Role of spontaneous breathing trial in predicting successful extubation in premature infants

Sanjay Chawla, Girija Natarajan, Matthew Gelmini, S. Nadya J. Kazzi

Research output: Contribution to journalArticlepeer-review

46 Scopus citations


Background The ability of clinicians to predict successful extubation in mechanically ventilated premature neonates is limited. Identifying objective criteria for predicting successful extubation may reduce the incidence of failed extubation and the duration of mechanical ventilation. Objective To evaluate the validity of objective measures of lung function and spontaneous breathing trial (SBT) in predicting successful extubation among premature neonates with attempted extubations within the first 3 weeks of life. Methods Respiratory compliance (Crs) along with SBT was performed prior to elective extubations within 3 weeks of age in premature infants ≤32 weeks. Extubation was considered successful if patients remained extubated for >72 hr. Ventilator settings including mean airway pressure (MAP), set rate, and fraction of inspired oxygen (FiO2) 24 hr after re-intubation were compared with pre-extubation settings, in patients requiring re-intubation. Results Thirty-nine of 49 infants (80%) were successfully extubated. Of 41 babies who passed SBT, only 5 infants failed extubation. SBT had 92% sensitivity, 50% specificity, 88% positive predictive, and 63% negative predictive value for successful extubation. Crs was comparable between infants who were successfully extubated and those who were not. Conclusions A SBT prior to extubation may be a practical objective adjunct in predicting successful extubation in ventilated premature infants.

Original languageEnglish
Pages (from-to)443-448
Number of pages6
JournalPediatric Pulmonology
Issue number5
StatePublished - May 2013


  • compliance
  • extubation
  • premature
  • ventilation


Dive into the research topics of 'Role of spontaneous breathing trial in predicting successful extubation in premature infants'. Together they form a unique fingerprint.

Cite this