We often wonder if we should place an arterial line before transporting a patient by air, so we were very interested to read the study by McMahon et al. 1 that found poor agreement between invasive and non-invasive mean and systolic arterial blood pressure readings, but with no clear diminution of accuracy during flight compared with measurements taken before transport. Perhaps, as could be inferred from the results, aeromedical transport per se is not an indication for invasive arterial pressure monitoring. However, we should like more details before being sure of this conclusion. Two hundred and fifty-one patients were eligible for the study by virtue of having arterial catheters in place. Of these, only 56 (22%) were included in the analysis. Patients were excluded if simultaneous measurements of intra-arterial (IABP) and non-invasive (NIBP) blood pressure were not possible for the duration of the transfer. Therefore, 195 patients presumably had one or other monitor ‘fail’. Knowing more about these patients and the modes of such failures would be very helpful. Could it be that NIBP measurement simply does not work in flight in more than 75% of patients? Secondly, the patients were transported by either fixed or rotary wing platforms. Vibration in each platform type is quite different, so it would be useful to analyse the data stratified by platform type. Even if McMahon's dataset is too small for definitive conclusions, this would facilitate a better subsequent study. It would be useful to know a little more of the statistical analysis that concluded no significant difference between pre-flight and in-flight readings. The data presented in Fig. 4 of the paper seem to show that in-flight NIBP readings were somewhat more random. For example, was there a difference in the proportion of readings that exceeded a clinically unacceptable threshold? It was also unclear how unwell these patients were. This might easily be analysed by reporting how many patients were receiving vasoactive medication. Vasopressors are known to affect the accuracy of radial vs femoral invasive pressures 2, 3. The authors have assumed that IABP is the ‘gold standard’ measurement, which is not unreasonable. In many parts of the world, invasive pressure monitoring is standard-of-care for patients requiring inotrope infusions, so the comparison of IABP and NIBP in transport becomes moot. Even if we still have little idea of optimal blood pressure targets in critical illness 4, the reference point by which this question might be answered will be IABP. However, there is some evidence that clinicians can recognise a poorly-representative arterial pressure waveform, and when they supplement their therapeutic judgments with non-invasive measurements, their patients receive fewer blood transfusions, vasopressor infusions and antihypertensive medications, without worse outcomes 5. Finally, it was unclear where the arterial catheters were sited. Radial artery catheters commonly over- or under-read femoral arterial pressure 2, 3 and if McMahon's patients had a mixture of catheter sites the variable correlation with NIBP they observed is less surprising. In summary, we think Mc Mahon et al. present an important question with immediate practical implications, but an incomplete answer suggesting the need either for further analysis of their data or a larger study with prespecified stratification.