It was hypothesised that the time to detect Mycobacterium tuberculosis in liquid culture of sputum from patients with pulmonary tuberculosis may be a better indicator for the duration of respiratory isolation than sputum smear status. Pre-treatment and during-treatment sputum acid-fast bacilli (AFB) smear and culture results were reviewed in 284 patients with pulmonary tuberculosis. The time to detect M. tuberculosis in liquid culture (TTD-TB) was the number of days from inoculation of the Mycobacterial Growth Indicator Tube to culture detection and visualisation of AFB. The median (interquartile range) TTD-TB for smear group 0 (no bacilli seen) was 14 (12-20) days. This value was used as the standard at which release from isolation could be permitted. In smear group 4 (>9 AFB per high-power field (hpf) in sputum specimens before treatment) patients, the TTD-TB exceeded 14 days after a median of 25 days of treatment. The current authors recommend that patients in smear groups 1 and 2 (1-9 AFB per 100 hpf and 1-9 AFB per 10 hpf in sputum specimens before treatment, respectively) receive treatment in respiratory isolation for 7 days, provided the risk of drug resistance is low. Smear group 3 (1-9 AFB per hpf) and 4 patients should receive treatment in respiratory isolation for 14 and 25 days, respectively. These criteria would have reduced the duration of respiratory isolation by 1,516 days in the 143 study participants with sputum smear-positive pulmonary tuberculosis. Provided clinical and radiographical criteria are satisfactory, use of the time to detect Mycobacterium tuberculosis in liquid culture could enable the duration of respiratory isolation to be predicted from the pre-treatment sputum smear grade. The recommendations enable isolation to end well before sputum becomes smear negative, with considerable benefits to patients and healthcare providers.