Integrating validated scores with clinical parameters to predict NIV weaning outcomes: a prospective study in the respiratory ICU

Early prediction of successful weaning from non-invasive ventilation (NIV) remains challenging in patients with acute respiratory failure. This study evaluated clinical, laboratory, and severity-score predictors of NIV weaning outcomes among patients admitted to the Respiratory Intensive Care Unit (RICU). This prospective observational cross-sectional study included 150 adult patients with acute respiratory failure, admitted to the RICU and required NIV for > 24 h between April 2024 and March 2025. Clinical characteristics, arterial blood gases, laboratory findings, ventilator settings, and APACHE II score were recorded. HACOR score was assessed after one hour of NIV discontinuation. Patients were classified into successful or failed weaning groups. Successful weaning occurred in 99 patients, while 51 experienced weaning failure. Failure was significantly associated with older age (p = 0.001), decompensated cor pulmonale (p = 0.003), lower hemoglobin and serum albumin levels (p = 0.008 and p = 0.005, respectively), higher APACHE II score (18.3 ± 3, p < 0.001), and higher HACOR score (8.1 ± 2.1, p < 0.001). Previous invasive mechanical ventilation ≥ 2 times was also significantly associated with failure (p = 0.001). Multivariate analysis identified age, serum albumin, APACHE II score, HACOR score, and previous invasive ventilation as independent predictors of weaning outcome. Receiver operating characteristic analysis showed good predictive performance for HACOR (AUC 81.6%) and APACHE II (AUC 76.2%) scores (p < 0.001 for both). Older age, poor nutritional status, previous invasive ventilation, high APACHE II score, and HACOR score independently predict NIV weaning failure. The incorporating of HACOR score into NIV weaning protocols may improve clinical decision-making and potentially reduce adverse outcomes associated with delayed recognition of respiratory deterioration.