Article Text

Download PDFPDF

Mask leak in one-person mask ventilation compared to two-person in newborn infant manikin study
  1. M B Tracy1,2,
  2. J Klimek1,2,
  3. H Coughtrey1,2,
  4. V Shingde1,2,
  5. G Ponnampalam1,
  6. M Hinder1,
  7. R Maheshwari1,
  8. S K Tracy2,3
  1. 1Nepean Hospital Sydney West Area Health Service, Sydney, Australia
  2. 2University of Sydney, Sydney, Australia
  3. 3Royal Hospital for Women, Sydney, Australia
  1. Correspondence to Dr Mark Tracy, Nepean Neonatal Intensive Care Unit, Nepean Hospital, Derby Street Kingswood 2747, PO Box 63, Penrith 2751 Sydney, Australia; mark.tracy{at}


Aim To compare a new two-person method (four hands) of delivering mask ventilation with a standard one-person method using the Laerdal self-inflating bag (SIB) and the Neopuff (NP) infant resuscitator in a manikin model.

Background Recent studies of simulated neonatal resuscitation using bag and mask ventilation techniques have shown facemask leak levels of 55–57% in expert hands.

Methods 48 participants were randomly paired and instructed to give mask ventilation for a 2-min period as single-person resuscitators, then as two-person paired resuscitators at set pressures for NP and set parameters for SIB. Airway pressure, flow, inspiratory tidal volume, expiratory tidal volume and mask leak were recorded.

Results A total of 21 578 inflations were recorded and analysed. For SIB, mask leak was greater (11.5%) with single-person compared to two-person (5.4%; mean difference 6.1%, 95% CI 1.5 to 10.7, p<0.01). For NP, mask leak was greater for single-person (22.2%) compared to two-person (9.1%; mean difference 13.1% 95% CI 3.6 to 22.6, p<0.01). For single-person mask ventilation, mask leak was greater with NP (22.2%) compared to SIB (11.5%; mean difference 10.7%, 95% CI 1.4 to 19.7, p<0.01). For two-person mask ventilation, mask leak was greater for NP (9.1%) compared to SIB (5.4%; mean difference 3.7%, 95% CI 0.1 to 6.4, p<0.05).

Conclusions Two-person mask ventilation technique reduces mask leak by approximately 50% compared to the standard one-person mask ventilation method. NP mask ventilation has higher mask leak than Laerdal SIB for both single- and two-person technique mask ventilation.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Newborn resuscitation is a frequent (5–16% of all newborns)1 2 and life-saving procedure. The critical factor in a successful resuscitation is adequate ventilation.3 The need for progressing resuscitation beyond ventilation to adding cardiac compressions and epinephrine use is infrequent (<0.2% of births) and may signal a failure of adequate ventilation technique in as many as 50% of these resuscitations.4 Despite the obvious need for evidence-based practice leading to training and implementation, there remains little randomised controlled evidence for device type and method of training.5

Recent studies investigating mask leak in modified newborn manikin models have shown concerning levels of mask leak (>50%) in expert hands.6,,8 Intensive test, retraining and retest improved mask leak to only 32–33% in the same group.8 Mask leak is unintended, not recognised by the person undertaking the resuscitation and is likely to be highly variable. Breath by breath variation in mask leak may result in under-inflation to over-inflation with the potential for barotrauma and/or volutrauma particularly in the preterm infant. Preterm animal studies have shown excessive tidal volumes (TVs) to rapidly injure lungs.9 10 Finer et al11 have recently described airway obstruction in the majority of very low birth-weight (VLBW) babies studied who received face mask ventilation during resuscitation at birth with a median of 14 consecutive mask ventilation breaths obstructed. Mask leak combined with the potential of airways obstruction in VLBW babies potentially further complicates effective resuscitation. Resuscitators have difficulty achieving adequate ventilation in a test lung model with changing compliance, detecting and adjusting ventilation using airways pressure alone as a guide.12

What is already known on this topic

  • Neonatal manikin studies have shown consistnely large and widely variable mask leak.

  • Adult resuscitation guidelines suggest a two person four handed method of mask ventilation is superior to one person providing mask ventilation.

What this study adds

  • This study shows significant reduction in mask leak using a two person four handed mask ventilation method.

  • The two handed four finger rim pincer grip reduces mask leak by approximately 50% compared to a single operator holding the mask and providing the ventilation.

Teaching resuscitation and sustaining proficiency in these skills is a vexed issue, particularly in low-volume birthing units. A recent comprehensive survey of all birthing units in the state of New South Wales and Australian Capital Territory revealed that the majority of births occurred in rural or urban non-tertiary centres where one third of health personnel felt inadequately trained.13

In our series of studies in mask ventilation (paper 1), we measured delivered pressures and TV comparing four mask ventilation devices: Laerdal 240-ml self-inflating bag (SIB), a 500-ml flow-inflating anaesthetic bag (vital signs), a t-piece resuscitator (Neopuff (NP); Fisher & Paykel Health Care, Auckland, New Zealand) and ventilator-delivered mask ventilation (VDMV). In that study, we found that the SIB and flow-inflating bags delivered potentially dangerously elevated TV and minute ventilation (MV) compared to that measured with NP and VDMV. That study also demonstrated two further interesting findings. First, a large coefficient of variation was seen in peak inspiratory pressure (PIP), positive end-expiratory pressure (PEEP) and TV, all of these being statistically smallest with the VDMV mode. Second, although that study was not designed to study mask leak, we noted that the mask leak with VDMV using the Dräger Babylog 8000 plus ventilator (Dräger, Lubeck, Germany) was consistently <10% and frequently zero. As the stiff ventilator circuit in the VDMV mode did not require support by the resuscitator, the natural mask hold was that of a two-handed four-finger rim grip. We hypothesised that a two-person, four-handed, bag and mask method may reduce mask leak.

This hypothesis is supported by adult resuscitation studies in the 1980s which determined four-handed two-person mask ventilation to be superior in delivering larger and appropriate PIP and MV in manikin models.14 15 Adult resuscitation guidelines recommend two-person, four-handed, bag–mask ventilation recognising that single-operator technique is more difficult.16 There is one report by Davidovic et al comparing one- versus two-person bag and mask ventilation with older infant and child-sized manikins showing superiority with the two-person method.17 Current newborn and paediatric guidelines make no mention of two-person bag and mask ventilation.18 19 Two current international newborn and paediatric guidelines make no mention of two-person bag and mask ventilation. The UK's Resuscitation Council Newborn Life Support Manual does describe a two-person method as superior with the mask hold utilising jaw thrust and either thumbs only pressing down on the upper flat ridge of Laerdal mask or thumb and first finger on the flat ridge.20

The aim of this study was to compare mask leak and delivered ventilation between single-person and two-person (four-handed) mask ventilation, using two common devices: the 240-ml Laerdal SIB and the t-piece NP on a modified Laerdal Advanced Life Support Training (ALST) baby manikin (Laerdal Medical, Oakleigh, Australia).

Materials and methods


Staff of a newborn intensive care nursery at a major metropolitan teaching hospital (Nepean Hospital) in Western Sydney were invited to participate. All staff tested had previously had extensive training in neonatal resuscitation and demonstrated proficiency annually in the locally run neonatal intensive care unit resuscitation course. This course uses the American Academy of Pediatrics neonatal resuscitation program.21 The mask hold taught for single person is the two-point top hold as described by Wood et al.8 The staff were classified as senior medical (consultants, n=5), junior medical (fellows and registrars, n=11), senior neonatal nurse (n=5) and neonatal intensive care nurse (n=27).

Ventilation devices

The Laerdal Infant Silicone device is a 240-ml SIB with a 35-cm H2O pressure-release valve (Laerdal Medical). The NP t-piece resuscitator (Fisher & Paykel Health Care) is a gas-pressure-driven device set at 10 l/min with a rotating valve to select PEEP and finger occlusion to deliver a selected PIP. A Laerdal round 0/1 size mask (Laerdal Medical) was used with both devices.

The model

We used the Laerdal ALST baby manikin model in this study. This device has dual lung “bags” connected to the “trachea” and an additional bag connected to an “esophagus” to simulate the leak into the stomach if there is incorrect head positioning. The Laerdal ALST manikin has a closed system with lung and stomach bags with no intended leak. The Laerdal ALST manikin has a hinged mandible allowing a realistic jaw thrust.22 The Laerdal ALST baby manikin was free of leak internally, and the structure of “esophageal” tube and “stomach” bag were specifically left intact to allow findings of this study to be easily generalised to the commercial Laerdal manikin. The static compliance of the model was calculated by measuring the inspired volume of the system; with esophageal tube blocked, when pressured to 25 cm H2O, this was 3.9 ml/cm H2O. This is comparable to that of a term newborn with healthy lungs.23 A Florian Respiratory Monitor (Acutronics Medical Systems, Zurich, Switzerland) pneumotach was placed between the Laerdal mask and the ventilating device. Percentage mask leak, using the Florian measurements, was simply defined as: (tidal volume inspired (TVinsp) − tidal volume expired (TVexp))/(TVinsp) × 100.

Recording equipment

Data from the Florian monitor was collected, via an analogue to digital converting device, using Spectra Software (Grove Medical, London, UK). The Florian monitor was calibrated with an external syringe of known volume and pressure/flow via a ventilator calibration analyser with pressure resolution of 0.1 cm H2O with pressure accuracy of ±0.5% and flow calibration with resolution of 0.1 l/min with accuracy of ±1% (RT-200; Timeter Instrument, Allied Healthcare Products, St Louis, Missouri, USA).

Instructions to participants

Staff members were randomly assigned to a two-person team. The task was to provide mask ventilation according to current International Liaison Committee on Resuscitation (ILCOR)-based Australian Resuscitation Council guidelines24 with additional instructions of aiming to provide a rate of 60 inflations per minute and either adequate chest excursion (when using the SIB) or a preset PIP of 18 cm H2O and PEEP of 5 cm H2O (when using the NP).

Single-person resuscitation

Each person was asked to perform mask ventilation of the Laerdal ALST baby manikin, using both devices (randomly sequenced), each for 2 min with 2-min rest between devices. The data for the two resuscitators were combined giving a total of 4-min data of single-person mask ventilation for each device.

Two-person resuscitation

The pair performed two-person mask ventilation for 2 min with one member holding the mask on the upper rim at four points equidistant applying balanced force down while the second to fifth fingers curl under the mandible to apply jaw thrust lift the face into the mask and correct head positioning (figs 1 and 2). The four-point grip applied in this manner does not distort the rim of the mask as maybe the case with a single-handed mask application with only two fingers applied in a pincer grip. The other person stands to the side of the manikin and uses the ventilation device. The pair then swapped roles and a further 2 min of mask ventilation was recorded. This was performed for both devices, and data for the two paired resuscitations were combined giving a total of 4-min data of two-person mask ventilation for each device.

Figure 1

Two-handed four-finger rim mask hold caudal view.

Figure 2

Two-handed four-finger rim mask hold side view.

Statistical analysis

To allow statistical comparison of single-person mask ventilation and two-person mask ventilation, data analysis is described below. Data were analysed using Stata V.MP10 (Stata Corp, College Station, Texas, USA). Combined single-person data from each pair of participants were averaged to compare to the combined two-person data. Paired t tests were used to examine differences between single-person compared to two-person for each device mode and to examine differences between SIB and NP with single-person and two-person ventilation technique. Differences between means were analysed with repeated-measure analysis of variance reporting p values for adjusted F test using Box's conservative ε with two sets of variables.

  1. Between group variables of device type

  2. Level of resuscitator

Interaction terms were examined between level of resuscitator and device mode.

p Values <0.05 were considered significant, and differences between paired means were presented with a 95% CI. This study was approved by the Sydney West Area Health Service Human Research and Ethics Committee.


Forty-eight experienced neonatal staff participated in 24 random pairs. A total of 21 578 inflations were analysed with 5679 inflations for SIB single-person, 4807 inflations for NP single-person, 5498 inflations for SIB two-person and 5594 inflations for NP two-person mask ventilation.

Comparison between single- and two-person mask ventilation using SIB (table 1): mask leak for SIB with single-person ventilation (11.5%) was significantly greater than for two-person ventilation (5.4%) with a paired difference of 6.1% (95% CI 1.5 to 10.7, p<0.01).

Table 1

Mask leak and delivered ventilation comparing single- and two-person mask ventilation with Laerdal SIB

Comparison between single- and two-person mask ventilation using NP (table 2): significant differences were found between groups for mask leak and PEEP. Mask leak reduced from a mean of 22.2% for single-person to 9.1% with a paired means difference of 13.1% (95% CI 3.6 to 22.6, p<0.01). PEEP for two-person ventilation was significantly higher at 3.8 cm H2O compared to 3.5 cm H2O for single-person with a difference of 0.3 cm H2O (95% CI 0.1 to 0.6 p<0.05).

Table 2

Mask leak and delivered ventilation comparing single and two person mask ventilation with NP

Comparison between SIB and NP single-person mask ventilation (table 3): SIB had significantly lower mask leak than NP: 11.5% compared to 22.2% with a difference of 10.7% (95% CI 1.4 to 19.7, p<0.01). SIB had significantly higher PIP than the NP, 20.1 cm H2O compared to 16.6 cm H2O with a difference of 3.5 cm H2O (95% CI 1.3 to 5.6, p<0.01). TVexp was higher for the SIB (mean 36.8 ml) compared to NP (21.7 ml) with a difference of 15.1 ml (95% CI 11.7 to 18.4, p<0.001).

Table 3

Mask leak and delivered ventilation comparing SIB and NP with single person mask ventilation

Comparison between SIB and NP two-person mask ventilation (table 4): significant differences although smaller remained in mask leak and TVexp but with no statistical difference in PIP and RR. Mean mask leak for SIB was 5.4% compared to 9.1% for NP with a difference of 3.7% (95% CI 0.1 to 6.4, p<0.05). The mean TVexp for SIB was 35.1 ml compared to 23.8 ml for NP with a mean difference of 11.3 ml (95% CI 7.2 to 15.6, p<0.001).

Table 4

Mask leak and delivered ventilation between SIB and NP with two-person mask ventilation


There were no significant differences between level of resuscitator among medical/nursing or junior/senior staff. There were no significant interactions between single- or two-person staff and device mode (data not shown).


Our study demonstrates that it is possible, in a manikin model, to halve, on average, mask leak with either SIB or NP by using a two-person method of mask ventilation (fig 3). O'Donnell et al have elegantly shown that airway pressure measurements do not inform the resuscitator about the presence of mask leak.6 Previous studies of single-person mask ventilation technique aiming to reduce mask leak were able to show a reduction to around 30%.8 A 30–50 percentage mask leak does pose the very real possibility of critical under-ventilation for any individual newborn requiring resuscitation. Our method of mask hold for two-person method differs slightly from that described in the UK Resuscitation Council Newborn Life Support Manual. Our method applies firm equal pressure on the upper rim edge (figs 1 and 2) balancing against the jaw thrust obtained from the second to fifth fingers curling under the jaw. In our experience, done correctly, there is no deformation of the mask rim as might occur with one-handed rim pincer grasp. This is confirmed in our results as one would expect greater mask leak with the two-person method if mask rim deformation was occurring rather than less leak as shown. Adult ILCOR guidelines16 and the UK Resuscitation Council Newborn Life Support Manual20 have recommended two-person mask ventilation. The adult ILCOR recommendations are on the strength of similar adult manikin studies to ours.14 15

Figure 3

Box and whisker plot of percentage mask leak by device and one or two persons.

Our study showed findings consistent with others6 that NP has more mask leak on average than SIB which maybe due to the positive pressure of the NP throughout expiration as well as inspiration. This is indeed a worrying issue as t-piece resuscitators such as NP which, because of their ability to deliver PEEP and more consistent PIP and TVexp, are becoming a popular choice of resuscitation device and are being deployed widely before the randomised controlled trial evidence of their efficacy. Given appropriate starting pressures with a t-piece resuscitator, potential complicating issues of airways obstruction in VLBW newborns11 and large unrecognised mask leak may lead to a lack of response to mask ventilation and unnecessary escalation to endotracheal intubation or cardiac compression and epinephrine use. Assessment of effective ventilation by colorimetric CO2 detector11 or TV measurements may prevent this cascade.12

Our study has shown lower levels of mask leakage than the series of studies by Wood et al and O'Donnell et al.6,,8 25 Their adaptation of the Laerdal Resusci baby manikin removed the Laerdal “lung” bags and replaced this with a Dräger 50-ml test lung with compliance equivalent to a newborn with acute respiratory distress syndrome (RDS). Our model utilised the Laerdal “lung” bags which had a static compliance equivalent to that of a healthy term lung. The majority of the 5–16% of all newborns requiring some form of resuscitation1 2 do not have RDS. We reasoned it was important to test mask leak and delivered ventilation in a non-RDS compliant model. Further, we speculate that mask leak may be greater in a poorly compliant model simulating RDS which may explain our finding of much lower average leak compared to data from Wood et al7 8 and O'Donnell et al.6 25 More sophisticated models with the ability to load elastic resistance factors may allow practitioners to be taught on more disease-focussed manikins in future.12

The strength of this study lay in the large number of inflations analysed in an expert group of users in addition to the minimal alteration of the Laerdal ALST baby manikin, which we hope will allow better extrapolation of our results to unaltered commercial Laerdal ALST baby manikins. Limitations are shared with other manikin studies of the extent one can generalise to actual human newborn resuscitations; however, practice in simulation models are the best system of teaching to date and strongly advocated.26


Neonatal mask ventilation is a technique that requires mastering, one that can be simulated in a manikin model. Previously simulated resuscitation studies suggest that mask leak is very common and significant in magnitude. We have described a simple two-person technique for mask ventilation that, in a manikin model, halves mask leak when using either the SIB or NP devices. This is the first study to support the use of two persons in mask ventilation in newborn infants. This technique is applicable to any level of unit in the first-world or third-world countries where two persons are available to resuscitate the newborn. Further studies in newborns are recommended.


We thank the staff of the Nepean Newborn Intensive Care Unit Nepean Hospital Sydney West Area Health service for their participation.



  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the Human Research and Ethics Committee Sydney West Area Health Service, Nepean Campus.

  • Provenance and peer review Not commissioned; externally peer reviewed.