Counts of Campylobacter spp. on U.S. Broiler Carcasses

Transcription

1 1034 Journal of Food Protection, Vol. 69, No. 5, 2006, Pages Copyright, International Association for Food Protection Counts of Campylobacter spp. on U.S. Broiler Carcasses NORMAN J. STERN 1 * AND STEPHEN PRETANIK 2 1 U.S. Department of Agriculture, Agricultural Research Service, South Atlanta Area, Poultry Microbiological Safety Research Unit, Russell Research Center, Athens, Georgia 30604; and 2 National Chicken Council, Washington, D.C , USA MS : Received 31 August 2005/Accepted 18 November 2005 ABSTRACT Foodborne Campylobacter-associated gastroenteritis remains a public health concern, and the Centers for Disease Control and Prevention suggests that improperly handled poultry is the most important source of this human disease. In response to these concerns, 10 of the largest U.S. poultry integrators cooperatively determined the incidence and counts of Campylobacter on processed broiler carcasses. Prior to conducting the survey, laboratory personnel were trained in a direct Campy-Cefex plating procedure for enumeration of the organism. Before and after the survey enumeration, consistency in reporting was compared among the participating laboratories. Participating laboratories were able to consistently estimate inoculated concentrations of Campylobacter in carcass rinses. Within the central study, we determined the potential exposure of U.S. consumers to Campylobacter spp. associated with broiler carcasses during a 13-month period. Among each of the 13 participating poultry complexes, rinses from 25 randomly selected fully processed carcasses were sampled monthly from individual flocks. Among 4,200 samples, approximately 74% of the carcasses yielded no countable Campylobacter cells. Campylobacter spp. were isolated from approximately 3.6% of all commercially processed broiler carcasses at more than 10 5 CFU per carcass. Acceptable counts of these organisms on raw poultry carcasses remain to be determined. Nevertheless, this survey indicates industry recognition of its responsibility to assess and reduce public exposure to Campylobacter through broiler chickens. Campylobacter spp. is reported as one of the most frequent causes of foodborne bacterial illness in the United States. The Centers for Disease Control and Prevention (CDC) estimates that approximately 1.6 million cases occur each year, and approximately 15 people per 100,000 are infected by the agent each year (3, 11). Several food animals are involved in transmitting this pathogen to humans (1). Harris and coworkers (6) suggested that poultryborne transmission accounts for close to 50% of the human cases of Campylobacter-associated illness. Consequently, the poultry industry wishes to control and reduce the incidence and counts of this organism associated with raw poultry product at the end of the processing line. Campylobacter is a microaerophillic gram-negative organism that thrives within the intestinal tract of warmblooded hosts (13, 17). Within the intestinal tract of broiler chickens, concentrations frequently can reach 10 6 CFU g 1 of feces and often exceed 10 8 CFU g 1 of feces (13). Thus, during the transport and processing of broilers, contamination with substantial concentrations of Campylobacter can and does occur. The organism cannot proliferate at room or refrigeration temperatures and, in general, is considered very fragile and susceptible to most antimicrobial treatments (17, 18). Therefore, the pathogen numbers residing on the carcasses leaving the processing plants represent the highest counts associated with the product. The Campylobacter concentrations on the carcass are related indirectly to the potential for human exposure. Ingestion of relatively low numbers ( 800 cells) of Campylobacter is re- * Author for correspondence. Tel: ; Fax: ; nstern@saa.ars.usda.gov. quired to cause human disease (2). Reduced counts of the organism on carcasses would reduce the likelihood of disease. In response to this important health concern, the U.S. food industry established an initiative to determine Campylobacter spp. concentrations presently occurring on poultry carcasses across the United States. When coupled with an understanding of what constitutes a risk-acceptable count of this organism on poultry carcasses, the magnitude and approach for further industry response can then be assessed. Intervention strategies must still be formulated to reduce and control the transmission of this organism associated with poultry carcasses. MATERIALS AND METHODS Survey protocol. Thirteen broiler chicken processing plants in the area of the United States that has the highest density of poultry producing centers volunteered for and were enrolled in this study. These plants were owned by 10 of the nation s largest integrated poultry operations. Once each month for 13 consecutive months (September 2003 through September 2004), one flock per processing plant was selected for sampling with methods previously described (19, 20). Twenty-five processed broiler carcasses (approximately 1 of every 1,000 birds) that had undergone standard evisceration and rinsing along the processing lines were sequentially aseptically removed from the water-ice cooled chiller tanks. Carcasses were then individually rinsed with 400 ml of Butterfield s buffer for 1 min, and approximately 100 ml of the rinsate was transferred to an aseptic container, which was packed on ice in an insulated cooler and sent by overnight delivery service to the microbiology laboratory that conducted the analysis. Microbiology. Microbiological methods entailed directly plating serial dilutions from the rinse materials onto surfaces of

2 J. Food Prot., Vol. 69, No. 5 CAMPYLOBACTER COUNTS ON BROILERS 1035 Campy-Cefex medium and incubating the plates for 36 to 48 h at 42 C under reduced atmospheres (21). A 0.1-ml aliquot of each dilution was spread plated directly onto duplicate agar plates. Therefore, for 400 ml of carcass rinsate, an average of 4,000 CFU per carcass was needed for detection. Colonies with the translucent appearance typical of Campylobacter were observed with a phase-contrast microscope ( 1,000) for corkscrew morphology and rapid characteristic corkscrew-like movement. These suspect colonies were then subjected to a highly specific latex agglutination assay to confirm the presence of the organism (Campylobacter jejuni, Campylobacter coli, or Campylobacter lari) (Integrated Diagnostics, Inc., Baltimore, Md.). The monthly laboratory results were sent to the National Chicken Council (Washington, D.C.), encoded, and forwarded to the Poultry Microbiological Safety Research Unit (Athens, Ga.) for analysis. Campylobacter check samples. Personnel from 11 cooperating laboratories were trained in the above procedures to enumerate Campylobacter spp. from broiler rinses. Prior to the above industry survey study, check samples were sent to the cooperating laboratories to determine the reproducibility of the method among and between laboratories after 1 month of practice within their own facilities. Four strains of C. jejuni (AL-22, BL-1, BH-6, and CL-11) previously isolated from broilers (14) were struck onto Brucella agar and grown for 24 h at 42 C under a microaerobic atmosphere. Swabs of each strain were resuspended in Butterfield s buffer and serially diluted. These serial dilutions were added to chicken rinsates to provide logarithmic serial dilutions distributed to the testing laboratories. Encoded inoculated ice-chilled rinsate samples were sent by courier service overnight to the participating laboratories ( 100 ml per sample container). The coded samples were evaluated by each participating laboratory to estimate the organism counts in each inoculated sample and determine the accuracy and reproducibility of the method. The data were collated at the Russell Research Center; the code indicating the original inoculation concentration was deciphered, and the variation in reporting distribution was determined. The entire check procedure was repeated after the survey study had been completed to assess whether greater consistency among the laboratories had been gained during the study. For comparison with the newly trained laboratories, in only the check sample study, the Agricultural Research Service laboratory (no. 12), which had the most experience with enumeration of Campylobacter, was included. Only the industrial laboratories were involved in the 13-month commercial carcass enumeration study. Statistical analysis. Counts of CFUs at each dilution were averaged, and estimations of Campylobacter concentrations per carcass were calculated and recorded. For example, at 40 and 46 CFU per 0.1 ml of the initial dilution, the calculation would be the average of 43 CFU 0.1 ml ml carcass CFU carcass 1. For statistical purposes, samples providing no detectable Campylobacter were assigned the negligible value of All bacterial counts calculated from the rinsate were expressed as CFU of Campylobacter per broiler carcass and transformed to mean standard deviation (SD) log values for subsequent data analysis. The reporting individual laboratory and the reported mean counts at each inoculation dilution in the check sample studies were tested to determine statistical variation (P 0.05) using the SAS (version 8) analysis of variance, with significance defined at the 95% confidence level. We compared results from each laboratory with those of every other laboratory by using the Scheffé S method and compared the results from the control laboratory (no. 12) with those of the other laboratories by using the Dunnett procedure. RESULTS The counts of recovered Campylobacter spp. from inoculated carcass rinsates are reported in Table 1. For the inoculum of 10 3 CFU per carcass, concentrations reported by laboratory (lab) 9 were significantly different (P 0.05) from those of the control lab 12, and for the inoculum of 10 4 CFU per carcass concentrations reported by labs 1 and 11 were significantly different (P 0.05) from those of the control lab 12. For the inoculum of 10 5 CFU per carcass, none of the laboratories reported significantly different concentrations (P 0.05). For the inoculum of 10 6 CFU per carcass, concentrations reported by labs 2, 3, and 11 were significantly different (P 0.05) from those of the control lab 12. The mean sample concentrations reported for each inoculation concentration were significantly different from one another (P 0.05). Following the 13-month survey trial, inoculated check rinsate samples were once again sent to the participating laboratories for enumeration. The inoculation concentrations of Campylobacter provided were lower than desired (Table 2). Although samples with lower concentrations of inoculum were provided, the laboratories were able to distinguish between the serial dilutions (P 0.05). For uninoculated samples and those inoculated at 10 2 CFU per carcass, results from none of the laboratories were significantly different (P 0.05) from one another. At 10 3 CFU per carcass, concentrations reported by lab 10 were significantly different (P 0.05) from those of the control lab 12. At 10 4 CFU per carcass, concentrations reported by lab 3 were significantly different (P 0.05) from those of the control lab 12. At 10 5 CFU per carcass, concentrations reported by lab 3 were significantly different (P 0.05) from those of control lab 12. Lab 2 did not receive their shipment of samples in a timely manner, so their data were not considered for these check samples. The prevalence and concentrations of Campylobacter spp. in carcass rinsates from 13 U.S. commercial poultry processing complexes by month, September 2003 through September 2004, are reported in Table 3. Among the 4,200 carcasses studied, 1,094 (26%) were positive for this pathogen. Variation in the frequency of positive carcasses ranged from a low of 15% in March to a high of 43% in February. The concentrations of Campylobacter spp. varied by sample month, with a mean of less than 10 CFU per carcass in March and about 390 CFU per carcass in October. Although Campylobacter was not detectable in 74% of the carcasses, concentrations as high as CFU per carcass were observed. Table 4 provides the prevalence and concentrations of Campylobacter spp. in poultry carcass rinsates from each of the 13 U.S. commercial poultry processing complexes, September 2003 through September There was substantial difference in reported frequency of carcass contamination among the complexes. No contaminated carcasses were detected in complex I, but 74% of the carcasses sampled from complex H were contaminated with Campylobacter spp. Likewise, mean concentrations ranged from nondetectable (complex I) to log CFU per carcass

3 1036 STERN AND PRETANIK J. Food Prot., Vol. 69, No. 5 TABLE 1. Concentrations of Campylobacter spp. in samples of inoculated chicken carcass rinsate as recovered by 12 participating laboratories, August 2003 a Inoculum Laboratory samples Lab 1 Lab 2 Lab 3 Lab 4 Lab 5 Lab 6 Lab 7 Lab 8 Lab 9 Lab 10 Lab 11 Lab 12 Mean SD b None A B C D LA c E LA a Each laboratory was given four coded samples for each of the five inoculation concentrations. b Means with different letters are significantly different (P 0.05). c LA, laboratory accident (data not recorded). (complex E). Variation of Campylobacter concentrations per carcass within a flock was quite high (data not shown). The stratified frequency of estimated Campylobacter spp. concentrations among these poultry carcasses sampled from September 2003 through September 2004 is shown in Table 5. The percentage of highly contaminated carcasses asymptotically approached zero but only after the concentrations reached 10 8 CFU per carcass. In total, 3.6% of the carcasses sampled yielded concentrations of 10 5 CFU per carcass. DISCUSSION Variation in the counts of Campylobacter spp. reported was observed among the laboratories for the inoculated check samples (Tables 1 and 2). Because the samples were derived from poultry rinsates, it was reasonable to anticipate that natural contamination might account for bacterial recovery from the noninoculated samples. However, in general the counts reported were low among these noninoculated samples. The notable exception was the result from lab 4, which reported a surprisingly high count of CFU per carcass for one of its samples (Table 1). Both lab 3 and lab 4 reported high counts in one sample each from the noninoculated group in September 2004 (Table 2). Overall, for both sets of check samples, the participating laboratories were able to distinguish between the various dilutions of the inoculum (P 0.05). The limitations of direct plating as compared with enrichment technologies were recognized but accepted in the present study. Line et al. (10) found that estimations of Campylobacter spp. in broiler rinsates by direct plating onto Campy-Cefex agar was not significantly different from an most-probable-number (MPN) procedure employing Hunt s (U.S. Food and Drug Administration) Campylobacter selective enrichment broth followed by recovery on modified Campylobacter charcoal differential agar. More comprehensive methodologies for detecting and enumerating Campylobacter are available (16, 22), such as the MPN-enrichment methodologies (10). However, this more rigorous and expensive approach did not provide statistical improvement for reporting organism counts (10). Additionally, time is lost and further expenses are incurred when such an approach is applied to Campylobacter. For example, an additional 24 to 48 h of enrichment incubation is required before the culture can be streaked for isolation. The potential for clerical error also increases greatly with MPN protocols. Enrichment procedures may provide higher rates of recovery, but these higher rates have not been observed consistently (10). Better molecular and immunological Campylobacter detection methods may be available but have not been widely accepted among poultry microbiological laboratories. Therefore, we chose to employ a comparatively simple method for estimating Campylobacter numbers in this survey. In our survey, approximately 74% of the processed carcasses sampled did not yield Campylobacter by the direct plating method (Tables 3 and 4). Using various methods, other researchers have reported Campylobacter prevalences

4 J. Food Prot., Vol. 69, No. 5 CAMPYLOBACTER COUNTS ON BROILERS 1037 TABLE 2. Concentrations of Campylobacter spp. in samples of inoculated chicken carcass rinsate as recovered by 11 of the 12 participating laboratories, September 2004 a Inoculum Laboratory samples Lab 1 Lab 3 Lab 4 Lab 5 Lab 6 Lab 7 Lab 8 Lab 9 Lab 10 Lab 11 Lab 12 Mean SD b None A A B C D a Each laboratory was given five coded samples for each of the five inoculation concentrations. Lab 2 received shipped samples 1 day late; therefore, their data were not considered. b Means with different letters are significantly different (P 0.05). TABLE 3. Prevalence and concentrations of Campylobacter spp. in carcass rinsates from 13 U.S. commercial poultry processing complexes by month, September 2003 through September 2004 Month No. of samples tested No. (%) of positive samples Mean SD Median Maximum 2003 September (36) October (33) November (31) December (26) January (31) February (43) March (15) April (24) May (16) June (19) July (18) August (30) September (18)

5 1038 STERN AND PRETANIK J. Food Prot., Vol. 69, No. 5 Table 4. Prevalence and concentrations of Campylobacter spp. in carcass rinsates from each of 13 U.S. commercial poultry processing complexes (A through M), September 2003 through September 2004 Complex No. of samples tested No. (%) of positive samples Mean SD Median Maximum A (60) B (14) C (20) D (14) E (34) F (9) G (31) H (74) I J (25) K (8) L (24) M (25) of 48, 94, and 98% among broiler carcasses (6, 8, 12). In Northern Ireland, Wilson reported that the range of carcass contamination was 47 to 81%, depending upon the producer (23). Corry and Atabay (5) indicated that approximately 80% of poultry sold in the United Kingdom was contaminated by Campylobacter, sometimes at concentrations of thousands of cells per square centimeter. Among 241 whole raw chickens from retail outlets, 83% yielded Campylobacter (8). Hood et al. (7) indicated that concentrations higher than 10 6 CFU per carcass may be encountered on market poultry. Smeltzer (12) also found carcasses contaminated with Campylobacter at up to 10 5 CFU per carcass. The low frequency observed in the current study is noteworthy and may be due to the use of 50 ppm chlorine in the chiller water. The same methods were used for assessing broiler carcasses contamination in north Georgia in 2001, and 84.7% of the 450 carcasses assayed from nine flocks were positive for Campylobacter (19). Those chickens tested were 56 days old when processed. Consequently, the incidence and concentration of the organism would be predictably greater than that among typical 42-day-old birds. In that study, the mean log CFU per flock ranged from to The overall mean was log CFU per sample of carcass rinsate. What accounted for the variation among results reported by different processing complexes within the study? TABLE 5. Frequency of estimated Campylobacter spp. concentration among 4,200 U.S. poultry carcasses sampled from September 2003 through September 2004 Concn of Campylobacter spp. (CFU/carcass) % of carcasses tested Certain complexes never found Campylobacter-contaminated carcasses. Other complexes had considerably higher contamination rates. Further retraining of laboratory personnel regarding the sampling protocol might address errors in enumeration, but real differences in processing plant operations may have accounted for the variation observed (9). Within the means reported for each flock, individual carcass contamination differed by as much as 2 log units. Such observations suggest that relatively small contributions to these numbers might be accounted for by cross-contamination. Otherwise, homogeneity of Campylobacter numbers would have been expected. We still do not know how to explain such variation in numbers on carcasses within a flock. Over the past 10 years, the U.S. poultry industry has reduced Campylobacter spp. contamination of processed broilers by more than 10-fold (19). CDC reports indicate that during the same period of time human disease caused by this pathogen has been substantially reduced (3). Adherence to hazard analysis critical control point (HACCP) procedures within the poultry processing plants has played a large role in this reduction. Because Campylobacter can contaminate carcasses within the feather follicles (4), the limits of HACCP intervention may have already been reached. On-farm counts of the organism have remained constant from 1995 through at least 2001 (19). In the current study, 3.6% of the carcasses yielded Campylobacter at 10 5 CFU. Therefore, the next incremental reduction in the presence of the organism may need to occur during the animal production phase. Our study provided part of the data for a science-based regulatory position. When compared with other European countries using similar quantitative methods, the United States appears to have lower Campylobacter counts per carcass (7, 15). We still must learn what level of reduction will be required to ensure public safety from this pathogen. By overlaying the data provided in this study with the counts determined to be of public health concern, the industry will learn what improvements must be made. Likewise, the regulators will have a basis for setting a position incumbent

6 J. Food Prot., Vol. 69, No. 5 CAMPYLOBACTER COUNTS ON BROILERS 1039 upon the industry. Further reductions in the sporadically high Campylobacter counts found on individual carcasses may be of substantial importance. ACKNOWLEDGMENTS The authors thank the following U.S. poultry companies for their participation in this study: Fieldale Farms, Foster, George s, Goldkist, Keystone, Perdue Farms, Pilgrim s Pride, Sanderson Farms, Tyson, and Wayne Farms. The authors also extend appreciation to Ms. Susan Brooks and Ms. Latoya Wiggins for their invaluable technical support in this study. REFERENCES 1. Alterkruse, S. F., J. M. Hunt, L. K. Tollefson, and J. M. Madden Food and animal sources of human Campylobacter jejuni infection. J. Am. Vet. Med. Assoc. 204: Black, R. E., M. M. Levine, M. L. Clements, T. P. Hughes, and M. J. Blaser Experimental Campylobacter jejuni infection in humans. J. Infect. Dis. 157: BookRags.com Campylobacteriosis. Available at: bookrags.com/sciences/biology/campylobacteriosis-wmi.html. 4. Chantarapanont, W., M. E. Berrang, and J. F. Frank Direct microscopic observation of viability of Campylobacter jejuni on chicken skin treated with selected chemical sanitizing agents. J. Food Prot. 67: Corry, J. E. L., and H. I. Atabay Poultry as a source of Campylobacter and related organisms. J. Appl. Microbiol. 90:96S 114S. 6. Harris, N. V., N. S. Weiss, and C. M. Nolan The role of poultry and meats in the etiology of Campylobacter jejuni/coli enteritis. Am. J. Public Health 76: Hood, A. M., A. D. Pearson, and M. Shahamat The extent of surface contamination of retailed chickens with Campylobacter jejuni serogroups. Epidemiol. Infect. 100: Jorgensen, F., R. Bailey, S. Williams, P. Henderson, D. R. Wareing, F. J. Bolton, J. A. Frost, L. Ward, and T. J. Humphrey Prevalence and numbers of Salmonella and Campylobacter spp. on raw, whole chickens in relation to sampling methods. Int. J. Food Microbiol. 76: Kazwala, R. R., J. D. Collins, J. Hannan, A. P. Crinion, and H. O Mahony Factors responsible for the introduction and spread of Campylobacter jejuni infection in commercial poultry production. Vet. Rec. 126: Line, J. E., N. J. Stern, C. P. Lattuada, and S. T. Benson Comparison of methods for the recovery and enumeration of Campylobacter from freshly processed broilers. J. Food Prot. 64: Mead, P. S., L. Slutsker, V. Dietz, L. F. McCaig, J. S. Bresee, C. Shapiro, P. M. Griffin, and R. V. Tauxe Food-related illness and death in the United States. Emerg. Infect. Dis. 5: Smeltzer, T. I Isolation of Campylobacter jejuni from poultry carcasses. Aust. Vet. J. 57: Stern, N. J., J. S. Bailey, L. C. Blankenship, N. A. Cox, and F. McHan Colonization characteristics of Campylobacter jejuni in chick ceca. Avian Dis. 32: Stern, N. J., P. Fedorka-Cray, J. S. Bailey, N. A. Cox, S. E. Craven, K. L. Hiett, M. T. Musgrove, S. Ladely, D. Cosby, and G. C. Mead Distribution of Campylobacter spp. in selected U.S. poultry production and processing operations. J. Food Prot. 64: Stern, N. J., K. L. Hiett, G. A. Alfredsson, K. G. Kristinsson, J. Reiersen, H. Hardardottir, H. Briem, E. Gunnersson, F. Georgsson, R. Lowman, E. Berndtson, A. M. Lammerding, G. M. Paoli, and M. T. Musgrove Campylobacter spp. in Icelandic poultry operations and human disease. Epidemiol. Infect. 130: Stern, N. J., and J. E. Line Comparison of three methods for recovery of Campylobacter spp. from broiler carcasses. J. Food Prot. 55: Stern, N. J., and J. E. Line Campylobacter, p In B. W. Lund, T. C. Baird-Parker, and G. W. Gould (ed.), The microbiological safety and quality of food, vol. II. Aspen Publishers, Gaithersburg, Md. 18. Stern, N. J., J. E. Line, and H.-C. Chen Campylobacter, p In F. P. Downes and K. Ito (ed.), Compendium of methods for the microbiological examination of foods, 4th ed. American Public Health Association, Washington, D.C. 19. Stern, N. J., and M. C. Robach Enumeration of Campylobacter spp. in broiler feces and in corresponding processed carcasses. J. Food Prot. 66: Stern, N. J., P. J. Rothenberg, and J. M. Stone Enumeration and reduction of Campylobacter jejuni in poultry and red meats. J. Food Prot. 48: Stern, N. J., B. Wojton, and K. Kwiatek A differential-selective medium and dry ice generated atmosphere for recovery of Campylobacter jejuni. J. Food Prot. 55: Wassenaar, T. M., and D. B. Newell Genotyping of Campylobacter spp. Appl. Environ. Microbiol. 66: Wilson, I. G Salmonella and Campylobacter contamination of raw retail chickens from different producers: a six year survey. Epidemiol. Infect. 129: