Efficacy of Amplicor PCR for the diagnosis of tuberculosis in respiratory specimens other than sputum.

A total of 832 respiratory specimens not including the sputum (402 bronchial lavages, 241 bronchial brushing specimens, 136 pumping lavages, 41 pleural effusions, and 12 others) from 462 patients were assayed using the Roche Amplicor Mycobacterium tuberculosis test for amplification and identification of M. tuberculosis, M. avium and M. intracellulare (Amplicor PCR). The results were compared with those obtained using conventional microscopy and cultivation methods. Each patient had little or no sputum and showed an abnormal chest X-ray shadowing of unknown cause. No patients had previously undergone antituberculous therapy. Of the specimens obtained, 24 were both PCR and culture positive, 786 were both PCR and culture negative, 11 were PCR positive and culture negative, and 11 were PCR negative and culture positive. Based on these results, the sensitivity and specificity of Amplicor PCR were determined to be 68.67% and 98.6%, respectively, when compared with culture of respiratory specimens not including the sputum. After correcting for discrepancies due to differences in patient clinical data, the sensitivity of Amplicor PCR was found to be 68.6%, and the specificity to be 99.9%; the corresponding values for culture were 66.7% and 100%, and those for smear were 9.8% and 100%. Thus, Amplicor PCR was shown to possess a similar sensitivity to culture and to be a highly specific technique for the diagnosis of tuberculosis in the respiratory system using non-sputum specimens within hours in patients showing little or no sputum and abnormal chest X-ray shadowing of an indeterminant cause.