BackgroundCollecting sputum specimens are a challenge in infants and young children. We assessed performance and safety of induced sputum (IS) collection in this population, embedded in a prospective study evaluating respiratory cryptosporidiosis in Malawian children with diarrheal disease.MethodsWe assessed sputum quality and correlation with detection of cryptosporidium, and evaluated safety and adverse events in 162 children.ResultsAmong 159 stool specimens tested, 34 (21%, 95% CI 15.0 – 28%) were positive for Cryptosporidium spp. There were160 IS and 161 nasopharyngeal (NP) specimens collected. The majority of IS specimens 122/147 (83%) were clear in appearance, and 132/147 (90%) were of good quality. Among the respiratory specimens tested, 10 (6.3%, 95% CI 2.5 - 10) IS and 4 (3% (95% CI 0 – 5)) NP were positive for Cryptosporidium spp. When stool cryptosporidium PCR was the gold standard, IS PCR sensitivity was higher (29 %, 95% CI 22 – 37) compared to NP PCR (12%, 95% CI 7 – 17) for detection of Cryptosporidium spp. One (0.4%) adverse event occurred, a drop in oxygen saturations at 30-minute post procedure evaluation. Consciousness – level, median respiratory rate and oxygen saturations were unchanged, before or after IS.ConclusionsIS provides good quality specimens, is more sensitive than NP specimens for diagnosis of respiratory cryptosporidiosis, and collection can be done safely in children hospitalized with diarrheal disease.