On-line version ISSN 2224-9435
J. S. Afr. Vet. Assoc. vol.86 n.1 Cape Town 2015
Catriona H. LyleI; Cornelius H. AnnandaleII; Johan GouwsIII; Paul S. MorleyI, IV
IDepartment of Companion Animal Clinical Studies, University of Pretoria, South Africa
IIDepartment of Production Animal Studies, University of Pretoria, South Africa
IIIDepartment of Veterinary Tropical Diseases, University of Pretoria, South Africa
IVCollege of Veterinary Medicine and Biomedical Sciences, Colorado State University, United States
Salmonellosis is a common healthcare-associated infection in large-animal hospitals, and surveillance forSalmonella is an integral part of comprehensive infection control programmes in populations at risk. The present study compares the effectiveness of two culture techniques for recovery of Salmonella from environmental samples obtained in a large-animal referral veterinary hospital during a Salmonella outbreak. Environmental samples were collected using household cleaning cloths that were incubated overnight in buffered peptone water (BPW). Aliquots of BPW were then processed using two different selective enrichment and culture techniques. In the first technique (TBG-RV-XLT4) samples were incubated at 43 °C in tetrathionate broth and then Rappaport-Vassiliadis broth before plating on XLT4 agar. The second technique (SEL-XLD) involved incubation at 37 °C in selenite broth before plating on XLD agar. Salmonella was recovered from 49.7% (73/147) of samples using the TBG-RV-XLT4 technique, but only 10.2% (15/147) of samples using the SEL-XLD method. Fourteen samples (9.5%) were culture-positive using both methods, and 73 (49.7%) were culture-negative using both techniques. There were discordant results for 60 samples, including 59 that were only culture-positive using the TBG-RV-XLT4 method, and one sample that was only culture-positive using the SEL-XLD method. Salmonella was much more likely to be recovered using the TBG-RV-XLT4 method, and there appeared to be five times more false-negative results using the SEL-XLD technique. Environmental contamination with Salmonella may be underestimated by certain culture techniques, which may impair efforts to control spread in veterinary hospitals.
Salmonella is a common cause of nosocomial infection in large-animal hospitals and may result in hospital closure or restricted admissions (Benedict, Morley & Van Metre 2008; Dallap Schaer, Aceto & Rankin 2010; Steneroden et al. 2010; Tillotson et al. 1997). A number of large-animal hospitals routinely culture patient faeces and the hospital environment for Salmonella in order to detect nosocomial infection at the earliest stage (Dallap Schaer et al. 2010; Steneroden et al. 2010). This targeted surveillance to detect environmental contamination has been shown to be an important tool in detecting and mitigating outbreaks of Salmonella in veterinary hospitals (Dallap Schaer et al.2010; Steneroden et al. 2010; Tillotson et al. 1997).
Routine, non-targeted sampling and culture of environmental samples collected from healthcare environments is not recommended by the Centers for Disease Control and Prevention (CDC) and other authorities (Sehulster et al. n.d.), but targeted microbiological sampling (e.g. methods used specifically to detect Salmonella contamination in order to detect or help mitigate nosocomial outbreaks) is considered a useful aid to infection control processes (Dallap Schaer et al. 2010; Steneroden et al. 2010; Tillotson et al. 1997) as long as there is a predefined protocol for sample collection and culturing and the analysis and interpretation of results are used as a basis for determining whether there is a need for intervention and which specific actions should be taken (Sehulster et al. n.d.).
A number of factors can affect surveillance results, including sample site selection, sampling frequency, sampling technique, sample storage and culture technique (Corrente et al. 2004; Ewart et al. 2001; Harvey & Price 1983; Ruiz et al. 1996). Isolation of Salmonella from faeces and the environment requires use of a selective enrichment medium and subsequent plating on selective agar. There are many options for media and culture techniques and methods for isolation of Salmonella from faeces and environmental samples, since they are not standardised (Davies et al. 2000; Waltman & Mallinson 1995). These methods vary greatly amongst laboratories and little information has been published about direct comparisons of techniques (Davies et al. 2000; Love & Rostagno2008; World Organisation for Animal Health 2010; Rostagno et al. 2005; Singer et al. 2009; Voogt et al. 2002).
Inadequate Salmonella culture techniques that produce false-negative results may delay implementation of strategies to control nosocomial disease and so facilitate spread of Salmonella within the hospital and between patients (Dallap Schaer et al. 2010). The purpose of this study was to compare the recovery of Salmonella from environmental samples collected during a Salmonella outbreak in an equine hospital using two different culture methods.
Materials and methods
In 2012 during an outbreak of healthcare-associated Salmonella infections at a large-animal referral hospital there was an opportunity to directly compare different culture methods for recovery of Salmonella from environmental samples. Environmental samples were collected from the animal care areas of the hospital and cultured using two different culture methods. One method was adapted from that previously described for environmental surveillance in a veterinary teaching hospital (Burgess, Morley & Hyatt 2004), and the other was the existing method being used in the diagnostic laboratory associated with the referral hospital. Samples were collected using commercially available household cleaning cloths, which were incubated overnight in buffered peptone water (BPW). Samples of BPW were then cultured by two different methods. In the first technique (TBG-RV-XLT4) samples were incubated at 43 °C in tetrathionate broth with brilliant green (TBG) and then Rappaport-Vassiliadis broth (RV) before plating on xylose-lysine-tergitol (XLT4) agar. The second technique (SEL-XLD) involved incubation in selenite broth (SEL) at 37 °C before plating on xylose-lysine-deoxycholate (XLD) agar.
Sample site determination
Environmental samples were collected from 147 sites in the Onderstepoort Veterinary Academic Hospital. Sample sites were selected prior to sampling, considering hospital layout, traffic patterns for patients and personnel and how spaces are used in care for animals. These sites included areas used by animals only (e.g. stables), areas used by people only (e.g. offices) and areas used by both people and animals (e.g. treatment rooms). Within these sites hand-contact areas (e.g. telephones, keyboards, door handles) and foot-contact areas (e.g. floors) were sampled separately. High-risk areas such as high-traffic areas and those housing patients with gastro-intestinal disease were selected for more extensive sampling than low-risk areas (e.g. low-traffic areas).
Samples were collected with dry new household cleaning cloths (Supawipes®, Pick 'n Pay, Johannesburg, South Africa [SA]) in a similar manner to that previously described for collection of environmental samples in large-animal hospitals (Burgess et al. 2004). A single cloth was used at each site. Cloths were attached to a modified commercial floor mop to facilitate collection of floor and wall samples. This mop was disinfected with 70% ethanol between sampling of each site. Hand-contact areas were sampled with a cloth held in a gloved hand. After sampling each cloth was placed in a labelled sterile plastic bag (Whirl-Pak, Nasco, Fort Atkinson, WI, United States of America [USA]). Personnel changed gloves between each sample collection. An effort was made to sample approximately 80% of the surface area of each sample site.
BPW (Oxoid Ltd., Basingstoke, Hampshire, United Kingdom [UK]) (90 mL) was added to each labelled plastic bag containing a cloth. These samples were incubated for 24 hrs at 37 °C and samples were then vortexed and aliquots were cultured using two different methods in parallel (TBG-RV-XLT4 method and SEL-XLD method). For the TBG-RV-XLT4 method, 100 uL of iodine iodide solution was added to 9 mL TBG (Selecta Media, Randburg, Gauteng, SA). The iodine iodide solution was prepared by mixing 6 g iodine (Merck, Modderfontein, Gauteng, SA) and 5 g potassium iodide (Merck, Modderfontein, Gauteng, SA) in 20 mL sterile distilled water. One mL of BPW was then added to the TBG and incubated for 24 hrs at 43 °C. The TBG was then vortexed and 0.1 mL was transferred to 10 mL of RV (Oxoid Ltd, Basingstoke, Hampshire, UK) and incubated for 24 hrs at 43 °C before it was vortexed and streaked for isolation on XLT4 agar (Selecta Media, Randburg, Gauteng, SA) and incubated overnight at 43°C. Suspect colonies (pink colonies with or without black centres) were streaked for isolation onto Columbia blood agar plates (Selecta Media, Randburg, Gauteng, SA) and XLT4 agar and incubated for 24 hrs at 37 °C (Columbia blood agar) and 43 °C (XLT4 agar). For the SEL-XLD method, 1 mL of BPW was transferred to 9 mL of SEL (Oxoid Ltd, Basingstoke, Hampshire, UK) and incubated overnight at 37 °C before it was vortexed, streaked for isolation onto XLD agar (Oxoid Ltd, Basingstoke, Hampshire, UK) and incubated overnight at 37 °C. Suspect colonies (pink colonies with or without black centres) were streaked for isolation onto Columbia blood agar plates and XLD agar and incubated for 24 hrs at 37 °C.
For both methods identification of Salmonella enterica isolates was confirmed by biochemical testing using a commercial kit (API10S, BioMirieux, Marcy I'Etoile, Rhône-Alpes, France). If samples cultured positive using both methods, the isolate from the TBG-RV-XLT4 method was stored at -70 °C in brain-heart infusion broth (Oxoid Ltd, Basingstoke, Hampshire, UK) pending serotyping. Positive isolates from the SEL-XLD method were only stored and serotyped if the sample did not culture-positive using the TBG-RV-XLT4 method.
Positive and negative control samples were created by culturing American Type Culture Collection (ATCC) reference strains of Salmonella Typhimurium (ATCC 13311) (Teddington, Middlesex, UK) and Escherichia coli(ATCC 25922) respectively on blood agar. These cultures were sampled using cotton-tipped swabs (The Scientific Group, Vorna Valley, Gauteng, SA). The tips of the swabs were then cut-off, placed in sterile plastic bags containing BPW and processed using the same methods as for the environmental samples. Laboratory personnel were not blinded as to the identity of the control samples.
Data management and analysis
Results of both culture methods were entered in a computer spreadsheet. Descriptive statistics were calculated, and differences in recovery by the different techniques were compared using McNemar's test and Cohen's kappa.
Positive and negative control samples were culture-positive and culture-negative respectively for S. enterica using both culture techniques. A total of 50.3% (74/147) of the environmental samples were culture-positive for S.enterica, with the TBG-RV-XLT4 method detecting S. enterica in 49.7% (73/147) of samples, and SEL-XLD detecting S. enterica in 10.2% (15/147) of samples. Overall 9.5% (14/147) of samples were culture-positive on both techniques, 49.7% (73/147) of samples were culture-negative using both methods, 40.1% (59/147) were only culture-positive using the TBG-RV-XLT4 method, and the one remaining sample (0.7%, 1/147) was only culture-positive using the SEL-XLD method. This marked discordance was found to be statistically significant using the McNemar test (P < 0.001), and the kappa statistic was 0.18.
Six different serotypes were identified amongst the isolates recovered in this study: S. Heidelberg, S. Kibusi, S. Kottbus, S. Orion, S. Typhimurium and S. Virchow (Table 1). The serotype of three isolates could not be determined by the reference laboratory using typing antisera. The majority (81.1%, 60/74) of S. enterica isolates obtained from culture-positive samples were serotype S. Kibusi.
This study showed that there was a marked difference in the ability to recover Salmonella from environmental samples, dependent on the culture method used. The TBG-RV-XLT4 culture method was more sensitive than the SEL-XLD culture method, with the latter method resulting in a large number of false-negative results. In a survey of laboratories in the USA regarding methods used for Salmonella culture from poultry samples, various formulations of selenite broth were used in > 1/3 of laboratories when culturing environmental samples (Waltman & Mallinson 1995). Veterinarians should therefore inquire about the methods being used at diagnostic laboratories for Salmonella culture, and this information should be used when selecting which laboratories to use.
False-negative results from environmental cultures can have important adverse effects on hospital infection control programmes. Firstly, if the environmental contamination is associated with healthcare-associated infections in patients, false-negative results for environmental cultures can contribute to the impression that salmonellosis is community-acquired rather than hospital-acquired, and so delay identification and mitigation of a nosocomial outbreak (Dallap Schaer et al. 2010). Secondly, false-negative environmental results may result in reopening of facilities that are still contaminated, with subsequent recurrence of nosocomial infection (Schott et al. 2001; Tillotson et al. 1997), which could necessitate repeat facility closure. Lastly, false-negative results delay the identification and hence cleaning and disinfection of contaminated areas (Dallap Schaer et al. 2010).
Early identification of contaminated stables in hospitals aids in resolving contamination. However, use of additional culture steps also has consequences, as this means it takes longer to obtain negative results. For example, even though the TBG-RV-XLT4 method had much greater sensitivity in this study when compared with the SEL-XLD method, results were obtained 24 hrs sooner using the SEL method. Delays in obtaining culture results can create important logistical problems in busy hospitals where patient turn-over is high. However, the magnitude of the difference in sensitivity outweighs the difference in culture times when considering the safety of patients and personnel.
A previous report demonstrated a difference in Salmonella recovery between two environmental sampling and culture systems in a large-animal hospital (Ruple-Czerniak et al. 2014). However, it was not possible to determine from that study the relative contributions of the sampling technique and the culture method to the overall detected difference in Salmonella recovery. This study demonstrates a clear difference in Salmonella recovery that can be attributed to culture methods.
A number of studies have investigated the use of different media for isolation of Salmonella from pigs, humans, chickens, reptiles and food products (Corrente et al. 2004; Davies et al. 2000; Love & Rostagno 2008; Oboegbulem 1993; Rostagno et al. 2005; Ruiz et al. 1996; Voogt et al. 2002). Comparisons of selective and enrichment media for isolation of Salmonella from horses are very limited. In a comparison of tetrathionate (TET) enrichment, RV selective incubation and SEL selective incubation of horse faeces, TET was found to be superior to RV for detection of Salmonella, and no Salmonella was isolated from samples selectively incubated in SEL (Babuet al. 2008).
This study suggests similar enhanced Salmonella recovery from an equine hospital environment using TET enrichment and RV selective incubation when compared with SEL selective incubation. The majority of Salmonellaisolates obtained by environmental sampling were S. Kibusi (81%, 60/74). Different Salmonella serotypes vary in their tolerance for selective media and it is possible that the differences between the two techniques may not have been the same for other serotypes (Singer et al. 2009). However, TET broths have been shown to yield a much higher test sensitivity for isolation of Salmonella than SEL broths for a large number of serotypes (Carlson & Snoeyenbos 1974). In addition, it is possible that different serotypes were identified by the two different culture methods, but as positive isolates using the SEL-XLD method were only stored and serotyped if they were negative on the TBG-RV-XLT4 method it was not possible to determine this.
This study was carried out in a veterinary hospital during a nosocomial outbreak of salmonellosis and illustrated that the extent of environmental contamination would have been underestimated using the SEL-XLD method. This may have resulted in a less aggressive approach to the outbreak and so facilitated further spread of the organism and subsequent nosocomial infection. Hospital contamination was extensive at the time of this study; nevertheless a large difference was identified between culture methods. It is possible that the differences between the two methods may have been even further exacerbated if there were only low levels of environmental contamination. For example, it has been shown that large bacterial counts of S. enterica subsp. arizonae and lower numbers of competitors result in reduced ability to detect differences in efficacy of selective media (Snoeyenbos & Carlson1972).
There were a number of differences between the two culture methods. It was not possible to determine the relative contributions of the different components of each culture method to the final observed difference in culture results. Further investigation would be required to determine if this difference was caused (all or in part) by the use of TBG enrichment versus SEL selective incubation, by use of two versus one selective enrichment steps, or by incubation at 43 °C versus 37 °C.
In summary, this study identified an important difference in Salmonella recovery from hospital environmental samples dependent on culture method. Use of a more sensitive culture technique should facilitate earlier recognition and mitigation of environmental Salmonella contamination, which may ultimately prevent the need for hospital closure.
The authors thank Erick Kapp and Debbie Landman for laboratory work and all Veterinary Academic Hospital staff that assisted with sample collection, in particular, Cursti Windt.
The authors declare that they have no financial or personal relationships which may have inappropriately influenced them in writing this article.
C.H.L. (University of Pretoria) conceptualised and designed the experiment, collected the samples, analysed the data and compiled the manuscript. C.H.A. (University of Pretoria) and P.S.M. (Colorado State University) conceptualised and designed the experiment and edited the manuscript. J.G. (University of Pretoria) performed the experiment and edited the manuscript.
Babu, N., Singh, B.R., Agrawal, R.K., Vijo, T.V., Srivastava, S.K. & Yadav, M.P., 2008, 'Prevalence of Salmonella in equids determined by microbiological culture, standard tube agglutination test and PCR', Haryana Veterinarian 47, 58-63. [ Links ]
Benedict, K.M., Morley, P.S. & Van Metre, D.C., 2008, 'Characteristics of biosecurity and infection control programs at veterinary teaching hospitals', Journal of the American Veterinary Medical Association 233, 767-773. http://dx.doi.org/10.2460/javma.233.5.767 [ Links ]
Burgess, B.A., Morley, P.S. & Hyatt, D.R., 2004, 'Environmental surveillance for Salmonella enterica in a veterinary teaching hospital', Journal of the American Veterinary Medical Association 225, 1344-1348. http://dx.doi.org/10.2460/javma.2004.225.1344 [ Links ]
Carlson, V.L. & Snoeyenbos, G.H., 1974, 'Comparative efficacies of selenite and tetrathionate enrichment broths for the isolation of Salmonella serotypes', American Journal of Veterinary Research 35, 711-718. [ Links ]
Corrente, M., Madio, A., Friedrich, K.G., Greco, G., Desario, C., Tagliabue, S. et al., 2004, 'Isolation of Salmonellastrains from reptile faeces and comparison of different culture media', Journal of Applied Microbiology 96, 709-715. http://dx.doi.org/10.1111/j.1365-2672.2004.02186.x [ Links ]
Dallap Schaer, B.L., Aceto, H. & Rankin, S.C., 2010, 'Outbreak of salmonellosis caused by Salmonella entericaserovar Newport MDR-AmpC in a large animal veterinary teaching hospital', Journal of Veterinary Internal Medicine 24, 1138-1146. http://dx.doi.org/10.1111/j.1939-1676.2010.0546.x [ Links ]
Davies, P.R., Turkson, P.K., Funk, J.A., Nichols, M.A., Ladely, S.R. & Fedorka-Cray, P.J., 2000, 'Comparison of methods for isolating Salmonella bacteria from faeces of naturally infected pigs', Journal of Applied Microbiology 89, 169-177. http://dx.doi.org/10.1046/j.1365-2672.2000.01101.x [ Links ]
Ewart, S.L., Schott, H.C., 2nd, Robison, R.L., Dwyer, R.M., Eberhart, S.W. & Walker, R.D., 2001, 'Identification of sources of Salmonella organisms in a veterinary teaching hospital and evaluation of the effects of disinfectants on detection of Salmonella organisms on surface materials', Journal of the American Veterinary Medical Association 218, 1145-1151. http://dx.doi.org/10.2460/javma.2001.218.1145 [ Links ]
Harvey, R.W. & Price, T.H., 1983, 'Effect of stool storage at room temperature on salmonella isolation from faeces', Journal of Hygiene (London) 91, 267-271. http://dx.doi.org/10.1017/S0022172400060277 [ Links ]
Love, B.C. & Rostagno, M.H., 2008, 'Comparison of five culture methods for Salmonella isolation from swine fecal samples of known infection status', Journal of Veterinary Diagnostic Investigtion 20, 620-624. http://dx.doi.org/10.1177/104063870802000514 [ Links ]
Oboegbulem, S.I., 1993, 'Comparison of two enrichment media and three selective media for isolation of salmonellae from fresh chicken carcass rinse fluids and sewer swabs', International Journal of Food Microbiology 18, 167-170. http://dx.doi.org/10.1016/0168-1605(93)90221-2 [ Links ]
Rostagno, M.H., Gailey, J.K., Hurd, H.S., Mckean, J.D. & Leite, R.C., 2005, 'Culture methods differ on the isolation of Salmonella enterica serotypes from naturally contaminated swine fecal samples', Journal of Veterinary Diagnostic Investigation 17, 80-83. http://dx.doi.org/10.1177/104063870501700118 [ Links ]
Ruiz, J., Nunez, M.L., Diaz, J., Lorente, I., Perez, J. & Gomez, J., 1996, 'Comparison of five plating media for isolation of Salmonella species from human stools', Journal of Clinical Microbiology 34, 686-688. [ Links ]
Ruple-Czerniak, A., Bolte, D.S., Burgess, B.A. & Morley, P.S., 2014, 'Comparison of two sampling and culture systems fro detection of Salmonella enterica in the environment of a large animal hospital', Equine Veterinary Journal 46(4), 499-502. Doi: 10.111/evj.12193. [ Links ]
Schott, H.C., 2nd, Ewart, S.L., Walker, R.D., Dwyer, R.M., Dietrich, S., Eberhart, S.W. et al., 2001, 'An outbreak of salmonellosis among horses at a veterinary teaching hospital', Journal of the American Veterinary Medical Association 218, 1152-1159, 1100. [ Links ]
Sehulster, L.M., Chinn, R.Y.W., Arduino, M.J., Carpenter, J., Donlan, R., Ashford, D. et al., n.d., Guidelines for environmental infection control in health-care facilities. Recommendations from CDC and the Healthcare Infection Control Practices Advisory Committee (HICPAC). Chicago, IL: American Society for Healthcare Engineering/American Hospital Association; 2004, viewed 15 May 2014, from http://www.cdc.gov/hicpac/pdf/guidelines/eic_in_hcf_03.pdf [ Links ]
Singer, R.S., Mayer, A.E., Hanson, T.E. & Isaacson, R.E., 2009, 'Do microbial interactions and cultivation media decrease the accuracy of Salmonella surveillance systems and outbreak investigations?', Journal of Food Protection 72, 707-713. [ Links ]
Snoeyenbos, G.H. & Carlson, V.L., 1972, 'Comparative efficiency of tetrathionate and selenite enrichment broths for the isolation of Arizona serotypes', Avian Diseases 16, 756-766. http://dx.doi.org/10.2307/1588755 [ Links ]
Steneroden, K.K., Van Metre, D C., Jackson, C. & Morley, P.S., 2010, 'Detection and control of a nosocomial outbreak caused by Salmonella newport at a large animal hospital', Journal of Veterinary Internal Medicine 24, 606-616. http://dx.doi.org/10.1111/j.1939-1676.2010.0484.x [ Links ]
Tillotson, K., Savage, C.J., Salman, M.D., Gentry-Weeks, C.R., Rice, D., Fedorka-Cray, P.J. et al., 1997, 'Outbreak of Salmonella infantis infection in a large animal veterinary teaching hospital', Journal of the American Veterinary Medical Association 211, 1554-1557. [ Links ]
Voogt, N., Nagelkerke, N.J., Van De Giessen, A.W. & Henken, A.M., 2002, 'Differences between reference laboratories of the European community in their ability to detect Salmonella species', European Journal of Clinical Microbiology and Infectious Diseases 21, 449-454. http://dx.doi.org/10.1007/s10096-001-0685-x [ Links ]
Waltman, W.D. & Mallinson, E.T., 1995, 'Isolation of Salmonella from poultry tissue and environmental samples: A nationwide survey', Avian Diseases 39, 45-54. http://dx.doi.org/10.2307/1591981 [ Links ]
World Organisation for Animal Health, 2010, Manual of Diagnostic Tests and Vaccines for Terrestrial Animals, chapter 2.9.9, Salmonellosis. Paris: World Organisation for Animal Health (OIE). [ Links ]
Private Bag X04
Received: 10 Mar. 2015
Accepted: 04 June 2015
Published: 13 Aug. 2015