In healthy humans arterial PO2 is almost constant, showing virtually no oscillations within the respiratory cycle [1]. In Acute Respiratory Distress Syndrome (ARDS) [2], arterial PO2 is thought to oscillate widely due to breath-by-breath variations in pulmonary shunt as a consequence of atelectasis, when alveoli open and collapse during each respiratory cycle. Commercially-available fibre optic oxygen sensors [3] are not suitable for clinical intravascular use, and may not be sufficiently fast to detect these arterial PO2 oscillations at elevated respiratory rate (RR). We have developed an in house, fast (response time ~ 50 ms) fibre optic intravascular sensor [4] (which can be placed into standard arterial cannulae) to detect oxygen oscillations at different simulated RR, and compared the performance of our sensor with a commercially-available sensor (response time ~ 470 ms; Foxy-AL300, Ocean Optics, USA). The sensors were tested in an extracorporeal circuit comprising two standard paediatric oxygenators (Medos-HILITE1000) through which minimally heparinised lambs’ blood was circulated at 39° C via two coupled peristaltic pumps [5]. This provided two independent parallel circuits, in which PO2 was 5 or 50 kPa, and PCO2 was 5 kPa; these values were constantly controlled by a precision gas mixing pump, and monitored through samples for gas analysis before each experiment. A computer-controlled electronic switch opened and closed solenoid valves that diverted the circuits exposing the sensor to PO2 values oscillating between 5 and 50 kPa, simulating arterial PO2 oscillations in ARDS, where severe cyclical shunt exists. The simulated RRs studied were 10, 30 and 50 breaths per minute (bpm). Figure 1 shows results from the PO2 oscillations’ tests in blood. The in house and the AL300 sensors performed comparatively well at a simulated RR of 10 bpm, where they recorded the whole PO2 oscillation amplitude from 5 to 50 kPa. In contrast, only the in house sensor was able to detect the full dynamic range of the oxygen oscillation at simulated RRs of 30 and 50 bpm, where the AL300 sensor detected only about 70% of the actual PO2 oscillation amplitude. Improvements in fibre optic sensing technology and materials may afford the development of an intravascular sensor for real-time detection of rapid oxygen oscillations, and hence cyclical atelectasis, in ARDS. Further studies are needed to validate the use of this fast intravascular technology in vivo and in clinical trials of whether this information can better inform individualised ventilation management in adult and paediatric intensive care units.