Is the U-shaped curve still of relevance to oxygenation of critically ill patients?

Virtually every drug has an effective and toxic dose and oxygen is no exception. Molecular oxygen has been referred to as “janus-headed”, inasmuch as it is vital for mitochondrial respiration and toxic due to formation of reactive oxygen species (ROS). ROS share the “friend and foe character”, being both toxic and vital for host defence mechanisms [1]. Over 90% of oxygen consumption is used for adenosine triphosphate production via mitochondrial oxidative phosphorylation; however, approximately 1–3% of the oxygen consumption is utilised at complexes I and III of the electron transport chain to generate superoxide anions. Whilst ROS play a vital role in numerous homeostatic mechanisms, in excess these molecules can cause substantial harm to the framework of cells. Hyperoxia increases the rate of superoxide anion production [2] and once our innate antioxidant systems become overwhelmed, oxidative damage occurs. Severe hypoxia can lead to bioenergetic failure and cell death, but paradoxically, moderate hypoxia can also initiate an increase mitochondrial ROS production. Thus, both hyperoxia and hypoxia can initiate oxidative damage and a pro-inflammatory reaction. This is particularly pronounced during ischaemia-reperfusion injury (e.g. resuscitation after cardiac arrest) and/or disturbed cellular oxygen utilisation (e.g. sepsis). Consequently, hyperoxia normally exacerbates ischaemia-reperfusion reperfusion injury. Lung parenchyma is especially vulnerable to oxidative damage and this pulmonary oxygen toxicity usually presents as pneumonitis, eventually leading to haemorrhagic pulmonary oedema. Given this inescapable biological framework, it is vital that we administer the right dose of oxygen (neither too little or too much) to critically ill patients, in order to minimise the harm this potentially lethal drug may cause them.