Comparison of the Efficacy of a Sulfuric Acid–Sodium Sulfate Blend and Lactic Acid for the Reduction of Salmonella on Prerigor Beef Carcass Surface Tissue

More than 1 million illnesses occur each year in the United States due to ingestion of food contaminated with nontyphoidal Salmonella (19). Moreover, this bacterial pathogen is estimated to be the leading cause of foodborne illness–associated hospitalizations and deaths (19). According to latest surveillance data compiled by the Centers for Disease Control and Prevention, Salmonella was the confirmed etiologic agent of 140 foodborne outbreaks in 2014 (7).

Cattle are one of the known reservoirs of Salmonella, and numerous studies (2, 4, 5, 14, 18, 20) have reported on the prevalence of this pathogen in or on cattle hide and fecal samples, carcass surfaces before and after exposure to various decontamination treatments, beef trimmings, lymph nodes, and ground beef. In 2015, Salmonella was recovered from 1.32% of 1,290 bench trim samples, 1.74% of 3,218 manufacturing trim samples, and 2.91% of 11,047 ground beef samples collected from establishments regulated by the U.S. Department of Agriculture, Food Safety and Inspection Service (24). Over 38 years (1973 to 2011), consumption of beef was associated with 96 salmonellosis outbreaks, with ground beef as the implicated vehicle in 22 (23%) of these outbreaks (15). Of particular concern are foodborne-related infections with Salmonella strains that are multidrug resistant. Infection with multidrug-resistant Salmonella may be associated with an increased risk of hospitalization and treatment failure of infected individuals (6, 23). In 2011, consumption of ground beef contaminated with a Salmonella Typhimurium strain that was resistant to eight antibiotics resulted in 20 cases of salmonellosis across seven states (6). The hospitalization rate in this outbreak was 47% (6).

Various chemical decontamination interventions are used by beef processing facilities in North America for pathogen control on surfaces of carcasses, cuts, and beef trimmings. The efficacy of these antimicrobial chemicals in reducing Salmonella contamination levels, as well as contamination levels of Escherichia coli O157:H7 and non-O157 Shiga toxin–producing E. coli, on beef tissue surfaces has been extensively evaluated (1, 8–13, 17, 21, 26). Among these chemicals, lactic acid (LA) is one of the most commonly used organic acids in commercial beef processing plants, with numerous studies demonstrating its antimicrobial effect (1, 8–10, 12, 17, 21, 26, 27). Nevertheless, the beef industry continues to look for novel chemical treatments for use as part of a multiple hurdle system to effectively control pathogen contamination on carcasses and resulting beef products.

A commercially available blend of sulfuric acid and sodium sulfate (SSS), which was recently shown to effectively reduce inoculated Salmonella populations on chicken wings (22), is approved for use as an antimicrobial treatment of meat and poultry products (25). Published data on the efficacy of SSS against Salmonella on beef carcass surface tissue are limited. Therefore, the objective of this study was to evaluate the efficacy of SSS, applied at two temperatures, against Salmonella inoculated onto prerigor beef carcass surface tissue and to compare its efficacy to LA.

The inoculum used in this study consisted of a mixture of six Salmonella serotype strains of bovine origin, namely, Salmonella Agona, Salmonella Anatum, Salmonella Newport FSL S5-436, Salmonella Reading, Salmonella Saintpaul, and Salmonella Typhimurium DT104 var. Copenhagen. Three of the Salmonella serotype strains were multidrug resistant (i.e., Salmonella Newport, Reading, and Typhimurium DT104 var. Copenhagen), and three were antibiotic susceptible (i.e., Salmonella Agona, Anatum, and Saintpaul) (3, 11). The Salmonella Newport strain was kindly provided by Dr. Martin Wiedmann (Department of Food Science, Cornell University, Ithaca, NY), while the five remaining strains were previously isolated by Bacon et al. (2, 3) and were available in our laboratory's culture collection. All six strains were hydrogen sulfide producers, as indicated by the formation of black-centered colonies on xylose lysine deoxycholate agar (XLD; Acumedia, Neogen Corporation, Lansing, MI).

Working cultures of the inoculum strains were maintained on XLD agar. Two days prior to the start of each trial, a single colony of each strain was separately inoculated into 10 mL of tryptic soy broth (TSB; Difco, BD, Sparks, MD) and incubated at 35°C for 22 to 24 h. Afterwards, a 0.1-mL aliquot of the original broth culture was transferred to 10 mL of fresh TSB and again incubated at 35°C (22 to 24 h). On the day of the experiment, broth cultures of all six Salmonella strains were combined, and cells were harvested by centrifugation (3,220 × g, 20 min, 4°C; Eppendorf, model 5810 R, Brinkmann Instruments Inc., Hamburg, Germany). Cell pellets were washed with 10 mL of phosphate-buffered saline (pH 7.4; Sigma-Aldrich, St. Louis, MO), centrifuged again, and the final cell pellets resuspended in 60 mL of phosphate-buffered saline. The concentration of the inoculum mixture was approximately 8 to 9 log CFU/ml.

On three separate days, 20 pieces of prerigor (warm) beef carcass surface brisket tissue were collected directly from the processing line of a commercial beef processing plant in northern Colorado. The beef tissue samples were obtained from carcasses that had been subjected to a hot water treatment but prior to treatment with LA. Samples were placed in insulated containers and transported to the laboratory of the Center for Meat Safety & Quality (Department of Animal Sciences, Colorado State University, Fort Collins) within 1 h of collection. Two sections (10 by 10 cm) were cut from each warm (25 ± 5°C) beef brisket tissue sample. One section per piece of brisket tissue was left untreated (control), while the second section was spray treated with one of four antimicrobial treatment groups (described in the next section). Only the external adipose side of each sample was inoculated, and this was done by randomly distributing 0.2 mL (approximately 10 drops) of the inoculum over the surface. The target inoculation level was 6 to 7 log CFU/cm². Inoculated samples were left undisturbed for 15 min to allow for cell attachment and were then either subjected to one of the antimicrobial treatments or, in the case of untreated control samples, were microbially analyzed for initial bacterial counts.

The LA (l-isomer; Purac America, Lenexa, KS) and SSS (Centron [formerly, AFTEC 3000], Zoetis Inc., Florham Park, NJ) solutions were prepared at approved concentrations or pH level (25) and were applied per manufacturer recommendations. The four antimicrobial treatments evaluated in the study were (i) unheated SSS (pH 1.1, 21°C), (ii) heated SSS (pH 1.1, 52°C), (iii) unheated LA (4%, 21°C), and (iv) heated LA (4%, 52°C). Treatments were applied by using a custom-built spray cabinet (Chad Co., Olathe, KS) designed to simulate a commercial beef carcass spray cabinet. Individual beef tissue samples were suspended on a hook and spray treated for 5 s (15 lb/in², 33 mL/s flow rate). Following treatment, excess liquid was allowed to drip off of samples for 10 min before being processed for microbial analysis.

Untreated (control) and treated beef tissue samples were transferred to Whirl-Pak filter bags (Nasco, Fort Atkinson, WI) to which 100 mL of Dey-Engley neutralizing broth (Difco, BD) was added. Samples were mechanically agitated (2 min; Masticator, IUL Instruments, Barcelona, Spain) and then serially diluted in 0.1% buffered peptone water (Difco, BD). Appropriate dilutions were surface plated in duplicate onto tryptic soy agar (TSA; Acumedia, Neogen Corporation), for determination of total bacterial counts, and onto XLD agar, for determination of inoculated Salmonella counts. Colonies were manually counted after incubation of plates at 25°C for 72 h (TSA) or 35°C for 24 h (XLD agar). Counts were recorded and converted to log CFU per square centimeter. On each experiment day, five uninoculated beef samples were also analyzed for any naturally present hydrogen sulfide–producing microflora.

Uninoculated beef tissue samples treated with the designated antimicrobial treatments were analyzed for pH. These samples were diluted with deionized water (1:5 sample-to-water ratio) and mechanically agitated (Masticator) for 2 min. The pH of samples was measured with a calibrated pH meter fitted with a glass electrode (Denver Instruments, Arvada, CO).

Three repetitions of the study were conducted on three separate days. The study was designed as a randomized complete block with a 2 × 2 factorial arrangement (n = 15 per treatment). Each repetition (test day) was considered as a blocking factor. Separate analyses were performed for each treatment (unheated SSS, heated SSS, unheated LA, and heated LA) to evaluate the antimicrobial efficacy against the total bacterial populations and inoculated Salmonella populations. This was done by comparing the counts of untreated and treated samples using a pairwise t test (mixed procedure of SAS, Version 9.3; SAS Institute Inc., Cary, NC). Bacterial populations were expressed as least-squares means. To compare the antimicrobial efficacy between treatments, counts for untreated samples served as a covariate, and counts for treated samples were adjusted and then compared to determine the main effect of chemical (SSS or LA), solution temperature (unheated or heated) and their interaction. Data were analyzed by using the mixed procedure of SAS, with chemical, solution temperature, and their interaction as independent variables. The pH data from treated samples were analyzed to determine treatment effects on the final pH of treated beef samples. All differences were reported by using a significance level of α = 0.05.

Naturally occurring hydrogen sulfide–producing microflora were not recovered (1 CFU/cm² detection limit) from the uninoculated samples analyzed (data not shown); therefore, colony counts obtained from the XLD agar were those of the Salmonella inoculum used in this study. Bacterial counts of the inoculated untreated (control) and treated prerigor beef carcass surface tissue samples are shown in Table 1. Counts from the nonselective (TSA) and selective (XLD) agars were similar for the control samples, which indicated that total bacterial counts recovered with TSA consisted largely of the inoculated Salmonella strains. Spray treatment of beef tissue samples with unheated (21°C) or heated (52°C) solutions of SSS or LA effectively (P < 0.05) reduced microbial populations. More specifically, initial total bacterial counts (6.3 log CFU/cm²) were reduced by 1.7 and 2.0 log CFU/cm² by the unheated and heated SSS treatments, respectively, and by 1.6 and 1.9 log CFU/cm² following treatment with the unheated and heated LA solutions, respectively (Table 1). In comparison, numerically greater reductions (by 0.3 to 0.8 log CFU/cm²) were observed based on counts recovered with the selective agar, and were likely owing to the inability of sublethally injured cells to recover and grow in the presence of the selective ingredients included in the XLD agar formulation. As seen in Table 1, spray treatment of beef samples with the unheated and heated solutions of SSS reduced (P < 0.05) initial Salmonella counts (6.2 log CFU/cm²) by 2.0 and 2.3 log CFU/cm², respectively, and for the unheated and heated LA treatments, initial pathogen counts (6.1 to 6.2 log CFU/cm²) were reduced (P < 0.05) by 2.4 log CFU/cm², irrespective of solution temperature.

Inoculation level of the test organism may affect the antimicrobial efficacy of treatments in vitro, as a higher inoculation level may result in proportionately higher reductions and a lower inoculation level may lead to relatively lower proportional reductions in bacterial counts. Because the inoculation level was slightly different for each treatment in the present study (Table 1), counts of initial inoculation levels on untreated control samples served as a covariate, and the least-squares means of counts for treated samples were adjusted. Therefore, only the adjusted counts were used to compare the main effects of chemical, solution temperature, and their interaction.

As indicated by our data, the interaction of chemical (SSS or LA) and solution temperature (unheated or heated) was not significant (P > 0.05); therefore, only the main effects of chemical and solution temperature are presented (Table 2). The chemical solution temperature had a significant (P < 0.05) effect on adjusted least-squares mean total bacterial counts but not (P > 0.05) on adjusted least-squares mean Salmonella counts (Table 2). Spray treatment of beef tissue samples with heated (52°C) chemical solutions resulted in a 0.3 log CFU/cm² greater (P < 0.05) reduction in total bacterial counts as compared with reductions obtained with unheated (21°C) solutions. Note, however, that a 0.3-log difference is generally not considered biologically significant (16). The adjusted least-squares mean total bacterial counts of samples treated with SSS did not differ (P > 0.05) from those of LA-treated beef samples, indicating that use of either compound would be expected to be equally effective against total bacterial populations. In contrast, chemical type had a significant (P < 0.05) effect on adjusted least-squares mean Salmonella counts. More specifically, Salmonella counts from LA-treated samples were 0.3 log CFU/cm² lower (P < 0.05) than those of SSS-treated samples.

Comparison of the pH data of treated beef samples showed that solution temperature did not (P > 0.05) have an effect on the final pH of treated samples, whereas chemical type did (P < 0.05; data not shown). Beef carcass tissue samples spray treated with LA had a pH (pH 4.18) that was 0.53 pH units lower (P < 0.05) than the pH of SSS-treated samples (pH 4.71).

Use of LA for reducing bacterial pathogen contamination on beef carcasses and derived cuts has been extensively investigated (1, 8–10, 12, 17, 21, 26). Factors that may affect the decontamination efficacy of LA in inoculated challenge studies include the type and temperature of the beef tissue used, inoculation method, inoculation level, attachment time of the tested pathogen, solution concentration and temperature, and treatment application method and parameters (26). Cutter and Rivera-Betancourt (8) reported 3.2-log CFU/cm² reductions of Salmonella on the surface of beef flanks that were sprayed (15 s, 125 ± 5 lb/in², and 4.8 L/min flow rate) with 2% LA (35 ± 2°C). In another study (21), immersion of inoculated (approximately 4 log CFU/cm²) adipose or lean muscle surfaces of beef cheek meat in 5% LA for 1 min reduced Salmonella levels by 2.0 and 1.4 log CFU/cm², respectively. When lean beef trimmings were immersed for 30 s in 25°C or 55°C LA (5%), inoculated Salmonella counts (3.0 to 3.3 log CFU/cm²) were reduced by 1.3 to 1.5 log CFU/cm² (25°C solution) and 1.6 to 1.9 log CFU/cm² (55°C solution) (10).

A limited number of studies have evaluated the antimicrobial efficacy of SSS. A 30-s immersion treatment of beef trimmings in SSS (pH 1.2) resulted in 0.5- to 0.7-log CFU/cm² reductions of Salmonella Typhimurium and Salmonella Newport (11). Salmonella levels on beef cheek meat samples were reduced by 1.0 to 1.5 log CFU/cm² following an immersion treatment in 1% SSS for 1, 2.5, or 5 min (21). In another study (22), chicken wings immersed for 10 or 20 s in SSS (pH 1.1) lowered inoculated Salmonella populations by 0.8 to 0.9 and 1.1 to 1.2 log CFU/mL of rinse solution, respectively.

In conclusion, the results of the study indicate that spray treatment of prerigor beef carcass surface tissue with SSS at pH 1.1 or 4% LA effectively reduces Salmonella contamination. In addition, under the experimental conditions of our study, heating of the chemical solutions to 52°C was found not to have a practical effect on their antimicrobial effectiveness against Salmonella.