Higher versus lower oxygenation targets in COVID-19 patients with severe hypoxaemia (HOT-COVID) trial: Protocol for a secondary Bayesian analysis

Background: Respiratory failure is the main cause of mortality and morbidity among ICU patients with coronavirus disease 2019 (COVID-19). In these patients, supplemental oxygen therapy is essential, but there is limited evidence the optimal target. To address this, the ongoing handling oxygenation targets in COVID-19 (HOT-COVID) trial was initiated to investigate the effect of a lower oxygenation target (partial pressure of arterial oxygen (PaO₂) of 8 kPa) versus a higher oxygenation target (PaO₂ of 12 kPa) in the ICU on clinical outcome in patients with COVID-19 and hypoxaemia. Methods: The HOT-COVID is planned to enrol 780 patients. This paper presents the protocol and statistical analysis plan for the conduct of a secondary Bayesian analysis of the primary outcome of HOT-COVID being days alive without life-support at 90 days and the secondary outcome 90-day all-cause mortality. Furthermore, both outcomes will be investigated for the presence heterogeneity of treatment effects based on four baseline parameters being sequential organ failure assessment score, PaO₂/fraction of inspired oxygen ratio, highest dose of norepinephrine during the 24 h before randomisation, and plasma concentration of lactate at randomisation. Conclusion: The results of this pre-planned secondary Bayesian analysis will complement the primary frequentist analysis of the HOT-COVID trial and may facilitate a more nuanced interpretation of the trial results.