Diagnosis of Tuberculosis in the Wild Boar (Sus scrofa): A Comparison of Methods Applicable to Hunter-Harvested Animals

BACKGROUND To obtain robust epidemiological information regarding tuberculosis (TB) in wildlife species, appropriate diagnostic methods need to be used. Wild boar (Sus scrofa) recently emerged as a major maintenance host for TB in some European countries. Nevertheless, no data is available to evaluate TB post-mortem diagnostic methods in hunter-harvested wild boar. METHODOLOGY/PRINCIPAL FINDINGS Six different diagnostic methods for TB were evaluated in parallel in 167 hunter-harvested wild boar. Compared to bacteriological culture, estimates of sensitivity of histopathology was 77.8%, gross pathology 72.2%, PCR for the MPB70 gene 66.7%, detection of acid-fast bacilli (AFB) in tissue contact smears 55.6% and in histopathology slides 16.7% (estimated specificity was 96.7%, 100%, 100%, 94.4% and 100%, respectively). Combining gross pathology with stained smears in parallel increased estimated sensitivity to 94.4% (94.4% specificity). Four probable bacteriological culture false-negative animals were identified by Discriminant Function Analysis. Recalculating the parameters considering these animals as infected generated estimated values for sensitivity of bacteriology and histopathology of 81.8%, gross pathology 72.7%, PCR for the MPB70 gene 63.6%, detection of AFB in tissue contact smears 54.5% and in histopathology slides 13.6% (estimated specificity was 100% for gross pathology, PCR, bacteriology and detection of AFB in histopathology slides, 96.7% for histopathology and 94.4% for stained smears). CONCLUSIONS/SIGNIFICANCE These results show that surveys for TB in wild boar based exclusively on gross pathology considerably underestimate prevalence, while combination of tests in parallel much improves sensitivity and negative predictive values. This finding should thus be considered when planning future surveys and game meat inspection schemes. Although bacteriological culture is the reference test for TB diagnosis, it can generate false-negative results and this should be considered when interpreting data.