[Limits of nucleic acid amplification tests for diagnosis of smear negative pulmonary tuberculosis]

To assess the usefulness of commercial kits of nucleic acid amplification test (NAAT) for diagnosis of smear negative (SN) pulmonary tuberculosis.

Retrospective study of patients who were diagnosed as, or suspected of pulmonary tuberculosis during 3 years from January 1996 to December 1998 in Fukujuji Hospital which has 100 beds for tuberculosis patients.

145 smear-negative culture-positive pulmonary tuberculosis patients are entered to our analysis. The DNA-based amplification test kit (Amplicor Mycobacterium Tuberculosis Test (AMPL), Roche Diagnostic Systems, Basel, Switzerland) detected 39.2% (20/51, 95% confidence interval (CI): 25.8-52.6%) of smear-negative culture-positive (SNCP) pulmonary tuberculosis cases. The RNA-based amplification test kit (Gen-Probe Amplified Mycobacterium Tuberculosis Direct Test (AMTDT), Gen-Probe Inc., San Diego, Calif., USA) detected 40.5% (15/37, 95% CI: 24.7-56.3%) of SNCP pulmonary tuberculosis cases. For both NAATs (AMPL and AMTDT), between two groups with and without the NAAT at diagnosis of SNCP pulmonary tuberculosis, there was statistical difference in culture-positive rate (proportion of positivity in sputum culture tests at diagnosis), but no statistical difference in maximum number of colony of Mycobacterium tuberculosis (MTB). When stratified for the culture-positive rate, adjusted sensitivity for SNCP patients was 44.2% (AMPL) and 40.4% (AMTDT) respectively. On the other hand, among 245 patients with sputum AMPL positive results during the 3 years, 8 were smear-negative culture-negative (SNCN), only one out of these 8 cases was judged as true active tuberculosis without treatment. Among 89 patients with sputum AMTDT positive results, 7 were SNCN, and 3 out of them were judged as true active tuberculosis without treatment.

Usefulness of commercial NAAT kits (AMPL and AMTDT) to diagnosis SN pulmonary tuberculosis is limited in the point of sensitivity.