A snapshot of the oxygenation of mechanically ventilated patients in one Australian intensive care unit.

Hyperoxaemia in patients undergoing mechanical ventilation (MV) has been found to be an independent predictor of worse outcome and in-hospital mortality in some conditions. Data suggests that a fraction of inspired oxygen (FiO2) of 0.4 or lower may produce hyperoxaemia although it is commonly accepted without adjustment in ventilator settings. The primary aim of this study was to observe current practice at one Australian tertiary intensive care unit (ICU) with regard to prescription and titration of oxygen (O2) in patients undergoing MV, in particular whether they received higher FiO2 than required according to arterial blood gas (ABG) results, and whether there was FiO2 titration as a response to initial ABG results during the 12 hours following. A retrospective observational study of 151 ICU patients undergoing MV between November 2013 and February 2014 was conducted, with ABGs as the primary outcome measure. There were 250 ABG measures, with mean FiO2 0.38 (range 0.3-1.0) and mean PaO2 114 mmHg (standard deviation 36). Over all observations, 197 (79%) were of FiO2 ≤0.4, however no patients were ventilated on room air (FiO2 0.21) and 114 (46%) were in the hyperoxaemic range. Oxygen titration (up or down) occurred in 31% of patients. Morning ABGs were taken at a time suggested by ICU guidelines, and on review of these measures, the mean FiO2 was lower than that purported to create toxicity. Subsequently, almost one-third of the cohort had their FiO2 titrated, however there was a floor effect whereby 39%-43% of the cohort received an FiO2 of 0.3.