Predicting failure of high flow nasal cannula in bronchiolitis: a systematic review

Abstract

Background: In some infants with bronchiolitis High-Flow Nasal Cannula (HFNC) oxygen therapy fails and other respiratory support is needed.

Aim: Systematically review evidence around prognostic factors associated with HFNC failure in bronchiolitis.

Methods: We included observational studies of infants <2 years with bronchiolitis assessing prognostic factors for HFNC treatment failure. We included relevant randomised controlled trials when HFNC treatment arm data were analysed as a cohort study. We evaluated study quality using CASP critical appraisal tools. We summarized independent prognostic factors from each study and used GRADE tool to assess degree of confidence in the summary results.

Results: We included 13 studies (6/13 multicentre, 7/13 observational [6/7 retrospective]), involving 3,114 infants (85% had bronchiolitis). Studies were of reasonable quality, but considerable heterogeneity and low generalisability rendered overall quality of evidence very low. Reporting of methods, analysis and results were often inadequate. No infant characteristics (age/weight/comorbidity; 10 studies), or clinical observations/severity scores (6 studies) consistently predicted HFNC failure. When reported, markers of respiratory acidosis at HFNC initiation more consistently demonstrated statistically significant association with treatment failure (pH in 4/5 studies; pCO2 in 3/5 studies).

Conclusion: No factors consistently predict failure of HFNC in bronchiolitis, but the quality of the evidence is very low. Respiratory acidosis may predict treatment failure but should be tested in prospective multicentre studies before being used to select respiratory support modalities for individual infants with bronchiolitis.