OBJECTIVE: To identify risk factors for nosocomial Salmonella infections among hospitalized horses. DESIGN: Longitudinal study. ANIMALS: 1,583 horses hospitalized in an intensive care unit between January 1992 and June 1996. PROCEDURE: Survivor functions were used to estimate time to shedding salmonellae for various Salmonella serotypes. Survival analysis was then used to determine how variables associated with patient management, environmental conditions, hospital conditions, and other disease processes affected the risk of nosocomial Salmonella infection. RESULTS: 78 horses shed Salmonella organisms: 35 shed Salmonella krefeld, 26 shed S typhimurium, and 17 shed other Salmonella serotypes. Mean time from admission to shedding was significantly longer for horses shedding S krefeld or S typhimurium than for horses shedding other Salmonella serotypes. Therefore, infection with S krefeld or S typhimurium was considered nosocomial. Seven variables were found to be significantly associated with risk of nosocomial Salmonella infection: mean number of horses in the hospital shedding S krefeld during the 4 days prior to and the day of admission, mean number of horses shedding S typhimurium during this period, a diagnosis of large colon impaction, withholding feed, number of days fed bran mash, duration of treatment with potassium penicillin G, and mean daily ambient temperature. CLINICAL IMPLICATIONS: Results suggest that risk of nosocomial Salmonella infections is greater for horses with large colon impactions. In addition to implementing hospital protocols that minimize cross contamination between patients, strategies to reduce the risk of nosocomial Salmonella infection should include minimizing use of potassium penicillin G and regulation of environmental temperature in the hospital.
OBJECTIVE: To identify risk factors for nosocomial Salmonella infections among hospitalized horses. DESIGN: Longitudinal study. ANIMALS: 1,583 horses hospitalized in an intensive care unit between January 1992 and June 1996. PROCEDURE: Survivor functions were used to estimate time to shedding salmonellae for various Salmonella serotypes. Survival analysis was then used to determine how variables associated with patient management, environmental conditions, hospital conditions, and other disease processes affected the risk of nosocomial Salmonella infection. RESULTS: 78 horses shed Salmonella organisms: 35 shed Salmonella krefeld, 26 shed S typhimurium, and 17 shed other Salmonella serotypes. Mean time from admission to shedding was significantly longer for horses shedding S krefeld or S typhimurium than for horses shedding other Salmonella serotypes. Therefore, infection with S krefeld or S typhimurium was considered nosocomial. Seven variables were found to be significantly associated with risk of nosocomial Salmonella infection: mean number of horses in the hospital shedding S krefeld during the 4 days prior to and the day of admission, mean number of horses shedding S typhimurium during this period, a diagnosis of large colon impaction, withholding feed, number of days fed bran mash, duration of treatment with potassium penicillin G, and mean daily ambient temperature. CLINICAL IMPLICATIONS: Results suggest that risk of nosocomial Salmonella infections is greater for horses with large colon impactions. In addition to implementing hospital protocols that minimize cross contamination between patients, strategies to reduce the risk of nosocomial Salmonella infection should include minimizing use of potassium penicillin G and regulation of environmental temperature in the hospital.