Backgrounds: The bacteriologic diagnosis in pulmonary tuberculosis is confirmative, even though successful rates in self-ex-pectorated sputum are limited. In the previous studies, sputum specimens induced by hypertonic saline nebulization facilitated bacteriologic diagnosis with higher sensitivity over those from self-expectorated sputum. The benefits of the sputum induction were investigated in the diagnosis of pulmonary tuberculosis. Methods: A prospective randomized case-control study in one hospital. The subjects highly suspicious of pulmonary tuber-culosis were asked to provide 3 pairs of sputum specimens in 3 consecutive days. The first pairs of the specimens were obtained either by self-expectoration (SE) from the next day of the visit or sputum induction with 7% saline nebulization at presentation (HS) and the other pairs were collected in the same ways. The samples were used for acid fast bacilli staining, mycobacterial culture and PCR. The outcomes of bacteriologic detections were compared. Results: Seventy one patients were assigned either into SE (35 subjects, age of 52±17, 69% male) and HS (36, 54±17, 47%). Forty one subjects with pulmonary tuberculosis (57.7%) and 3 with non-tuberculous mycobacterial infection (4.2%) were diagnosed. At presentation, 21 (60%) of SE and 20 (55.6%) of HS revealed scanty amounts of sputum (p=0.705). In the myco-bacterial detection, HS (13 subjects, 36.1%) and SE (13, 37.1%) showed no difference (p=0.928). Consequently, 22 (62.9%) of SE and 19 (52.8%) of HS were clinically diagnosed and treated as pulmonary tuberculosis.