Sputum appearance doesn’t change TB test accuracy, study finds

Diagnosing tuberculosis (TB) commonly relies on testing sputum, the mucus people cough up from their lungs. The most widely used molecular test for this purpose is Xpert MTB/RIF Ultra (Xpert). Concern has existed among clinicians and laboratory staff that the visible quality of a sputum sample—whether it looks watery, mucoid, purulent, or salivary—might affect the chance of detecting TB. To investigate this, researchers working with the R2D2 TB Network and ADAPT studies screened consecutive people with symptoms of TB in seven countries: India, the Philippines, Vietnam, Nigeria, South Africa, Uganda, and Zambia. As corresponding author A. Cattamanchi and colleagues report, 1,855 participants provided 2–3 sputum samples for testing. The study group included 798 women (43%), 348 people living with HIV (19%), and nearly all participants (1,795; 97%) had a cough lasting two weeks or more. The team focused on the first sputum sample from each participant, grading its macroscopic appearance using standardized procedures performed by trained research staff, and then running that sample on Xpert to compare results across different sputum types.

The study collected data on Xpert results and compared them to culture-based testing using a microbiological reference standard (MRS). Overall, 313 of the 1,855 participants (17%) had a positive Xpert result. Most first sputum samples were classed as salivary (83%). Xpert positivity varied by macroscopic grade: 16.1% among salivary samples and 31.2% among purulent samples. The researchers used logistic regression to test whether sputum grade predicted Xpert positivity after adjusting for demographic and clinical variables, and they found no significant independent association between any sputum grade and Xpert positivity. The performance of Xpert against the culture-based MRS was high across all sample types. Reported sensitivities were salivary 89%, mucoid 91%, mucopurulent 87%, and purulent 100%, while specificity exceeded 98% for all sputum grades. These findings show that Xpert MTB/RIF Ultra reliably detected TB across a range of visible sputum qualities.

The key message from this multi-country study is that the macroscopic appearance of sputum should not determine whether a sample is tested with Xpert MTB/RIF Ultra. Because sensitivity remained high even for salivary samples and specificity was over 98% across all grades, the authors conclude that molecular testing can be applied to all sputum samples regardless of their visible quality. Practically, this supports laboratory and clinic workflows that accept and test the first available sputum rather than discarding or delaying testing until a thicker or purulent sample is produced. For busy or resource-limited settings, that could simplify specimen collection instructions to patients and reduce missed opportunities to diagnose TB. The study’s standardized grading, the use of a culture-based microbiological reference standard (MRS), and the large, diverse sample across seven countries strengthen confidence in applying these findings broadly.