1 EPIDEMIOLOGY OF NOSOCOMIAL Salmonella INFECTIONS IN HOSPITALIZED HORSES By ABEL B. EKIRI A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2008
2 2008 Abel B. Ekiri
3 To my late dad, Dr. Richard Bulamu, who untiringly contributed to my education.
4 ACKNOWLEDGMENTS I thank my graduate committee members, Drs. Robert J. Mackay, David E. Freeman, Jack M. Gaskin and Jorge A. Hernandez for their trem endous support. Special thanks go to my major professor, Jorge A. Hernandez for his invalu able support and guidance in my professional development. I extend my gratitude to Katherine Henry for her support at various levels. I also thank Maria von Chamier for her help with data collect ion. I appreciate the help of Sharon Hewitt and Marsha Swilley in providing access to necessary medical records. Lastly, I would like to thank my wife, Annet Kirabo, for her love and care. I also thank my Mom, Josephine Birabwa, and my brothers and sisters for their moral support and encouragement.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........6 ABSTRACT....................................................................................................................... ..............7 CHAPTER 1 LITERATURE REVIEW.........................................................................................................9 Etiology....................................................................................................................... ............10 Pathogenesis................................................................................................................... ........11 Diagnosis...................................................................................................................... ..........14 Epidemiology................................................................................................................... .......18 Prevention, Control, and Management of Salmonella in Horses............................................25 2 EPIDEMIOLOGY OF NOSOCOMIAL Salmonella INFECTIONS.....................................30 Materials and Methods.......................................................................................................... .31 Study Population.............................................................................................................31 Primary cases............................................................................................................32 Nosocomial cases.....................................................................................................32 Control horses..........................................................................................................33 Study Design...................................................................................................................33 Collection of Fecal Samples............................................................................................34 Microbiological Procedures for Detection of Salmonella Organisms.............................34 Data Collection................................................................................................................35 Statistical Analysis..........................................................................................................36 Results........................................................................................................................ .............37 Objective 1.................................................................................................................... ...37 Objective 2.................................................................................................................... ...39 Discussion..................................................................................................................... ..........39 APPENDIX A QUESTIONAIRE FOR SA LMONELLOSIS STUDY..........................................................52 LIST OF REFERENCES............................................................................................................. ..54 BIOGRAPHICAL SKETCH.........................................................................................................61
6 LIST OF TABLES Table page 2-1 Characterization of 16 hor ses classified as nosocomial cases in objective 1....................44 2-2 Characterization of 21 hor ses classified as nosocomial cases in objective 2....................45 2-3 Objective 1 Caseload, number of horses shedding Salmonella at admission, fecal samples collected, and hospital durat ion in case and control horses.................................46 2-4 Objective 2 Caseload, number of horses shedding Salmonella at admission, fecal samples collected, and hospital durat ion in case and control horses.................................46 2-5 Objective 1 Frequency di stribution of host factors, hos pital procedures, crude odds ratios (OR), and 95% confidence intervals (CI) of investigated risk factors among case and control horses......................................................................................................47 2-6 Objective 2 Frequency di stribution of host factors, hos pital procedures, crude odds ratios (OR), and 95% confidence intervals (CI) of investigated risk factors among case and control horses......................................................................................................49 2-7 Objective 1 Multivaria ble conditional logistic regression model for nosocomial Salmonella infections in hospitalized horses.....................................................................51 2-8 Objective 2 Multivaria ble conditional logistic regression model for nosocomial Salmonella infections in hospitalized horses.....................................................................51
7 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science EPIDEMIOLOGY OF NOSOCOMIAL Salmone lla INFECTIONS IN HOSPITALIZED HORSES By Abel B. Ekiri May 2008 Chair: Jorge Hernandez Major: Veterinary Medical Sciences The objectives of this study were to exam ine the relationship between abdominal surgery and nosocomial Salmonella infections, and to examine the re lationship between high caseload in combination with abdominal surgery and nosocomial Salmonella infections in horses hospitalized with signs of gastro intestinal tract disease. A cas e control study was conducted at the Large Animal Hospital, University of Florida. 51 and 105 horses for objective 1 and 2 respectively with signs of gastrointestinal disease we re enrolled for the study. To accomplish the first objective, 1 to 4 control horses were matched to each nosocomial case by admission date of a primary case (i .e., a horse that tested positively to Salmonella at admission and later during hospitalization). The frequency of abdominal surgery and other investigated exposure factors was compared between horses classifi ed as cases and controls. In the second objective, 4 control horses were matched to each nosocomial case by admission year. The frequency of high caseload, abdominal surg ery and other factors was compared between horses classified as cases and controls. The odds of nosocomial Salmonella infection were 8 times higher in horses that underwent abdominal surgery, compared to ho rses that did not (OR = 8.20; 95% CI = 1.11,
8 60.24; P = 0.03). High caseload alone or in combinat ion with surgery was not associated with high risk of infection. Abdominal surgery was identified as a risk factor for nosocomial Salmonella infection. Horses that undergo abdominal surg ery require enhanced infection c ontrol and preventative care. Risk of nosocomial Salmonella infections may be reduced by implementation of preventative measures immediately after surgery, such as wa rd isolation (e.g., use of gloves, gowns, plastic boots and footbaths).
9 CHAPTER 1 LITERATURE REVIEW Fecal shedding of salmonella e by hospitalized horses is an important problem for large animal hospitals because of the risk of nosocomial Salmonella infections. In the past, several veterinary teaching hospitals have closed te mporarily because of outbreaks of nosocomial Salmonella infections in hospitalized horses.1,2,3,4,5,6 Consequences of outbreaks of nosocomial Salmonella infections can be severe, re sulting in human infections,4 equine fatalities,4 disruption of hospital routine,3 loss of teaching cases, loss of revenue,5 hospital renovation costs,5 and the potentially devastating effects of lawsuits. Salmonellosis is often suspected to be of nosoc omial origin when an infection is identified after animals have been hosp italized for 72 hours or longer or when the serotype and antimicrobial susceptibility pattern match those of a serotype previously identified as causing nosocomial infection.2,3,5,7 In a previous study,8 infection with Salmonella enterica ser Krefeld or Salmonella enterica ser. Typhimurium was considered noso comial if the mean time from admission to shedding was 4 days. In another study,4 affected horses were considered to have nosocomial colonization if there was no exacerbati on of the primary disease after admission to the hospital (i.e., if a fever or diar rhea did not develop) or to have nosocomial infection if clinical signs of salmonellosis were identif ied in addition to the primary di sease. In addition, cultures that yielded a S. enterica ser. Typhimurium isolate with an antimicrobial susceptibility profile identical to that of the isolate recovered from a point-source foal were initially assumed to indicate nosocomial infection. In that study, pulsed field ge l electrophoresis (PFGE) of S. enterica ser. Typhimurium isolates, yielded results supportive of nosocomial infection for most affected horses.
10 Three studies have attempted to identify risk factors associated with nosocomial Salmonella infections. In a study3 at a veterinary teaching hospital following an outbreak of salmonellosis due to S. enterica ser. Saint-paul a presenting complaint of colic, nasogastric intubation, and treatment with anti biotics were identified as risk factors for nosocomial infection. A second study9 was later carried out at the same hosp ital and horses that tested positive for Salmonella sp. other than the outbreak strain ( S. enterica ser. Saint-paul) were enrolled. A presenting complaint of colic, nasogastric intuba tion, and treatment with antibiotics were again identified as risk factors for nosocomial Salmonella infection.9 In a different study,8 impaction of the large colon, treatmen t with potassium penicillin G, change in feed, and high ambient temperature were associated with nosocomial Salmonella infections. Hospital procedures, high number of horses shedding Salmonella, and a high caseload are believed to affect the risk of nosocomial infections.4,5,10 However, no epidemiological study using objective research methods has been done to investigate these factors. Identifying these risk factors would guide the formulation of infec tion control measures that minimize the risk of nosocomial Salmonella infection in susceptible hospitalized horses. Etiology Salmonellosis is caused by a variety of strains of Salmonella species, all of which are gram-negative, motile, non-lactose fermenting rods belonging to the family Enterobacteriaceae, facultative anaerobic bacteria capable of living in the intestinal tract.11 All Salmonella species are classified as S. enterica or S. bongori. Salmonellae are subdivided into serogroups and then further into serotypes or serovars ba sed on testing with specific antisera.12 The most important serogroups in veterinary medici ne are A, B, C, D, and E.13 In the United States, the types of salmonellae isolated from samples of ill equids are reported annually by the National Veterinary Services Laboratory (NVSL) in Ames, Iowa.12 Within the species of Salmonella enterica over
11 2,000 different serotypes of veterinary and medical importance affect the gastrointestinal tract and result in diarrhea alone, or diarrhea in conjunction with fe ver, anorexia, depression, and shock.11 The most frequently isolated serotypes in horses include S. enterica ser. Typhimurium S. enterica ser. Anatum S. enterica ser. Newport S. enterica ser. Krefeld and S. enterica ser. Agona .7 The only host-adapted serova rs reported for horses is Salmonella sp. ser. Abortusequi, causing abortion between 5 a nd 10 months of gestation.14 Pathogenesis Once salmonellae infect a host, the first line of defense encountered is the acid barrier of the stomach.15 Organisms that survive the acid barrier travel to the small intestine, where they are exposed to secretory products of the intestine (such as IgA, defensins, bile salts, and intestinal mucus) and to intestinal microflora that prevent bacteria from penetrating enterocytes.16 Environmental factors in the intestin al lumen (such as oxygen concentration, osmolarity, and pH) affect the expression of Salmonella invasion genes (which determine the release of bacterial products requ ired for invasion of host cells).17 After crossing the mucus layer of the small intestine, salmonellae interact with both enterocytes and microfold cell (M cells).17 Once attached to the enterocytes or M cells, th e organisms are rapidly internalized. Salmonellae have the ability to induce endocytocis (ruff ling) in otherwise nonphagocytic cells and enterocytes. This process involves the formati on of large membranes around the bacteria by the host cell (enterocytes and M cells), as well as cy toskeletal rearrangements within the host cell itself. Once inside the cells, salmonellae migrate toward the submucosa of the small intestine where they interact with macrophages and lymphocytes.18 Salmonellae have the ability to survive within macrophages and use them to spread beyond the small intestine.15 They are transported via the lymphatics to th e submucosal lymphoid tis sue and then to the thoracic duct, where they ente r the systemic circulation.
12 During the asymptomatic phase of the inf ection, the organisms are localized to the intestine and replicate within macrophages and epithelial cells. Afte r becoming enveloped by macrophages, vacuoles called phagosomes are form ed from the fusion of the ends of membrane ruffles and lysosomes contained within the macr ophage. When a critical number of organisms have replicated, clinical signs result from the secretion of cytokines by the infected cells.19 A characteristic feature of Salmonella infection is the induction of an early inflammatory response in the intestinal epithelium, resulting in th e infiltration of polymor phonuclear leukocytes. The induction of such a response is likely due to the production of cytokines or other proinflammatory molecules by natu ral killer cells and macrophage s. The inflammatory response contributes to the pathophysiology of the inf ection, characterized by in flammatory diarrhea.18 Many genes that are important in the virulence of S. enterica ser. Typhimurium are located in two Salmonella Pathogenicity Islands (SPI) in the bacterial genome. To date, 5 SPI have been described; and two of them, SPI120 and SPI2,21 have been the most studied. Both of these regions encode specialized secretion systems, called type III secretion systems; and they are conserved in a wide variety of pathogenic serotypes.20 The effector proteins are injected through the needle complex formed by the type III secretion apparatus into the hosts cell cytosol causing several changes (including alteration of signaling pathways, cell death, inflammation, or alteration of phagocytosis). SPI1 is required for invasion a nd encodes a type III secretion system that translocates effector proteins in to the cytosol of the host cells, acting in the initial stages of infection; these effector proteins ar e not required for systemic disease.18 A type III secretion system is also encoded by SPI2, which translocates effector proteins that act in later stages of infection where they are important to the survival of organisms within macrophages and essential to sustained systemic infection.20,21
13 Other essential elements in the pathogenesis of salmonellosis are the regulatory proteins like PhoO/PhoQ that control the sy nthesis of multiple proteins at the level of gene transcription. These proteins regulate genes important for surv ival in macrophages, resistance to cationic antimicrobial proteins and acid pH and invasion of epithelial cells.22 Other regulatory genes implicated in the pathogenesis include: crp/cya, which regulates catabolite repression and surface proteins through adenylate cyclase; ompR/envZ, the regulators of porin gene transcription; katF, an alternative bacterial -factor that regulates catalase production; and spv, a virulence locus associated with cellular cytotoxi city. Surface molecules are also im portant in the pathogenesis of salmonellosis. The Vi antigen prevents antibodymediated opsonization, increases resistance to peroxide, and confers resistance to complement activation by th e alternative pathway and to complement-mediated lysis. The lipid A portion of the lipopolysaccharide (LPS) component of the bacterial outer membrane is a potent toxin for mammalian cells and an essential virulence determinant for S. enterica ser. Typhimurium in mice.15 Both innate and adaptive i mmunity are important against Salmonella infections.23 The principal route of natura l infection is oral, whereby the or ganism encounters components of the innate immune system. After surviving these co nditions, it travels to the mucosal surface where it encounters factors that kill the organism outri ght or inhibit replication, such as mucus which forms a physical barrier, lysozyme lactoferrin, and lactoperoxidase.24 During the initial stages of Salmonella infection, the hosts innate immune res ponse is conducted by na tural killer cells, natural killer T cells, neut rophils and macrophages. This first line of defense involves production of high levels of ga mma interferon (IFN) and tumor necrosis factor alpha (TNF) produced primarily by natural killer cells and macrophages. Macropha ge derived cytokines act by enhancing the bactericidal capacity of phagoc ytes, facilitating antigen presentation, and
14 influencing the T helper cell po larization of the immune response.23 The adaptive immune response with both humoral and cell-mediated i mmune responses, is involved in the acquired resistance to Salmonella infection.7 The humoral response involves production of IgA by plasma cells. IgA is the principal anti body isotype involved in mucosal immunity, and acts by binding to surface antigens and preventing attachme nt and penetration of salmonellae.16 Cell-mediated immunity is activated by the diffe rent antigens expresse d in salmonellae, whic h induce a specific T helper cell response. On the basis of differe ntial cytokine expression, both T helper 1 (Th1) and T helper 2 (Th2) type responses can be iden tified in most cases of infection by facultative intracellular pathogens.19 Th1 cells produce cytokines such as interferon (IFN) and interleukin 12 (IL-12), which regulate cell-me diated protective immune responses against intracellular organisms. In contrast, Th2 cells pr oduce cytokines such IL-4, IL-5, IL-10 and IL13 which regulate humoral immune res ponses which may not be protective.19,25 Salmonellae are intracellular bacteria whic h can survive in macrophages and dendritic cells.18 Consistent with the notion th at Th1/Th2 ratio predicts the competence of the response to intracellular pathogens, there is so me early evidence that invasive Salmonella generally elicit a Th1 immune response; however, th ere are also reports describing both Th1 and Th2 responses to Salmonella with the predominant response reflecti ng the particular pathway of antigen processing (MHC-II versus MHC-I dependent ) in macrophages and dendritic cells.19 Another adaptive immune response in th e early stages of infection is cell-mediated immunity by T lymphocytes, which function as effector cells, and T lymphocytes that function as regulatory cells in the late stages of disease.26 Diagnosis Diagnosis is based on clini cal signs and isolation of Salmonella organisms from feces, blood, or tissues.12 Several diagnostic techniques have been used to detect salmonellae. Aerobic
15 bacterial fecal culture and polymerase chain reac tion (PCR) are the most frequently used. Other techniques used include enzyme-linked imm unoabsorbent assay (ELI SA), pulse field gel electrophoresis (PFGE), plasmid profile analysis, phage typin g, gross and histopathological findings. Sensitivity of bacterial culture is limited by many factors including the method used to collect the sample, amount of sample submitted, when sample is collected in the course of disease, seasonal variation in shedding of the organisms, and met hod of bacteriological culture.27,28,29,30 It is easier to isolate th e organism early in the course of disease. It is more difficult to isolate Salmonella organisms when feces are watery thus it may be best to attempt culture prior to onset of watery diarrh ea or as soon as diarrhea is resolved.31 Submission of 1-2 g of feces for culture has been more successful in identifying Salmonella than has culturing fecal swabs. Salmonellae cannot be consistently cultu red from feces, therefore, a minimum of 5 consecutive culture-negative samples should be collected before considering a horse a nonshedder.32 Culturing of rectal mucosal biopsies in creases the probability of isolating the organism; however, the technique is not without risk to the horse.12 If the animal does not survive, microbiologic culture of the wall of the cecum, large colon, and ileum, mesenteric lymph nodes, and spleen should be more sens itive than microbiologic culture of feces.33 Laboratory identification of organisms of the genus Salmonella is achieved by biochemical tests and the serotype is confir med by serologic testing. Specimens are plated on any of several non-selective and selective ag ar media (blood, MacConkey, eosi n-methylene blue, bismuth sulfite, Salmonella-Shigella, and brilliant green agars) as well as into enrichment broth such as selenite or tetrathionate After 24 hours of incubati on in enrichment broth at 37 C, the isolate is subsequently subcultured onto the various agars and incubated for 24 hours at 37 C. Non-lactose
16 fermenting, hydrogen sulphide pr oducing colonies are then se lected, isolated for purity, inoculated onto urea agar slan ts, and incubated at 37 C for 24 hours to determine urease activity. Urease-negative organisms are iden tified using a commercial system (API 20E, Biomeriex Vitek, Inc). Bioc hemical identification of Salmonella has been simplified by systems that permit the rapid testing of 10 to 20 diffe rent biochemical parameters simultaneously. Following biochemical identification, the presumptive identification of Salmonella can be confirmed by antigenic analysis of O and H antigens using polyvalent and specific antisera. Serogroups of approximately 95% of all clinical isolates can be determined with the available group A-E typing antisera. In addition, Salmonella isolates are tested for antimicrobial sensitivity by using the minimal inhibitory concentration method and commercially prepared plates (e.g., Radiometer America, Westlake, OH). Salmonella isolates are then sent to a central or reference laboratory, such as the National Veterinary Se rvices Laboratory, Ames, Iowa for confirmation and serotyping.34 Compared to bacterial culture, PCR is a more sensitive diagnostic test that has been used for the detection of salmonellae either alone or in conjunction with bacterial culture. The advantages of using PCR in fecal samples are that it is a relatively rapid molecular identification test, it is highly sensitive and it requires th e submission of fewer samples compared to bacteriologic cultures.35 Bacteriologic cultures re quired at least 3 to 5 serial fecal samples to accurately assess an animals shedding status.36,37 The disadvantages of PCR are that: (1) it cannot discriminate between organisms that are a live or dead or between pathogenic infections and transient bacteria; (2) the organism is not available for serogrouping, serotyping and antimicrobial sensitivity testing (important in formation to determine if the infection is nosocomial in origin); and (3) it has lo wer specificity than bacterial culture.35,36 Since PCR uses
17 specific target genes located in specific parts of the bacterial genome, one of the potential disadvantages that this diagnostic test has is that bact eria can lose the target genomic sequence, resulting in false negative results.36,38 PCR has been used for environmental monitoring in hospitals where extreme sensitivity is needed, in attempts to identify locations that harbor salmonellae.35,36 Enzyme-linked immunoabsorbent assay (ELISA) is another test used for the detection of Salmonella antigens. Several ELISA tests are available in commercial kits; they are based on the O antigen (serotype-specific) and have been used for screening milk from bulk tanks in commercial dairy farms.39 ELISA has also been used for detection of antibodies against salmonellae in blood serum samples in hum ans, pigs, cattle and poultry.40 Some limitations of this assay include cross-r eactivity between serotypes41 and difficulties in detecting certain serotypes, especially those that are poorly invasive.40 ELISA has not been used on feces due to cross-reactivity with other enteric organisms.41,42 Pulse field gel electrophoresis (PFGE) and phage typing have been used in outbreak investigations.2,4,6,43,44 PFGE has been used to characterize biotypes of Salmonella enterica isolates. This procedure provides information a bout the bacterial genotyp e by separating variably sized fragments of chromosomal DNA after diges tion with 1 or more re striction endonucleases.4 Plasmid profiles have been used as markers to identify strains and assess the impact of improvements in hospital operation on nosocomial Salmonellosis.43,45,46,47 Plasmid profile analysis proved to be more sensitive than either serotyping or antimicrobial susceptibility testing in identifying Salmonella isolates.46 Gross lesions are most commonly observed in the cecum and ascending colon, but they also can be found in the small intestine and sm all colon. Lesions range from fluid-filled bowel
18 with mucosal edema to severe, diffuse fibrinonecr otic ulceration or sloughing of the mucosa with petechial and ecchymotic hemorrhages observed on th e serosal surface of the intestinal tract, the epicardium, and adrenal glands.7 Histopathologic findings often in clude necrosis and blunting of villous tips of the intestinal epithelium and inf iltration of the lamina propria with inflammatory cells, predominantly polymorphonuclear cells.7 Epidemiology The most common sources of infection are ot her horses, contaminated feed and water, carrier birds, rodents and other farm animal species that excrete the bacteria.7,29,48 Horses shedding salmonellae in their feces without showing clinical signs of salmonellosis are referred to as subclinical or asymptomatic shedders This group of animals may be not shed the organisms until exposed to stressful conditions.49 Asymptomatic shedders are an important source of contamination for ot her horses and the environment.50,51 Salmonellae can persist in the environment for protracted periods and have been recovered from contaminated soil after more than 300 days and from water after 9 months.52 These organisms can be killed by desiccation and exposure to sunlight but can survive in dried f ecal matter for as long as 30 months. Freezing will not necessarily kill the bacteria, pa rticularly if they are in food or other organic matter. They have been isolated from contaminated ice cream after more than two years.27 The transmission of salmonellae occurs most often by the fecal-oral route, although infection may also take place through the mucous membranes of the eyes and the nose via aerosol droplets. Several non-host adapted Salmonella serotypes have been report ed to infect horses of all ages and breeds; however, foals are more susceptible.53,54 In a previous study53 at a veterinary teaching hospital, foals with gast rointestinal tract disease were more likely to shed salmonellae than were adult horses with gastrointestinal tr act disease. In a study following an outbreak of neonatal salmonellosis, mares were found to be th e source of infection for foals and the absence
19 of clinical signs in mares allowed for incr eased exposure of foals through environmental contamination.51 Breeding farms are susceptible to outbreaks caused by salmonellae or other enteric infections because of the concentra tion of numerous immunologi cally immature horses.55 Other key elements that influence whether sa lmonellosis develops are the availability and population density of susceptible hos ts and the infective dose of th e pathogen. For these reasons, veterinary hospitals, breeding farms and other faci lities that may have a high density of horses are most vulnerable to the development of outbreaks attributable to salmonellae.55 In a previous study,32 three types of Salmonella -infected horses were described: carriers without fecal shedding, carriers with fecal shedding but withou t diarrhea and shedders with diarrhea. Carriers are horses infe cted with salmonellae; this gr oup of animals may not shed the organisms until exposed to stressful conditions.32,49 Carriers without fecal shedding are difficult to detect by bacterial culture; results may be negative because of dilution of the organism or because the organism is shed intermittently.28,32 Some carriers have been identified as subclinical or asymptomatic with fecal shedding but wit hout diarrhea. These have been considered a potential source of contamination not only to the environment but also to other animals.56 Shedders with diarrhea, just like asymptomatic shedders, contaminate the environment and can lead to Salmonella epidemics in young and stressed animals on broodmare farms and in veterinary hospitals.32 Several studies14,51,57 have described the prevalence of Salmonella shedding in diverse populations of horses, including those in breeding farms and slaughter houses where large numbers of congregated horses are vulnerable to infection and sh edding. In one study,57 the national prevalence of fecal shedding of salmonellae on equine operations in the United States was estimated to be 0.8% and prevalence of salmonellae in grain or other concentrate used for
20 horse feed was 0.4%. In another study51 that followed an outbreak of salmonellosis in foals, the prevalence of fecal shedding of S. enterica ser. Ohio was 27.8% and 35.1% in mares and foals respectively. Studies based on fecal cultu res reported prevalence between 0.8 and 20 %,53,57,58,59,60,61 and studies using PCR identificat ion reported 17 to 71.4% prevalence.28 Prevalence of Salmonella shedding in horses admitted to veterinary teaching hospitals as determined by fecal culture has been reported to range from 1.4 to 20 %3,53,60 and 0.8 to 3% on farms, in stables and other types of operations.57,61 Higher prevalence has been reported in studies where PCR was used as the diagnostic te chnique due to the higher sensitivity of this test.35,36 Variable prevalence patterns of Salmonella shedding have been observed, depending on the geographic region, season of the year, and the diagnostic method used to identify the organisms.8,35,62 It is difficult to compare results from the different studies due to the different populations of horses tested (g eneral hospital population, horse s with clinical signs of salmonellosis, horses presenting with colic or ot her gastrointestinal conditions), diagnostic techniques (bacterial culture, PCR), types of sa mples tested (feces, lymph nodes, rectal biopsies) and number of samples tested per horse (1 or more). Presenting complaint of colic,3,50,60 long distance transportation,49 change in diet while hospitalized,63 withholding feed,8 use of shared instruments such as nasogastric tubes or rectal thermometers,2,3,5 and antimicrobial therapy8,9,32,64 have been identified as risk factors associated with isolation of salmonellae from horses in several US veteri nary teaching hospitals. In a study3 at a veterinary teaching hospital following an outbreak of salmone llosis, horses admitted because of colic were 2.2 times as likely to have S. enterica ser. Saint-paul isolated as those admitted for other reasons and horses in which na sogastric tubes were passed were at 3.9 times greater risk of having S. enterica ser. Saint-paul isolated, compared with horses that were not intubated. In the
21 same study, horses receiving parenteral antibiotics were at 10.9 times greater risk of having S. enterica ser. Saint-paul isolated than were horses not receiving parenteral antibiotics. A second study9 was later carried out at the same hosp ital and horses that tested positive for Salmonella sp. other than the outbreak strain ( S. enterica ser. Saint-paul) were enrolled. In that study, horses admitted because of colic were 4.2 times as likely to have salmonellae isolated as those admitted for other reasons and horses in which nasogastri c tubes were passed were at 2.9 times greater risk of having salmonellae isolated, compared with horses that were not intubated. In addition, horses treated with antibiotics parenterally were at 6.4 times greater risk, and those treated with antibiotics orally and parenterally were at 40 times greater risk of developing salmonellosis, compared with horses not receiving su ch treatment. In a previous study,49 transportation was found to play a role in reactivating Salmonella infection when ponies that were orally inoculated with S. enterica ser. Typhimurium were exposed to st ress by long distance transportation. In another study63 at a veterinary teaching hospital, change in diet during hospitalization was found to be associated with fecal shedding of salmonellae. The season of the year has also been considered a risk factor due to the higher prevalence reported in summer months8,57,62 or in fall.32 Hot weather and high humidity associated with prolonged transportation can be important stre ss factors leading to a higher risk of Salmonella shedding.8,62 A high ambient temperature increases the likelihood that horses will shed Salmonella in their feces, and high ambient temperatur e is a risk factor for development of nosocomial Salmonella infections in horses.8,65 Not surprisingly, the percentage of horses shedding salmonellae in their feces is highest during the hot months of the year.65 Veterinary teaching hospitals are at a high risk of nosocomial Salmonella infection in horses because of exposure of the hospital population to a common source of salmonellae.
22 Carriers are constantly reintroduced, the environm ent is persistently contaminated, and a large population of vulnerable horses is at risk. In some instances, vete rinary teaching hospitals have been forced to close temporarily because of outbreaks of salmonellosis in horses. In 1981-1982, the University of California-Davis, Veteri nary Medicine Teaching Hospital (VMTH) experienced an outbreak of nos ocomial salmonellosis due to S. enterica ser. Saint-paul .3 The outbreak extended from June 1981 to April 1982 and resulted in severe disruption of hospital routine. In 1995, an outbreak of nosocomial salmonellosis due to S. enterica ser. Infantis at Colorado State University VMTH resulted in an estimated $500,000 in lost revenues and facility renovation.5 The original source of the organism causing this outbreak was not determined. In 1996,4,36 another outbreak of equine salmonellosis o ccurred at the Michigan State University VMTH; unique features of the outbreak included a high case fatality rate and zoonotic infection. Of the 18 horses associated with nosocomial in fection, 8 (44%) died while hospitalized. In addition, the S. enterica ser. Typhimurium isolate from a vete rinary student had an antimicrobial resistance pattern identical to the outbreak stra in. Pulse field gel electro phoresis patterns also suggested that the student was expos ed to the outbreak strain. In 2000,6 an outbreak of salmonellosis due to a multi-d rug resistant strain of S. enterica ser. Typhimurium occurred at Purdue University veterinary te aching hospital resulting in closure of the hos pital for a period of ten weeks. The index case was identified as a foal that presented with diarrhea in August 1999. In this outbreak, Salmonella isolates were characterized us ing antimicrobial susceptibility testing, PFGE and phage typing.6 In previous studies,4,5,6 environmental contamination has been identified as a source or reservoir of nosocomial Salmonella infections and plays a major role in spreading infections. In a study at a veterinary teaching hospital,4 persistence of S. enterica ser. Typhimurium in the
23 environment was identified as the source of nos ocomial infection for several horses. The pointsource of infection was a foal that had been hospitalized and S. enterica ser. Typhimurium was later isolated from hospital pers onnel, shared equipment, and stal ls. In that study, environmental samples were tested for Salmonella sp. using bacterial culture and PCR. To determine the similarity among S. enterica ser. Typhimurium isolates a nd the likelihood of nosocomial infection, antimicrobial suscepti bility and PFGE patterns were co mpared to the pattern for the isolate recovered from the point-source foal. In another study,6 environmental contamination was suggested to be the source of infection for other horses during an outbreak of salmonellosis in a teaching hospital. The primary case was a horse that presented with co lic and was shedding multi-drug resistant S. enterica ser. Typhimurium Environmental samples were tested for Salmonella sp. using bacter ial culture and PCR. S. enterica ser. Typhimurium was isolated from stall drains, surgery pads, forklift tires and the ambulatory garage floor. In this study, the similarities in serotyping, antibiogram, phagetyping and PFGE patterns were used to indicate that a common source strain of S. enterica ser. Typhimurium was responsible for environmental contamination. In a different study,5 environmental contamination contri buted to the wide spread nature of infection during an outbreak of S. enterica ser. Infantis. The original source of S. enterica ser. Infantis was not determined; howev er, during the outbreak, S. enterica ser. Infantis was isolated from hospital workers hands, rectal thermometers mice trapped in the hospital facility, and mats in stalls and recovery rooms. In this study, sero types of isolates were used to determine the similarity between collected isol ates and the outbreak strain. Three epidemiologic studies3,8,9 have attempted to investigate risk factors associated with nosocomial Salmonella infections in hospitalized horses. In a study3 at the University of
24 California VMTH following an outbreak of salmonellosis due to S. enterica ser. Saint-paul a presenting complaint of colic, nasogastric int ubation, and treatment w ith antibiotics were identified as risk factors for S. enterica ser. Saint-paul nosocomial infection. In that study, cases were classified as horses from which S. enterica ser. Saint-paul had been isolated and controls were horses from which fecal samples may or may not have been submitted for bacterial culture. Control horses were randomly selected from horses discharged from the hospital during the month in which a case developed. A careful review of selection of controls in this study revealed some validity issues that warrant discussion. Ther e is a possibility that asymptomatic shedders were enrolled as controls since no t all controls were tested for Salmonella sp. In addition, there is a likelihood that a control horse admitted at the be ginning of a particular month was matched to a case horse admitted at the end of the same month. Hospital conditions may vary during the month; for example, horses admitted in a particul ar month may not necessarily be exposed to the same number of horses shedding salmonellae. At the same hospital, a second study9 was conducted using a si milar design. A presenting complaint of colic, nasogastric intubation, and trea tment with antibiotics we re again identified as risk factors for nosocomial Salmonella infections. In this study, cas es were horses that tested positive for Salmonella sp. other than the outbreak strain ( S. enterica ser. Saint-paul). Control horses were not tested for Salmonella and consisted of two groups ; one group included the total equine population excluding Salmonella cases and the second group consisted of horses randomly selected from horses discharged from the hospital during the month in which a case developed. A careful review of se lection of controls in this study revealed the same validity issues as previously discussed.
25 In a third study,8 diagnosis of large colon impactions withholding feed, number of days fed bran mash, treatment with potassium penicil lin G, and mean daily ambient temperature were identified as risk factors for nosocomial Salmonella infections in hospitalized horses. In that study, nosocomial cases were defined as horses that tested positively to S. enterica ser. Krefeld or S. enterica ser. Typhimurium 96 hours after admission. In all three studies,3,8,9 it was not clear if time of exposure to primary cases was comp arable between nosocom ial cases and control horses. Prevention, Control, and Management of Salmonella in Horses Patient and environmental surveillance and enfo rcement of infection control protocols in hospitals are needed to prevent outbreaks of salmonellosis.6 A number of veterinary teaching hospitals have established surveillance and infe ction control programs that are directed at minimizing the exposure of susceptible hosts to in fective doses of salmonellae. At the UF LAH, as part of the hospitals survei llance and infection control program fecal samples are collected at admission from all horses that are presented with diarrhea alone, or fever and a leucopenia. Fecal samples are also collected at admission from fo als less than 6 months of age and accompanying mares. Thereafter, samples are collected ever y 48 hours (i.e., Monday, Tuesday, and Friday) until the patient is discharged. In addition, fecal samples are collected from horses that develop diarrhea or fever and leucopenia after admissi on. For some horses, additional samples (e.g., every 12 to 24 hours) may be collected at the discretion of the attending clinician. Routine monthly environmental sampling is carried out to evaluate cleaning and disinfection procedures. In addition, environmental sampling is conducted wh enever there is evidence that a nosocomial Salmonella infection has occurred in the hospital, and when a positive environmental sample is isolated.
26 At Michigan State University VMTH, as part of the infection contro l hospital procedures, fecal samples are collected on the day of admi ssion and at various times thereafter from all horses with evidence of gastrointestinal tract abnormalities.36 At the attending clinicians discretion, fecal samples are collected from horses w ithout clinical signs of gastrointestinal tract disease that are considered to be at risk for shedding salmonellae (e.g., neonatal foals with systemic disease, mare and foal pairs when only one of the pair has diarrhea, and horses treated with antimicrobials for long periods). Stalls that house horses with di arrhea or that shed salmonellae in their feces are sampled. Environm ental samples are also collected from other hospital areas that are co nsidered at risk for Salmonella contamination such as surgery rooms, anesthesia induction and recovery rooms.36 At Purdue University ve terinary teaching hospital,6 as part of establis hed hospital infection control procedures, fecal samples are collected fro m horses that present wi th diarrhea, or that develop diarrhea with leucopenia or fever after admission. Such hor ses are placed in isolation, and fecal samples collected daily beginning on the da y of isolation, until at least five samples are collected. The hospital environment is sampled, targeting various sites including surfaces of stalls. Several studies on salmonellosis in horses ha ve shown the use of general principles of isolation, disinfection, and traffic control to be effective in the management of Salmonella outbreaks in hospitals.2,4,5 The most commonly used protocol ha s been isolation of horses with clinical signs of salmonellosis and horses with a high risk of shedding salmonellae. Most veterinary hospitals maintain isolation units for this purpose, and horses are considered infectious and contagious until proven otherwise. A number of methods are employed to control and prevent microbial contamina tion during isolation including th e use of barrier precautions
27 such as examination gloves, protective coveralls and disposable boots when handling infected horses. Footbaths and footmats have been s hown to be effective in decreasing bacterial contamination in veterinary hospital environments.66,67 Effective cleaning and disinfection of contaminat ed environments has been one of the most important measures in preventing and controlling salmonellosis. Several guidelines for the use of different disinfectants and disinf ection techniques for ma terials, stalls and horse facilities have been published.36,68,69 Thorough cleaning of areas with fecal contamination such as stalls, water buckets or automatic watering apparatuses and dr ains are important measures that have been used.55 The use of bleach in the environment after initial cleaning procedur es is effective for additional elimination of environmental bacteria, as it has been shown to be the most effective product in eliminating detectable Salmonella sp. from hospital surfaces.36 Traffic control measures have been used to control and prevent the spread of Salmonella infection. Elements of traffic control include: th e designation of individuals to deal with sick animals only; the cleaning of health y horses stalls before cleaning th e stalls of sick animals, and control of movement be tween and through barns.5 Other management practices dictate restricting the number of personnel and attendant s entering the isolation stalls. Additional management practices recommended for controlling microbial contamination include use of separate instruments (thermomet ers, nasogastric tubes, twitches) and cleaning tools (grooming tools, manure carts forks, brooms, and shovels) for suspicious animals and their stalls.53 Horses returning from hospitalizatio n are prime candidates for shedding Salmonella or developing an acute infection with diarrhea.70 Returning horses may include horses that tested positive for Salmonella during hospitalization and those still showing signs of diarrhea. In horses
28 shedding Salmonella fecal shedding may persist for days or weeks. Diarrhea may be due to factors such as Salmonella infection, may follow surger y of the large intestine.71 At the farm, principles of isolation, hygiene, disinfection, and traffic control are applied as well. While it may not be practical to implement all of the following measures on every farm, the more closely they are adhered to, the lesser the risk of disease outbreaks. It is important to involve all your personnel, including the farm veterinarian, in developing a feasible plan. Upon isolation, carrying out the following measures, w ill limit environmental contamination: Wear gloves while handling the horse a nd wash hands thoroughly after handling the horse or anything that ha s been in contact with it. Alcohol-based hand sanitizers have been shown to effectively control contamination and can be readily placed stall side. Wear protective clothing (gowns, shoes) prior to entering the stall and either disposed afterwards or attach to the stall for the next person to use.72 This clothing should not be worn when handling othe r horses on the property and should be washed separately. Use different mucking tools in sick animals stalls; alternatively, stalls of healthy horses should be cleaned first, with the sick animals st alls cleaned last. The same rule applies for grooming tools and any other equipment used on sick animals.72 Feces from sick horse stalls should neve r be spread on fields, but disposed off by composting. This is done in a way that will not contaminate ground water and is fenced off from other horses. Remove feces as often as possible to minimize contamination of soil and noncleanable surfaces in the stall. Provide a plastic bag to allow for separa te disposal of your gloves and materials used in treatment of your horse. Disinfect stalls (where app licable), equipment, clothing and any towels used in a barn with sick horses with a chemical that is effective against salmonellae in the presence of organic matter and on all surfaces involved. Use foot baths where applicable. These are pl aced in front of the isolated stall and can be replenished as required. It should be noted however, that high organic loads reduce the effectivene ss of disinfectants.
29 Curtail traffic of people so that only speci fied individuals deal with sick animals and do not handle others. An alternative is to work with healthy animals first, then don protective clothing prior to working with sick animals. Veterinarians, farriers, and other personnel who have to travel be tween barns should visit the barn with sick animals last (unless there is an emergency). Ensure that vehicles going between and through barns to deliver bedding and feed also have the same traffic pattern of goi ng to non-affected barns first, leaving the sick animals barn last.72
30 CHAPTER 2 EPIDEMIOLOGY OF NOSOCOMIAL Salmonella INFECTIONS Salmonella shedding in hospitalized horses can lead to outbreaks of nosocomial salmonellosis if adequate surveillance and infectio n control procedures are not in place. Because the frequency of Salmonella shedding can be high in hospita lized horses, veterinary hospitals have instituted surveillance and infection control programs to reduce the risk of nosocomial infection.4,8,53,60 Identification of risk factor s associated with nosocomial Salmonella infection in hospitalized horses is important so that effective control and preventative measures can be instituted to reduce the risk of disease transm ission and potential outbreaks. Previous studies 3,8,9 have provided an epidemiologic framework for i nvestigation of risk f actors associated with nosocomial Salmonella infections in hospitaliz ed horses. In two studies3,9 conducted in the 1980s, horses treated with antimicrobials, horses in tubated with nasogastric tubes, and horses with a presenting complaint of colic were at high risk of nosocomial Salmonella infection. In a third study,8 risk of nosocomial Salmonella infection was greater in horses with large colon impactions, as well as in horses treated w ith potassium penicillin G. In these studies3,8,9 however, research methods used were inconsistent and ha d several limitations. For example, in the first two studies,3,9 some control horses were tested for Salmonella but others were not. In addition, in all three studies,3,8,9 it was not clear if time of exposure to primary cases was comparable between nosocomial cases and control horses. Abdominal surgery and high casel oad are recognized as importan t contributors expected to increase the risk of nosocomial Salmonella infection in hospitalized horses.8,60 Previous studies, however, failed to identify abdominal surgery or high caseload as predisposing risk factors for nosocomial Salmonella infection.3,8,9 Currently, testing of horse s for early detection of
31 Salmonella shedding during hospitalizati on is a common practice in ve terinary hospitals. This new scenario provides an opportunity to re-asse ss the epidemiological aspects of nosocomial Salmonella infections in hospitalized horses. For ex ample, it is possible that previous studies3,9 failed to identify abdominal surgery as a risk f actor because many surgical inpatients were subclinically infected with Salmonella and were not detected becaus e they were not tested; this subpopulation of horses could have b een misclassified as susceptibl e controls. The objectives of this study were: (i) to examine the relations hip between abdominal surgery and nosocomial Salmonella infections; and (ii) to examine the rela tionship between high caseload in combination with abdominal surgery and nosocomial Salmonella infections in horses hospitalized with signs of gastrointestinal tract disease. Materials and Methods Study Population All equine inpatients admitted to the Univer sity of Florida Veterinary Medical Center (UF VMC) between January 1, 2002 and December 31, 2006 with signs of gastrointestinal tract disease were eligible for incl usion in the study. This subpopulati on of horses was targeted for early detection of Salmonella shedding as part of the UF VMC surveillance and infection control program. Fecal samples were collected with in 12 hours after admission and submitted for bacterial culture; thereafter, additional sa mples were collected every 48 hours (i.e., Monday, Wednesday and Friday) until the patient was discharg ed from the hospital. Horses that had tissue samples submitted for diagnosis of Salmonella sp. such as blood, intes tines, joint fluids and abscess were excluded, as well as horses with incomplete data (missing serogroup, unavailable/untypable serotype, an timicrobial susceptibility patte rns, admission date, discharge date, or sampling date). In a ddition, horses hospitalized for < 72 hours or with < 2 fecal samples collected for bacterial culture were excluded.
32 Primary cases Horses that tested positively for Salmonella on fecal samples collected at admission were classified as primary cases. In addition, horses w ith clinical signs of salmonellosis at admission (i.e., diarrhea, fever and leucopenia) that tested positively for Salmonella sp. on fecal samples collected later during ho spitalization (i.e., on second, third samp les, etc) but with no evidence of nosocomial infection were classified as primary ca ses. To rule out the possibility of nosocomial infection, surveillance data (i.e., Salmonella serotype, antimicrobial susceptibility patterns, stall location, time overlap (admission and discharg e dates) as well as results of monthly environmental samplings were reviewed) to classi fy this group of horses as primary cases. Nosocomial cases Hospital surveillance data (i.e., Salmonella serotype, antimicrobial susceptibility patterns, time and space) were reviewed to classify horse s as nosocomial cases. In this study, nosocomial cases were horses that tested negatively for Salmonella upon admission and positively 48 hours after hospitalization. The source of nosocomial infection was a primary case that tested positively to Salmonella upon admission or during hospitaliza tion and that shared the same serotype and antimicrobial susceptibility patterns as the nosocomial case; in addition, there was a time overlap between admission and discharge among the primary case and the nosocomial case during hospitalization. Another source of nosocomial infection was environmental contamination. Horses that tested negatively for Salmonella sp. upon admission but positively thereafter, and that shared the same serotype and antimicrobial susceptibility patterns as a positive environmental sample collected during the time of hospitalization were also classified as nosocomial cases.
33 Control horses Horses that tested negatively to Salmonella sp. on fecal samples collected within 12 hours of admission and on all later samples collect ed every 48 hours during hospitalization were classified as controls. Study Design This study was designed as a matched case-control study. To accomplish the first objective, 1 to 4 control horses were matched to each nosocomial case by admission date of the primary case 2 days (or by collection date of the positive environmental sample 2 days). Matching by admission date of the primary case 1 day (or by collection date of the positive environmental sample 1 day) was considered, bu t the number of eligible controls did not reach a ratio of controls to cases of 1:1. In addi tion, matching by admission date of the nosocomial case was considered, but an examination of medi cal records revealed that many control horses were never exposed to the primary case (e.g., hos pital duration was not the same in nosocomial cases and control horses). Twenty eight hor ses were classified as nosocomial cases; environmental contamination was considered th e source of infection in 5 cases. Twelve nosocomial cases were excluded due to lack of matching control horses or missing medical records. Final enrollment in objective 1 in cluded 16 nosocomial cases and 35 controls. Frequency of abdominal surgery, high caseload and other investigated exposure factors was compared between horses classified as cases and controls. To acc omplish the second objective, 4 control horses were matched to each nosocomial case by admission year. Elig ible control horses were numbered 1 through n. A computer programa was used to select random numbers (horses) needed in this study (n + 4). Final enrollment in objective 2 in cluded 21 cases and 84 controls.
34 Collection of Fecal Samples Fecal samples were collected from horses included in the study by hospital technicians within 12 hours after admission and submitted for bacterial culture of Salmonella sp. Thereafter, additional samples were collect ed every 48 hours as part of the hospitals su rveillance and infection control program until the patient was discharged from the hospital. Fecal samples that were collected after regular busin ess hours were refrigerated at 4 C prior to submission. In some horses, additional fecal samples were collected (e.g., every 12 to 24 hours) at the discretion of the attending clinician. If a foal was admitted to the hospital with its mare, fecal samples were collected from both the foal and mare. A Ho spital Infection Control Officer (ABE) was responsible for overseeing the collection of fecal samples, microbiologic procedures, and collection and analysis of epidemiologic data throughout the study period. Microbiological Procedures for Detection of Salmonella Organisms Bacterial culture of fecal samples for detection of Salmonella organisms was performed at the UF VMC Clinical Microbiology Laborator y. For selective enrichment, 2 to 5g of fresh feces was placed in 10ml of selenite broth,b and the broth was incubated at 37 C, and 5% CO2 for 24 hours. The following day, the selenite broth wa s subcultured on Hektoen enteric agar plates.c Plates were incubated at 37 C for 24 hours. Non-lactose fermenting, H2S producing colonies were selected and isolated. These colonies were then inoculated on urea agar and lysine iron agar slants, and incubated at 37 C for 24 hours. __________ a Research Randomizer. www.randomizer.org b Selenite broth, Hardy Diagnostics, Santa Maria, California c Hektoen enteric agar, Hardy Dia gnostics, Santa Maria, California
35 Identification of urease-negative and H2S producing organisms was established through use of a commercially available identification system.d Serogroup of Salmonella isolates was determined by means of agglutination; polyvalent (A-I & Vi) and groupspecific (A-E) Salmonella O antiserae were used. Salmonella isolates were tested for antimicr obial susceptibility by means of MIC method with commercially prepared plates.f Serotyping of Salmonella isolates was performed at the USDA National Veterinary Services Laboratories in Ames, Iowa. Data Collection A structured questionnaire was developed fo r collection of epidemiologic data including: age; gender; breed; admission date; discharge date; duration of hospitalizati on (days); caseload at admission; caseload during the 5-day period that included 4 days prior to and the day of Admission; number of horses shedding Salmonella at admission; number of horses shedding Salmonella during the 5-day period that included 4 days prior to and the day of admission; number of fecal samples collected for Salmonella culture; presenting complaint (e.g., colic, diarrhea, fever, anorexia); diagnosis (e.g., la rge colon impaction, enteritis, small colon impaction, rotavirus diarrhea); clinical proc edures during hospitaliz ation including, surgery, nasogastric intubation, rectal palpation, and abdominocentesis; use of anti-inflammatories, sedatives/analgesics and antimicrobials duri ng hospitalization; numbe r of patient-personnel contacts during the first 3 days after admission; and Salmonella serotype isolated. __________ d API-20E system, Biomerieux Sa, Marcy IEtoile, France e Bacto-Salmonella antisera, DIFCO Laboratories, Detroit, Michigan f NMIC30 plate, MicroScan Dade Behring, Siemens Medical Solutions Diagnostics, Deerfield, Illinois
36 Statistical Analysis Conditional logistic regression was used to model the odds of being a case as a function of investigated risk factors in this study. Initial screening of potential risk factors for nosocomial infection was performed by use of univariable conditional logistic regression or the Wilcoxon rank sum test (for continuous variables such as caseload, number of horses shedding Salmonella number of fecal samples collected, and duration of hospitalization). Initiall y, variables associated ( P 0.20) with the outcome of intere st (nosocomial infection) were entered into the model, and a forward stepwise approach was us ed to identify which variables we re associated with infection using two-sided P -values to-enter and to-remove of 0.05 and 0.10, respectively. Duration of hospitalization and number of fecal samples subm itted for bacterial culture were included as required variables in the final model (objective 1) because they can influence the probability of detecting Salmonella Similarly, number of horses sheddi ng at admission, number of fecal samples submitted, as well as number of patient care-personnel contacts were included as required variables in the final model (objective 2). The interaction terms of: abdominal surgery and antimicrobial use; abdominal surgery and pa tient care-personnel contacts (objective 1); and caseload and abdominal surgery ( objective 2) were included in the final model and tested for significance ( P 0.05). Fit of the model to the data was assessed by visual examination of residual plots (standardized delt a-beta values versus observati on number and delta-beta versus fitted values). Case-control sets that had horse s with extreme delta-beta values and low fitted values were excluded from the analysis to evalua te their influence on estimated odds ratios (OR). In the final model, adjusted OR and 95% conf idence intervals (CI) were reported. The OR was used as an epidemiologic measure of associati on between a risk factor and risk of nosocomial Salmonella infection. Thus, if a particular factor was not associated w ith risk of infection, the OR was 1. The greater the departure of the OR from 1 (either larger or smaller), the stronger the
37 association was between the f actor (e.g., abdominal surgery) and risk of infection. Under objective 2, median number of horses shedding Salmonella at admission was compared between times of high and low caseload by using the median test. Results Objective 1 Sixteen horses were classified as nosocomial cases and 35 as controls. S. enterica ser Newport (5/16) was the most commonly isolat ed serotype among nosocomial cases, followed by Reading (3), Anatum (2), Saint-paul (1), Meleagridis (1), Java (1), Litchfield (1), Javiana (1) and Braenderup (1) (Table 2-1). Median number of hospital days from time of admission to when the first positive sample was collected was 3 days. Median number of days exposed to the primary case was 4 days in cases and 3 days in controls ( P = 0.13) (Table 2-3). Median number of fecal samples collected was 4 samples in cases and 3 samples in controls ( P = 0.07). Median number of days spent in the hosp ital was 7 days in cases a nd 5 days in controls ( P = 0.16). Median number of patient care-personnel contacts during the first 3 days after admission was 29 in cases and 28 in controls ( P = 0.61). Median number of patient ca re-personnel contacts during the first 3 days after admission was 32 in horses that underw ent abdominal surgery a nd 26 in horses that did not ( P = 0.04). Eight of the 16 nosocomial cases unde rwent surgery, and 6 of the 8 surgical cases did not develop diarrh ea during hospitalization. Using univariable conditional logistic regression analysis, age, withholding of feed, abdominal surgery, use of sedatives, and use of antimicrobials had P values 0.20 (Table 2-5), and were included in the multivariable analysis The variables for duration of hospitalization, number of fecal samples collected and number of patient care-personnel c ontacts during the first 3 days after admission were forced into the fi nal model. In the multivariable analysis, the variable for abdominal surgery was retained in the model ( P 0.10) (model 1) (Table 2-7).
38 Addition of the two-way interaction terms: a bdominal surgery and use of antimicrobials; abdominal surgery and patient care-personnel cont acts did not contribute to the final model, and these terms were removed from the model. Hors es that underwent surgery were more likely to become infected with Salmonella during hospitalization than horse s that did not undergo surgery, after controlling for duration of hospitalization, number of feca l samples collected, and number of patient care-personnel contacts during the fi rst 3 days after admission (OR = 8.2; 95% CI = 1.11, 60.24; P = 0.03). A visual examination of residuals revealed that the de lta-beta values for the variable of abdominal surgery were not extr eme (i.e., not larger than 1), supporting overall goodness-of-fit. Analysis of residuals (set of cases and controls w ith the largest delta-beta value and the lowest fitted value) indicated the existence of infl uential observations. However, abdominal surgery was still signific ant after removal of the set of cases and controls identified as the most influential observations (OR = 33.75; 95% CI = 0.89, 1267.85; P = 0.05). Because all horses that underwent abdominal surgery were treated with antimicrobials, a second model examined the risk of infection in this group of horses, as well as in horses that did not have surgery but were treated with antimicrobials, co mpared to horses that did not have either exposure (Table 2-7). In this second model, the odds of infection were 15 times higher in horses that underwent abdominal surgery and were treated with antimicrobials, co mpared to horses that did not have either exposure (OR = 15.35; 95% CI = 1.35, 173.93; P = 0.02). The odds of infection were 5 times higher in horses that did no t have abdominal surgery but were treated with antimicrobials; but this association was not significant (OR = 5.04; 95% CI = 0.28, 89.11; P = 0.26). Finally, we examined the potential intera ction effect between abdominal surgery and number of patient care-personnel contacts, but this effect was not relevant or significant (OR = 0.96; 95% CI = 0.78, 1.17; P = 0.71).
39 Objective 2 Twenty one horses were clas sified as nosocomial cases a nd 84 as controls. Median number of horses shedding Salmonella at admission was not different during times of high caseload (26 to 54 inpatients) (2 shedders) or lo w caseload (15 to 25 inpa tients) (2 shedders) ( P = 0.65). We failed to identify high caseload alone or in combination with surgery as a risk factor for nosocomial Salmonella infection in hospitalized horses afte r controlling for number of horses shedding at admission and number of samples collected (Model 3; OR = 2.81; 95% CI = 0.31, 25.32; P = 0.35) (Table 2-8). Abdominal surgery was c onfirmed as a risk factor associated with nosocomial Salmonella infection (Model 4; OR = 4.98; 95% CI = 1.55, 15.95; P < 0.01). Discussion This study provides epidemiologic evidence that equine inpatients with gastrointestinal disease that undergo abdominal surger y are at high risk of nosocomial Salmonella infection. The study results, however, do not support the hypothesis that high casel oad alone or in combination with abdominal surgery is a predisposing factor for nosocomial Salmonella infection in hospitalized horses at the UF VMC. The strengths of this study in clude the use of nosocomial cas es and susceptible (control) horses exposed to primary cases, the use of controls defined as Salmonella negative horses as determined by bacterial culture, th e number of fecal samples collect ed and tested, as well as the duration of hospitalization which was comparable between cases and controls. The small number of cases is a study limitation, which affected the precision of the risk estimates of nosocomial infections associated with abdominal surgery. The small sample size also affected our ability to adequately assess the potential interaction effect s of high caseload and abdominal surgery on risk of nosocomial Salmonella infections. Finally, the study popula tion was restricted to equine inpatients presenting with signs of gastrointestin al disease and mare/foal pairs with foals less
40 than 6 months of age. Thus, our study result s cannot be generalized to the entire equine population at the UF VMC. In the present study, horses that underwent a bdominal surgery were more likely to become infected with Salmonella sp. during hospitalization. To our knowledge, this is the first observational study that provides ev idence that abdominal surgery is a risk factor for nosocomial Salmonella infection in hospitalized ho rses. Three previous studies3,8,9 failed to identify abdominal surgery as a risk factor for nosocomial infection. In 2 studies,3,9 control horses were not necessarily cultured for Salmonella organisms. and therefore some horses classified as controls might have been primary or nosocom ial cases. In our study, 8 of the 16 nosocomial cases underwent surgery, and 6 of the 8 surg ical cases did not develop diarrhea during hospitalization. In the aforementioned 2 studies,3,9 this subpopulation of horses could have been classified as controls, making the identification of surgery as a risk factor difficult. Finally, in the third study,8 similar to the previous 2 studies,3,9 it was not clear if time of exposure to primary cases was comparable between nosocomial cases and susceptible (control) horses. Thus, the interpretation of the epidemio logic analysis of that study,8 especially for abdominal surgery as a risk factor, was difficult. Horses that undergo abdominal surger y suffer substantial amounts of stress.32 The surgical procedure in itself causes stress to th e patient; in addition, the large colon or cecum may be emptied and lavaged, feed may be withhe ld or changed, antimicrobial drugs may be administered, and various de grees of ileus may develop.8,32,49,73,74,75, These events further increase stress in surgical patients and also alter the function of normal gastrointestinal microflora.8,32,49 Major surgery is associated with se vere alterations of the host-defense mechanisms.76,77,78 In humans who underwent partial gastrectomy, surgic al stress rapidly
41 depressed monocyte mCD14 and HLA-DR expressi on in comparison with preanesthesia levels.76 CD14 plays a key role in transmitting LPS signals intracellularly and ul timately activating TNFproduction.77 Loss of cell surface HLA-DR has been suggested to reduce the antigenpresenting capacity of monocytes, resu lting in impaired T cell stimulation. These alterations suppress the innate immune sy stem during the perioperative and postoperative periods.76,79 Other studies on surgical stress in humans have de monstrated that an impaired immune system increased the risk of developing systemic inflammatory response syndrome, sepsis, and multiple organ failure.76,80,81 Studies in mice have illustrated that surgical stress due to laparatomy produces significant impairment of cell-mediated immunity.10,78 In view of the effects of abdominal surgery on the immune system in humans and mice, it is possible that surgical stress similarly suppresses both the innate and adaptive immune system of equine surgical patients. As a result, this may increase their susceptibility to nosocomial Salmonella infections. In this study, assessment of the potential interaction effe ct of abdominal surgery and antimicrobial use on nosocomial infection was difficu lt. The assessment of this interaction effect requires the risk estimation of 4 groups of horses: (i) horses unexposed to surgery and antimicrobials; (ii) horses exposed to surgery on ly; (iii) horses exposed to antimicrobials only; and (iv) horses exposed to both factors. In our study, all horses that underwent surgery were treated with antimicrobials. The expected joint effect of surgery and antimicrobials was difficult to estimate because our study population did not include the group of horses exposed to surgery only. We examined a second model using an indicat or variable with 3 categories to estimate the risk of infection in horses exposed to antimic robials and horses exposed to both surgery and antimicrobials, compared to horses not exposed to both factors. The odds of infection were 5 and 15 times higher in horses exposed to antimicr obials and horses exposed to both surgery and
42 antimicrobials, respectively, compar ed to horses not exposed to both factors. In this study, if the odds of infection due to surgery alone were 2, then the estimated odds ratio of 15 due to exposure to both factors would be an indication that there was an interaction effect; implying, the joint effect of surgery and antimicrobials on infection exceeded that by the additive or multiplicative effect of these two factors. The association between antimicrobial use and nosocomial Salmonella infections in hospitalized horses ha s been established in previous studies.3,8,9 The normal intestinal flora is an import ant line of defense ag ainst colonization by potentially pathogenic bact eria that are ingested.82 In addition, it averts overgrowth of already present opportunistic microorganisms.83,84 Normal intestinal flora prevent the attachment and multiplication of pathogenic micro-organisms on mucosal surfaces and their invasion into epithelial cells and circulation.82 Administration of antimicrobial agents, therapeutically or as prophylaxis, causes disturbances in the ecological balance between the host and the normal microflora.85 Antimicrobials eliminate intestinal flora that are antagonistic to Salmonella ; this may explain the association between expos ure to antimicrobial drugs and nosocomial Salmonella infections in horses that are susceptib le and previously tested negative for Salmonella Adverse changes in the functi on of intestinal tract flora as a result of antimicrobial use appear to make surgical patients more susceptible to colonization with Salmonella sp. Although frequency of patient care-personne l contacts during the first 3 days after admission was higher in horses that underwent abdominal surgery th an in horses that did not, we failed to identify an interaction effect between abdominal surgery and nu mber of patient carepersonnel contacts on risk of nos ocomial infection. This finding indicates that horses that undergo abdominal surgery are at high risk of nos ocomial infection, and that this risk is not
43 increased nor confounded by a high frequency of patient care-personnel contacts observed in surgical cases. We failed to identify high caseload alone or in combination with abdominal surgery as a risk factor for nosocomial Salmonella infection in hospitalized horses. This result can be explained in part by the fact th at the number of horses shed ding salmonellae at the time of admission during times of high or low caseload was not different. Another explanation is that the UF VMC surveillance and infec tion control program is well accepted by hospital personnel, and the level of compliance is high. It is possible th at high caseload may have an effect on risk of nosocomial infection when the number of shedders in the hospital is hi gh and the size of the technician task force is not adequate. In summary, results from this epidemiol ogic study indicate that abdominal surgery increases the risk of nosocomial Salmonella infections in hospitali zed horses with gastrointestinal disease. In light of this finding, horses that unde rgo abdominal surgery require enhanced infection control and preven tative care. Risk of nosocomial Salmonella infections may be reduced by implementation of preventative measures immediately following surgery, such as ward isolation (i.e., use of gloves, gowns, plas tic boots and footbaths). Minimizing patient carecontact of this highly susceptib le group of inpatients may also reduce the risk of nosocomial infections. In addition, due to the adverse effect s of antimicrobials on the intestinal microflora, justification of antimicrobial us e in surgical patients should be carefully evaluated prior to prescription.
44 Table 2-1. Characterization of 16 horses cla ssified as nosocomial cases in objective 1 Variable Cases (n) Diarrhea No Yes Fever No Yes Leucopenia No Yes Not known Salmonella serogroup B C1 C2 D E Not A-E Salmonella serotype Newport Saint-paul Meleagridis Java Anatum Litchfield Javiana Braenderup Reading Fecal sample number that tested positively first 2 3 4 5 6 Day of week when first test positive was collected Monday Tuesday Wednesday Thursday Friday Number of hospital days when first test positive sample was collected 10 6 10 6 8 6 2 5 1 6 1 3 0 5 1 1 1 2 1 1 1 3 5 7 2 1 1 5 1 3 1 6 3 (2, 4) Data reported as median (1st and 3rd quartiles)
45 Table 2-2. Characterization of 21 horses cla ssified as nosocomial cases in objective 2 Variable Cases (n) Diarrhea No Yes Not Known Fever No Yes Not Known Leucopenia No Yes Not known Salmonella serogroup B C1 C2 D E Salmonella serotype Newport Saint Paul Meleagridis Java Anatum Litchfield Javiana Braenderup Reading Miami Rubislaw Fecal sample number that tested positively first 2 3 4 5 6 Day of week when first test positive was collected Monday Tuesday Wednesday Thursday Friday Saturday Number of hospital days when first test positive sample was collected 14 6 1 14 6 1 11 7 3 6 1 7 2 5 6 1 2 1 3 1 1 1 3 1 1 8 8 3 1 1 6 1 4 1 8 1 3 (2, 4) Data reported as median (1st, 3rd quartiles)
46 Table 2-3. Objective 1 Caseloa d, number of horses shedding Salmonella at admission, fecal samples collected, and hospital durat ion in case and control horses Variable Cases n = 16 Controls n = 35 P value Caseload at admission Caseload at admission plus 4 days before admission Number of horses shedding Salmonella at admission Number of horses shedding Salmonella at admission plus 4 days before admission Mean number of horses shedding Salmonella at admission plus 4 days before admission Time of exposure to the primary case (days) Number of fecal samples collected Duration of hospitalization (days) Number of patient care-personnel contacts during the first 3 days after admission 24 (19, 28) 124 (107, 140) 3 (2, 4) 4 (3, 6) 0.7 (0.6, 1.4) 4 (2, 7) 4 (3, 6) 7 (4, 11) 29 (24, 39) 29 (23, 30) 129 (113, 143) 4 (3, 5) 4 (3, 7) 0.8 (0.6, 1.4) 3 (2, 4) 3 (2, 4) 5 (4, 8) 28 (18, 40) 0.19 0.85 0.15 0.38 0.42 0.13 0.07 0.16 0.61 Data reported as median (1st and 3rd quartiles) Table 2-4. Objective 2 Caseloa d, number of horses shedding Salmonella at admission, fecal samples collected, and hospital durat ion in case and control horses Variable Cases n = 21 Controls n = 84 P value Caseload at admission Caseload at admission plus 4 days before admission Number of horses shedding Salmonella at admission Number of horses shedding Salmonella at admission plus 4 days before admission Mean number of horses shedding Salmonella at admission plus 4 days before admission Number of fecal samples collected Duration of hospitalization (days) 25 (19, 28.5) 123 (108, 139.5) 3 (2, 4) 3 (2, 6) 0.6 (0.4, 1.3) 3 (2, 6) 7 (4, 9) 25 (22, 30) 126 (109, 144) 2 (1, 3) 2 (1, 4) 0.4 (0.2, 0.8) 3 (2, 4) 5 (4, 7) 0.29 0.45 0.02 0.04 0.03 0.05 0.30 Data reported as median (1st & 3rd quartiles)
47 Table 2-5. Objective 1 Frequenc y distribution of host factors, hospital procedures, crude odds ratios (OR), and 95% confidence intervals (CI) of investigated risk factors among case and control horses Variable Cases n = 16 Controls n = 35 OR 95% CI P value Age Foal Adult Sex Female Male Gelding Duration of hospitalization 3 to 5 days 6 + days Presenting complaint Colic Fever Diarrhea Anorexia Pneumonia Others* History of antimicrobial use prior to admission No Yes Housing Isolation barn ICU Barn A Barn B Withholding feed No Yes Surgery No Yes Anti-inflammatory use No Yes Sedative/analgesic use No Yes Not known 5 11 6 7 3 5 11 12 1 2 0 0 1 14 2 3 2 11 0 4 12 8 8 3 13 3 12 1 4 31 16 6 13 22 13 21 2 0 1 2 8 23 6 2 8 24 1 17 18 29 6 10 25 17 18 0 1.00 0.24 1.00 2.74 0.50 1.00 4.43 1.00 0.78 ND ND ND 0.20 1.00 0.62 1.00 0.32 0.49 ND 1.00 2.54 1.00 4.29 1.00 1.19 1.00 4.50 ND Reference 0.04, 1.36 Reference 0.67, 11.13 0.08, 2.95 Reference 0.88, 22.36 Reference 0.06, 8.87 ND ND ND 0.02, 1.87 Reference 0.11, 3.45 Reference 0.03, 3.23 0.09, 2.63 ND Reference 0.67, 9.5 Reference 1.07, 17.12 Reference 0.22, 6.44 Reference 0.86, 23.43 ND NA 0.10 NA 0.15 0.45 NA 0.07 NA 0.84 ND ND ND 0.15 NA 0.59 NA 0.34 0.40 ND NA 0.16 NA 0.03 NA 0.83 NA 0.07 ND Others include corneal ulcer, uveitis, umblical hernia, and normal mare/foal with sick foal/mare
48 Table 2-5. Continued Variable Cases n = 16 Controls n = 35 OR 95% CI P value Nasogastric tubing Normal Abnormal Not done Rectal exam Normal Abnormal Not done Abdominocentesis Normal Abnormal Not done Antimicrobial use No Yes Diarrhea No Yes Fever No Yes Leucopenia No Yes Not known Number of patient personnel contacts during first 3 days 7 to 28 contacts 29 + contacts Diagnosis Small colon impaction Large colon impaction Enteritis Ileal impaction Large colon displacement Others* 2 6 8 2 7 7 1 4 11 4 12 10 6 10 6 8 6 2 7 9 0 0 5 0 3 8 8 9 18 7 21 7 8 3 24 22 13 26 9 26 9 21 4 10 19 16 3 5 0 5 0 21 1.00 1.88 1.57 1.00 0.8 3.78 1.00 9.61 3.73 1.00 5.92 1.00 1.60 1.00 1.79 1.00 3.92 0.55 1.00 1.24 ND ND ND ND ND ND Reference 0.31, 11.41 0.28, 8.65 Reference 0.08, 7.25 0.42, 33.94 Reference 0.5, 182.79 0.34, 41.09 Reference 1.22, 28.64 Reference 0.43, 5.87 Reference 0.47, 6.74 Reference 0.70, 21.97 0.09, 3.22 Reference 0.31, 4.82 ND ND ND ND ND ND NA 0.48 0.59 NA 0.84 0.23 NA 0.13 0.28 NA 0.02 NA 0.47 NA 0.38 NA 0.11 0.50 NA 0.75 ND ND ND ND ND ND Others include ileal strangulation, rotavirus diarrhea, rhodococcus pneumonia, cecal impaction, corneal ulcer, nephrosplenic entrap ment, ileocecal intususcep tion, enterolithiasis, gastric ulceration, umbilical hernia, colon torsi on, peritonitis, and normal mare/foal with sick foal/mare
49 Table 2-6. Objective 2 Frequenc y distribution of host factors, hospital procedures, crude odds ratios (OR), and 95% confidence intervals (CI) of investigated risk factors among case and control horses Variable Cases n = 21 Control n = 84 OR 95% CI P value Age Foal Adult Sex Female Male Gelding Duration of hospitalization (d) 1 to 5 > 5 Presenting complaint Fever Colic Diarrhea Anorexia Pneumonia Others* History of antimicrobial use prior to admission No Yes Not known Housing Isolation barn ICU Barn A Barn B Withholding feed No Yes Surgery No Yes Anti-inflammatory use No Yes Sedative/analgesic use No Yes Not known 6 15 7 7 7 9 12 1 17 2 0 0 1 8 3 0 4 2 15 0 5 16 12 9 5 16 7 13 1 18 66 39 19 26 48 36 3 46 9 3 2 17 60 15 9 14 12 56 2 34 50 68 16 26 58 45 37 2 1.00 0.66 1.00 2.26 1.51 1.00 1.72 1.00 3.90 1.89 ND ND 0.51 1.00 0.64 ND 1.00 0.64 1.04 ND 1.00 2.30 1.00 3.05 1.00 1.43 1.00 2.41 3.33 Reference 0.21, 2.06 Reference 0.62, 8.24 0.47, 4.78 Reference 0.67, 4.38 Reference 0.43, 35.49 0.16, 22.08 ND ND 0.02, 9.05 Reference 0.16, 2.55 ND Reference 0.09, 4.42 0.30, 3.65 ND Reference 0.72, 7.35 Reference 1.12, 8.3 Reference 0.47, 4.34 Reference 0.82, 7.01 0.26, 41.41 NA 0.47 NA 0.21 0.47 NA 0.25 NA 0.22 0.61 ND ND 0.65 NA 0.52 ND NA 0.65 0.93 ND NA 0.15 NA 0.02 NA 0.51 NA 0.10 0.34 Others include weight loss, abscess, corneal ulcer, and normal mare/foal with sick foal/mare
50 Table 2-6. Continued Variable Cases n = 21 Control n = 84 OR 95% CI P value Nasogastric tubing Normal Abnormal Not done Rectal exam Normal Abnormal Not done Abdominocentesis Normal Abnormal Not known Antimicrobial use No Yes Diagnosis Enteritis Small colon impaction Large colon impaction Ileal impaction Large colon displacement Others* 2 7 12 3 10 8 4 4 13 7 14 5 0 1 0 3 12 13 22 49 21 38 25 17 7 60 41 43 8 3 14 3 4 48 1.00 2.13 1.62 1.00 1.93 2.41 1.00 2.41 0.92 1.00 1.81 1.00 ND 0.19 ND 2.99 0.77 Reference 0.37, 12.23 0.32, 8.25 Reference 0.46, 8.07 0.52, 11.18 Reference 0.47, 12.29 0.27, 3.10 Reference 0.68, 4.76 Reference ND 0.02, 1.78 ND 0.38, 23.62 0.23, 2.55 NA 0.39 0.55 NA 0.36 0.25 NA 0.28 0.89 NA 0.22 NA ND 0.14 ND 0.29 0.67 Others include volvulus, strangulating lipoma, hemoperitoneum, lactose intolerance, failure of passive transfer, infiltrative bowel disease, normal mare/foal with sick foal/mare, and colic of unknown cause
51 Table 2-7. Objective 1 Multivariable conditiona l logistic regression model for nosocomial Salmonella infections in hospitalized horses Variable Adjusted OR* 95% CI* P value Model 1 Surgery No Yes Duration of hospitalization 3 to 5 days 6+ days Number of fecal samples collected Number of patient personnel contacts during first 3 days after admission Model 2 Surgery (no) and antimicrobial use (no) Surgery (no) and antimicrobial use (yes) Surgery (yes) and antimicrobial use (yes) Duration of hospitalization 3 to 5 days 6+ days Number of fecal samples collected Number of patient personnel contacts during first 3 days after admission 1.00 8.20 1.00 1.05 1.57 0.99 1.00 5.04 15.35 1.00 1.06 1.35 0.97 Reference 1.11, 60.24 Reference 0.10, 10.73 0.82, 3.00 0.94, 1.03 Reference 0.28, 89.11 1.35, 73.93 Reference 0.11, 9.59 0.69, 2.65 0.91, 1.02 NA 0.03 NA 0.96 0.17 0.68 NA 0.26 0.02 NA 0.95 0.37 0.32 OR = odds ratio; 95% CI = 95% confidence interval Table 2-8. Objective 2 Multivariable conditiona l logistic regression model for nosocomial Salmonella infections in hospitalized horses* Variable Adjusted OR** 95% CI** P value Model 3 Caseload Low (15-25 inpatients) High (26-54 inpatients) Surgery No Yes Caseload x Surgery Model 4 Caseload Low High Surgery No Yes 1.00 0.56 1.00 2.95 2.81 1.00 0.84 1.00 4.98 Reference 1.13, 2.34 Reference 0.60, 14.41 0.31, 25.32 Reference 0.26, 2.67 Reference 1.55, 15.95 NA 0.42 NA 0.18 0.35 NA 0.77 NA < 0.01 Adjusted for number of horses shedding at admission and number of samples collected ** OR = odds ratio; 95% CI = 95% confidence interval
52 APPENDIX QUESTIONAIRE FOR SALMONELLOSIS STUDY 1. General information Medical record number: Age of primary patient: Foal Adult Gender of primary patient: Female Male Gelding Foal and mare admitted together: Yes No Not known Breed: Admission date: Discharge date: Duration of hospitalization: 2. Presenting complaint: Colic Fever Diarrhea Anorexia Pneumonia Others 3. Diagnosis: Small colon impaction Large colon impaction Enteritis Ileal impaction Large co lon displacement Others 4. History of antimicrobial use prior to admission: No Yes Not known If yes, antimicrobial used: Gentamicin Penicillin G Procaine Pen G Potassium Other 5. Housing upon admission: Isolation Unit ICU Barn A Barn B 6. Clinical parameters of Salmonellosis during hospitalization: Diarrhea Fever Leucopenia 7. Withholding of feed: No Yes Not known 8. Surgical Procedures: No Yes If yes, type of procedure: Exploratory surgery Enterotomy Resection Other 9. Treatment/procedures during hospitalization: Anti-inflammatory use: No Yes Not known If yes, Banamine DMSO Phenylbutazone Other Sedative/analgesic use: No Yes Not known If yes, Acepromazine Butorphanol Detomidine Xylazine Other Nasogastric tubing on admission: Normal Abnormal No fluid Not known
53 Rectal exam on admission: Normal Abnormal Not done Not known Abdominocentesis on admission: Normal Abnormal No fluid Not known Antimicrobial use: No Yes Not known If yes, Gentamicin Penicillin G Procai ne Pen G Potassium Other 10. Caseload: At admission At admission + 4days before admission 11. Total number of horses she dding Salmonella at admission: 12. Total number of horses shedding Salmonella at admission + 4 days before admission: 13. Mean number of horses shedding Salmonella at admission + 4 days before admission: 14. Information on Salmonella isolates: Number of fecal samples collected: Fecal sample number that tested positive first: Date of first test positive: Day of week when first test positive sample was collected: Number of hospital days when first test positive sample was collected: Serogroup: Serotype:
54 LIST OF REFERENCES 1. Castor ML, Wooley RE, Shotts EB, et al. Charasteristics of Salm onella isolated from an outbreak of equine salmonellosis in a veterinary teaching hospital. Eq. Vet. Sci 1989;9:236241. 2. Hartmann FA, Callan RJ, McGuirk SM, et al. Control of an outbreak of salmonellosis caused by drug-resistant Salmonella anatum in horses at a veterinary hospital and measures to prevent future infections. J Am Vet Med Assoc 1996;209:629-631. 3. Hird DW, Pappaioanou M and Smith BP. Case control study of risk factors associated with isolation of Salmonella saintpaul in hospitalized horses. Am J Epidemiol 1984;120:852-864. 4. Schott HC, Ewart SL, Walker RD, et al. An outbreak of salmonellosis among horses at a veterinary teaching hospital. J Am Vet Med Assoc 2001;218:1151-1159. 5. Tillotson K, Savage CJ, Salman MD, et al. Outbreak of Salmonella infantis infection in a large animal veterinary teaching hospital. J Am Vet Med Assoc 1997;211:1554-1557. 6. Ward MP, Brady TH, Couetil LL, et al. Investigation and control an outbreak of Salmonella typhimurium in a population of hospitalized horses. Vet Microbiol 2 005;107:233240. 7. Spiers SJ. Salmonellosis. Vet Clin North Am Equine Pract 1993;9:385-397. 8. House JK, Mainar-Jaime RC, Smith BP, et al. Risk factors for nosocomial Salmonella infection among hospitalized horses. J Am Vet Med Assoc 1999;214:1511-1516. 9. Hird DW, Casebolt DB, Carter JD, et al. Risk factors for salmonellosis in hospitalized horses. J Am Vet Med Assoc 1986;188:173-177. 10. Qui-shi L and Guizhen Y. Influence of surgical stress on cellu lar immunity and the induction of plastic adherent suppressor cells of spleen in mice. Immunol Invest. 1986;15:419430. 11. Farmer JJ. Enterobacteriaceae: Introduc tion and identification. In: Murray PR, Baron EJ, Pfaller MA, eds. Manual of clinical microbiology 6th ed. Washington, DC: American Society for Microbiology, 1995;438-449. 12. Kahn CM and Line S. Salmonellosis. In: Gyles CL, ed. The Merck Veterinary Manual Whitehouse station, New Jersey: MERCK & Co, 2005;157-234. 13. Smith BP. Salmonellosis. In: Smith BP, ed. Large animal internal medicine St Louis: CV Mosby Co, 1991;818-822. 14. Madic J, Hajsig B, Sostaric B, et al. An outbreak of abortion in mares associated with Salmonella abortusequi infection. Equine Vet J 1997;29:203-233.
55 15. Mastroeni P, Chabalgoity JA, Dunstan SJ, et al. Salmonella: Immune Responses and Vaccines. Vet J 2000;161:132-164. 16. Michetti P, Mahan MJ, Slauch JM, et al. Monoclonal secr etory immunoglobulin A protects against oral challenge with the invasive pathogen Salmonella typhimurium Infect Immun 1992;60:1786-1792. 17. Santos RL, Zhang S, Tsolis RM, et al. Morphologic and molecula r characterization of Salmonella typhimurium infection in neonatal calves. Vet Pathol 2002;39:200-215. 18. Santos RL, Tsolis RM, Baumler AJ, et al. Pathogenesis of Salmonella-induced enteritis. B J Med Biol Res 2003;36:1-12. 19. Karem KL, Kanangat S and Rouse BT. Cy tokine expression in the gut associated lymphoid tissue after oral administration of attenuated Salmonella vaccine strains. Vaccine 1996;14:1495-1502. 20. Darwin KH and Miller VL. Molecular basi s of the interaction of Salmonella with the intestinal mucosa. Clin Microbiol Rev 1999;12:405-428. 21. Hensel, M. Salmonella pathogenicity island 2. Mol Microbiol 2000;36:1015-1023. 22. Hohmann EL, Oletta CA, Killen KP, et al. phoP-phoQ-deleted Salmonella typhi (TY800) is a safe and immunogenic single dose typhoid fever vaccine in volunteers. J Infect Dis 1996;173:1408-1414. 23. Kirby AC, Yrlid U and Wick MJ. The inna te immune response differs in primary and secondary Salmonella infection. J Immunol 2002;169:4450-4459. 24. Dietrich G, Griot-Wenk M, Metcalfe IC et al. Experience with registered mucosal vaccines. Vaccine 2003;21:678-683. 25. Holt PS. Host susceptibility, resistance a nd immunity to Salmonella in animals. In: Wray C and Wray A, eds. Salmonella in domestic animals 1st ed. London, UK: CABI Publishing, 2000;73-87. 26. Weintraub BC, Eckman L, Okamoto S, et al. Role of and T cells in the host response to Salmonella infection as demonstrated in T-cell-receptor-defici ent mice of defined Ity genotypes. Infect Immun 1997;65:2306-2312. 27. Guthrie RK. Salmonella. Boca Raton: CRC Press, 1992;65. 28. McCain CS and Powell KC. Asymptoma tic salmonellosis in healthy adult horses. J Vet Diagn Invest 1990;2:236-237. 29. Smith BP. Salmonellosis in Ruminants. In: Smith BP, ed. Large animal internal medicine 2nd ed. St Louis: Mosby, 1996;894-898.
56 30. Smith BP: Salmonella infections in horses, Compend Cont Educ Pract Vet 3: S4S16,1981. 31. Cohen ND. Salmonellosis. In: Coloha n PT, Merrit AM, Moore JN, et al, eds. Equine medicine and surgery 5th ed. St. Louis: Mosby, 1999;749-753. 32. Smith BP, Reina-Guerra M and Hardy AJ. Prevalence and epizootiology of equine salmonellosis. J Am Vet Med Assoc 1978;172:353-356. 33. Kopf N: Rectal exam ination of the colic patient. In: Robinson NE, ed. Current therapy in equine medicine 2 Philadelphia: WB Saunders, 1987;23-27. 34. Bopp CA, Brenner FW, Wells JG, et al. Es cherichia, Shigella and Salmonella. In: Murray PR, Baron EJ, faller MA, et al, eds. Manual of clinical microbiology Washington D.C: ASM Press, 1999;459-474. 35. Cohen ND, Martin LJ, Simpson RB, et al. Comparison of polymerase chain reaction and microbiological culture for detection of sa lmonellae in equine feces and environmental samples. Am J Vet Res 1996;6:780-786. 36. Ewart SL, Schott HC, Robinson RL, et al Identification of s ources of Salmonella organisms in a veterinary teaching hospital and evaluation of the effects of disinfectants on detection of Salmonella organisms on surface materials. J Am Vet Med Assoc 2001;218:11451151. 37. van Duijkeren E, Flemming C, van Oldr uitenborgh-Oosterbaan MS, et al. Diagnosing salmonellosis in horses. Culturing of mu ltiple versus single faecal samples. Vet Q 1995;17:6366. 38. Kurowski PB, Traub-Dargatz JL, Morley PS et al. Detection of Salmonella spp in fecal specimens by use of real-tim e polymerase chain reaction assay. Am J Vet Res 2002;6:12651268. 39. Veling J, Barkema HW, van der Schans J, et al. Herd-l evel diagnosis for Salmonella enterica subsp. enterica serovar Dublin in fection in bovine dairy herds. Prev Vet Med 2002;53:31-42. 40. Barrow PA. Serological diagnosis of Salm onella by ELISA and other tests. In: Wray C and Wray A, eds. Salmonella in domestic animals 1st ed. London, UK: CABI Publishing, 2000; 407-427. 41. Van Wisen RL, van Nes A, Keuzenkamp D, et al. Monitoring of transmission of Salmonella enterica serovars in pigs using bacteriologi cal and serological detection methods. Vet Microbiol 2001;80:267-274.
57 42. Sibley J, Yue B, Huang F, et al. Comp arison of bacterial enri ched-broth cultura, enzyme linked immunosorbent assay, and brot h cultura-polymerase Chan reaction techniques for idenifying asymptomatic infect ions with Salmonella in swine. Can J Vet Res 2003;67:219224. 43. Amavisit P, Markham PF, Lightfoot D, et al. Molecular epidemiology of Salmonella Heidelberg in an equine hospital. Vet Microbiol 2001;80:85-98. 44. van Duijkeren E, Wannet WJB, Heck MEOC, et al. Serotypes, phagetypes and antibiotic susceptibilities of Salmonella strains isolated from horses in the Netherlands from 1993 to 2000. Vet Microbiol 2002;86:203-212. 45. Ikeda JS and Hirsh DC. Common plas mid encoding resistance to ampicillin, chloramphenicol, gentamicin, and trimethoprim-s ulfadiazine in two se rotypes of Salmonella isolated during an outbreak of equine salmonellosis. Am. J. Vet. Res 1985;46:769-773. 46. Rumschlag HS and Boyce JR. Plasmid pr ofile analysis of salmonellae in a large animal hospital. Vet Microbiology 1987;13:301-311. 47. Walker RL, de Peralta TL, Villanueva MR, et al. Genotypic and phenotypic analysis of Salmonella strains associated with an outbreak of equine neonatal salmonellosis. Vet Microbiol 1995;43:143-150. 48. Allen GP and Bryans JT. In: Pandey R, ed. Progress in veterinary microbiology and immunology. Basel: Karger, 1986;78-144. 49. Owen RR, Fullerton J and Barnum DA. Effects of transportation, surgery, and antibiotic therapy in ponies infected with Salmonella. Am J Vet Res 1983;44:46-50. 50. Palmer JE, Benson CE and Whitlock RH. Salmonella shed by horses with colic. J Am Vet Med Assoc 1985;187:256-257. 51. Walker RL, Madigan JE, Hird DW, et al. An outbreak of equine neonatal salmonellosis. J Vet Diagn Invest 1991;3:223-227. 52. Palmer JE and Whitlock RH. Salmonello sis. In: Colohan PT, Mayhew IG, Merritt AM, et al, eds: Equine medicine and surgery 4th ed. Goleta, Calif: American Veterinary Publications, 1991;643-647. 53. Ernst NS, Hernandez JA, MacKay RJ, et al. Risk factors associated with fecal Salmonella shedding among hospitalized horse with signs of gastrointe stinal tract disease J Am Vet Med Assoc 2004;225:275-281. 54. Palmer JE. Salmonellosis. In: Current Therapy in Equine Medicine 2 Philadelphia, USA: WB Saunders Company, 1987;88-92. 55. Murray MJ. Salmonellosis in horses. J Am Vet Med Assoc 1996;209:558-560.
58 56. Palmer JE, Benson CE. Salmonella shedding in the equine, in proceedings Am Assoc Avian Pathol Int Symp Salmonella 1985;161-164. 57. Traub-Dargatz JL, Garber LP, Fedor ka-Cray PJ, et al. Fecal shedding of Salmonella spp by horses in the United States during 1998 and 1999 and detection of Salmonella spp in grain and concentrate sour ces on equine operations. J Am Vet Med Assoc 2000;217:226-230. 58. Begg AP, Johnston KG, Hutchins DR, et al Some aspects of the epidemiology of equine salmonellosis. Aust Vet J 1988;65:221-223. 59. House JK, Smith BP, Wildman TR, et al. Isolation of Salmonella organisms from the mesenteric lymph nodes of horses at necropsy. J Am Vet Med Assoc 1999;215:507-510. 60. Kim LM, Morley PS, Traub-Dargatz JL, et al. Factors associ ated with Salmonella shedding among equine colic patients at a veterinary teaching hospital. J Am Vet Med Assoc 2001;218:740-748. 61. Losinger WC, Traub-Dargatz JL, Garber LP, et al. Factors a ssociated with fecalshedding of Salmonella spp by horses on US operations. Arq Bras Med Vet Zootec 2002;54:109116. 62. Carter JD, Hird DW, Farver TB, et al. Salmonellosis in hospitalized horses: seasonality and case fatality rates. J Am Vet Med Assoc 1986;188:163-167. 63. Traub-Dargatz JL, Salman MD and Jones RL. Epidemiologic study of salmonellae shedding in the feces of horses and potential risk factors for developmen t of the infection in hospitalized horses. J Am Vet Med Assoc 1990;196:1617-1622. 64. Baker JR and Leyland A. Diarrhea in the horse associated with stress and tetracycline therapy. Vet Rec 1973;93:583-584. 65. House JK, Smith BP, House A, et al. Imp lications of Salmonella infection in horses; diagnostic and epidemiological considerations in management and prevention of disease outbreaks, in Proceedings 14th Annu Vet Med Forum 1996;535-537. 66. Dunowska M, Morley P, Patterson G, et al. Evaluation of the efficacy of a peroxygen disinfectant-filled footmat for reduction of bacteria l load on footwear in a large animal hospital setting. J Am Vet Med Assoc 2006;228:1935-1939. 67. Stockton KA, Morley PS, Hyatt DR, et al. Evaluation of the effects of footwear hygiene protocols on nonspecific bacterial contaminat ion of floor surfaces in an equine hospital. J Am Vet Med Assoc 2006;228:1068-1073. 68. Dwyer RM. Back to basics: disinfecting stables. The Horse 1999;40:67-80. 69. Dwyer RM. Disinfecting equine facilities. Rev Sci Tech Off Int Epiz 1995; 14:403418.
59 70. Conboy HS. Preventing contagious equine diseases, in proceedings AAEP 2005;5153. 71. Cohen ND and Honnas CM. Risk factors a ssociated with developm ent of diarrhea in horses after celiotomy for colic: 190 cases (1990-1994). J Am Vet Med Assoc 1996;209:810-813. 72. Dwyer RM. Control and prevalence of foal diarrhea outbreaks, in proceedings AAEP 2001;47:472. 73. Argenzio RA. Functions of the equine la rge intestine and their interrelationship in disease. Cornell Vet 1975;65:303-330. 74. Clarke LL, Roberts MC and Argenzio RA Feeding and digestive problems in horses. Physiologic responses to a concentrated meal. Vet Clin North Am Equine Pract 1990;6:433-450. 75. Linerode PA and Goode RL. The effect s of colic on the micr obial activity of the equine large intestine, in Proceedings 16th Annu Conv Am assoc Equine Pract 1970;321-341. 76. Kawasaki T, Ogata M, Kawasaki C, et al. Surgical stress induces endotoxin hyporesponsiveness and an early decrease of mo nocyte mCD14 and HLA-DR expression during surgery. Anesth Analg 2001;92:1322-1326. 77. Ziegler-Heitbrock HW, Ulevitch RJ. CD 14: cell surface recepto r and differentiation marker. Immunol Today 1993;14:121-125. 78. Zellweger R, Ayala A, Zhu X, et al. Effect of surgical trauma on splenocyte and peritoneal macrophage immune function. J Trauma 1995;39:645-650. 79. Kawasaki T, Ogata M, Kawasaki C, et al Effects of epidural anaesthesia on surgical stress-induced immunosuppression during upper abdominal surgery. Br J Anaesth 2007;98:196203. 80. Astiz M, Saha D, Lustbader D, et al. Monocyte response to bacterial toxins, expression of cell surface receptors, and release of anti-inflammatory cytokines during sepsis. J Lab Clin Med. 1996;128:594-600. 81. Cheadle WG, Hershman MJ, Wellhausen SR, et al. HLA-DR antigen expression on peripheral blood monocytes correlat es with surgical infection. Am J Surg 1991;161:639-645. 82. Tlaskalova-Hogenova H, Yuekova L, Lodinova-Zadnikova R, et al. Mucosal immunity: its role in defense and allergy. Intl Arch Allergy Immunol 2002;128:77-89. 83. Koterba A, Torchia J, Silverthorne C, et al. Nosocomial infections and bacterial antibiotic resistance in a uni versity equine hospital. J Am Vet Med Assoc 1986;189:185-191. 84. Lolekha S. Consequences of trea tment of gastrointestinal infections. Scand J Infect Dis, Suppl 1986;49:154-159.
60 85. Sullivan A, Edlund C and Nord CE. Effect of antimicrobial agents on the ecological balance of human microflora. The Lancet Infectious Diseases 2001;1:101-114.
61 BIOGRAPHICAL SKETCH Abel Ekiri received a degree in veterinary medicine in 2002 from Makerere University, Kampala, Uganda. Thereafter, he worked as a field veterinarian with Heifer Project International, an organization that is involved with modernizing dairy farming in third world countries. In June 2003, he enrolled in a one-year dairy manageme nt course at California Polytechnic State University, San Luis Obispo. In September 2006, he joined the University of Florida, Large Animal Hospital as an Infection Control Officer, and graduate rese arch assistant. He pursued a Master of Science degree with emphasis in c linical epidemiology under the supervision of Dr. Jorge Hernandez. His research focused on epidemiology of nosocomial Salmonella infections in hospitalized horses.