Preterm Supraglottic Airways Succeed in Neonatal Resuscitation Simulations

Authors: S. Myrnerts Höök, E. Lindén, T. Alfvén, N. J. Pejovic

Abstract

Aim

Advanced airway management for preterm infants is limited. Supraglottic airway devices may provide an alternative when face‐mask ventilation fails or intubation is not feasible. Preterm‐sized devices are now accessible. This study aimed to evaluate ventilation outcomes in manikins to determine whether progression to clinical testing is warranted.

Methods

This observational manikin study included healthcare workers at Sachs Children and Youth Hospital, Stockholm, Sweden, in March 2022. Insertion time, success rates, mask leakage, ventilatory rate, tidal volumes, and airway pressures were assessed using three manikins (2200, 950, 500 g) with one commercially available device (i‐gel size 1) and three prototype sizes (neo‐i‐gel 0.85, 0.75, 0.65). Mask leakage was categorised as good (< 25%), acceptable (25%–49%), or unacceptable (≥ 50%). Participant satisfaction was recorded using a Likert scale.

Results

Across 243 sessions conducted by nine physicians and 18 nurses, optimal device sizes were 0.85 for the 2200 g and 950 g manikins, and 0.75 for the 500 g manikin, based on mask leakage and tidal volume performance. High insertion success and consistent ventilatory rates were found. Participant satisfaction was positive.

Conclusion

Preterm‐sized supraglottic airway devices appeared feasible for simulated neonatal resuscitation. Further clinical studies are urgently needed to validate their safety and effectiveness in clinical practice.

Summary This was the first study to assess ventilation outcomes using three new preterm‐sized supraglottic airway devices in preterm manikins.The new devices provided better fits than existing models, guided by mask leakage and tidal volume performance.Healthcare workers emphasized the need for preterm‐sized supraglottic airway devices but raised safety concerns for extremely preterm neonates, underscoring the importance of clinical studies to confirm safety, usability and effectiveness.

Introduction

Of the estimated 132 million neonates born in 2023, approximately 13.4 million were born preterm [1]. Premature birth, birth asphyxia and infections remain the leading causes of neonatal death. Complications arising from preterm birth were the leading cause of death among children under five, accounting for approximately 940.000 deaths and an estimated 620.000 neonatal deaths each year were attributed to intrapartum‐related events in 2019 [2]. Surviving preterm neonates often faced long‐term health consequences [3].

Immediate resuscitation is needed when a neonate is not breathing. The cornerstone of resuscitation is positive pressure ventilation (PPV), which is traditionally administered through a face mask or endotracheal intubation (ETI). Face‐mask ventilation presents several challenges, including airway obstruction [4], mask leakage [5] and interruptions during ventilation efforts [6]. This can impede the delivery of adequate tidal volumes during resuscitation. ETI ventilation is more effective, but is considered a complex technique and is mostly used by trained neonatologists and anaesthesiologists.

Geographic disparities in neonatal care further exacerbate these challenges. The highest rates of preterm birth and birth asphyxia occur in low‐income and middle‐income countries, where access to skilled personnel capable of performing ETI is limited. The World Health Organisation estimates that up to 75% of preterm deaths could be prevented by implementing basic, cost‐effective interventions [7]. Therefore, identifying accessible alternatives to face masks and ETI is essential, particularly in resource‐limited settings in both low‐income and high‐income settings.

Supraglottic airway devices (SADs) offer a less invasive alternative to ETI and can be used by a broader range of healthcare providers [8]. SADs have proven effective for achieving PPV in term neonates, potentially reducing the need for ETI. International Neonatal Resuscitation Guidelines recommend SADs for neonates ≥ 34 weeks of gestation and/or ≥ 2000 g when a face mask or ETI fail to deliver adequate PPV [9]. However, no guidelines exist for preterm neonates due to the unavailability of appropriately sized devices. ETI is also the standard method for administering surfactants to preterm neonates with respiratory distress syndrome. A less invasive method for surfactant administration via SADs is being explored in several settings [10, 11, 12, 13].

To our knowledge, no studies have been published that have evaluated the feasibility of using SAD for preterm neonates. Before clinical trials, manikin testing is essential to assess insertion time, success rates, mask leakage, tidal volume and user acceptability. While manikins cannot replicate preterm anatomy or physiology, they can provide valuable preliminary data regarding the performance of new devices.

This study aimed to evaluate ventilation‐related outcomes in manikins to determine if progression to clinical testing is warranted.

Subjects and Methods

This observational manikin study was conducted at the neonatal ward of Sachs's Children and Youth Hospital, South General Hospital, Stockholm, Sweden, between 3 March and 31 March, 2022. Healthcare workers from the Neonatology and Paediatrics departments were invited to participate. The inclusion criteria were clinically active medical doctors or registered nurses involved in neonatal resuscitation, with written informed consent obtained before participation. The sole exclusion criterion was refusal to sign a non‐disclosure agreement, as the prototypes were not yet approved for clinical use.

Four neonatal manikins were initially selected to approximate preterm and term neonatal Lifecast Body Simulation (St. Albans, UK) full term 2200 g (LC2200), pre‐term 950 g (LC950), micro‐preemie 500 g (LC500) and Laerdal Medical (Stavanger, Norway) Premature Anne 440 g. These models are designed to simulate real‐life clinical scenarios and are used in neonatal medical training [14].

The airway devices tested included three newly developed, production‐ready preterm‐compatible neo‐i‐gel prototypes, size 0.85, 0.75, and 0.65, from Intersurgical Ltd. (Wokingham, UK) and the commercially available i‐gel size 1 (Figure 1) [15]. Data collection was performed using the Monivent Neo Training from Monivent AB (Gothenburg, Sweden), with a sensor module measuring airflow and providing continuous feedback via an iPad application [16]. To minimise performance bias, visual feedback from the iPad screen and red and green colour‐coded indicators on the sensor module were concealed from participants throughout the sessions. Neonatal resuscitation simulation was performed on a GE Healthcare (Chicago, IL, USA) Panda Infant Warmer.

The manikins were selected based on pre‐trial evaluations conducted by the research team, as well as by previously published data on manikin anatomical fidelity [17]. The optimal SAD fit for each manikin was determined based on mask leakage performance. A mask leakage threshold of less than 50% was defined as acceptable. Leakage levels were classified as Grade A for less than 25%, considered good, Grade B for 25% to 49%, considered acceptable, and Grade C for 50% or greater, considered unacceptable. Details of the pre‐trial evaluation outcomes are summarised in Table 1.

Pre‐trial assessments were conducted to identify optimal SAD‐manikin combinations. The Laerdal 440 g manikin was excluded due to unreliable data capture by the Monivent system, likely attributable to its wide oesophageal design causing excessive leakage.

Procedural dry runs were performed with two neonatal nurses to standardise study protocols and ensure fidelity to clinical practice. Based on these simulations and in alignment with local neonatal resuscitation guidelines, ventilatory settings were set at a peak inspiratory pressure (PIP) of 25 cm water (H2O). The positive end‐expiratory pressure (PEEP) was set at 5 cm H2O.

The participants completed a questionnaire capturing demographics, professional background, resuscitation experience, and prior familiarity with SADs. Eligible participants were assigned unique study numbers to ensure anonymity, stored separately from the session data. Before each session, the researcher (EL) prepared, charged and tested the equipment. Both the manikin's mouth and the airway devices were lubricated with a glycerol‐based lubricant prior to insertion, in accordance with the manufacturer's recommendation. This was done to compensate for the absence of mucus in manikins. When they arrived, the participants were briefed, provided signed consent and non‐disclosure agreements, and were then given up to 10 min to familiarise themselves with the equipment. PIP and PEEP settings were verified before each session.

Each participant ventilated the three manikins selected, guided by the pre‐trial evaluation (Table 1), representing approximate birth weights of 2200, 950, and 500 g. The corresponding SAD sizes tested were 0.65, 0.75, 0.85, and 1.0. The order of device‐manikin combinations was randomised for each participant to minimise potential learning or fatigue bias.

For every combination, participants performed two ventilation cycles. Ventilation was first performed at a self‐selected rate to simulate typical clinical practice. It was then performed at a standardised rate of 60 breaths per minute, guided by a metronome in accordance with local neonatal resuscitation protocols.

This sequence was repeated for each assigned SAD–manikin combination to ensure comprehensive data collection from every participant. Incorrect SAD insertions were identified when the Monivent sensor module failed to detect ventilation data. No corrective interventions were made by the data collector during the sessions. This was in keeping with real‐world conditions, where immediate feedback may be absent.

The outcome measures included insertion success and insertion time, mask leakage, ventilatory rate, inspired tidal volume, expired tidal volume, PIP and PEEP. Insertion success was defined by successful data capture from the sensor module, with insertion time measured from SAD entry to the first ventilated breath. Mask leakage was calculated Inspired tidal volume−expired tidal volume/inspired tidal volume×100expressed as a percentage. Tidal volumes were recorded in ml/kg and ventilatory rates were expressed in breaths per minute. PIP and PEEP values were continuously monitored.

All Monivent data were exported from the iPad and securely stored. After each session, participants rated their satisfaction with the devices on a five‐point Likert scale ranging from very bad to very good and were encouraged to provide free‐text feedback on device performance or procedural aspects.

Statistics and Ethics

Summary statistics for all outcome measures included insertion time, mask leakage, ventilatory rate, inspired tidal volume, expired tidal volume, PIP and PEEP were presented as means with standard deviation (SD). For each manikin, the effects of SAD size and ventilation for the first versus the second minute on outcome measures were analysed using linear mixed‐effect models. SAD size and ventilation period were treated as fixed effects, while participant identification number was included as a random effect to account for repeated measures.

Effect sizes were reported as mean difference and 95 per cent confidence interval. SAD size 0.65 was used as the reference category. The statistical significance level was set to five per cent. Statistical analysis was performed using R 4.0 from R Foundation for Statistical Computing (Vienna, Austria) [18].

The Swedish Ethical Review Authority approved the study number 2022‐00089‐01. All participants provided written informed consent and could withdraw at any time. Data were collected anonymously, and individual results were not compared, as the study focused on product evaluation.

Results

The study included 27 participants, nine physicians and 18 nurses. The majority, representing 78%, were aged between 26 and 55 years. Eleven participants, representing 41%, had completed their most recent resuscitation training within the last 6 months. Twelve, or 44%, had completed training more than 6 months earlier. Four participants, accounting for 15%, were instructors. Most participants reported performing resuscitation procedures either weekly or monthly, accounting for 33%. Three participants, representing 11%, reported annual resuscitation experience, and five did not specify a frequency because they had recently returned from maternity or sick leave. Sixteen participants, or 59%, had never performed a SAD insertion, whereas 11 participants, representing 41%, reported performing SAD insertion approximately once per year.

Each participant completed nine ventilation sessions, resulting in a total of 243 sessions. Almost all insertions with the studied SADs on the selected manikins were successful. Four sessions, accounting for 1.6%, were excluded. Three were excluded because of excessive leakage exceeding 90%, which prevented accurate data two with LC500 plus SAD 0.75, one with LC2200 plus SAD 0.85 and one because of a technical error resulting in irretrievable data.

Insertion time and mask leakage are summarised in Table 2. Mean insertion time was approximately 6 s. Notably, insertion times were longer when using SAD size 1 compared to size 0.65 on the LC2200 manikin, and when using size 0.85 compared to size 0.65 on the LC950 manikin, as detailed in (Table S1).

Overall mask leakage values ranged from 8% to 60%. On the LC2200 manikin, mask leakage remained within the good range, defined as less than 25%, for all SAD sizes. For the LC950 manikin, leakage was classified as good for size 0.85, while sizes 0.75 and 0.65 demonstrated acceptable leakage levels, ranging between 25% and 49%. In contrast, on the LC500 manikin, mask leakage was deemed unacceptable, exceeding 50%, when using size 0.65, whereas size 0.75 yielded acceptable leakage levels. There were no statistically significant differences in mask leakage observed between the first and second minutes of ventilation (Table S2).

Mean ventilation rate ranged from 53 to 58 ventilations/min with no major differences between the different SAD and manikin combinations. No clinically relevant variations were observed between different device sizes or between self‐paced and standardised ventilation periods. All measured ventilation rate values remained within acceptable clinical ranges during both ventilation phases (Tables S3 and S4).

Inspired and expired tidal volumes are presented in Table 3. Mean inspired tidal volume values ranged from 4.8 to 16.7 mL/kg, with the highest volumes recorded using size 0.65 on the LC500. It was lower with size 0.85 and 0.75 compared to size 0.65 on the LC950, and much lower for size 0.75 compared to size 0.65 on the LC500 (Table S5).

Mean expired tidal volume values ranged from 2.2 to 6.9 mL/kg. It was higher with size 0.85 compared to size 0.65 on LC950, and size 0.75 compared to size 0.65 on LC500 (Table S6).

No significant differences in inspired and expired tidal volume were found between the first and second minutes of ventilation (Tables S5 and S6).

PIP and PEEP values are summarised in Table 4. Mean PIP ranged from 32.1 to 37.7 cm H2O, consistently exceeding the set PIP of 25 cm H2O across all manikins. An inverse correlation was observed between PIP and manikin size, with higher pressures recorded in smaller manikins. Mean PEEP values ranged from 5.8 to 6.1 cm H2O, closely aligning with the set PEEP of 5 cm H2O with only minor fluctuations between sizes and manikins. No significant differences in PIP or PEEP were observed between the first and second minutes of ventilation (Tables S7 and S8).

Participant satisfaction was rated as very good by 21 participants and good by six. Free‐text responses revealed a consensus on the clinical need for preterm‐sized SADs, although some expressed safety concerns about their use in extremely preterm neonates. The most common challenge was determining the correct insertion depth in manikins. A participant noted that the large plastic disc at the top of the SAD could press against the neonate's mouth, potentially causing injury.

Discussion

This observational manikin study evaluated the feasibility of preterm‐sized SADs in simulated neonatal resuscitation scenarios. The findings indicate that SAD size 0.85 was optimal for manikins representing 2200 and 950 g preterm infants. Size 0.75 was optimal for the 500 g manikin based on mask leakage and tidal volume performance. The combination of SAD size 0.65 in the LC500 resulted in unacceptable mask leakage, consistent with pre‐trial evaluations where size 0.85 and size 1 were also unsuitable for the LC500, and size 1 was excluded for the LC950. Importantly, mask leakage below 30% was achievable with at least one SAD size for each manikin, though the variability between sizes was most pronounced in the smaller manikins.

The recommended inspired tidal volume for neonatal ventilation is 4–8 mL/kg, but further research is needed to establish guidelines for preterm neonates [19]. In this study, inspired tidal volume values for the LC2200 and LC950 manikins largely aligned with the recommended range. By comparison, the LC500 manikin consistently exhibited elevated inspired tidal volume values, reflecting potential overventilation. This discrepancy may be attributed to the duration of inspiration, which was determined by each participant. Typically, inspiration accounts for approximately one‐third of the ventilatory cycle, with passive expiration occurring over the remaining two‐thirds. However, if the inspiration duration remained constant across all manikins, smaller manikins would receive a disproportionately higher volume of inspired air relative to their lung capacity. Because lung volume is finite and excess air must be displaced, this would likely result in relevant mask leakage. Our findings support this, where the SADs tested on the LC500 had the highest leakage. These observations suggest that future clinical implementation of preterm‐sized SADs should be accompanied by targeted training on ventilatory rate. Such training should also address pressure control to mitigate the risk of volutrauma and barotrauma. However, whether SAD mask leakage in manikins correlates with mask leakage during resuscitation in real‐life newborns is still unknown.

The existing literature on neonatal SAD use remains limited, particularly concerning preterm populations. To our knowledge, only one published study, from our research group, has radiologically evaluated the anatomic positioning of three preterm‐sized SADs in manikins. However, it was conducted after the present study, and thus its results could not be considered in our study design [20]. A previous manikin study that compared face masks and the i‐gel size 1 SAD for ventilation concluded that the SAD was superior in achieving successful PPV and also more user‐friendly [21]. Early‐phase trials indicated improved outcomes, such as reduced time to spontaneous breathing [22]. However, later‐phase trials did not demonstrate significant superiority over a face mask in terms of major clinical outcomes, although SADs were deemed safe even when used by midwives [8]. The reports of mask leakage around 40% with both SADs and face masks further highlight the challenges of achieving optimal ventilation, although SADs have been associated with better heart rate responses during initial resuscitation efforts [23]. A systematic review suggested that SADs are more effective than face masks in delivering PPV and reducing the need for ETI in term neonates, particularly in low‐income and middle‐income countries [24].

Our study adds to this growing body of evidence by focusing on preterm‐compatible SADs. It highlights their potential feasibility and ease of use, even among healthcare providers with limited prior experience. We included both physicians and nurses with diverse neonatal resuscitation backgrounds. This approach strengthens the generalizability of the findings to real‐world clinical settings.

Limitations

This study had a number of inherent limitations. Critical factors were missing, such as mucosal compliance, airway tone, and physiological feedback. Challenges related to manikin design, including rigid tongues and unrealistic airways, were noted by participants and may have influenced device placement and performance assessments. One researcher (NJP) is a neonatologist and struggled to intubate the LC500 manikin before the study, noting unrealistic airway angles that made visualisation difficult. Several participants found it challenging to determine proper SAD placement compared to live neonates. There is a possibility that smaller SADs were inserted too deeply. In real‐life resuscitation, increasing pulse and spontaneous breathing play crucial roles in optimising placement, factors not replicable in manikin studies.

Equipment‐related limitations were also identified. The GE Panda infant warmer, though widely utilised in neonatal care, demonstrated variability in PIP, consistently exceeding the set value of 25 cm H2O. This aligns with previous reports of unintended PIP spikes associated with this device, particularly in situations of low pulmonary compliance, which is typical during initial resuscitation of apneic neonates [25, 26]. When compared with the similar device Neopuff Infant T‐Piece Resuscitator (Fisher & Paykel Healthcare, New Zealand), the GE Panda infant warmer delivered a higher plateau pressure, potentially leading to increased tidal volumes and increasing the risk of barotrauma (Figures S1 and S2). While mask leakage, the primary outcome, was unlikely to be affected by these variations, the potential for elevated PIP to increase tidal volumes warrants caution in clinical translation. Furthermore, difficulties in the precise adjustment of PIP and PEEP contributed to session variability. After starting data collection, we found a previous report indicating that the Panda infant warmer may produce unintended PIP spikes, which could pose a safety risk for neonates [25]. Our data, processed by Monivent AB staff, showed that lower pulmonary compliance correlated with larger PIP spikes, a particular concern given that apnoeic neonates have lower lung compliance during initial PPV [26]. Finally, manikin studies lack the real‐life stress and contextual factors of clinical resuscitation. Participants focused solely on ventilation. In actual resuscitations, they must also contend with distressed parents, material preparation, and the urgency of reviving a non‐breathing neonate. All of this may impact SAD insertion and ventilation success. In high‐resource settings, ventilation is often a shared task with alternating providers. This may not be feasible in resource‐constrained environments where trained personnel are few. Given these limitations, we would like to emphasise the uncertainty regarding whether the results of this manikin study can be directly translated to human neonates.

Conclusion

This study provides an important initial assessment of preterm‐sized SADs for neonatal resuscitation. Optimal device sizes were identified based on mask leakage and tidal volume performance, with size 0.85 suitable for the 2200 and 950 g manikins and size 0.75 for the 500 g manikin. High participant acceptance highlights the potential usability of these devices in clinical practice. Preterm‐sized SADs could improve access to effective ventilation and facilitate surfactant administration in preterm neonates with respiratory distress syndrome, particularly in settings where endotracheal intubation expertise is limited. By enabling a broader range of healthcare providers to perform advanced airway management, these devices may contribute to reducing neonatal morbidity and mortality. While SADs are not intended to replace ETI in all settings, they offer a valuable adjunct when immediate specialist support is unavailable. Overall, these findings provide a strong foundation for urgently needed clinical trials to validate the safety, effectiveness and practical application of preterm‐sized SADs in neonatal care.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information