A55-10 A Descriptive Analysis of Pressure Support Ventilation in Adult Intensive Care Patients

Abstract Rationale Pressure support ventilation (PSV) is the dominant ventilatory mode for weaning but it is poorly understood. Previous analyses using twice-daily ventilator data demonstrated half of patients receiving PSV were exposed to high tidal volumes (8ml/kg) and two-thirds to over-assistance (RSBI 37 breaths/min/L). Additionally, those ventilated with PS 12 had greater ventilator asynchrony. We analysed granular intensive care (ICU) data, measured hourly, to describe PSV practice at three UK centres, between 2014-2025. We hypothesised that hourly data more accurately reflect real-world PSV practice and aimed to demonstrate clinical dynamics masked by twice-daily data. Methods Patients aged over 18 years who received mechanical ventilation for any indication were included. Input data included demographics, ventilatory, blood gas, laboratory, vital signs and vasopressor use. Derived variables including rapid shallow breathing index (RSBI) were calculated. Ventilation modes were categorised into controlled, combined (e.g., synchronised intermittent mandatory ventilation) and assisted modes, which included PSV variations. Over-assistance was defined by respiratory rate 17 or RSBI 37 breaths/min/L. Comparison between groups was performed using the Kruskall-Wallis (continuous) or Chi-squared (categorical variables) tests. Results Of 9057 invasive ventilated patients, median age was 65 IQR 52-77, 66.3% were male and median ICU length of stay was 7 3-16 days. Of total time invasively ventilated, 33.8% was spent in assisted mode. The most common pressure support (PS) levels were 5cmH2O (21.0%) and 10cmH2O (19.4%). Coefficient of variance plot of PS showed stability throughout assisted ventilation in most cases. Mean tidal volume during PSV was 6.9ml/kg 5.8-8.2ml/kg actual body weight (ABW), compared with 6.5ml/kg 5.7-7.5ml/kg in controlled and 6.3ml/kg 5.5-7.2ml/kg in combined modes, respectively (p 0.001). PSV delivered a tidal volume 8ml/kg ABW in 22.3%, and PS was changed in only 6.1% of these instances in the next hour. Over-assistance was found in 34.2% of PSV by RSBI and 27.3% by respiratory rate criteria, with PS changed in 5.7% and 5.8% of cases respectively, in the next hour. Patients with PS 12 for more than 10% of assisted ventilation had a relative risk of extubation success of 0.87 95% CI 0.80-0.94, Figure 1. Conclusions Patients spend a third of total time invasively ventilated in assisted modes, with over a fifth receiving a tidal volume 8ml/kg and a third over-supported by RSBI criteria. PS levels were rarely changed, even in patients with high tidal volumes or over-support. Our results build a case for more responsive management of PSV and should be externally validated. This abstract is funded by: Asthma and Lung UK/ HCA Healthcare