Objectives: To assess the yield of repeated sputum induction for the diagnosis of active tuberculosis in patients who do not produce spontaneous sputum, or with smear-negative spontaneous samples.
Methods: Induced sputum was examined with fluorescent microscopy, two amplification methods (PCR Amplicor MTB, and MTD2), and cultured for mycobacteria using liquid (Bactec 12B) and Lowenstein-Jensen media. Bronchoscopy and collection of other specimens were performed at the discretion of the treating physician.
Results: A total of 1115 sputum inductions performed in 500 patients without adverse events yielded an adequate specimen in 1113 (99.8%), and microbiological confirmation in 43 of 44 (98%) culture-positive active TB cases. Yield increased with repeated sputum induction. The cumulative yield for acid-fast bacilli smear and mycobacterial culture was 64% and 70% respectively for one, 81% and 91% for two, 91% and 99% for three, and 98% and 100% for four induced samples. Yield of PCR also increased with the greater number of induced samples tested.
Conclusions: Repeated sputum induction could considerably improve diagnostic accuracy for pulmonary TB.