Sputum scarcity and respiratory sample availability among children with presumptive tuberculosis in high burden countries: a systematic review and meta-analysis.
Mary Gaeddert, Jennifer Habbes, Julian Meister, Ashlyn Beecroft, Amelie von Saint Andre-von Arnim, Beate Kampmann, Mikashmi Kohli, Claudia M Denkinger, et al. (11 authors)
medRxiv : the preprint server for health sciences · 2025-11
Abstract
BACKGROUND: Tuberculosis (TB) diagnosis typically relies on testing sputum samples, but children often cannot produce sputum. Our review investigated the collection of self-expectorated and induced sputum, and alternative methods of sampling including gastric and nasopharyngeal aspirates, among children evaluated for presumed TB in healthcare facilities.
METHODS: We searched PubMed, Embase, Cochrane Library, Web of Science, and clinical trials databases from January 2010 to June 2024. Studies not reporting sufficient information on respiratory sampling or not conducted in high TB burden countries were excluded. Summary data was extracted, and the risk of bias was assessed. Sputum scarcity was defined as the proportion of children who could not provide a sample among those attempting. The pooled estimate of sputum scarcity was calculated by random effects meta-analysis. The review protocol was registered with PROSPERO (CRD42023473882).
FINDINGS: The search identified 6,751 records and 36 studies were included which enrolled 14,018 children from 14 high burden countries. Respiratory sampling methods varied by age. In children under 5, a pooled estimate of 4% (95% CI: 2-6%, n=5) provided self-expectorated sputum, but collection was not attempted routinely. In studies collecting one or two self-expectorated spot sputum samples in children 5 to 15 years, the pooled estimate of sputum scarcity was 38% (95%CI: 20-55%, n=6). Studies performing sputum induction assisted by nasopharyngeal suctioning in children under 15 had pooled scarcity of 3% (95% CI:0-6%, n=8). For studies performing gastric aspirates, the median proportion without a sample was 0.0% (95% CI: 0·0-0·8%, IQR: 0·0-2·0%, n=23).
INTERPRETATION: Collecting respiratory samples in children with presumed TB is complex and age-dependent. Children, especially under 5, are often unable to produce self-expectorated sputum and depend on alternative methods, such as induction or gastric aspirates. TB diagnostics using samples that are easier to collect from children in resource-limited settings are needed.
FUNDING: Gates Foundation (INV-069540).