565. Entropy-based Source Attribution of Salmonella: Trends in Human Infection
Session: Poster Session: Hospital-acquired and Transplant Infections
Friday, October 30, 2009: 12:00 AM
Room: Poster Hall A
Background: Animals are an important reservoir for Salmonella bacteria, and zoonotic Salmonella infections represent an important cause of diarrheal illness in the U.S. Attributing human salmonellosis to likely original animal sources can further our understanding of common Salmonella reservoirs and their impact on public health.
Methods: We used an entropy-based source attribution model to examine trends in human salmonellosis. JMP software was used to analyze non-human Salmonella isolate data reported to CDC through the National Veterinary Services Laboratory (NVSL) and to calculate serotype-level source entropy (a measure of concentration). Single animal source attribution, based on entropy analysis, was applied to human Salmonella isolate data reported to CDC through the Public Health Laboratory Information System (PHLIS) to describe trends over time.
Results: During 1970-2006, 1,299,128 human Salmonella isolates were reported to CDC. Of those, 219,426 (17%) were serotypes highly attributed to a single animal source. Among these human isolates attributed to a single animal source, poultry-associated serotypes experienced the largest increase from 62.7% (1,442/2,300) of isolates in 1974, to 91.3% (9,888/10,834) in 1994 and remained between 85% and 90% through 2006. The largest decrease occurred among porcine-associated serotypes, falling from 26.5% (593/2,232) of human isolates in 1973, to 1.8% (140/7,544) in 2005. Reptile-associated serotypes increased from 0.6% (29/4,610) of human isolates in 1984, to 8% (487/5,823) in 2003.
Conclusion: Applying an entropy-based source attribution model to Salmonella surveillance data provides a novel approach toward quantifying the public health impact of important animal sources of Salmonella. These results may guide future pathogen reduction programs and provide the basis for more focused source specific prevention measures. Other analyses are needed to determine the routes by which Salmonella reach humans.
Ezra Barzilay, MD1, Richard Bishop, AS2, Robert Hoekstra, PhD2, Colin Schwensohn, MPH2 and  C. A. Schwensohn, None..
E. J. Barzilay, None..
R. D. Bishop, None..
R. M. Hoekstra, None., (1)Enteric Diseases Epidemiology Branch, CDC, Atlanta, GA, (2)CDC, Atlanta, GA