Sputum scarcity among adolescents and adults with presumptive tuberculosis: a systematic review and meta-analysis.

BACKGROUND: The diagnosis of tuberculosis (TB) typically relies on being able to produce a sputum sample for microbiological testing. However, sputum scarcity, the inability to self-expectorate an adequate sputum sample for TB testing, is a well-known concern. Our systematic review and meta-analysis investigated the proportion of sputum scarcity among adolescents and adults being evaluated for presumptive TB in healthcare facilities.

METHODS: We searched PubMed, Embase, Cochrane Library, Web of Science, and clinical trials databases with no language restrictions from January 2010 to October 2023 using terms for 'TB', sputum, and diagnostic studies. We excluded studies with participants aged <15 years, that enrolled patients already providing sputum, or not adequately reporting information on sputum collection. Published summary data was extracted, and the risk of bias was assessed. Summary estimates for the proportion of sputum scarcity were calculated overall and by pre-specified sub-groups. The pooled proportion of sputum scarcity was calculated by random effects meta-analysis. The review protocol was registered on PROSPERO (CRD42023473882).

FINDINGS: Our search identified 9895 records, of which 114 studies were included and 81 were rated as a low risk of bias. The median proportion of sputum scarcity across all 114 studies was 6.0% (95% CI: 2.9-9.1%, IQR: 0-19.9%). In subgroup meta-analyses limited to studies collecting one or two self-expectorated spot sputum samples, the pooled estimate of sputum scarcity was 23% (95%CI: 14-33%, n=27). Sputum scarcity was higher in PLHIV sub-groups. The pooled estimate of sputum scarcity in studies enrolling only PLHIV was 24% (95% CI: 15-33%, n=9) for collection of one or two self-expectorated spot sputum samples. Sputum scarcity was the highest in PLHIV inpatients or with advanced disease, with 32% (95% CI: 22-41%, n=5) unable to provide one or two self-expectorated samples. Patients without HIV had the lowest pooled estimate of scarcity, with 12% (95% CI: 3-21%, n=5) unable to provide a self-expectorated sample of any number or collection time. And studies using sputum induction to collect one or two spot samples had a pooled scarcity of 10% (95% CI: 0-21%, n=12).

INTERPRETATION: Sputum scarcity is seen in nearly a quarter of patients being evaluated for TB and this compromises TB detection. These findings support the ongoing work to develop non-sputum TB tests.

FUNDING: Gates Foundation (INV-069540).