Setting: Community-wide active case finding for tuberculosis (TB) using Xpert® MTB/RIF as the primary screening tool, Ca Mau Province, Viet Nam.
Objectives: To determine whether macroscopic sputum quality characteristics (sputum colour and volume) can be used to predict Xpert MTB-negative sputum and hence exclude sputum samples from testing.
Design: Field staff conducted household visits to approximately 51,200 adults in 58 villages randomly selected from throughout the province. Sputum samples from all screened participants who were able to produce ⩾1 ml sputum underwent macroscopic sputum assessment and were tested with Xpert.
Results: Of the 21,624 sputum samples tested, 159 (0.74%) were Xpert MTB-positive; 93% of the samples were 1-2 ml and nearly all were mucoid (93%) or mucopurulent (5.7%). One salivary sample was Xpert MTB-positive (2.0% of all salivary samples). The lowest positive predictive value for any sputum volume or colour characteristic was 0.66%. This was not substantially different from the overall prevalence of positive sputum Xpert MTB (0.74%).
Conclusion: Sputum colour and volume cannot be used to predict the presence or absence of M. tuberculosis in sputum detected using Xpert. These sputum quality parameters cannot therefore be used to exclude sputum samples from testing for TB.