How many sputum samples should be examined during follow-up of tuberculosis patients in Myanmar: two or one?

Setting: Ten selected microscopy centres in Sagaing Region, Myanmar, functioning under an external quality assurance system with no reported major errors. Objective: To assess the pattern of serial sputum results (NN, both smear-negative; NP, first smear-negative and second smear-positive; PN, first smear-positive and second smear-negative; and PP, both smear-positive) among follow-up sputum microscopy examinations of tuberculosis (TB) patients (end of intensive phase, mid-continuation phase and end of treatment) conducted from 1 November 2017 to 15 April 2018. Design: Cross-sectional study using secondary data (laboratory registers). Results: Of 2001 examinations, 94 (4.7%) were smear-positive: 66 PP (3.3%), 12 PN (0.6%) and 16 NP (0.8%); 75% of NP results were scanty. The proportion of NP results was 0.8% (95%CI 0.5-1.3), i.e., 125 smears (95%CI 77-200) were required to detect one additional smear-positive result in the second sample. Of the 16 NP results (15 patients), 14 were tested using Xpert^® MTB/RIF and none had rifampicin resistance. During the continuation phase of treatment, 13 became smear-negative, one remained smear-positive and one had unknown follow-up smear status. Conclusion: The benefit of the second sputum sample for monitoring anti-tuberculosis treatment was negligible. Given the favourable resource implications (reduced laboratory workload and costs), we recommend changing the policy from two sputum smears to one during follow-up sputum examinations of TB patients.