preferences suggest that the highest FiO 2 group (Group 5) was not a comparable group to the others, and possibly they were much sicker. The link between hyperoxia and potential harms in tissue-under-threat is mechanistically well established, but this is not necessarily true for high FiO 2 use. 10 Therefore, I believe the implication of high intraoperative FiO 2 use as a cause is currently unjustifiable.