High-flow nasal oxygen versus nasopharyngeal oxygen insufflation during upper aerodigestive tract endoscopy: protocol for a multicentre randomised controlled non-inferiority trial (OPTIGO)

Abstract

Introduction

Upper aerodigestive tract endoscopy is routinely performed for the diagnosis and staging of head and neck cancers. These procedures are commonly conducted under general anaesthesia with spontaneous ventilation to optimise surgical exposure and avoid tracheal intubation. However, maintaining adequate oxygenation remains challenging, particularly in patients with altered airway anatomy due to prior surgery or radiotherapy.

Standard practice often involves a preliminary laryngoscopy to insert a nasopharyngeal oxygen catheter, which may increase procedural complexity and interfere with surgical conditions. High-flow nasal oxygen (HFNO) has emerged as a promising alternative for maintaining oxygenation in various perioperative settings.

The OPTIGO trial aims to determine whether HFNO is non-inferior to nasopharyngeal oxygen insufflation in preventing intraoperative oxygen desaturation during upper aerodigestive tract endoscopy.

Methods and analysis

The OPTIGO study is a multicentre, prospective, randomised, open-label, non-inferiority trial conducted in three French centres. A total of 610 adult patients scheduled for endoscopic procedures of the pharynx, larynx and oesophagus will be randomised in a 1:1 ratio to receive either HFNO or standard oxygenation via a nasopharyngeal oxygen catheter.

The primary outcome is the occurrence of intraoperative oxygen desaturation, defined as a peripheral oxygen saturation ≤85% at any time from induction to the end of the procedure.

Secondary outcomes include intraoperative complications, postoperative laryngeal pain, surgeon comfort, duration of the procedure, anaesthetic drug consumption, and perioperative gas exchange.

Ethics and dissemination

This study has been approved by the independent ethics committee CPP Est 1 (10 April 2025; EudraCT 2025-A00278-41). The trial will be conducted in accordance with the Declaration of Helsinki and Good Clinical Practice Guidelines.

Results will be disseminated through presentations at scientific conferences and publication in peer-reviewed journals.

Trial registration number

CPP Est 1 EudraCT 2025-A00278-41 (Clinical trial NCT07004699).

Strengths and limitations of this study.

• The pragmatic design reflects real-life clinical practice across different centres, enhancing external validity.

• The large sample size and randomised design will provide robust evidence with limited risk of bias.

• The exclusion of patients with body mass index ≥35 kg/m² may limit generalisability to obese populations.

• Capillary rather than arterial blood gas measurements may reduce the precision of gas exchange assessment.

Introduction

Head and neck cancers represent a significant health burden, with approximately 15 000 new cases diagnosed annually in France alone.¹ Endoscopic examination of the pharynx, larynx, and oesophagus (panendoscopy or upper aerodigestive tract endoscopy) is a key diagnostic and staging procedure in these patients.

These procedures are typically performed under general anaesthesia while maintaining spontaneous ventilation.² This approach improves surgical exposure and avoids the need for tracheal intubation, which may be technically challenging in patients with altered airway anatomy due to prior surgery, radiotherapy or tumour-related changes.

However, maintaining adequate oxygenation during these procedures remains a major concern. Current practice often involves an initial laryngoscopy performed by the anaesthesia team to insert a nasopharyngeal oxygen catheter positioned above the vocal cords. While effective, this approach has several drawbacks. It requires an additional airway instrumentation step, may increase the risk of dental or mucosal injury and can interfere with the surgeon’s field of view.³

High-flow nasal oxygen (HFNO) therapy has gained widespread use in recent years across intensive care, emergency medicine and perioperative settings.⁴ HFNO delivers heated and humidified oxygen at high flow rates (up to 70 L/min), allowing improved oxygenation through mechanisms including reduction of dead space, generation of low levels of positive airway pressure and improved matching of oxygen delivery to inspiratory demand.

Recent studies suggest that HFNO may be effective in maintaining oxygenation during upper airway procedures, potentially avoiding the need for invasive airway instrumentation.⁵ Given the relatively short duration and minimally invasive nature of upper aerodigestive tract endoscopy, HFNO could represent a simpler and less invasive alternative to standard oxygenation techniques.

The OPTIGO trial is designed to evaluate whether HFNO is non-inferior to nasopharyngeal oxygen insufflation in preventing intraoperative oxygen desaturation during these procedures.

Methods and analysis

Trial design and setting

The OPTIGO trial is an investigator-initiated, publicly funded, national, multicentre, prospective, randomised, open-label, non-inferiority trial with two parallel groups.

Participants will be randomly assigned in a 1:1 ratio to receive either HFNO therapy (experimental) or standard oxygenation using a nasopharyngeal oxygen catheter (control) (figure 1). Randomisation will be stratified by centre.

Figure 1. Diagram of the OPTIGO study. ENDO-PLO, endoscopy of the pharynx, larynx and oesophagus; HFNO, high-flow nasal oxygen.

The study will be conducted in three French centres, including both university and non-university hospitals, to enhance external validity.

Patient and public involvement

Patients and the public were not involved in the design, conduct, reporting or dissemination plans of this research.

However, the study protocol was reviewed and approved by an independent ethics committee (CPP Est 1 EudraCT 2025-A00278-41), which includes patient representatives. These representatives contributed to assessing the relevance of the research question, the acceptability of the study procedures and the clarity of the patient information and consent documents .

All participants will receive detailed oral and written information (see onlinesupplemental files S1 S2) prior to inclusion, and their burden related to study participation has been carefully considered. At the end of the study, participants will have the opportunity to access the overall results on request(figure 2).

Figure 2. Schedule of enrolment, interventions and assessments. ENDO-PLO, endoscopy of the pharynx, larynx and oesophagus; NRS, Numerical Rating Scale; PCo2, partial pressure of C02 in arterial; ScO2, Estimated oxygen saturation in the capillaries.

Eligibility criteria

Eligible patients will be screened during the pre-anaesthesia consultation and will receive both oral and written information. A minimum reflection period of 48 hours will be ensured before obtaining written informed consent.

Inclusion criteria

• Adult patients (≥18 years).

• Scheduled for upper aerodigestive tract endoscopy (pharynx, larynx, oesophagus).

• ASA physical status I to IV.

• No requirement for preoperative supplemental oxygen

• Patients benefiting from a Social Security system or benefiting from it through a third party.

• Free subject, without guardianship or curatorship or subordination.

• Having given their informed consent.

Exclusion criteria

• Inability to understand the study or provide informed consent.

• Combined surgical procedure requiring tracheal intubation

• Body mass index (BMI) ≥35 kg/m².

• Severe chronic obstructive pulmonary disease (GOLD stage IV).

• Known allergy to one of the anaesthetic agents (remifentanil, propofol, lidocaine).

• Persons benefiting from enhanced protection, namely minors, pregnant or breastfeeding women, persons deprived of their liberty by a judicial or administrative decision, adults under legal protection.

• Baseline peripheral oxygen saturation (SpO₂) ≤92% on arrival in the operating room prior to randomisation.

Assignment of intervention

Randomisation

The randomisation list will be generated by an independent statistician using a computer-based random-number generator and implemented through a secure web-based system.

The patients will be randomly allocated to receive either HFNO therapy or standard oxygenation using a nasopharyngeal oxygen catheter in a 1:1 ratio, using a dedicated, password-protected website. Given the nature of the interventions, blinding of clinicians is not feasible.

Interventions

General anaesthesia management

All patients will receive standardised anaesthesia using target-controlled intravenous anaesthesia (TIVA) of remifentanil and propofol. Anaesthetic depth will be adjusted to maintain spontaneous ventilation throughout the procedure.

Standard monitoring will include electrocardiography, non-invasive blood pressure, pulse oximetry, and respiratory rate. Preoxygenation will be considered adequate when the fraction of expired oxygen exceeds 90%. The depth of anaesthesia will not be monitored directly by a specific sensor such as the bispectral index (BIS).

Only the patient’s respiratory rate and ability to tolerate ventilation and surgery will guide the administration of the desired intraoperative targets using the TIVA model.

In case of loss of spontaneous ventilation or severe desaturation, airway management will be performed according to standard clinical practice, including mask ventilation or tracheal intubation if required.

Experimental group: HFNO

• Preoxygenation using HFNO at 20 L/min and fraction of inspired oxygen (FiO₂) of 100%

• During the procedure: HFNO at 50 L/min and FiO₂ 100%

• Oxygen delivered via heated and humidified high-flow nasal cannula

Control group: standard oxygenation

• Preoxygenation using a face mask at 15 L/min and FiO₂ 100%

• Direct laryngoscopy performed to introduce a nasopharyngeal oxygen catheter positioned above the vocal cords

• During the procedure: oxygen delivered at 8 L/min and FiO₂ 100%

Surgical procedure

Suspension laryngoscopy and biopsies will be performed according to standard practice in both groups. No electrocoagulation or laser procedures will be performed during the study.

Topical anaesthesia of the airway will be achieved using 2% lidocaine administered via a multiperforated catheter.

Outcomes

The primary outcome is the occurrence of intraoperative oxygen desaturation, defined as at least one episode of SpO₂ ≤85%, measured continuously from induction of anaesthesia to the end of the procedure during surgery.

The secondary outcomes are:

• Intraoperative complications (eg, bleeding, dental injury).

• Postoperative laryngeal pain assessed using a 0–10 Numerical Rating Scale (NRS), with 0 anchored to “no pain” and 10 anchored to “worst pain imaginable”.

• Surgeon comfort assessed with a 0–10 NRS, with 0 anchored to “very uncomfortable” and 10 to “very comfortable”.

• Duration of the procedure (from the suspension laryngoscopy to the end of procedure).

• Duration of anaesthesia (from preoxygenation to awakening).

• Total dose of anaesthetic agents (remifentanil and propofol), standardised by body weight.

• Per- and postoperative gas exchange (partial pressure of oxygen in arterial, PaCO₂, pH, Estimated oxygen saturation in the capillaries). Intraoperative blood gas measurements are taken exactly 2 min after the start of the procedure. From the moment the surgeon exposes the larynx with a rigid laryngoscope, a stopwatch is started and blood gas samples are taken. For postoperative blood gas tests, we wait until the patient is fully awake, and these are taken after all forms of oxygenation have been stopped. At least 2 min after the patient has woken up and after all forms of oxygen have been stopped, the postoperative blood gas is taken).

Statistics

Sample size

A non-inferiority margin of 7% was selected based on clinical judgement, reflecting the maximum acceptable difference that would not compromise patient safety. Considering that 13.1% of patients⁵ will present a desaturation episode (SpO₂ ≤85% during surgery), with an absolute non-inferiority margin set to 7%, a one-sided type I error of 5%, a power of 80% and a 5% rate of missing data (missing outcomes, consent withdrawal, post-randomisation exclusions, etc.), a total sample size of 610 patients (305 per group) is required.

Flowchart

A flowchart will be decomposed according to the following sets: the screened set (patients having signed the informed consent), the intention-to-treat (ITT) set (patients who were randomised), the full analysis (FA) set (previous ITT population excluding participants who failed to meet major entry criteria⁶ and participants with no data post-randomisation or with missing primary outcome), and the per-protocol set (previous FA population, excluding participants who did not receive the allocated intervention). The effects and the related differences between two consecutive sets will be reported. The causes of the related differences will also be listed. The withdrawals will be exhaustively described in the flowchart.

All the analyses will be performed based on the per-protocol set (main analysis). According to the non-inferiority design, analyses will also be performed on the FA set (secondary analysis).

Statistical analysis

Baseline characteristics and outcomes will be summarised in each group and marginally. The number of missing values will be presented—no imputation will be performed. The means and the SD will be presented for continuous covariates. The effects and the percentages of patients in each category will be presented for categorical covariates. The balance diagnostics of baseline characteristics between the two groups will be achieved by estimating standardised differences.⁷ A two-proportions z-test will be performed to compare the primary binary outcome. Pearson’s Chi-2 test or Fisher’s exact test will be performed to compare other binary outcomes, and Student’s t-test will be performed to compare continuous outcomes. Bilateral p-values <5% will be considered statistically significant for outcomes, except for the primary outcome, where statistical significance will be achieved for p value<5% by using one-sided alternative hypothesis.

We plan to adjust on stratification factors (recruitment centre) and baseline risk factors of the outcome in order to limit bias and to increase power.^{8 9} To avoid the exclusion of patients with missing values on the baseline predictors, multivariate imputation by chained equations will be performed five times. For each complete sample, 1000 bootstrap samples will be generated. We will use G-computation to estimate the average treatment effect (ATE) for each bootstrap sample. First, we will estimate the outcome models. No interaction between the randomisation group and the predictors will be tested. The elastic net regularisation¹⁰ will be used to select the predictors to consider for adjustment. The penalty parameters will be estimated by minimising the 20-fold cross-validation mean square errors. Binary (respectively continuous) outcomes will be modelled with logistic (respectively linear) regressions. Second, based on the outcome models, we will predict the outcomes if all participants were assigned to the experimental group (HFNO) and then if all participants were assigned to the control group (oxygen cannula). ATE will be defined as the difference of the obtained proportions or means for the two adjusted predictions. CIs of the ATE estimations will be non-parametrically estimated from the 5000 imputed bootstrap samples.¹¹

Ethics and dissemination

This research has been approved by the independent ethics committee CPP Est 1 on 10 April 2025 (approval number EudraCT 2025-A00278-41) and will be carried out according to the principles of the Declaration of Helsinki and Good Clinical Practice Guidelines. The results of this study will be disseminated through presentation at scientific conferences and publication in peer-reviewed journals.

Data availability statement

The study dataset is the property of the sponsor (Centre Hospitalier Universitaire de Poitiers, France) and will be made available to the steering committee and participating investigators. In agreement with European and French law, pseudonymised data cannot be open.¹² For data sharing, a request can be made to the principal investigator (corresponding author). Data transfer agreement can be proposed after the agreement of the steering committee of the study and the completion of collaborative forms with the sponsor. Trial monitoring will take place every 40 patients, and will be carried out by a clinical research associate dedicated to the study but not involved in its design.

Discussion

This multicentre randomised controlled trial will be the first to compare HFNO with nasopharyngeal oxygen insufflation for maintaining oxygenation during upper aerodigestive tract endoscopy performed under general anaesthesia with spontaneous ventilation.

Maintaining adequate oxygenation during these procedures remains a key challenge. Current practice often relies on the insertion of a nasopharyngeal oxygen catheter under direct laryngoscopy, which requires an additional airway manipulation and may interfere with surgical exposure. HFNO represents a less invasive alternative that could simplify airway management while preserving adequate oxygenation.

Previous studies have demonstrated the efficacy of HFNO in various perioperative settings, including upper airway procedures, although comparisons have mainly been made with face mask ventilation rather than with nasopharyngeal oxygen insufflation.⁵ Therefore, the present trial addresses an important gap in the literature by directly comparing these two strategies in a randomised design.

If HFNO is shown to be non-inferior to standard oxygenation, this could have several clinical implications. First, it could reduce the need for preliminary laryngoscopy performed solely for oxygenation purposes, thereby limiting airway instrumentation and its associated risks, such as dental or mucosal injury. Second, it may improve surgical conditions by avoiding the presence of a nasopharyngeal catheter in the operative field. Finally, HFNO could streamline perioperative workflows by simplifying anaesthetic management.

The strengths of this study include its randomised controlled design, multicentre setting, and large sample size, which together enhance internal and external validity. In addition, the pragmatic design reflects routine clinical practices, increasing the applicability of the results.

However, several limitations should be acknowledged. First, the open-label design may introduce performance bias, although the primary outcome is an objective and continuously monitored parameter. Second, the exclusion of patients with BMI ≥35 kg/m² may limit the generalisability of the findings to obese populations, who are at higher risk of perioperative hypoxaemia. Third, the use of capillary rather than arterial blood gas measurements may reduce the accuracy of gas exchange assessment.

In conclusion, the OPTIGO trial will provide robust evidence on the effectiveness of HFNO compared with standard oxygenation during upper aerodigestive tract endoscopy. The results may contribute to optimising airway management strategies in this setting and improving both patient safety and procedural conditions.

Trial status

The current protocol is Version 4 dated on 10 April 2025. The trial is currently recruiting patients. The inclusion started on 11 July 2025 and the number of patients included to date (14 January 2026) is 37. The estimated length of inclusion is 24 months.