Comparative analysis of tuberculin and defined antigen skin tests for detection of bovine tuberculosis in buffaloes ( Bubalus bubalis ) in Haryana state, India

Background Bovine tuberculosis (bTB) is a chronic disease that results from infection with any member of the Mycobacterium tuberculosis complex. Infected animals are typically diagnosed with tuberculin-based intradermal skin tests according to World Organization of Animal Health which are presently in use. However, tuberculin is not suitable for use in BCG-vaccinated animals due to a high rate of false-positive reactions. Peptide-based defined skin test (DST) antigens have been identified using antigens (ESAT-6, CFP-10 and Rv3615c) which are absent from BCG, but their performance in buffaloes remains unknown. To assess the comparative performance of DST with the tuberculin-based single intradermal test (SIT) and the single intradermal comparative cervical test (SICCT), we screened 543 female buffaloes from 49 organized dairy farms in two districts of Haryana state in India. Results We found that 37 (7%), 4 (1%) and 18 (3%) buffaloes were reactors with the SIT, SICCT and DST tests, respectively. Of the 37 SIT reactors, four were positive with SICCT and 12 were positive with the DST. The results show that none of the animals tested positive with all three tests, and 6 DST positive animals were SIT negative. Together, a total of 43 animals were reactors with SIT, DST, or both, and the two assays showed moderate agreement (Cohen’s Kappa 0.41; 95% Confidence Interval (CI): 0.23, 0.59). In contrast, only slight agreement (Cohen’s Kappa 0.18; 95% CI: 0.02, 0.34) was observed between SIT and SICCT. Using a Bayesian latent class model, we estimated test specificities of 96.5% (95% CI, 92–99%), 99.7% (95% CI: 98–100%) and 99.0% (95% CI: 97–100%) for SIT, SICCT and DST, respectively, but considerably lower sensitivities of 58% (95% CI: 35–87%), 9% (95% CI: 3–21%), and 34% (95% CI: 18–55%) albeit with broad and overlapping credible intervals. Conclusion Taken together, our investigation suggests that DST has a test specificity comparable with SICCT, and sensitivity intermediate between SIT and SICCT for the identification of buffaloes suspected of tuberculosis. Our study highlights an urgent need for future well-powered trials with detailed necropsy, with immunological and microbiological profiling of reactor and non-reactor animals to better define the underlying factors for the large observed discrepancies in assay performance, particularly between SIT and SICCT.