A prospective observational study comparing lung ultrasound and clinical scores ability in prediction of non invasive respiratory support outcome in acute respiratory failure

Abstract

Abstract Background Non-invasive oxygenation techniques lower the risk of ventilator-associated complications in cases of acute respiratory failure (ARF). Employing specific scoring systems derived from lung ultrasound (U/S) can predict failure of non-invasive respiratory support (NIRS), thereby reducing the use of mechanical ventilation (MV) and its associated complications. Objectives To assess the ability of different chest U/S scores to predict failure of NIRS methods in comparison with other scores, e.g., HACOR score, Acute Physiology and Chronic Health Evaluation (APACHE II), Respiratory rate–Oxygenation index (ROX), and the partial pressure of arterial oxygen to fraction of inspired oxygen (PaO2/FiO2) ratio. Methods Eighty-two patients presenting with ARF (type I and II) were managed with high-flow nasal cannula (HFNC) and non-invasive ventilation (NIV). Routine laboratory tests were performed. Lung ultrasound score (LUS), diaphragmatic thickening fraction (DTF), ROX, HACOR, APACHE II, and PaO2/FiO2 ratio were calculated on admission and after 48 h. Results Among the 82 patients enrolled in this study, 20 (16 succeeded and 4 failed) were treated with NIV, and 62 (46 succeeded and 16 failed) were treated with HFNC. Comparisons between LUS, DTF, ROX, HACOR, APACHE II, and P/F ratio were performed. ROX index, HACOR score, APACHE II score, and DTF were able to predict failure of NIRS, and also predicted failure after 48 h of admission, in addition to the P/F ratio. Conclusion ROX index, P/F ratio, DTF, and APACHE II were able to predict failure of NIRS, with APACHE II being the most accurate on admission day. LUS was not significant.