Failure Prediction of High-Flow Nasal Cannula at the Conventional Oxygen Therapy Phase in the Emergency Department

Abstract

<b><i>Introduction:</i></b> The use of high-flow nasal cannula (HFNC) in patients with acute hypoxemic respiratory failure has been increasing in the emergency department (ED). However, studies are lacking on the prediction of HFNC failure before therapy initiation in the ED. We investigated whether the existing indices, such as the ratio of pulse oximetry oxygen saturation/fraction of inspired oxygen to respiratory rate (ROX) and ratio of ROX index to heart rate (ROX-HR), can accurately predict HFNC failure at the conventional oxygen therapy phase in the ED. <b><i>Methods:</i></b> This retrospective single-center study included patients treated with HFNC in the ED. The ROX and ROX-HR indices were calculated before initiating HFNC. An estimated fraction of inspired oxygen was used for conventional oxygen therapy. We plotted each index’s receiver operating characteristics curve and calculated the area under the curve (AUC) for diagnostic capacity. The optimal cutoff values were assessed using the Youden index. The primary outcome was HFNC failure, defined as intubation in the ED. <b><i>Results:</i></b> Among the 97 included patients, 25 (25.8%) failed HFNC therapy in the ED. The ROX and ROX-HR indices measured before initiating HFNC showed AUCs of 0.709 and 0.754, respectively. A ROX index of &lt;5.614 and a ROX-HR index of &lt;6.152 were associated with a high risk of intubation, even after correcting for confounding variables. <b><i>Conclusion:</i></b> The ROX and ROX-HR indices measured before initiating HFNC provide a relatively fair predictive value of HFNC failure in the ED.