O157 STEC are responsible for about 112,752 cases of illness annually (24). More than 200 virulent non-O157 serotypes have been isolated from outbreaks, sporadic cases of hemolytic uremic syndrome, and severe diarrhea in the United States and other countries (11). In the United States, six O groups (comprising 13 serotypes) have been described by the CDC to be the cause of 71% of non-O157 STEC disease (11). These serotypes have been identified as O26:H11 or nonmotile (NM); O45:H2 or NM; O103:H2, H11, H25, or NM; O111:H8 or NM; O121:H19 or H7; and O145:NM. The true number of illnesses caused by non-O157 STEC could be underestimated because detection and isolation of non-O157 STEC in stool and foodstuffs is laborious and time-consuming, with only about 4% of clinical laboratories routinely screening for these pathogens. Previous studies have shown that beef cattle hides and feces carried non-O157 STEC at a prevalence of 4.6 and 55.9%, representing a potential source of beef carcass contamination (17). Barkocy-Gallagher et al. (6) reported that the prevalence (56.6%) of non-O157 STEC on cattle hides is about the same as the prevalence of E. coli O157:H7 (60.6%). The prevalence (8%) of non-O157 STEC was reported on carcasses after the application of multiple hurdle * Author for correspondence. Tel: 402-762-4224; Fax: 402-762-4149; E-mail: [email protected]. { Mention of trade names or commercial products in this article is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. The USDA is an equal opportunity provider and employer. 1207 Journal of Food Protection, Vol. 75, No. 7, 2012, Pages 1207–1212 doi:10.4315/0362-028X.JFP-11-531 interventions (5). Bosilevac et al. (7) reported that domestic boneless beef trim used for ground beef in the United States was contaminated with non-O157 STEC at a prevalence of 30%, while prevalence in U.S. commercial ground beef was 24.3% (8). Clearly, non-O157 STEC threaten consumers’ health as well as cause economic loss because of illnesses, medical costs, or productivity losses. Furthermore, the impending implementation of U.S. Department of Agriculture, Food Safety and Inspection Service (FSIS) regulations to consider non-O157 STEC serogroups O26, O45, O103, O111, O121, and O145 as adulterants in certain raw beef products the same as for E. coli O157:H7 creates the need for approaches to control these top six non-O157 STEC. Although numerous interventions targeting E. coli O157:H7 have been developed and implemented to decontaminate meat and meat products during the harvesting process, the information on efficacy of these interventions against nonO157 STEC strains is limited. This study was designed to determine the effectiveness of existing antimicrobial interventions currently used in the meat industry for inactivating non-O157 STEC on fresh beef as compared with their effectiveness against E. coli O157:H7. MATERIALS AND METHODS Bacterial strains, growth conditions, and preparation of inocula. Strains of non-O157 E. coli serotypes O26 (:H11, 3891 and :H11, 3392, both human isolates), O45:H2 (01E-1269, human isolate), O45 WDG3 (isolated from cattle hide), O103 (:H2, 2421, human isolate), O111 (:NM, 1665, human isolate and :NM, ECRC 3007:85), O121:H19 (O1E-2074, human isolate), O121:H7 (isolated from ground beef), and O145 (:NM, GS5578620 and a ground beef isolate), and E. coli O157:H7 (ATCC 43895 and FSIS no. 4) from the U.S. Meat Animal Research Center (USMARC) culture collection were grown for 16 to 18 h at 37uC in nutrient broth (BD, Sparks, MD). Each strain was adjusted with 0.1% peptone solution to a cell concentration of approximately 1.5 | 10 CFU/ml, with a spectrophotometer value of 600 nm. Two cocktail mixtures were used in this study because there were no commercial immunomagnetic beads available for serogroups O45 and O121, making them difficult to separate from the other STEC strains during enumeration and detection. Cocktail mixture 1 consisted of an equal volume of each strain of O26, O103, O111, O145, and O157 to form a nine-strain cocktail mixture. Cocktail mixture 2 consisted of an equal volume of each strain of O45, O121, and O157 to form a six-strain mixture. These two cocktails were each diluted to approximately 1.5 | 10 or 1.5 | 10 CFU/ ml for high and low inoculations, respectively. Cocktail mixture 1 was diluted in maximum recovery diluent, while cocktail mixture 2 was diluted with purge to provide a typical background flora. Purge was aseptically collected from vacuum-packaged beef subprimals that had been stored at 220uC and then thawed at 4uC. The inocula were then placed in an ice bath while processing each day’s samples to restrict further cell growth before use. Sample collection and inoculation. Prerigor beef flanks (cutaneous trunci muscle, 16 flanks for each treatment) were collected from a local beef cattle processing plant (80 flanks for mixture 1 and 80 for mixture 2) within 25 min postexsanguination and transported to the USMARC laboratory within 2 h in insulated containers. One intervention treatment per cocktail mixture combination was processed per day. Each day 16 flanks were divided into two groups of 8 flanks. The first group was inoculated with high levels of inocula, while the second group was inoculated with low levels of inocula. Before inoculation, each flank was divided into 16 25-cm sections by using a template (10 by 10 cm), sterile cotton swab, and edible ink. An aliquot of 50 ml of either 1.5 |10 (high) or 1.5 |10 (low) CFU/ml of either cocktail mixture 1 or cocktail mixture 2 was inoculated on individual 25-cm sections, spread over the area with a sterile cell spreader, and let stand undisturbed for 15 min at room temperature to allow bacterial cell attachment before subjecting the flanks to antimicrobial treatments. The final cell concentrations on meat surfaces for low and high inoculation were approximately 5 |10 and 5 |10 CFU/cm, respectively. Intervention treatments. The antimicrobial compounds that were used in this study are generally regarded as safe approved, and the applied concentrations were within the recommended range. All the antimicrobial compounds were prepared according to the manufacturers’ recommendations. The following four antimicrobial treatments were applied to the inoculated fresh beef flank tissue for 15 s: (1) acidified sodium chlorite (1,000 ppm; Ecolab, St. Paul, MN), (2) peroxyacetic acid (200 ppm; Ecolab), (3) lactic acid (4%; Purac, Chicago, IL), and (4) hot water (85uC) with a model spray wash cabinet with three oscillating spray nozzles (SS5010, Spray Systems Co., Wheaton, IL) at 60 cycles per min. Hot water (85uC at nozzles) was sprayed at 15 lb/in, while the other antimicrobial compounds were freshly prepared with water (22 to 25uC), sprayed at 20 lb/in and dripped for 30 s. The distance between the spray nozzles and the beef flank tissue was 17 cm. For the first set before subjecting each beef flank to antimicrobial treatment, four 25-cm tissue sections were randomly excised and placed individually into four filtered bags (Whirl-Pak, Nasco, Ft. Atkinson, WI) to serve as controls. After treatments, another four 25-cm tissue sections were excised and individually placed in another four filtered bags. For the second set, eight bags (four untreated control and four treated tissue samples) were stored for 48 h at 2 to 4uC before enumeration to determine residual effect on antimicrobial treatments. The first set of bags was enumerated within 10 min after the treatments. Enumeration and culture. After 10 min posttreatment, untreated control and treated tissue samples (25-cm section) were neutralized by adding 50-ml of Dey-Engley broth (BD, Sparks, MD) supplemented with 0.3% soytone, 0.25% sodium chloride, and homogenized for 1 min with a stomacher (BagMixer 400, Interscience, Weymouth, MA). For one set of high inoculation samples, a 1-ml aliquot of each sample was transferred into a 2-ml cluster tube and serially 10-fold diluted with maximum recovery diluents (BD). Appropriate dilutions were spiral plated on differential USMARC chromogenic agar plates and were enumerated on nonselective medium for aerobic plate count (APC) by using Petrifilm (3M, St. Paul, MN). The chromogenic medium was prepared as follows: Bacto Peptone (BD), 17.0 g/liter; Proteose Peptone (BD), 3.0 g/liter; sodium chloride (Sigma, St. Louis, MO), 5.0 g/liter; crystal violet (Sigma), 1.0 mg/liter; sorbose (Sigma), 6.0 g/liter; raffinose (Sigma), 6.0 g/liter; phenol red (Sigma), 20 mg/liter; bromothymol blue (Sigma), 1.5 mg/liter; and Bacto agar (BD), 15 g/liter. The medium pH was adjusted to 7.4 ¡ 0.1 and autoclaved for 10 min at 115uC. The medium was cooled to 50uC before adding filter-sterilized bile salts no. 3 (BD), 3 g/liter; 5-bromo-4-chloro-3-indoxyl-b-D-galactopyranoside (Gold Biotechnology, St. Louis, MO), 0.05 g/liter; isopropyl-b-D-thiogalactopyranoside (Sigma), 0.05 g/liter; novobiocin (Sigma), 5 mg/liter; and 1208 KALCHAYANAND ET AL. J. Food Prot., Vol. 75, No. 7