This prospective, randomized controlled trial was conducted at the First Affiliated Hospital of Xi’an Jiaotong University in China between February 2023 and July 2024.

The study protocol adhered to the Consolidated Standards of Reporting Trials (CONSORT) guidelines (12). The procedure was conducted in accordance with the Helsinki Declaration-2013.¹ The study was approved by the ethics committee of the First Affiliated Hospital of Xi’an Jiaotong University, Xi’an, Shaanxi, China (approval number: XJTU1AF2022LSK-402) on 25 October 2022, and was registered at the Clinicaltrials.gov (NCT05701839) on 28 December 2022 before the first participant was enrolled. Written informed consents were obtained from all the participants.

There were slight changes from the primary protocol. The threshold of ORI to perform assist ventilation was changed from 0.1 to zero in ORI monitoring groups. Because the SpO₂ was usually 100% when ORI was 0.1, there may have excessive intervention in primary protocol. The threshold of SpO₂ to perform assist ventilation was changed from 95 to 98% in non-ORI monitoring groups. Because the primary outcome was the incidence of SpO₂ less than 95%, it seems that there has been no attempt to reduce the primary outcome in primary protocol. The changes were occurred on January 2023 before the first participant was enrolled and decided by all authors. There was not accumulated outcome data before the protocol amendment occurred.

Participants scheduled for hysteroscopy under deep sedation were assessed for eligibility. Participants complied with all the following criteria could be enrolled in this study: aged between 18 and 60 years, the predicted procedure duration was more than 10 min. Participants met any of the following criteria could not be enrolled: body mass index (BMI) more than 30 kg/m², predicted difficult airway, comorbidities including upper respiratory tract infection, pneumonia, obstructive sleep apnea syndrome, chronic obstructive pulmonary disease, asthma, coagulopathy, ongoing anticoagulant therapy, nasal deformity, and pregnant.

Based on whether ORI monitoring was adopted or not, and the oxygen supply methods during procedure [through face mask (FM) or nasopharyngeal airway (NPA)], participants were randomly divided into ORI+FM group, ORI+NPA group, non-ORI+FM group, and non-ORI+NPA group in a 1:1:1:1 ratio using random number table generated by Excel software. The specific method of random grouping was as follows: entered a set of data from 1 to 400 in column A, entered the formula “=RAND()” in column B1, and dropped down formula from B1 to B400 to generate random data, arranging the data in column B in ascending order. Participants were numbered serially according to the order of admission, the number above the line A100 (include A100) were divided into the ORI+FM group, the number between A101-A200 were divided into ORI+NPA group, the number between A201-A300 were divided into non-ORI+FM group, and the number below the line A301 (include A301) were divided into non-ORI+NPA group (The randomization list was provided as a Supplementary Table S3). The randomization codes were kept in sealed envelopes, they were opaque, tamper-proof and sequentially numbered. They were prepared and maintained by clinical trial assistant in our department who was not involved in this study. Once participants entered the operation room, the envelopes were given in order and opened by anesthesiologists. Participants were not informed of which oxygen supply methods and monitoring methods to use, the NPA was inserted after sedation. Outcome assessors were independent to anesthesiologists, they just assessed the severity of sore neck, sore jaw, and sore throat, they also calculated the time-weighted average duration of assist ventilation and performed the statistical analysis. Participants and outcome assessors were all blinded to group assignment.

Participants fasted solid and liquid for 8 and 2 h respectively, electrocardiogram, SpO₂, blood pressure were monitored, the blood pressure measured every 3 min. ORI was monitored in ORI monitoring groups by finger-cot sensor which was worn on the participant’s right index finger, it was connected to ORI monitor (Masimo corp., Irvine, CA, USA). In non-ORI monitoring groups, the finger-cot sensor was worn, but the ORI was not monitored. 5 μg of sufentanil was injected firstly, followed by target-control infusion of propofol at a plasma concentration of 2–3 μg/mL, additional sufentanil was administered at the discretion of anesthesiologist. Sedative degree was evaluated using Ramsay Sedation Scale, the score should equal or more than five, namely participants had a sluggish response or no response to shaking or loud sound (13).

All the participants inhaled oxygen through FM at a flow rate of 5 L/min spontaneously for 5 min before sedation. In ORI+FM and non-ORI+FM groups, oxygen was supplied through FM continuously after sedation. In ORI+NPA and non-ORI+NPA groups, the NPA was inserted after sedation and oxygen was supplied through it. The oxygen flow rate was all of 5 L/min. Airway intervention includes chin lift, jaw thrust, neck extension, and assist ventilation. In ORI monitoring groups, assist ventilation through FM was performed when ORI dropped to zero, it lasted for another 10 s after ORI restored to 0.1. In non-ORI monitoring groups, assist ventilation through FM was performed when SpO₂ dropped to 98%, it lasted for another 10 s after SpO₂ restored to 100%. If assist ventilation was ineffective, laryngeal mask airway (LMA) would be inserted and mechanical ventilation would be performed, the data would be analyzed as the full analysis set. If systolic blood pressure was more than 180 mmHg, 20 μg of nitroglycerin would be used. If systolic blood pressure was less than 90 mmHg, 6 mg of ephedrine would be used.

The primary outcome was the incidence of low SpO₂, which defined as SpO₂ less than 95% during procedure. The secondary outcomes include: (1) Average duration of assist ventilation. (2) Time-weighted average duration of assist ventilation, which defined as the average duration of assist ventilation divided by procedure duration. (3) The severity of sore neck, sore jaw, and sore throat. The severity was evaluated using Visual Analogue Scale (VAS), 0 indicated no discomfort, 1–3 indicated mild discomfort, 4–6 indicated moderate discomfort, 7–10 indicated severe discomfort. (4) The airway injury, which defined as bleed or bloody secretion found on the tip of NPA or in the participants’ mouth or nose.

The primary outcome was the incidence of low SpO₂. Previous study showed that the incidence of SpO₂ less than 95% during deep sedation for hysteroscopy was 51.2% (14), our pilot study showed that the incidence of SpO₂ less than 95% with ORI monitoring was 34.4% (11/32). Using these data to calculate the sample size to achieve a statistical power of 0.9 and alpha error of 0.05 using two-sided test, accounting for 10% dropouts, both the ORI monitoring groups and the non-ORI monitoring groups required 200 participants.

The statistical analysis was performed using SPSS 26.0 software. We analyzed participants who were randomized as the full analysis set. The per protocol set, a subset of the full analysis set, included participants without major protocol deviations. The participants characteristics were analyzed according to the intention to-treat (ITT) principle. Due to major protocol deviations and the excessive missing proportion, the ITT analysis could not be performed for other data.

Data were assessed for normality using the Kolmogorov–Smirnov test. The normally distributed data were presented as mean ± SD, the non-normally distributed data were presented as median [interquartile range (IQR)], the categorical data were presented as number (percentage). Inter group comparison, the normally distributed data were compared using the Student t-test for independent samples, the non-normally distributed data were compared using the Kruskal-Wallis test, the categorical data were compared using the chi-square test. Intra group comparison using the repeated measures ANOVA. p-value of less than 0.05 was considered statistically significant. The Bonferroni correction was applied to adjust the statistical results in multiple comparisons.

For primary outcome, the absolute risk reduction (ARR) and number needed to treat (NNT) were also calculated. For secondary outcomes analysis, the comparisons of sore neck, sore jaw, sore throat, and airway injury were confirmatory, while the comparisons of incidence of SpO₂ ≤ 90%, average duration of assist ventilation, time-weighted average duration of assist ventilation were exploratory.

Between February 2023 and July 2024, there were 1,187 participants scheduled for hysteroscopy in our hospital, in which 417 participants were eligible for this study, excluding 17 participants who declined to participate, finally 400 participants were enrolled and randomized in this study, 395 participants were completed the intervention and entered the statistical analysis as the per protocol set. The CONSORT flow diagram was showed in Figure 1. The details of the five excluded participants were described in Supplementary Table S1.

The demographic data were showed in Table 1. There were not significant differences of demographic data, comorbidities and preoperative hemoglobin between groups. Operational and anesthesia characteristics were showed in Table 2. there also were not significant differences of procedure types, procedure duration, anesthesia duration, and anesthetics dosage between groups.

The vital signs during anesthesia were showed in Figures 2a–c and Supplementary Table S2. The SpO2, mean arterial pressure (MAP), and heart rate (HR) were recorded in three time points: 5 min after participants lied flat in operation room (baseline value), the lowest value during procedure (lowest value), 5 min after participants woke up (recovery value). The baseline and recovery values of SpO₂ were measured during oxygen inhalation through FM at a flow rate of 5 L/min. Intra group comparison, the recovery value of SpO₂ was higher than baseline and lowest values in all the groups, the baseline value of SpO₂ was also higher than lowest value in non-ORI monitoring groups. The baseline and recovery values of MAP and HR were higher than lowest value in all the groups. Inter group comparison, the lowest value of SpO₂ was higher in ORI monitoring groups than in non-ORI monitoring groups. The lowest value of SpO₂ was also higher in non-ORI+NPA group than in non-ORI+FM group. There were not significant differences of baseline and recovery value of SpO₂, all values of MAP and HR between groups.

The primary and secondary outcomes were showed in Table 3. The incidence of low SpO₂ was lowest in ORI monitoring groups, followed by non-ORI+NPA group, it was highest in non-ORI+FM group, there was not significant difference between ORI+FM group and ORI+NPA group. The ARR between ORI monitoring groups and non-ORI monitoring groups was 0.41 (41%) (95%CI: 0.322–0.498), the NTT was 3. The incidence of SpO₂ ≤ 90% was also compared, it was highest in non-ORI+FM group, there were not significant differences between the other three groups. The ARR between ORI monitoring groups and non-ORI monitoring groups was 0.13 (13%) (95%CI: 0.073–0.187), the NTT was 8. The average duration of assist ventilation was shorter in ORI monitoring groups than in non-ORI monitoring groups. Time-weighted average duration of assist ventilation was shortest in ORI+NPA group, followed by ORI+FM and non-ORI+NPA groups, it was longest in non-ORI+FM group. The number of ORI dropped to zero were 2 (1–3) in ORI+FM group and 1.5 (0–3) in ORI+NPA group, there was not significant difference between groups.

There were not severe cardiovascular events, as well as aspiration, LMA insertion, mechanical ventilation, and unplanned admission in our study. The sore throat was more serious in ORI+NPA and non-ORI+NPA groups than in ORI+FM and non-ORI+FM groups. There were not significant differences of sore neck and sore jaw between groups. NPA insertion could induce mild airway injury, the incidence was 5.6% (11/197).

To explore which population could benefit more from ORI monitoring, we performed subgroup analysis of primary outcome according to baseline characters, age ≤ 40 years or > 40 years, with diabetes or not, baseline SpO₂ ≤ 95% or > 95%, procedure duration ≤ 25 min or > 25 min. The results were showed in Table 4. Compare to ORI monitoring groups, the incidences of low SpO₂ were higher in non-ORI+FM group in all subgroup analysis, they were higher in non-ORI+NPA group in age ≤ 40 years, without diabetes, baseline SpO₂ ≤ 95%, and all procedure duration subgroups. ARR and NNT analysis showed that patients under the age of 40, and with longer procedure duration were benefit more from ORI monitoring.