LAUSR

© 2019 The Egyptian Journal of Chest Diseases and Tu Background Weaning has been classified into simple, difficult, and prolonged. Objective The aim of this study was to assess predictors of prolonged weaning (PW) and survival in the respiratory ICU (RICU). Patients and methods Fifty-three patients admitted to RICU and requiring mechanical ventilation (MV) were included in the study. According to the classification of weaning, patients were categorized into simple [20 (37.7%)], difficult [20 (37.7%)], and PW [13 (24.5%)]. Factors associated with PW and outcomes were assessed. Results Multiple factors were associated with PW: decreased serum Ca (P<0.0001), Mg (P<0.0001), and albumin (P=0.001); increased urea (P=0.001), aspartate aminotransferase (P=0.0005), and alanine aminotransferase (P=0.0001); increased PaCO2 (P=0.04); decreased SaO2% (P<0.0001) at beginning of the first weaning trial; at least 2 comorbidities (P=0.02); increased duration of MV (P<0.0001); and increased ICU stay (P<0.0001). There was no significant association between PW and modes of MV (P=0.2); moreover, alveolar–arterial gradient decreased in patients with PW but was not significant (P=0.57). By multivariate analysis, PW was associated with low serum Mg (odds ratio: 0.03, 95% confidence interval: 0.003–0.24, P=0.001), and SaO2% change (between baseline and first weaning trial) (odds ratio: 0.68, 95% confidence interval: 0.54–1.04, P=0.03). Regarding baseline serum Mg and tidal volume, the cutoff point for predicting PWwas less than or equal to 1.5mg/dl and less than or equal to 6.3ml/kg body weight, respectively, with sensitivity of 61.5 and 77%, respectively, specificity of 92.5 and 55%, respectively, positive predictive value of 72.7 and 36%, respectively, negative predictive value of 88.1 and 88%, respectively, accuracy of 77 and 66%, respectively, and P values of less than 0.0001 and less than 0.02, respectively. The cutoff points of PaCO2 of at least 49 mmHg and SaO2% of less than or equal to 91 at beginning of the first weaning trial and SaO2% change of at least 5 (between baseline and first weaning trial) were strongly associated with PW (P<0.02, <0.0001, and <0.003, respectively). PW was significantly associated with mortality (P<0.0001). Conclusion PW was associated with multiple factors; when subjected to multivariate analysis, only serum Mg and SaO2% change (between baseline and first weaning trial) significantly predict PW. Increased ICU stay and mortality were significantly associated with PW. Clinical implications Prolonged MV is associated with higher frequency of tracheostomy, reintubation, longer ICU stay, and poor outcome. Detection and correction of PW associated factors may improve the outcome.