Patients with suspected pulmonary tuberculosis (PTB) are usually placed in respiratory isolation awaiting three sputum smear microscopy results for acid fast bacilli (3AFB). GeneXpert MTB/RIF® (Xpert) on a pooled sample from two sputa may allow for more rapid de-isolation.To compare the sensitivity and negative predictive value (NPV) of Xpert performed on a single pooled sputum sample ('pooled Xpert') to 3AFB, in order to exclude PTB in patients placed in respiratory isolation METHODS: Hospital inpatients in respiratory isolation for possible PTB were prospectively enrolled. Three expectorated sputum samples were obtained for smear microscopy. Two of the same samples had 0.5 mL removed from each and pooled for 'pooled Xpert'. We assessed the diagnostic accuracy of 'pooled Xpert' and 3AFB compared to liquid culture at eight weeks as the reference standard.Of 56 participants, nine (16.1%) were diagnosed with PTB. Compared to liquid culture 'pooled Xpert' had a sensitivity of 88.9% (95% confidence interval [CI], 57 to 99%) and NPV of 97.9% (95% CI, 89 to 99%). 3AFB had a sensitivity of 66.7% (95% CI, 35 to 88%) and NPV of 93.5% (95% CI, 83 to 98%).A single 'pooled Xpert' was non-inferior to 3AFB with a strong trend towards greater sensitivity and better NPV. These findings support the use of a single 'pooled Xpert' as an effective rapid screening approach for ruling out PTB in low incidence settings. Its value in high incidence settings and optimal combination with smear microscopy and culture warrant further evaluation.