Survival of Campylobacter jejuni and Campylobacter coli on Retail Broiler Meat Stored at 220, 4, or 12uC and Development of Weibull Models for Survival

Survival of Campylobacter jejuni and Campylobacter coli isolated from broiler meat was investigated and modeled on retail breast meat. Meat portions were inoculated with C. jejuni or C. coli at 6.4 to 6.8 log CFU/g followed by storage at 220uC for 84 days or at 4 or 12uC for 14 days. Kinetic data within a species and temperature were fitted to the Weibull model. When $70% of the residuals were in an acceptable prediction zone from 21 (fail-safe) to 0.5 (fail-dangerous) log units, the model was considered to have acceptable performance. Survival of Campylobacter was highest at 4uC, lowest at 12uC, and intermediate at 220uC. Survival of C. jejuni and C. coli was similar at 220uC but was lower (P , 0.05) for C. jejuni than for C. coli at 4 and 12uC. The Weibull model provided acceptable predictions for four of six sets of dependent data with unacceptable performance for survival of C. jejuni at 220 and 12uC. A difference in survival was observed between the two strains of C. jejuni tested. Comparison of Weibull model predictions with data for C. jejuni archived in ComBase revealed mostly unacceptable performance, indicating that C. jejuni and C. coli survival on raw broiler breast meat differs from published results for other strains and growth media. Variation in Campylobacter survival among replicate storage trials was high, indicating that performance of the models can be improved by collection of additional data to better define the survival response during storage at temperatures from 220 to 12uC. Although commercially processed broiler chickens go through a wide variety of steps during processing to reduce microbial contaminants (30), several studies have revealed that retail broiler meat is frequently contaminated with Campylobacter spp. (10, 31). This contamination occurs during processing when carcasses come in direct contact with fecal matter and then commingle in the chiller tank (22, 28, 42). In an attempt to reduce contamination and improve the shelf life of broiler carcasses, rapid chilling methods have been developed by the poultry industry. In the United States, immersion chilling is the typical method used to reduce the carcass temperature. In other countries, air chilling and evaporative air chilling are more common (39). All three cooling methods are effective for rapidly reducing the temperature of the carcasses, which have similar prevalences of microbial contamination (4, 21, 26). However, El-Shibiny et al. (14) found that these methods may enhance the survival of foodborne bacterial pathogens, including Campylobacter, during the shelf life of broiler meat. Beyond the rapid chilling methods, retail broiler meat is subjected to variable freezing and refrigeration temperatures during storage, transportation, and display in retail outlets and in consumers’ refrigerators. Because of the relatively high prevalence of Campylobacter spp. found in retail broilers and the low infective dose required to cause human disease (38), an understanding of the ability of Campylobacter spp. to survive refrigeration and freezing is directly relevant to designing new strategies to improve food safety and public health. The survival of Campylobacter spp. on broiler meat also may be an important factor for at-home contamination due to improper food handling. Several researchers have reported the effects of refrigeration and freezing on the survival of Campylobacter jejuni (5, 24, 37), but all of these studies have included chicken skin as the product evaluated to determine survival. In one study, broth medium was used instead of poultry meat (7). Very few studies carried out in the last 20 years have addressed the survival of C. jejuni in broiler meat. Similarly, the survival of Campylobacter coli has been studied only in inoculated chicken skin (14). Although C. coli may represent 20% or more of all Campylobacter * Author for correspondence. Tel: 334-229-8449; Fax: 334-229-6709; E-mail: [email protected]. { These authors contributed equally. 1438 Journal of Food Protection, Vol. 73, No. 8, 2010, Pages 1438–1446 species isolated from retail broiler meat (25), there have been no studies in the last 15 years that have addressed the survival of this pathogen in retail broiler meat. The objective of the present study was to investigate and model the survival rate of C. jejuni and C. coli isolates that were obtained from retail broilers, inoculated onto boneless, skinless broiler breast meat, and stored at 4, 12, and 220uC for various time periods. Kinetic data within a species and temperature were fitted to the Weibull model, and performance of the models was assessed using the acceptable prediction zone (APZ) method, which classifies a model as providing acceptable predictions of the test data when $70% of the residuals fall in an APZ (29). MATERIALS AND METHODS Bacterial strains, culture conditions, and typing method. C. jejuni 971 and 1065 and C. coli 947 and 956 were isolated from retail broiler meat and identified using described multiplex PCR assays (31). These isolates were recovered from stock cultures (280uC in Brucella broth supplemented with 30% glycerol and 5% lysed horse blood) by filtration through a 0.65-mm-pore-size Millipore filter (Fisher Scientific, Billerica, MA) and onto modified Campy-Cefex (mCC) agar supplemented with 5% lysed horse blood (33). Cultures were incubated at 42uC for 48 h under microaerobic conditions (10% CO2, 5% O2, and 85% N2; Airgas, Radnor, PA), which were provided by an evacuation replacement system (MACSmics Jar Gassing System, Microbiology International, Frederick, MD) in anaerobic jars. All strains were typed using a pulsed-field gel electrophoresis (PFGE) protocol described elsewhere (32). During the trials, isolates were collected at the initial, middle, and final sampling points and were typed using the same PFGE protocol. Retail broiler meat and inoculum preparation. Boneless, skinless broiler breast meat was purchased from a local retail store. The meat was aseptically cut into 30-g (¡1 g) pieces and grouped into runs consisting of 16 pieces. Meat samples (16 pieces) were spread onto sanitized trays and allowed to dry in a biological safety II laminar flow cabinet for 20 min. Inocula were prepared using colonies grown on mCC agar plates for 24 h at 42uC under microaerobic conditions and then dissolved into 4.5 ml of phosphate-buffered saline (PBS). Suspension concentrations were standardized to optical densities of 1.5 (¡0.2) at 600 nm and transferred into sanitized spray bottles. The inoculum was supplemented with 15.5 ml of sterile PBS to obtain a final volume of 20 ml, with a final level of approximately 7 log CFU/ml. The inoculum level was checked for each strain and for each replicate for each temperature. Meat samples were evenly inoculated on all sides until the inoculum was exhausted, allowed to dry in a biological hood for 60 min, and then transferred to individual Ziploc freezer bags (Glad Products Company, Oakland, CA), which were stored at 4uC. Survival experiments. Samples stored at 4 and 12uC were placed in an MIR 252 incubator (Sanyo North America Corporation, San Diego, CA), and two samples were removed from each trial for enumeration at day 0 and every 2 days for up to 14 days. Samples stored at 220uC were placed in a freezer (Thermo-Kool, Laurel, MS). Two samples were then removed from each run for enumeration at day 0 and every 14 days for 84 days. Samples removed from 220uC storage were allowed to thaw at room temperature (,25uC) for 1 h. All samples were then aseptically transferred to individual sterile plastic bags (Whirl-Pak, Nasco, Fort Atkinson, WI) and stomached for 1 min in a 1:2 (wt/ vol) meat:broth ratio of Bolton broth supplemented with 5% lysed horse blood. Bacterial counts. Surviving Campylobacter cells were enumerated by direct plating. Samples were serially diluted in sterile PBS (1:9) and spread plated on mCC agar in duplicate. The average of two duplicate plates and the average of two samples (two pieces of meat) were used to calculate the surviving number of cells per replicate. Enrichment samples and plates were incubated at 42uC under microaerobic conditions for 48 h. For enumeration by direct plating, CFUs for the last countable spread plate were recorded. If the enriched sample was positive and no Campylobacter colonies were found during enumeration, a value of 10 CFU/g of meat was assigned for that sample. For survival at 220uC, three replicate experiments were run with C. jejuni 1065, and three replicates were run with C. coli 947. For survival at 4uC, three and one replicate experiments were run with C. jejuni 971 and 1065, respectively, and three and one replicate experiments were run with C. coli 947 and 956, respectively. For survival at 12uC, three and one replicate experiments were run with C. jejuni 1065 and 971, respectively, and three and one replicate experiments were run with C. coli 947 and 956, respectively. The time 0 count was determined right before placing the meat at the test temperature and right after taking the meat from a 24-h storage at 4uC. Typing of bacterial strains. PFGE patterns of the inoculated strain and the strains surviving at the end of the experiment were compared. PFGE was performed as described elsewhere (32). Salmonella Choleraesuis Braenderup H9812 (ATCC BAA-664) cut with the restriction enzyme XbaI was used as the DNA size marker (1). Pair comparisons and cluster analyses were performed using the Dice correlation coefficient and the unweighted pair group with mathematical average clustering algorithm. The position tolerance for band analysis was set at 3%, and a cutoff of 90% DNA relatedness was used to determine whether isolates were similar. Survival modeling. The GInaFit program was used to identify an appropriate survival model for the data (16). This program allowed comparison of 10 microbial survival models for their goodness of fit to the data. After this evaluation (results not shown), replicate data (n ~ 24; eight samples for each of three storage trials) for survival of C. jejuni or C. coli on raw chicken breast meat stored at 220uC for 0 to 84 days, 4uC for 0 to 14 days, or 12uC for 0 to 14 days were fitted to the Weibull model using version 5.0 of the Prism software program (GraphPad Software, Inc., San Diego, CA):