Authors: Shu-Ting Yang, Hao-Wei Chung, Hsiu-Lin Chen
Categories: Research, High-flow nasal cannula, Nasal continuous positive airway pressure, Premature infant
Nasal continuous positive airway pressure (NCPAP) is widely used for premature infants with respiratory distress syndrome (RDS). A high-flow nasal cannula (HFNC) provides positive end-expiratory pressure using high-flow oxygen; however, the variability in distending pressure is a primary concern. This study evaluated the feasibility and safety of a newly designed protocol for NCPAP weaning with cyclic HFNC use for premature infants.
Premature infants with RDS using NCPAP support who were ready for weaning were enrolled. The weaning protocol used cyclic NCPAP with HFNC every 3 h for 3 days in the neonatal intensive care unit. The heart rate (HR), respiratory rate (RR), pulse oximetry (SpO2), transcutaneous carbon dioxide (PtcCO2), and cerebral tissue oxygen saturation (StO2) at the end of NCPAP with HFNC support were recorded once daily for 3 days.
From June 2019 to April 2021, 46 premature infants (27 male, 19 female) were enrolled. The mean gestational age and birth body weight were 28.7 ± 2.6 weeks and 1181 ± 354 g, respectively. No statistically significant differences in the HR, RR, SpO2, and cerebral StO2 during NCPAP weaning with HFNC were observed. However, the mean PtcCO2 with NCPAP was statistically significantly lower than that with HFNC (46.9 ± 6.0 mmHg vs. 47.9 ± 6.4 mmHg, P = 0.02).
The feasibility and safety of the NCPAP weaning protocol with cyclic HFNC for premature infants are acceptable in this preliminary study. Due to the limited number of participants, further studies are required for more comprehensive analysis.
This prospective observational case study was approved by the Human Experiment and Ethics Committee of our hospital (approval KMUHIRB-SV(I)-20180059; approval January 11, 2019).
Keywords: Nasal continuous positive airway pressure, High-flow nasal cannula, Premature infant
Respiratory distress syndrome (RDS) is a common cause of neonatal intensive care unit (NICU) admission for premature infants [1–3]. Non-invasive respiratory support for premature infants includes nasal intermittent positive pressure ventilation, nasal continuous positive airway pressure (NCPAP), and high-flow nasal cannula (HFNC) use. NCPAP has been widely used to treat RDS in premature infants by providing continuous positive pressure to prevent alveolar prolapse and stabilize functional residual capacity [1–3]. However, the need for prongs that completely fit the nostrils may damage the nasal mucosa and septum, and the higher positive airway pressure may also induce complications such as pneumothorax, pneumomediastinum, and abdominal distension [1, 2].
Conversely, the HFNC provides positive end-expiratory pressure (PEEP) using high-flow oxygen (2–8 L/min) for infants [1, 2, 4, 5]. The advantages of the HFNC include a lower risk of injury to the nares, increased comfort, and ease of use. However, the variability in PEEP is one of the primary concerns associated with HFNC use [1, 2, 4, 5]. The HFNC has been considered an NCPAP weaning device (step-down) for premature infants [1, 2, 4, 5]. This study aimed to design and evaluate the feasibility and safety of a new protocol involving cyclic HFNC use for NCPAP weaning among premature infants.
We designed a prospective observational study involving premature infants (gestational age [GA] < 37 weeks) with RDS using NCPAP support who were ready for weaning. Participants were enrolled between June 6, 2019 and April 13, 2021. The evaluation criteria for NCPAP weaning included a fraction of inspired oxygen (FiO2) of 0.21, PEEP of 4–5 cm H2O, with relatively stable vital signs, and no episodes of apnea while supported by NCPAP support for 3 days. The exclusion criteria were term infants and preterm infants with major birth defect. The major birth defect included chromosome anomalies, congenital heart defects, neural tube defects, and congenital gastrointestinal defects (such as gastroschisis or omphalocele).
This prospective observational case study was approved by the Human Experiment and Ethics Committee of our hospital (approval KMUHIRB-SV(I)-20180059; approval January 11, 2019). All experiments were performed in accordance with relevant guidelines and regulations. Written informed consent was obtained from the parents of the included premature infants.
The 3-day weaning protocol comprised alternating 3-hour sessions of NCPAP and cyclic HFNC in each participant in the NICU at our hospital, beginning from 30 AM with NCPAP. The flow rate of HFNC was adjusted to match the pressure levels during its use, as measured by a GiO Digital Pressure Gauge (GIO 6, GaleMed Corporation, Taipei, Taiwan) at the distal end of the nasal cannula. The pressure was set to correspond with the PEEP used during the preceding NCPAP treatment. The flow rate of HFNC was maintained at 4–6 L per minute throughout the study. The heart rate (HR), respiratory rate (RR), pulse oximetry (SpO2), transcutaneous carbon dioxide (PtcCO2), and cerebral tissue oxygen saturation (StO2) were recorded for 0.5 h. Then, the respiratory support was changed to HFNC use for 3 h starting at 30 AM. We again recorded the HR, RR, SpO2, PtcCO2, and cerebral StO2 for 0.5 h (Fig. 1). The primary outcomes were the differences in the mean values of HR, RR, SpO2, PtcCO2, and cerebral StO2 recorded during NCPAP and HFNC. After cyclic use for 3 days, neonatologists in the NICU evaluated the clinical condition of the infants and assessed if they could tolerate continuous HFNC support until complete weaning was achieved.
Fig. 1 Daily weaning protocol for 3 days
This cyclic approach was chosen in line with the standard nursing care schedule in our NICU, which follows a 3-hour cycle. We also believed that providing respiratory support via HFNC in 3-hour intervals would offer more consistent support to preterm infants than the shorter cycles would [6]. The initiation of cyclic HFNC use for 3 days in enrolled preterm infants was based on the inclusion criterion of no apnea occurrences while on NCPAP support for 3 consecutive days. We monitored the frequency of apnea during the cyclic use of NCPAP and HFNC. The absence of apnea during the 3-day cyclic period was considered indicative of successful weaning from NCPAP to HFNC.
However, some infants could not successfully wean from NCPAP using the HFNC during the protocol due to observed respiratory distress, which included a heart rate (HR) > 160 beats per min, a respiratory rate (RR) over 60 cycles per min, an oxygen saturation (SpO2) below 90%, or PtcCO2 above 60 mmHg during the 3-day weaning period. Therefore, we further divided the study participants into those successfully weaned from NCPAP using the HFNC within 3 days (success group) and those weaned after > 3 days (failure group).
Data on the sex, GA, birth body weight (BW), delivery mode, Neonatal Therapeutic Intervention Scoring System score, Apgar score, medication used to treat apnea, respiratory therapy condition, post-menstrual age (PMA), BW when starting weaning from NCPAP, apnea frequency, vital sign changes, and possible adverse effects associated with the HFNC of all enrolled infants were collected.
This study used the Optiflow System HFNC (Fisher & Paykel Optiflow System Healthcare, Auckland, New Zealand); short binasal prongs with different sizes were chosen based on the infant’s BW. NCPAP was administered using the Babi.Plus^®^ Bubble CPAP system (GaleMed Corporation, Taipei, Taiwan). The SenTec Digital Monitoring System (SenTec AG, Therwil, Switzerland) was used to measure PtcCO2. The FORE-SIGHT Oximeter MC-2000 Series Cerebral Oximeter (CAS Medical Systems, Inc., Branford, CT, USA) was used to measure cerebral StO2.
Data recording and evaluation were performed using JMP 10 software (SAS Institute Inc., Cary, NC, USA). The rank sum test was performed to compare numerical variables of primary outcomes during NCPAP and HFNC and the characteristics and clinical outcomes in the success and failure groups. Univariate regression analyses were performed to analyze the factors associated with the PtcCO2 of preterm infants using NCPAP and HFNC. The chi-square test was performed to compare categorical variables in the success and failure groups. In contrast, the rank sum test was performed to compare numerical variables of primary outcomes, respiratory outcomes, and prognoses in the success and failure groups.
Fifty premature infants admitted to the NICU at our university hospital who met the inclusion criteria underwent the weaning protocol after written informed consent was received from their parents. Four participants were excluded due to severe complications, such as severe pulmonary hypertension or hydrocephalus (Fig. 2). Therefore, 46 participants were included in this study. No adverse effects were observed among the remaining 46 infants. The mean GA was 28.7 weeks (standard deviation (SD), ± 2.6 weeks), whereas the mean birth BW was 1181 g (SD, ± 354 g). The characteristics and clinical outcomes of the 46 infants are shown in Table 1.
Fig. 2 Flow chart of the study design
The mean PMA during NCPAP weaning was 35.6 weeks (SD, ± 2.3 weeks), and the mean BW was 1990 g (SD, ± 527 g). The mean apnea frequency during the 3-day cyclic use of NCPAP and HFNC was less than once daily. The primary outcomes (mean ± SD) of 46 enrolled infants, in whom NCPAP and HFNC support were administered, revealed no statistical differences in the HR (162 ± 13 beats per min under NCPAP, 163 ± 13 beats per min under HFNC, P = 0.34), RR (46 ± 12 cycles per min under NCPAP, 46 ± 13 cycles per min under HFNC, P = 0.71), SpO2 (97 ± 3% under NCPAP, 97 ± 3% under HFNC, P = 0.46), and cerebral StO2 (75.3 ± 5.8% under NCPAP, 74.9 ± 5.2% under HFNC, P = 0.58) during NCPAP weaning with HFNC. The median of SpO2 and cerebral StO2 under NCPAP and HFNC were 97% vs. 97% and 75.8% vs. 75.5%. The PtcCO2 was higher in this population with HFNC support (47.9 ± 6.4 mmHg) than with NCPAP support (46.9 ± 6.0 mmHg; P = 0.02).
Table 2 shows the associated factors influencing the differences in PtcCO2 with cyclic use of NCPAP and HFNC on univariate regression analysis. None of these factors affected the differences in PtcCO2.
Eighteen premature infants could not be completely weaned from NCPAP to HFNC after 3 days of protocol and required an extended period of weaning (defined as the failure group). There were no statistical differences in infants’ characteristics and clinical outcomes in the success and failure groups (Table 3). The mean apnea frequency during the 3-day course of cyclic use of NCPAP and HFNC in both groups was less than once daily.
Table 4 presents the primary outcomes of infants in the success and failure groups with NCPAP or HFNC support. The median of SpO2 under NCPAP or HFNC in the success and failure groups were 98% vs. 97% and 98% vs. 97%, respectively. The median of cerebral StO2 under NCPAP or HFNC in the success and failure groups were 76.0% vs. 75.7% and 75.9% vs. 74.6%, respectively. The cerebral StO2 with HFNC support was lower for infants in the failure group than in the other groups.
Table 5 shows infants’ respiratory outcomes and prognoses in the success and failure groups. The failure group required more NCPAP weaning days, HFNC usage days, and total days of respiratory therapy. The PMA at discontinuation of NCPAP and HFNC was greater in the failure group than in the success group.
Our study investigated the feasibility and safety of the NCPAP weaning protocol with the cyclic use of an HFNC. This protocol showed no differences in HR, RR, SpO2, and cerebral StO2; however, the PtcCO2 was higher during HFNC use than during NCPAP use. No adverse effects were observed during this study.
The differences in PtcCO2 among premature infants during NCPAP and HFNC use in our study may be associated with lower respiratory tract support pressure stability with the HFNC than with NCPAP [4, 5]. Although the difference in PtcCO2 levels between HFNC and NCPAP support was statistically significant—with PtcCO2 being slightly higher in HFNC support (47.9 ± 6.4 mmHg) compared to NCPAP (46.9 ± 6.0 mmHg; P = 0.02)—the clinical relevance of this difference may be considered minor. This suggests that while NCPAP provides more stable expiratory pressure and thus more effective CO2 washout, HFNC still serves as a viable option for weaning from NCPAP due to its sufficient support capabilities. Notably, most studies have focused on comparing NCPAP and HFNC support as the main respiratory therapy for premature infants after birth. Lampland et al. found no differences in the HR and arterial oxygen saturation during NCPAP and HFNC use in premature infants, but the respiratory rate was higher with the HFNC [7]. However, Taha et al. reported that HFNC use resulted in higher mortality and bronchopulmonary dysplasia rates, prolonged hospital stay, and longer time for oral feeding than NCPAP use among newborns [8]. Therefore, the American Academy of Pediatrics suggested HFNC support as an alternative respiratory therapy for infants after extubation rather than the main respiratory therapy for premature infants after birth [9, 10]. As previously mentioned, HFNC support has been suggested as an accompanying respiratory therapy for NCPAP weaning in premature infants because of the variable stability of the airway support pressure [1–3, 5, 11–15]. Consequently, we inferred that the PtcCO2 was higher with HFNC support than with NCPAP support, possibly due to the greater stability of the airway support pressure of NCPAP.
There were no obvious differences in cerebral StO2 with NCPAP or HFNC support among the 46 premature infants in our study. Sett et al. found no obvious differences in cerebral StO2 when performing NCPAP weaning with HFNC use in premature infants [16]. Bemdich et al. found no influence of cerebral StO2 and cerebral blood flow with PEEP of 3–8 cm H2O [17]. Combining our findings with those of the literature mentioned above, it is evident that using the HFNC for NCPAP weaning did not influence cerebral StO2. Therefore, the SpO2 and cerebral StO2 did not negatively influence premature infants when the HFNC was used during the NCPAP weaning process, which indicates the safety of the cyclic use of HFNC for NCPAP weaning.
Our study design involved the cyclic use of NCPAP and HFNC every 3 h for NCPAP weaning for premature infants with RDS; however, 18 (39%) premature infants required cyclic use for > 3 days to achieve complete NCPAP weaning (failure group). We inferred that lower GA, lower BW, higher NTISS, and late-onset sepsis may be associated with a higher failure rate due to the unstable condition of premature infants. However, the results showed no significant differences in clinical outcomes and complications, which may have been influenced by the small sample size (Table 3). Additionally, we were unable to measure lung function directly in these preterm infants, some of whom might inherently have delayed lung maturation. A lower cerebral StO2 was observed in the failure group during NCPAP weaning using the HFNC for 3 days, indicating more unstable oxygenation with HFNC support (Table 4); therefore, the infants required more number of days to achieve successful NCPAP weaning (Table 5). Sett et al. found no obvious differences in the cerebral StO2 when performing NCPAP weaning with 6 cm H2O using the HFNC at 8 L/min; however, they adjusted the fraction of inspired oxygen for the HFNC to maintain the SpO2 at approximately 92–95%, which could be why their results were different from ours [16]. The PMA at discontinuation of NCPAP or HFNC use was older, and the total number of days of respiratory therapy was longer in the failure group (Table 5). The length of hospitalization and PMA at discharge were higher in the failure group, and the longer respiratory therapy course might influence both groups. However, the statistical results showed no obvious difference between the success and failure groups, which might be associated with the small sample size in our study (Table 5).
This study had some limitations. The small sample size might have influenced the analysis results. In this prospective study, the HFNC devices were funded by Kaohsiung Medical University Hospital. Due to the limitations of the grant, we were only able to purchase approximately 50 devices. Additionally, this study is structured similarly to a cross-over study, rather than a comparative study, which precludes calculating the difference in treatment effects. As a result, this is a preliminary study. Previously, using HFNC was self-paid in Taiwan; however, because the health insurance of Taiwan began covering the use of the HFNC in 2022, further analyses with larger sample sizes can be considered in the future.
The feasibility and safety of the weaning protocol from NCPAP using cyclic HFNC support for premature infants are acceptable. During the weaning process, there were no significant changes in HR, RR, SpO2, or StO2, although PtcCO2 levels were higher with HFNC support. No adverse effects related to HFNC use were observed, and there was no increase in apnea frequency. However, due to the limited number of participants, further studies with larger samples are required for more comprehensive analysis.
Not applicable.
S.-T.Y. and H.-L.C conceived and designed of the study. S.-T.Y. and H.-W.C acquired the data of the study. S.-T.Y. and H.-L.C analyzed and interpreted the data of the study. S.-T.Y. drafted the manuscript. H.-L.C revised the manuscript critically for important intellectual content. All authors read and approved the final manuscript.
This study was supported by grants from our hospital in 2018 (KMUH107-7M26 and KMUH107-7R44).
No datasets were generated or analysed during the current study.
This prospective observational case study was approved by the Human Experiment and Ethics Committee of our hospital (approval KMUHIRB-SV(I)-20180059; approval January 11, 2019). All experiments were performed in accordance with relevant guidelines and regulations. Written informed consent was obtained from the parents of the included premature infants.
Not applicable.
The authors declare no competing interests.
No datasets were generated or analysed during the current study.