A five-year series on the use of noninvasive ventilation as a weaning tool from invasive ventilation

Data from randomised-controlled trials suggest that use of noninvasive ventilation (NIV) as a weaning tool decreases mortality, length of stay (LOS) and duration of invasive ventilation. We aimed to evaluate if increased use of NIV after high-risk extubation over time was associated with improvement in outcomes in the clinical setting. In this prospective cohort study of all invasively ventilated patients (n=2316) in our medical intensive care unit (ICU) between 2006-2010, we performed time-trend analyses using simple logistic and linear regression to determine temporal changes in post-extubation NIV use, ventilator days, ICU LOS, reintubation rates, and ICU mortality. During this time we had gradually implemented the use of NIV post-extubation in patients with risk factors for extubation failure. The proportion of patients receiving NIV post-extubation increased from 2006-2010: 21.1%, 22.0%, 28.4%, 34.6% and 32.8% respectively (OR 1.20 with each passing year, 95% CI 1.12-1.29). The commonest indication for NIV was high-risk extubation despite a successful spontaneous breathing trial (79.1%). Increased NIV use was associated with temporal decreases in invasive ventilator days and ICU LOS but these were neither statistically nor clinically significant. There were significant associated decreases in overall reintubation rates (OR 0.87 per year, 95% CI 0.76-0.99) and ICU mortality (OR 0.82 per year, 95% CI 0.76-0.87). Although greater use of NIV post-extubation did not reduce ventilator days and ICU LOS in a general medical ICU population, there were associated improvements in reintubation rates and mortality. Confounders included a heterogeneous population and sedation protocols.