Comparison of Three Non-invasive Ventilation Interfaces in Preterm and Term Neonates: A Retrospective Cohort Study

Summary

This study of 187 preterm and term neonates demonstrates that systematic rotation between nasal prongs and masks reduces moderate-to-severe nasal injuries by 60% without compromising respiratory efficacy. This strategy also improves oxygenation and shortens the duration of continuous positive airway pressure compared with continuous interface use.

Abstract

Non-invasive ventilation (NIV) is the primary mode of support for neonatal respiratory distress; however, nasal injury remains a frequent complication. The optimal interface strategy for minimizing nasal injury while maintaining respiratory efficacy remains uncertain. This retrospective cohort study of 187 neonates compared three interface strategies: nasal prongs (n = 63), nasal mask (n = 61), and an alternating prongs/mask protocol (n = 63). Primary outcomes included NIV failure and nasal injury severity. NIV failure rates were comparable across groups (p = 0.72); however, the alternating group demonstrated a significant reduction in moderate-to-severe nasal injury (9.5%) compared with continuous prongs (23.8%, p = 0.03) and continuous masks (18.0%, p = 0.19). Furthermore, the alternating strategy was associated with significantly improved oxygenation and shorter median durations of continuous positive airway pressure (CPAP) (15 vs. 18 vs. 20 h, p = 0.04). Systematic interface rotation every 2–4 h minimizes iatrogenic injury without compromising efficacy and should be integrated into standard neonatal care.

Introduction

Nasal injury remains a significant iatrogenic complication of neonatal non-invasive ventilation (NIV), with incidence rates reported as high as 90%¹. Non-invasive ventilation has revolutionized neonatal respiratory care over the past two decades and has become the preferred initial approach for respiratory support in preterm and term infants^2,3. The shift from invasive mechanical ventilation to NIV strategies has been driven by compelling evidence demonstrating reduced rates of bronchopulmonary dysplasia, shorter hospitalizations, and improved neurodevelopmental outcomes⁴. Current guidelines from the American Academy of Pediatrics and the European Consensus Guidelines recommend continuous positive airway pressure (CPAP) as first-line therapy for respiratory distress syndrome, with NIV failure rates ranging from 25%–40%, depending on gestational age and underlying pathology^5,6.

Despite these advances, the optimal interface for NIV delivery remains controversial. Recent systematic reviews have failed to demonstrate clear superiority of any single interface type, although individual studies suggest potential benefits of specific approaches in selected populations^7,8. The heterogeneity in study designs, patient populations, and outcome measures has limited the ability to draw definitive conclusions regarding interface selection.

Nasal prongs are widely used for their reliable pressure delivery; however, their rigid structure often creates concentrated pressure points at the nasal septum and the columella^9,10,11. In contrast, nasal masks distribute pressure over a larger area but may introduce challenges such as increased gas leak and pressure variability^12,13,14,15.

The concept of interface rotation aims to combine the advantages of both systems. Early investigations suggest that systematic rotation protocols can reduce injury rates^16,17. This study addresses the lack of direct comparisons across a broad cohort of preterm and term neonates, particularly highlighting extremely low birth-weight infants, who are at the highest risk due to skin fragility and the need for prolonged NIV duration^18,19,20. By evaluating these strategies, this research provides a framework for proactive interface management, guiding clinical decision-making to balance respiratory efficacy with the prevention of iatrogenic injury. It is hypothesized that the underlying physiological mechanisms involve a reduction in localized inflammatory cascades and optimization of tissue perfusion; however, these mechanisms require further validation²¹.

This study aimed to compare the incidence and severity of nasal injuries among three interface strategies; evaluate the impact on respiratory parameters and inflammatory markers; test whether interface rotation improves gas exchange via varied mechanical stimuli; and identify predictors of NIV success across different interface approaches.

Protocol

This study was performed in accordance with the principles of the Declaration of Helsinki. Ethical approval was granted by the Institutional Review Board of the First Affiliated Hospital of Xiamen University (Approval No. [2025] KYLSZ (049)) prior to data collection. De-identified data extraction forms were used to maintain patient confidentiality. Given the retrospective nature of this study, the ethics committee waived the requirement for written informed consent.

Inclusion criteria. Neonates admitted to the facility between 2020 and 2024 were eligible for inclusion if they had a gestational age of ≥24 weeks, required non-invasive ventilation (NIV) within the first 72 h of life, and exhibited respiratory distress identified by clinical signs, including a respiratory rate >60 breaths/min, grunting, or a Silverman–Anderson score ≥3. The study intentionally included extremely preterm infants (24 weeks) and extremely low birth-weight infants (<1,000 g) to reflect the full clinical scope of NIV application.

Exclusion criteria. Infants with major craniofacial anomalies or those requiring immediate intubation were excluded.

1. Subject screening and preparation

The electronic medical database was searched for neonates admitted between 2020 and 2024 with a gestational age of ≥24 weeks who required non-invasive ventilation (NIV) within the first 72 hours of life. Respiratory distress was identified based on clinical signs, including a respiratory rate >60/min, grunting, or a Silverman–Anderson score ≥3. Infants with major craniofacial anomalies or those requiring immediate intubation were excluded. The inclusion of extremely preterm infants (24 weeks) and extremely low birth-weight infants (<1,000 g) was intentional to reflect the full clinical scope of NIV application in the unit.

2. Non-invasive ventilation interface application

The appropriate interface size was determined using a calibrated measuring guide to assess the nasal columellar distance and nostril diameter. The skin was cleaned and dried, followed by application of a thin hydrocolloid barrier. The barrier was cut into an ‘H-shape’ for nasal prongs or a ‘reverse heart-shape’ for nasal masks. The interface was then placed and secured with the fixation system, with Velcro straps adjusted to allow one finger-breadth of space to prevent excessive pressure.

Interface rotation was performed every 2–4 h for the alternating group. The timing within this interval was individualized based on clinical judgment and synchronized with routine clustered care to minimize unnecessary handling and physiological instability. This approach prioritized infant comfort, stability during handling, and the preservation of nasal skin integrity, as assessed at visual checkpoints.

During each interface switch, a 5-min pressure-relief period was provided by delivering oxygen via a low-flow cannula positioned 1–2 cm above the nares.

3. Visual checkpoints and monitoring

Visual checkpoints ensured that nasal prongs did not contact the base of the nose and that the nasal mask did not impinge on the eyes. Patients were monitored for clinical stabilization, including a reduction in respiratory rate and maintenance of oxygen saturation between 91%–95%. Blood gas samples were obtained at baseline and at 2, 6, 12, and 24 h after initiation to measure pH (potential of hydrogen), partial pressure of arterial oxygen (PaO₂), and partial pressure of arterial carbon dioxide (PaCO₂). Primary outcomes included NIV failure (intubation within 72 h) and nasal injury severity.

4. Post-procedural data analysis

All statistical analyses were performed after completion of clinical data collection to maintain procedural flow. Categorical and continuous variables were entered into statistical software for analysis. The normality of continuous data was assessed using the Shapiro–Wilk test. Clinical outcomes were compared across the three interface groups using ANOVA, the Kruskal–Wallis test, or the chi-square test, as appropriate. The Bonferroni correction was applied for multiple comparisons of secondary outcomes (blood gas parameters and inflammatory markers) to adjust the significance threshold. Multivariable logistic regression was used to identify independent predictors of nasal injury, adjusting for birth weight and CPAP duration.

Results

Baseline characteristics and demographics

The study cohort comprised 187 neonates with a median gestational age of 33+2 weeks (interquartile range [IQR]: 29+4 to 34+5) and a mean birth weight of 881.1 ± 525.7 g (range: 580–2050 g). The inclusion of extremely preterm infants (<28 weeks) and very low birth weight infants (<1,500 g) reflects the broad clinical application of NIV in the neonatal intensive care unit.

Baseline characteristics, including the distribution of infants with birth weights <1,000 g, were well balanced across the three intervention groups (Table 1). The distribution of gestational ages showed clustering around 34+4 weeks (n = 8, 4.3%), 33+2 weeks (n = 7, 3.7%), and 30+2 weeks (n = 7, 3.7%), reflecting a predominantly preterm and very low birth weight cohort. Primary reasons for prematurity varied widely, with preeclampsia/eclampsia (n = 13, 7.0%), threatened preterm labor (n = 10, 5.4%), and prelabour rupture of membranes (n = 15, 8.0%) being the most common.

Baseline characteristics were well balanced across the three intervention groups (Table 1). The nasal prong group had a slightly higher representation of extremely low birth weight infants (<1,000 g; 38.1% vs. 32.8% vs. 33.3%, p = 0.79), while the alternating group included more infants born via cesarean section (76.2% vs. 68.9% vs. 66.7%, p = 0.52). Apgar scores at 1, 5, and 10 min were comparable, with median scores of 10 at 10 min across all groups. Initial respiratory diagnoses showed a similar distribution, with respiratory distress syndrome predominating (47.6% prongs, 49.2% mask, 46.0% alternating).

Characteristic	Nasal Prongs (n=63)	Nasal Mask (n=61)	Alternating (n=63)	p-value
Gestational age, weeks				0.84
Median (IQR)	33+1 (29+2, 34+4)	33+3 (29+5, 34+6)	33+2 (29+4, 34+5)
<32 weeks, n (%)	24 (38.1)	22 (36.1)	23 (36.5)
Birth weight, g				0.91
Mean ± SD	865.3 ± 512.4	892.7 ± 539.2	885.4 ± 528.1
<1000g, n (%)	24 (38.1)	20 (32.8)	21 (33.3)
Delivery mode				0.52
Cesarean, n (%)	43 (68.3)	42 (68.9)	48 (76.2)
Vaginal, n (%)	19 (30.2)	18 (29.5)	15 (23.8)
Forceps/Vacuum, n (%)	1 (1.6)	1 (1.6)	0 (0)
Apgar scores
1 min, median (IQR)	8 (7, 9)	9 (8, 9)	9 (8, 10)	0.42
5 min, median (IQR)	9 (9, 10)	10 (9, 10)	10 (9, 10)	0.67
10 min, median (IQR)	10 (9, 10)	10 (10, 10)	10 (10, 10)	0.88
Primary diagnosis				0.95
RDS, n (%)	30 (47.6)	30 (49.2)	29 (46.0)
TTN, n (%)	18 (28.6)	16 (26.2)	19 (30.2)
Pneumonia, n (%)	8 (12.7)	9 (14.8)	8 (12.7)
Other, n (%)	7 (11.1)	6 (9.8)	7 (11.1)

Table 1: Baseline Characteristics and Demographics. This table summarises the maternal and neonatal baseline characteristics, including gestational age, birth weight, delivery mode, Apgar scores, and primary respiratory diagnosis across the nasal prongs, nasal mask, and alternating groups.

Ventilation parameters and duration

Initial CPAP settings were uniform across groups, with starting pressures of 5–6 cmH₂O. The alternating group demonstrated more rapid weaning, achieving pressure reduction to 5 cmH₂O at a median of 8 days (IQR: 6–12) compared with 12 days (IQR: 8–16) for continuous prongs and 10 days (IQR: 7–14) for continuous mask groups (p = 0.03).

Total CPAP duration differed significantly, with the alternating group requiring a median of 15 days (IQR: 6–30.25) compared with 20 days (IQR: 8–35) for the prongs group and 18 days (IQR: 7–32) for the mask group (p = 0.04).

Methylxanthine use for apnoea management was comparable across groups (69.8% prongs; 72.1% mask; 68.3% alternating; p = 0.89), with caffeine citrate as the sole agent used. The proportion of infants requiring escalation to bilevel support was lowest in the alternating group (14.3%) compared with prongs (20.6%) and mask (18.0%) groups, although this difference did not reach statistical significance (p = 0.62) (Table 2).

Parameter	Nasal Prongs	Nasal Mask	Alternating	p-value
Initial CPAP (cmH2O)	5.8 ± 0.6	5.9 ± 0.5	5.8 ± 0.6	0.72
Maximum CPAP (cmH2O)	7.2 ± 0.9	7.1 ± 0.8	6.9 ± 0.7	0.14
Time to wean (days)	12 (8-16)	10 (7-14)	8 (6-12)	0.03
Total CPAP duration (days)	20 (8-35)	18 (7-32)	15 (6-30.25)	0.04
Bilevel support required, n (%)	13 (20.6)	11 (18.0)	9 (14.3)	0.62
Methylxanthine use, n (%)	44 (69.8)	44 (72.1)	43 (68.3)	0.89

Table 2: NIV Settings and Duration. This table details the non-invasive ventilation (NIV) parameters applied, including initial and maximum continuous positive airway pressure (CPAP) levels, time to wean, total CPAP duration, and the requirement for bilevel support or methylxanthine.

Primary outcomes: non-invasive ventilation failure and nasal injury

Non-invasive ventilation failure rates within 72 h were statistically similar across groups: 15.9% (10/63) in the prongs group, 13.1% (8/61) in the mask group, and 11.1% (7/63) in the alternating group (p = 0.72).

The most common reasons for intubation included worsening respiratory acidosis (48%), recurrent apnoea (28%), and pneumothorax (16%). Time to failure showed no significant difference, with most failures occurring within the first 24 h, regardless of interface type.

Nasal injury assessment revealed marked differences between groups (Table 3, Figure 1). Any degree of nasal injury occurred in 68.3% of the prongs group, 57.4% of the mask group, and 38.1% of the alternating group (p = 0.003). Moderate-to-severe injuries (score ≥4) were significantly reduced in the alternating group (9.5%) compared with continuous prongs (23.8%, p = 0.03) and showed a trend toward reduction compared with continuous mask (18.0%, p = 0.19).

The anatomical distribution of injuries differed by interface type. Prong-related injuries predominantly affected the nasal septum (85%) and columella (62%), whereas mask-related injuries were more evenly distributed across the nasal bridge (55%), alae (40%), and philtrum (35%). The alternating group showed fewer injuries at all sites, with most presenting as mild erythema that resolved within 48 h of NIV discontinuation.

Outcome	Nasal Prongs	Nasal Mask	Alternating	p-value
NIV Failure
Overall, n (%)	10 (15.9)	8 (13.1)	7 (11.1)	0.72
Time to failure (hours)	18 (12-28)	20 (14-30)	22 (16-32)	0.81
Nasal Injury
Any injury, n (%)	43 (68.3)	35 (57.4)	24 (38.1)	0.003
None	20 (31.7)	26 (42.6)	39 (61.9)
Mild (score 1-3)	28 (44.4)	24 (39.3)	18 (28.6)
Moderate (score 4-6)	12 (19.0)	9 (14.8)	5 (7.9)
Severe (score >6)	3 (4.8)	2 (3.3)	1 (1.6)
Injury Location
Nasal septum	37 (86.0)*	12 (34.3)	10 (41.7)	<0.001
Columella	27 (62.8)*	8 (22.9)	6 (25.0)	<0.001
Nasal bridge	5 (11.6)	19 (54.3)*	8 (33.3)	<0.001
Nasal alae	8 (18.6)	14 (40.0)	5 (20.8)	0.08
*Percentage calculated from those with any injury in each group

Table 3: Primary Outcomes: NIV Failure and Nasal Injury. This table reports the primary outcomes of the study, including the incidence and timing of NIV failure, as well as the frequency, severity, and anatomical distribution of nasal injuries for each intervention group. Percentages for injury locations are calculated based on the number of infants with any nasal injury within their respective groups.

Figure 1: Nasal injury rates by interface type. The chart presents the incidence of any nasal injury and moderate-to-severe injuries across the continuous nasal prongs (n = 63), continuous nasal mask (n = 61), and alternating interface (n = 63) groups. Statistical comparisons were performed using the chi-square test. Please click here to view a larger version of this figure.

Blood gas improvements and physiological response

All groups demonstrated significant improvements in oxygenation and ventilation parameters. Pre-treatment PaO₂ values were comparable (mean: 50.0 ± 16.5 mmHg overall), with post-treatment improvements most pronounced in the alternating group. Mean PaO₂ increase at 6 h was 18.2 ± 7.3 mmHg for the prongs, 20.1 ± 8.1 mmHg for the mask, and 22.4 ± 6.9 mmHg for the alternating group (p = 0.01).

Carbon dioxide clearance showed similar patterns, with baseline PaCO₂ averaging 55.1 ± 10.4 mmHg across all groups. The alternating group achieved more rapid normalization, reaching PaCO₂ < 50 mmHg at a median of 4 h compared with 6 h for both continuous groups (p = 0.02). The pH correction followed expected patterns, with all groups achieving pH > 7.30 within 2–4 h of NIV initiation.

Inflammatory markers demonstrated notable patterns. Baseline C-reactive protein (CRP) levels were available for 94 of 187 infants (50.3%). Missingness occurred because CRP was measured based on clinical suspicion of infection. A sensitivity analysis comparing infants with and without CRP data showed no significant differences in gestational age, birth weight, or primary respiratory diagnosis (all p > 0.05).

CRP reductions were observed at 48 h across all groups. The alternating group showed the greatest reduction (median decrease: 3.8 mg/L) compared with prongs (2.9 mg/L) and mask (3.1 mg/L), although variability limited statistical significance (p = 0.24) (Supplementary Table 1, Figure 2).

Figure 2: Physiological parameter changes by interface type. Data represent (A) Partial pressure of arterial oxygen (PaO₂) change at 6 h, (B) Partial pressure of arterial carbon dioxide (PaCO₂) change at 6 h, (C) pH (potential of hydrogen) values at baseline and 6 h after initiation of non-invasive ventilation, and (D) C-reactive protein (CRP) change at 48 h. Panels A–C plot the full dataset (prongs, n = 63; mask, n = 61; alternating, n = 63). Panel D displays the subset of infants with available CRP measurements (total n = 94). Statistical significance was assessed using one-way ANOVA (for PaO₂, PaCO₂, and pH) and the Kruskal-Wallis test (for CRP). Please click here to view a larger version of this figure.

Predictors of outcomes and subgroup analysis

Multivariable logistic regression identified several independent predictors of NIV failure. Lower gestational age (odds ratio [OR]: 1.42 per week decrease; 95% confidence interval [CI]: 1.18–1.71), initial FiO₂ requirement >0.50 (OR: 3.24; 95% CI: 1.67–6.29), and PaCO₂ >65 mmHg at 2 h (OR: 2.89; 95% CI: 1.44–5.82) were significantly associated with failure. Interface type was not an independent predictor after adjustment for these factors.

For nasal injury, longer CPAP duration emerged as the strongest predictor (OR: 1.08 per h; 95% CI: 1.05–1.11), followed by birth weight <1,000 g (OR: 2.34; 95% CI: 1.28–4.27). The protective effect of the alternating interface remained significant after adjustment (OR: 0.31; 95% CI: 0.15–0.64, compared with continuous prongs).

Correlation analysis revealed relationships between physiological parameters and outcomes. Body weight showed a positive correlation with PaO₂ improvement (r = 0.42, p < 0.001), while CPAP duration inversely correlated with initial pH (r = −0.38, p < 0.001) (Supplementary Figure 1). The relationship between interface type and outcomes was consistent across weight categories, although the absolute risk reduction for nasal injury with an alternating interface was greatest in infants < 1,000 g (28.6% vs. 52.4% with continuous prongs, NNT = 4.2).

DATA AVAILABILITY:

The complete de-identified raw dataset generated and analyzed in the current study is provided as Supplementary File 1

Supplementary Figure 1: Correlation analysis of physiological parameters. (A) Birth weight vs partial pressure of arterial oxygen (PaO₂) improvement at 6 h, showing a positive correlation (r = 0.42, p < 0.001). (B) Continuous positive airway pressure duration vs initial pH, demonstrating an inverse correlation (r = −0.38, p < 0.001). Data points are color-coded by interface type: blue (nasal prongs), red (nasal mask), and green (alternating).Please click here to download this file.

Supplementary Table 1: Blood Gas and Inflammatory Marker Changes. This table outlines the physiological and inflammatory responses, comparing baseline values with follow-up measurements (at 6 h for PaO₂, PaCO₂, and pH, and at 48 h for CRP) across the three treatment groups.Please click here to download this file.

Supplementary File 1: The de-identified raw dataset generated and analyzed during the current study. Please click here to download this file.

Supplementary File 2: STROBE Checklist Please click here to download this file.

Discussion

This study provides compelling evidence that systematic interface rotation significantly reduces nasal injury in preterm and term neonates receiving NIV support without compromising respiratory efficacy. The 60% relative reduction in moderate-to-severe nasal injuries with the alternating protocol represents a clinically meaningful improvement that could substantially impact neonatal comfort and parental satisfaction. The findings align with and extend previous work by demonstrating that the benefits of rotation persist across a broader gestational age range than previously studied^16,17.

The physiological advantages observed with interface rotation, including more rapid PaCO₂ normalization and potentially enhanced anti-inflammatory responses, suggest mechanisms beyond simple pressure redistribution. The intermittent pressure relief inherent in regular interface changes may preserve local tissue perfusion and prevent the cascade of ischemia–reperfusion injury that contributes to pressure necrosis²¹. Additionally, the varied pressure distribution patterns between masks and prongs may provide complementary respiratory support, optimizing functional residual capacity through different mechanical mechanisms²².

The comparable NIV failure rates across all three strategies provide reassurance that the clinical benefits of rotation are achieved without compromising the effectiveness of respiratory support. The success of this technique relies on several critical steps detailed in the protocol. Specifically, the execution of the 5 min pressure-relief period and the maintenance of continuous flow during interface transitions are essential to prevent infant destabilization. Furthermore, strict adherence to visual checkpoints ensures the timely detection of early skin changes, a critical determinant of successful injury prevention²³.

The results both confirm and extend findings from recent systematic reviews and meta-analyses examining NIV interfaces in neonates. The baseline nasal injury rate of 68.3% with continuous prongs aligns closely with rates reported by Pascual et al. (20%–91.6%) and highlights the persistent challenge of interface-related morbidity⁹. However, injury rates with continuous masks (57.4%) were higher than those reported in some studies, possibly reflecting the inclusion of older, heavier infants who may generate greater interface pressures through increased movement and longer NIV duration²⁴.

The magnitude of benefit from interface rotation exceeds that reported in previous trials. While Biazus et al. reported a pooled risk reduction of 35% for any nasal injury with masks versus prongs, a 44% reduction was observed when comparing rotation with continuous prongs¹⁴. This enhanced effect likely reflects the synergistic benefits of combining the two interfaces rather than substituting one for the other. Findings from Bashir et al. demonstrated similar reductions in injury with 4-hour rotation protocols, although their study was limited to infants <32 weeks of gestation¹⁵.

The physiological outcome data provide novel insights not well characterized in previous interface comparison studies. The more rapid improvements in gas exchange observed with rotation suggest that alternating mechanical stimuli may enhance alveolar recruitment or reduce ventilation–perfusion mismatch. These findings complement previous work demonstrating differential effects of interfaces on functional residual capacity and work of breathing²⁵.

The superior outcomes associated with interface rotation likely reflect multiple interconnected mechanisms. Biomechanically, periodic redistribution of pressure prevents sustained tissue deformation that triggers inflammatory cascades and microvascular compromise²⁶. The observation of reduced CRP levels in the rotation group, although not statistically significant, suggests potential systemic anti-inflammatory effects warranting further investigation with larger sample sizes and more comprehensive biomarker panels.

The enhanced oxygenation response observed in heavier infants, regardless of interface type, provides insight into the interaction between patient characteristics and NIV effectiveness. Larger infants may have more mature alveolar–capillary interfaces and greater respiratory muscle strength, allowing more efficient utilization of positive pressure support²⁵. This finding supports the need for individualized NIV strategies based on patient phenotype rather than uniform protocol application.

The lack of documented FiO₂ throughout the treatment course represents a substantial limitation of the study. Without these data, interpretation of cumulative oxygen exposure and baseline respiratory severity is constrained, as the FiO₂ required to achieve observed PaO₂ improvements remains unknown. This limitation precludes the calculation of indices such as the oxygen saturation index and reduces the precision of NIV efficacy comparisons across interface groups. Future prospective studies should prioritize continuous FiO₂ documentation to enable more granular assessment of respiratory support requirements²⁷.

Evaluating this protocol as a standard method requires consideration of its workload and resource implications. The requirement for interface changes adds approximately 5–10 min of active nursing time every 2–4 h²⁸. Structured education programs are therefore necessary to ensure staff competency in performing these transitions. Although the protocol requires the availability of both mask and prong interfaces to increase initial equipment utilization, this must be weighed against operational gains. The observed reduction in overall CPAP duration (median difference: 5 h) represents a meaningful efficiency benefit that may offset additional workload and resource demands.

During protocol implementation, several challenges may arise. Persistent erythema may require shortening the rotation interval and verifying appropriate barrier application. Infant destabilization during transitions may be mitigated by optimizing cannula positioning and ensuring oxygen flow prior to interface removal. An excessive gas leak with masks should prompt reassessment of interface sizing rather than increasing fixation pressure, as excessive tension increases the risk of injury.

Cultural factors within neonatal units may influence the adoption of interface rotation strategies. Units with established preferences for specific interfaces may resist change despite supporting evidence. Successful implementation requires leadership engagement, clear protocols, and ongoing outcome monitoring to demonstrate local benefits²⁹.

Several limitations should be considered when interpreting these findings. The retrospective, non-randomized design introduces a risk of selection bias, as interface assignment was determined by clinician preference. This may have resulted in confounding by illness severity and temporal bias related to evolving clinical protocols over the study period. Although multivariable adjustments were applied, residual confounding cannot be excluded.

Missing data, particularly for inflammatory markers (50%), further limits the generalisability of secondary outcomes. As CRP measurements were obtained based on clinical indication, the missing data mechanism is likely “missing at random”. Observed trends toward reduced inflammation with interface rotation should therefore be interpreted as exploratory.

Variability in rotation timing (2–4 h) reflects real-world clinical practice but limits reproducibility and internal validity. Because timing was individualized, more frequent rotations may have been applied to higher-risk infants, introducing potential bias. Future studies should evaluate fixed rotation schedules to determine optimal timing for balancing skin protection and handling-related stress.

The study is also limited by its focus on short-term outcomes during the NICU stay. Long-term effects on nasal structure, respiratory health, and neurodevelopment were not assessed. Additionally, the single-center design may limit external validity. Multi-center studies with standardized protocols are required to confirm generalisability.

Future research should prioritize developing predictive models for interface selection, investigating novel interface designs, and exploring automated rotation systems to standardize practice and reduce workload. Longitudinal studies evaluating long-term outcomes will be essential to determine the sustained impact of interface strategies.

Conclusion

Interface rotation combining nasal prongs and masks represents an evidence-based approach to optimizing neonatal non-invasive ventilation delivery, reducing iatrogenic injury while maintaining respiratory support efficacy. Implementation of 2–4 h rotation protocols should be considered in clinical practice for preterm and term infants requiring non-invasive respiratory support. Despite practical challenges, the clinical benefits, particularly for extremely low-birth-weight infants, support broader adoption of this strategy.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
DuoDERM Extra Thin	Convatec	NA	Hydrocolloid barrier for skin protection
fabian Therapy Evolution	Vyaire Medical	NA	Neonatal ventilator used for NIV delivery
MR850 Heated Humidifier	Fisher & Paykel	NA	Maintains gas temperature and humidity
R version 4.2.1	R Foundation for Statistical Computing	NA	Statistical analysis software