High-flow nasal oxygen versus face-mask ventilation for rapid sequence induction in non-elective surgical patients: a randomized controlled trial

Abstract

Background

Patients undergoing non-elective surgery frequently require rapid sequence induction (RSI) because of concerns related to recent oral intake, impaired gastric emptying, or uncertain fasting status. Previous randomized controlled trials have reported mixed results regarding the effectiveness of high-flow nasal cannula (HFNC) in preventing oxygen desaturation during RSI. This study evaluated the effect of HFNC compared with facemask ventilation (FMV) during RSI in a non-elective surgical population.

Methods

In this randomized controlled study, adult ASA I–II patients undergoing non-elective surgery were allocated to HFNC (100% O₂, 60 L·min⁻¹) or FMV (FiO₂ 1.0, 10 L·min⁻¹) during RSI. A modified RSI protocol was applied, including predefined induction agents, cricoid pressure, gentle facemask ventilation as needed, and cuffed tracheal intubation. The primary outcome was peri-intubation oxygenation. Secondary outcomes included CO₂ accumulation, hemodynamic variables, gastric volume, patient comfort, and adverse events. Trial registration was performed retrospectively.

Results

Among 214 analyzed patients (107 per group), no desaturation events (SpO₂ < 94%) occurred in the HFNC group versus 12.1% in FMV (p < 0.001). HFNC yielded higher PaO₂ after preoxygenation (445.7 ± 16.8 vs. 314.2 ± 14.5 mmHg) and after intubation (405.5 ± 17.7 vs. 236.5 ± 58.5 mmHg; both p < 0.001). End-tidal oxygen concentrations were also higher with HFNC (86.0 ± 11.4% vs. 75.7 ± 12.3%; p < 0.001). CO₂ accumulation, hemodynamic variables, gastric volume, and aspiration incidence were similar between groups. Mild nasal dryness reduced comfort with HFNC, and one patient withdrew because of discomfort; overall tolerance remained acceptable.

Conclusion

HFNC appeared to improve oxygenation compared with FMV in this low-risk, non-elective surgical cohort. Within the limits of this physiologically stable population, HFNC was well tolerated and did not increase measured gastric or hemodynamic complications.

Trial registration

ClinicalTrials, NCT06879600. Retrospectively registered on 05 March 2025, https://clinicaltrials.gov/study/NCT06879600.

Introduction

Patients undergoing non-elective surgery are often categorized as being at risk of aspiration because of concerns related to recent oral intake, impaired gastric emptying, or uncertain fasting status. The rapid sequence induction (RSI) technique is frequently required to minimize this threat, as it is specifically designed to reduce the time between loss of consciousness and securement of the airway with a cuffed endotracheal tube [1, 2]. However, a major challenge in the management of these patients is that traditional RSI protocols often restrict or omit face-mask ventilation (FMV) to avoid gastric insufflation and the subsequent risk of regurgitation and aspiration [3]. As a result, patients are particularly vulnerable to hypoxaemia once apnoea ensues, with peri-intubation desaturation posing a significant risk of cardiac arrest, organ dysfunction, or even death. Pulmonary aspiration continues to be a leading cause of anaesthesia-related morbidity and mortality [1].

Given these risks, contemporary airway management guidelines from authoritative bodies such as the American Society of Anesthesiologists (ASA) [4] and the Difficult Airway Society (DAS) [5] have underscored the paramount importance of maintaining adequate oxygenation throughout airway manipulation. The 2022 ASA Difficult Airway Guidelines advocate continuous oxygen delivery during airway interventions, while the DAS guidelines explicitly prioritize maintenance of oxygenation and recommend universal preoxygenation and apnoeic oxygenation, particularly in scenarios involving unanticipated difficult tracheal intubation, which are more commonly encountered in non-elective surgical settings.

Although FMV has traditionally served as the cornerstone of preoxygenation, its use during RSI remains controversial. Classic RSI protocols omit manual ventilation after induction to minimise aspiration risk [1], and even when gentle ventilation with cricoid pressure is permitted, mask ventilation is frequently interrupted during laryngoscopy [5]. An effective mask seal can be challenging in non-elective surgical settings, and high inspiratory pressures may inadvertently insufflate the stomach despite cricoid pressure, further complicating oxygenation management.

High-flow nasal cannula (HFNC) oxygenation has emerged as an appealing alternative, capable of delivering heated, humidified 100% oxygen at high flow rates (40–70 L.min⁻¹) via a wide-bore nasal interface [2, 6]. This approach enables continuous oxygenation during both spontaneous breathing and apnoea, while also providing unobstructed access to the oral cavity for airway interventions. In addition to improving oxygenation through high inspired oxygen concentrations, HFNC may facilitate alveolar recruitment by generating low levels of positive airway pressure and reducing alveolar collapse during apnea. Notably, clinical evidence suggests that HFNC generates only modest positive airway pressure that does not significantly increase gastric insufflation, as demonstrated by McLellan et al., who found no increase in gastric volume after 30 min of HFNC at 70 L.min⁻¹ [7]. Several randomized trials and meta-analyses have demonstrated that HFNC improves physiological oxygenation parameters, including higher PaO₂ and end-tidal oxygen concentrations, and may prolong safe apnea time compared with conventional facemask preoxygenation [8, 9]. However, clinical outcomes remain inconsistent. Large randomized studies, including the multicenter trial by Sjöblom et al., reported no significant reduction in peri-intubation desaturation, suggesting that improvements in PaO₂ do not consistently translate into clinically meaningful reductions in SpO₂ desaturation or aspiration events [10, 11].

These discrepancies likely reflect heterogeneity in study populations, RSI techniques (classic versus modified), definitions of desaturation, and baseline patient risk. Importantly, relatively few studies have specifically examined HFNC in non-elective surgical cohorts undergoing a standardized modified RSI protocol, in which preoxygenation and neuromuscular blockade are routine, mask ventilation is avoided unless rescue ventilation is required for desaturation, and tracheal intubation is performed after adequate neuromuscular blockade. In such settings, physiologic reserve may be preserved, yet apnea-related hypoxemia can still occur.

Therefore, this randomized controlled trial aimed to compare HFNC with conventional FMV for preoxygenation and apneic oxygenation during modified RSI in adult patients undergoing non-elective surgery. We hypothesised that HFNC would provide superior arterial oxygenation and prolong the duration of safe apnoea without increasing the risk of gastric insufflation or aspiration, thereby enhancing the safety of airway management in this particularly vulnerable patient population.

Methods

Study design and ethical considerations

This prospective, single-centre, randomised controlled trial was designed to compare high-flow nasal cannula oxygenation with conventional face-mask ventilation for preoxygenation and apnoeic oxygenation during rapid sequence induction in patients undergoing non-elective surgery.

Ethical approval and registration

The study was approved by the Institutional Review Board of Hanoi Medical University (approval number: 1185/GCN-HMUIRB) and registered at ClinicalTrials.gov (NCT06879600). It was conducted in accordance with the Declaration of Helsinki and CONSORT guidelines, and all participants provided informed consent.

Study setting

The trial was conducted in the Department of Anesthesiology and Resuscitation at Phu Tho General Hospital, Vietnam, between July 2023 and August 2025. All procedures took place in standard operating rooms equipped with advanced monitoring systems and ultrasound devices for assessing gastric volume.

Personnel and data collection

Airway management and data collection were carried out by anaesthesiologists experienced in both HFNC and FMV techniques, ensuring consistency and reliability in the application of study protocols and outcome assessments.

Participants

Eligible individuals were adults aged 18 years or older undergoing non-elective surgery necessitating general anesthesia with rapid sequence induction and tracheal intubation. Inclusion criteria comprised American Society of Anesthesiologists (ASA) physical status I–II and Mallampati class I–III. Exclusion criteria included ASA ≥ III, anticipated difficult airway, facial deformity or challenges with mask seal, existing respiratory disease, pregnancy, known allergy to anesthetic agents, or refusal to participate.

Randomization and sample size

Participants were randomly assigned in a 1:1 ratio to either the HFNC or FMV group utilizing a computer-generated randomization list prepared by an independent statistician. Group allocation was concealed using sequentially numbered, opaque, sealed envelopes. These envelopes were stored in the operating suite and opened only after patient eligibility had been confirmed and written informed consent obtained, immediately before induction of anesthesia. The anesthesiologist responsible for intubation was informed of group assignments; however, data collectors and outcome assessors remained blinded. Sample size calculations, based on a pilot study estimating desaturation rates (SpO₂ < 94%) of 15% for the FMV group and 2% for the HFNC group, indicated that at least 98 patients per group would be required to achieve 80% power with a two-sided α of 0.05. To compensate for potential dropouts or protocol deviations, 108 patients were enrolled per group, yielding a total sample size of 216.

Interventions

All patients received standard monitoring, including ECG, noninvasive blood pressure, SpO₂, end-tidal CO₂ (EtCO₂), and transcutaneous CO₂ (TcCO₂) (SenTec Digital Monitor, Switzerland). Left radial arterial cannulation was performed under local anesthesia upon arrival in the operating room, before induction of anesthesia.

HFNC group

Preoxygenation was provided for 3 min using an HFNC system (Optiflow™, Fisher & Paykel Healthcare, New Zealand) delivering heated, humidified 100% oxygen at 60 L.min⁻¹ and 37 °C. Cannula size was determined based on the dimensions of the patient’s nares. HFNC therapy was maintained throughout induction and intubation, ensuring uninterrupted oxygen delivery during both spontaneous respiration and periods of apnea.

FMV group

Preoxygenation was performed for 3 min using a well-fitted face mask (Economy, Intersurgical, Fontenay-sous-Bois, France) connected to an anesthesia machine (Aisys CS2, GE Healthcare, Finland) with a fresh gas flow of 10 L.min⁻¹ and FiO₂ = 1.0, without inspiratory assistance or PEEP. After administration of the neuromuscular blocking agent, the face mask was removed, and routine mask ventilation was avoided during the apneic period; however, gentle rescue ventilation was permitted when necessary to prevent hypoxemia (SpO₂ ≤ 94%).

Anesthesia and intubation protocol

All patients underwent a standardized modified RSI protocol using fentanyl (2 µg.kg⁻¹), propofol (2 mg.kg⁻¹), and rocuronium (1 mg.kg⁻¹). Tracheal intubation was performed with conventional direct laryngoscopy using a Macintosh blade, with patients positioned supine in the sniffing position. Airway management was conducted by attending anesthesiologists with more than five years of experience in emergency anesthesia and prior familiarity with both HFNC and facemask ventilation techniques.

Preoxygenation was performed according to group allocation. Cricoid pressure was applied after loss of consciousness and maintained until confirmation of tracheal tube cuff inflation. Tracheal intubation was initiated 90 s after rocuronium administration. A maximum of three intubation attempts was permitted. If peripheral oxygen saturation (SpO₂) decreased to ≤ 94% at any point, the intubation attempt was aborted, and rescue facemask ventilation was provided in accordance with institutional safety protocols.

Measurements and data collection

Arterial blood gas samples were obtained at three time points: baseline (before preoxygenation), after 3 min of preoxygenation, and immediately after successful intubation.

Continuous monitoring encompassed SpO₂, EtO₂, EtCO₂, and TcCO₂ parameters. Heart rate, mean arterial pressure, and oxygenation indices were systematically recorded at designated intervals. Ultrasound assessment was employed to determine gastric antral cross-sectional area (CSA) and gastric residual volume (GRV), both prior to preoxygenation and following intubation. After preoxygenation, patient comfort was evaluated using a numeric rating scale (NRS; 0–10, with 0 indicating worst discomfort and 10 indicating optimal comfort). Adverse events—including aspiration, nasal irritation, hypotension, and arrhythmia—were carefully documented.

Outcomes

The primary outcome was oxygenation performance, evaluated through the incidence of oxygen desaturation (SpO₂ < 94%) during periods of apnea and intubation, as well as alterations in PaO₂ and EtO₂ levels. Secondary outcomes encompassed carbon dioxide metrics (PaCO₂, TcCO₂, EtCO₂), comfort scoring, gastric insufflation, aspiration occurrences, and additional adverse events.

Statistical analysis

Data were analyzed using SPSS version 22.0 (IBM Corp., Armonk, NY, USA). Continuous variables were presented as mean ± SD or median (IQR) and compared using Student’s t-test or Mann–Whitney U test. Categorical variables were expressed as counts (percentages) and analyzed using χ² or Fisher’s exact test. Repeated-measures data were analyzed using two-way ANOVA with Bonferroni post hoc correction. Linear regression analysis was used to assess correlations between apnea time and changes in PaO₂ or PaCO₂, and interaction effects between group and apnea duration were tested. A p-value < 0.05 was considered statistically significant. All figures were generated using GraphPad Prism (version 8.0; GraphPad Software, San Diego, CA, USA) and R software (version 4.5; R Foundation for Statistical Computing, Vienna, Austria).

Results

A total of 227 patients underwent screening, with 216 individuals subsequently randomized into two groups: 108 assigned to the HFNC cohort and 108 to the FMV cohort. One participant from the HFNC group discontinued participation due to intervention-related discomfort, while one individual from the FMV group was excluded because of missing data. Consequently, the final analysis comprised 214 patients, with 107 in each group (Fig. 1).

Fig. 1.

CONSORT Flow chart

Baseline demographic and clinical characteristics were comparable between the two groups, with no significant differences in age, sex, BMI, comorbidities, surgical type, fasting duration, or pre-anesthetic gastric ultrasound findings (Table 1).

Table 1.

Demographic characteristics and gastric ultrasound assessment

	HFNC group (n = 107)	FMV group (n = 107)	p-value
Age (yr)	49.4 ± 14.9 [19–77]	48.1 ± 15.3 [19–76]	0.515
Weight (kg)	60.7 ± 7.2 [45–75]	61.2 ± 6.8 [46–75]	0.633
Height (cm)	163.1 ± 5.9 [150–173]	163.7 ± 6.6 [147–178]	0.513
BMI	22.74 ± 1.66 [17–27]	22.75 ± 1.55 [19–27]	0.966
Gender (Male/Female)	63/44	64/43	0.889
ASA grade (I/II/III)	54/53/0	58/49/0	0.584
Comorbidity
Hypertension, n (%)	17 (15.9)	13 (12.1)	0.431
Diabetes, n (%)	10 (9.3)	9 (8.4)	0.810
Type of surgery, n (%)
Abdominal	106 (99.1)	104 (97.2)	0.313
Orthopaedic	1 (0.9)	3 (2.8)
Pre-op bowel obstruction/ileus. yes	26 (12.1)	20 (18.7)	0.318
Fasting time for solids (h)	8.6 ± 2.2 [4.00–14.00]	8.9 ± 3.0 [4.0–13.0]	0.410
Fasting time for fluids (h)	5.8 ± 1.7 [2.5–11.0]	6.2 ± 1.9 [3.0–10.0]	0.130
Pre-anes CSA (cm2)	8.5 ± 1.6 [4.6–11.7]	8.6 ± 1.7 [4.3–12.3]	0.589
Pre-anes GRV (ml)	88.1 ± 26.9 [14–125]	91.6 ± 30.7 [8–135]	0.374
Pre-anes GRV (ml.kg−1)	1.47 ± 0.47 [0.22–2.47]	1.51 ± 0.51 [0.15–2.56]	0.558
Gastric content on ultrasound, n (%)
Clear fluid	86 (80.6)	86 (80.6)	1
Solid	21 (9.8)	21 (9.8)
Antral grading system, n (%)
Grade 0	5 (4.7)	6 (5.6)	0.940
Grade 1	16 (15.0)	15 (14.0)
Grade 2	86 (80.4)	86 (80.4)

Data are expressed as mean ± SD [range] or number (proportion)

BMI Body mass index, ASA American Society of Anesthesiologists classification, CSA Cross-sectional area of the antrum, GRV Gastric residual volume, Pre-anes Pre-anesthesia, Student’s t-test Chi-squared test. or Fisher’s exact test

A summary of airway assessment and intubation parameters is provided in Table 2. No statistically significant differences were observed between groups regarding Mallampati classification, thyromental distance, mouth opening, neck mobility, or Cormack–Lehane grade. Additionally, intubation time, apnea duration, and the number of intubation attempts were comparable across groups, suggesting similar procedural complexity and operator proficiency.

Table 2.

Airway assessment and intubation conditions

	HFNC group (n = 107)	FMV group (n = 107)	p-value
Modified Mallampati class, n (%)
I	42 (39.3)	43 (40.2)
II	52 (48.5)	51 (47.7)	0.989
III	13 (12.1)	13 (12.1)
IV	0	0
Thyromental distance, n (%)
< 6 cm	2 (1.9)	1 (0.9)	0.844
6–7 cm	24 (22.4)	24 (22.4)
> 7 cm	81 (75.7)	82 (76.6)
Mouth opening, n (%)			0.775
< 2 cm	0	0
2–4 cm	7 (6.5)	6 (5.6)
> 4 cm	100 (93.5)	101 (94.4)
Neck movement, n (%)			0.561
Normal	105 (98.1)	106 (99.1)
Limited	2 (1.9)	1 (0.9)
Cormack–Lehane grade, n (%)			0.088
I	80 (74.8)	85 (79.4)
II	23 (21.5)	13 (12.1)
III	4 (3.7)	9 (8.4)
IV	0	0
Intubation time (s)	70 ± 49 [38–255]	68 ± 49 [38–238]	0.840
Apnoea time (s)	160 ± 49 [128–345]	158 ± 48 [128–328]	0.829
Intubation attempts	1 (1–1[1-3])	1 (1–1[1-3])	0.865

Data are presented as number (proportion). mean ± SD [range]. or median (IQR [range]); Chi-squared or Fisher’s exact test. Student’s t-test. or Mann–Whitney U test

Oxygenation outcomes

Oxygenation profiles during preoxygenation, apnea, and intubation are shown in Fig. 2.

Fig. 2.

Oxygenation profiles during preoxygenation, apnea, and tracheal intubation. a The lowest SpO₂ during apnea and tracheal intubation. b Arterial partial pressure of oxygen (PaO₂) at baseline (before preoxygenation), after 3 min of preoxygenation, and immediately after intubation. c End-tidal oxygen (EtO₂) in the first breath after intubation. Data are expressed as means with SD. Differences were estimated by two-way repeated-measures ANOVA with Tukey’s post hoc test (**p < 0.001, between groups; ^##p < 0.001; within the group). d Relationship between apnea duration and the decrease in arterial oxygen partial pressure (ΔPaO₂ = PaO₂ after preoxygenation – PaO₂ after intubation). Linear regression analysis (y = ΔPaO₂, x = apnea time [s]) was performed for both groups. Comparison of regression slopes showed that the HFNC group exhibited a significantly flatter gradient (p < 0.001), indicating more stable PaO₂ maintenance during apnea

Throughout the apneic and intubation phases, oxygen desaturation (SpO₂ < 94%) was not observed in any patient within the HFNC group. In contrast, 13 patients (12.1%) in the FMV group experienced desaturation (p < 0.001; Fig. 2a).

Baseline PaO₂ values were similar between groups. After 3 min of preoxygenation, PaO₂ was significantly higher in the HFNC group compared with the FMV group (445.7 ± 16.8 mmHg vs. 314.2 ± 14.5 mmHg; adjusted p < 0.001). A significant group × time interaction was observed [F(1, 212) = 2885, p < 0.001; Fig. 2b].

During periods of apnea and intubation, PaO₂ decreased in both groups but remained significantly higher in the HFNC group (405.5 ± 17.7 mmHg vs. 236.5 ± 58.5 mmHg; p < 0.001), with a significant group × time interaction [F(1, 212) = 41.33, p < 0.001].

Consistently, end-tidal oxygen (EtO₂) in the first breath after intubation was also greater with HFNC (86.0 ± 11.4%) than with FMV (75.7 ± 12.3%; p < 0.001; Fig. 2c).

Correlation analysis demonstrated a strong relationship between apnea duration and oxygen decline (ΔPaO₂ = PaO₂ after preoxygenation – PaO₂ after intubation). In the FMV group, PaO₂ decreased sharply with increasing apnea time (r = 0.967, p < 0.001; slope = 1.22), whereas in the HFNC group, the decline was much slower (r = 0.630, p < 0.001; slope = 0.11; Fig. 2d). Comparison of slopes confirmed a significantly flatter gradient in the HFNC group (p < 0.001), indicating more stable oxygen maintenance during apnea.

Multiple linear regression analysis demonstrated that both the type of intervention group and the duration of apnea independently predicted PaO₂ (p < 0.001). Specifically, allocation to the HFNC group was associated with a 16.34 mmHg higher PaO₂ compared to the FMV group (B = 16.340, p < 0.001), and each additional second of apnea corresponded to a 0.986 mmHg decline in PaO₂ (B = 0.986, p < 0.001). Furthermore, the negative interaction term (B = –0.350, p < 0.001) indicated that HFNC significantly reduced the rate of PaO₂ decline over time. Collectively, these results confirm that HFNC not only improved preoxygenation but also mitigated the extent of oxygen desaturation during periods of apnea.

Carbon dioxide exchange

Figure 3 summarizes the carbon dioxide parameters.

Fig. 3.

Carbon dioxide exchange during preoxygenation, apnea, and tracheal intubation. a Arterial partial pressure of carbon dioxide (PaCO₂) was checked at baseline (before preoxygenation), after 3 min of preoxygenation, and immediately after intubation. b The Transcutaneous Carbon Dioxide (TcCO₂) was monitored continuously. c End-tidal carbon dioxide (EtCO₂) in the first breath after intubation. Data are expressed as means with SD. Differences were estimated by two-way repeated-measures ANOVA with Tukey’s post hoc test (**p < 0.001, between groups; ^##p < 0.001; within groups). d Relationship between apnea duration and the increase in arterial carbon dioxide partial pressure (ΔPaCO₂ = PaCO₂ after intubation – PaCO₂ after preoxygenation). To compare the rates of CO₂ accumulation between the HFNC and FMV groups during apnea, linear regression analysis was performed for each group (y = ΔPaCO₂, x = apnea time [s]). The regression slopes were not significantly different (p for interaction = 0.78)

Baseline and post-preoxygenation PaCO₂ values were comparable between groups. During apnea, PaCO₂ increased progressively in both cohorts; however, the increase was significantly smaller in the HFNC group (43.8 ± 1.83 mmHg) than in the FMV group (47.43 ± 1.82 mmHg; adjusted p < 0.001). A significant group × time interaction was observed (F(1, 212) = 145.9, p < 0.001; Fig. 3a).

Baseline and post-preoxygenation TcCO₂ values were also similar between groups. TcCO₂ demonstrated a comparable temporal pattern but remained lower in the HFNC group at the end of intubation (42.5 ± 2.1 mmHg vs. 44.3 ± 2.1 mmHg; p < 0.001; Fig. 3b), again with a significant group × time interaction (F(1, 212) = 147.5, p < 0.001).

Consistently, EtCO₂ measured at the first breath after intubation was lower in the HFNC group (39.3 ± 2.9 mmHg) than in the FMV group (43.6 ± 3.4 mmHg; p < 0.001; Fig. 3c).

A significant correlation was found between apnea duration and CO₂ increase (ΔPaCO₂ = PaCO₂ after intubation – PaCO₂ after preoxygenation) in both groups (FMV: r = 0.588, p < 0.001; HFNC: r = 0.455, p < 0.001; Fig. 3d). However, the interaction term between apnea time and group was not significant (B = –0.001, p = 0.78), indicating similar rates of CO₂ accumulation between HFNC and FMV. Therefore, while HFNC reduced overall hypercapnia, it did not produce a significant change in the rate at which hypercapnia increased during apnea.

TcCO₂ showed a strong positive correlation with PaCO₂ (r = 0.626, p < 0.001). Linear regression analysis revealed PaCO₂ as a significant predictor of TcCO₂ (F = 129.9, p < 0.001), with the equation: TcCO₂ = 6.254 + 0.831 × PaCO₂.

Hemodynamic, metabolic, and adverse events

Hemodynamic, metabolic, and safety data are summarized in Table 3. Heart rate and mean arterial pressure remained stable and did not differ significantly between groups at any time point.

Table 3.

Hemodynamic, metabolic, and adverse events

Variables		HFNC group (n = 107)	FMV group (n = 107)	p-value
Heart rate (bpm)	T0	94.5 ± 5.4	93.6 ± 6.3	0.261
	T6	83.1 ± 6.6	82.9 ± 7.5	0.831
	T11	90.3 ± 4.6	90.8 ± 5.9	0.474
MAP (mmHg)	T0	94.5 ± 5.7	94.9 ± 5.6	0.707
	T6	93.2 ± 4.9	93.4 ± 4.7	0.864
	T11	95.6 ± 5.1	96.0 ± 5.0	0.619
pH	T0	7.426 ± 0.029	7.428 ± 0.026	0.577
	T6	7.454 ± 0.028	7.452 ± 0.023	0.579
	T11	7.377 ± 0.028	7.369 ± 0.027	0.029
BE (mEq/L)	T0	−0.406 ± 0.875	−0.617 ± 0.821	0.071
	T6	−0.674 ± 0.891	−0.885 ± 0.881	0.084
	T11	−0.733 ± 1.29	−0.872 ± 0.907	0.367
Lactates (mmol/L)	T0	1.29 ± 0.05	1.28 ± 0.05	0.066
	T6	1.29 ± 0.12	1.28 ± 0.15	0.861
	T11	1.31 ± 0.05	1.32 ± 0.05	0.444
∆CSA (cm2)		0.172 ± 0.092	0.192 ± 0.078	0.088
∆GRV (ml)		1.726 ± 0.455	1.785 ± 1.005	0.577
∆GRV (ml/kg)		0.0289 ± 0.008	0.029 ± 0.016	0.690
Aspiration (n, %)		1 (0.93%)	1 (0.93%)	1
Nasal dryness/nasal mucosal irritation		7 (6.54%)	1 (0.93%)	0.065
Comfort Score, median (IQR)*		3 (1–5)	4 (3–5)	0.084

Data are expressed as mean ± SD [range] or number (proportion)

MAP Mean arterial pressure, BE Base excess, CSA Cross-sectional area of the antrum (∆CSA = CSA after intubation – CSA baseline), GRV Gastric residual volume (∆GRV = GRV after intubation – GRV baseline)

^*Numeric Rating Scale (NRS, 0–10; 0 = worst discomfort, 10 = best comfort)

Student’s t-test. Chi-squared test or Fisher’s exact test

Metabolic parameters, including arterial pH, base excess (BE), and lactate, showed no clinically relevant differences. Both groups exhibited a small decrease in pH after intubation, consistent with transient respiratory acidosis from brief apnea, which was slightly greater in the FMV group (7.369 ± 0.027 vs. 7.377 ± 0.028; p = 0.029).

The alterations in antral cross-sectional area (ΔCSA) and gastric residual volume (ΔGRV) after intubation showed no difference between the groups (p > 0.05). No aspiration events occurred in either group.

There were no significant differences in patient comfort scores between groups. Patients in the HFNC group reported mild discomfort due to nasal dryness or mucosal irritation, whereas some patients in the FMV group experienced a sense of claustrophobia from the face mask.

Disscusion

In this randomized controlled trial, HFNC oxygenation demonstrated clear superiority over conventional FMV in maintaining oxygenation during rapid sequence induction in non-elective surgery patients. HFNC effectively prevented oxygen desaturation events, achieved higher arterial levels during both preoxygenation and the apneic period, and attenuated the decline in PaO₂ associated with prolonged apnea. Although carbon dioxide accumulation occurred in both groups, HFNC modestly reduced overall CO₂ retention without altering its rate of increase. Hemodynamic and metabolic parameters remained stable, and there were no significant differences in gastric insufflation, aspiration, or other adverse events. One patient discontinued HFNC because of discomfort; however, overall tolerance was acceptable.

Patients with non-elective surgery undergoing RSI are particularly prone to hypoxemia due to the limited opportunity for ventilation between induction and intubation. Effective preoxygenation and apneic oxygenation are therefore essential to extend safe apnea time. In this study, HFNC was associated with more effective oxygenation than facemask ventilation, as reflected by higher PaO₂ and EtO₂ levels, absence of desaturation events, and a slower decline in PaO₂ during apnea, consistent with previous reports [6]. Jo et al. similarly demonstrated higher PaO₂ before and after intubation with HFNC compared with facemask preoxygenation [12], and Tang et al.’s meta-analysis showed a pooled mean PaO₂ difference of + 63 mmHg favoring HFNC [2]. The higher EtO₂ observed with HFNC in our study is physiologically expected, as continuous high-flow oxygen delivery is maintained during laryngoscopy and apnea, facilitating ongoing alveolar oxygen exchange and limiting alveolar denitrogenation. This finding parallels the OPTIMASK trial, in which HFNC combined with a standard facemask improved EtO₂ compared with facemask oxygenation alone [13]. Prior studies have also reported improved apnea tolerance with HFNC, reflected by prolonged safe apnea times [12, 14], which likely contributes to the higher PaO₂ and EtO₂ levels observed in our cohort.

With respect to desaturation events, compared with facemask ventilation, the use of HFNC was associated with a significantly lower incidence of SpO₂ values below 94% during the apneic period in low-risk, non-elective surgical patients. This finding is consistent with several clinical studies [13, 15], although it contrasts with some reports [16, 17], likely due to differences in study populations, particularly those with higher oxygen consumption. For example, Lodenius et al. reported no cases of SpO₂ < 93% in the HFNC group compared with 12.5% in the facemask group among non-obese and non-pregnant emergency surgical patients [15], while the OPTIMASK trial demonstrated a lower incidence of SpO₂ ≤ 95% with HFNC (1% vs. 6%) in non-pregnant patients, including obesity [13]. The beneficial effects of HFNC may be explained by the generation of low-level positive airway pressure that mitigates anesthesia-induced alveolar collapse and preserves functional residual capacity, together with continuous delivery of high-FiO₂ oxygen that facilitates apneic oxygenation. In contrast, studies in pregnant populations have not shown a significant difference [16], likely reflecting the combined effects of reduced functional residual capacity and increased whole-body oxygen consumption. Moreover, differences in desaturation thresholds across studies may further influence statistical outcomes [17]. Collectively, these findings suggest that HFNC can effectively maintain oxygenation during prolonged apnea, potentially offering a safety advantage in unanticipated difficult tracheal intubation, which may occur in non-elective surgical patients.

Carbon dioxide accumulation during apnea occurred in both groups, as reflected by increases in arterial PaCO₂ and transcutaneous TcCO₂, which is expected during cessation of ventilation. Although HFNC was associated with a modest attenuation of overall CO₂ retention, the rate of PaCO₂ increase did not differ significantly between the HFNC and FMV groups. This finding aligns with the meta-analysis by Tang et al., which demonstrated no significant difference in post-intubation PaCO₂ between HFNC and facemask oxygenation strategies [2]. Previous studies suggest that high-flow oxygen may modestly limit CO₂ accumulation primarily through washout of upper airway dead space and reduction of CO₂ rebreathing from the nasal and pharyngeal cavities [12, 18]. Additionally, HFNC may facilitate oxygen binding to hemoglobin and promote CO₂ elimination [19, 20]. However, in the setting of RSI, where apnea duration is brief, these effects are likely limited and did not translate into a significant reduction in CO₂ retention in our study. Consequently, the degree of respiratory acidosis remained mild. Continuous TcCO₂ monitoring provided a practical, noninvasive surrogate for arterial CO₂ trends, allowing early identification of excessive CO₂ accumulation when necessary.

Hemodynamic stability was observed throughout induction in both study arms. Heart rate and mean arterial pressure remained within clinically acceptable ranges, with no significant intergroup differences. This finding is consistent with prior reports. Jo et al. found no difference in blood pressure or heart rate between HFNC and mask groups during induction in general surgery patients [12] or even in patients with a high risk of hemodynamic instability [21]. In the present study, the mild respiratory acidosis and variability in PaO₂ associated with apnea did not translate into clinically relevant cardiovascular changes. Importantly, this hemodynamic stability likely reflects the physiologically robust nature of the enrolled cohort, which consisted of low-risk, non-elective patients without severe comorbidities, rather than a specific protective hemodynamic effect of HFNC. Within this context, HFNC appeared hemodynamically well tolerated during RSI. with no adverse sympathetic or vagal effects observed.

We also examined potential gastric effects of high-flow oxygen, given concerns that positive airway pressure from HFNC might insufflate the stomach. In this study, baseline gastric volumes were relatively low, and no increase in gastric residual volume or signal of aspiration risk was observed with HFNC during RSI. These findings are consistent with meta-analytic data showing no significant difference in regurgitant aspiration events between HFNC and facemask oxygenation [2], as well as with clinical series reporting an absence of severe HFNC-related adverse events [10, 12, 22]. Although certain reports have highlighted concerns regarding reflux associated with HFNC, our controlled RSI setting, which includes airway protection through the use of a cuffed tube, seems to reduce this risk effectively. In practical terms, HFNC did not compromise gastric safety in our non-selective surgical patients.

Contrary to several prior studies [23] reporting better subjective tolerance with HFNC, our results showed slightly lower comfort scores in the HFNC group compared with FMV, although the difference was not statistically significant. This may be explained by nasal dryness or mild mucosal irritation from high flow rates (60 L.min⁻¹), as previously noted by McLellan et al. [7]. In contrast, FMV patients occasionally experienced claustrophobia or mask pressure discomfort, which may have offset the advantage of nasal oxygenation [24]. Despite these minor differences and one patient discontinuing HFNC because of discomfort, both methods were generally well tolerated. Importantly, the brief duration of preoxygenation in RSI (approximately 3 min) likely limited the clinical relevance of these subjective sensations. Overall, HFNC remained feasible and acceptable for non-selective surgical RSI use.

These findings indicate that HFNC is a superior option for preoxygenation and apneic oxygenation in RSI for non-selective surgical patients, as it increases oxygen reserves and prevents desaturation, thereby extending the safe apnea period for challenging intubations. This could improve first-pass success rates and reduce hypoxic complications. The lack of adverse hemodynamic or gastric effects means HFNC can be applied without additional risk in high-acuity settings. Incorporating HFNC into RSI protocols can improve patient safety over traditional facemask ventilation, with acceptable tolerance.

The strengths of this study lie in its use of objective and serial physiological measurements to characterize oxygenation and ventilation during RSI. These include repeated arterial blood gas analyses at predefined time points, continuous transcutaneous carbon dioxide monitoring, and end-tidal oxygen measurements, allowing detailed assessment of PaO₂ and PaCO₂ dynamics during apnea. In addition, pre-induction gastric ultrasound provided an objective evaluation of gastric volume, and all patients underwent a standardized induction and intubation protocol. Together, these methodological features strengthen the internal validity of the study and provide a robust physiological framework for evaluating HFNC during RSI.

However, several limitations should be acknowledged. First, this study intentionally enrolled patients with preserved baseline physiological status and adequate oxygen reserve, which likely contributed to the high absolute PaO₂ values observed. As such, the findings may not be directly applicable to non-elective surgical patients with significant comorbidities, such as obesity or chronic respiratory disease, who often exhibit reduced functional residual capacity and impaired gas exchange. Second, the study was registered retrospectively, which precludes independent confirmation that all outcomes and analyses were fully prespecified and should be recognized as a methodological limitation. Third, although the sample size was sufficient for physiological oxygenation endpoints, it remained modest, and the study relied on surrogate rather than clinical outcomes (e.g., aspiration pneumonia). Blinding was not feasible, and performance-related biases may have arisen from variability in facemask seal quality or tolerance of high-flow nasal oxygen. Operator familiarity with HFNC may also have influenced performance, consistent with reports of an initial learning curve [12]. Finally, the generalizability of our results may be limited in resource-constrained settings, where access to HFNC devices, continuous TcCO₂ monitoring, or arterial blood gas analysis may be restricted, potentially affecting feasibility and safety.

Future research should address these limitations through larger, multicenter trials to validate our findings and to evaluate clinically meaningful outcomes, including hypoxemic injury and aspiration-related complications. Comparative studies examining different HFNC settings, timing of initiation, or combinations with other airway support strategies may further refine its use during RSI. In addition, focused investigations in specific patient subgroups—such as individuals with obesity, chronic lung disease, or limited physiological reserve—are needed to clarify differential benefit. Importantly, future studies should also incorporate cost-effectiveness analyses, evaluating the cost–benefit balance of HFNC across varying clinical contexts and patient populations, including resource-limited settings, to better inform clinical decision-making and guideline development.

Conclusion

In this low-risk, non-elective surgical cohort, HFNC was associated with improved oxygenation compared with conventional facemask ventilation during rapid sequence induction, without evidence of increased gastric insufflation or hemodynamic instability. One patient discontinued HFNC because of discomfort; however, overall tolerance was acceptable. Given the relatively low baseline gastric residual volumes and the physiological stability of the study population, these findings should be interpreted within this specific context.

Abbreviations

HFNC

High-flow nasal cannula

FMV

Face-mask ventilation

RSI

Rapid sequence induction

ASA

American Society of Anesthesiologists

DAS

Difficult Airway Society

PaO₂

Partial pressure of oxygen in arterial blood

PaCO₂

Partial pressure of carbon dioxide in arterial blood

FiO₂

Fraction of inspiratory oxygen

SpO₂

Saturation of peripheral oxygen

EtO₂

End-tidal oxygen

EtCO₂

End-tidal carbon dioxide

TcCO₂

Transcutaneous CO₂

BMI

Body mass index; CSA, cross-sectional area

GRV

Gastric residual volume

NRS

Numeric rating scale

Authors’ contributions

Concept and study design: NDL, BMH Data acquisition: BMH, TMP, NTN, NDL, VTS Data analysis and interpretation: BMH, NDL, NDT Writing-original draft: BMH, NDL, NDT Writing, review, and editing: BMH, LDTS, TMP, NTN, NDL, NDT, VTS All authors have read and approved the final manuscript.

Funding

None declared.

Data availability

The data supporting this study’s findings are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Institutional Review Board of Hanoi Medical University (approval number: 1185/GCN-HMUIRB). Written informed consent was obtained from all participants. The trial was conducted in accordance with the Declaration of Helsinki and reported following the CONSORT guidelines.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.