Abstract: Objective  To compare and analyze the traditional method of identifying M.  tuberculosis complex with PCR method, and to provide a scientific basis for clinical application.  Methods  The isolates of 313 sputum samples from patients diagnosed with pulmonary tuberculosis in Kashgar Thoracic Hospital from April 2010 to May 2011 were identified M. tuberculosis complex using traditional method (p-nitrobenzoic acid and thiophene-2-carboxy-lic acid hydrazide) and PCR method, respectively. 100 isolates including M. tuberculosis, M. bovis and M. africanum type Ⅰ were chosen purposely to examine the reliability of these two methods using the spoligotyping test.  Results  253 M. tuberculosis, 60 M. bovis were identified by the traditional method. 306 M. tuberculosis, 1 M. bovis and 6 M. africanum type Ⅰ were identified by PCR method. The accordance rate of these two methods was 79.87% (250/313). Although there was no statistical difference between the two methods in the identification of M. tuberculosis complex (χ²=5.05, P=0.08), the difference was statistically significant in the identification of M. bovis (χ²=4.16, P=0.04). 100 isolates purposively chosen were further identified as 90 M. tuberculosis, 2 M. bovis and 6 M. africanum type Ⅰ by Spoligotying test. According to the result of the Spoligotyping test, the accordance rates by the traditional method and PCR were 38.78% ［(2+36)/98］ and 98.98% ［(1+90+6)/98］, respectively.  Conclusion  The traditional method of strain identification of M. tuberculosis complex is  time consuming, tedious, and inaccurate, and the PCR method can accurately and quickly identified M. tuberculosis complex to subspecies.

Key words: Mycobacterium tuberculosis, Mycobacterium bovis, Culture media, Bacteriological techniques, Polymerase chain reaction, Comparative study