Time needed to achieve changes in oxygen concentration at the T-Piece resuscitator during respiratory support in preterm infants in the delivery room

Abstract

OBJECTIVE:

To measure the time needed to achieve changes in fraction of inspired oxygen concentration (FiO₂) from the oxygen blender to the facemask during simulated neonatal resuscitation.

METHOD:

Two oxygen analyzers were placed at each end of the T-Piece. During simulated ventilation, the duration to achieve the set oxygen concentration at the facemask was measured. This was repeated at different gas flow rates (5 L/min, 8 L/min or 10 L/min) and different FiO₂ changes (0.21 to 1.0 to 0.21, with stepwise increases and decreases in 0.05, 0.1 and 0.2 increments).

RESULTS:

A total of 1134 measurements (378 measurements for each flow) were recorded. Overall, the mean (± SD) time required to achieve FiO₂ changes at 5 L/min, 8 L/min and 10 L/min was 36±15 s, 31±14 s and 28±14 s, respectively.

CONCLUSION:

There was a lag time of approximately 30 s to achieve the FiO₂ at the facemask. This delay needs to be considered when making serial adjustments to FiO₂ during neonatal resuscitation.

Optimal management of oxygen during neonatal resuscitation is particularly important because of the evidence that either insufficient or excessive oxygenation can be harmful to newborn infants (1,2). Current resuscitation guidelines recommend beginning with air rather than 100% oxygen in term infants receiving respiratory support (1). In preterm infants, air or blended oxygen and air may be administered judiciously and, ideally, guided by pulse oximetry during respiratory support (1,2). However, if there is no increase in heart rate despite effective ventilation or if oxygenation (guided by oxygen saturation [SpO₂]) remains unacceptable, a higher fraction of inspired oxygen (FiO₂) should be considered (1,2). The aim of the present study was to measure the time needed to achieve changes in the delivery of FiO₂ from the oxygen blender to the face mask.

METHODS

The present study was performed at The Royal Alexandra Hospital (Edmonton, Alberta), a tertiary perinatal centre admitting approximately 350 infants with a birth weight of <1500 g to the neonatal nursery annually. A 50 mL test lung (Draeger, Germany) was attached to a T-piece device (Giraffe Warmer, GE Health Care, Canada). A T-piece device is a continuous-flow, pressure-limited device with a built-in manometer and a positive endexpiratory pressure valve. An oxygen analyzer (Maxtec, USA) was placed immediately after the oxygen blender of the Giraffe Warmer. A second oxygen analyzer was placed just before the test lung (Figure 1). Positive pressure ventilation was provided according to the local default settings, with a peak inflation pressure of 24 cmH₂O, positive end-expiratory pressure of 6 cmH₂O and a respiratory rate between 40 breaths/min and 60 breaths/min in all trials; however, gas flow was changed to 5 L/min, 8 L/min or 10 L/min. During positive pressure ventilation, the FiO₂ at the oxygen blender was changed from 0.21 to 1.0 to 0.21, with stepwise increase and decrease in increments of 0.05, 0.1 and 0.2, always starting at 0.21. The duration (in seconds) until the set oxygen concentration was achieved at the test lung was recorded. Each measurement was repeated three times for each group: with test lung attached and no leak (group 1); with 50% leak within the test lung (group 2) to simulate median leak during mask ventilation (3); and without a test lung attached to the T-piece (group 3) to simulate respiratory support via continuous positive airway pressure. The oxygen analyzers were calibrated before the experiments and have an accuracy of ±3% (manufacturer data). All values were measured with a stopwatch and recorded onto a spreadsheet. The response time of the two oxygen analyzers were measured and deducted from the duration in seconds. Data are presented as mean ± SD. Data were compared using an ANOVA for repeated measures with a Bonferroni post-test. P values are two-sided and P<0.05 was considered to be statistically significant. Statistical analyses were performed using Stata (Intercooled 10, Statacorp, USA).

Figure 1).

Experiment setup with T-piece, two oxygen analyzers and test lung

RESULTS

A total of 1134 measurements were recorded, with 378 measurements for each flow rate. Overall, the time required to achieve changes in FiO₂ at 5 L/min, 8 L/min and 10 L/min was 36±15 s, 31±14 s and 28±14 s, respectively. Table 1 shows the time needed to achieve changes in FiO₂. Although the time to achieve FiO₂ changes was reduced with higher gas flow, the differences were not significant. Overall, the time needed to achieve change in oxygen concentration was similar within groups at different flow rates during stepwise increase or decrease at 0.05, 0.1 or 0.2. In addition, no difference in the time needed to achieve a change in oxygen concentration from 0.21 to 1.0 or 1.0 to 0.21 was observed at any flow rate and between study groups with or without leak and test lung.

TABLE 1.

Overall time to reach oxygen concentration, s

Change in fraction of inspired oxygen	Oxygen flow rate, L/min
	5	8	10
0.21 to 1.0	42±13	39±15	33±12
1.0 to 0.21	45±13	40±13	34±10
0.21 to 1.0 in 0.05 increments	29±6	24±6	20±6
1.0 to 0.21 in 0.05 increments	32±5	24±7	21±6
0.21 to 1.0 in 0.1 increments	34±13	31±11	28±12
1.0 to 0.21 in 0.1 increments	43±18	38±17	35±17
1.0 to 0.21 in 0.2 increments	33±11	28±11	28±10
1.0 to 0.21 in 0.2 increments	42±18	38±22	34±20

Data presented as mean ± SD

DISCUSSION

Since 2010, the neonatal resuscitation guidelines advise that administration of supplementary oxygen in newborn infants should be titrated against preductal SpO₂ values (1,4). However, lung aeration and establishment of a functional residual capacity needs to occur before supplementary oxygen will be effective (5–7). Therefore, effective ventilation is required before alterations to oxygen therapy can be considered (7). Dawson et al (7) recently described an approach to guide supplementary oxygen delivery during neonatal resuscitation: once effective ventilation is confirmed, if SpO₂ remains below the 10th percentile, then FiO₂ should be increased until SpO₂ is above the 10th percentile. Ongoing supplemental oxygen should be given to target the 50th percentile and avoid high SpO₂ >90% (7). Goos et al (8) observed considerable variations of SpO₂ from the published target ranges within the first minutes after birth. The variation in the first minutes after birth were likely due to an inability to control the SpO₂, whereas later deviations were due to weaning, pauses in respiratory support (ie, intubation) and over-exposure to oxygen (8), as well as the physiological instability during resuscitation. Our unpublished observations are similar to those of Goos et al (8) – clinical staff struggle to stay within the target range and adjust FiO₂ levels rather quickly (eg, every 10 s to 20 s) to stay within the published references ranges. We aimed to measure the time needed to achieve FiO₂ changes during respiratory support. Our study shows that it takes approximately 30 s until the FiO₂ concentration change at the blender is similar to the oxygen delivery at the face mask. There were several limitations to the present simulated study including variations in equipment set-up (eg, using a neopuff, bagger or even different length of tubing for gas). Detailed examination of the rate of FiO₂ change or how the FiO₂ fluctuated during the interval would be interesting. Furthermore, there are other clinical factors that will contribute to the translation of our findings into clinical practice, in which the titration of FiO₂ depends on the change in pulse oximetry in the patient. The clinical factors include additional delay between the time the gas enters the lungs and changes in pulse oximetry (eg, lung recruitment, rate of bagging, inadequate airway penetration, poor perfusion, delays in obtaining accurate oximetry readings). Nonetheless, our findings demonstrate the significant delay in the change in oxygen exposure at the facemask after the change in oxygen concentration at the blender. A ‘patience’ approach is important to avoid flip-flop hypoxia-hyperoxia fluctuations.

CONCLUSION

In the present simulated ventilation study investigating the titration of FiO₂ at the oxygen blender, there was a lag time of approximately 30 s to achieve the FiO₂ at the facemask. This delay needs to be considered when making serial adjustments to FiO₂ during neonatal resuscitation.