Drug-Induced Sleep Endoscopy in adults with Sleep Disordered Breathing: Technique and the VOTE Classification System

Abstract

Drug-induced sleep endoscopy (DISE) offers an unique evaluation of the upper airway. After pharmacologic induction of unconscious sedation, fiberoptic endoscopic examination of the pharynx can characterize the pattern of upper airway obstruction. The authors describe DISE techniques and the VOTE Classification system for reporting DISE findings. The VOTE Classification focuses on the primary structures that contribute to upper airway obstruction and represents a common language to describe the patterns of obstruction during DISE. The latter can facilitate the scientific evaluation of DISE, including its role in directing treatment.

INTRODUCTION

Assessment of the pattern of upper airway obstruction is critical to successful surgical treatment of snoring and obstructive sleep apnea (OSA). A variety of diagnostic techniques have been developed to examine pattern of obstruction; each has important strengths and weaknesses. Traditionally, OSA surgical evaluation techniques are performed during wakefulness and primarily involve static observations rather than the dynamic assessments during sleep. As such, they may not be ideal methods to assess the upper airway during breathing while asleep. The variety and complexity of vibrations and collapse events in the upper airway during sleep depend on multiple factors including sleep stage, muscle tone, body position, head and neck position, and variations in lung volumes.

Drug-induced sleep endoscopy (DISE) has been performed for decades in numerous leading centers. Originally described by Croft and Pringle as “sleep nasendoscopy” in 1991, the evaluation requires pharmacologic induction of sedation and flexible fiberoptic endoscopy to visualize upper airway obstruction and/or snoring.¹

Hohenhorst and Kezirian renamed the procedure DISE to reflect the three key features of this method of assessment: (1) the use of various pharmacologic agents to achieve unconscious sedation; (2) the goal of reproducing upper airway physiology similar to that occurring during natural sleep; and (3) endoscopic upper airway evaluation. Other terms referenced in the literature include sleep endoscopy, sleep nasendoscopy, somnoendoscopy, somnoscopy, sedated endoscopy, and propofol sleep endoscopy.

DISE has been a major advance in the personalized medicine approach to treatment of OSA in individuals who do not tolerate positive airway pressure therapy. DISE has gained widespread adoption, with ongoing research continually advancing its diagnostic technique and clinical applications. Multiple studies have examined its safety, feasibility, validity, and reliability as a diagnostic tool.^2–6

Ultimately, the role of DISE as an evaluation technique depends on whether it can identify patients who will be more or less likely to benefit from specific interventions. Studies have examined the association between DISE findings and outcomes of upper airway surgery,^7–11 hypoglossal nerve stimulation,^12–14 and mandibular repositioning appliances (MRA).^15–18 Most of these studies are from single institutions, but the more-recent, larger international multicenter cohort studies have identified preoperative DISE findings associated with outcomes of surgical interventions. For upper airway surgery, the following findings were each associated with a lesser odds of surgical response: partial or complete obstruction related to the oropharyngeal lateral walls and complete tongue-related obstruction.⁸ For unilateral hypoglossal nerve stimulation, complete oropharyngeal lateral wall-related obstruction was associated with a lesser odds of response, but complete tongue-related obstruction was associated with a greater odds of response.¹²

Innovative work is examining the importance of DISE findings during administration of positive airway pressure¹⁹ and therapeutic interventions, including nerve stimulation.^20,21

INDICATIONS

DISE is designed to improve OSA treatment in patients who do not tolerate positive airway pressure therapy, especially when surgery or MRA therapy is being considered as a potential treatment modality. DISE may specifically be required to determine candidacy for unilateral hypoglossal nerve stimulation, based on exclusion of those with complete concentric collapse at the level of the velum (palate).²²

Like any test, DISE should be used if it will change the treatment plan. This is less likely if other evaluations have suggested treatment with positive airway pressure, weight loss, or positional therapies, although DISE could inform recommendations regarding these approaches. For example, DISE can includes changes in body position that may be associated with the pattern of obstruction,²³ and weight loss may specifically improve airway obstruction related to the oropharyngeal lateral walls or tongue.²⁴ Conversely, those with markedly-enlarged tonsils (3+ or 4+ on the Brodsky scale)²⁵ will likely undergo tonsillectomy as part of their surgical treatment, with DISE probably less useful to treatment selection; DISE would be expected to show that the tonsils are not contributing to upper airway obstruction, while these markedly-enlarged tonsils will limit evaluation of other potential structures contributing to obstruction, simply because the tonsils will occupy a substantial portion of the airway.

Contraindications to DISE include a high American Society of Anesthesiologists score (ASA) (≥3), in whom anesthesia and upper airway surgery would be associated with increased risks;²⁶ allergies to sedative agents, such as propofol or midazolam; and marked elevation in body mass index (that can make surgery or MRAs less likely to achieve favorable outcomes).

TECHNIQUE

DISE is best performed in a controlled environment, with basic cardiorespiratory monitoring (pulse oximetry, blood pressure, electrocardiogram) and the ability to administer supplemental oxygen, if needed.

Patients should remain nil per os (NPO) before the DISE to prevent regurgitation and reduce aspiration risk. In order to reduce salivation and improve visualization during endoscopy, atropine or other anti-cholinergic agents can be administered 30 minutes before starting the procedure; however, caution is advised in patients with a history of cardiovascular disease or difficulty with urination.

We believe that topical anesthetic (1% lidocaine) with or without decongestant may be administered to the more-open nasal cavity prior to the procedure. Although it is possible to block reflex pharyngeal muscle activation with topical nasal anesthesia, the doses required to block those reflexes are much higher than one would use in a standard DISE exam (one study required 4% aerosolized lidocaine to block almost all nasal and pharyngeal sensation, including an absence of any gag reflex).²⁷

Body position is the subject of some debate. Historically, DISE has been performed with the patient should in a supine position on an operating table or bed. With more awareness of the impact of body position on OSA and the pattern of obstruction, another approach is to examine the patient in other body positions, such as the lateral decubitus position, whether in addition to the supine position or in lieu of it. With the assistance of personnel normally present during a DISE, it is possible to move the patient’s body position safely. In some cases, using a supine body position with turning the head to the side may reveal some of the same changes in DISE findings as movement of the entire body to the lateral body position.²⁸ Regardless of body position, the patient should be placed in conditions similar to those during sleep at home (e.g., use and placement of one or two pillows, with or without dentures in place). It can be helpful for the entire team to view the flexible endoscopic examination, and recording (ideally with audio recording as well) for later review can be helpful. The lights should be dimmed and the room quiet to minimize wake-promoting stimuli.

Anesthetic depth is one of the most important aspects of obtaining a valuable DISE exam. The target depth of sedation is the transition from consciousness to unconsciousness (defined by loss of response to verbal stimulation). Because individuals have differential susceptibilities to propofol, the required dosage can vary widely. Slow stepwise induction of sedation is required to avoid oversedation. Deeper levels of sedation are associated with progressive decreases in upper airway dilator muscle tone and neuromuscular reflex activation that both increase airway collapsibility.²⁹ The transition to unconscious sedation achieved with propofol may be a close approximation to natural sleep, based on previous research showing that the transition to unconsciousness is associated with changes in upper airway collapsibility (passive critical closing pressure), bispectral index score (BIS, based on frontal EEG activity), and genioglossus muscle tone. The most-detailed study of propofol unconscious sedation showed that normal individuals have decreases in genioglossus tone to 10% of maximum awake activity,³⁰ which is one-half to one-third of the level during natural NREM sleep in normals but greater than during REM sleep in normals and those with OSA.¹⁷ While unconscious sedation under propofol may not a perfect simulation of natural sleep, pharyngeal dilator muscle activity appears to lie somewhere between non-REM and REM sleep.

Commonly-used sedative agents include propofol, midazolam, or a combination of both medications (often starting with a low loading dose of midazolam (3–4 mg), followed by propofol). The pharmacological properties of both drugs are shown in Table 1. If midazolam is used, a flumazenil injection should be readily available in case reversal of the benzodiazepine effect (excessive sedation or respiratory compromise) should be necessary. Dexmedetomidine has been proposed and used in certain centers because it induces EEG changes similar to natural sleep,³¹ but our experience has been that this medication does not produce pharyngeal physiologic changes consistent with sleep, making it a poor choice for DISE.

Table 1:

Pharmacological properties of midazolam and propofol.

Midazolam	Propofol
Benzodiazepine derivative	2–6 Disopropylphenol
Large therapeutic range	Small therapeutic range
Active Metabolites (accumulation)	No accumulation
Functional half life 45 min.	Functional half life 4–6 min.
Elimination half life 150 min.	Elimination half life 55 min.
Central muscle relaxation	Central breathing depression
Paradoxical reaction in 1%	Hypopharyngeal reflex depression

Propofol may be administered in a number of ways. Target-controlled infusion is an algorithm-based method of propofol administration (relying on age, gender, height, weight, and other parameters) that achieves a predictable propofol serum concentration more rapidly. Target-controlled infusion is not available in the United States and many countries. Alternatives methods of administration include continuous infusion, boluses, or a combination (often a bolus followed by continuous infusion). Greater doses are generally required for patients with alcohol or recreational drug use, use of sedatives, various mental health diagnoses, or chronic insomnia.

Regardless of the choice of sedative agent, slow titration to the target depth of sedation - without oversedation – is critical. BIS (target range typically 50 to 70)³⁰ or similar monitoring is not required but can assist with determination of the depth of sedation, although it is important to note that clinical signs of adequate sedation (e.g., snoring, oxygen desaturation, and airway obstruction) will generally occur before the scores on monitoring systems reflect the clinical changes.

Once an adequate depth of sedation has been achieved, the flexible endoscope is introduced into the nasal cavity. The nasal cavity, pharynx, and larynx are visualized. The examination generally involves assessment of the pattern of obstruction, encompassing multiple cycles of what is often progressive airway obstruction, followed by arousal with airway opening, similar to natural sleep. Assessment should include evaluation of obstruction throughout the entire pharynx, typically requiring movement of the endoscope to enable assessment and the level of the velum and the oropharynx/hypopharynx (often called the hypopharyngeal, retroglossal, or retrolingual region in the sleep surgery literature). The assessment is often performed for at least 3 minutes but may be longer if there is variation in the pattern of obstruction. The assessment can be repeated with changes in body position.

During the DISE, maneuvers such as a chin lift (manual closure of the mouth to bring the teeth into baseline occlusion) (Figure 1) or a jaw thrust (Esmarch Maneuver) (Figure 2) can be performed, with reassessment of the airway during each maneuver. The jaw thrust is a gentle advancement of the mandible by approximately 5 mm, mimicking the effect of an MRA. It has been shown that the degree of opening of the pharyngeal airway during jaw thrust is associated with MRA outcomes.¹⁸ There are devices that allow variable degrees of mandibular advancement, and some of these have been used during DISE to inform MRA settings;³² their utility is not entirely clear at this time.

Figure 1 -.

Chin lift, a manual closure of the mouth

Figure 2 –

Jaw thrust, or Esmarch maneuver, a gentle advancement of the mandible by up to approximately 5mm.

Positive airway pressure administration during DISE has been performed for decades in isolated cases with unexplained incomplete resolution of OSA with positive airway pressure. More recently, it has been advanced as a potential additional assessment that could be performed routinely. Positive airway pressure would generally be administered following baseline evaluation and other maneuvers, involving a nasal mask and a bronchoscopy adapter introduced into the circuit in order to permit fiberoptic endoscopy. Various techniques have been described, with a straightforward approach described in the literature.³³ Findings of the so-called DISE-PAP assessment may be associated with outcomes of unilateral hypoglossal nerve stimulation³⁴ and can also characterize anatomic reasons for incomplete efficacy of positive airway pressure therapy. Ongoing research will determine optimal techniques and its role in assessment.

DISE may also enable evaluation of the potential impact of other therapeutic interventions. Currently limited to research settings, it has proven possible to examine visual and physiologic changes in the airway during neurostimulation, both of the hypoglossal nerve²¹ and ansa cervicalis.²⁰

VOTE CLASSIFICATION

The VOTE Classification (updated version in Table 2) was developed to characterize the pharyngeal structures that primarily contribute to airway obstruction in OSA, either alone or in combination: velum, oropharyngeal lateral walls (including the tonsils), tongue, and epiglottis.³⁵ Prior to its introduction, there were various classification systems, ranging from overly simple to complex, that may have ignored structures, included non-pharyngeal structures, or only allowed for specific combinations.^36–38 For example, region-based classifications, including that proposed by Fujita to include palatal and hypopharyngeal obstruction, ignored differences in obstruction pattern related to various hypopharyngeal structures, such as the oropharyngeal lateral walls, tongue, and epiglottis.

Table 2:

The VOTE classification

	DIRECTION
LEVEL	A-P	LATERAL	CONCENTRIC
Velum
Oropharyngeal Lateral Walls
Tongue
Epiglottis

Degree of obstruction: 0: no obstruction (no vibration, <50%); 1: partial obstruction (vibration 50–75%); 2: complete obstruction (collapse, >75%); x: not visualized

Since its adoption, the VOTE Classification has become the most-widely-used system for describing the pattern of obstruction in adults. The VOTE Classification may be an oversimplification that overlooks some interactions between structures, but it is a foundation for further study of pharyngeal obstruction in OSA and for assessment of the response of upper airway structures to interventions. Because DISE is inherently a qualitative, not quantitative, assessment of vibration and obstruction events, the qualitative assessment of degree of obstruction (none, partial, complete) was preferred, as opposed to problematic evaluation of exact percentages of obstruction related to these structures. DISE is not designed to determine OSA severity (this is based on a diagnostic sleep study of natural sleep). Objective measurement of pharyngeal physiology can be performed with DISE or natural sleep, often with administration of positive airway pressure.

The VOTE Classification has facilitated the scientific evaluation of DISE and communication across multiple institutions and studies. Our experience suggests that a focus on structures enables examination of two central questions: treatment selection and the association between DISE findings and treatment outcomes—for surgery, mandibular repositioning appliances, or combined therapy. In contrast to most surgical evaluation techniques, DISE not only uniquely offers a dynamic evaluation of the upper airway during conditions that ideally mimic natural sleep but also enables visualization of specific structures that contribute to upper airway obstruction. This structure-based characterization is the foundation of DISE and the VOTE Classification.

The Structures of the VOTE Acronym

Our experience with DISE suggests that a selected group of structures contribute to pharyngeal upper airway narrowing and/or obstruction in sleep disordered breathing, individually or in combination. The VOTE Classification evaluates the potential contribution of each of these structures and both the degree and configuration of airway narrowing.

Velum (Figure 3 a–d).

Figure 3. Velum-related Obstruction.

a) no obstruction

b) partial AP obstruction

c) complete AP obstruction

d) partial lateral obstruction

e) complete lateral obstruction

f) partial concentric obstruction

g) complete concentric obstruction

Velum-related obstruction occurs at the level of the palate, uvula, or lateral pharyngeal wall tissues at the level of the velopharynx. Because these are not entirely distinct entities—both anatomically and on DISE—the VOTE Classification groups them together. Airway closure related to the velum can occur with collapse in an anteroposterior (AP), concentric, or, less often, lateral configuration.

Oropharyngeal Lateral Walls, including Tonsils (Figure 4 a–e).

Figure 4. : Oropharyngeal Lateral Wall-Related Obstruction.

a) no obstruction

b) partial obstruction

c) complete obstruction

d) partial obstruction primarily related to 3+ tonsils

The oropharyngeal lateral walls combine the tonsils and the lateral pharyngeal wall soft tissues, including musculature and the adjacent parapharyngeal fat pads. The oropharyngeal lateral walls collapse in a lateral configuration, although this may occur in combination with collapse of other structures such as the tongue, creating what can appear as a concentric pattern. In the presence of lateral wall collapse, it can be difficult (but certainly not impossible) to determine whether the tonsils or lateral walls are playing the more-important role, reflecting potential subtypes. Importantly, the distinction can have important implications for treatment selection and outcomes, as enlarged tonsils would be treated with removal in conjunction with palate surgery. While the VOTE Classification is based on DISE findings alone, the examination of tonsil size and lateral pharyngeal wall tissues during routine oral cavity examination can be invaluable in making a determination of potential contributions of each structure. Obstruction related to the oropharynx can occur with collapse in a lateral configuration but not in an anteroposterior or concentric configuration

Tongue (Figure 5 a–d).

Figure 5. Tongue-related Obstruction.

a) partial AP obstruction

b) complete AP obstruction

c) hypertrophic tongue tonsils

d) effect chin lift

Tongue-related obstruction is common during DISE, and it occurs in the anteroposterior configuration only. Tongue-related obstruction in OSA patients may occur because of greater tongue volume, narrowing of the space that the tongue occupies (for example, with mandibular insufficiency or retrognathia), or solely related to the decrease in tongue muscle tone that occurs in natural sleep.

Epiglottis.

Epiglottis-related obstruction can occur in either the anteroposterior (Figure 6a) or lateral (Figure 6b) configurations. Anteroposterior collapse can occur with folding of the superior portion of the epiglottis (appearing related to decreased structural integrity of the epiglottis) or posterior displacement of the entire epiglottis against the posterior pharyngeal wall (apparently normal epiglottic structural integrity). The lateral configuration is characterized by a lateral folding or involution and is consistent with a central vertically oriented crease of decreased rigidity of the epiglottis, typically in the setting of oropharyngeal lateral wall-related obstruction. The epiglottis may be underrecognized as a factor in patients with sleep disordered breathing, and it is estimated that 15% of OSA patients demonstrate a significant epiglottic contribution to airway obstruction, independent of tongue-related obstruction that produces airway obstruction posterior to the epiglottis.^39,40 DISE may provide a unique assessment of the epiglottis, as its apparent role in upper airway obstruction is not identified with other evaluation techniques (awake fiberoptic examination, Friedman Stage).⁴¹

Figure 6. Epiglottis-related Obstruction.

a) Partial anteroposterior obstruction

b) Complete anteroposterior obstruction

c) Partial lateral obstruction

d) Complete lateral obstruction

Other Structures

Airway obstruction in sleep disordered breathing can be related to other structures, although these less commonly play a key role in OSA. In rare cases, obstruction in the nasal cavities (e.g. by massive nasal polyps, adenoid hypertrophy or nasopharyngeal neoplasms) or the larynx (e.g. in substantial edema related to reflux or previous radiation or with vocal fold immobility) can be visualized. Fortunately, these are usually detected during awake fiberoptic examination prior to DISE. If identified during DISE, they are noted separately.

Degree of Airway Narrowing

The VOTE Classification involves a qualitative assessment of the degree of airway narrowing related to each of the VOTE structures, with categories of:

none (no vibration of the involved structure and with airway dimensions sufficient to permit normal airflow);

partial (vibration of the involved structure causing obstruction or other structures and with reduced airflow); or

complete (complete or near-complete obstruction, with no or markedly-limited airflow).

Differentiating between the three categories is not always clear, and the degree of obstruction is scored separately for each VOTE structure (as with configuration). At one level in the upper airway, a partial collapse (vibration, snoring) might be present, while at the other level no or complete collapse might be detected. Although qualitative, the evaluation of degree of obstruction has been demonstrated as having fair to substantial reliability (poorest for the velum).^8,12

The VOTE Classification does not preclude additional assessments. In fact, VOTE scoring can be performed separately in different body positions, mouth closure (Figure 1), Esmarch Maneuver (Figure 2), or administration of positive airway pressure.

COMPLICATIONS

There have been no severe side effects or complications related to DISE reported in the literature. In the more than 10,000 combined DISE cases performed by the authors, there have been no cases of endotracheal intubation, tracheostomy, or use of flumazenil. Oversedation can occur, requiring airway management with the use of temporary airway support (without or with positive airway pressure) or, in rare cases, a laryngeal mask airway. Saliva aspiration can occur, but it is uncommon and usually not dangerous; however, it may compromise the procedure, due to extreme coughing. Laryngospasm is extremely rare. Regurgitation and aspiration of gastric contents are theoretically possible and could need specific treatment.

To prevent complications in high-risk patients, such as those with body mass index > 35 kg/m2, problematic anatomical features (short neck, Modified Mallampati Position IV), DISE may be performed concurrent with PAP administration.

CONCLUSION

DISE is an important evaluation technique for patients with OSA who are unable to tolerate positive airway pressure. The VOTE Classification is based on the specific structures that may contribute to airway obstruction in a specific patient. The VOTE Classification provides a common language for characterizing the pattern of obstruction. DISE examination can include separate assessments with changes in body or head position, mouth closure, Esmarch Maneuver, or other interventions that simulate therapies. Studies, include some multicenter cohort studies, have identified specific DISE findings associated with outcomes of upper airway surgery, hypoglossal nerve stimulation, and MRAs.