Article Text
Abstract
Aim The aim was to compare resuscitators' estimates of tidal volume (VT) and face mask leak with measured values during positive pressure ventilation (PPV) of newborn infants in the delivery room.
Patients and methods The authors measured inflating pressures and VT delivered using a respiratory function monitor, and calculated face mask leak. After 60 s of PPV, resuscitators were asked to estimate VT and face mask leak. These estimates were compared with measurements taken during the previous 30 s.
Results The authors studied 20 infants who received a mean (SD) of 21 (6) inflations during the 30 s. The median (IQR) expired tidal volume (VTe) delivered was 8.7 ml/kg (5.3–11.3). VTe varied widely during each resuscitation and between resuscitators. Five resuscitators could not estimate VTe, one overestimated and 14 underestimated their median delivered VTe. The median (IQR) face mask leak was 29% (16–63%). Leak also varied widely during each resuscitation and between resuscitators. One resuscitator could not estimate mask leak, four overestimated leak and 15 underestimated leak.
Conclusion During face mask ventilation in the delivery room, VT and face mask leak were large and variable. The resuscitators were unable to accurately assess their face mask leak or delivered VT.
Statistics from Altmetric.com
Introduction
Approximately 3–6% of all newborn infants receive some form of respiratory support at birth.1 Positive pressure ventilation (PPV) is commonly used in the delivery room and is a cornerstone of respiratory support after birth.2 3 An international consensus statement recommends that infants with inadequate breathing or bradycardia be given PPV via a face mask with a self-inflating bag, flow-inflating bag or T-piece device.4 5 The purpose of PPV is to establish a functional residual capacity and deliver an appropriate tidal volume (VT) to achieve effective gas exchange.6 Adequacy of ventilation is then judged by assessing the heart rate.4 5 However, if the heart rate does not increase, chest wall movements should be assessed to gauge adequacy of ventilation.4 5 The resuscitator judges the delivered VT by observing chest wall movements during PPV.4 5 The VTs are not routinely measured.
A human observational study reported a mean VT of 6.5 ml/kg in spontaneous breathing preterm infants in the first minutes of life.7 When assisted ventilation is required, a peak inflating pressure (PIP) is chosen with the assumption that this will deliver an appropriate VT.4 5 However, lung compliance and therefore the PIP required to deliver an appropriate VT vary in the minutes after birth.8 9 It is likely that there are even greater differences between infants as the mechanical properties of the lung vary with gestational age and disease states.6 In addition, many infants breathe during PPV adding to the inconsistency of VT delivered with a set PIP.10 11 Therefore, relying on a fixed PIP and subjective assessment of chest wall movement may result in either under- or over-ventilation.10 12 Animal studies have shown that PPV with VT >8 ml/kg or inflations with large VTs can damage the lungs.13,–,16
What is already known on this topic
▶ Assessment of mask ventilation relies on a clinical impression of ‘adequate chest rise’ and increased heart rate.
▶ In manikin studies resuscitators had large and unrecognised leaks around the face mask during positive pressure ventilation.
▶ Lung injury can occur if tidal volumes are greater than 8 ml/kg.
What this study adds
▶ Assessment of face mask leak and delivered VT is inaccurate.
▶ Face mask leak varies between resuscitators and between inflations, while delivered VTs also vary considerably and bear little relationship to peak inflation pressure.
▶ A respiratory function monitor might help improve face mask leak detection and optimise VT delivery.
Face mask seal is an important determinant of successful PPV. Manikin studies have shown that resuscitators were unaware of large and variable leaks between the face and mask during PPV.17,–,19 Large leaks may lead to inadequate VTs during PPV in the delivery room and may result in ineffective ventilation.
We hypothesised that resuscitators are unable to accurately assess the delivered VT and face mask leak during PPV. The aims of this study were as follows: (1) to measure VT and face mask leak during PPV in the delivery room and (2) to compare these measurements with resuscitators' estimates of these parameters.
Patients and methods
We conducted this observational study at The Royal Women's Hospital, Melbourne, Australia, a tertiary perinatal centre where approximately 6000 infants are delivered and more than 100 infants <1000 g birth weight are admitted to the neonatal intensive care unit annually. Infants were enrolled from February 2008 to November 2008. This study was approved by The Royal Women's Hospital Research and Ethics Committees and written parental consent was obtained.
Infants of <32 weeks' gestation were eligible for this study if the clinicians judged they had inadequate breathing immediately after birth and so provided face mask PPV. Infants with a congenital abnormality or whose parents did not consent were excluded. When the research team were available they attended deliveries to make recordings. The research team did not participate in the clinical care of the infant.
Face masks
All infants received PPV via a round silicone size 00 face mask (Laerdal, Stavanger, Norway).
Ventilation devices
Infants received mask PPV with either a T-piece device (Neopuff Infant Resuscitator; Fisher & Paykel Healthcare, Auckland, New Zealand) or a self-inflating bag (Laerdal). Both devices are commonly used in our hospital. The T-piece is a continuous flow, pressure-limited device with a built-in manometer and a positive end expiratory pressure (PEEP) valve. The default settings used were a gas flow of 8 l/min, PIP of 30 cm H2O and PEEP of 5 cm H2O. The 240 ml silicone self-inflating bag was used with an 8 l/min gas flow and without a manometer or a PEEP valve.
Respiratory function monitor
A Florian respiratory function monitor (RFM) (Acutronic Medical Systems, Zug, Switzerland) was used to measure inflating pressures and gas flow. The monitor measured pressure directly from the circuit and measured gas flow with a hot wire anemometer flow sensor. The flow sensor was placed between the ventilation device and the face mask. The monitor automatically calculates the VT passing through the sensor by integrating the flow signal.
Data acquisition and analysis
The gas flow, VTs and airway pressure data measured with the RFM were recorded at 200 Hz using a laptop computer with Spectra software (Grove Medical, London, UK), a program specifically designed for recording physiological data. In the delivery room, neither the RFM nor the computer screen were visible to the resuscitators and the monitor's alarm was disabled.
We examined every inflation given to each infant and calculated the leak from the mask by expressing the volume of gas that did not return through the flow sensor during expiration as a percentage of the volume that passed through the flow sensor during inflation19:
Resuscitators' assessment of VT and leak
If PPV was performed for at least 60 s the resuscitator was asked four questions while they continued PPV:
Do you think you have a leak around the face mask?
If yes, estimate the percentage of your leak.
How well do you think the chest is moving? (‘appropriately’, ‘too low’, ‘too much’, ‘not at all’, ‘not sure’).
Estimate the VT you are delivering in ml/kg.
All answers could include: ‘don’t know', ‘cannot do it’ or ‘not sure’.
Questions were asked when the resuscitator was comfortable with the progress of the resuscitation. Resuscitators were given the opportunity to decline to participate in the study but all were willing to contribute. All resuscitators had received neonatal resuscitation training prior to participation in this study and were familiar with both ventilation devices
Statistical analysis
The waveforms of pressure, flow and VT were analysed and the VT and mask leak for each inflation were measured for the 30 s prior to the questioning. The face mask leak was retrospectively corrected for body temperature, pressure and water vapour saturation.20 The mean/median values obtained were compared to the estimates of the resuscitator. Results are presented as mean (SD) for normally distributed continuous variables and median (IQR) for variables with a skewed distribution. Data were analysed using STATA v 10 (STATA, College Station, Texas, USA).
Results
A total of 163 eligible infants were born between February and November 2008. A delay in registering recording equipment after relocation of the hospital meant that 69 infants were not studied. The research team was not notified of the impending delivery of 35 infants. A further 34 did not receive PPV. We therefore recorded 25 infants who received PPV in the delivery room. The recording was of poor quality in five cases. The characteristics of the remaining 20 infants are presented in table 1. Fifteen infants were ventilated with a T-piece and five with a self-inflating bag.
A total of 419 inflations were analysed, with a mean (SD) of 21 (6) inflations per infant during 30 s. PPV was started at a median (IQR) of 91 (44–122) s after birth. The need for PPV was determined by the resuscitator. The median (range) level of neonatal resuscitation experience was 31 (4–144) months. After answering the questions the resuscitator was told the current expired tidal volume (VTe) and face mask leak.
PIP and PEEP
Mean (SD) PIP was 26.3 (8.8) cm H2O. The mean (SD) PEEPs with the Neopuff and Laerdal self-inflating bag were 4.0 (2.9) cm H2O and 0.6 (0.3) cm H2O, respectively.
Expiratory VT
The median (IQR) expiratory VT (VTe) was 8.7 (5.3–11.3) ml/kg. The VT varied widely between inflations and resuscitators (figure 1A). Five resuscitators could not estimate the VT, one overestimated and the remaining 14 underestimated the median VT they delivered (figure 1B).
Face mask leak
The median (IQR) face mask leak, as a percentage of the inspired VT, was 29% (16–63%) (figure 2A). The face mask leak varied widely between inflations and resuscitators. One resuscitator could not estimate the leak, one correctly estimated the leak and four overestimated the leak, while the remaining 14 underestimated it (figure 2B).
Correlation between peak pressure and expired VT
There was only a weak relationship between each PIP and VT. For example at a PIP of approximately 30 cm H2O, the VTe ranged from 0 to 30 ml/kg (figure 3).
Judgement of chest wall movements
Comparison between resuscitators' judgement of chest rise and delivered VTe showed a median (IQR) delivered VTe for ‘appropriate’ chest rise of 9.6 (7.1–12.9) ml/kg, ‘too low’ of 2.4 (1.5–7.6) ml/kg, ‘not at all’ of 8.3 (1.3–9.9) ml/kg and ‘not sure’ of 8.6 (7.6–10.5) ml/kg. Resuscitators who judged chest rise as ‘too low’ more accurately assessed the chest rise (figure 4).
Discussion
Resuscitation of an infant born at <32 weeks' gestation is an emotional and stressful event, even for experienced neonatologists. Decisions have to be made quickly and resuscitators have to be skilled in clinical assessment and resuscitation. However, observational studies have shown the clinical assessment of colour and heart rate is unrealiable.21 22 The accuracy of clinical assessment of adequacy of ventilation (ie, chest rise) is unknown.
To our knowledge this is the first study which compares clinical assessments with measurements of pressure, VT and face mask leak in newly born infants during PPV in the delivery room. We found resuscitators delivered highly variable VTs with both ventilation devices, from almost 0 to 31 ml/kg. In our study the median delivered VTe was 8.7 ml/kg. VTs of this order or above might be excessive and cause lung damage.13,–,16 The assessment of chest rise by the resuscitator was an unreliable estimate of VTe.
We found airway pressure was a poor proxy for the delivered VTe. Often the desired PIP was not achieved because of large leaks. This has also has been shown in manikin studies.19 Clinicians should not assume that effective ventilation is occurring when the desired pressures are achieved.23,–,26 Measurement of VTe during PPV may be more useful than relying on an inflating pressure or clinical judgment to ensure that over- or under-ventilation is not occurring.18
Face mask leak is both a common and an unrecognised problem during PPV which can lead to failure of ventilation.17 24 26 We found that face mask leak varied widely from almost 0 to 100%. In addition, resuscitators were unable to accurately self-assess their leak. These findings are consistent with those of manikin studies during PPV.17 24 In manikin studies, a RFM allowed resuscitators to adjust the face mask position during PPV and reduce leak by more than half.18
We speculate that face mask leak during PPV might be beneficial and if there was no leak the VTe might have been even higher. The problem, as we have shown, is that the VT varies and is unknown. A display showing PIP and VT allows the resuscitator to alter face mask position to diminish face mask leak and also adjust the PIP to optimise VT delivery.
This study has several limitations. Although the number of enrolled infants was small, we found a wide variation in VTe and leak during PPV. One could argue that a median of 31 months' resuscitation experience caused this wide variation. However, in manikin studies operators had wide variations in VTe delivery and face mask leak regardless of training.17 25
Conclusion
During mask ventilation in the delivery room, leak and VT are very variable and correlated weakly with the PIP. On average the VTs were high at 8.7 ml/kg and in a range that might injure the lungs.
The measured face mask leak and delivered VT correlated poorly with the estimated values of each operator. Resuscitators were unable to accurately assess their face mask leak or VT.
Using a respiratory monitor in the delivery room to continuously measure and display face mask leak and VT might improve the effectiveness of neonatal resuscitation.
Acknowledgments
The authors would like to thank Kevin I Wheeler for his helpful reviews of the manuscript and Connie Wong for her help during patient recruitment.
References
Footnotes
-
Funding GMS and JAD are supported in part by a Royal Women's Hospital Postgraduate Research Degree Scholarship. GSM is supported in part by a Monash International Postgraduate Research Scholarship. PGD is supported in part by an Australian National Health and Medical Research Council Practitioner Fellowship. This study was supported by Australian National Health and Medical Research Council Program Grant No. 384100.
-
Competing interests None.
-
Ethics approval This study was conducted with the approval of The Royal Women's Hospital Research and Ethics Committees, Melbourne, Australia.
-
Provenance and peer review Not commissioned; externally peer reviewed.