Drug‐Induced Sleep Endoscopy in Children Predicts OSA Severity

ABSTRACT

Background

While adenotonsillectomy (AT) is first line for pediatric obstructive sleep apnea (OSA), treatment failure is common. Drug‐induced sleep endoscopy (DISE) provides dynamic airway assessment, but its role in predicting OSA severity requires clarification. This study correlates DISE findings with polysomnographic severity.

Objective

To determine the association between DISE‐based obstruction patterns and OSA severity in children.

Methods

This prospective study included 95 children with adenotonsillar hypertrophy. All underwent preoperative Level III polysomnography followed by DISE before AT. Two blinded otolaryngologists independently reviewed DISE recordings, classifying obstruction via the VOTE system (degree 0–2; pattern: concentric, lateral, or anteroposterior). Statistical analysis employed non‐parametric tests, relative risk calculation, and ROC curves.

Results

Severe OSA (AHI > 10) was significantly associated with Brodsky grade IV tonsils (40.7% vs. 16.7%) and severe adenoid hypertrophy (44.1% vs. 16.7%). DISE identified severe obstruction at the velopharynx (p = 0.003) and tongue base (p = 0.036) as specific predictors of severe OSA, with high specificities (91% and 97.1%) and positive predictive values (87.7% and 90.9%). Concentric collapse was the most frequent velopharyngeal pattern (54.2%). Obesity was significantly linked to severe tongue base obstruction (38.9% vs. 5.4%, p = 0.001) and higher AHI. ROC analysis for tongue base obstruction combined with the Obstructive Apnea Index (OAI) showed high diagnostic accuracy (AUC = 0.811).

Discussion and Conclusion

DISE is a valuable tool for predicting pediatric OSA severity. Obstruction at the velopharynx and tongue base, particularly concentric collapse and in the context of obesity, are key risk factors. Integrating DISE findings with PSG enhances preoperative risk stratification, aiding in the selection of candidates who may benefit from personalized, multilevel surgical interventions.

Summary

• What Are the Significant Findings of the Study?

  • ∘
    Velopharynx and tongue base obstruction observed through DISE were significantly associated with pediatric OSA severity. Obesity increased the risk of severe tongue base collapse and higher AHI, confirming the value of combined anatomical and anthropometric assessment.

• What Does This Study Add?

  • ∘
    This study demonstrates DISE's utility, especially using VOTE classification, in preoperative risk stratification for pediatric OSA. Incorporating Z‐score and polysomnography allows a personalized approach to identify high‐risk children and improve outcomes through targeted therapeutic strategies.

1. Introduction

Obstructive sleep apnea (OSA) is characterized by recurrent episodes of partial or complete upper airway obstruction during sleep. It is a relatively common condition in the pediatric population, with a reported prevalence ranging from 1% to 5.7% [1, 2]. Left untreated, pediatric OSA can lead to numerous comorbidities, including arterial hypertension, attention‐deficit/hyperactivity disorder (ADHD), autonomic dysregulation, neurobehavioral disturbances, and a significant reduction in overall quality of life [3, 4].

According to current international guidelines, adenotonsillectomy (AT) is the first‐line treatment for pediatric OSA [5]. However, the persistence of OSA postoperatively remains a significant clinical concern, with residual disease reported in 30% to 75% of cases, particularly among high‐risk subgroups [6, 7]. These include children with Down syndrome, obesity, those older than 7 years, and those with severe preoperative OSA [8].

Although these risk factors are well‐established, therapeutic responses to AT very substantially, and the mechanisms underlying persistent postoperative OSA are not fully understood [9]. In this context, drug‐induced sleep endoscopy (DISE) has emerged as a valuable diagnostic tool, allowing for dynamic assessment of the upper airway under pharmacologically induced sleep that closely mimics natural sleep. Consequently, DISE is increasingly utilized in cases of residual OSA following AT to identify multilevel sites of obstruction that may contribute to persistent symptoms [10, 11].

Despite its potential for preoperative risk stratification and for informing individualized treatment strategies, evidence regarding the predictive accuracy of DISE for postoperative outcomes following AT remains inconclusive [12]. Although several studies support its utility in surgical planning, the lack of standardized protocols and validated prognostic markers continues to limit its widespread adoption [13].

The integration of clinical data, individual anatomical characteristics, and DISE findings may facilitate the development of more robust predictive models to optimize therapeutic decision‐making. This is particularly relevant for high‐risk populations, such as children with obesity, those of African descent, patients older than 7 years, and children with severe OSA or craniofacial abnormalities [14, 15].

The degree of tonsillar obstruction observed during DISE demonstrates a stronger correlation with polysomnographic parameters than assessments conducted during wakefulness [16]. Dynamic visualization of the upper airway during pharmacologic sleep provides a more realistic representation of its behavior during natural sleep and enables a comprehensive evaluation of obstruction beyond the tonsillar region [17, 18].

Therefore, this study aims to investigate the clinical and polysomnographic parameters associated with preoperative obstruction severity, as assessed by the standardized VOTE classification (Velopharynx, Oropharynx, Tongue base, Epiglottis). The VOTE classification is one of the most widely adopted frameworks for the topographic characterization of upper airway obstruction in DISE [19]. By correlating DISE findings with polysomnographic indices, this study seeks to enhance risk stratification and contribute to the development of individualized treatment approaches for pediatric OSA.

2. Materials and Methods

The study protocol was approved by the institutional ethics committee (CAAE: 42477014.2.0000.5411). Written informed consent was obtained from the parents or legal guardians of all participants, and assent was provided by children over the age of 9 years, in the presence of parents due to explanation.

This prospective interventional study was conducted in the Department of Otorhinolaryngology at the University Hospital of the Botucatu Medical School, São Paulo State University (UNESP), Brazil, between 2018 and 2020. A total of 95 children aged 3 to 11 years were enrolled. The study population comprised children presenting with symptoms of upper airway obstruction, including snoring, oral breathing, witnessed apneas, and restless sleep. All participants exhibited adenotonsillar hypertrophy graded as Brodsky III or IV and were clinically indicated for adenotonsillectomy.

Inclusion criteria were: age between 3 and 11 years, presence of snoring and/or oral breathing, and adenotonsillar hypertrophy (Brodsky grades III–IV). Exclusion criteria encompassed congenital heart disease, chronic pulmonary disease, neuromuscular disorders, and genetic syndromes.

On the night prior to surgery, all participants were admitted and underwent a Type III polysomnography (PSG) study to confirm or exclude obstructive sleep apnea (OSA). Recorded parameters included the Apnea‐Hypopnea Index (AHI), central apnea index, obstructive apnea index (OAI), hypopnea index, oxygen desaturation index, and snoring. OSA severity was classified according to the American Academy of Sleep Medicine (AASM) guidelines [20] as follows: mild (AHI ≥ 1 and < 5), moderate (AHI ≥ 5 and < 10), and severe (AHI ≥ 10). All PSG studies were scored by trained sleep technicians and interpreted under the supervision of a sleep medicine specialist.

Immediately prior to adenotonsillectomy, all patients underwent drug‐induced sleep endoscopy (DISE) to dynamically evaluate upper airway obstruction under sedation. The procedure was performed in the operating room with an anesthesiologist in attendance. Patients were placed in the supine position, and anesthesia was induced with sevoflurane (6% in oxygen) via face mask until loss of the palpebral reflex. Intravenous propofol (1 mg/kg) was administered as needed to maintain an adequate depth of sedation. A 2.7‐mm flexible nasopharyngoscope (Karl Storz, Tuttlingen, Germany) lubricated with 10% lidocaine gel was used for airway visualization. All endoscopic examinations were digitally recorded (Storz system) and independently reviewed by two experienced otolaryngologists who were blinded to patient data.

Upper airway obstruction was classified using the VOTE system. The degree of obstruction at each level—velopharynx, oropharynx (lateral pharyngeal walls), tongue base, and epiglottis—was graded as follows: (0) no obstruction (< 50%), and no vibration; (1) partial obstruction (50% to 75%) with vibration; (2) complete obstruction (> 75%) or not visible [18, 21]. In cases of complete velopharyngeal obstruction, the predominant collapse pattern was characterized as concentric (circumferential), lateral, or anteroposterior. Oropharyngeal obstruction was characterized by lateral collapse, resulting from constriction of the oropharyngeal lateral walls, influenced by both palatine tonsil size (Brodsky I‐IV) and redundant soft tissue in the area [21]. Adenoid hypertrophy was assessed via preoperative flexible nasopharyngoscopy and categorized as non‐severe (< 75% nasopharyngeal airway obstruction) or severe (≥ 75% airway obstruction).

Interrater agreement, assessed using the Kappa statistic, was substantial (κ > 0.875). Descriptive statistics for qualitative variables are presented as absolute and relative frequencies. For quantitative variables, measures of central tendency and dispersion (mean, standard deviation, median, minimum, and maximum) were computed. The Kolmogorov‐Smirnov test (applied given N < 100) indicated a non‐normal data distribution; consequently, nonparametric statistical tests were employed.

Associations between categorical variables were analyzed using the chi‐square test. Sensitivity, specificity, and relative risk (RR) were calculated to evaluate the strength of variable associations with outcomes and the predictive accuracy of the applied scales. The Mann‐Whitney U test was used for comparisons between dichotomized groups to determine statistical significance.

A comparative analysis was performed for each DISE level by dichotomizing obstruction as “severe” (complete obstruction) or “non‐severe” (partial or no obstruction) in relation to OSA persistence or resolution. Receiver operating characteristic (ROC) curve analysis was conducted to determine the optimal AHI cutoff value for predicting OSA severity. Sensitivity and specificity were calculated based on the mean and median values of the corresponding variables. A two‐sided p‐value of less than 0.05 was considered statistically significant. All analyses were performed using SPSS Statistics version 22 (IBM Corp., 2019), Minitab version 21.2 (Minitab LLC, 2022), and Microsoft Excel Office 2010.

3. Results

A total of 95 patients were included in the study. All participants underwent preoperative polysomnography (PSG) and drug‐induced sleep endoscopy (DISE), followed by adenotonsillectomy (AT). The cohort was 56.8% male. Age distribution was homogeneous, with a coefficient of variation (CV) below 50%. The mean age was 6.8 years (95% CI: 6.36 to 7.16 years). The median Apnea‐Hypopnea Index (AHI) and Obstructive Apnea Index (OAI) were 10.4 events/hour and 9.4 events/hour, respectively.

The relationships between sex, tonsil size (Brodsky scale), adenoid size, and OSA severity were examined. While no significant association was found between sex and OSA severity (p = 0.441), both tonsil size and adenoid hypertrophy demonstrated significant correlations. Compared to the non‐severe OSA group, patients with severe OSA had a greater proportion of Brodsky grade IV tonsils (40.7% vs. 16.7%) and severe adenoid obstruction (≥ 75% of the nasopharyngeal cavity) (44.1% vs. 16.7%) (Table A1). No significant correlations were identified between sex, age, AHI, and OAI (p = 0.650).

Analysis of dichotomized groups based on PSG and DISE data revealed a significant association between OSA severity, and the degree of upper airway obstruction (UAO) classified by the VOTE system. Endoscopic findings showed that obstruction at the velopharynx (p = 0.003) and tongue base (p = 0.036) were significantly associated with severe preoperative OSA, with specificities of 91% and 97.1%, and positive predictive values (PPV) of 87.7% and 90.9%, respectively. Furthermore, patients with severe OSA had a 63% higher relative risk (RR) for velopharyngeal obstruction and a 56% higher RR for tongue base obstruction (Table A2).

Analysis of the pattern of severe velopharyngeal obstruction revealed that concentric collapse was the most prevalent (n = 13 [54.2%]), followed by lateral (n = 7 [29.2%]) and anteroposterior (n = 4 [16.7%]) patterns.

To assess the association between OSA severity and UAO according to body weight, the Z‐score was used to determine weight‐for‐age percentile. A Z‐score > 2 (approximately the 97th percentile) was defined as indicative of obesity. Comparison of UAO between obese and normal‐weight children revealed a significant difference only at the tongue base: severe tongue base obstruction was observed in 5.4% of normal‐weight children, compared to 38.9% of obese children (p = 0.001) (Table A3). Obese patients also had a higher risk for severe OSA, with a mean AHI of 15.76 events/hour versus 9.80 events/hour in children with a normal weight‐for‐age (p = 0.024) (Table A4).

Regarding the sites of UAO, only the velopharynx and tongue base showed statistical significance on ROC curve analysis. The area under the curve (AUC) was 0.811 for the model comparing OAI with tongue base obstruction, indicating the strongest diagnostic performance for preoperative OSA severity (p = 0.001) (Figure 1). This corresponds to an 81% probability of correctly identifying severe OSA when significant tongue base obstruction is observed during DISE (Table A5). The optimal cutoff point for diagnosing severe OSA—defined as the threshold maximizing both sensitivity and specificity—was a mean OAI of 6.10 events/hour, yielding 100% sensitivity and 60.5% specificity.

Figure 1.

ROC curve analysis determining the optimal AIH cutoff value to predicts OSA Severity.

4. Discussion

This study analyzed 95 patients who underwent preoperative polysomnography and drug‐induced sleep endoscopy (DISE) to evaluate the relationship between upper airway obstruction (UAO) at the velopharyngeal and tongue base levels and the severity of (OSA). The demographic distribution revealed a male predominance (56.8%) and relative age homogeneity within the preschool range, consistent with the typical epidemiological profile of pediatric OSA [4].

Polysomnographic data showed a median Apnea–Hypopnea Index (AHI) of 10.4 events/hour and an Obstructive Apnea Index (OAI) of 9.4 events/hour, indicating a high prevalence of severe OSA in this population. No significant correlations were found between sex, age, AHI, or OAI (p = 0.650), suggesting that neither age nor sex are decisive predictors of OSA severity in this cohort. This finding aligns with recent literature emphasizing anatomical and physiological factors over demographic characteristics in pediatric OSA [3, 22, 23].

A significant association was observed between OSA severity and both tonsil size (Brodsky grade, p = 0.01) and the degree of adenoid obstruction (p = 0.006). Children with severe OSA had a markedly higher prevalence of Brodsky grade IV tonsils (40.7% vs. 16.7%) and severe adenoid obstruction (≥ 75% of the nasopharyngeal space) (44.1% vs. 16.7%) compared with non‐severe cases. These findings reinforce the ongoing value of static anatomical assessment during preoperative evaluation [24, 25].

Nevertheless, the strong and independent associations observed for dynamic velopharynx and tongue base collapse during DISE suggest that an integrated assessment—combining static anatomical and dynamic functional findings—provides superior risk stratification. This is consistent with emerging evidence that DISE identifies sites and patterns of obstruction often missed during awake examination [26, 27], particularly in complex pediatric cases where multilevel collapse contributes to persistent OSA [28].

In dichotomized analyses, obstruction at the velopharynx (p = 0.003) and tongue base (p = 0.036) were both significantly associated with OSA severity. Several authors have highlighted the frequent occurrence of multilevel upper airway obstruction, involving the velopharynx and lateral pharyngeal walls in approximately 35% of cases [22]. Combined velopharynx and tongue base collapse has been reported in up to 60% of patients, with major clinical implications, especially among younger children. Such findings are frequently detected during preoperative DISE and are strongly associated with higher AIH [29]. Following adenotonsillectomy, approximately 45.5% of patients may continue to experience oxygen desaturation related to tongue base obstruction, with a significant correlation (β = −0.73; p = 0.004) [30].

While the VOTE classification effectively identifies the primary level and degree of obstruction, it does not routinely differentiate between concentric, lateral, or anteroposterior collapse patterns. In our study, these specific collapse configurations were analyzed. Among 24 children with severe velopharyngeal obstruction, concentric collapse was the most prevalent pattern (54.2%), followed by lateral (29.2%) and anteroposterior (16.6%) collapse. This distribution aligns with previous pediatric DISE findings, where concentric collapse predominates in children with severe preoperative OSA [31, 32]. Although the VOTE system accurately categorizes obstruction level and severity, adding collapse‐pattern analysis provides a more refined airway phenotype. This enhanced phenotyping may improve prognostic accuracy and inform surgical decision‐making.

The (DISE) demonstrates the capacity to alter the surgical plan in up to 50% of cases, with multilevel obstruction identified in approximately 60% of these patients, confirming its critical utility in planning adjuvant procedures [29, 33]. Obesity is well‐established as the primary risk factor for residual (OSA) post‐(AT). The underlying pathophysiology—characterized by adipose deposition at the base of the tongue and pharynx—demands intervention beyond tonsillar removal to avoid insufficient treatment resulting from an incomplete diagnosis [34].

In this context, Rivero et al. [35] provide pivotal evidence regarding the safety and efficacy of lingual tonsillectomy for persistent pediatric OSA. When guided by DISE, this procedure can reduce the (AHI) by approximately 50%, with a positive surgical response observed in 70% of cases [36]. These collective findings robustly support the efficacy of DISE‐directed base‐of‐tongue surgery, motivating its incorporation into clinical guidelines [37, 38]. Furthermore, a comparative study demonstrated that Expansion Sphincter Pharyngoplasty in pediatric patients achieved surgical cure in 64% of cases post‐operatively, with greater AHI reduction compared to AT alone, demonstrating its safe and effective application for persistent lateral pharyngeal wall collapse in complex pediatric populations [39, 40]. Although conducting a randomized clinical trial comparing “AT versus AT + Pharyngoplasty + Lingual Tonsillectomy” in obese children presents logistical and ethical complexities, the convergence of available evidence provides a solid foundation for implementing a multilevel surgical approach.

Evidence from adult studies indicates that concentric velopharynx collapse predicts suboptimal outcomes following isolated palatal surgery, likely due to circumferential muscular involvement and dynamic airway instability [19, 41]. Although pediatric data remain limited, the high prevalence of concentric collapse in our cohort suggests that similar mechanisms may underlie persistent OSA following adenotonsillectomy. Recent pediatric investigations have emphasized that multilevel and concentric collapse patterns may predict incomplete surgical response, underscoring the potential need for adjunctive or staged interventions [37, 42]. Consequently, detailed endoscopic characterization may hold prognostic significance and should be incorporated into preoperative protocols, particularly for patients with high apnea–hypopnea indices or craniofacial comorbidities.

When comparing VOTE classification findings with polysomnographic diagnoses, severe OSA was observed in 63% of patients with even partial velopharynx collapse. Similarly, those with tongue base obstruction had a 56% higher risk of developing severe OSA, both with high specificity. Despite the widespread use of the VOTE system to describe DISE findings, variability persists in its predictive performance. Nonetheless, studies have demonstrated that total palatal collapse is associated with significantly higher AHI values [43]. Among children with persistent OSA after adenotonsillectomy, velopharynx collapse—especially with lateral wall involvement—was predominant. This pattern is often underestimated in awake evaluations and better delineated through DISE [29].

Weight‐for‐age assessment using Z‐scores identified values > 2 as indicative of obesity. A significant association was found between elevated Z‐scores and severe tongue base obstruction (p < 0.001), likely related to adipose tissue deposition in the tongue, as previously demonstrated by MRI studies in obese adults with OSA [44]. Comparison of Z‐score data with polysomnographic parameters revealed that AHI—but not OAI—was significantly higher in overweight children (mean 15.76) than in normal‐weight peers (mean 9.80; p = 0.024). This finding suggests that airflow limitation related to overweight status is common, and hypoventilation may be underrecognized. Obesity increases the risk of OSA fivefold in children (AHI ≥ 2) compared to those with normal weight (RR = 4.9; p = 0.0002) [45] and raises the likelihood of OSA persistence after adenotonsillectomy (AHI > 5) up to sevenfold (OR = 7.1; p = 0.001). Some of these patients may ultimately require postoperative CPAP therapy [46, 47]. In obese children, the correlation between tongue base collapse and OSA severity (as measured by AHI) appears particularly strong [48], reinforcing the multifactorial nature of pediatric OSA.

Receiver operating characteristic (ROC) curve analysis combining tongue base obstruction and OAI demonstrated high diagnostic accuracy, with an area under the curve (AUC) of 0.811, excellent sensitivity (100%), and moderate specificity (60.5%). The optimal diagnostic cutoff for OAI was 6.10 events/hour, approximating the threshold for moderate‐to‐severe OSA. This high AUC underscores the diagnostic value of DISE as a highly sensitive tool for predicting preoperative OSA severity, particularly in children with significant obstruction, and its role in identifying patients at risk of deep desaturation events [49]. Recent artificial intelligence models suggest that incorporating anatomical variables like those assessed during DISE could further enhance predictive accuracy, potentially approaching an AUC of 1.0 [50].

Collectively, these results demonstrate that sleep endoscopy is a valuable tool for predicting OSA severity, particularly when assessing obstruction at the velopharynx and tongue base levels.

4.1. Clinical Implications

The findings of this study reinforce the importance of a multidimensional approach in the preoperative evaluation of pediatric OSA. The high specificity observed highlights the utility of DISE in stratifying OSA severity. As the procedure can be safely performed by a trained otorhynolaryngologist under perioperative sedation, it represents a practical adjunct to routine evaluation. When integrated with polysomnographic and anthropometric data, DISE can aid in selecting candidates for adenotonsillectomy and in identifying patients who may benefit from additional interventions, such as tongue base reduction procedures.

4.2. Limitations and Future Directions

This study has certain limitations. The sample was drawn from a single tertiary center and included patients presenting with otorhynolaryngologic complaints and tonsillar hypertrophy, which may limit generalizability. The absence of postoperative follow‐up to evaluate adenotonsillectomy outcomes further restricts interpretation of surgical efficacy. Future studies with larger, multicenter cohorts and longitudinal follow‐up are warranted to validate these findings and refine prognostic criteria for surgical decision‐making. Additionally, long‐term studies evaluating whether specific collapse patterns—such as concentric velopharynx collapse—hold similar prognostic significance in children could inform more individualized surgical planning. Integration of additional assessment tools, such as the modified Cormack–Lehane score and the IPSES scale, may also provide valuable prognostic insights, particularly in cases of persistent OSA after adenotonsillectomy.

5. Conclusion

The results of this study showed that (DISE), using the VOTE classification scale, is an important predictor of pediatric OSA severity, especially when observed with velopharynx and tongue base obstruction. Combined to the DISE, obesity, a known risk factor for OSA severity and persistence after AT, the use of Z‐score proved to be a useful tool for weight‐for‐age assessment and its relationship with OSA severity. These findings reinforce the need for a personalized preoperative evaluation. An integrated approach combining polysomnography, sleep endoscopy, and anthropometric assessment may improve the risk stratification and optimize patient selection for different therapeutic strategies.

Author Contributions

Antônio Carlos Marão: conceptualization; methodology; validation; formal analysis; investigation; data curation; writing – original draft preparation; visualization; project administration. Neemias Santos Carneiro: methodology; validation; formal analysis; investigation; data curation; writing – original draft preparation; visualization. Renato Battistel Santana: validation; formal analysis; investigation; data curation; writing – original draft preparation; visualization. Camila de Castro Corrêa: methodology; validation; formal analysis; investigation; data curation; writing – original draft preparation; visualization. Silke Anna Theresa Weber: conceptualization; methodology; validation; formal analysis; investigation; data curation; writing – original draft preparation; visualization; supervision; project administration. All authors have read and agreed to the published version of the manuscript.

Ethics Statement

The study was submitted and approved by the local ethics committee (CAAE 42477014.2.0000.5411), all parents and participants older than 9 years of age signed a written consent.

Conflicts of Interest

The authors declare that there are no conflicts of interest related to this study. They have no financial, institutional, or personal relationships that could have influenced the work reported in this manuscript. No affiliations with any organization or publisher have biased the research.

Appendix 1. Tables

Table A1.

Association of gender, tonsil size, and adenoid size with OSA severity.

Characteristic	Severe OSA	Nonsevere OSA	p‐value
Gender, n (%)			0.441
Male	32 (54.2%)	22 (61.1%)
Female	27 (45.8%)	14 (38.9%)
Brodsky grade, n (%)			0.01*
Grade III	35 (59.3%)	30 (66.7%)
Grade IV	24 (40.7%)	6 (33.3%)
Adenoid obstruction, n (%)			0.006*
< 75% of cavum (nonsevere)	33 (25.4%)	30 (36.1%)
≥ 75% of cavum (severe)	26 (74.6%)	6 (63.9%)

Abbreviation: OSA, obstructive sleep apnea.

Table A2.

Characterization of OSA severity through description of obstruction levels (VOTE) and degrees of obstruction.

Obstruction levels	Degrees of obstruction	OSA severe	OSA no severe			Statistics
		N	%	N	%	p‐value	RR	S	SP	PV +	PV –
Velopharynx	Severe	21	35.6	3	8.3	0.003*	1.63 (1.18 −2.26)	35.6	91.7	87.5	46.5
	Not severe	38	64.4	33	91.7
Oropharynx/lateral pharyngeal wall	Severe	33	55.9	19	52.8	0.764	1.05 (0.76 – 1.44)	55.9	47.2	63.5	39.5
	Not severe	26	44.1	17	47.2
Tongue base	Severe	10	16.9	1	2.8	0.036*	1.56 (1.03 – 2.36)	16.9	97.2	90.9	41.7
	Not severe	49	83	35	97.2
Epiglottis	Severe	1	1.7	0	0	0.432	1.62 (0.49– 5.41)	1.7%	100%	100	38.3
	Not severe	58	98.3	36	100

Abbreviations: p, p‐value; PV, predictive value (%); RR, relative risk; S, sensitivity (%); SP, specificity (%).

Table A3.

Relationship between Z‐score and sonoendoscopy and airway obstruction points (severe/nonsevere).

Obstruction points		Nonobese children (Z‐score < 2)		Obese children (Z‐score ≥ 2)		p‐value
		N	%	N	%
Velopharynx	Severe	16	21.6%	6	33.3%	0.296
	Not severe	58	78.4%	12	66.7%
Oropharynx/lateral pharyngeal wall	Severe	39	52.7%	12	66.7%	0.285
	Not severe	35	47.3%	6	33.3%
Tongue base	Severe	4	5.4%	7	38.9%	< 0.001*
	Not severe	70	94.6%	11	61.1%
Epiglottis	Severe	1	1.4%	0	0%	0.620
	Not severe	73	98.6%	18	100%
	Not severe	73	98.6%	18	100%

Table A4.

Comparison between the Z‐score and polysomnographic parameters for pre‐TA severity (AHI and OAI).

		Media	Median	SD	N	CI	p‐value
AHI pre	< 2	11.44	6.0	8.25	74	1.88	0.024*
	≥ 2	15.76	10.4	14.83	18	6.85
Pre OAI	< 2	9.80	4.7	16.34	71	3.80	0.417
	≥ 2	14.99	9.4	17.24	18	7.96

Abbreviations: AHI, apnea hypopnea index; CI, confidence interval; N, population; OAI, obstructive apnea index; SD, standard deviation; TA: tonsillectomy and adenoidectomy

Table A5.

Area of the ROC curve for severity diagnosis, comparing polysomnographic data, and VOTE scale.

Obstruction points	AHI pre							OAI pre
	(AUC)	p‐value	LL	UL	S (%)	E (%)	M	(AUC)	p‐value	LL	Ul	S (%)	E (%)	M
Velopharynx	0.689	0.006	0.566	0.812	75.0	69.0%	8.9	0.769	< 0.001	0.656	0.881	77.3	71.4	7.60
Oropharynx/lateral pharyngeal wall	0.583	0.164	0.468	0.698	—	—	—	0.567	0.269	0.449	0.685	—	—	—
Tongue base	0.790	0.002	0.643	0.936	81.8%	79.8%	14.65	0.811*	0.001	0.720	0.903	100.0	60.5	6.10*
Epiglottis	0.883	0.189	0.818	0.948	—	—	—	0.681	0.534	0.586	0.777	—	—	—

Abbreviations: A, area under the curve; AHI, apnea hypopnea index; AUC, area under the curve/*, Represents the best index; E, specificity; LI, lower LI, lower limit; M, media; OAI, obstructive OAI, obstructive apnea index; p, p‐value; S, sensitivity; UL, upper limit.

Carneiro N. S., Marão A. C., Santana R. B., Corrêa C. d. C., and Weber S. A. T., “Drug‐Induced Sleep Endoscopy in Children Predicts OSA Severity,” World Journal of Otorhinolaryngology ‐ Head and Neck Surgery 0 (2026): e70098. 10.1002/wjo2.70098.

Data Availability Statement

Data are not available due to privacy restrictions.