Individualised pulse oximetry limits in neonatal intensive care
Editor—Gupta et al 1 have successfully demonstrated that they were unable to accurately predict PaO 2 from saturation monitoring even after standardising from a previous measurement. The rest of the conclusions presented in their discussion are however based upon interpretation of other research findings which is not further supported by their own study.
They correctly point out that the poor relation between SpO 2 and PaO 2 is related to differing proportions of fetal haemoglobin, Pco 2, and acid–base balance. This begs a question, which their discussion fails to address, on whether the PaO 2 or the SpO2 is the most useful index of oxygenation. It is certainly the case that normal in utero PO 2 is within a range which they would describe as “hypoxic”. On the other hand, in the presence of 100% fetal haemoglobin, a saturation monitor should, in these circumstances, correctly indicate adequate saturation.
The authors also remark that transcutaneous oxygen monitoring is “a better way of non-invasively assessing PaO 2”. They provide no evidence for this remark. It is certainly a common experience to find a saturation monitor alarming high when a transcutaneous monitor is apparently recording a normal or even low PO 2 because of an undetected poor contact. It is also not the case that transcutaneous oxygen monitoring, particularly on the extremely premature infant, is entirely “non-invasive”.
In the long run, the purpose of oxygen monitoring is to detect degrees of hypoxia which are likely to cause acidosis or tissue damage and levels of hyperoxia which may risk retinopathy of prematurity. To date there would appear to be no study comparing different measurement methods with respect to these outcomes. However, the authors' own discussion of the reasons for the poor correlation between SpO2 and PaO 2 provides an excellent theoretical argument in favour of the former over the latter!
Dr Yoxall and Dr Shaw respond: Dr Clifford has raised the important question of whether PaO 2 or SpO2 is the best index of arterial oxygenation. The answer to this question is, of course, unknown and the aim of our study was not to attempt to provide an answer. Most of the work defining hypoxia and harmful hyperoxia in neonates was performed in the era prior to pulse oximetry and therefore defines these situations in terms of partial pressure rather than oxyhaemoglobin saturation. The guidelines for good practice in the management of neonatal respiratory distress syndrome published by BAPM and RCP state that arterial blood sampling is the “gold standard” for assessing arterial oxygenation.1-1 In the absence of evidence to the contrary, we would agree with this.
Pulse oximetry is very widely used during neonatal intensive care. Our study has shown that it is not possible to accurately predict PaO 2 from SpO2 even after standardising from a previous measurement. The information provided by measurements of PaO 2 and SpO2 is different and should not be interpreted as interchangeable. As SpO2 monitoring is non-invasive, semicontinuous, and has a rapid response time, it useful for monitoring trends and particularly in detecting episodes of sudden deoxygenation.
Monitoring of PaO2 is possible using transcutaneous monitoring and this is a better way of continuously monitoring PaO2 than trying to extrapolate from the SpO2 signal. The ease with which pulse oximetry can be applied has led to the virtual abandonment of transcutaneous monitoring by many units. There are technical difficulties with transcutaneous monitoring as pointed out by Dr Clifford, but these can be overcome if the staff caring for the babies are familiar with the technique and use it on a routine basis. The anecdotal experiences of those units that continue to use transcutaneous monitoring as part of their clinical routine is that it remains an extremely useful technique.
The purpose of maintaining adequate arterial oxygenation is to provide oxygen to meet the metabolic demands of the baby, prevent pulmonary hypertension, and avoid oxygen toxicity. We do not have a satisfactory method of defining what constitutes adequate oxygenation during intensive care at present and it is necessary to take into account other variables, such as haemoglobin concentration or tissue perfusion, which also determine oxygen delivery when we assess our patients.
