RESUMO
BACKGROUND: Collecting sputum specimens is a challenge in infants and young children. We assessed the 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. METHODS: We assessed the sputum quality and correlation with detection of Cryptosporidium spp. and evaluated safety and adverse events in 162 children. RESULTS: Among 159 stool specimens tested, 34 (21%, 95% CI 15.0 to 28%) were positive for Cryptosporidium spp. There were 160 IS and 161 nasopharyngeal (NP) specimens collected. IS and NP specimen collection was performed for each patient. 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 to 10%) IS and 4 (3%, 95% CI 0 to 5%) NP were positive for Cryptosporidium spp. When stool cryptosporidium PCR was the gold standard, IS PCR sensitivity was higher (29%, 95% CI 22 to 37%) compared with NP PCR (12%, 95% CI 7 to 17%) for detection of Cryptosporidium spp. One (0.4%) adverse event occurred, consisting of a drop in oxygen saturations at the 30-min postprocedure evaluation. Consciousness level, median respiratory rate and oxygen saturations were unchanged, before or after IS. CONCLUSIONS: IS provides good quality specimens, is more sensitive than NP specimens for diagnosis of respiratory cryptosporidiosis, and collection can be performed safely in children hospitalized with diarrheal disease.