|Dey, Bhabani - USDA, FSIS, WASHINGTON DC|
|Hsieh, Ching-Lu - ISL,VISITING SCI.,TAIWAN|
Submitted to: Poultry Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: September 24, 2002
Publication Date: February 1, 2003
Citation: DEY, B.P., CHEN, Y.R., HSIEH, C., CHAN, D.E. DETECTION OF SEPTICEMIA IN CHICKEN LIVERS BY SPECTROSCOPY. POULTRY SCIENCE. 82:199-206. 2003. Interpretive Summary: To reduce the incidence of human illness associated with the consumption of meat and poultry products, a new project called HACCP-Based Inspection Models Project (HIMP) is being tested by FSIS. The HIMP requires zero tolerance for the two food safety categories (FS-1 and FS-2) related to infection and fecal contamination. One of the major defects for which chicken carcasses are removed from the processing line is septicemia/toxemia (septox) condition. To establish a procedure for differentiating normal chickens from chickens with septox condition by machine, spectral measurements of normal and diseased livers were collected and analyzed. Using a neural network classifier, normal and septox livers were correctly differentiated by spectroscopy at a rate of 96%. A 100% correlation was established between the spectroscopic identification and the subset of normal and septox samples whose conditions were histopathologically diagnosed. Implementing the concept of this study for detection of septox carcasses would improve inspection speed, accuracy, and overall effectiveness, reduce cost and variability, and minimize problems of human error, motion injury, attention deficiency, and fatigue. This research would be of great interest to FSIS, poultry processors, and research engineers who are developing systems for the implementation of HACCP programs.
Technical Abstract: To establish a procedure for differentiating normal chickens from chickens with septicemia/toxemia (septox) condition by machine inspection under the HIMP model system, spectral measurements of three hundred chicken livers, of which half were normal and half were condemned due to septox conditions, were collected and analyzed. Neural network classification of the spectral ldata after principal components analysis (PCA) indicated that normal and septox livers were correctly differentiated by spectroscopy at a rate of 96%. Analysis of the data established 100% correlation between the spectroscopic identification and the subset of samples, both normal and septox, that were histopathologically diagnosed. In an attempt to establish the microbiological etiology of the diseased livers, isolates from 30 livers indicated that the poultry carcasses were contaminated mostly with Coliform present in the environment, hindering the isolation of fpathogenic microorganisms. Therefore, to establish the etiology of diseased livers, a strict aseptic environment and procedure for sample collection is required.