A novel nasal cannula versus conventional face mask for preventing hypoxemia during flexible bronchoscopy in children: a randomized controlled trial

Abstract

Maintaining adequate oxygenation during pediatric bronchoscopy is critical for patient safety and procedural success. This study evaluates a novel supraglottic oxygen delivery technique designed to improve intraoperative oxygenation via a less invasive, more efficient approach. In this single-center, prospective, randomized controlled trial, children undergoing elective bronchoscopy were randomly assigned to receive oxygen through a novel nasal cannula (NNO) or conventional face mask (CMO), using computer-generated allocation sequences with concealed envelopes. The primary outcome was intraoperative hypoxemia (SpO₂ < 90% for > 10 s). Secondary outcomes included minimum SpO₂, severe hypoxemia (SpO₂ < 80%), frequency of mask ventilation due to desaturation, operator satisfaction with airway access and visual field, and time to first hypoxemic episode. Safety was assessed by comparing adverse event rates. In the intention-to-treat analysis of 199 randomized children, the incidence of hypoxemia was significantly lower in the NNO group than in the CMO group (6.0% vs. 18.2%; RR = 0.33, 95% CI 0.14–0.80; p = 0.008). Per-protocol analysis yielded consistent results (5.1% vs. 20.5%; p = 0.001). After multivariable adjustment, NNO remained a significant independent protective factor against hypoxemia (aOR = 0.272, 95% CI 0.098–0.753; p = 0.012). Secondary outcomes also favored NNO: lowest recorded SpO₂ was higher (94.9% vs. 93.5%, p = 0.003), mask interventions were fewer (2 vs. 13, p = 0.005), and time to hypoxemia was significantly prolonged (aHR = 0.203, 95% CI 0.075–0.547; p = 0.002).

Conclusion: In children with American Society of Anesthesiologists (ASA) physical status I–II undergoing flexible bronchoscopy, the novel nasal cannula was more effective than the conventional face mask in preventing hypoxemia and reducing procedural interruptions.

Trial registration: The Chinese Clinical Trial Registry (chiCTR.gov): No. ChiCTR2500113734, Retrospectively registered, Date of registration: December 2, 2025.

What is Known: • Maintaining adequate oxygenation is essential for ensuring patient safety and procedural success during pediatric flexible bronchoscopy. • Conventional face masks (CMO) are widely used for oxygen delivery; however, they may be insufficient in preventing intraoperative hypoxemia.

What is New: • A novel nasal cannula (NNO) significantly reduces the incidence of hypoxemia (6.0% vs. 18.2%) compared to CMO in children undergoing flexible bronchoscopy. • The use of NNO is associated with higher minimum SpO₂ levels, a longer time to onset of first hypoxemic event, and fewer procedural interruptions due to rescue ventilation, thereby providing a less invasive and more effective oxygenation strategy.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00431-026-06891-w.

Introduction

Bronchoscopy is a critical medical procedure in which a bronchoscope is introduced transorally or transnasally through the glottis into the trachea and bronchial tree, enabling direct visualization of airway anatomy and real-time assessment of dynamic physiological and pathological changes [1]. Diagnostically, it allows for precise identification of a wide range of pulmonary conditions, including endobronchial lesions, infections, congenital anomalies, and airway malacia [2, 3]. Therapeutically, it facilitates minimally invasive interventions such as foreign body removal, targeted drug delivery, endobronchial biopsy, and lavage, thereby reducing the need for more invasive surgical procedures [4–6]. In recent years, the application of bronchoscopy has grown increasingly prominent in pediatric respiratory medicine. Given the distinct anatomical constraints and developmental physiology of the pediatric airway, bronchoscopy offers unparalleled diagnostic clarity and therapeutic access [7, 8]. It has become an essential component in the evaluation and management of complex pediatric lung diseases, supporting early diagnosis, individualized treatment planning, and improved clinical outcomes [9–12].

Compared to adults, the pediatric airway is anatomically and physiologically distinct, featuring a narrower luminal diameter, more fragile mucosal tissues, and reduced respiratory reserve [13, 14]. Pediatric patients exhibit a higher basal metabolic rate (BMR), with an oxygen consumption of 6–8 mL/kg/min—markedly exceeding the adult reference value of 4 mL/kg/min [15, 16]. These inherent vulnerabilities render pediatric bronchoscopy a technically demanding procedure that requires specialized expertise in procedural execution, appropriate equipment selection, and meticulous anesthetic management [17]. The primary anesthetic approaches used in pediatric bronchoscopy include deep sedation, general anesthesia, and local anesthesia [18]. However, due to the limited cognitive and emotional self-regulation capacity in young patients, local anesthesia is often poorly tolerated and may lead to significant psychological distress, thereby limiting its clinical applicability. Furthermore, while general anesthesia ensures airway control and procedural immobility, it is associated with increased risks of hemodynamic instability, prolonged recovery, and adverse respiratory events—particularly in vulnerable pediatric populations—making it less favorable in many routine scenarios [19, 20]. As a result, deep sedation has emerged as the predominant anesthetic strategy in clinical practice. It provides adequate sedation and amnesia while preserving spontaneous ventilation and airway reflexes, thus offering a favorable balance between patient comfort, procedural safety, and hemodynamic stability [21].

In adult airway management, nasal-based strategies such as transnasal humidified rapid-insufflation ventilatory exchange (THRIVE) have been validated to yield substantial improvements in oxygenation. Patel et al. [22] were the first to report in adult cohorts that THRIVE significantly extends safe apnea time. Subsequent pediatric retrospective studies further corroborated the efficacy of nasal high-flow oxygenation for maintaining oxygenation during pediatric airway procedures [23].

During pediatric bronchoscopy, spontaneous ventilation is frequently inadequate to maintain sufficient oxygenation, necessitating the provision of supplemental oxygen throughout the procedure. Commonly employed methods include nasal cannula and simple face mask oxygen delivery [20, 24]. However, these conventional techniques provide limited inspired oxygen concentration and are particularly prone to inefficacy in the presence of upper airway compromise—such as tongue base retraction or nasal obstruction—leading to suboptimal oxygenation. Moreover, face mask administration can physically interfere with bronchoscope insertion and manipulation, often requiring repeated repositioning or temporary removal, which results in intermittent oxygen desaturation and hemodynamic instability. Alternative strategies such as laryngeal mask airway (LMA) ventilation or endotracheal intubation offer more reliable oxygenation and airway control. Nevertheless, their implementation typically requires deep sedation or neuromuscular blockade, which may increase the risk of respiratory depression, hemodynamic fluctuations, and other anesthesia-related complications [25, 26]. Consequently, while effective, these approaches entail a trade-off between improved oxygenation and heightened procedural risk, particularly in vulnerable pediatric populations.

To mitigate the risk of respiratory depression or respiratory distress—commonly induced by excessive sedative dosing or airway stimulation during bronchoscopy—selecting an appropriate oxygen delivery strategy is of critical importance. In response to this clinical challenge, our research team has developed a modified oxygen administration technique for pediatric bronchoscopy: a slender, extended oxygen catheter is inserted transnasally and advanced to the supraglottic region, enabling targeted and direct oxygen delivery into the upper airway. We hypothesize that, compared with conventional face masks, the nasally placed supraglottic oxygen catheter will: (1) deliver a more stable and elevated effective inspired oxygen fraction (FiO₂) with reduced air entrainment; (2) generate mild pharyngeal positive pressure via continuous airflow to extend safe apnea time—thereby maintaining upper airway patency and functional residual capacity (FRC); and (3) wash out anatomical dead space, thus more effectively preventing hypoxemia during pediatric flexible bronchoscopy. This approach not only enhances the effective inspired oxygen concentration but also reduces dependence on deep sedation or neuromuscular blocking agents. By maintaining spontaneous ventilation while ensuring stable oxygenation, the method demonstrates significant potential for improving intraoperative safety, minimizing anesthesia-related complications, and facilitating procedural efficiency.

Materials and methods

Research design

This study employed a single-center, prospective, randomized, controlled trial with a parallel-group design to systematically evaluate and compare the efficacy and safety of a novel oxygen delivery catheter specifically designed for pediatric bronchoscopy versus conventional face mask oxygenation in maintaining intraoperative oxygen saturation. The study protocol was prior approved by the Institutional Ethics Review Board of our hospital (Approval Number: cqbykyll-20250110–20, January 13, 2025) and it was planned and conducted in accordance with the ethical guidelines set forth in the Declaration of Helsinki. Written informed consent was obtained from the legal guardians of all participants after comprehensive explanation of the study objectives, procedures, and potential risks.

We wish to transparently declare that the registration of this trial was delayed due to the research team’s initial unfamiliarity with the specific platform’s requirements. Prior to study initiation, ethical approval was obtained from our hospital’s ethics committee, and the trial was prospectively registered in the National Medical Research Registration and Filing Information System in compliance with national regulations (registration number: MR-50–25–080717). Patient enrollment, which commenced on January 15, 2025, was conducted strictly according to the principles of a prospective randomized controlled trial (RCT). To ensure transparency and adherence to international reporting guidelines, retrospective registration was completed in the Chinese Clinical Trial Registry (registration number: ChiCTR2500113734). We confirm that no unplanned modifications to the study protocol have occurred since the onset of participant enrollment.

Sample size calculation

The sample size was determined using a priori power analysis. Based on an assumed hypoxemia incidence of 20–30% in the control group [27] and 10% in the experimental group, a sample size of 89 participants per group was required to detect a statistically significant difference with 80% power at a two-sided α level of 0.05. To account for an anticipated dropout rate of approximately 10%, the final sample size was increased to 99 participants per group, resulting in a total sample size of 198 participants.

Study population

Participants were consecutively enrolled from children undergoing flexible bronchoscopy in the Department of Pediatrics at our institution between January 2025 and October 2025.

Inclusion criteria are as follows: (1) age 0–15 years; (2) American Society of Anesthesiologists (ASA) physical status classification I or II; (3) scheduled for elective flexible bronchoscopy; and (4) written informed consent provided by legal guardians.

Exclusion criteria are as follows: (1) known or anticipated difficult airway; (2) severe cardiac or pulmonary insufficiency; (3) active upper respiratory or nasal bleeding; (4) craniofacial abnormalities that compromise mask sealing or nasal cannula placement; and (5) documented hypersensitivity to components of the study devices or administered medications.

Randomization and blinding

Eligible participants were randomly assigned in a 1:1 ratio to either the experimental group (novel nasal cannula for oxygen delivery, NNO) or the control group (conventional face mask oxygen delivery, CMO) using a computer-generated random sequence. Allocation concealment was ensured through sealed, opaque envelopes managed by an independent data administrator not involved in patient recruitment, intervention delivery, or outcome assessment. Due to the nature of the interventions, blinding of endoscopists and anesthesiologists was not feasible. However, personnel responsible for intraoperative monitoring data collection and post-procedure statistical analysis remained blinded to group assignment to minimize detection bias.

Intervention procedures

All participants received standardized anesthesia induction and maintenance. Prior to induction, penehyclidine hydrochloride was administered to reduce airway secretions and prevent bradycardia. Anesthesia was induced and maintained with propofol, supplemented with fentanyl for analgesia; dosages were weight-adjusted according to clinical guidelines [28, 29].

Experimental group (NNO): following anesthesia induction, the novel oxygen delivery catheter was inserted transnasally and advanced to the supraglottic region, then secured in place with adhesive tape and connected to a humidified oxygen source at a flow rate of 5 L/min. Catheter specs (internal diameter, length) and insertion depth were pre-determined based on the pediatric patient’s height, age, and weight, with a standard reference of the earlobe-to-nasal tip distance. Confirmation of its supraglottic position was achieved via two methods: (a) direct bronchoscopic visualization during insertion, confirming the catheter tip was positioned above the epiglottis; (b) post-placement, a mark was made on the catheter at the nasal vestibule and the device was secured, enabling real-time monitoring of catheter displacement. The bronchoscope was introduced via the oral route or contralateral nasal passage. Catheter position was verified during advancement to ensure patency and optimal oxygen delivery before proceeding with the procedure.

Control group (CMO): after induction, a standard open-face mask was applied over the mouth and nose and connected to the same oxygen source at 5 L/min. During bronchoscope insertion and manipulation, the mask was temporarily lifted or repositioned as per routine clinical practice and promptly restored during procedural pauses to support ventilation.

Outcome measures

Primary outcome: incidence of intraoperative hypoxemia, defined as peripheral oxygen saturation (SpO₂) < 90% persisting for more than 10 s, measured from the time of first bronchoscope passage through vocal cords until it is completely withdrawn from mouth.

Secondary outcomes included: lowest recorded intraoperative SpO₂ value; incidence of severe hypoxemia (SpO₂ < 80%); number of interventions required to restore oxygenation (e.g., mask ventilation due to desaturation), predefined triggers for mask ventilation: Despite cessation of bronchoscopy and escalation of the fraction of inspired oxygen (FiO₂), SpO₂ either continued to decline or remained persistently below 90% for ≥ 15 s; operator satisfaction regarding airway device interference and procedural field visibility, assessed using a 100-mm visual analog scale (VAS); and time from procedure initiation to first episode of hypoxemia.

Intention-to-treat analysis and missing data handling

The primary analysis of this study strictly adhered to the intention-to-treat (ITT) principle, where all 199 randomized children were included in the analysis according to their original randomization assignments, regardless of subsequent deviations from the protocol. For primary and secondary outcomes, missing data were managed using a conservative strategy designed to maximize the robustness of the superiority conclusion—all imputations were systematically biased toward the null hypothesis (i.e., assuming the novel nasal cannula [NNO] was non-superior to conventional mask oxygenation [CMO]). The specific rules are detailed below:

Binary outcomes (hypoxemia/severe hypoxemia): in the NNO group, the 1 child with missing data was conservatively classified as an event case (i.e., hypoxemia occurred) to avoid overestimating the protective effect of the experimental intervention. In the CMO group, the 11 children with missing data due to procedure cancellation or crossover were conservatively classified as non-event cases to avoid overestimating the risk associated with the control intervention.

Continuous outcome (minimum SpO₂): missing values in the NNO group were imputed using the 5th percentile of observed values from the per-protocol (PP) population of the same group. Missing values in the CMO group were imputed using the 95th percentile of observed values from the PP population of the same group. For children with cancelled procedures, if pre-procedural baseline SpO₂ was recorded, this value was used instead.

Count outcome (Number of mask ventilation interventions): Missing values in the NNO group were imputed using the 95th percentile of observed values within the group. In the CMO group, children with cancelled procedures were assigned a value of 0 (no intervention); children who crossed over intraoperatively were assigned a value of 1 (one intervention).

This set of “worst-case scenario for superiority” imputation methods was employed for the primary ITT analysis. As sensitivity analyses, we additionally report: (a) Results of the per-protocol (PP) analysis; (b) ITT analysis results using alternative imputation methods (e.g., last observation carried forward).

Statistical analysis

This study employed the intention-to-treat (ITT) and per-protocol (PP) principles for analysis. The primary analysis was the ITT analysis, where all participants were analyzed according to their original randomization groups. The sensitivity analysis was the PP analysis, which only included participants who fully complied with the study protocol and had complete primary outcome data, to evaluate the treatment effect under ideal compliance conditions. All analyses were conducted using IBM SPSS Statistics for Mac, Version 31.0.1.0 (IBM Corp., Armonk, N.Y., USA). A two-sided p-value < 0.05 was considered statistically significant. The normality of all continuous variables was assessed using the Shapiro–Wilk test. Continuous variables are presented as mean ± standard deviation or median (interquartile range), where appropriate, the distribution of these variables was visualized using box-and-whisker plots and compared between groups using independent samples t-tests or Mann–Whitney U tests. Time-to-event analysis for the first onset of hypoxemia was performed using the Kaplan–Meier method, and the curves were compared with the log-rank test. The treatment effect was quantified as a hazard ratio (HR) with a 95% confidence interval (CI) using univariate Cox proportional hazards regression. Subsequently, a multivariate Cox proportional hazards model was fitted to estimate the adjusted hazard ratio (aHR) and its corresponding 95% CI. The model was adjusted for age and weight as covariates, which were selected a priori based on clinical relevance. For cases without hypoxemia, the mean procedural duration of the study cohort (21.1 min) was defined as the censoring time. Categorical variables are expressed as frequency (percentage) and analyzed using chi-square tests or Fisher’s exact test, where applicable. The risk ratio (RR) and corresponding 95% CI for the primary outcome (hypoxemia incidence) were calculated to quantify the treatment effect. Subsequently, multivariate logistic regression analysis was performed to estimate the adjusted odds ratio (aOR) and its corresponding 95% CI. Covariates included for adjustment were age and weight.

Results

Baseline characteristics of the study population

Between January 2025 and October 2025, a total of 210 pediatric patients were assessed for eligibility. Of these, 11 were excluded for the following reasons: failure to meet inclusion criteria (n = 6) or guardian refusal to participate (n = 5). The remaining 199 children were randomly assigned to either the experimental group (NNO, n = 100) or the control group (CMO, n = 99). All randomized participants were included in the intention-to-treat (ITT) analysis: 100 in the novel nasal cannula (NNO) group and 99 in the conventional mask oxygenation (CMO) group. For the per-protocol (PP) sensitivity analysis, 99 participants in the NNO group and 88 in the CMO group were analyzed. One participant in the NNO group was excluded due to incomplete data recording. In the CMO group, 11 participants were excluded: 8 due to pre-procedural cancellation (i.e., before bronchoscope insertion) and 3 due to intraoperative crossover to the NNO device following recurrent hypoxemia. The participant flow through each phase of the trial is illustrated in Figs. 1 and 2.

Fig. 1.

Flowchart of the patient enrollment process

Fig. 2.

CONSORT flow diagram of participant allocation and analysis

No statistically significant differences were observed between the two groups with respect to baseline characteristics, including age, sex, body weight, American Society of Anesthesiologists (ASA) physical status classification, presence of comorbidities, and preoperative arterial blood gas parameters (PO₂, PCO₂) and SpO₂(all p > 0.05; Table 1). These findings confirm that the groups were well-balanced at baseline and thus considered comparable for comparative analysis.

Table 1.

Baseline characteristics and comparison between study groups (the intention-to-treat (ITT) population)

	NNO	CMO		Test statistic value	p	Statistical methods
Subjects, n	100	99
Age, years	6.05 ± 3.67	5.79 ± 3.62	U = 4735.5		0.597	Mann–Whitney U
Sex, n (%)			x2 = 0.048		0.827	Chi-square
Male	53 (53.0%)	54 (54.5%)
Female	47 (47.0%)	45 (45.5%)
Weight, kg	23.72 ± 14.89	21.89 ± 11.63	U = 4803.0		0.717	Mann–Whitney U
ASA, n (%)			x2 = 1.825		0.177	Chi-square
p1	38 (38.0%)	47 (47.5%)
p2	62 (62.0%)	52 (52.5%)
Non-underlying diseases, n	100	99
Preoperative PO2, mmHg	101.55 ± 24.576	101.63 ± 22.756	U = 4805.0		0.721	Mann–Whitney U
Preoperative PCO2, mmHg	35.086 ± 3.6851	34.190 ± 3.9968	t =  − 1.644		0.102	Independent samples t-tests
Preoperative SpO2, %	98.602 ± 1.1286	98.799 ± 0.9404	U = 4625.5		0.424	Mann–Whitney U
Operative duration, min	21.01 ± 5.20	21.15 ± 6.62	U = 4225.0		0.614	Mann–Whitney U

This table presents the baseline characteristics of all randomized participants, irrespective of their subsequent adherence to the treatment protocol

Primary outcome

The primary outcome was the incidence of intraoperative hypoxemia. In the intention-to-treat (ITT) analysis (Table 2), the incidence of hypoxemia was significantly lower in the NNO group than in the CMO group (6.0% vs 18.2%; risk ratio [RR] 0.33, 95% CI 0.14–0.80; p = 0.008). This corresponds to a 67% reduction in absolute risk—equivalently, a 67% relative risk reduction. The per-protocol (PP) analysis yielded consistent and numerically stronger estimates (5.1% vs 20.5%; RR 0.25, 95% CI 0.10–0.64; p = 0.001), representing a 75% relative risk reduction.

Table 2.

Primary outcome

	Analysis	NNO	CMO	Chi-square	p	RR (95%CI)	RRR
Incidence of Hypoxemia, n(%)	ITT	6 (6.0%)	18 (18.2%)	6.961	0.008	0.33 (0.14–0.80)	67%
	PP	5 (5.1%)	18 (20.5%)	10.248	0.001	0.25 (0.10–0.64)	75%

In multivariable logistic regression adjusting for age and weight (Table 3), novel nasal oxygenation remained independently associated with reduced hypoxemia risk in both the ITT (adjusted odds ratio [aOR] 0.272, 95% CI 0.098–0.753; p = 0.012) and PP analysis (aOR 0.179, 95% CI 0.060–0.538; p = 0.002). Neither age nor weight was significantly associated with hypoxemia risk in either model (all p > 0.05).

Table 3.

Adjusted primary outcome

Analysis	Variable	aOR	95%CI	p
ITT	Oxygen delivery method	—	—	—
	NNO vs. CMO	0.272	0.098–0.753	0.012
	Age (per 1-year increase)	0.685	0.442–1.063	0.092
	Weight (per 1-kg increase)	1.006	0.867–1.868	0.937
PP	Oxygen delivery method	—	—	—
	NNO vs. CMO	0.179	0.060–0.538	0.002
	Age (per 1-year increase)	0.663	0.416–1.055	0.083
	Weight (per 1-kg increase)	1.008	0.860–1.181	0.923

Secondary outcome measures

The lowest intraoperative oxygen saturation (SpO₂) was significantly higher in the NNO group than in the CMO group, with consistent findings across both ITT and PP analysis. In the ITT population, the median lowest SpO₂ was 94.85% (IQR: 92.8–97.8) in NNO versus 93.5% (IQR: 91.4–95.7) in CMO (Mann–Whitney U = 3747.000, p = 0.003). The PP analysis yielded a numerically larger difference: NNO median 94.9% (IQR: 92.9–97.9) vs. CMO median 93.2% (IQR: 91.2–94.9) (Mann–Whitney U = 2712.500, p < 0.001) (Table 4 and Fig. 3).

Table 4.

Secondary outcomes

	Analysis	NNO	CMO	Test Statistic Value	p	Statistical Methods
Lowest SpO2 (%) Median; IQR	ITT	94.85; 92.8–97.8	93.5; 91.4–95.7	U = 3747.000	0.003	Mann–Whitney U
	PP	94.9; 92.9–97.9	93.2; 91.2–94.9	U = 2712.500	< 0.001	Mann–Whitney U
Incidence of severe hypoxemia, n (%)	ITT	2 (2%)	5 (%)	/	0.431	Fisher Exact Test
	PP	1 (1.0%)	5 (5.7%)	/	0.108	Fisher Exact Test
Frequency of mask ventilation, n	ITT			U = 4448.000	0.005	Mann–Whitney U
0		98	87
1		2	11
2		0	1
	PP			U = 3854.000	0.005	Mann–Whitney U
0		98	79
1		1	8
2		0	1
Surgical field exposure VAS	PP	9.32 ± 0.26	8.74 ± 0.51	U = 1192.500	< 0.001	Mann–Whitney U
Interference from the oxygen delivery VAS	PP	9.73 ± 0.18	8.46 ± 0.38	U = 0.000	< 0.001	Mann–Whitney U
Adverse events, n (%)	ITT	2 (2%)	1 (1%)	/	1.000	Fisher Exact Test

Lowest SpO₂: the lowest intraoperative oxygen saturation; frequency of mask ventilation: the number of bronchoscopy interruptions due to hypoxemia requiring mask ventilation; surgical field exposure VAS, interference from the oxygen delivery VAS: surgeons’ satisfaction scores regarding surgical field exposure and interference from the oxygen delivery device

Fig. 3.

Comparison of the lowest intraoperative oxygen saturation

For severe hypoxemia (predefined as SpO₂ < 85% for ≥ 10 s), no significant between-group differences were observed in either ITT or PP analysis (Table 4). In the ITT population, 2 (2.0%) NNO patients and 5 (5.1%) CMO patients experienced severe hypoxemia (Fisher’s Exact Test, p = 0.431). The PP analysis showed a similar trend: 1 (1.0%) NNO patient vs. 5 (5.7%) CMO patients (Fisher’s Exact Test, p = 0.108). This lack of significance may be attributed to the small sample size and restriction to American Society of Anesthesiologists (ASA) physical status I or II patients.

Kaplan–Meier survival analysis (Fig. 4) demonstrated a significantly delayed onset of hypoxemia in the NNO group compared to CMO. The median time to first hypoxemia was 4.3 min (95% CI: 3.66–4.94) in NNO versus 2.0 min (95% CI: 1.58–2.42) in CMO (log-rank test: χ² = 10.996, p < 0.001). Cox proportional hazards regression confirmed a strong protective effect of NNO: the unadjusted hazard ratio (HR) was 0.218 (95% CI: 0.081–0.587, p = 0.003), and after adjusting for age and weight, the adjusted HR (aHR) remained significant at 0.203 (95% CI: 0.075–0.547, p = 0.002), corresponding to a 79.7% relative risk reduction. Age was marginally associated with lower hypoxemia risk (aHR 0.687, 95% CI: 0.448–1.053, p = 0.085), while weight showed no significant association (aHR 1.009, 95% CI: 0.871–1.169, p = 0.904) (Table 5).

Fig. 4.

Kaplan–Meier curves for time to hypoxia. The difference between groups was assessed by the log-rank test (Note: the survival curve was constructed using data from 199 patients. Among these, 175 patients who did not develop hypoxemia were censored at 21.1 min (mean surgical duration))

Table 5.

Adjusted hazard ratios of factors associated with time to first hypoxemia

Variable	aHR	95%CI	p
Oxygen delivery method	—	—	—
NNO vs. CMO	0.203	0.075–0.547	0.002
Age (per 1-year increase)	0.687	0.448–1.053	0.085
Weight (per 1-kg increase)	1.009	0.871–1.169	0.904

The frequency of mask ventilation interventions (a proxy for procedural interruptions) was significantly lower in NNO than in CMO, with consistent results in both ITT and PP analysis (Table 4). In the ITT population, 98 NNO patients required 0 interventions, 2 required 1, and 0 required ≥ 2; in contrast, 87 CMO patients required 0, 11 required 1, and 1 required 2 (Mann–Whitney U = 4448.000, p = 0.005). The PP analysis showed a similar pattern: 98 NNO patients with 0 interventions, 1 with 1, and 0 with ≥ 2; vs. 79 CMO patients with 0, 8 with 1, and 1 with 2 (Mann–Whitney U = 3854.000, p = 0.005). Clinically, all 13 episodes requiring mask ventilation necessitated complete bronchoscope withdrawal and bag-valve mask ventilation, whereas only 1 NNO patient required such intervention—disrupting procedural continuity and surgical field visualization in CMO. These procedural interruptions directly led to disruption of operational continuity and loss of surgical field visualization.

Surgeons’ satisfaction scores regarding surgical field exposure and interference from the oxygen delivery device were significantly higher in NNO than in CMO (9.32 ± 0.26 vs. 8.74 ± 0.51, p < 0.001; and 9.73 ± 0.18 vs. 8.46 ± 0.38, p < 0.001, respectively) (Table 4). Both groups exhibited high scores with small standard deviations, which reflect that the quality of surgical field visualization during pediatric flexible bronchoscopy (FB) is largely dependent on procedural continuity. Interruption of FB for mask ventilation due to hypoxemia—necessitating bronchoscope withdrawal—results in complete loss of the visual field, thereby reducing satisfaction scores. Conversely, maintenance of procedural continuity enables stable, high-quality visualization. These scores thus provide intuitive evidence that the novel catheter group experienced fewer hypoxemia-related procedural interruptions.

There was no significant difference in the overall incidence of adverse events between the two groups (p = 1.000, Fisher’s Exact Test) (Table 4). Two patients (2%) in NNO experienced mild, self-limiting nasal mucosal bleeding, and one patient (1%) in CMO developed mild laryngospasm; both were promptly resolved with intervention. However, the small number of events limits definitive conclusions about safety equivalence. No severe adverse events—including bronchospasm, bradycardia, vomiting, pulmonary aspiration, or requirement for advanced airway support—occurred in either group.

Discussion

Hypoxemia is a common and potentially serious complication during pediatric flexible bronchoscopy [30, 31]. The results of this prospective, randomized controlled trial demonstrate that the novel nasal cannula oxygen delivery method is significantly superior to conventional face mask oxygenation in maintaining adequate intraoperative oxygenation.

The most striking finding of this study is the significantly lower incidence of hypoxemia with novel nasal oxygenation in both intention-to-treat and per-protocol analyses. In the primary intention-to-treat analysis, hypoxemia occurred in 6.0% of NNO patients versus 18.2% of CMO patients, representing a 67% relative risk reduction. The per-protocol sensitivity analysis showed even stronger results (5.1% vs. 20.5%). This superiority can be attributed to several key physiological and procedural advantages. Firstly, this device is fixed in place through the nasal cavity and can continuously supply oxygen without interruption. Unlike conventional masks, which often need to be repositioned or may not seal properly when the bronchoscope passes through the oropharynx, creating a critical “oxygen saturation drop window,” the new nasal catheter allows for continuous oxygen supply during the insertion of the bronchoscope through the mouth or nose, providing a more stable oxygen reserve throughout the FB process and eliminating the risk of early oxygenation decline. Second, the system delivers oxygen, which effectively flushes the anatomical dead space in the nasopharynx, and generates a mild positive airway pressure [32, 33]. This proposed mechanism aligns with the findings reported by Patel et al. [22]. This physiological positive pressure helps maintain airway and alveolar patency, improves ventilation/perfusion matching, and counteracts sedation-induced atelectasis and mechanical airway obstruction caused by the bronchoscope—mechanisms not achievable with conventional, poorly sealed, and intermittently used face masks [34, 35]. However, the specific values of airway positive pressure that can be generated by different oxygen flow rates remain to be further studied.

The benefits of the novel oxygen delivery system extend beyond the primary outcome of hypoxemia prevention. Secondary outcomes further underscore its clinical utility. The NNO significantly reduced procedural interruptions due to hypoxemia compared to CMO. In CMO, nearly one-tenth of cases required complete procedural cessation and bag-valve mask ventilation—a necessary intervention that prolonged procedure duration, increased anesthetic management complexity, and elevated the risk of mucosal injury secondary to repeated bronchoscope insertion and withdrawal. By enabling more stable and continuous oxygenation, the novel nasal cannula effectively maintained procedural continuity and operational efficiency during FB. The resulting reduction in interruptions facilitated more consistent surgical field visualization, a finding supported by significantly higher operator satisfaction scores for both field exposure and device-related interference; this directly accounts for the observed difference in VAS scores. It should be noted that the VAS score is inherently subjective. Future studies may integrate the VAS score with objective metrics to enhance the robustness and interpretability of outcome assessments. Importantly, the clinical benefits of the novel nasal cannula extend beyond physiological stabilization (e.g., SpO₂ maintenance); they encompass meaningful improvements in procedural workflow, diagnostic accuracy, procedural precision, patient safety, and resource utilization [36, 37]. In this study, the CMO group reduced mask ventilation leading to surgical interruption by 13 times. Assuming that each interruption prolongs the operation time by an average of 2 min, in a center with an annual operation volume of 200 cases, if all use this catheter, theoretically, about 6.7 h of operating room time can be saved and corresponding economic benefits can be achieved. A detailed cost–benefit analysis needs to be conducted in a larger-scale study.

Moreover, Patients in NNO also exhibited higher minimum intraoperative oxygen saturation levels, and a significantly delayed time to the first hypoxemic event (log-rank p < 0.001), indicating a larger “physiological safety margin”. The hazard ratio of 0.218 further underscores the strength of this intervention, indicating a 78.2% reduction in the instantaneous risk of hypoxemia. Subsequently, we adjusted for key confounding factors (age, body weight). The adjusted model was significantly superior to the unadjusted model (p = 0.002). Additionally, the adjusted hazard ratio (aHR = 0.203) was highly consistent with the adjusted odds ratio (aOR = 0.272) derived from logistic regression, which strengthens the robustness of the study findings. Furthermore, the primary finding of a significant difference in time-to-event, as assessed by the log-rank test, remains robust and clinically meaningful. This buffer is particularly valuable when managing acute complications such as bleeding or bronchospasm, providing critical time for intervention [38–40]. Given that severe hypoxemia is a leading contributor to intraoperative cardiopulmonary arrest, maintaining higher baseline oxygenation substantially reduces the risk of life-threatening events—especially crucial in pediatric patients, who have limited physiological reserves and are prone to rapid clinical deterioration [41–43].

While the between-group difference in the incidence of severe hypoxemia (SpO₂ < 80%) did not reach the conventional threshold for statistical significance (p = 0.431), a trend toward fewer events was observed in NNO (2 vs. 5 cases). Given the overall low incidence of such severe events, this study may have been underpowered to detect between-group differences. However, when combined with the significant difference in the more clinically relevant endpoint of rescue mask ventilation for hypoxemia (2 vs. 13 episodes, p = 0.005), the novel nasal cannula can be reasonably concluded to offer superior protection against clinically actionable oxygenation disturbances.

Our findings align with an increasing body of evidence from adult endoscopy supporting the use of specialized transnasal oxygen devices [44–47]. However, this study is the first RCT in routine paediatric bronchoscopy (extends THRIVE). Children’s heightened oxygen consumption and reduced functional residual capacity render them particularly vulnerable to the limitations of conventional oxygen delivery methods, suggesting that the observed benefits may be even more pronounced in this age group [48–50].

This study has several limitations. First, the retrospective registration of the trial, which may compromise the prospective transparency of the study protocol. Second, blinding of operators was not feasible due to the nature of the intervention, potentially introducing performance bias. However, the objectivity of the primary endpoint—oxygen saturation—helps mitigate this concern. All outcome data were collected and recorded by independent research assistants who were not involved in the surgical procedure. Third, the study was conducted at a single center with a relatively small sample size and primarily included patients classified as ASA physical status I or II. This study did not directly measure the actual oxygen concentration inhaled by patients under the two oxygen administration methods. Although both groups used the same oxygen flow rate (5 L/min), due to differences in equipment design principles and patients’ breathing patterns, the FiO₂ actually reaching the alveoli might have been different. Therefore, generalizability to children with more severe comorbidities—such as significant pulmonary parenchymal disease or cardiac conditions—requires further investigation. Future studies should evaluate efficacy in these high-risk subgroups and assess the cost-effectiveness of routine implementation. We also plan to integrate capnography into the cannula system to enable real-time, non-invasive monitoring of ventilation, thereby further enhancing patient safety during bronchoscopic procedures.

Conclusions

In conclusion, this single-center randomized controlled trial enrolling children with ASA physical status I or II provides preliminary evidence that the novel nasal oxygen delivery catheter demonstrated superior efficacy in preventing hypoxemia compared to conventional face mask oxygenation during pediatric flexible bronchoscopy. It may potentially improve procedural safety by reducing hypoxemic events, although no significant difference was observed in the incidence of serious adverse events between the two groups. This device represents a promising alternative oxygenation strategy. Further multi-center studies are warranted to validate its efficacy and safety in diverse clinical settings and among high-risk (ASA physical status III or IV) pediatric populations.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

NNO

Novel nasal cannula oxygen delivery

CMO

Conventional face mask oxygen delivery

BMR

Basal metabolic rate

THRIVE

Transnasal humidified rapid-insufflation ventilatory exchange

LMA

Laryngeal mask airway

SpO₂

Peripheral oxygen saturation

FiO₂

Inspired oxygen fraction

PEEP

Positive end-expiratory pressure

FRC

Functional residual capacity

RCT

Randomized controlled trial

ASA

American Society of Anesthesiologists

VAS

Visual analog scale

HR

Hazard ratio

CI

Confidence interval

aHR

Adjusted hazard ratio

RRR

Relative risk reduction

aOR

Adjusted odds ratio

FB

Flexible bronchoscopy

Authors’ contributions

Conceptualization: Jun Chen; Methodology: Jun Chen; Formal analysis: Jun Chen, Zhoubo Han; Investigation: Jun Chen, Kui Zheng; Writing – original draft: Jun Chen; Writing – review & editing: Zhoubo Han; Supervision: Wenbing Zhang.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

All the analysis results are given within the manuscript. The study was performed prospectively and according to the regulations of the institution review board, such data could not be openly shared. Request for the dataset supporting our results can be made via corresponding author and will be given by the corresponding author after approval.

Declarations

Ethics approval and consent to participate

This study protocol was approved by the Medical Ethics Committee of Bishan Hospital of Chongqing Medical University (cqbykyll-20250 110–20, January 13, 2025), and it was planned and conducted in accordance with the ethical guidelines set forth in the Declaration of Helsinki. Retrospective registration was completed in The Chinese Clinical Trial Registry (No. ChiCTR2500113734), Date of registration: December 2, 2025. All participants signed informed consent after receiving details about the study from the research team.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.