Finding a needle in a haystack: Evaluation of Cartridge-Based Nucleic Acid Amplification Test in comparison with microscopy and culture for diagnosis of Extrapulmonary Tuberculosis - A study from a tertiary care centre, Kerala

Introduction. Mycobacterium tuberculosis has held an exceptional presence in medical research, especially in recent years (1). While pulmonary infections largely remain the focus of literature and public health initiatives (2), extrapulmonary tuberculosis (EPTB) infections have been gaining prominence due to their increasing prevalence, particularly in populations characterized by a low incidence of tuberculosis (2-4). Although EPTB infections, such as tuberculous meningitis, carry a high mortality risk (5), delays in diagnosis and treatment persist. These delays are driven by their diverse, non-specific clinical presentations, coupled with the variable performance of diagnostic modalities like the Cartridge-Based Nucleic Acid Amplification Test (CBNAAT) (6-10). While general literature regarding the use and efficacy of CBNAAT in diagnosing EPTB exists (2,11-15), studies focusing on its sensitivity and specificity in paucibacillary specimens are an evolving niche (16). Since there is a greater prevalence of paucibacillary EPTB specimens in our setting, where such an analysis has not been conducted, we performed a retrospective study to understand the overarching diagnostic performance of our existing modalities in EPTB specimens. Methods. An observational retrospective study was conducted with the data collected between 2019 and 2022 at a tertiary care centre in central Kerala, India. A total of 105 specimens - categorized as pleural, lymph node, pus, tissue, cerebrospinal fluid (CSF), peritoneal and others - from presumptive EPTB cases were included in the study. The samples underwent acid-fast bacilli staining microscopy, liquid culture and CBNAAT utilizing the GeneXpert platform. The results were recorded, and the data were analysed with P<0.05 considered statistically significant. Observations and Results. In the 105 samples collected for the study, 6 (5.7%) were positive on Ziehl-Nielsen staining, 26 (24.8%) samples grew M. tuberculosis complex on liquid culture and CBNAAT detected M. tuberculosis in 16 (15.2%) specimens. The sensitivity and specificity of CBNAAT, with culture as the gold standard, were found to be 50.0 and 96.2%, respectively. The sensitivity and specificity of CBNAAT for pus, peritoneal (including tissue and fluid aspirate), pleural fluids, CSF, lymph nodes and other miscellaneous sample types ranged from 33.3 to 100% and 85.7 to 100%, respectively. Among smear-negative specimens, the sensitivity and specificity of CBNAAT were 42.9 and 96.2%, respectively, whereas in smear-positive specimens, the sensitivity increased to 80.0%. Conclusion. CBNAAT is a viable tool for the detection of M. tuberculosis; however, in paucibacillary specimens, its use along with other methods like smear and culture could lead to better identification of the mycobacteria.