Rationale: Knowledge of the potential for aerosol release of Mycobacterium tuberculosis (Mtb) during disease, treatment, recovery, and asymptomatic carriage is fundamental to understanding tuberculosis (TB) transmission. Objectives: To quantify viable aerosolized Mtb from TB clinic attendees. Methods: 102 presumptive TB patients from two informal settlements in Cape Town, South Africa, were classified by laboratory, radiological, and clinical features into three mutually exclusive groups: A. Sputum GeneXpert-positive TB (n=52), B. Sputum GeneXpert-negative TB (n=20), and C. TB not diagnosed (n=30). A respiratory aerosol sampling chamber was used to collect exhaled Mtb organisms over a 15-minute period at baseline, and at 2-months and 6-months post initial presentation. Viable bacilli were enumerated based on incorporation of the fluorescent probe, DMN-trehalose. Measures and Main Results: Mtb was isolated from 92%, 90% and 93% at baseline; 87%, 74%, 71% at 2 weeks; 53%, 47% and 46% at 2 months; and 32%, 25%, 22% at 6-months for groups A, B, and C, respectively. Median Mtb counts (ranges) reduced from baseline to 6 months from 10(1-38), 5(1-31), and 9(1-38) to 3(1-28), 4(3-18), and 2(1-14) for groups A, B, and C, respectively. TB symptoms resolved in all 3 groups. Conclusions: Aerosolized Mtb was isolated from almost all TB patients at baseline and reduced during treatment. Small numbers of viable organisms remained in almost a fifth of patients completing 6-months treatment. Aerosolized Mtb may be a useful metric for modifying standard TB therapy. Aerosolized Mtb in group C may reflect exacerbation of an existing infection or transient Mtb infection not reaching a clinical threshold for TB diagnosis, consistent with recent models proposing cyclic subclinical disease states.