Sensitivity and specificity of bronchoalveolar lavage and protected catheter brush methods for isolating bacteria from foals with experimentally induced pneumonia caused by Klebsiella pneumoniae.

Abstract

One indication for referral of horses to veterinary hospitals is for diagnosis of the microbiologic cause of pneumonia, particularly when the initial treatment fails. Although endoscopic methods have long been available for microbiologic sample collection, accuracy of these methods under these conditions have not been studied in detail. We compared the bacteria isolated from samples obtained by bronchoalveolar lavage (BAL) with those obtained by protected catheter brush (PCB) from foals with unilateral pneumonia induced by inoculation with Klebsiella pneumoniae. As part of previously described clinical trials, foals were administered antimicrobial therapy IM (n = 15) or vehicle IM (n = 7), and collection of distal airway secretion samples was conducted during the treatment period. Sensitivity and specificity of the sample collection methods were assessed by comparison of the isolates from BAL or PCB samples with isolates from tissue of the inoculated lung lobe, which was the most severely affected lung region. Sensitivity and specificity of BAL for recovery of K pneumoniae (challenge strain) and Streptococcus zooepidemicus (common secondary pathogen) was 90 and 69%, respectively, compared with 76 and 85%, respectively, for the PCB method. Sensitivity was significantly (P = 0.03) higher for BAL (100%) than for PCB (69%) for recovery of K pneumoniae (P = 0.03) from lungs. However, difference in the sensitivity of these methods for recovery of S zooepidemicus was not significant. In conclusion, BAL was a more reliable method for recovery of bacteria from the lungs in chronically infected foals that received antimicrobial treatment.