In our study, we aimed to investigate the effectiveness of SJOV via WNJ can reduce the incidence of hypoxia in patients under deep sedation during ERCP. According to our results, SJOV can effectively improve oxygenation during ERCP in deeply sedated semiprone patients. Nasopharyngeal airway can relieve upper airway obstruction to some extent, but for its tip far from glottis, it is too difficult to maintain oxygenation under suppressed breathing. We found that the incidence of hypoxemia was significantly higher in Group N, which is consistent with the results of Han S J [
20], while the incidence of hypoxemia was lower in Group W (20%) and Group WS (4%). WNJ can ensure adequate oxygenation under deep sedation, which significantly reduces the incidence of hypoxia compared to the commonly used oxygen supplies via a nasopharyngeal airway [
11]. The WNJ is a rapidly inserted device that requires no previous experience or practice to use and it is well tolerated by lightly sedated patients [
21]. In our study, each WNJ was placed appropriately in Group W and Group WS. Easy implementation may allow the WNJ with SJOV to be used in emergent airway management as a rescue device [
22]. Based on this feature, SJOV via WNJ is considered especially appropriate for the resuscitation of injured people who suffered from acute trauma in war fields [
21].
The SJOV system used via WNJ is shown in Fig.
1(2). We confirmed the location of the tip of the WNJ by duodenal rectoscope, concluding that the gap of the epiglottis and uvula was the optimized location for WNJ [Fig.
1(4)]. If the location of WNJ was appropriate, the wave of P
ETCO
2 would be regular, and the best position of distal end of WNJ between epiglottis and vulvar usually provide highest P
ETCO
2. Oxygenation was better maintained in Group WS than in Group W (Table
2), suggesting that the high-pressure jet pulse promotes ventilation via WNJ. Thus, we inferred that WNJ maintains adequate oxygenation performed primarily by SJOV instead of by recovering spontaneous respiration or relieving an obstructive airway. Furthermore, an animal study confirmed that the SpO
2 could be maintained over 95% for more than 20 min by SJOV without the need for assisted mask ventilation in an apnoeic pig [
23]. The additional utilization of a built-in CO
2 monitoring catheter on the WNJ may help detect depressed or apnoeic breathing which often occurred after induction of anesthesia. Therefore, SJOV via WNJ lasting for 5 min at the beginning of the procedure has a good effect on relieving hypoxemia caused by hypoventilation. Although SJOV can maintain the desired oxygenation for up to 1 h according to previously reported clinical trials, there is insufficient evidence to support the use of SJOV for a longer duration [
24]. In our study, all procedures during ERCP were performed within one hour. We excluded patients whose procedure time exceeded 2 h. The maximum application time and potential airway mucosal inflammation remain to be investigated in the future. The incidences of intraoperative adverse events except SRAEs and subclinical respiratory depression were comparable among all three groups (Table
3). The data on non-invasive arterial pressure and heart rate are listed in Supplementary Table
2, and all patients’ hemodynamic were stable. There was no barotrauma or airway injuries among three groups, and there were no significant differences in the incidence of nose bleeding, nausea or vomiting, dysphoria and pharyngalgia (Table
4). Compared with Group N, the application of SJOV via WNJ significantly increased the incidence of xerostomia from 2 to 20% (
P = 0.002) at 5 min after the procedure, but there were no significant differences between two groups at 30 min and 24 h after the procedure. The utilization of humidified oxygen could potentially decrease the incidence of xerostomia after using SJOV. As hyoscine was intravenously injected preprocedure, xerostomia may inevitably be attributed to the gland secretion inhibition caused by hyoscine, but it generally disappeared within 30 min postoperatively without any treatment. All of the above complications were tolerable and manageable without difficulties, similar to the results of previous clinical studies [
10,
11]. In addition to conveniently using WNJ for SJOV, there are varying techniques reported to perform SJOV, such as the application of a soft-suction catheter [
25] and a Cook airway exchange catheter [
26]. Compared to the aforementioned techniques, an advantage of WNJ is that it can measure the P
ETCO
2 of the patients, which may help detect problems such as respiratory insufficiency or mechanical failure during anaesthesia in time for timely intervention. A device named the Hague Airway can also monitor P
ETCO
2, but its inability to prevent airway obstruction make it limited clinical applications [
27]. Furthermore, a high-flow nasal cannula (HFNC), a new oxygen delivery device, is adapted to maintain oxygenation and humidification ventilation. The high flow of gas delivered by the HFNC through the nasopharynx and airways generates a positive end-expiratory pressure, which increases the effective alveolar ventilation, thereby increasing respiratory efficiency and improving oxygenation, but the HFNC is unable to correct the upper airway obstruction induced in the semi-prone position. In addition, HFNC does not directly monitor P
ETCO
2, and the apparent SpO
2 may mask the risk of carbon dioxide accumulation. The LMA® Gastro™ Airway is an airway technique, which could improve airway control, prevent hypoxia and avoid the need for intubation [
18]. However, based on our clinical experience, there were difficulties to keep LMA in appropriate position when turned patients to semiprone position. With the advantages of a more open ventilating system, fewer complications, and a lower requirement for spontaneous breathing [
3], SJOV has shown its versatility in the emergency airway to support sufficient ventilation [
21] and in difficult airways to achieve desirable oxygenation. In our study, the incidence of SRAEs for patients undergoing ERCP under deep sedation were successfully decreased through SJOV via WNJ. Furthermore, lower incidences of muscle twitch, cough and body movement were observed intraoperatively in our study, which indicated that SJOV has the potential to reduce body movements for safer procedures.
Some limitations exist in our study. First, our study is a single-blinded trial with potential biases of outcome assessment, but the objective parameter of hypoxemia might correct for the single-blindness in the present study. Second, the sedation strategy in our study consisted of propofol, remifentanil and dexmedetomidine, so our results cannot be applied to patients with other sedative strategy. Third, we only enrolled patients aged 18 to 65 years with ASA classifications from I to III. The available evidence has demonstrated that the pharmacokinetics and pharmacodynamics of medicine are significantly affected by age [
28]. A previous study showed that older age, higher BMI, higher ASA class and longer procedure duration led to higher rates of SRAEs [
29]. Therefore, future studies should focus on specific and high-risk patients under deep sedation who may benefit from SJOV in different procedures. Furthermore, the combination use of some useful current monitors such as Oxygen Reserve Index may make the use of SJOV via WNJ safer, which needs further validation.