A retrospective study to predict failure of high-flow oxygen therapy for acute hypoxic respiratory failure

Objective The objective of this study is to analyse the characteristics of patients who fail high-flow nasal cannula (HFNC) therapy for the treatment of acute hypoxaemic respiratory failure and to identify predictors of treatment failure. Methods This was a single-centre, retrospective, observational study. Clinical data from 388 patients with acute hypoxaemic respiratory failure were analysed. Patients were divided into two groups: the high flow oxygen therapy success group ( HFNC-S group ) and the high flow oxygen therapy failure group ( HFNC-F group ). The primary endpoint was the need for escalation of respiratory support to tracheal intubation in the enrolled patients. The demographic data, laboratory tests, blood gas analysis data, CT severity scores, and disease severity scores were analysed to determine the difference between patients who were successful and those who failed HFNC treatment. Univariate and multivariate logistic regression models were used to assess potential predictors of failure of HFNC for patients with acute hypoxaemic respiratory failure. Results The mean age of patients enrolled was 67.97 ± 14.40 years. The HFNC-S group comprised 256 patients, while the HFNC-F group had 132 patients. The PSI score, CURB-65 score, CPIS score, CT score and SOFA score of the HFNC-F group were found to be significantly higher than those of the HFNC-S group. Within 12 hours of the initiation of treatment, the HFNC-F group exhibited significantly lower oxygen saturation index (PaO2/FiO2) and significantly higher respiratory rate than the HFNC-S group. Additionally, the HFNC-F group exhibited significantly higher levels of C-reactive protein (CRP), platelet count (PLT), D-dimer, interleukin-10 (IL-10), total bilirubin (TB) and creatinine (CB) than the HFNC-S group. Conversely, the HFNC-F group exhibited significantly lower albumin levels than the HFNC-S group. In a multivariate logistic regression analysis model, the CT score, SOFA score, IL-1β and albumin were identified as independent predictors of failure of high-flow nasal oxygen therapy. Conclusion High-flow oxygen can effectively treat patients with acute hypoxaemic respiratory failure. Chest CT severity score, SOFA score, IL-1β and albumin were independent predictors of failure of high-flow nasal oxygen therapy.