Clinical Applications of Closed-Loop Oxygen Control

Abstract

Guidelines recommend titrating oxygen to maintain oxygen saturation (SpO2) within a target range. Clinical studies report that manual oxygen titration processes are inaccurate, exposing patients to risks associated with both inadequate and excessive oxygen. The studies presented in this thesis aim to investigate the effect of an automated closed-loop oxygen control (CLOC) system on the proportion of time within a target SpO2 range across different hospital settings, using high-flow nasal oxygen therapy (HFNOT), non-invasive ventilation (NIV) and continuous positive airway pressure (CPAP) therapies. Two randomised controlled trials (RCTs) investigated CLOC compared with manual titration in patients receiving HFNOT. Clinical settings were an Intensive Care Unit (ICU) following cardiac surgery, and an Emergency Department (ED). Median (interquartile range [IQR]) proportion of time within target SpO2 range for ICU was 97.7% (87.9 to 99.2%) for CLOC (n=28), and 91.3% (77.1 to 96.1%) for manual oxygen control (n=28); mean difference (95%CI) 4.8% (1.6 to 10.3%), p=0.01, favouring CLOC. In ED, the proportion was 96.4% (92.5 to 99.4%) for CLOC (n=25), and 89.9% (69.8 to 97.2%) for manual oxygen control (N=24); mean difference (95% CI) 8.0% (1.7 to 16.9), p=0.01, favouring CLOC. These findings demonstrate that CLOC increases time spent within the target SpO2 range compared with manual oxygen control with HFNOT in both the ICU and ED settings, thereby broadening the clinical settings in which CLOC improves the accuracy of oxygen delivery. Two further RCTs assessed CLOC's ability to achieve the target SpO2 range across HFNOT, NIV, and CPAP therapies. Twelve outpatients with cardio-respiratory disease and chronic hypoxaemia received CLOC with each therapy. SpO2 increased into the target range faster with NIV and CPAP than HFNOT, but time spent within the target SpO2 range across the three therapies was equivalent. In 68 hospitalised participants requiring HFNOT, NIV or CPAP, the median (IQR) proportion of time within the target SpO2 range with CLOC (n=35) was 90.7% (81.9 to 96.8%) compared to manual oxygen control (n=33) 76.3% (64.6 to 87.8%); mean difference 12.64% (5.18 to 18.83%), p<0.001. A non-significant interaction term (P-interaction = 0.94) suggested that this difference did not depend on whether HFNOT, NIV, or CPAP was used. These studies demonstrate that CLOC maintains SpO2 within the target range with similar efficacy using HFNOT, NIV and CPAP therapies. A systematic review and meta-analysis of RCTs comparing CLOC to manual oxygen control in hospitalised adults using low-flow oxygen therapy (LFOT) or HFNOT identified that the mean proportion of time within target SpO2 range was 81.2 to 94.9% for CLOC, and 51.3 to 84.0% for manual oxygen control; mean difference (95% CI) 21.2% (18.5 to 23.8%), p <0.0001 (I2=87.9), favouring CLOC. This study demonstrates that, when oxygen is delivered via low- or high-flow therapy, CLOC improves adherence to a target SpO2 range in hospitalised adults. This series of studies demonstrates that CLOC improves the accuracy of non-invasive oxygen titration compared to usual care across a range of delivery methods and clinical settings, and may represent the optimal non-invasive oxygen titration strategy in acute inpatient care.