OBJECTIVES: To explore the potential for reducing the procedural duration of the string test for the diagnosis of tuberculosis (TB) using microscopic observation drug susceptibility (MODS) culture. METHODS: Twelve patients already diagnosed with pulmonary TB, four each with sputum smear acid-fast bacilli grade 1+, 2+ and 3+, underwent four consecutive string tests of varying intra-gastric downtime (IGDT) of 30 min, 1, 2 and 4 h. Each retrieved string was cut into three - one oesophageal and two gastric sections. Eluates from one of the gastric sections and the oesophageal section were cultured in MODS after a decontamination procedure; eluate from the other gastric section was cultured in MODS with no decontamination. RESULTS: No significant difference was observed in the retrieval efficacy of Mycobacterium tuberculosis (P = 0.29) or time to positive MODS culture (P = 0.80) among string tests of varying IGDTs. Every patient with a sample that was positive after a 4-h IGDT also had positive culture of a 1-h IGDT sample. A pre-inoculation sample decontamination step significantly reduced culture contamination (P < 0.001). CONCLUSION: In smear-positive patients, reducing the IGDT to 1 h did not affect the M. tuberculosis retrieval efficacy of the string test. Future evaluations in non-expectorating human immunodeficiency virus and paediatric populations should include a 1-h IGDT. © 2008 The Union.