Serodiagnosis of Mycobacterium avium infections in pigs.

The aim of this study is the development and evaluation of a serodiagnostic assay for Mycobacterium avium (MA). After screening MA lipid fractions in an ELISA format, a polar lipid fraction was selected as antigen because of its superior recognition by serum antibodies in experimentally infected pigs. The resulting MA-ELISA was evaluated as an alternative for detection of MA infection by traditional pathological examination of pig lymph nodes for granulomatous lesions by meat inspectors. By comparing with bacteriological examination, the MA-ELISA showed significantly better sensitivity (69%) as compared to pathological examination (31%) in experimentally infected pigs. The MA-ELISA also appeared significantly more specific in a set of serum samples from MA negative pigs: only 1 out of these 153 serum samples reacted positive, whereas 99 (65%) of these had displayed false positive results by detection of lymph nodes lesions that appeared not to be associated with MA (Komijn et al., 2007). The MA-ELISA was subsequently evaluated using serum samples from two farms with pigs known to be infected with MA. Bacteriological examination of the sub-maxillary and mesenteric lymph nodes showed that 56% (103/184) and 35% (41/117) of the pigs, respectively were positive for MA in these farms. In the first farm, 16% (29/184) of the pigs tested positive in MA-ELISA and 31% (57/184) by pathological examination. On the contrary, in the second farm, more pigs tested positive 17% (15/117) in MA-ELISA with 8% (9/117) positivity by pathological examination. Taking the results on both farms together, the sensitivity of the MA-ELISA was 14% and the specificity 83%, whereas the sensitivity of the pathological examination was 31% and the specificity 86%. For practical reasons use of a serological test as the MA-ELISA may be preferred over pathological or bacteriological examinations. Our studies in experimentally infected and negative "field" sera indicate that the MA-ELISA is significantly more specific and more sensitive than detection by classical pathological examination. However, the studies in two MA infected farms show a variable picture with pathological examination overall performing better. Study in a wider range of "positive" farms will be needed to provide a more comprehensive view of the quality of both tests for detection of MA in infected farms. At the same time further optimization of MA-ELISA with use of lipid antigens from a broader range of serotypes may improve its performance in the face of infections with different MA serotypes.