Extubation Failure and Duration of Respiratory Support Postextubation in Children: Australian and New Zealand Multicenter, ICU Registry Data, 2014–2021

OBJECTIVES: To compare postextubation failure and duration of noninvasive respiratory support postextubation in children receiving continuous positive airway pressure (CPAP) or high-flow nasal cannula (HFNC). DESIGN: Retrospective cohort using the Australian and New Zealand Pediatric Intensive Care (ANZPIC) registry. SETTING: Ten PICUs and 24 mixed adult/pediatric units. PATIENTS: All children extubated between 2014 and 2021 who required noninvasive respiratory support. INTERVENTIONS: None. MEASUREMENTS: Postextubation failure rate requiring reintubation and duration of noninvasive respiratory support. Secondary outcomes were length of stay (LOS) in the PICU, time from extubation to discharge from the PICU, hospital LOS, and prevalence of death during hospital admission. RESULTS: Of 31,910 extubation episodes in 23,930 children, 7,156 children received noninvasive respiratory support. Propensity matching yielded 1418 pairs for CPAP and HFNC. We failed to identify an association between postextubation failure and the use of CPAP vs. HFNC (107/1418 [7.6%] vs. 115/1418 [8.1%]; rate ratio, 0.93 [95% CI, 0.72-1.25]). There was an associated difference in median (interquartile range [IQR]) duration of postextubation respiratory support when CPAP was compared with HFNC, 72.0 hours (IQR, 30.5-72.0 hr) vs. 30.5 hours (IQR, 17.2-72.0 hr); median difference, 41.5 hours (95% CI, 37.7-45.2 hr). Median PICU LOS on CPAP vs. HFNC was 10.2 days (IQR, 5.1-20.9 d) vs. 7.1 days (IQR, 3.9-14.3 d); median difference, 2.2 days (95% CI, 1.6-2.8 d). Use of HFNC as opposed to CPAP was associated with shorter PICU LOS postextubation and hospital LOS. Last, we failed to identify an association between postextubation support and in-hospital mortality. CONCLUSIONS: In the ANZPIC registry, 2014-2021 dataset, we failed to identify an association between type of postextubation form of respiratory support and postextubation failure rate. However, being supported with CPAP as opposed to HFNC was associated with longer duration of postextubation respiratory support in the PICU.