Dear Editor,I would like to thank Spindelboek and colleagues [1] for their letter in response to the recent editorial [2] that questions the optimal FiO2 during advance cardiopulmonary resuscitation (CPR).Their observational study raises some challenging hypotheses. However, I would question whether the evidence it provides is sufficient to prevent a prospective randomised control trial of normoxic verses hyperoxic inspired gas mixtures during advanced CPR. Their retrospective analysis of 145 of 1005 (14 %) patients, who they stratified into arterial oxygen tension groups, all received an FiO2 of 1.0 from the earliest possible time point. Thus the stratification is likely to be the consequence of different cardiorespiratory pathologies in the patients. I would conjecture that it is this difference in pathology which has determined outcomes and not the arterial oxygen tension, which in this case acts merely as a surrogate marker. This may also be reflected in differences between the gro ...