Beyond VOTE: The New Frontier of Drug-Induced Sleep Endoscopy

Abstract

Alternatives to positive airway pressure therapy, including surgery, represent an important area of research. Specifically, predictors of response to surgical therapy remains underdeveloped. Drug-induced sleep endoscopy (DISE) holds promise as a diagnostic to tool to identify patient-specific causes of airway collapse. Herein, we present a novel, standardized approach which combines anatomic and physiologic measurements during DISE. Our DISE platform measures airflow, airway compliance, airway collapsibility as well as structural drivers of collapse. Taken together, these inputs provide a comprehensive framework to further inform the surgeon in providing personalized care of the patient with obstructive sleep apnea.

Background

Obstructive sleep apnea (OSA) is a disorder that involves repetitive cycling of complete or partial blockage in airflow during sleep. In addition to poor sleep quality and daytime sleepiness, the resulting hypoxia and sympathetic activation place patients at an increased risk for neurocognitive decline and a myriad of cardiometabolic sequelae.^1,2 The primary therapy for OSA is positive airway pressure (PAP), and while highly efficacious, regular adherence ranges from 17 to 54%.³ Consequently, there exists a compelling need for treatment alternatives to PAP therapy.

While PAP serves as a pneumatic stent of the entire pharynx, PAP alternatives act on specific pharyngeal subsite(s). The clinician must identify the specific mechanisms of airway collapse to provide personalized treatment options. As OSA is a state-dependent phenomenon and natural sleep evaluations pose logistical challenges, the advent of drug-induced sleep endoscopy (DISE) in 1991 offered promise as a dynamic upper airway evaluation for the sleep clinician.⁴ Initially described as “nasopharyngoscopy with small doses of midazolam”, the field steadily sought to standardize both anaesthesia and otolaryngology parameters. In 2011, Kezirian et al. presented the velum, oropharyngeal, tongue base, epiglottis (VOTE) classification, which has become the most widely used classification scheme.⁵

VOTE describes the four major regions of obstruction: the velum (soft palate, uvula, or lateral pharyngeal walls), oropharyngeal lateral walls (palatine tonsils, pharyngeal muscles, and parapharyngeal fat), tongue base, and epiglottis. Obstruction is classified by configuration (anterior-posterior, lateral, concentric) and degree. The degree of obstruction for each structure is described as none (no vibration of the structure and <50% airway narrowing), partial (vibration, 50–75% narrowing), or complete (obstruction, >75% narrowing). While the VOTE system provides a construct for sleep surgeons in which to interpret DISE, its ability to predict surgical outcomes has been limited.^6,7 The common explanation stems from the ostensible differences between natural and sedated sleep including the effects of anaesthesia depth on airway collapsibility⁸ and the inability to replicate REM sleep⁹. There appears, however, to be an equally important limitation of the VOTE score: the reliance on a surgeon’s visual impression of an endoscopic video recording. Obstructive sleep apnea is a disorder clinically defined by quantifiable measures of airflow reduction --- not based on airway imaging. In addition to the lack of quantifiable measures, the site of collapse from the VOTE score may not necessarily reflect the cause of collapse. In these ways, the VOTE classification bears 3 important limitations:

1. Visually-based evaluation

2. Subjective grading

3. Description of site(s), not driver, of collapse

Herein, we present a quantifiable, physiologic complement to the standard DISE evaluation which seeks to address the above limitations with a distinct focus on improving patient selection for the sleep apnea surgeon. Specifically, our setup provides a novel paradigm to assess (a) severity of collapse (b) site(s) of airflow obstruction and (c) patterns of flow obstruction

Rationale for Measurements

A. Airflow (Pneumotachometer)

Sleep medicine uses airflow reduction to define both obstructive apneas (>90% reduction) and obstructive hypopneas (>30% reduction). As DISE is meant to simulate obstruction seen during natural sleep, airflow measurements --- rather than visual impression of collapse --- can be used to identify the extent and severity of obstruction and its impact on ventilation. Nasal airflow is measured directly with a pneumotachometer that is placed in-line with the the nasal mask/breathing circuit (see Figure 1). In addition, the use of simultaneous oximetry and epiglottic catheters (measure of respiratory effort) parallels a Type III sleep study or home sleep test.

Fig. 1.

Drug-Induced Sleep Endoscopy (DISE) Recording Platform

The drug-induced sleep endoscopy characterization of airway dynamics (DISE-CAD) protocol and recording platform integrates nasolaryngoscopy video with synchronized upper airway physiology tracings including nasal pressure (P_N), nasal airflow (Flow), pharyngeal manometry via catheters at the distal soft palate (P_US) and inferior to the epiglottis (P_DS), and oxygen saturation (SpO₂). Depicted is a positive airway pressure (PAP) run with inspiration plotted in the downward direction. Of note, discrete phases of sleep-disordered breathing are observed including respiratory cycling (i.e. repeated obstructive events) (P_N [0–5] cm H₂O), stable flow limitation (P_N [5–8] cm H₂O), non-flow limited breathing or “DISE PhOP” (P_N = 9 cm H₂O), and maximal pharyngeal cross-sectional area (P_N = 13 cm H₂O).

Of note, our group has observed several patients in whom the airway appeared to be obstructed by the epiglottis, yet airflow was completely unobstructed (non-flow-limited). These examples underscore the importance of an airflow-based assessment of airway patency as a means of reconciling DISE and polysomnography findings.

B. Compliance & Collapsibility (Airway Opening Pressures & Pressure Catheters)

VOTE scoring has been shown to have inter-rater reliablilty from poor to good^10,11. This variability poses inherent problems for use as a primary outcome variable. By measuring the pharyngeal opening pressure (PhOP), and subsite opening pressures (e.g. palatal opening pressure (POP)), the DISE-CAD setup provides objective measures of airway compliance, defined as the change in area for a given change in pressure..^12,13 The critical pressure (Pcrit) of the airway can also be extracted from the pressure-flow relationship across a range of nasal pressure from atmospheric to DISE-POP.

In addition, the two individual pressure catheters demonstrate sites of collapse and airflow obstruction. They can also be used to measure the collapsibility of each site encompassed by the VOTE score, and hence, can be used to delineate primary and secondary sites of pharyngeal collapse. The ordinal description of collapse sites may have implications for surgical management.¹⁴ Additional insight can be gain by measuring the impact of pharyngeal neuromuscular¹⁵ and patterns of airflow obstruction (i.e., negative effort dependence)¹⁶, which remain beyond the scope of this paper.

C. Driver of Collapse (Imaging)

Soft palate obstruction has been shown on sleep MRI to be secondary to collapse of the junctional tongue (intersection of oral and oropharyngeal tongue) in nearly half of cases.¹⁷ As significant soft palate obstruction occurs in > 80% of DISE exams,¹⁸ the driver of soft palate collapse is paramount for optimal surgical selection. Identifying tongue-driven soft palate collapse on traditional DISE is challenging, relying on collapse patterns (anterior-posterior soft palate collapse) and tongue position (posteriorly positioned with inspiratory glossoptosis). Unlike nasolaryngoscopy, ultrasound during DISE provides real-time imaging of the tongue-soft palate interface (see Figure 2). In this way, the ultrasound imaging during DISE complements the findings of both DISE nasolaryngoscopy and the pre-operative CT imaging (next section).

Fig. 2.

DISE Tongue Ultrasound

Ultrasound image during DISE with probe placed in submental midline. Note presence of anterior tongue dorsum as hyperechoic line (*) while posterior tongue dorsum appears hypoechoic and diffuse (**). The echogenicity of the tongue dorsum differentiates air contact (*) vs soft palate contact (**).

Standard Clinical Assessment

All patients are seen in the Penn CPAP Alternatives Clinic for consultation. Previous sleep studies, including hypnogram and tables, are analysed prior to clinic visit. PAP data is also obtained, when available. Following a comprehensive history and physical exam, DISE is recommended to either better understand the drivers of airway obstruction or assess candidacy for hypoglossal nerve stimulation (HGNS).

Before DISE, each patient undergoes a specialized head and neck non-contrast CT scan. With the patient lying supine, neck neutral, lips closed and capture on exhalation, the radiographic images provide a standardized assessment of the bony, soft tissue and airway anatomy. Figure 3 shows a patient’s CT with the junctional tongue opposing the soft palate. This patient’s subsequent DISE exam demonstrated tongue-driven soft palate collapse.

Fig. 3.

Computed Tomography – Sleep Protocol

Non-contrast, bone window, midsagittal section of patient undergoing CT sleep protocol. Note apposition of junctional tongue along course of soft palate.

Drug-Induced Sleep Endoscopy – Characterization of Airway Dynamics (DISE-CAD)

Pilot Studies

In 2018, our research group became interested in identifying predictive factors for HGNS, a relatively novel procedure with variable outcomes. Lee et al¹⁹ studied the use of an individual’s therapeutic PAP levels, a marker for airway collapsibility shown to predict outcomes of oral appliance therapy,^20–22 as a potential predictor of HGNS success. The study data demonstrated that patients with low PAP requirements (<8cm H₂O) had significantly improved response to HGNS compared to patients with higher PAP requirements (mean apnea-hypopnea index reduction of 36.7 ± 22.7 versus 18.4 ± 23.4, P = .02). Subsequently, we sought to better understand the site-specific effects of positive pressure in a prospective manner. In order to do so, we identified a DISE outcome measure (pharyngeal opening pressure) analogous to the minimally-effective therapeutic PAP level obtained from natural sleep. In 2019, we conducted a DISE study which demonstrated the ability of a single, objective measurement --- palatal opening pressure (DISE-POP) --- to predict outcomes for HGNS beyond that of both the VOTE score and commonly used anthropometric measures.²³ Subsequently, our group formally launched a specialized protocol: Drug-Induced Sleep Endoscopy – Characterization of Airway Dynamics (DISE-CAD). DISE-CAD represents an integrated recording platform designed to extract clinically-relevant anatomic and physiologic features of each patient’s upper airway. The patient setup, anesthesia procedure and biologic measurements are described below.

Patient Setup

After informed consent (IRB833511), DISE-CAD patients proceed to the DISE suite. Once in the suite, 0.2mg of intravenous glycopyrrolate is administered by anesthesia. Once the patient is transferred to the procedure bed and placed in a semi-upright position, the more patent naris is topically medicated with 4% lidocaine and 0.05% oxymetazoline. The other naris is intentionally not medicated to limit afferent receptor-mediated breathing dysfunction.²⁴ Once the medicated cottonoids are removed, the patient is then reclined in a supine position to avoid inadvertent pharyngeal anesthesia. At this time, two Mikro-Cath^™ pressure catheters (Millar, Inc., Houston, TX, USA) are placed in the anesthetized nostril and advanced to the level of the nasopharynx (13 cm). A nasal CPAP mask (Pulmodyne, Indianapolis, IN) and pneumotachometer are secured and the nasolaryngscope is passed through the custom fitted airtight mask opening and into the nasal cavity (Fig. 1). Two liters/minute of supplemental oxygen via a nasal cannula (placed in the mouth) are provided. End-tidal CO₂ is sampled through the oral cannula to detect mouth breathing, and pulse oximetry is monitored.

Procedure Steps

Once the nasolaryngoscope reaches the soft palate, propofol anesthesia is administered using a probability ramp infusion system.²⁵ The pressure catheters are adjusted in that the distal catheter is placed in the retroepiglottic space and proximal catheter at the distal margin of the soft palate. When the patient demonstrates an obstructive hypopnea or apnea (based on the nasal airflow and epiglottic catheter inputs), the clinical end point is reached and the propofol infusion rate is maintained.

1. A baseline surveillance of the upper airway is performed, recording a minimum of 2 breaths at the soft palate, oropharyngeal lateral walls, tongue and epiglottis, respectively.

2. An ultrasound probe (C10–3/4 with Zonare S3; Mindray Medical, Mahwah, New Jersey) is concurrently placed in the midline of the submental region. The probe is used to image the interface of the tongue and soft palate. See Figure 2.

3. Nasal PAP ramp is performed starting at 2cm H₂O. The pressure is raised by increments of 1–2cm H₂O until inspiratory airflow limitation is abolished. This pressure defines the pharyngeal opening pressure (DISE-PhOP). The pressure is then increased step-wise by an additional 4cm H₂O to evaluate the maximal pharyngeal cross-sectional area (Amax).

4. The propofol infusion is increased by 10% to ready the patient for planned manipulations.

5. With body straps across the patient’s chest and hips, the bed is “airplaned” to the patient’s right at 45 degrees. The assistant turns the patient’s head an additional 45 degrees to simulate right-sided sleep. The upper airway surveillance sequence is repeated.

6. The bed and patient’s head are returned to the baseline position. The lower jaw is manually distracted by the assistant’s index finger such that incisors are edge-to-edge. An upper airway surveillance sequence is performed.

7. The jaw is released. The mandible is hinged close such that the teeth are in occlusion. An upper airway surveillance sequence is repeated.

8. Another nasal PAP ramp is performed with teeth in occlusion starting at 2cm H₂O. The pressure is raised by increments of 1–2cm H₂O until the abolishment of airflow limitation. This pressure is effectively the pharyngeal opening pressure (DISE-PhOP). The pressure is then increased by an addition 4cm H₂O to evaluate the maximal pharyngeal cross-sectional area (Amax).

9. The mandible closure is released. The camera is removed from the nasal cavity and placed intraorally. The tongue is examined, particularly its relationship to the hard and soft palate. The camera is withdrawn and procedure is terminated.

Limitations & Future Directions

In our early efforts to standardize the DISE procedure, several challenges exist including variation in both manipulations (lateral sleep, mandibular advancement) and depth of sedation. In addition, important validity (e.g. DISE-PhOP vs. therapeutic CPAP levels) and reliability (e.g. inter-rater measurements of DISE-PhOP) measures have yet to be studied. We are beginning to address the aforementioned limitations in a practical, stepwise fashion.

As our preliminary data using quantification of physiology during DISE has shown predictive capacity in HGNS-related outcomes,²³ we continue to expand our resources into the physiologic assessment during DISE. Currently, we are using 2 pressure catheters to partition the airway into a palatal and oropharyngeal airway. The next phase of airway pressure segmentation will include a custom, multi-sensor reusable catheter (HRM 6F custom esophageal catheter, Laborie Medical Technologies, Portsmouth, NH), creating 5 segments of distinct pressure profiles. This level of anatomic granularity enables structural assessment from the nasopharynx to supraglottic space and may improve characterization and clinical implications of multi-level collapse.

The validity of DISE in relation to natural sleep remains unanswered. In this way, our DISE setup will be recreated in our experimental sleep laboratory during natural sleep. The same subjects undergoing our DISE will complete a natural sleep assessment, enabling numerous, clinically-relevant comparisons related to site, pattern, intensity and driver of airway collapse. Establishing our DISE protocol before moving into natural sleep was a deliberate attempt to refine our physiologic assessment before tackling inherent challenges in naturally sleeping patients including patient tolerance and fluctuations in sleep stage and body position.

Ultimately, we plan to assemble the components of DISE-CAD as a turnkey clinical tool. This platform would allow clinicians to routinely assess key physiologic parameters that define the severity, sites and patterns of airway collapse. In this way, the otolaryngologist’s DISE exam can provide a comprehensive evaluation of predictive anatomic and physiologic traits, delivering personalized care for the OSA patient.