Unlike gastrointestinal endoscopy, flexible bronchoscopy (FB) involves direct tracheobronchial manipulation, which induces intense airway irritation. Patients may experience severe discomforts, including violent coughing, asphyxia, and a sense of impending death, as well as critical adverse events such as desaturation, circulatory overload, and malignant arrhythmias (1, 2). So sedation is recommended currently by most guidelines or expert consensus to be offered to all patients undergoing FB without contraindications, aiming to mitigate stress responses, optimize procedural conditions, and improve patient comfort (3–5).

Although mild-to-moderate sedation is generally recommended during FB to preserve spontaneous breathing, clinical evidence indicates desaturation incidence still approaches 56% even at this depth (6). Desaturation constitutes an inherent risk of FB due to the airway-sharing requirement of the bronchoscope and sedation-induced respiratory depression, with potential life-threatening severity. Consequently, the currently advocated mild-to-moderate sedation protocol represents a compromise between anesthetic efficacy and patient safety. This strategic limitation of sedation prioritizes preserving adequate spontaneous respiration and the patient’s responsiveness, thus enabling rapid corrective intervention during oxygen desaturation events via temporary withdrawal of the flexible bronchoscope and/or verbal commands to arouse the patient (7).

However, the mild-to-moderate sedation regimen frequently fails to effectively suppress the intense airway reflexes and severe coughing provoked by FB, exposing patients to recurrent procedural stimuli. This approach not only increases technical difficulty for the FB operator but also adversely impacts tolerance of patients, ultimately diminishing compliance with subsequent endoscopic examinations. While increasing sedation depth may theoretically enhance patient comfort and procedural conditions, this approach carries significant risks of exacerbated respiratory depression and elevated desaturation incidence (8, 9). Therefore, performing FB under deep sedation with conventional oxygen supplementation has been consistently considered unsafe and is not recommended in prevailing clinical guidelines and expert consensus documents.

The emergence of novel oxygen delivery systems, particularly high-flow nasal cannula (HFNC) oxygenation, presents a potential paradigm shift in this clinical situation. HFNC delivers heated, humidified, and oxygen-enriched air via nasal cannula at flow rates up to 70 L/min, offering multiple physiological benefits including maintenance of constant FiO₂, apneic oxygenation, pharyngeal dead space reduction, decreased work of breathing, positive end-expiratory pressure (PEEP) generation, and enhanced mucociliary clearance and gas exchange (10, 11). These physiological advantages, combined with the system’s simplicity and patient comfort, establish HFNC as an optimal solution not only for respiratory failure patients but also for perioperative oxygenation and ventilation support (11–13).

Accumulating clinical evidence demonstrates the effectiveness of HFNC in FB, with multiple randomized controlled trials reporting significant reductions in desaturation incidence and enhanced nadir SpO₂ values during procedures (14). Nevertheless, all these investigations were confined to mild-to-moderate sedation levels, with no existing literature addressing applications under deep sedation. Considering that deep sedation typically induces severe respiratory depression, current FB guidelines advocate for introducing airway devices (such as laryngeal masks or endotracheal intubation) with assisted ventilation. While HFNC, which has no effective ventilation function, is not yet mentioned. For this reason, in this randomized controlled trial, we aim to evaluate whether HFNC is beneficial for FB under deep sedation, with a focus on providing adequate oxygenation with acceptable carbon dioxide accumulation.

While HFNC has been extensively studied in FB procedures under mild-to-moderate sedation, no randomized controlled trials to date have systematically evaluated its potential benefits in the context of deep sedation. This might be the first study to indicate that HFNC generates significant improvements in both oxygenation and ventilation parameters during deep sedation FB, mainly manifested as an attenuation of desaturation risk with concurrent alleviation of carbon dioxide (CO₂) retention. Critically, HFNC enables safe implementation of non-intubated deep sedation bronchoscopy, optimizing procedural conditions and providing effective respiratory support even during anticipated severe sedation-induced respiratory depression.

Undoubtedly, deep sedation more effectively suppresses intense airway irritation and severe cough responses during FB, providing optimal operating conditions and enhanced patient comfort. However, severe respiratory depression and unacceptably high desaturation risk associated with deep sedation significantly outweigh these benefits (15, 16). Our results also confirm that during bronchoscopy under deep sedation, the cough VAS score was very low (median: 1–1.5), and the cough response was significantly lower than that reported in our previous study with a mild-to-moderate sedation regime for bronchoscopy, as well as less intraoperative hypertension and tachycardia (17). However, the desaturation incidence reached 57.14% in controls receiving conventional nasal catheter oxygen supply (5 L/min). This aligns with Wei et al.’s reported 84% desaturation incidence during deeply sedated (MOAA/S 1–2) diagnostic FB, where higher rates may relate to the inclusion of ASA grade 3 cases and propofol-based regimens (9). Due to the concerns of oxygenation, most guidelines recommend mild-to-moderate sedation for FB operation, as deep sedation with conventional oxygen supply measures is generally unsafe without endotracheal intubation or LMA assistance—but airway devices that increase complexity and cost (10, 18).

HFNC, as an advanced oxygen delivery system, enables non-intubated bronchoscopy under deep sedation while maintaining spontaneous respiration, which has been confirmed in our previous study with elderly patients undergoing FB examination (19). In this study, we find that HFNC significantly decreases desaturation risk during FB under deep sedation, with incidence rates showing a declining trend as flow rates increased from 25 to 65 L/min. This is very similar to previously reported outcomes of HFNC-supported FB under mild-to-moderate sedation (20–22). Specifically, Zhang et al. observed flow-dependent desaturation reduction during propofol-sulfentanil moderate sedation (flow rate from 10 to 60 L/min) (22). In patients receiving FB under deep sedation, however, our investigation revealed attenuated HFNC flow dependence for FB desaturation prevention: minimal intraoperative SpO₂ showed no significant differences among three subgroups, while desaturation incidence differed solely between HF25 and HF65 cohorts. This may be due to the fact that most patients unconsciously open their mouths during our FB procedure, especially after receiving sedation. Existing evidence confirms HFNC reduces FB-related desaturation through FiO₂ elevation and positive airway pressure generation (20, 23), subsequent studies reveal that though mouth opening has little effect on FiO₂ (24), HFNC-generated airway pressure can be significantly diminished with open-mouth breathing (mean: 2.3 cm H₂O at 45 L/min when closed vs. near-zero when open) (25). Since HFNC’s clinical benefits -including improved oxygenation, ventilation-perfusion matching, and reduced respiratory work -largely depend on airway pressure maintenance (26), open-mouth states during FB may attenuate both HFNC’s anti-desaturation efficacy and flow-dependent therapeutic mechanisms.

The efficacy of HFNC in significantly reducing the desaturation risk of FB under deep sedation, despite lacking active ventilation capacity, is quite beyond our initial expectations. This effect may mainly derive from two functions of HFNC, namely oxygen reservoir expansion via pre-oxygenation and apnoeic oxygenation (24–27). Apnea is prevalent during FB sedation, a 100% apnea occurrence was recorded by Minami et al. under fentanyl-midazolam sedation during bronchoscopy (26). HFNC’s oxygen reservoir expansion and apnoeic gas exchange are critical for bridging sedation-induced temporary respiratory depression, thereby preventing both desaturation events and ventilation intervention requirements. Clinical evidence validates its capacity to enhance oxygenation and extend safe apnea duration during anesthesia induction (11). Admittedly, other benefits of HFNC, such as nasopharyngeal dead space reduction, positive end-expiratory pressure (PEEP) effect generation, end-expiratory lung volume augmentation, and alveolar ventilation preservation, also contribute to improving oxygenation (11, 16, 20). Just as Patel A et al. elucidated, HFNC synergizes apnoeic oxygenation with continuous positive airway pressure and gaseous exchange through flow-dependent dead space flushing, achieving remarkable oxygenation improvement and desaturation risk reduction (28). However, even at maximal flow rates (65 L/min), HFNC still demonstrates a desaturation incidence of approximately 5% during deep sedation FB. This confirms that HFNC cannot entirely eliminate desaturation risks or replace assisted ventilation. Intensive intraoperative monitoring and preparation for assisted/controlled ventilation are still required.

CO₂ accumulation during deep sedation constitutes a significant clinical concern beyond desaturation risk (26, 28). Sedative agents can induce varying degrees of respiratory depression and hypoventilation, with established evidence demonstrating frequent PaCO₂ elevation during bronchoscopy, particularly under deep sedation rather than mild or moderate sedation (3, 4, 17). And the application of HFNC, which lacks effective ventilation capacity, raises ongoing debate regarding its role in modulating (even exacerbating) CO₂ retention during FB (29, 30). However, to date, the impact of HFNC on CO₂ accumulation during deep sedation of FB has not reported in the literature. The results of this study indicated significant end-procedure EtCO₂ elevation in all groups, confirming that respiratory depression and hypoventilation also exist even after awakening from deep sedation, and the application of HFNC could alleviate the accumulation of CO₂ to a certain extent, which was independent of the flow rate (25–65 L/min). These observations align partially with prior mild-to-moderate sedation studies suggesting HFNC’s dual benefit in oxygenation improvement and CO₂ reduction (31–33). Irmak et al. indicates HFNC minimizes CO₂ retention primarily through enhanced CO₂ clearance. This functions by continuously displacing exhaled upper airway air with low- CO₂ fresh gas, reducing rebreathing (32). A controlled trial confirmed HFNC’s superiority over standard nasal prongs in facilitating CO₂ removal and lowering Pt CO₂ (33). Additionally, HFNC alleviates respiratory depression and CO2 retention by reducing respiratory work and lung tissue metabolic demand while improving oxygenation and respiratory muscle function (31, 34). However, contrasting findings have been reported regarding HFNC’s effect on EtCO₂. While Irfan et al. (14) and Douglas et al. (15) found no significant difference in EtCO₂ levels between HFNC and conventional oxygen therapy post-surgery, patients with CO₂ retention tendencies may develop inadequate ventilation when receiving oxygen supplementation, particularly with high-flow 100% oxygen, resulting in elevated EtCO₂ levels. Su et al. reviewed five randomized controlled trials involving 257 patients and drew conclusions that HFNC may reduce the incidence of hypoxemic events and improve oxygenation in patients undergoing FB but has no significant effect on the EtCO₂ value at the end of the surgery (34). Divergent findings regarding HFNC’s influence on CO₂ retention during FB sedation may reflect methodological differences in sedation administration and airway patency (particularly the open-mouth state during FB procedures). Notably, in the open-airway model, HFNC at 10 L/min achieves optimal CO₂ clearance—but this effect diminished in closed system thus requiring higher flow rates for equivalent efficacy (35). Meanwhile, the open-mouth state compromises HFNC’s ability to generate flow-dependent pressure in the upper airways, thereby limiting its efficacy in promoting ventilation and reducing CO₂ retention (36, 37). This may explain the observed absence of flow-dependence in CO₂ alleviation in this study, though these speculations remain unsubstantiated and require further mechanistic investigation.

This study presents several limitations. Firstly, although our data analysis indicates a significant advantage of HFNC in reducing the risk of desaturation during FB examination under deep sedation, larger sample sizes, multicenter studies, as well as well-designed randomized controlled clinical trials are needed to confirm and support these findings. Secondly, the results of this study are only applicable to the relatively healthy population with ASA grade 1 or 2. Although patients with obvious obesity were not excluded in this study, the actually enrolled patients had a BMI up to 34 kg/m₂, so the results of this study are not applicable to patients with obvious obesity. Thirdly, the cocktail regimen of deep sedation used in this study consists of a basal hypnotic dose of midazolam, a relative higher dosage (more than 20ug /kg) of alfentanil, and a remedial dose of propofol, our previous studies and other literature showed that its anti-stress effect was more significant, while the respiratory inhibitory effect was relatively mild (38, 39). However, it is quite different from the conventional sedation regimen recommended by guidelines or expert consensus. Therefore, whether the results are also applicable to other sedation regimens needs to be verified by further research. Finally, arterial blood gas was not analyzed in this study. Although EtCO₂ values were monitored at two time points before and after surgery, and the existence of respiratory depression and CO₂ accumulation were observed in this study, it did not fully reflect the CO₂ changes throughout the whole FB procedure.

The major risks of deep sedation with preserved spontaneous breathing during FB should be emphasized as respiratory depression, desaturation and CO₂ accumulation. In this study, HFNC improved oxygenation and ventilation by significantly reducing desaturation risk in a weakly flow-dependent manner and partially alleviating CO₂ retention, thus enhancing sedation efficacy and patient comfort. However, HFNC cannot completely eliminate desaturation risk or replace assisted ventilation, necessitating continuous monitoring and preparation for advanced ventilatory support.