Understanding Correlation of Polysomnography Parameters with Drug Induced Sleep Endoscopy in Obstructive Sleep Apnea

Abstract

The aim of this study was to analyze drug-induced sleep endoscopy (DISE) findings performed in 64 patients and to evaluate the association of DISE findings with PSG parameters. This retrospective, single center, observational study included patients with obstructive sleep apnoea (OSA) who have undergone DISE as part of surgical planning. DISE was performed using dexmedetomidine infusion. The data were documented as per VOTE (velum, oropharynx, tongue base and epiglottis) classification. The patient characteristics and level 2 Polysomnography (PSG) findings were analyzed with DISE findings. Among 62 patients, mean AHI and lowest oxygen saturation levels were 39.68 ± 27.59 and 78.36 ± 9.38, respectively. Mean ESS, SSS and PSQI levels were 10.74 ±  − 4.96, 7.73 ±  − 1.52, and 8.92 +/− 4.99, respectively. A Single level of obstruction was observed in 4.8% patients, followed by 40.3%, 43.5%, and 11.3% were found to have 2, 3, and 4 levels of obstruction, respectively. All patients had palatal involvement, followed by the oropharyngeal (88.7%), the tongue base (59.7%), and the epiglottis (12.9%) obstruction. There was no significant correlation between partial collapse at velopharynx with AHI. However, complete collapse at the level of the oropharynx (p < 0.05) and the tongue base (p < 0.001) showed a statistically significant association with AHI. Also, a significant correlation was observed between the number of sites and AHI Grade (p < 0.0001). The study helps to understand the association of different patterns and degrees of anatomical obstruction in OSA with different PSG parameters. PSG and DISE findings are complimentary to each other in treatment planning and selection of surgical procedures.

Introduction

Pringle and Croft introduced drug-induced sleep endoscopy (DISE) for assessing obstructive sleep apnoea (OSA) as a tool for selecting cases for surgical intervention [1]. Patients with OSA, during the awake state, maintain the upper airway patency through protective reflexes, which increases upper airway dilator muscle activity. Thus, they compensate for an otherwise anatomically compromised upper airway. It is only possible to evaluate real time changes during snoring through DISE or cine- MRI [1, 2].

According to our experience cine-MRI is not easy or feasible due to multiple logistic and patient factors. We follow the protocol of DISE which is relatively easy, safe and result oriented.

DISE is increasingly being used as a pre-operative examination tool for patients undergoing surgery for OSA because it can provide three-dimensional evaluations of changes in the upper airway during pharmacologically induced sleep [3].

Materials and Methods

This was a retrospective, observational study. The medical records of patients who underwent DISE were reviewed. Data including, demographics, snoring history, comorbidities, body mass index, and ENT examination details were obtained in a patient case report form (CRF) from hospital source records. The polysomnography (PSG) report and DISE findings were also collected.

Ethical Committee clearance was taken on 14th April, 2016 XXX. Registration No: ECR/353/Inst/AP/2013.

Evaluation in the Outpatient Department

All patients with snoring underwent initial clinical evaluation in the outpatient department. Their snoring severities was assessed by Snoring Severity Score (SSS) [4]. Excessive day time sleepiness and quality of sleep assessment were assessed using two questionnaires—Epworth Sleepiness Scale (ESS) and Pittsburgh Sleep Quality Index (PSQI) [5]. Clinical parameters, such as body mass index (BMI) and blood pressure (BP) were documented. ENT examination was carried out focusing on nasal septal deviation, turbinate hypertrophy, adenoid hypertrophy, tonsillar size, palatal morphology, uvula, and tongue [6].

All snorers were advised to undergo PSG. The patients who were proved to have OSA (as per definition given by AASM) from PSG and who were potential candidates for surgical intervention were included in the study.

Polysomnography (PSG)

Patients were advised for level 2 home sleep study. The following parameters were monitored during PSG: central and occipital electroencephalogram (EEG), electrooculogram (EOG), submental is electromyography (EMG), nasal and oral airflow, thoracic and abdominal wall motion, body position, and electrocardiogram (ECG). Arterial oxygen saturation was monitored with a pulse oximeter. Tracing was scored using 30 s epoch. Hypopneas were scored per the American Academy of Sleep Medicine (AASM) definition [7]. Apnea hypopnea index (AHI) and lowest desaturation (LSAT) were noted for all patients.

Drug Induced Sleep Endoscopy (DISE)

DISE was performed in the operating theater after obtaining informed consent from the patients. Patients were asked to be nil per mouth for at least 6 h before the procedure. Dexmedetomidine (bolus dose of 0.5 µcg/kg) was injected 20 min before the procedure; followed by dexmed infusion (0.2–0.7 µcg/kg) by infusion pump during the procedure. Each patient’s nose was sprayed with 10% xylocaine spray 5 min before DISE. As each patient entered the procedure room, lights were dimmed, and was made to lie in a comfortable supine position on the operating table with a pillow. A bispectral index (BIS) monitor system was used to monitor the depth of sleep. As the patient started snoring, the flexible endoscope (Karl Storz flexible video-laryngoscope) with a diameter of 3.5 mm—was passed through the nose. The BIS score was maintained between 50 and 70. As the scope passed through the nose and nasopharynx, static obstructions were noted like deviated nasal septum and adenoid hypertrophy, hypertrophied salpingo-pharyngeal fold. Beyond the nasopharynx, dynamic obstructions were observed either at a single level or at multiple levels. The Average time taken to perform DISE was 10 to 15 min.

None of the patients had any adverse events/complications during or after the procedure. Throughout the procedure, the anesthetist monitored the basic cardiorespiratory parameters. For emergency management, oxygen delivery system and intubation tray were kept at the bedside. After the procedure, patients were kept under observation for about an hour. Once they were out of the sedation effect, they were discharged home.

Every patient’s endoscopy video was recorded. The patients selected for surgery, their plan was explained to them as decided based on the DISE findings.

VOTE Classification

The findings are recorded in a table according to the VOTE classification. This system focuses on the primary structures contributing to upper airway obstruction, either alone or in combination: the velum, oropharyngeal lateral walls (including the tonsils), tongue, and epiglottis [3]. Velum includes the soft palate, uvula, and lateral pharyngeal wall tissue at the level of the velopharynx. Airway closure related to the velum can occur with collapse in an anteroposterior, lateral, and concentric manner. The oropharyngeal level includes: the palatine tonsils and the lateral pharyngeal wall tissues, including the pharyngeal musculature and the parapharyngeal fat pads. In the original VOTE system, only the lateral oropharyngeal collapse had been considered; however, concentric collapse is not uncommon. Therefore, in the current study, we have documented circumferential oropharyngeal collapse too. Tongue base collapse is a common finding, and it results in anteroposterior narrowing of airway. Epiglottic collapse occurs either in the anteroposterior or in the lateral direction.

Results

Among the study population (n = 62), 87.1% were males and 12.9% were females. The age group ranged from 18 to 64 years. Baseline characteristics, including BMI, ESS, SSS, PSQI, and PSG parameters (AHI, LSAT) are shown in Table 1. The two most common comorbidities present among patients with OSA were hypertension and diabetes mellitus (12.5% each). Hypothyroidism was seen in 4.8% patients. The distribution of patients according to the number of anatomical sites involved as per DISE findings (VOTE classification) is depicted in Fig. 1.

Table 1.

Baseline analysis of patient’s characteristics, clinical and sleep study parameters (n = 62)

Variables	Mean ± Std. deviation
Age	38.37 ± 10.32
Weight (Kg)	79.82 ± 12.37
BMI (kg/m2)	27.00 ± 3.40
ESS	10.74 ± 4.96
SSS	7.73 ± 1.52
PSQI	8.92 ± 4.99
SBP mm Hg	128.81 ± 18.70
DBP mm Hg	77.23 ± 9.26
AHI per hour	39.68 ± 27.59
LSAT%	78.36 ± 9.38

BMI Body-mass index, ESS Epworth sleepiness scale, SSS Snoring severity scale, PSQI Pittsburg sleep quality index, SBP Systolic blood pressure, DBP Diastolic blood pressure, AHI Apnea–hypopnea index, LSAT Lowest saturation of arterial oxygen

Fig. 1.

Distribution of patients according to number of anatomical sites involved

All patients (100%) had a palatal involvement, followed by the oropharyngeal obstruction (88.7%), tongue base (59.7%), and epiglottis (12.9%) involvement. Anteroposterior collapse at the velum (p = 0.002) and tongue base collapse (p = 0.001) has shown a statistically significant correlation with AHI. (Table 2 and 3).

Table 2.

Association between the anteroposterior obstruction at the velum and AHI grades

Grade of AP collapse at velum	Grade of OSA as per the AHI			Total (62)
	Mild OSA (13) (%)	Moderate OSA (16) (%)	Severe OSA (33) (%)
0	23.08	87.50	60.61	37 (59.68%)
1	30.77	6.25	3.03	6 (9.68%)
2	46.15	6.25	36.36	19 (30.64%)

Mild OSA AHI 5–15, Moderate OSA AHI > 15–30, Severe OSA AHI > 30; figures in parenthesis denotes total number of patients in that group; AP Anteroposterior, OSA Obstructive sleep apnoea, AHI Apnea–hypopnea index

p = 0.002, Pearson Chi-Square value

Table 3.

Association between the tongue base obstruction and the AHI grades

Gradeof collapse at the Tongue base	Grade of OSA as per the AHI			Total (62)
	Mild OSA (13) (%)	Moderate OSA (16) (%)	Severe OSA (33) (%)
0	76.92	56.25	18.18	25 (40.32%)
1	23.08	31.25	27.27	17 (27.42%)
2	0	12.50	54.55	20 (32.26%)

Mild OSA AHI 5–15, Moderate OSA AHI > 15–30, Severe OSA AHI > 30. Figures in parenthesis denote total number of patients in that group; OSA Obstructive sleep apnoea, AHI Apnea–hypopnea index

p = 0.002, Pearson Chi-Square value

Table 4 shows the correlation of mean BMI and mean AHI with different patterns and degrees of anatomical obstruction as documented during DISE according to VOTE classification. No significant correlation was found between partial collapse and AHI. However, complete collapse at the level of oropharynx (p < 0.05) and tongue base (p < 0.001) showed a statistically significant association with AHI. BMI did not show any significant correlation with any different anatomical sites of obstruction.

Table 4.

Association between mean AHI and mean BMI and different patterns and degrees of narrowing of various anatomical levels in upper airway during DISE

Site, pattern, and degree of collapse	Mean AHI ± SE (n) when variable absent	Mean AHI ± SE (n) when variable present	p value	Mean BMI ± SE, when variable absent	Mean BMI ± SE, when variable present	p value
V, AP-1	43.79 ± 4.40 (37)	25.92 ± 13.9 (6)	0.151	27.0 ± 0.61 (37)	26.1 ± 1.23 (6)	0.599
V, AP-2	43.79 ± 4.40 (37)	48.19 ± 8.02 (19)	0.308	26.95 ± 0.61 (37)	27.38 ± 0.66 (19)	0.664
V, L-2	37.43 ± 3.81 (49)	48.19 ± 8.02 (13)	0.214	26.73 ± 0.49 (49)	28.01 ± 0.87 (13)	0.231
V, C-1	39.33 ± 4.75 (37)	8 (1)	0.292	27.19 ± 0.45 (37)	26.50 (1)	0.807
V, C-2	39.33 ± 4.75 (37)	41.55 ± 5.25 (24)	0.76	27.19 ± 0.45 (37)	26.73 ± 0.88 (24)	0.61
O, L-1	46.33 ± 5.03 (31)	30.93 ± 5.72 (18)	0.058	27.01 ± 0.61 (31)	27.16 ± 0.87 (18)	0.89
O, L-2	46.33 ± 5.03 (31)	35.95 ± 7.96 (13)	0.27	27.01 ± 0.61 (31)	26.73 ± 0.89 (13)	0.80
O, C-1	35.67 ± 4.56 (37)	32.0 ± 7.07 (8)	0.72	26.72 ± 0.59 (37)	27.58 ± 0.86 (8)	0.53
O, C-2	35.67 ± 4.56 (37)	52.03 ± 6.71 (17)	0.05	26.72 ± 0.59 (37)	27.31 ± 0.82 (17)	0.57
T-1	27.56 ± 5.21 (25)	36.66 ± 5/60 (17)	0.252	27.30 ± 0.77 (25)	27.05 ± 0.72 (17)	0.823
T-2	27.56 ± 5.20 (25)	57.41 ± 5.50 (20)	0.001	27.30 ± 0.76 (25)	26.58 ± 0.74 (20)	0.511
E	38.26 ± 3.82 (54)	49.28 ± 8.23 (8)	0.29	27.06 ± 0.46 (54)	26.55 ± 1.15 (8)	0.68

V velum, O Oropharynx, T Tongue base, E Epiglottis, AP Anteroposterior, L Lateral, C Circumferential, 1 Partial collapse (50–75%); 2 > 75% to complete collapse, AHI Apnea–hypopnea index, BMI Body mass index, SE Standard error

The patients (76.9%) who had one or two sites involvement belonged to the mild OSA group (Table 5). Among the patients with involvement of 3 or more sites, majority (75.8%) belonged to the severe OSA group. This established a significant correlation between AHI and the number of sites (p < 0.01). There was no such significant correlation between BMI and number of sites (p > 0.05).

Table 5.

Distribution of AHI per number of sites involved in DISE findings

Number of anatomical sites narrowing	Mild OSA AHI 5–15	Moderate OSA AHI > 15–30	Severe OSA AHI > 30		Total
1 and 2	10 (76.9%)	10 (62.5%)	8 (24.2%)		28 (45.2%)
3 and 4	3 (23.1%)	6 (37.5%)	25 (75.8)		34 (54.8%)
Total	13 (100%)	16 (100%)	33 (100%)		62 (100%)

AHI Apnea Hypopnea Index, Chi-Square Yates p value = 0.0054

Discussion

Obstructive sleep apnoea is a multi-factorial disease. It is related to sedentary lifestyle, weight gain and changing morphology of human facial skeleton in different races in last decades. In modern society incidence of OSA is increasing [8]. For successful management it is important to understand how upper airway behaves during sleeping. Every individual has different phenotypical features of upper airway [9].

After Fujita et al. published their work on UPPP the surgery got popularity. But as time passed many publications showed failure of UPPP. The reason is that UPPP cannot address lateral pharyngeal collapse and tongue base collapse. Then came the concept of multilevel obstruction and surgical correction at multiple level. Hence the success depends on patient selection and proper surgical planning [11]. One surgery cannot be solution for all. In this context we find DISE as the most useful tool to plan surgical procedures corroborated with PSG report.

We perform different palatoplasty techniques depending of degree and pattern of velopharyngeal collapse. Tongue base minimally invasive procedure is done for grade 1 collapse and maxilla mandibular advancement for higher grade tongue base obstruction with very high AHI (> 40).

After decades of DISE being a popular practice in many countries, there is no doubt that DISE complements the basic ENT examination and PSG in the evaluation of patients with sleep-related breathing disorders (SRBD). Now the question is whether the same insight can be gained from awake endoscopy. Majority of the studies have proved that DISE provides more information about the anatomical locations and pattern of obstruction and changes surgical decision-making compared to awake evaluation methods [10]. Campanini et al., demonstrated that awake and sedation evaluation was identical in only 25% of the cases [11]. These discrepancies mainly involved the oropharyngeal and laryngo-hypopharyngeal sites. Another similar study has shown a significant difference in the incidence of severe retro-lingual collapse identified via DISE compared to Muller’s maneuver [12].

There are certain patient characteristics and PSG outcomes that show an association between certain patterns and degrees of upper airway collapse during DISE. Our small-scale study helped to unravel these associations, thereby, understanding the pathogenesis of OSA and tailor multilevel surgery for patients. We found a significant correlation between number of sites of obstruction and AHI, but not with BMI. Complete oropharyngeal and tongue base collapse had statistically significant association with AHI. A similar result was obtained in two prospective studies conducted by Madeline et al. [13], and De Corso, et al. [14]. Their results suggested that multilevel collapse and complete collapse are significantly associated with higher AHI values. Madeline et al., (n = 100) showed that there was statistically significant association between complete concentric collapse and increased BMI [14]. However, De Corso et al., (n = 138) demonstrated no significant correlation between BMI and degree of collapse [15].

A recent review article included eight studies (with 535 patients), which directly compared awake examination data with DISE outcome in terms of the effect on surgical decision-making and surgical success [16]. It stated that surgical treatment changed after DISE in 50.24% cases. DISE guides in surgical planning and in selecting ideal candidates for oral devices [17, 18]. This is accomplished by mandibular advancement maneuver performed during DISE [19]. DISE can also be recommended as a patient selection tool for implanted upper airway stimulation (UAS) therapy [20, 21].The study conducted by Vanderveken et al., has shown that the absence of palatal complete concentric collapse during DISE may predict therapeutic success with UAS therapy.

Our recommendation is that moderate to severe OSA should be cautiously evaluated for multilevel narrowing, with special attention to the lateral pharyngeal wall and tongue base. For a better treatment outcome, patient characteristic, comorbidities, and PSG parameters should be considered along with the specific pattern and degree of collapse at various levels in DISE; none of these evaluations are sufficient alone, however they are complementary to each other. In case, DISE finding shows only partial collapse and, or single level collapse in a patient with very high AHI, it is advisable to repeat the PSG. Many times PSG report from one night study can be wrong due to scoring error, lesser sleep efficiecy etc. On the contrary, some cases may show severe obstruction in DISE while PSG parameters show mild OSA. In that case we must check BIS monitor score. Deeper sedation and too much propofol use can show exaggerated obstruction in DISE. Hence to get actual result from DISE, proper titration of the sedating agent is crucial. DISE performed in continuous positive airway pressure (CPAP) non-compliant patients can also provide valuable insights like epiglottis collapse, which is otherwise difficult to diagnose by awake endoscopy [22].

Conclusion

DISE is a valuable addition in the surgical planning of patients with OSA. It is safe and easy to perform procedure, helping the surgeons to make targeted strategies for a specific pattern and degree of collapse at various levels of the upper airway. This kind of assessment, when combined with patient characteristics, provides comprehensive knowledge on pathogenesis, and improves the success rate of OSA treatment.