Guidelines for newborn resuscitation have been invaluable in disseminating knowledge, enhancing skills and promoting research. Most guidance is built on the consensus on science and treatment recommendations of the International Liaison Committee on Resuscitation (ILCOR).1 This group will publish new guidance later this year. The ILCOR evidence evaluation has been a strong stimulus for research and a number of studies related to newborn resuscitation have been published in Archives of Disease in Childhood since the 2005 statement. The difficulties of conducting research on newly born infants have dictated that much of the available data comes from studies performed on manikins.

Schmölzer et al2 present their experience using a respiratory function monitor to assess tidal volume and mask leak during resuscitation of 20 preterm infants. They should be congratulated for producing some of the first respiratory data from newborn infants since the 1980s.3,–,7 Data were gathered from the onset of mask ventilation, then after 1 min the resuscitator (whose experience ranged from 4 months to 12 years) was asked to estimate mask leak, assess chest wall movement and estimate the tidal volume they were delivering. The authors regarded 4–8 ml/kg as an appropriate tidal volume, in keeping with observations from spontaneously breathing infants.8

When respondents considered themselves to be delivering appropriate chest wall movement, the majority of babies were being over-inflated, with a median tidal volume of >9 ml/kg (sometimes massively greater). Even when chest wall rise was considered to be absent, more than 50% of the infants were receiving a tidal volume in excess of 8 ml/kg. This suggests that visual assessment of chest wall rise during resuscitation is practically useless as a means of avoiding over-inflation or assessing the adequacy of ventilation. A large proportion of preterm infants are likely to be receiving initial ventilation that is clearly damaging in immature newborn animals.9,–,11 Present methods for confirming that ventilation is effective, such as a rise in heart rate and/or the detection of exhaled CO₂, provide no insurance against over-inflation. When added to observations from the same group regarding the weakness of clinical assessments of colour12 and heart rate13 during neonatal resuscitation, it is difficult to escape the conclusion that it is time to extend the monitoring technologies employed in the neonatal intensive care unit to the delivery room if we are really serious about optimising outcomes.

As with measurements of saturation and heart rate, we should be careful not to prejudge the most appropriate way to use this sort of information. There is a danger that un-researched targets may lead to inappropriate intervention. While avoiding over-inflation seems a worthwhile aim, it is also important to recognise that immature lungs recruit gradually over several minutes.14 Initial tidal volumes are delivered mainly to the airways and even apparently normal tidal volumes may be excessive if delivered in full from the start before ventilation can be distributed evenly.15

The infants in this study were ventilated either with a pressure-limited T-piece device or with a self-inflating bag and all were mask ventilated. There were quite substantial and variable leaks around the mask and judgements of mask leak were similarly unreliable. This confirms earlier findings in manikin studies.16 17 Again, a respiratory function monitor might provide useful feedback in this regard, but relationships between leaks, inflation pressures and tidal volumes were inconsistent. Leak is greater with continuous flow devices than with self-inflating bags.18 Some leak may be caused by a mask technique that obstructs the upper airway.19 Low tidal volume does not necessarily infer the need for increased inflation pressure. The authors speculate that leaks around the masks may be protective in reducing the risk of excessive inflation volumes. Certainly, manikin studies show that very high inflation pressures are generated with self-inflating bags in intubated models when there is no leak.20 Despite inconsistent leaks in this study, inspection of figure 4 suggests an important relationship between inflation pressure and tidal volume such that if avoiding excessive tidal volumes during initial stabilisation is an important goal, limiting inflation pressure is likely to be an important component. Other groups have found that most preterm infants can be stabilised with inflation pressures of less than 30 cm H₂O.21 22

The question is, can we use this technology to do better? This important work suggests that we can. It is still possible to demonstrate improved outcomes with different delivery room practices.23 In this paper, Schmölzer et al show that the people who thought that they were getting it right were delivering the highest tidal volumes. A blunderbuss approach to initial ventilation may be a bigger problem than is presently recognised. Continuous display of tidal volume information has been shown to help optimise manual ventilation in bench studies.24

The difficulties in conducting research in the delivery room are significant. Numbers from any one centre are likely to be small. Vyas et al5 attended 250 deliveries to obtain data on the nine term infants who required lung inflation; one must wonder whether they would have been able to conduct that work had they been required to obtain written, informed consent prior to participation. For the current study, Schmölzer et al attended the delivery of 59 preterm infants to obtain data from 20. If we are to progress our knowledge in newborn delivery room management, we must work together and argue the case for recruitment utilising a waiver of informed consent where circumstances prevent discussion in advance.25 Newborn resuscitation research is essential and meets all the international criteria for such a waiver.26,–,29