Publication:

Comparison of one versus two sputum specimens in the diagnosis of pulmonary tuberculosis

Background In 2007, the number of sputum smears required for diagnosis of pulmonary tuberculosis was reduced from three to two. In high burden settings such as India, this practice continues to be the referenced standard whilst using smear microscopy. The objective of the study was to assess the plausibility of relying on one sputum smear rather than two serial sputum smears for tuberculosis diagnosis while also examining inter-specimen grading discrepancies.

Methods A retrospective review of tuberculosis specimen registers (2012–2016) to compare sputum positivity and grading for incremental yield. Inter-specimen grading proportions were compared using McNemar’s test.

Results Incremental yield of the second specimen was 9.9% [95% CI 7.3%, 12.6%]. Differences in inter-specimen grading were not significant: nil (p=0.08), scanty (p=0.75), 1+ (p=0.66), 2+ (p=0.39), 3+ (p=0.15). Less than 25% of specimens showed a change in inter-specimen grading when classified as negative, low-grade, or high-grade results.

Conclusions Two serial sputum specimens for the diagnosis of pulmonary tuberculosis remains a valid practice while continuing to rely on conventional smear microscopy. Inter-specimen grading congruencies support further exploration into same-day two-specimen collections to expedite diagnosis.