Interpreting diagnosis outcomes for tuberculosis to timely and reliably predict non-tuberculosis mycobacteria isolation.

Introduction. Timely distinguishing non-tuberculous mycobacteria (NTM) from Mycobacterium tuberculosis is needed, but it is challenging.Hypothesis. Smear-positive and tuberculosis (TB) molecular-test-negative outcomes could timely and accurately predict NTM existence in the clinical specimen.Methodology. Laboratory outcomes of the smear test and TB molecular test outcomes were evaluated in a high TB and NTM prevalence setting. Additionally, the interferon-gamma release assay (IGRA) outcome was scrutinized to assess its supplementary value to the above strategy.Results. The smear-positive/Xpert MTB/RIF (Cepheid, USA) outcomes accurately predicted 91.67% (198/216) of the NTM isolation, while that of smear-positive/Simultaneous Amplification and Testing method (SAT-TB) (Rendu Biotechnology, China) negative outcomes was 84.5% (169/200). Applying these indicators to rule out TB could achieve an accuracy of up to 99.49% (3435/3453). Combining smear-positive, Xpert-negative and SAT-TB-negative outcomes increased the accuracy up to 95%. Adding a negative IGRA outcome to the indicators further increased the accuracy to over 96%, albeit at the cost of losing prediction sensitivity. When evaluating the strategy in NTM isolates, the indicators successfully predicted about 40% of these isolations with over 92% accuracy.Conclusion. A smear-positive/molecular TB test-negative outcome could timely and accurately predict NTM isolation in the given setting. This strategy could predict ~40% of the NTM isolations of the patients on their first day of hospital visit.