Impact of flow‐dependent imposed work of breathing for optimising neonatal CPAP with a T‐piece device

1. AIM

T‐piece systems (Neopuff®/Perivent®) with constant resistance technology for generating CPAP (continuous positive airway pressure) are frequently used during neonatal resuscitation. ¹ They may cause a significantly higher imposed work of breathing (iWOB; up to 90%) than other CPAP devices. ² Imposed work of breathing might contribute to respiratory exhaustion ³ and should thus be minimised.

A potential approach to decreasing iWOB of this widely used CPAP device during perinatal adaptation ¹ may be through reducing resistance. ⁴ , ⁵

We explored the flow‐dependency of the inspiratory iWOB (iWOB_insp) for two simulated patients with different CPAP settings and multiple flow levels.

2. METHODS

We configured a Neonatal Active Lung Model (NALM, Dr Schaller Medizintechnik, Dresden, Germany) to emulate two neonatal patients receiving CPAP (a healthy 1000 g pre‐term and a 3000 g newborn with respiratory impairment). A 3D‐printed dummy (data available upon request) was connected to a T‐piece device (Perivent®, Fisher & Paykel Healthcare, Auckland, New Zealand) using a typical neonatal facemask ¹ (35 resp. 42 mm, Fisher & Paykel). Before starting simulated breathing, the desired CPAP level was set using the graphical user interface of the NALM. Pressure fluctuations during simulated breathing were measured (.tdms data file) and imported using the Excel add‐in from National Instruments (TDM‐Importer Version 21.3.049496). As this type of NALM did not provide an automated calculation of iWOB, it was calculated as the integration of the pressure–volume curve (trapezoidal approximation) of Δp (difference of P _y (pressure on the y‐piece) to CPAP level) to ΔV (difference in lung volume) using custom scripts in Matlab (MATLAB R2022b, Natick, MA). Ventilator breaths were analysed based on the increase in P _Y during inspiration. Twenty‐five breaths in steady state (from the fifth inspiration onwards) were evaluated per simulated patient using various flow (8, 12 and 15 L/min) and CPAP settings (3, 5, 7 and 10 cmH₂O). As expiration requires active breathing only at extremely high expiratory resistance, we only calculated the inspiratory iWOB (iWOB_insp).

Between measurements, the ventilator was disconnected, gas flow was interrupted, and the simulated breathing stopped. Measurements were excluded if CPAP levels before/after starting simulated breathing differed by >5%.

Statistics were performed using a two‐factor analysis of variance (ANOVA) for considering not only different flow levels but also slightly different mask positions. Pairwise analysis was performed with the Tukey–Kramer test as a post hoc test for comparing the different flow levels of each simulated patient and CPAP level.

3. RESULTS

During the evaluation of 1200 simulated breaths, a significant iWOB_insp reduction (p < 0.005) was observed when increasing the flow (Figure 1). Thus, increasing flow rates from 8 to 15 L/min led to a statistically significant decrease in iWOB_insp by 21%–33% which is 0.08–0.19 mJ/breath for the simulated pre‐term, resp. 0.58–1.80 mJ/breath for the simulated term infant (Figure 1).

FIGURE 1.

Boxplots showing the impact of flow rates (L/min) on measured iWOB_insp at different CPAP (cmH₂O) levels. (A) Simulated pre‐term, (B) simulated term_p. Boxplots depict the measured iWOB_insp for a total of 1200 simulated breaths, comprising 600 in pre‐term and 600 in term infants. Flow rates were assessed at 8, 12 and 15 L/min, categorised CPAP levels of 3, 5, 7 and 10 cmH₂O. Physiological iWOB was defined based on in vivo data as 1.18 mJ/breath for simulated pre‐term infants and 4 mJ/breath for term infants (p < 0.005) at all flow levels.

Increasing flow rates even from 12 to 15 L/min reduced iWOB_insp by 6%–24%, which are 0.02–0.14 mJ/breath for the simulated pre‐term, resp. 0.20–0.82 mJ/breath for the simulated term infant. Data are available upon request.

4. DISCUSSION

While increased flow led to a decreased iWOB_insp in nearly every simulated scenario, no adverse effects regarding CPAP levels or the simulated patient were observed. Neither flow nor CPAP levels influenced the accuracy of the target CPAP level.

Consistent with others, ⁴ , ⁵ we observed a reduction in iWOB_insp with increasing flow. While Kuypers et al. examined two flow rates (8 L/min vs. 12 L/min) in a cohort of 54 infants, ⁵ and Wald et al. investigated a single simulated patient (Vtid = 6 mL), one CPAP level (6 cmH₂O) and two flow rates (6 L/min vs. 15 L/min), ⁴ our findings confirmed these results and investigated a wider range of flow rates and CPAP levels. Our data indicate that irrespective of tidal volume, CPAP level or flow rate, every simulated patient experienced a reduction in iWOB_insp with increasing flow. The missing significance of the decreased iWOB_insp (CPAP 3 cmH₂O, simulated pre‐term, Flow 12 vs. 15 L/min) may be due to the small pressure and tidal volume used. As CPAP 3 cmH₂O is rarely used in sick infants, a reduction of iWOB_insp in this setting is not always necessary. As we detected an increase in iWOB_insp (CPAP 7 cmH₂O, simulated pre‐term, Flow 8 vs. 12 L/min) only once, this might be a measurement inaccuracy and should be checked by other researchers.

Kuypers et al. observed no significant clinical alterations in their cohort of 54 patients when flow was increased from 8 to 12 L/min, while maintaining a CPAP level of 6 cmH₂O. ⁵ Considering that the reduction in iWOB_insp could be up to 45% of the total physiologic iWOB at higher CPAP levels (7/10 cmH₂O, data upon reasonable request), it might be possible to further reduce iWOB_insp through additional flow augmentation, particularly at higher CPAP levels.

Of course, the possible adverse effects of higher flow rates must be considered.

Limitations of this study include that although T‐piece devices are often used during resuscitation, we did not investigate the effects on PPV. Another limitation is that in vitro models may not fully replicate the complex physiological interactions seen in patients, potentially leading to performance differences compared to actual clinical settings (e.g. mask leak, apnea, irregular breathing, sighs). Additionally, only V _t and f were used to simulate different patients while compliance and resistance were kept constant. This approach facilitates a better comprehension of the relationship between V _t /f and flow but also limits the realistic representation of simulated patients (as this simulates a very healthy pre‐term and a very sick term infant). Nevertheless, this has no influence on respiratory pressure fluctuations in the t‐piece. Unfortunately, it was not possible to set a fixed tidal volume at the NALM, therefore, tidal volumes and consecutively iWOB_insp differed slightly.

Although the clinical benefit of increasing flow in CPAP devices is uncertain, these in vitro data showed its potential for simply decreasing the resistance and iWOB of this commonly used t‐piece device in neonatal CPAP.

AUTHOR CONTRIBUTIONS

Hanna Sterzik: Conceptualization; investigation; writing – original draft; methodology; data curation; software; formal analysis; visualization. Joerg Arand: Conceptualization; investigation; writing – review and editing; supervision; data curation; validation. Kriszta Molnar: Conceptualization; writing – review and editing; supervision; data curation; validation. Christian F. Poets: Writing – review and editing; validation; methodology; supervision. Bianca Haase: Conceptualization; investigation; funding acquisition; writing – original draft; methodology; visualization; writing – review and editing; software; formal analysis; supervision.

FUNDING INFORMATION

BH was supported by an intramural TÜFF (Tuebinger Frauenfoerderung; grant number 2742‐0) from the Faculty of Medicine, University of Tuebingen.

CONFLICT OF INTEREST STATEMENT

BH works as a consultant at PFM Cologne. CFP received funding from Fritz Stephan GmbH and Loewenstein Medical.

ETHICS STATEMENT

The trial did not involve humans.

DATA AVAILABILITY STATEMENT

STEP files for 3D‐printed nasopharyngeal space dummies, the custom‐written MATLAB script for calculation of iWOB and statistical analysis (with German comments) and the post hoc tables are available upon reasonable request.