PRE-HARVEST INTERVENTIONS FOR APPLICATION DURING POULTRY PRODUCTION TO REDUCE FOOD-BORNE BACTERIAL PATHOGENS
Location: Poultry Microbiological Safety Research
Title: INFLUENCE OF CULTURE METHODS ON RECOVERY OF SALMONELLA SEROTYPES
Submitted to: Meeting Abstract
Publication Type: Proceedings
Publication Acceptance Date: October 5, 2011
Publication Date: October 12, 2011
Citation: Cox Jr, N.A., Cason Jr, J.A., Cray, P.J. 2011. Influence of culture methods on recovery of Salmonella serotypes. Proceedings of the 46th National Meeting on Poultry Health and Processing. October 11-13, 2011, Ocean City, Maryland. p.119-122.
Current laboratory methods were developed to obtain the greatest number of Salmonella-positive samples at an acceptable cost and different types of samples can contain different microbial competitors, nutrients, and antimicrobial inhibitors. Each of these result in turn affect a different response regarding Salmonella growth and recovery. Further, the degree of stress or sub lethal injury that the bacteria have undergone are often underappreciated or more often ignored when selecting cultural methodology for Salmonella. Influence on serotype further compounds the problem and is discussed below.
Quite interestingly, even for a relatively well-understood sample type such as a chilled poultry carcass, there is no internationally recognized standard methodology for Salmonella detection, either for sample size or culture. Use of different cultural media is not simply a matter of laboratory choice as methods can be decided by regional, national, or international organizations. It is well known that many microbiologists/laboratories use a different culture method which further complicates comparison of recovery rates.
Classical cultural techniques for Salmonella detection generally include a nonselective preenrichment (which may be used depending on whether cells are stressed or are present in low numbers), a selective enrichment, isolation on selective agar media, biochemical screening with triple sugar and lysine iron agars, and serological confirmation with poly-O and poly-H antisera. Many different media formulations have been developed for the preenrichment, enrichment, and selective isolation steps in the detection sequence. A survey of diagnostic veterinary laboratories in the United States revealed that 17 different selective enrichment media were being used (Waltman & Mallinson 1995).There were also differences in incubation temperatures, whether samples were incubated for 24 h or 48 h or both, or whether samples were subjected to delayed secondary enrichment conditions. Selective enrichment cultures were inoculated onto 14 different plating media, with most labs using two or more different types of media to increase the likelihood of recovering Salmonella. Finally, the survey revealed considerable variation in numbers of colonies selected for further testing by the different laboratories, with many selecting and identifying only one colony.
Types of liquid and solid media have also changed over time. Lactose broth, Gram negative broth, Rappaport (Vassiliadis) broth, selenite cystine broth, and tetrathionate broth were used years ago. While all of these broths are still used to some degree significant composition modifications have occurred and selenite cystine broth is use the least. Plating media for Salmonella isolation included brilliant green, bismuth sulfite, Hektoen enteric, Salmonella-Shigella, and XLD agars. Some of these are rarely used today while others have been modified. Brilliant green with sulfapyridine added is now called BGS. XLD has had many modifications, including addition of sodium thiosulfate and Tergitol 4 to make XLT4 agar, one of the more popular agars used today. Food microbiology laboratories usually use two or more plating media to reduce the occurrence of false-negative results (Cox & Berrang 2000). Rappaport’s enrichment medium was modified to become the widely used Rappaport-Vassiliadis (RV) broth (Vassiliadis 1983) and was then further modified with the development of modified semisolid RV (MSRV) for selection of motile Salmonella (De Smedt et al. 1986). The main reason for development and adoption of different media has always been to isolate the maximum number of colonies of Salmonella serotypes other than Typhi. The widespread use of MSRV in Europe during the past 20 years may have played a major role in the reported differences in isolation rates and prevalence of different serotypes in various parts of the world. Additional sampling and methodology issues are discussed in a recent publication by the National Advisory Committee on Microbiological Criteria for Foods (NACMCF 2010).
Although the scientifically conservative position is to regard all Salmonella serotypes as human pathogens, there are known differences between serotypes in ability to survive stress, colonize animals, invade tissues, survive temperature fluctuations, and cause disease. Different serotypes also differ in ability to survive food processing conditions, with Enteritidis surviving better than Typhimurium and Infantis, and much better than Dublin (Hald et al. 2004).
In general, Salmonella methodology has been selected based on maximum recovery of positive colonies not on detecting specific Salmonella serotypes a sample. Many studies have demonstrated, however, that cultural techniques can influence the serotypes that are recovered from samples (Carrique-Mas & Davies 2008, Harvey & Price 1967, Kinde et al. 2004, Love & Rostagno 2008, Rostagno et al. 2005, Singer et al. 2009). Serotype bias has been documented since the 1950s. Enteritidis, for instance, is a major serotype in human illness, but when mixed in equal proportions and incubated overnight, Enteritidis is outgrown by Newport (Singer et al. 2009) and by Heidelberg and Senftenberg (Kinde et al. 2004). Growth of Enteritidis in the presence of other serotypes in feces or hatchery fluff is also reduced by the stress of drying (Cox et al. 2010c). The unequal competition between serotypes may have implications for pooling of samples in monitoring of Salmonella in poultry. Some samples are routinely pooled in neck skin and farm sampling in Europe, possibly distorting the picture of what serotypes are present in multi-serotype samples. In a study that sampled retail chicken wings and turkey necks by two culture methods, multiple serotypes were recovered from a high proportion of Salmonella-positive samples (Temelli et al. 2010). In a recent study in our laboratory, 49 of 52 broiler chicken carcasses rinsed immediately after defeathering were Salmonella-positive, with one serotype isolated from seven carcasses, two from 19 carcasses, three from 18 carcasses, and four from five carcasses (Cox et al. 2010a). Other studies have revealed isolation of multiple Salmonella serotypes from individual samples (FSANZ 2010, Jorgenson et al. 2002). Using multiple isolation media and picking three to five suspect colonies from each plate can result in the isolation of multiple Salmonella serotypes from a large percentage of individual samples. Picking only one colony per plate may underestimate the number of Salmonella serotypes that are present in a sample.
Two other recent research projects in our laboratory revealed different effects of plating media on serotype isolation from hatchery fluff and carcass rinses (Cox et al. 2010a,c). From the fluff samples, 455 presumptive Salmonella isolates were subtyped with antisera. Serogroup C1 (later identified as Lille) was isolated 34 times from BGS or Hektoen Enteric plates, but only once from XLT4 plates. Conversely, in the broiler carcass rinse study serotype Kiambu was detected 15 times on XLT4 but never on BGS. When the isolates were later restreaked, the Lille was determined to be a weak H2S producer that did not appear as typical Salmonella on XLT4 while the Kiambu was a strong H2S producer that formed smaller-than-average colonies on BGS. While the experience level of technicians is a major factor in recovery of Salmonella, experienced technicians can also overlook nontypical Salmonella colonies, even when multiple colonies are being selected.
Serotype differences may explain some of the difficulty in attributing human salmonellosis to different foods. Many studies have revealed a poor match between human and animal Salmonella serotypes (Heithoff et al. 2008, Jones et al. 2008, Kariuki et al. 2002, Pointon et al. 2008, Ravel et al. 2010, Sarwari et al. 2001, Schlosser et al. 2000, Stevens et al. 2009, Sumner et al. 2004,