Article Text
PDF
Abstract
Background T-piece is recommended for respiratory support during neonatal stabilisation. Bench studies have shown a delay >30 s in achieving changes in fraction of inspired oxygen (FiO2) at the airway when using the T-piece. Using a face mask adds dead space (DS) to the patient airway. We hypothesised that adding face mask to T-piece systems adversely affects the time required for a change in FiO2 to reach the patient.
Methods Neopuff (Fisher and Paykel, Auckland, New Zealand) and rPAP (Inspiration Healthcare, Croydon, UK) were used to ventilate a test lung. DS equivalent to neonatal face masks was added between the T-piece and test lung. Additionally, rPAP was tested with nasal prongs. Time course for change in FiO2 to be achieved at the airway was measured for increase (0.3–0.6) and decrease (1.0–0.5) in FiO2. Primary outcome was time to reach FiO2+/−0.05 of the set target. One-way analysis of variance was used to compare mean time to reach the primary outcome between different DS volumes.
Results In all experiments, the mean time to reach the primary outcome was significantly shorter for rPAP with prongs compared with Neopuff and rPAP with face mask DS (p<0.001). The largest observed difference occurred when testing a decrease in FiO2 with 10 mL tidal volume (TV) without leakage (18.3 s for rPAP with prongs vs 153.4 s for Neopuff with face mask DS). The shortest observed time was 13.3 s when increasing FiO2 with 10 mL TV with prongs with leakage and the longest time was 172.7 s when decreasing FiO2 with 4 mL TV and added face mask DS without leak.
Conclusion There was a delay in achieving changes in oxygen delivery at the airway during simulated ventilation attributable to the mask volume. This delay was greatly reduced when using nasal prongs as an interface. This should be examined in clinical trials.
- Intensive Care Units, Neonatal
- Resuscitation
- Neonatology
Data availability statement
Data are available upon reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Previous bench studies have shown a delay in achieving changes in fraction of inspired oxygen (FiO2) at the airway when using the T-piece for positive pressure ventilation.
WHAT THIS STUDY ADDS
We found that the delay in changes of FiO2 to reach the patient’s airway during simulated ventilation is largely related to the interface dead space (DS) added to the respiratory circuit. This delay can be greatly reduced by using nasal prongs as the interface. The differences in mean time between different DS volumes ranged from twofold to eightfold.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
These results raise questions about whether or not the face mask is the optimal interface for neonatal stabilisation.
Introduction
While most newborn infants breathe spontaneously at birth,1 many require respiratory support to establish regular breathing.2 This is more common in preterm infants and reaches 100% for the smallest preterm infants. Positive pressure ventilation (PPV) is delivered with devices such as conventional and variable flow T-pieces and self-inflating and flow-inflating bags.3 During stabilisation, supplemental oxygen is adjusted to maintain the desired oxygen saturation (SpO2) based on recommended targets.4 5 Both hypoxia and hyperoxia can be potentially harmful6 and recommendations for initial oxygen concentration during neonatal stabilisation have changed considerably during the last 20 years.7 8
Previous studies have shown clinically important delays (>30 s) in the time required for changes in fraction of inspired oxygen (FiO2) at the blender to reach the patient’s airway in simulated PPV tested with T-piece devices.9 10 Those studies focused on different gas flow rates and leakages but no reasonable explanation was identified as a cause for this delay. The use of face masks during neonatal stabilisation is well-established.11 However, providing PPV with a face mask can be challenging and ventilation can be compromised by leaks and airway obstruction.12 13 Using a face mask as the interface substantially increases dead space (DS) outside the patient’s airway. Nasal prongs are an alternative interface with very low DS. We hypothesised that the added DS of the airway interface (mask or prongs) would significantly affect the time taken for changes in FiO2 at the blender to reach the patient’s airway.
Methods
The experimental setup is shown in figure 1. The primary outcome was time to reach a FiO2 0.05 from the set target. For the two starting points, FiO2 0.30 and 1.0, this was the time to reach FiO2 of 0.55. The FiO2 0.05 was similar to the 5% margin between the two sensors used by Dekker et al.9 The respiratory support systems tested were Neopuff (Fisher and Paykel, Auckland, New Zealand) and rPAP (Inspiration Healthcare, Croydon, UK). The DS of the face mask was simulated by connecting straight 15 mm connectors to the T-piece. The inner space of Laerdal face masks (Laerdal Global Health, Stavanger, Norway) size 00 and 0 was measured by injecting water into the masks. The volume was 18 mL for size 00 and 31 mL for size 0. To make the simulation more clinically relevant, we decreased the estimated added mask DS by 5 mL for each mask on the basis that parts of the face will protrude into the face mask and compression of the face mask for seal may reduce the volume. The DS added was 13 mL and 26 mL to represent the volume of the size 00 and 0 masks respectively. The medium short binasal prongs used with the rPAP had an estimated volume of less than 1 mL. The respiratory support device and added interface DS/rPAP prongs were connected leak-free to a custom-built three-dimensional printed connector for the O2 sensor (5 mL DS) and this was connected directly to a 50 mL Dräger Bellows MP20050 (Dräger, Lübeck, Germany). (online supplemental figure 1). There was no humidification circuit.
Supplemental material
The equipment setup used for the experiment. Fresh gas flow was set to 10 litres per minute. FiO2 was adjusted with a blender. Dead space was added by connecting straight cylinders to the Neopuff and rPAP. The respiratory system was connected to an O2 sensor and directly to a 50 mL Dräger Bellows test lung. Tests were performed at two levels of tidal volumes which were altered by applying or removing clamps to the test lung. Measurements were done with and without leaks. FiO2, fraction of inspired oxygen.
To simulate clinical problems with leakage during mask ventilation, the tests were performed both with and without leakage. Leakage between the resuscitation system and the oxygen sensor was generated through a 0.80*50 mm steel needle (21G Sterican, Braun, Melsungen, Germany) that could be turned on using an intravenous line valve. Tidal volumes (TVs) and leaks were measured by a respiratory function monitor (RFM), Monivent Neo100 (Monivent, Gothenburg, Sweden). With the leak activated, the relative leak for 10 mL TV was 29.9% (SD 1.8, for 12 inflations) and for 4.0 mL 47.5% (SD 5.8, 12 inflations). The absolute leakage was the same for both TVs. The Monivent connector piece was removed before the experiments were started to avoid introducing DS to the circuit.
All tests were performed using PPV at 60 breaths per minute with a peak inspiratory pressure (PIP) of 24 cm H2O and positive end-expiratory pressure (PEEP) of 6 cm H2O. Fresh gas flow was set to 10.0 L/min. All flows and pressures were calibrated using Fluke VT650 Gas Flow Analyser (Fluke, Washington, USA). Tests were performed at two levels of TVs which were altered by applying or removing clamps to the test lung to restrict movement and reduce compliance. The experiments were performed at 4 mL and 10 mL TV. These volumes represent the TV used to ventilate infants of appropriate weight for Laerdal face mask sizes 00 and 0, delivering 6 mL/kg per inflation for infants weighing 750 g and 1750 g respectively. The 13 mL DS was used for the 4 mL TV and 26 mL for the 10 mL TV. The less clinically relevant combinations (13 mL DS with 10 mL TV and 26 mL DS with 4 mL) are available as supplement (online supplemental table 1).
Supplemental material
FiO2 was adjusted from 0.30 to 0.60 and from 1.00 to 0.50. FiO2 was adjusted manually using a blender and the time was set to zero when FiO2 deviated from baseline at an oxygen sensor positioned after the blender. FiO2 was measured using chemical sensors (R-17MED, Teledyne Analytical Instruments, California, USA). Data was collected until equilibrium was reached at the patient interface or a maximum of 3 min. Data was acquired using NI NI-9205 and LabVIEW 2015 (National Instruments, Austin, Texas, USA) at 100 Hz.
Each experiment was repeated three times to ensure a low variance of the measuring method. Data were compiled and analysed in SPSS (IBM, Armonk, New York, USA). One-way analysis of variance (ANOVA) was used to compare means in seconds to reach FiO2 0.55 for all combinations of TV, presence of a leak or not and upwards and downwards titration of oxygen. This resulted in eight ANOVA tests. Post hoc comparisons to reveal within group difference were done using Bonferroni. Assumptions were checked with normal probability plots and the assumption of equality of variance was tested with Levene’s test. Statistical significance was set to p<0.05.
Results
Mean time to reach FiO2 within 0.05 of the set target was significantly shorter for rPAP with prongs compared with Neopuff and rPAP with higher DSs representing facemask volumes. This was seen in all measurements, in both set TVs and with or without leakage (table 1). The largest difference was found when testing the decrease in FiO2 with 10 mL TV without leakage where it took 18.3 s for rPAP with prongs compared with 153.4 s for Neopuff with face mask DS (p<0.001, 95% CI 134.4 to 135.7) (table 1). This difference was mainly related to the interface DS and not the type of resuscitation device.
Time until reaching FiO2 0.05 from the set target when decreasing FiO2 from 1.0 to 0.5 and increasing from 0.3 to 0.6. One-way ANOVA for comparison. * Post hoc test Bonferroni
There was a delay of >10 s in reaching the set oxygen concentration at the patient interface in every measurement (table 1, figures 2 and 3). Comparisons between different DS volumes, for both TVs, both T-piece devices, with and without leakage, were all statistically significant, p=<0.001, in one-way ANOVA, post hoc analysis with Bonferroni. Statistical analyses and 95% CI are available as a supplement (online supplemental tables 2,3).
Oxygen measured at the patient interface for the first 90 s after adjustments from FiO2 1.0 to 0.5, downward titration and from FiO2 0.3 to 0.6, upward titration. On the left simulations with tidal volume 4 mL and dead space 13 mL: On the right simulations with tidal volume 10 mL and dead space 26 mL. Orange line; rPAP with prongs (<1 mL dead space). Blue line; rPAP with 13 mL and 26 mL added dead space. Green line; Neopuff with 13 mL and 26 mL added dead space. Measurements with leaks. FiO2, fraction of inspired oxygen.
Oxygen measured at the patient interface for the first 90 s after adjustments from FiO2 1.0 to 0.50, downward titration and from FiO2 0.3 to 0.6, upward titration. On the left simulations with tidal volume 4 mL and added dead space 13 mL: On the right simulations with tidal volume 10 mL and added dead space 26 mL. Orange line; rPAP with prongs (<1 mL dead space). Blue line; rPAP with 13 mL and 26 mL added dead space. Green line; Neopuff with 13 mL and 26 mL added dead space. Measurements without leaks. FiO2, fraction of inspired oxygen.
Discussion
This bench study shows that with current equipment used for neonatal stabilisation at birth there is a substantial time delay between when FiO2 is adjusted at the blender and when the new intended FiO2 is achieved at the patient interface. This is particularly prolonged with the system DS of a commonly used face mask interface and much reduced with a low DS nasal prong interface. Our results indicate that during PPV in the smallest infants, when changing the O2 concentration at the blender from 30% to 60%, the concentration may reach 40% first after 30 s and 50% after almost 1.5 min to only come close to the target concentration of 60% nearly 3 min after adjustment. The same is likely to apply for downward titration from 100% to 50% where the O2 concentration may be at 80% 30 s after changes made at the blender and at 60% after approximately 2 min. Two previous studies identified a time lag in oxygen adjustments8 9 but did not measure the additional substantial effect of the mask DS in the set-up. These time delays are clinically important because they are a lot longer than clinicians manually titrating FiO2 against SpO2 or automated oxygen adjustment systems would allow between FiO2 changes.9 14 These time delays may explain the observation that clinicians are frequently undershooting and later overshooting the target SpO2 in the delivery room. This may result in unnecessary oxygen exposure which in turn may contribute to hyperoxia in the first few minutes of life. This may also explain why many preterm infants are not able to achieve SpO2 of 80% at 5 min after birth. SpO2<80% at 5 min is associated with death and severe intraventricular hemorrhage (IVH) in preterm infants <32 weeks gestational age (GA).15 The results also highlight the potential value of further clinical studies comparing different interfaces for respiratory support during initial stabilisation.
The reason we opted to use the time to reach a FiO2 0.05 from the set target as the primary outcome is that a change in FiO2 over time is exponential in our setup. That means that the final few percent of change takes a long time with the majority of change occurring earlier. This is also in harmony with methods used earlier.9
In a study from 2017, Hishikawa et al reported delayed oxygen titration during face mask CPAP in term infants. Their explanation was that due to DS in the face mask, ventilation was less efficient. When leaks were added to their model by making holes in the face mask, oxygen titration time was shorter.16 Similar results were reported by Dekker et al where leaks reduced the time needed to achieve the desired FiO2 after adjustments using the T-piece system during PPV.9 This concurs with our results on the shorter time needed when leaks were introduced in the model compared with measurements without leaks. Studies on PPV with face mask have shown that leakage is common and is often very large.12 13 However, with the increasing use of RFM during stabilisation the users get feedback on the leak in real time which leads to less leakage.17 This facilitates stable PEEP and the establishment of functional residual capacity (FRC). However, reduced leaks will increase the delay of accurate oxygen delivery to the patient after adjustments.
The ratio between DS and TV (DS/TV) in mechanically ventilated preterm infants has been shown to be around 0.4–0.8.18 Placing a face mask on the infant during stabilisation adds DS volume to the anatomical DS. The DS of the neonatal face mask greatly exceeds the anatomical DS making the DS/TV ratio during face mask ventilation for the infants unfavourable. The smallest infants with respiratory distress syndrome (RDS) and non-compliant lungs will have a reduced TV before surfactant is given. FRC will also be low initially and will then gradually increase. All of these factors may influence oxygen delivery after adjustments.
In all the experiments performed, we used ventilation rates of 60 with an inspiratory to expiratory (I:E) ratio of 1:1 and a PEEP of 6 cm H2O. If lower inflation rates were used, which are included in some international guidelines,4 the oxygen adjustment time may be longer. The delay in change in FiO2 due to DS will also apply when providing PEEP/CPAP to a spontaneously breathing infant with a face mask. The delay could be different compared with the presented PPV results depending on spontaneous respiratory rate, lower mask leakage and wash out of DS. Furthermore, the titration frequency and increments will affect the oxygen concentration at the airway level. All of the above should be investigated in future research, preferably in clinical settings.
Limitations
This study reflects the delay observed in vitro. These results need to be confirmed in a clinical study. The added DS in our respiratory circuit model was a straight cylinder and not in the face mask shape. Different types and brands of face masks will have different volumes. However, the experiment was only designed to show the relationship between added DS and TV and the effect that this has on oxygen delivery time after adjustments. In the clinical setting leakage, inflation rate, I:E ratio, TV, PEEP, glottic closure, intermittent spontaneous breathing and FRC will be variable and this could not be accounted for in our lab settings.
Conclusion
There is a substantial delay in the time taken for changes in FiO2 made at the blender to reach the patient’s airway during simulated ventilation when face mask DS is added to the respiratory circuit used for neonatal stabilisation. This should be recognised in guidelines and research protocols. Using nasal interfaces with low DS might be more appropriate for infants to minimise the risk for hypo-oxygenation and hyper-oxygenation during initial stabilisation.
Data availability statement
Data are available upon reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
Not applicable.
Acknowledgments
Mathilda Stenmark, MD, helped with designing and producing figure 1.
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
X @kapadianeo
Contributors KG: Study design, data collection, data and statistical analysis, manuscript writing and review. BJS, EEF, VK: Study design, manuscript writing and review. TD: Study design, data collection, equipment development, manuscript writing and review. SD: Study design, manuscript writing and review. Study guarantor.
Funding The study was funded by Region Stockholm ALF agreement, clinical research appointment (2020-0302).
Competing interests TD is one of the designers of the original rPAP respiratory support system.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.