Nearly a quarter of TB patients can’t produce sputum, study finds

Diagnosing tuberculosis (TB) in routine care usually depends on a patient being able to cough up a sputum sample that can be tested in the lab. But some people cannot produce an adequate sputum sample — a problem called sputum scarcity — and when that happens TB can be missed. To measure how common this problem is, Mary Gaeddert and colleagues carried out a systematic review and meta-analysis of studies done in healthcare settings from January 2010 to October 2023. They searched PubMed, Embase, Cochrane Library, Web of Science, and clinical trials databases with no language restrictions, and they registered their plan on PROSPERO (CRD42023473882). The team excluded studies of children under 15, studies that only enrolled people already providing sputum, and any reports that did not give clear information about sputum collection. The review pulled together published summary data, assessed the risk of bias in included studies, and aimed to estimate the share of adolescents and adults being evaluated for presumptive TB who could not provide sputum for testing.

The search returned 9,895 records and after screening the authors included 114 studies, of which 81 were rated as low risk of bias. They combined results using a random effects meta-analysis and reported both medians and pooled estimates in pre-specified subgroups. Across all 114 studies the median proportion of sputum scarcity was 6.0% (95% CI: 2.9-9.1%, IQR: 0-19.9%). However, when they focused on studies that attempted to collect one or two self-expectorated spot sputum samples the pooled estimate of sputum scarcity rose to 23% (95% CI: 14-33%, n=27). People living with HIV (PLHIV) had higher rates of scarcity: among studies enrolling only PLHIV the pooled estimate for one or two self-expectorated spot sputum samples was 24% (95% CI: 15-33%, n=9), and it was 32% (95% CI: 22-41%, n=5) for PLHIV who were inpatients or had advanced disease. By contrast, studies of patients without HIV found the lowest pooled estimate — 12% (95% CI: 3-21%, n=5) unable to provide a self-expectorated sample. Studies that used sputum induction to collect one or two spot samples had a pooled scarcity of 10% (95% CI: 0-21%, n=12).

These findings make plain that sputum scarcity is not a rare hiccup but a frequent barrier to TB diagnosis in everyday clinical settings, and the problem is worse in people with HIV and in those who are sicker or hospitalized. If nearly a quarter of people being tested for TB cannot provide the key sample, many cases may be missed or diagnosis delayed, undermining efforts to control transmission and start treatment promptly. The results also show that sputum induction reduces but does not eliminate scarcity, suggesting that relying only on sputum-based testing will leave gaps. The authors conclude that the scale of the problem supports ongoing work to develop non-sputum TB tests and alternative diagnostic pathways so clinicians can find TB in people who cannot cough up sputum. The study was funded by the Gates Foundation (INV-069540), underlining global interest in improving TB detection tools and strategies.