Recent Progress in Controlling Salmonella in Veterinary Hospitals

Subclinical and clinical Salmonella infections are reasonably common among large animal patients. Segregation based on the severity of systemic illness and clinical signs is useful for separating high-risk patients. However, this method alone may not identify all shedding animals. Environmental contamination with Salmonella is common in areas close to patients shedding Salmonella, and environmental microbiology can help to detect and minimize risks related to nosocomial spread. Peroxygen disinfectants show great promise as effective disinfectants in footbaths and for aerosol dispersal for general environmental disinfection.

1. Introduction

Biosecurity personnel at the James L. Voss Veterinary Teaching Hospital (JLV-VTH) at Colorado State University have been engaged in ongoing investigations to refine and objectively evaluate risks and control measures regarding Salmonella enterica at the JLV-VTH. Major outbreaks of nosocomial Salmonella infections have been documented in several large veterinary hospitals during the past decade [1-4]. Nosocomial infections with Salmonella have also been reported in human health care settings [5-8]. The Large Animal Hospital at the JLV-VTH was closed in 1996 and 2001 to prevent further nosocomial spread of Salmonella among hospitalized animals. The 1996 outbreak resulted in a partial closing and later, a total hospital closure of the large animal facility over a 3-mo period; it cost more than an estimated $500,000. Total losses were actually greater, because this estimate only includes expenses related to mitigation and does not include opportunity losses such as lost revenues. The results of a survey conducted in 1997 among veterinary teaching hospitals showed that 12 of 18 responders reported 18 outbreaks of nosocomial disease in the period between 1985 and 1996. Seventy-eight percent of outbreaks were attributed to Salmonella infection, and six of these outbreaks resulted in hospital closures [9]. Although zoonotic transmission has not been reported at the JLV-VTH in the past decade, Salmonella can cause serious illness in humans and has been seen in association with other reported outbreaks. After the 1996 outbreak at Colorado State University, a comprehensive surveillance and prevention program was initiated to control risks associated with Salmonella and other important infectious agents. Continued evaluation and refinement of this ongoing biosecurity program, including implementation of targeted research, has improved our understanding of Salmonella spread and persistence in the environment as well as efficacy of various control measures to prevent nosocomial Salmonella infections. These findings, with emphasis on novel strategies for environmental monitoring and disinfection, are presented below.

2. Materials and Methods

Patient Monitoring

Fecal samples were obtained from all large animal patients admitted to the JLV-VTH for ≥1 day and cultured to identify Salmonella. Samples were collected at the time of admission and every Monday, Wednesday, and Friday throughout hospitalization. Fecal samples were cultured for Salmonella using standard laboratory procedures. Suspect Salmonella isolates were serogrouped using commercial polyvalent O antisera and group-specific antisera. Salmonella isolates were serotyped by the U.S. Department of Agriculture National Veterinary Services Bacteriology Laboratory [a]. Patient signalment and hospitalization information was collected from medical records. Information collected regarding patients’ health and management during the 48 h preceding each sample collection included categorical assessment of the patients’ overall systemic illness status, presence of soft fecal consistency or diarrhea, fever, leukopenia, anesthesia or surgery, antimicrobial treatment (none, oral, parenteral, topical, or ophthalmic), significant reduction in dietary intake, body systems affected, stabling location when sampled, and attending service. Factors affecting the likelihood that patients were shedding Salmonella were analyzed using logistic regression with generalized estimating equations to control for repeated sampling of horses. Factors associated with the likelihood of Salmonella recovery in univariable models (p < 0.20) were included in multivariable modeling using backward stepping to determine which variables were retained in the final model (p < 0.10).

Environmental Monitoring

Environmental samples were collected approximately once a month at an average of 60 locations throughout the hospital using a commercially available electrostatic dust collection wipe [b]. Sites sampled in each location included the floor surface in designated rooms and hand-contact surfaces (door knobs, handles, keyboards, telephones, medical instruments, etc.). Floor samples were collected using the commercial sweeper, which was disinfected with 70% ethanol and allowed to dry between uses. After collection, the wipes were cultured for Salmonella using an enrichment protocol (incubation in buffered peptone water for 24 h at 35°C followed by a standard Salmonella culture). Suspect colonies were evaluated using Salmonella polyvalent O antisera and group-specific antisera. Salmonella isolates were serotyped at the U.S. Department of Agriculture National Veterinary Services Laboratories [a].

Footbath Study

An investigator wearing rubber boots [c] walked in a serpentine pattern for 2 min in a bedded stall contaminated through routine stabling of an adult cow. Two disinfectants, quaternary ammonium [d] and a peroxygen compound [e], were prepared according to manufacturers’ instructions and used in the stall. Tap water was used as a control treatment for comparison. Sterile swabs moistened with neutralizing broth were used to swab four marked areas on each boot. Pre-disinfection bacterial counts were estimated by sampling sites on one boot after contamination. Post-disinfection sampling was performed on the second boot 7 min after stepping out of the footbath. Ten pairs of boots were sampled for each type of solution tested. Differences between treatments were analyzed using regression analysis [f]. Separate analyses evaluated bacterial counts estimated at 24 or 48 h using either a tryptic soy agar (TSA) plate or MacConkey agar.

Peroxygen Disinfectant[e] Misting

Approximately 5.6 x 107 cfu of Salmonella [g] were inoculated onto each of 40 transparencies (36 cm2 each) and allowed to dry at room temperature for 14 h. The transparency squares were then placed in various locations throughout the food animal ward. The locations included 20 vertical surfaces (low and high) and 20 horizontal surfaces (low and high). The entire facility was then misted with ~180 L of 4% peroxygen disinfectant [e] using backpack foggers [h] to distribute the disinfectant. Transparency squares were collected 30 min after all the misting had been completed. The squares were submerged in neutralizing broth [i] and plated on TSA plates using a spiral plater [j]. The estimated concentration of recovered bacteria was calculated based on colony counts read the next day. Positive and negative controls samples were processed simultaneously with the test samples.

3. Results

Salmonella Shedding Study

As a part of the surveillance program, 3504 samples were submitted from 1417 patients (1508 separate admissions). Salmonella isolates were recovered from 3.5% of samples (124/3504), which was 5.5% of patients (77/1396). Crude Salmonella recovery rates were highest among bovine patients, followed by caprine, equine, and camelid patients. The highest percentage of positive patients were housed in the calf isolation ward, followed by equine isolation, equine colic ward, food animal ward, and main equine ward. Animals between 3 and 12 mo of age were least likely to shed Salmonella. The highest percentage of positive samples was collected during the period from the fourth to the seventh day of hospitalization. In the multivariable model, the patient characteristic most strongly associated with higher rates of Salmonella recovery was the subjective characterization of severity of patients’ systemic illness, regardless of the body systems affected. Diarrhea occurring within 48 h of sampling was also associated with an increased rate of Salmonella recovery but having other types of gastrointestinal disease was not. Antimicrobial treatment by a parenteral route within 48 h of sampling was associated with an increased rate of Salmonella recovery, but oral treatment was associated with a lower recovery rate. Variables that were not retained in the final model included fever, leukopenia, history of anesthesia or surgery, reduction in feed intake, body system affected, and gender.

Footbath Study

There was no statistically detectable reduction in bacterial counts from boots after quaternary ammonium disinfectant treatment; however, a >70% reduction in bacterial counts was observed after treatment with the peroxygen disinfectant [e]. There was a detectable increase in bacterial counts on boots treated with water compared with pre-treatment counts.

Environmental Monitoring

Salmonella isolates were recovered from 59 sites throughout the VTH. Similar Salmonella isolates were frequently recovered from fecal samples and the immediate environment of Salmonella positive patients.

Peroxygen Disinfectant [e] Misting

Ninety-five percent (38/40) of the transparencies showed a >6 log10 reduction in bacterial load after Virkon S [e] misting; this represents a 1,000,000-fold reduction in bacterial numbers.

4. Discussion

Salmonella infections represent a major risk to hospitalized large animal patients as well as to the operational status of a large animal hospital. At JLV-VTH, food animal and equine patients are segregated as are patients with gastrointestinal diseases. Additionally, patient management is affected by results of active surveillance for Salmonella shedding using the system described here. Our ongoing studies are used to gain a better understanding of which patients are more likely to shed Salmonella so that future monitoring efforts can be more cost effective and still efficiently identify and segregate Salmonella-shedding horses. The patient parameter showing the strongest association with fecal shedding of Salmonella was the subjective categorization of severity of systemic illness. Antimicrobial treatment by a parenteral route within 48 h of sampling was associated with an increased rate of Salmonella recovery, but oral treatment was associated with a lower recovery rate. This association may have been confounded by the clinicians’ decisions to select parenteral routes of administration in patients with gastrointestinal diseases and use oral treatment in patients with less severe systemic illness. A higher percentage of samples collected from equine patients housed in isolation were positive for Salmonella compared with samples from the main equine barn, reinforcing that our current separation procedures are successful in isolating Salmonella shedding horses. However, if we had relied on clinical criteria typically associated with salmonellosis, such as presence of one or more of leukopenia, fever, or diarrhea, to determine which patients should be cultured, we estimate that we would have detected ~55% fewer positive samples. Similarly, collection of samples from only colic horses and patients housed in isolation would have resulted in detecting 80% fewer positive samples. Currently, we do not know how big of a risk these horses without signs of gastrointestinal diseases would have posed to the other patients, and this aspect of Salmonella biology warrants further investigation. In a study by Traub-Dargatz et al. [10], fecal shedding of Salmonella in non-colic horses was detected, but it was not accompanied by nosocomial spread of Salmonella. Despite this uncertainty regarding the risks imposed by non-colic Salmonella shedding horses, at JLV-VTH, we currently feel that microbiological monitoring of all large animal patients is warranted in the teaching hospital environment and that exclusion of any group of patients could result in increased risk of environmental contamination and potentially increased risk of nosocomial Salmonella infections. In a compromise intended to decrease some costs of microbiological monitoring and maintain adequate rigor in our surveillance program, we have recently decreased the frequency of sampling. At present, all large animal patients at JLV-VTH are cultured for Salmonella at admission and each Tuesday and Friday.

Environmental monitoring is another useful tool aimed at maximizing control of nosocomial Salmonella infections. We have found that the electrostatic wipe [b] culture method is very sensitive at detecting environmental contamination with Salmonella [11]. The electrostatic wipes [b] are readily available in many stores and are very easy to use, which makes this method extremely practical. Salmonella isolates with the same serogroup, serotype, and antibiograms as isolates obtained from hospitalized patients were obtained from the environment on several occasions. In addition, we obtained positive environmental samples that were not accompanied by positive fecal samples from the hospitalized patients. This most likely indicates the presence of a Salmonella shedding patient that was not detected based on fecal cultures. Detection of environmental contamination can alert clinicians, students, and cleaning crews to the potential of such a situation and result in increased efforts at cleaning and adherence to other biosecurity precautions.

Footbaths are one of the most common control measures used in efforts to reduce trafficking of pathogenic microorganisms in livestock operations and veterinary hospitals. However, few other objective investigations are available documenting the efficacy of footbaths in reducing bacterial counts on footwear. The aim of the current study was to evaluate the efficacy of footbath under conditions commonly encountered in veterinary hospitals. It is generally recommended that 10 - 15 min of contact time be used to obtain maximal decontamination with common disinfectants. Although personnel would generally never stand in footbaths for 15 min, antimicrobial activity should theoretically continue even after stepping out of footbaths, unless the disinfectant solution was rinsed from surfaces. Thus, we elected to collect microbiological samples from the boots not directly after the use of a footbath, but 7 min after use. The results of this study suggested that Virkon S [e] is superior to the quaternary ammonium compound in reducing bacterial contamination on footwear after briefly stepping in and out of the footbath. As a result, beginning early in 2004, we have begun to use peroxygen disinfectant [e] in all footbaths at the JLV-VTH.

Cold misting with peroxygen disinfectant [e] is another infection control measure that has recently been implemented at JLV-VTH. Misting with peroxygen disinfectant [e] resulted in a >1,000,000-fold reduction in bacterial counts on transparencies experimentally contaminated with Salmonella. Although the dynamics of Salmonella survival on different types of surfaces was not evaluated in these experimental settings, the use of transparencies allowed for simultaneous testing of a large number of sites within the treated area. Only two transparencies showed minimal or no reduction in bacterial counts after misting. This may have reflected human error during misting. The benefits of routine use of cold misting with peroxygen disinfectant [e] in veterinary hospital settings would include its wide-range antimicrobial action and the significantly decreased labor time required to disinfect large areas. Also, misting would potentially minimize microbial contamination in hard-to-reach areas. Peroxygen disinfectant [e] has very low toxicity and is biodegradable; thus, it’s environmentally friendly [12]. It has been reportedly used in the presence of animals without adverse effects. However, there are no reports in peer-reviewed journals regarding peroxygen disinfectant [e] safety of peroxygen compounds when used in this manner over a period of time. Until such data become available, it would be prudent to remove animals from fogged buildings whenever possible. The manufacturer of peroxygen disinfectant [e] recommends fogging as a way of controlling respiratory infections in veterinary settings. Although this aspect of cold fogging was not evaluated in this study, it would seem to be a potential additional benefit of routine fogging or misting with peroxygen disinfectant [e], particularly in areas where animals that show signs of infectious respiratory diseases are stabled.

Funding for these projects was provided by James L. Voss Veterinary Teaching Hospital and the U.S. Department of Agriculture, the Cooperative State Research Education and Extension Services for the Colorado State University Program for Economically Important Infectious Animal Diseases at Colorado State University. The authors thank Denise Bolte and the CSU Diagnostic Laboratories.

Footnotes

1. United States Department of Agriculture National Veterinary Services Bacteriology Laboratory, Ames, IA 50010.
2. Swiffer, Proctor & Gamble, Cincinnati, OH 45202.
3. Tingley, Plainfield, NJ 07080.
4. A464N, Airkem Professional Products, St. Paul, MN 55118.
5. Virkon S, Antec International, Sudbury, Suffolk CO10 2XD, UK.
6. PROC GENMOD, Statistical Analysis Software, version 8.2, SAS Institute Inc., Cary, NC 27513.
7. Salmonella enterica serovar Typhimurium, Sarb #65, Salmonella Genetic Stock Center, Calgary, Alberta, Canada T2N 1N4.
8. Model 430 motorized mist blower, Solo, Newport News, VA 23605.
9. D/E Broth, BD Difco, Franklin Lakes, NJ 07417.
10. Model D spiral plater, Spiral Biotech, Inc., Norwood, MA 02062.