Effect of post-extubation high-flow nasal cannula on reintubation in elderly patients: a retrospective propensity score-matched cohort study.

Background: Studies of mechanically ventilated patients with a low risk of reintubation have suggested that the use of high-flow nasal cannula (HFNC) oxygen therapy reduces the risk of reintubation compared with conventional oxygen therapy (COT). However, the effect of HFNC following extubation in elderly patients with a high risk of reintubation remains unclear. Methods: All consecutive medical intensive care unit (ICU) patients aged >65 years who were mechanically ventilated for >24 h were prospectively registered between July 2017 and June 2018. Control was obtained from a historical database of patients attending the same ICU from January 2012 to December 2013. A total of 152 patients who underwent HFNC after planned extubation according to institutional protocols (HFNC group) were compared with a propensity-matched historical control group who underwent COT (n = 175, COT group). The primary outcome was the proportion of reintubated patients within 48 h after planned extubation. Results: One hundred patients from the HFNC group and 129 patients from the COT group were matched by a propensity score that reflected the probability of receiving HFNC, and all variables were well matched. Post-extubation respiratory failure (41.0% versus 33.3%, p = 0.291) and reintubation rate within 48 h (16.0% versus 11.6%, p = 0.436) did not differ between the HFNC and COT groups. However, decreased levels of consciousness as a sign of post-extubation respiratory failure (27.0% versus 11.7%, p = 0.007) were significantly increased in the HFNC group compared with the COT group. Conclusion: Among elderly patients who underwent planned extubation, HFNC was not associated with a decrease in the risk of reintubation. Further prospective study evaluating the clinical benefits of post-extubation HFNC in elderly patients is needed. The reviews of this paper are available via the supplemental material section.