Introduction

Obstructive sleep apnea (OSA) is a disorder characterized by repetitive collapse of the upper airway during sleep. OSA is the most common sleep-related breathing disorder in North America [1]. Standard treatment for OSA is continuous positive airway pressure therapy (CPAP). However, up to 50% of patients with OSA are unable to tolerate CPAP obligating alternative treatments such as surgery, upper airway stimulators, or oral appliances [2•].

Because upper airway obstruction can occur at multiple levels, successful targeted surgical management of OSA requires precise localization of the culprit area [3]. Since the late 1970s, researchers have leveraged endoscopy as a valuable tool to directly visualize the upper airway and identify anatomical segments of obstruction. There are at least 10 diagnostic modalities, of which lateral cephalometry, Muller’s maneuver, and drug-induced sleep endoscopy (DISE) are most popular [4]. However, DISE has been the most used technique to identify specific areas of obstruction and patterns of airway collapse that serves as a baseline screening tool for both surgical and non-surgical options [4].

There is a wide array of agents available for sedating patients undergoing DISE. However, the patient must be sedated to a level mimicking natural sleep. The most often utilized agents are propofol, midazolam, and dexmedetomidine [3]. A more detailed review of these agents will be addressed in a later section.

The protocols for DISE differ greatly between institutions, in large part due to the heterogeneity of available options. Furthermore, much of the current literature on this topic exists in otolaryngology and surgery journals instead of anesthesiology publications. In this paper, we provide a standardized DISE protocol centered around and geared toward the anesthesiologist.

Indications

Indications for DISE include patients with OSA, those patients that are unable to tolerate CPAP (a population thought to be as high as 50%), or those being considered for surgery, oral appliances, or positional therapy.

Contraindications

While no absolute contraindications exist unique to DISE, relative contraindications include drug allergy to selected sedative agents, pregnancy, and medical comorbidities resulting in significant airway compromise.

Complications

No catastrophic events have been reported in patients undergoing DISE; endotracheal intubation is extremely rare and the need for emergent surgical airway interventions have not been reported [2•]. Therefore, DISE is a relatively safe procedure.

Technique

Pre-Procedure Preparation, Setting, and Monitoring

DISE is usually performed in the operating room or procedure suite, but can also be performed in various outpatient settings, depending on the availability of personnel and appropriate equipment for safely administering sedation. Standard ASA monitoring is used. While the odds of endotracheal intubation are low, equipment to definitively secure the airway as well as trained personnel should be readily available in case endotracheal intubation is needed. Supplemental oxygen may not be necessary but should also be readily available. In the preoperative setting, anticholinergic agents, such as glycopyrrolate, can be administered intravenously up to 30 minutes before the procedure to reduce secretions and improve visualization [5]. Topical vasoconstrictors such as oxymetazoline as well as local anesthetics such as lidocaine can be administered intranasally to facilitate passage of the endoscope and minimize risk of bleeding [6, 7].

Depth of Sedation and Monitoring

The depth of sedation during DISE plays a crucial role in accurately identifying the anatomical segment(s) obstructing the upper airway. Evaluation with lighter sedation may underestimate the obstruction, whereas over sedation may overestimate obstruction. A deep level of sedation (as defined by the ASA) is required to facilitate the procedure, in which the patient is purposefully responsive to only repeated or painful stimuli [8]. During the assessment, the proceduralist observes 2-3 repeat cycles of snoring, hypoxia, obstruction with apnea, and breakthrough snoring. While the most important gauge of sedation depth is clinical, Bispectral index score (BIS) may provide additional monitoring of sedation. A potential target score of 61-70 will ensure a targeted level of sedation and optimize muscle tone and upper airway collapsibility [9]. While BIS systems provide an added level of monitoring for the anesthesiologist, we forgo this measurement at our institution as clinically gauging sedation has yielded sufficiently reliable results.

It should also be noted that many European institutions implement DISE using targeted control infusion (TCI) systems [10•]. These systems allow for precise control of plasma levels of various drugs to provide a predictable and stable level of plasma drug concentration and thereby effect. However, these systems are not approved for use in the United States and will not be further discussed in this paper.

Airway Evaluation

A variety of classification schemes have been described to characterize and standardize DISE findings. The Velum Oropharynx Tongue Base Epiglottis (VOTE) Classification was proposed as a standard for DISE scoring because it incorporates the 4 major structures that most often contribute to airway obstruction. They are the Velum (palate), Oropharyngeal lateral walls (including the tonsils), Tongue, and Epiglottis [2•]. This system incorporates the degree and configuration of airway narrowing related to these structures that are each composed of multiple components [2•]. Furthermore, this evaluation can be performed in conjunction with various maneuvers and position changes that garner additional information.

Drugs for Sedation: An Overview

The most common sedative agents used for DISE procedures are propofol, midazolam, and dexmedetomidine (Table 1).

Table 1 A comparison of major aspects and effects of common sedative agents used in DISE [2•, 10•, 11] (Adapted from Butterworth JF, Mackey DC, Wasnick JD. Morgan and Mikhail's clinical anesthesiology cases. New York: McGraw Hill Lange; 2020. 171-85 p.)

Propofol

Although propofol’s exact mechanism of action for sedation is unknown, it acts as a global Central Nervous System (CNS) depressant that directly activates gamma aminobutyric acid-A (GABA-A) receptors [3, 11]. Useful in performing DISE, it is a fast-acting drug with a large volume of redistribution thereby allowing a rapid recovery after an infusion [3]. Importantly for DISE, propofol sedation acts through different mechanisms than natural sleep. However, sedation with propofol importantly resulted in the presence of slow waves that are found in natural NREM sleep, the stage where obstruction most commonly occurs [3].

It is important to recognize that propofol also has dose-dependent central inhibitory effects on respiratory muscles. In one study, Mahmoud et al. found statistically significant differences between low and high doses of propofol with differences in the sites and degrees of obstruction [12]. However, these differences were not sufficiently important to influence the measured upper airway dimension or decision making regarding site-specific interventions [12]. Furthermore, Yoon et al. showed that there was excellent agreement and correlation regarding upper airway collapse for all obstruction sites regardless of depth of sedation in patients sedated with propofol versus dexmedetomidine during DISE [13]. Accordingly, while propofol and drug induced sedation, in general, will artificially decrease upper airway tone and induce obstruction, propofol can still be used reliably in the clinical setting.

Midazolam

Midazolam is a GABA-A receptor agonist that produces sedative, hypnotic, anxiolytic, anticonvulsant, and muscle relaxant effects [14]. Like propofol, it acts rapidly, and its effect dissipates quickly. Additionally, there is the potential for respiratory depression and decreased upper airway tonicity. One study noted similar rates of airway collapse compared to propofol in all areas of the upper area except the velum [15, 16]. However, it has also been described to induce higher pharyngeal muscle tone relaxation compared to propofol, highlighting the need for further investigation [15]. When compared to dexmedetomidine, researchers found higher incidences of tongue base collapse with midazolam. Regarding its effects on sleep, like propofol, it also preferentially places patients into an OSA-prone Non-Rapid-Eye-Movement (NREM) stage 1 and 2, making it an appropriate anesthetic agent for DISE [17].

Co-Induction with Midazolam and Propofol

Some groups use midazolam as an induction agent followed by a propofol infusion for sedation maintenance [3]. While there are possible benefits to this method including added anxiolysis and the noted synergy between propofol and midazolam, it remains unclear as to whether the use of midazolam reduces the required propofol dosage or prolongs recovery times [3]. Therefore, further research and standardization of this drug combination is needed.

Dexmedetomidine

Dexmedetomidine is a selective alpha-2 adrenoreceptor agonist. Notably, unlike propofol and midazolam, its sedative and analgesic effects have almost no effect on the respiratory drive [18]. It has also been suggested that dexmedetomidine does not affect the tonicity of the upper airway, or at least less than midazolam and propofol. In fact, one study found dexmedetomidine did not induce upper airway muscle collapsibility significantly enough to cause an increase in the apnea-hypopnea index [19]. The effect of sedation on the tonicity of the upper airway is an area that requires more research.

Dexmedetomidine does, however, have its drawbacks. First, it is a slower-acting medication that takes longer to induce a target level of sedation. While propofol acts within 15-30 seconds, and midazolam < 60 seconds, dexmedetomidine’s onset of action ranges from 5-10 minutes and includes a longer period of recovery [6].

There is also some evidence that dexmedetomidine is not reliable when used as the sole anesthetic for DISE. In one study, half of the patients receiving dexmedetomidine for DISE required supplemental propofol despite being at a maximum dose of 1.5 mcg/kg/h [20]. Considering its pharmacokinetic profile and limited efficacy, dexmedetomidine is not the agent of choice at our institution.

Protocols

Currently Existing Protocols

DISE originated in Europe, and use TCI systems, for which no such conversion exists to implement such a protocol in the United States. However, several protocols that do not use TCI have been proposed. They use propofol, midazolam, or dexmedetomidine. Previously published propofol bolus regimens utilize a loading dose of 30–50 mg with subsequent 10 mg boluses every 2 min [10•] or a 1 mg/kg load followed by 20 mg every 2 min [10•]. Alternatively, propofol can be run at an infusion of 100-150 mcg/kg/min supplanted by 20-50 mg boluses as needed [10•]. Midazolam can be bolused at 0.03 mg/kg and again after 2-5 minutes, followed by 0.015 mg/kg after 5 min as needed [10•]. Dexmedetomidine can also be loaded as an infusion of 1.5 mcg/kg over 10 minutes followed by a maintenance infusion of 1.5 mcg/kg/hr [10•].

Our Protocol

Our institution has performed more than 600 DISE procedures over the past six years using the protocol below (Fig. 1). It is performed in an ambulatory surgical setting and has not resulted in any adverse patient outcomes.

Fig. 1
figure 1

Our proposed protocol for DISE based on an initial propofol bolus tansitioned into an infusion. Non-Invasive Blood Pressure (NIBP), Saturation of peripheral Oxygen (SpO2), End-Tidal Carbon Dioxide (ETCO2)

Preoperative medications

0.2 mg glycopyrrolate IV 15-30 minutes before start of procedure.

Monitoring

Non-Invasive Blood pressure (NIBP), pulse oximetry, electrocardiogram (ECG), End-Tidal Carbon Dioxide (ETCO2) via nasal cannula

Intraoperative Procedure

  • When prompted by the surgeon, begin sedation regimen as follows:

    • 0.5-0.6 mg/kg propofol IV bolus (max dose 60 mg)

    • Begin propofol infusion at 140-160 mcg/kg/min

    • Increase infusion every 2-3 minutes as directed by surgeon until proper level of sedation has been achieved as described below

  • Observe patient reaction to passage of distal chip endoscope through the nasopharynx and be ready to titrate sedation as needed or directed. The patient should be responsive to repeated physical stimulus. The surgeon will also monitor for appropriate sedation based clinically on the degree of upper airway collapsibility and patient comfort.

  • Surgeon will perform their upper airway examination. Be mindful of the patient’s positioning during manipulation of the patient’s head and neck as well as supine versus lateral lie.

  • Stop propofol infusion at the direction of the surgeon.

Discussion

DISE—initially called “sleep nasendoscopy”—was developed in 1990 in London and has since been used to provide a snapshot of the various anatomical segments of the oropharyngeal airway in an unconscious person. DISE has proven to be a reliable tool to aid Eye-Nose-Throat (ENT) physicians to guide treatment options in patients with OSA [6, 21, 22]. However, additional investigation is needed to standardize DISE techniques, training, and interpretation. Additionally, a standardized sedation protocol has not yet been described [16].

Among the myriad of drugs used to induce sedation, we have reviewed the most used agents: propofol, midazolam, and dexmedetomidine. Midazolam is rapid-acting but may influence upper airway tone and predispose to artificial collapse. Dexmedetomidine takes longer to induce sedation in patients but seems to spare upper airway tone more than propofol or midazolam and does not disturb natural sleep architecture. Considering this, propofol may be a desirable alternative that has a fast onset/offset and provides enough sedation without artificially collapsing the upper airway to a large degree.

Conclusions

This paper serves as a relevant synopsis of DISE for the anesthesiology clinicians and proposes a protocol for sedation during DISE that we have found to be safe at our institution. From 2018-2023, our group has performed over 600 cases utilizing this technique with favorable results, no serious adverse events, and only one episode of post-operative nausea and vomiting. In short, propofol’s rapid onset and elimination coupled with providing deep sedation makes it an ideal candidate for DISE. However, future studies are needed to better understand its ability to artificially decrease upper airway tone and induce obstruction.