Reply to ‘The Oxygenation Benefits Provided by High‐Flow Nasal Oxygen in Bronchoscopy May be Underestimated’

We thank Zhou and Cao for their insightful comments on our study comparing high-flow nasal oxygen (HFNO) with low-flow oxygen (LFO) during bronchoscopy [1]. We are happy to clarify the points they raised on breathing pattern, open-mouth breathing, and cough. Zhou and Cao observe that certain breathing patterns, particularly open-mouth breathing and coughing, can affect the fraction of inspired oxygen (FiO2). Higher respiratory rate and an open mouth can decrease the FiO2 using HFNO or LFO. Yaroshetskiy et al. have shown that under extreme conditions (hyperventilation/mouth-opening), the FiO2 measured in the hypopharynx may fall significantly [2]. We acknowledge these physiological phenomena; however, our trial was designed to minimise their impact in a real-world bronchoscopy setting. In our study, patients were not hyperventilating. The respiratory rates were similar between the HFNO and LFO groups, with a modest reduction in the HFNO group during EBUS procedures. Notably, during nasal bronchoscopy, we observed that patients generally kept their mouths closed. This likely limited oxygen dilution. Even in the linear EBUS subset using an oral approach, we found that HFNO maintained oxygenation as effectively as LFO. All patients received intravenous midazolam and fentanyl for conscious sedation (mean total 2 mg midazolam and 46 μg fentanyl). The resulting sedation level was on average RASS 0 to −1 (alert/calm to lightly sedated). In addition to sedation, we confirm that topical lidocaine was used to anaesthetise the upper airway. We monitored cough frequency and found no significant difference between the two groups. Patients receiving HFNO coughed a mean of 11 times per procedure versus 13 times with LFO (p = 0.90). Overall, under light sedation and topical anaesthesia, patients experienced minimal cough. This aligns with the high comfort scores reported by patients and the minimal procedural interruptions. We have previously demonstrated the feasibility of transtracheal FiO2 sampling [3]. This measure would have added complexity to this study. Instead, we ensured that baseline FiO2 was comparable between groups. Our primary outcome was the maintenance of oxygen saturation. Notably, both arms began with the same FiO2 setting and were titrated as required. This approach is a strength of our trial as previous studies did not match FiO2 between groups. Some gave HFNO at 100% O2, whereas the LFO group used a lower FiO2, which confounds the interpretation. Our study eliminated FiO2 mismatch as a confounder. Despite theoretical variations in delivered FiO2 caused by mouth opening or coughing, our randomised trial found no significant differences in oxygenation or safety outcomes between HFNO and LFO. Both devices were equally effective at maintaining oxygenation. We also observed no differences in desaturations, cough frequency, or other adverse events. HFNO offered practical advantages, including a lower need for oxygen adjustments and a lower respiratory rate in the EBUS subset. These did not translate into significant differences in hypoxaemia or patient outcomes. In summary, our trial was designed to reflect real-world clinical practice while maintaining scientific rigour and feasibility. The evidence indicates that both HFNO and conventional LFO are safe and effective for oxygenation during bronchoscopy under conscious sedation. The authors have nothing to report. The authors declare no conflicts of interest.