Comparison of several challenge models for studies in avian colibacillosis.

In previous studies, the embryo lethality assay (ELA) discriminated between virulent and avirulent avian Escherichia coli isolates, and also proved to be highly correlated with mortality and morbidity results of the intravenous (IV) challenge model. In the current study, the same 20 avian E. coli isolates were used in subcutaneous (subQ) and intratracheal (IT) chicken challenge models in order to determine whether the results from the prior ELA challenges and/or the IV challenge model correlate with these models. The correlation observed between the two previous ELA trials and the combined mortality/morbidity percentages of the subQ challenge model were r = 0.792, P > 0.0001 for the first ELA trial and r = 0.738, P = 0.0002 for the second ELA trial. The IV challenge results were more highly correlated with the subQ challenge results (mortality/morbidity comparison, r = 0.894, P < 0.0001). The IV challenge mortality results were slightly correlated (r = 0.4810, P=0.0319) with the IT challenge results. Several of the isolates differed in their ability to produce mortality and/or morbidity with the different challenge models. The mortality/morbidity results of the IV and subQ challenges and the mortality results of the ELA were all positively correlated with the ability of an E. coli isolate to produce Colicin V (ColV) (r = 0.7131, P = 0.0004). The IT mortality results were slightly correlated with the production of ColV (r = 0.455, P = 0.049). The IT challenge results were only slightly correlated with resulting IV mortality and ColV production. Previous results indicate that the ELA correlates extremely well with the IV challenge model. The current study demonstrates that ELA also correlates well with the subQ challenge model. Overall, the conclusion of this study is that the ELA, IV, and subQ challenge models similarly demonstrate the ability to discriminate between virulent and avirulent avian E. coli isolates.