Diagnostic performance and costs of Capilia TB for Mycobacterium tuberculosis complex identification from broth-based culture in Bangkok, Thailand.

OBJECTIVES: Broth-based culture (BBC) systems are increasingly being used to detect Mycobacterium tuberculosis complex (MTBC) in resource-limited. We evaluated the performance, time to detection and cost of the Capilia TB identification test from broth cultures positive for acid-fast bacilli (AFB) in Thailand. METHODS: From October-December 2007, broth cultures that grew AFB from specimens submitted by district TB clinics to the Bangkok city laboratory were tested for MTBC using Capilia TB and standard biochemical tests. Isolates that were identified as MTBC by biochemical tests but not by Capilia TB underwent repeat testing using Capilia TB, Accuprobe (Gen-Probe, San Diego, CA, USA) and sequencing. Costs of time, labour, infrastructure and consumables for all procedures were measured. RESULTS: Of 247 isolates evaluated, the sensitivity of Capilia TB was 97% and its true specificity 100% compared with biochemical testing. The median time from specimen receipt to confirmed MTBC identification was 20 days (range 7-53 days) for Capilia TB and 45 days (range 35-79 days) for biochemical testing (P < 0.01). Six isolates that were Capilia TB negative but positive by biochemical testing were confirmed as MTBC and mutations in the mpb64 gene were detected in all. The unit cost of using Capilia TB was 2.67 USD that of biochemical testing was 8.78 USD. CONCLUSIONS: In Thailand, Capilia TB had acceptable sensitivity and specificity, was lower in cost and had shorter turn-around times. Laboratories investing in BBC should consider Capilia TB for identification of MTBC, after validation of performance in their setting.

Performance of the BACTEC MGIT 960 compared with solid media for detection of Mycobacterium in Bangkok, Thailand.

Controlled trials have demonstrated that liquid media culture (LMC) is superior to solid media culture for diagnosis of Mycobacterium tuberculosis (MTB), but there is limited evidence about its performance in resource-limited settings. We evaluated the performance of LMC in a demonstration project in Bangkok, Thailand. Sputum specimens from persons with suspected or clinically diagnosed tuberculosis were inoculated in parallel on solid (Lowenstein-Jensen [LJ]) and liquid (mycobacterial growth indicator tube [MGIT 960]) media. Biochemical tests identified isolates as MTB or nontuberculosis mycobacteria (NTM). Of 2566 specimens received from October 2004 to September 2006, 1355 (53%) were culture positive by MGIT compared with 1013 (39%) by LJ. Median time to growth for MGIT was significantly less than LJ: 11 versus 27 days. Of 1417 isolates detected by at least 1 media, 1255 (86%) were identified as MTB and 162 (11%) NTM. MGIT improved speed and sensitivity of MTB isolation and drug susceptibility testing, regardless of HIV status.