Article Text

Download PDFPDF

Randomised comparison of two neonatal resuscitation bags in manikin ventilation
  1. Monica Thallinger1,
  2. Hege Langli Ersdal2,
  3. Crescent Ombay4,
  4. Joar Eilevstjønn3,
  5. Ketil Størdal5
  1. 1Faculty of Medicine, Institute of Clinical Medicine, Institute for Experimental Medical Research, University of Oslo, Nydalen, Oslo, Norway
  2. 2Department of Anaesthesiology & Intensive Care, Stavanger University Hospital, Stavanger, Norway
  3. 3Laerdal Medical AS, Strategic Research, Stavanger, Norway
  4. 4Principal Tutor at Haydom School of Nursing, Haydom, Manyara, Tanzania
  5. 5Division of Epidemiology, Norwegian Institute of Public Health, Nydalen, Oslo, Norway
  1. Correspondence to Dr Monica Thallinger, Faculty of Medicine, Institute of Clinical Medicine, Institute for Experimental Medical Research, University of Oslo, P.O. Box 4956, Nydalen, Oslo 0424, Norway; monica.thallinger{at}


Objective To compare ventilation properties and user preference of a new upright neonatal resuscitator developed for easier cleaning, reduced complexity, and possibly improved ventilation properties, with the standard Laerdal neonatal resuscitator.

Design Eighty-seven Tanzanian and Norwegian nursing and medical students without prior knowledge of newborn resuscitation were briefly trained in bag-mask ventilation. The two resuscitators were used in random order on a manikin connected to a test lung with normal or low lung compliance. Data were collected with the Laerdal Newborn Resuscitation Monitor. The students graded mask seal and ease of air entry on a four-point scale ranging from 1 (‘difficult’) to 4 (‘easy’) and stated which device they preferred. (Equipment from Laerdal Global Health and Laerdal Medical).

Results For upright versus standard resuscitator and normal lung compliance, mean expiratory lung volume was 15.5 mL vs 13.9 mL (p=0.001), mean mask leakage 48% vs 58% (p<0.001), and mean airway pressure 20 cm H2O vs 19 cm H2O (p=0.003), respectively. For low lung compliance, mean expiratory lung volume was 8.6 mL vs 8.1 mL (p=0.045), mean mask leakage 53% vs 62% (p<0.001), and mean airway pressure 21 cm H2O vs 20 cm H2O (p=0.004) for upright versus standard. The upright resuscitator was preferred by 82% and 68% of students during ventilation with normal and low lung compliance, respectively (p=0.001).

Conclusions Expiratory volumes were higher, mask leakage lower, and mean airway pressure slightly higher with upright versus standard resuscitator when ventilating a manikin. The majority of students preferred the upright resuscitator.

  • Neonatology
  • Resuscitation
View Full Text

Statistics from

What is already known on this topic

  • Manual bag-mask ventilation (BMV) during newborn resuscitation is a critical intervention and a life-saving skill to learn and master.

  • BMV, using standard resuscitators, is difficult with irregular tidal volumes and high degree of leakage around the mouth.

  • Research suggests that even skilled healthcare providers show inadequate BMV skills.

What this study adds

  • A new upright resuscitator gave higher volumes and lower mask leak compared with a standard resuscitator on a manikin model.

  • The participants, who did not have previous experience on neonatal ventilation, preferred the upright resuscitator.


Approximately one million newborns died on the day of birth in low-income countries in 2013.1 ,2 Neonatal mortality risk in low-income countries is approximately 7 times higher than in high-income countries and 10 times higher in sub-Saharan countries.2 It is estimated that 24% of these deaths are related to birth asphyxia.2 ,3 Improved methods of teaching and improved, cheaper and simpler-to-use equipment are needed to reduce this problem.4 Helping Babies Breathe (HBB) is a newborn resuscitation teaching programme under dissemination in many low-income countries potentially reducing neonatal mortality.5 In Tanzania, the reported reduction in neonatal mortality was caused by immediate stimulation of non-crying newborns, since mortality was unchanged among infants receiving bag-mask ventilation (BMV).5

BMV is one of the most difficult parts of newborn resuscitation as experienced in HBB.6 ,7 It is crucial when resuscitating asphyxiated newborns whereas chest compressions seldom are required if airways and ventilation are handled appropriately.8 Research suggests that even skilled paediatricians or anaesthesiologists show inadequate BMV skills.9 Research related to optimal ventilation pressures, volumes and frequency is limited, with many recommendations derived by consensus or extrapolated from adult models or cardiac arrest; not appropriate for newborns.10

In 1954 Henning Ruben invented the self-inflating bag,11 ,12 later modified and described as the Air Mask Bag Unit in 1957, with design and shape mostly unchanged until now. T-piece resuscitators and anaesthesia bags came in use at a later stage in newborn resuscitation.13 ,14 A resuscitator shaped differently than the standard bag-mask with a vertical shape (‘upright’) and fewer parts has recently been designed (figure 1).

Figure 1

Picture of the Laerdal Upright Resuscitator (left) and standard Laerdal Neonatal Resuscitator (right).

This study aims to compare standard Laerdal Neonatal Resuscitator (used today) and Laerdal Global Health Upright Resuscitator with respect to expiratory lung volume, mask leakage, mean airway pressure (MAP) and user preference. Midwives or traditional birth attendants with little or no resuscitation training most frequently attend deliveries in low-resource settings. In six African countries half of the births occurred outside of facilities, and in facilities only 2–12% of birth attendants had neonatal resuscitation training.15 As the upright resuscitator primarily is constructed for low-resource settings, we chose to study inexperienced students instead of trained neonatologists.


Eighty-seven 1st-year and 2nd-year students volunteered for the study: 6 medical and 35 nursing students in Norway and 46 nursing students in Tanzania. Students were recruited through social media and presentations during lectures. Exclusion criteria were previous knowledge of neonatal resuscitation, experience with BMV or neonatal ward experience. All 41 Norwegian volunteers (among approximately 100 students) were included. Four of 50 Tanzanian volunteers (among approximately 120 students) were excluded because of labour ward experience. All students signed a written consent form after receiving oral and written information about the study and data management.


All students received standardised one-to-one 10-min information including reasons for ventilating newborns, for the research question, background information and a brief introduction to the HBB algorithm. They were not informed which resuscitator was in clinical use or newly designed before practicing on the manikin for 5 min stressing the following points:

  • To open the airway, keep the head in neutral position with the nose up, and place the mask over mouth and nose.

  • Three-finger hand grip: Two fingers on top of the mask with index finger over the chin and thumb over the nose, middle finger under the chin without applying pressure on the trachea. A contracting force between the three fingers help seal the mask-face interface.

  • Three-finger grip of the other hand squeeze the bag until adequate chest rise.

A three-item questionnaire to be filled out after the testing was presented pre testing to make them aware of what we wanted their personal opinion on (see below under ‘Devices, study design and testing’). Predefined primary outcomes were differences in expiratory lung volumes between the upright and standard resuscitators with normal and low lung compliance, respectively. Secondary outcomes were mask leakage, MAP and questionnaire answers.

Devices, study design and testing

Upright Resuscitator 320 mL (Cat. no. 846050 Laerdal Global Health, Stavanger, Norway) ( (figure 1) was compared with standard Laerdal Neonatal Resuscitator 220 mL (Cat. no 846040 Laerdal Medical, Stavanger, Norway) in a randomised cross-over design in a laboratory setting using the NeoNatalie Resuscitation Manikin (approximately 2 kg) (Laerdal Global Health) connected to the Laerdal Newborn Lung Simulator (Laerdal Medical). The upright resuscitator was designed to be less complex to assemble and disassemble, with fewer parts. The new Laerdal clear cushioned face mask (Size 1) that comes with the upright resuscitator was used with both resuscitators to eliminate the influence of its assumed improved qualities. Both resuscitators have a pop-off valve set at 40 mbar with gradually increasing leak from approximately 25 mbar, in compliance with ISO 10651-4 (International Organization for Standardization).

The lung simulator has two compliance settings: ‘normal’ of 0.8 mL/cm H2O simulating aerated lungs, and ‘low’ of 0.1–0.2 mL/cm H2O simulating fluid-filled lungs based on data from first breaths manual ventilations postdelivery by Hull,16 Miner et al,17 and Field et al.18 The lung simulator system was sealed for leakage with approximately 90 mL total volume of lungs and tubes.

Ventilation pressure and expiratory and inspiratory volumes were measured with a flow sensor (Acutronic Medical Systems, Hirzel, Switzerland) integrated in the Laerdal Newborn Resuscitation Monitor. The sensor measures airflow with hot wire anemometer technology, has negligible flow resistance, 1 mL dead space, and was attached between the resuscitator (with its pop-off valve) and the face mask. The sensor technology is identical to that used in the Florian Respiratory Function Monitor (Acutronic Medical Systems, Hirzel, Switzerland), extensively used during newborn resuscitation. Laerdal Newborn Resuscitator Monitor calculated mask leakage automatically from inspiratory minus expiratory lung volume. Data from each ventilation were recorded separately (equipment from Laerdal Global Health and Laerdal Medical, Stavanger, Norway).

To standardise the ventilation frequency, a 50 bpm metronome was used.

Students chose one out of two sealed envelopes for random allocation of resuscitator sequencing in the cross-over study. They first attempted ventilation with normal lung compliance. Each resuscitator was used for two 1 min ventilation cycles giving a total of 4 min of ventilation (table 1).

Table 1

Randomisation table for resuscitator

The participants thereafter answered the following in writing for each resuscitator:

  1. How easy was it to get the mask sealed well? (1=difficult, 2=somewhat difficult, 3=somewhat easy, 4=easy)

  2. How easy was it to see chest rise on the manikin? (1=difficult, 2=somewhat difficult, 3=somewhat easy, 4=easy)

  3. Which bag do you prefer? (upright or standard).

Thereafter, lung compliance was switched from normal to low, and they again chose one of two sealed envelopes for a second 4-min random-device ventilation session. An identical questionnaire was filled out for the low-compliance sequence.

The first five ventilations of each ventilation cycle were excluded from analysis assuming that they would include mask adjustment and hand grip adaptation. The next 30 ventilations were extracted and analysed. Using each resuscitator twice gave 120 (60 for each resuscitator) ventilations per student to be analysed. Expiratory volumes above 50 mL (considered maximum realistic volume in this set-up) were excluded from the main analysis if caused by an error in the ventilation detection algorithm (ventilation end point detected erroneously in some ventilations with very high mask leakage). Consequently, 86 ventilations for normal compliance (25 for upright and 62 for standard) and 103 for low compliance (36 for upright and 67 for standard) were excluded from the main analysis, and the full data set used in a sensitivity analysis.

Data analysis

Descriptive data were analysed with mean, SDs, medians and ranges. To ensure normal distribution of the test variable (difference between resuscitators), these were plotted and the Kolmogorov-Smirnov test performed.

We estimated resuscitator type (upright vs standard) effects on expiratory lung volume, MAP, peak inspiratory pressure (PIP) and mask leakage during each 1 min cycle. Because repeated measurements for each student do not meet the criteria of independency, we fitted linear mixed-effects regression models with random intercept for each student.

Questionnaire data were analysed by Wilcoxon signed rank test (questions 1 and 2) and by a one-sample binomial test (question 3). We used SPSS V.20.0 statistical software package (IBM SPSS, Chicago, Illinois, USA) and Stata V.12 (StataCorp LP, Texas, USA).

Ethical considerations

The study was approved by The National Institute for Medical Research in Tanzania and The Regional Committee for Medical and Health Ethics, Western Norway.


Ventilation data from 5 of 87 participants were excluded due to missing ventilation data (monitor not turned on). For normal lung compliance 9460 ventilations were analysed and 9303 for low compliance.

Ventilation outcomes, normal and low lung compliance

Table 2 presents ventilation data for the resuscitators and normal lung compliance. The upright resuscitator resulted in significantly higher expiratory volumes, lower mask leakage and higher MAP with no difference in PIP compared with the standard resuscitator.

Table 2

Ventilatory outcomes with normal lung compliance

With low lung compliance expiratory lung volumes were also significantly higher with the upright resuscitator, but difference between resuscitators was smaller than with standard compliance. As with standard compliance, mask leakage was also lower and MAP higher for the upright resuscitator compared with the standard resuscitator (table 3).

Table 3

Ventilatory outcomes with low lung compliance

Sensitivity analysis including ventilation cycles with expiratory cycles >50 mL showed only minor differences compared with the main model (tables 2 and 3).


All students answered the questionnaire. More students indicated that it was easier to get mask seal (p<0.001) and to see chest rise with upright compared with the standard resuscitator (p=0.026 for normal compliance, p=0.003 for low compliance, table 4).

Table 4

Distribution of students’ answers of the four-point scale (in %)

The upright resuscitator was preferred by 68% when ventilating with normal lung compliance (p=0.001) and 82% with low compliance (p<0.001).


Expiratory volume was higher, mask leakage less and MAP higher with the upright resuscitator compared with the standard resuscitator on a manikin with normal and low lung compliance. More students preferred the upright resuscitator, and found it easier to get good mask seal and to observe chest rise compared with the standard resuscitator.

Higher expiratory volume and MAP without differences in PIP suggest that the differences were due to higher mask leakage with standard equipment. The leakage difference could not be due to the mask, as the same new mask was used with both resuscitators. The difference could be due to ergonomics with the upright resuscitator ‘standing’ over the face, with the weight of the bag pushing downwards on the face instead of being at an angle with its weight working to lift the mask from the face. Van Vonderen et al19 found that physicians and nurses used a mean force of approximately 2 kg during mask ventilation on a manikin with self-inflating bag and Neopuff. There were no changes in force used after attempts to correct mask leaks. We did not investigate the force applied in our study. With the upright design we might query the possibility of it increased adding force to the head. We would stress that mask-face handgrip is the same with both resuscitators, and the weight of the upright resuscitator did not appear to cause more airway compression. The potential of airway compression should be kept in mind in clinical evaluation and use.

The differences between the two resuscitators were statistically significant, but relatively small, and the clinical relevance remains unclear. The upright resuscitator has 100 mL more volume than the standard resuscitator, and this might have enabled the users to increase the volumes and maintain higher MAP. The bigger volume cannot explain lower mask leakage using the upright resuscitator.

Difficulties in neonatal mask ventilation with irregular tidal volumes and variable but high degrees of leakage are reported in clinical and manikin studies.20–24 Manikin study leakage is typically reported to be 55–70%; similar to what we found with standard equipment, but higher than with the upright resuscitator.21 ,22 This suggests that positioning might be easier to obtain with improved seal using the upright resuscitator. A few manikin studies have examined mask positioning, grip and methods during BMV.22 ,25 ,26 Wood et al22 found optimised technique of positioning and holding the face mask together with written instruction and demonstration could reduce leakage down to 32% with no differences between mask types. O'Donnell et al23 reported 65% leaks around the face mask. Wilson et al24 found no difference in mask leakage with three different mask holds. Mask and manikin face compatibility has not been objectively compared with the clinical situation. A manikin face can never fully mimic a newborn face, but to the three clinical authors with extended experience with newborn resuscitation, the NeoNatalie manikin seems realistic. The set-up has been used in hundreds of thousands of training situations around the world.

The test lung compliance settings of 0.8 mL/cm H2O and 0.1–0.2 mL/cm H2O were based on data from first manual ventilation attempts in newborns requiring resuscitation post delivery by Hull,16 Milner et al,17 and Field et al.18 Hull reported mean compliance of 1.0 (range 0.3–1.7) mL/cm H2O during the first ventilations in intubated newborns with inflations lasting up to 0.5 s.16 Milner et al17 found mean 0.5–0.9 mL/cm H2O during first ventilations in intubated newborns and 0.1 mL/cm H2O with face mask, both with inflations lasting up to 0.5 s. Field et al18 reported 0.3–0.7 mL/cm H2O for the first three face mask inflations each lasting approximately 1 s. These compliances are lower than mean values of 1–4 mL/cm H2O reported by others, who did not study the immediate postdelivery period.27–30 Compliance probably changes with time post delivery, even during the first few minutes with initial viscous fluid-filled airways and tension of air-fluid interfaces.

Although we excluded students with prior neonatal ward or BMV experience, some may have known which device was in clinical use or new. This could cause questionnaire and performance bias. They were not asked to cross-evaluate between high and low compliance, but it cannot be excluded that the evaluation with normal compliance, which always occurred first, could influence their opinion when ventilating with low compliance. While there were no differences in results between Norwegian and Tanzanian students despite different backgrounds, data cannot automatically be generalised to other groups.

Manikin study results cannot automatically be extrapolated to clinical situations. Manikins are effective learning tools, but can never provide a realistic picture of how neonates behave in real settings.24 Clinical studies are needed to translate these preliminary findings to outcomes. The results indicate that the upright resuscitator functions at least as good as standard equipment for inexperienced users and a clinical feasibility study seems warranted.

Most research on newborn resuscitation and ventilation occurs in high-resource settings, particularly regarding premature treatment where high-quality care is present. Equipment is needed that is easy to clean and maintain and adapted for low-resource settings with high neonatal mortality. In any setting, high or low resourced, mastering the skill of newborn ventilation is difficult. Improved functions such as less mask leakage may be valuable for any resuscitation setting.


The upright resuscitator provided higher expiratory volumes and MAPs with lower mask leak compared with the standard resuscitator, and the upright resuscitator was preferred by the majority of participants in both settings.


The authors thank Prof Petter Andreas Steen, University of Oslo, for his advice and critical review of this manuscript. The authors thank the 87 medical and nursing students who willingly and enthusiastically joined the research in Tanzania and Norway. The authors thank Associate Professor Kristin Myhre, International Coordinator Faculty of Health and Social Studies at Østfold University College for all her help in recruiting students, dispose facilities and excellent cooperation. The authors thank statistician Nina Gunnes at the Norwegian Institute of Public Health. The authors also thank Laerdal Global Health for providing equipment and technical support.


View Abstract


  • Contributors MT conceptualised and designed the study protocol, oversaw and carried out the study and data collection, data analysis/statistics, drafted the initial manuscript, reviewed and approved the final manuscript as submitted. KS contributed practical and theoretical support of the protocol and the study, data analysis/statistics, and reviewed and accepted the final manuscript as submitted. HLE provided theoretical support of the protocol and reviewed and approved the final manuscript as submitted. JE gave technical support of the equipment used, managed data processing, managed the figures for publication, and reviewed and approved the final manuscript as submitted. All authors contributed to refinement of the study protocol and approved the final manuscript. CO coordinated and supervised data collection, and reviewed and approved the final manuscript as submitted.

  • Competing interests MT has received unrestricted PhD grant from Laerdal Foundation. KS is supported by an unrestricted grant from Oak Foundation, Geneva, Switzerland. JE is an employer of Laerdal Medical.

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

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.