One of the benefits of HFNO is that it can easily and comfortably be applied to a patient who is awake until anesthesia is induced. A study compared HFNO with an oxygen flow of 60 L/min for 3 min with oxygenation via a facemask (with an oxygen flow of 10 L/min) in 10 adult volunteers who were awake [32]. The resulting mean EtO₂ was similar between both techniques (86 kPa for HFNO and 89 kPa for a facemask). Nevertheless, in the case of HFNO, the EtO₂ decreased to 49 kPa when the volunteers opened their mouths, indicating that HFNO may be effective only with the mouth closed.

In another observational study, HFNO with an oxygen flow of 70 L/min was applied to 21 adult volunteers who were awake, and it was observed that while the volunteers were closing their mouths, the EtO₂ increased from 14–17% to 78–92% in 3 min [33]. Nevertheless, in half of the volunteers, the EtO₂ did not reach 90%. In addition, 1 patient experienced discomfort, and the test was stopped midway; further, an additional 4 patients reported experiencing moderate discomfort (5 or 6 on a visual analogue scale). Although it may be technically difficult to measure the EtO₂ accurately during HFNO, it is also possible that HFNO may not reliably increase the EtO₂ during preoxygenation for 3 min. In addition, patients in respiratory distress situations inevitably need to breathe with an open mouth, and it may be difficult for them to follow instructions requiring them to close their mouths [34]. The optimal parameters for the application of HFNO, such as the duration of preoxygenation, breathing pattern, or mouth status, need to be determined. In reported studies, different durations, ranging between 3–5 min, of HFNO were applied before the start of anesthesia induction [10,32,35–37]. To determine the optimal preoxygenation time with HFNO, apart from the EtO₂, new criteria, such as the oxygen reserve index sensor, should be considered in future clinical trials.

In obese patients, when preoxygenation is carried out through a facemask, the safe apnea time period is as short as 1–3 min following preoxygenation compared with 7–10 min in healthy adult patients [18,38]. This is due to obesity-related physiological changes such as decreased functional residual capacity, increased oxygen consumption, and increased closing volume [15]. In a study involving 33 morbidly obese patients (BMI ≥ 35 kg/m2) [39], HFNO at a flow rate of 50 L/min with the patient’s mouth closed was superior to facemask oxygenation in achieving a high arterial partial pressure of oxygen (PaO₂). After 3 min of preoxygenation with the patient in a 30° head-up position, the median PaO₂ was 380 mmHg in the HFNO group and 337 mmHg in the facemask group. None of the patients developed complications due to epistaxis or gastric aspiration when using HFNO. Therefore, in obese patients, the efficacy of HFNO is increased by placing the patient in a head-up position.

In obstetric patients, it is particularly important to make sure that the EtO₂ reaches over 90% by employing at least 3 min of tidal volume breathing of 100% oxygen [19,40] because the rate at which difficult intubation occurs in these patients may be 10 times higher than in the general population [12,41]. In addition, the safe apnea time period is relatively short due to several pregnancy-related changes such as airway edema, decreased functional residual capacity, increased oxygen consumption, and a higher risk of aspiration, as well as due to emergency surgery [40].

A computational model that simulates the effects of apneic oxygenation during rapid-sequence induction of anesthesia in obstetric women has indicated that the increase in FiO₂ to up to 1.0 extends the safe apnea time [42]. The successful use of HFNO in a pregnant woman with severe cardiopulmonary disease has been reported in the case of a 27-year-old pregnant woman with acute respiratory distress syndrome and heart failure who required an emergency Cesarean section [11]. The patient had dyspnea and exhibited significant mouth breathing, and her SpO₂ was 80%. Oxygen at a flow rate of 9 L/min via a facemask increased the SpO₂ to 95%, and HFNO at a flow rate of 70 L/min with an FiO₂ of 1.0 further increased the SpO₂ to 98% within 5 min even though the patient breathed predominantly through her mouth. After the induction of anesthesia, tracheal intubation was successfully performed with the SpO₂ maintained at 98%.

In contrast, the results of another study have cast a doubt on the efficacy of HFNO [35] in pregnant women: following HFNO for 3 min (30 L/min for 30 s, followed by 50 L/min for 150 s), an EtO₂ of 90% was achieved in merely 44 out of 73 (60%) pregnant women. In addition, although HFNO received similar scores associated with the comfort of its use to those of a facemask, only 56% of women preferred HFNO to the facemask. These results are similar to those associated with non-pregnant patients.

In 48 neurosurgical adult patients [10], the efficacy in increasing the PaO₂ between HFNO with an oxygen flow of 50 L/min and an oxygen flow of 10 L/min through a facemask was compared. After 5 min of preoxygenation, the median PaO₂ was significantly higher in the case of HFNO (471 mmHg) than in that of facemask oxygenation (357 mmHg). However, after the induction of anesthesia, the PaO₂ decreased significantly in the HFNO group but not in the facemask group in which bag-mask ventilation was performed. This suggests that HFNO is more efficient than facemask oxygenation in ensuring a high PaO₂ before the induction of anesthesia, but HFNO is less effective than facemask ventilation in maintaining a high PaO₂. Nevertheless, 7 patients in the facemask group required the use of airway adjuncts such as oropharyngeal airways, but none of the patients in the HFNO group required them, suggesting that HFNO is more easily incorporated.

A male patient who had acute epiglottitis requiring emergency tracheal intubation [44] exhibited the following characteristics: an airway assessed as a Mallampati score of 2, a thyromental distance of > 6 cm, poor dentition, adequate mouth opening and neck range of movement, and an easily palpable cricothyroid membrane. Despite the administration of intravenous antibiotics and steroids, the airway obstruction progressed rapidly and required airway intervention. HFNO at a flow rate of 70 L/min was used in combination with the target-controlled infusion (TCI) of propofol, thereby allowing spontaneous breathing. The minimum oxygen saturation was 96% following tracheal intubation.

Oxygenation can be more difficult to perform in obese patients because the incidences of difficult mask ventilation and tracheal intubation may increase with obesity. The head-up position is helpful to maximize the efficacy of apneic oxygenation by reducing atelectasis and the subsequent pulmonary shunting [12,29]. Another study conducted on obese patients demonstrated that HFNO at a flow rate of 50 L/min resulted in the highest PaO₂ value after 3–5 min of preoxygenation in morbidly obese patients and that the PaO₂ decreased as the preoxygenation duration was extended up to 7 min [39]. Therefore, the appropriate duration of preoxygenation using HFNO remains to be determined.

Compared to adults, children are more prone to the risk of rapid desaturation in the order of seconds after the cessation of ventilation due to their decreased functional residual capacity, higher oxygen consumption, increased closing capacity, and higher risk of airway collapse [17,45]. A study has demonstrated that the mean desaturation time until the SpO₂ reached 90% was significantly shorter in children (160 s) than in adolescents (382 s), and it was substantially shorter in infants (97 s). In fact, the incidence of desaturation is common in children (4–10% during induction of anesthesia and 20% during tracheal intubation) [18]. Therefore, in children, optimal oxygenation strategies are crucial in order to extend the duration of safe apnea.

HFNO may extend the safe apnea time in children. In a study involving 48 children who were under 10 years with normal airways [46], in the facemask group, bag-valve ventilation was ceased and the jaw thrust was maintained whereas in the HFNO group, the flow rates of 2 L/kg/min, 35, 40, and 50 L/min were applied based on the body weight, and the jaw thrust method was performed during apnea. The main hypothesis was that the safe apnea time (time taken for the SpO₂ to reach 92%) in the case of HFNO would be more than double the length of the apnea time in the facemask group and the results confirm that this hypothesis was true.

In another study involving 60 children, aged 1–6 years, weighing 10–20 kg, HFNO at a flow rate of 2 L/kg/min with an FiO₂ of 1.0, HFNO at a flow rate of 2 L/kg/min with an FiO₂ of 0.3, and low-flow nasal oxygenation at a flow rate of 0.2 L/kg/min with an FiO₂ of 1.0 were compared [47]. Bag-valve mask ventilation was ceased when the target EtO₂ of > 90% was reached. Apnea was terminated when the SpO₂ was less than 95%, transcutaneous carbon dioxide (tcCO₂) reached 65 mmHg, or apnea duration reached 10 min. The SpO₂ had decreased to < 95% within 10 min in all the HFNO (FiO₂ 0.3) patients and in 3 of the low-flow nasal oxygenation patients (17%), but it did not decrease in any of the HFNO (FiO₂ 1.0) patients, indicating that HFNO at a flow rate of 2 L/kg/min with an FiO₂ of 0.3 is not effective. The tcCO₂ exceeded above 65 mmHg within 10 min in 16 (80%) HFNO patients with an FiO₂ of 1.0 and in 13 (72%) low-flow nasal oxygenation patients. In the case of HFNO with a flow rate of 2 L/kg/min and an FiO₂ of 1.0, the median safe apnea time was 7.6 min with a range of 5.2–10 min. Therefore, in children, HFNO with a flow rate of 2 L/kg/min and an FiO₂ of 1.0 can maintain the SpO₂ for 10 min; however, there is a higher risk of hypercapnea with this technique, and therefore, it may be useful for surgeries that last between 5–6 min. It is not known whether or not a higher flow rate (e.g., 4 L/kg/min) with an FiO₂ of less than 1.0 can be effective.

HFNO may also have a potential role during rapid-sequence induction of anesthesia in the operating room. In a study involving 80 adults who were receiving rapid-sequence induction of anesthesia, the efficacy of oxygenation was compared between HFNO at a flow rate of 70 L/min and facemask oxygenation at a rate of 10 L/min [36]. The lowest SpO₂ up to 1 min after tracheal intubation was similar between the techniques, but the SpO₂ decreased to < 96% in 7 patients (18%) in the facemask group whereas this was not observed in the HFNO group. None of the patients developed complications, such as the regurgitation of gastric contents, and there was no significant difference in the duration of the safe apnea time between HFNO (median: 116 s) and facemask (median: 109 s).

In another study involving 40 adult patients receiving rapid-sequence induction of anesthesia for emergency surgery, the efficacy between HFNO at the rate of 70 L/min and facemask oxygenation at the rate of 12 L/min were compared [48]. In the HFNO group, HFNO was maintained throughout the induction period whereas in the facemask group, the jaw was kept thrust out without bag-mask ventilation. In all the patients, the trachea was intubated successfully and there were no significant differences in the PaO₂ between the groups. Nevertheless, it took significantly longer to intubate the trachea in the HFNO group (the mean of 248 s) than in the facemask group (123 s). The authors of the report state that this difference was not due to any apparent difference in procedural difficulty; however, the difference was quite marked. One of the reasons might be that the presence of a nasal cannula mechanically or psychologically facilitated a more careful time-consuming laryngoscopy and tracheal intubation.

These results indicate that HFNO with the jaw thrust is an efficient method to prevent desaturation within 3–4 min of apnea during rapid-sequence induction of anesthesia in adults even if cricoid pressure is being applied.

Securing the airway is frequently more difficult outside the operating rooms than inside [49,50]. This is because outside the operating room, the unstable physiological status of patients, such as being affected by a cardiopulmonary disease, having a low cardiac output, or being in hypermetabolic states, limits the efficiency of preoxygenation and peroxygenation. In addition, emergency procedures, poor planning, less skilled staff, and the inadequate availability of equipment increase the incidence of difficult intubation to up to approximately 12% [43,51,52]. Therefore, tracheal intubation is associated with a high incidence of hypoxemia, which is the most commonly associated complication (approximately 19–26%) [51,53] that is linked to hemodynamic deterioration, cardiac arrest, and death [12,54]. The incidence of death related to airway management has been reported to be 38-fold higher in the emergency department and 58-fold higher in the ICU than in the operating rooms [12].

Meta-analyses have shown that oxygenation through a nasal cannula at various flow rates is effective in preventing desaturation in patients who received emergency tracheal intubation [20,43]. In a study involving 101 adult patients who required rapid-sequence induction of anesthesia and intubation in the ICU [52], the efficacy of HFNO was compared with that of a non-rebreathing bag reservoir facemask combined with a nasopharyngeal catheter. For the HFNO group, HFNO at a flow rate of 60 L/min with an FiO₂ of 1.0 was used for 3 min before the induction of anesthesia (preoxygenation) and was continued after the induction of anesthesia and during tracheal intubation. For the facemask group, 15 L/min of oxygen was provided for at least 3 min before the induction of anesthesia, and after the induction, the facemask was removed, a nasal cannula was placed, and 6 L/min of oxygen was insufflated. The median of the lowest SpO₂ was significantly higher in the HFNO group (100%, range: 95–100%) than in facemask group (94%; 83–99%). Episodes of severe hypoxemia (SpO₂ < 80%) were significantly lower in the HFNO group (2%) than in facemask group (14%); further, 1 cardiac arrest due to hypoxemia occurred in the facemask group. A multivariate analysis indicated that HFNO was an independent factor that prevented severe hypoxemia.

Nevertheless, HNFO may not be effective in improving oxygenation when tracheal intubation is required in hypoxemic patients. In a study involving 119 patients who had acute hypoxemic respiratory failure that required rapid-sequence induction and intubation (respiratory rate > 30 breaths/min, FiO₂ requirement > 50%, PaO₂/FiO₂ < 300 mmHg) [55], either HFNO at a flow rate of 60 L/min with an FiO₂ of 1.0 or facemask oxygenation at a flow rate of 15 L/min was used. There was no significant difference in the lowest SpO₂ level (92% in HFNO group and 90% in facemask group) and in the occurrence of severe hypoxemia involving an SpO₂ < 80% during tracheal intubation (16 [26%] in the HFNO group and 13 [22%] in the facemask group).

Similarly, another study that compared the efficacy of HFNO at a flow rate of 50 L/min and that of bag-valve-mask ventilation at a flow rate of 10 L/min in 40 critically ill patients who suffered hypoxemic respiratory failure (PaO₂/FiO₂ < 300 mmHg) [34], the lowest mean SpO₂ during tracheal intubation and the occurrence of severe hypoxemia (SpO₂ < 80%) were similar between the groups.

In a report on 49 adult patients who required rapid-sequence induction due to severe acute hypoxemic respiratory failure in the ICU (respiratory rate > 30 breaths/min, FiO₂ requirement ≥ 50%, PaO₂/FiO₂ < 300 mmHg) [56], the efficacy of preoxygenation with HFNO at a flow rate of 60 L/min with an FiO₂ of 1.0 together with non-invasive ventilation was compared with that of non-invasive ventilation performed on its own. HFNO plus non-invasive ventilation resulted in improved oxygenation and less frequent episodes of desaturation (SpO₂ < 80%). Therefore, HFNO combined with non-invasive ventilation can reduce the incidence of severe hypoxemia during rapid-sequence induction of anesthesia in patients with severe acute hypoxemic respiratory failure.

In a prospective observational study that investigated the efficacy of HFNO at a flow rate of 60 L/min during emergency tracheal intubation in 71 adult patients in the ICU, in the operating room, or in the emergency department, the median (range) apnea time was 80 (30–480) s [37]. Significant desaturation (reduction in SpO₂ to > 10% after induction of anesthesia) occurred in 5 patients (7%) who underwent acute respiratory failure (2 patients) or who had difficulty maintaining airway patency during apnea (3 patients). There were no complications resulting from the use of HFNO.

In the ICU, patients’ conditions in relation to their respiratory diseases and the techniques used in the control and intervention groups are varied. The intubation protocol including intubation equipment, patient positioning, preoxygenation approach, operators, and medications administered have not been standardized [57]. This may be a reason for the discrepancies among out-of-hospital studies. Moreover, the definitions of clinical outcomes vary between studies. For example, the standard that is taken into consideration to define hypoxemia is usually the SpO₂ value, which is the starting point of the steep slope on the oxygen dissociation curve of hemoglobin [58]. Nevertheless, desaturation can be defined differently, which may be another reason for the conflicting results. Therefore, the efficacy of apneic HFNO compared to that of low-flow nasal apneic oxygenation is still unclear in emergency airway management in the ICU [3,20]. At least it is now apparent that HFNO alone may not be beneficial to patients who suffer severe respiratory failure [34,55]. This may be because pulmonary shunting or easily collapsed airways diminish the benefits of HFNO. To overcome this limitation, other techniques that provide high positive airway pressure may be required [58].

One major potential problem in relation to apneic oxygenation via HFNO is excessive hypercapnea. A study has demonstrated that the partial pressure of carbon dioxide (PaCO₂) increased with the speed of 3 mmHg/min during apneic oxygenation via HFNO at a flow rate of 50 L/min [10]. The PaCO₂ was significantly higher during apnea with HFNO than during bag-mask ventilation, but the resultant rise in the PaCO₂ with HFNO seemed tolerable because the highest PaCO₂ value was approximately 65 mmHg after the completion of tracheal intubation. Flow rates of up to 70 L/min may be required to achieve the maximum clearance of carbon dioxide. In contrast, carbon dioxide accumulation was similar between HFNO and facemask preoxygenation during rapid-sequence induction during emergency surgery [36,48] potentially due to a short observation period and the increase in carbon dioxide at a more rapid rate during the first minute following the onset of apnea. The clearance of carbon dioxide during HFNO may be superior to that resulting from a low-flow oxygen delivery system [59].

In children aged up to 10 years [46], the mean tcCO₂ that was measured at the end of apnea was 62 mmHg (range 49–79 mmHg), and the mean rate of the increase in carbon dioxide was 2.4 mmHg/min (range 0.2–3.9 mmHg/min). However, the rate of increase in carbon dioxide was not significantly lower when HFNO was combined with the jaw thrust method compared to when jaw thrust was used in isolation during apnea. In a study involving children aged up to 6 years [47], the tcCO₂ increased at a median rate (range) of 4.1 (2.2–5.3) kPa/min in response to HFNO with an FiO₂ of 1.0, and the rate of increase in the tcCO₂ was similar to that observed in the low-flow nasal oxygenation group. It has also been demonstrated that the younger the child, the faster the increase in the tcCO₂ during apnea. These results imply that the efficiency in carbon dioxide clearance via HFNO is not prominent in children. This may be because the increase in carbon dioxide is higher in children than in adults and in younger children than in older children due to their higher metabolic demands.

The application of HFNO interrupts the earlier detection of a rise in carbon dioxide and of airway obstruction compared to when bag-mask ventilation is used. Therefore, the transcutaneous monitoring of carbon dioxide [10,46] and the use of oxygen reserve index sensors may help in minimizing this risk and in optimizing the utilization of HFNO.

Gastric insufflation is a theoretical complication resulting from HFNO because HFNO generates positive airway pressure. The increase in the nasal flow rate to 10 L/min is known to cause a 1.2 cmH₂O increase in the nasopharyngeal airway pressure, as observed when healthy volunteers breathed with their mouths closed when HFNO was being used [4], and thus, the airway pressure would increase to around 3 cmH₂O at 30 L/min and to around 12 cmH₂O at 100 L/min.

No serious complications, such as gastric insufflation, regurgitation, or pulmonary aspiration of gastric contents, have so far been reported even in the case of morbidly obese patients [36,39]. The number of studies on this subject is still limited, and thus, the true risk of gastric insufflation associated with the use of HFNO during anesthesia remains to be elucidated.

Reports on the use of HFNO have indicated various high-flow rates such as 50 L/min [10,34,35,39], 60 L/min [32,37,52,55,56], or 70 L/min [33,36,48]. One study has demonstrated that the median intratracheal FiO₂ significantly increased from 67% to 93% as the flow rate increased from 15 L/min to 45 L/min [60]. Similarly, another study has demonstrated that increasing the flow rate of HFNO from 10 L/min to 50 L/min can yield a higher FiO₂ [61]. Therefore, using a higher flow rate of greater than 50 L/min is advisable to obtain the maximal effects of oxygenation.

Generally, HFNO is well tolerated by patients who are awake [35,36], but some of them may experience moderate or severe discomfort when a flow of 50–70 L/min is used [10,33]. Therefore, it may be prudent to start HFNO with a relatively low flow rate (30–40 L/min) when a patient is awake and then to increase the flow rate (50–70 L/min). When a patient cannot tolerate a high flow, the flow should immediately be reduced to 30–40 L/min, and once the patient has lost consciousness, the flow should be increased back to 50–70 L/min [10,35,36,48]. Concerning children, the adequate flow rate has been indicated to be 2 L/kg/min for those weighing 0–15 (or 20) kg, 35 L/min for those weighing 15 (or 20) –30 kg, 40 L/min for those weighing 30–50 kg, and 50 L/min for those weighing > 50 kg [46,47].

As the FiO₂ increases, the safe apnea time gets extended. In addition, a study has demonstrated that an increase in the FiO₂ from 0.9 to 1.0 yielded a greater increase in the safe apnea time compared to an increase in the FiO₂ from 0.21 to 0.9 [62]. Therefore, the FiO₂ should usually be set to 1.0 in order to maximize the safe apnea time in adults including in morbidly obese patients [39] and in those requiring rapid-sequence induction of anesthesia [36,52].

In children, HFNO with an FiO₂ of 0.3 may not be as efficient in extending the safe apnea time when compared to HFNO with an FiO₂ of 1.0 [47]; therefore, when HFNO is used, the FiO₂ should be kept high.

There is still little evidence with regard to which breathing method is the best. In some studies, patients were asked to maintain normal breathing [32,33] whereas in other studies, deep breathing was preferred [35]. One clear characteristic is that the efficacy of HFNO in oxygenation is reduced when the patient’s mouth is open [2]. In a study [63] in which HFNO with a flow rate of 35 L/min was applied in postoperative cardiac surgery adult patients, the mean nasopharyngeal positive airway pressure was 2.7 cmH₂O when the patient’s mouth was closed whereas it was only 1.2 cmH₂O when the mouth was open. Similarly, in another study [64], the increase in the flow rates of HFNO resulted in increased mean airway pressures of 0.7 cmH₂O per every 10 L/min when the mouth was closed, and the increased rate of airway pressure significantly reduced to 0.4 cmH₂O when the mouth was open. Further [60], the median of mean airway pressure significantly increased from 0.4 cmH₂O to 2 cmH₂O as the oxygen flow was increased from 15 L/min to 45 L/min, but significantly decreased from 2 cmH₂O to 0.6 cmH₂O when the mouth was open. However, the FiO₂ was not significantly different between when the mouth was closed or open with HFNO at a flow rate of 45 L/min [60]. In a study in which the EtO₂ was measured during HFNO at a flow rate of 60 L/min [32], HFNO increased the EtO₂ when the patient’s mouth was closed (86 kPa) whereas it failed to increase the EtO₂ when the patient’s mouth was open (49 kPa).

Another crucial point to consider during HFNO is that the patency of the upper airway is essential to provide adequate oxygenation [17]. The patency of the upper airway needs to be maintained using a head tilt [37], jaw thrust [36,37,46–48], chin lift [36,37], or the insertion of a pharyngeal airway [47].

Theoretically, deep breathing through the nose with the mouth shut would be the best patient status for preoxygenation [32,33,35,39].

Patel and Nouraei [59] were the first to report the use of HFNO during airway surgery. They used the technique in 25 adult patients who were undergoing hypopharyngeal or laryngotracheal surgeries, such as for benign laryngeal conditions, for obstructive sleep apnea, or for benign or malignant head and neck conditions. Their study included 12 obese patients and 9 patients with stridor. The median BMI (range) was 30 (18–52) kg/m2. The median apnea time was 14 min, with the longest duration being 65 min. None of the patients exhibited desaturation, i.e., an SpO₂ of < 90%, when HFNO at a flow rate of 70 L/min was used. Further, none of the patients experienced complications, such as cardiac arrhythmia, due to an increase in carbon dioxide during apnea.

Booth et al. [70] employed a ‘traditional’ use of HFNO during airway surgery by maintaining spontaneous breathing. In 30 adult patients who underwent laryngotracheal surgeries due to laryngotracheal stenosis or papilloma, HFNO at a rate of 70 L/min was applied while anesthesia was maintained with TCI propofol, thereby maintaining spontaneous breathing. The median duration of spontaneous breathing (range) was 44 (18–100) min. Only 1 patient experienced desaturation due to a miscalculated remifentanil overdose, and tracheal intubation was required during the surgery. In another 3 patients, the SpO₂ decreased to less than 90% during laser surgery when the FiO₂ was reduced to 0.3. Nevertheless, the SpO₂ increased rapidly when the FiO₂ was increased back to 1.0.

In another report on 28 adult patients who were undergoing airway surgeries, apneic oxygenation was provided via a unique system called the Perioperative Insufflatory Nasal Therapy system at a flow rate of 80 L/min [71]. In 4 patients, the SpO₂ decreased to 85–90%, lasting for less than 2 min at several time points, following the injection of rocuronium. The oxygen saturation was increased by the jaw thrust method and by increasing the flow rate to 120 L/min following the removal of the suspension laryngoscope or the insertion of a supraglottic airway. A different HFNO system (Optiflow) has a maximal flow rate of only 70 L/min; therefore, the rescue technique of increasing the flow rate to up to 120 L/min is not feasible with this system.

HFNO has also been successfully used in patients with difficult airways. For example, in a 55-year-old male who had severe supraglottic-pharyngeal stenosis, HFNO with a flow rate of 70 L/min and an FiO₂ of 1.0 was applied, and the patient remained apneic for 26 min with the use of propofol, remifentanil, and rocuronium [72]. After the scar tissue was transected along the lateral epiglottic border resulting in the partial release of the stenotic lesion, the laser-safe tracheal tube was inserted through the opening. Until the time point at which the tracheal tube was inserted, the SpO₂ was maintained at greater than 90%. In another report on a male patient who was morbidly obese with a BMI 40 kg/m2 and who had a neck circumference of 45 cm, mouth opening of 2 finger breaths, and Mallampati score of 4, HFNO at a flow rate of 60 L/min allowed for the performance of a 14-min long surgery without desaturation occurring [73].

HFNO may also be used in children, including in premature babies, who are undergoing airway surgeries. For example, a report describes the successful use of HFNO (at the flow rate of 2 L/kg/min) in more than 30 children (aged 6 days to 13 years) undergoing airway surgeries (e.g., aryepiglottoplasty, subglottic cyst excisions, tracheal dilatations, and endoscopic cricoid splints) [74]. The FiO₂ was titrated to maintain the SpO₂ at 95–99%, and spontaneous breathing was successfully preserved during surgery.

A study reported on a premature male baby who had a subglottic web after repeated tracheal intubations. He had severe inspiratory stridor that required laser resection and dilatation under general anesthesia [68]. Anesthesia was induced with sevoflurane and maintained with an infusion of propofol and remifentanil. Rocuronium was used to induce a full neuromuscular blockade, and an uncuffed tube was inserted. The surgeon placed the Parsons laryngoscope using the suspension apparatus. For the resection of the subglottic web, the trachea was extubated until the SpO₂ fell to 80%, following which it was re-intubated. The apnea time was 39 and 41 s; however, it could be increased to 95 and 160 s when HFNO with a flow rate of 4 L/kg/min and an FiO₂ of 0.3 was applied. When HFNO with an FiO₂ of 1.0 was used, the apnea time extended to over 4 min. Following 2–4 min of apnea during HFNO, the EtCO₂ was measured as 48–66 mmHg. This case report indicates that HFNO is efficient in extending the safe apnea time in an infant undergoing laryngeal repair even with a low FiO₂ of 0.3.

One of the major complications associated with the use of HFNO during airway surgeries is the failure to maintain oxygenation. For example, in a report on the use of HFNO at a flow rate of 50 L/min during laryngomicrosurgery [75], the SpO₂ decreased to 72% in 1 of the 23 studied patients, and tracheal intubation was temporarily required. In this case, HFNO was restarted, and the surgery could be completed without further complications.

In 20 infants and children (aged between 5 days to 15 years) who were undergoing upper airway surgeries or dynamic airway assessments lasting between 3–61 min, oxygenation was maintained via HFNO, but 1 of the neonates required rescue tracheal intubation at 3 min following apnea due to desaturation of 77% [76]. In a 35-year-old man with a parapharyngeal mass and a narrowing of the trachea to 4 mm, an emergency surgical tracheostomy was performed under sedation while HFNO at a flow rate of 30 L/min was applied with the patient in a 40° headup position [77]. When the sedation level was deepened, the respiratory rate decreased, and complete airway obstruction occurred. The SpO₂ rapidly decreased to 89%, and an immediate tracheostomy was required. This case indicates that if the airway is obstructed, HFNO could delay but not prevent the onset of hypoxemia.

Hypercapnea is another major complication that occurs during airway surgeries because unlike short periods of apneic oxygenation before tracheal intubation, apneic oxygenation during airway surgery could last for as long as 40–60 min. There have been some reports on excessive hypercapnea requiring tracheal intubation or jet ventilation [78]. For example, in a case series on 30 non-obese patients who underwent elective short laryngeal procedures, HFNO at a flow rate of 70 L/min was applied while the patients were under apnea during surgery, and the surgeon inserted a rigid tubular laryngoscope [79]. In 1 of the patients, HFNO had to be stopped, and supraglottic jet ventilation was required because the PaCO₂ exceeded a predetermined threshold of 11 kPa.

Compared with low-flow oxygenation, HFNO enables a greater clearance of carbon dioxide [71]. When HFNO (70–80 L/min) is used in adults for apneic oxygenation, the EtCO₂ increases at a rate of 0.12–0.17 kPa/min [59,71,79]. By maintaining spontaneous breathing, the increase in EtCO₂ can be reduced further (e.g., 0.03 kPa/min) [70].

The increase in EtCO₂ progresses much faster in children than in adults. For example, in children weighing 10–20 kg, the mean rate of increase in tcCO₂ was 0.55 kPa/min [47]. Younger children exhibited a more rapid carbon dioxide increase [47]. In a case report on infants weighing 4 kg, the EtCO₂ increased from 35 mmHg to 51–52 mmHg after 250–251 s of apnea with HFNO at a rate of 4 L/kg/min [68].

One feature of the increase in carbon dioxide during apnea is that for an initial 1–2 min, the amount of carbon dioxide increases rapidly and then slows down [80]. For example, in a study [80], apnea increased the PaCO₂ at a rate of 12 mmHg/min in the first minute, followed by a slow rise to 3.4 mmHg/min.

The exact mechanism of carbon dioxide clearance during HFNO is unclear although cardiogenic oscillations, dead space gas mixing, and micro-ventilation induced by pharyngeal pressure variations appear to be important factors [81].

Airway fire is a well-known complication that can occur during head and neck surgeries using lasers. Ignition sources, such as diathermy or laser rays, may come into accidental contact with inflammable items, such as tracheal tubes or surgical drapes, in an oxygen rich environment, and surgical fire can occur [82]. If HFNO is used during laser airway surgeries, a high concentration of oxygen flows through the surgical field of the airway, making it easier to ignite when the electrosurgical devices come into contact with this fuel. A case report on a 65-yearold female patient has been presented [83]. She was scheduled to receive a palatal biopsy. She had titanium dental implants in order to attach a palatal obturator following a previous surgery. HFNO at a rate of 30 L/min and an FiO₂ of 1.0 was used while she was spontaneously breathing following the administration of propofol and fentanyl. A burn occurred on the diathermy shaft when the surgeon used a diathermy device to achieve hemostasis at the biopsy site. Fortunately, the tissue of the patient was not harmed. It is recommended that during the use of HFNO, the FiO₂ should be lowered to the minimum level [71]. One possible method is to titrate the FiO₂ to 0.3 using a gas blender [70]. It is recommended that clinicians be constantly vigilant concerning the risk of airway fires during laser or diathermy procedures.