Abstract

The identification of M.tb in smear negative samples has the inherent TB culture delays. These delays potentially affect treatment choices and contact tracing decisions particularly where TB prevalence is low and isolation of MOTT common. Faster molecular techniques are expensive and not universally available. The identification of cord formation1 in positive cultures gives an early indication of the likely TB culture result before the final reference laboratory identification (local audit data: M.tb 12.3±5.9 d, MOTT 13.7±11.6 d). We evaluated the presence of cord formation in AFB smears prepared from MGIT tubes and stained by auramine-phenol staining from positive liquid culture (BACTEC MGIT growth supplement medium) for the presumptive identification of M.tb. We prospectively evaluated 612 positive mycobacterial culture specimens from 316 patients over 3 years. All specimens were sent to our local reference laboratory for species identification. The identification of M.tb by the presence of cord formation was compared with the subsequent reference laboratory report. 420/426 samples showing cording by smear microscopy were confirmed by the reference laboratory as M.tb (PPV: 98.6%). 182/186 non-cording culture positive specimens were confirmed as MOTT (NPV: 97.9%). The 6/612 (0.9%) MOTT misidentified as M.tb by cording were M kansasii as previously described.1 The 4/612 (0.7%) M.tb misidentified as MOTT by the absence of cording were all 3+ sputum smear positive samples as previously described.2 The presumptive identification of M.tb and MOTT by the presence or absence of cording formation in liquid culture is both sensitive and specific. This technique has a potentially significant impact on treatment timing and contact tracing decisions.