Accuracy of a spontaneous breathing trial for extubation of neonates

BACKGROUND: Prevalence of extubation failure in neonates may be up to 80%, but evidence to determine if a neonate is ready for extubation remains unclear. We aim to evaluate a spontaneous breathing trial accuracy with minimum pressure support to predict success in neonates’ extubation and identify variables related to failures. METHODS: This is a diagnostic accuracy study based on a cohort study in an intensive care unit with all eligible newborn infants subjected to invasive mechanical ventilation for at least 24 hours submitted to the trial for 10 minutes before extubations. The outcome was failures of extubations, considered if reintubation was needed until 72 hours. RESULTS: The incidence of failure was 14.7%among 170 extubations. There were 145 successful extubations; of these, 140 also passed the trial with a sensitivity of 96.5%(95%CI: 92.1–98.9). Of the 25 extubations that eventually failed, 16 failed the test with a specificity of 64.0%(95%CI: 42.5–82.0). The negative predictive value was 76.2%, and the positive predictive value was 94%. In stratifying by weight, the accuracy was >98.7%for neonates weighting >2500 g, but 72.5%for those weighing