Synchronous Airway Lesions in Children with Obstructive Sleep Apnea

Abstract

Objective:

Drug-induced sleep endoscopy is established for evaluating the upper airway in pediatric obstructive sleep apnea (OSA), but the role of comprehensive lower airway assessment remains uncertain. We aimed to determine whether synchronous airway lesions identified during microlaryngoscopy and bronchoscopy (MLB) are associated with OSA severity in children undergoing adenotonsillectomy (AT).

Study Design:

Retrospective cohort study.

Setting:

Tertiary care academic children’s hospital.

Methods:

We reviewed MLB findings documented by three fellowship-trained pediatric otolaryngologists during ATs for OSA between January 2019 and December 2023. Children with craniofacial abnormalities, genetic syndromes, or tracheostomy dependence were excluded. Associations between airway findings, demographics, polysomnography, and surgical management were analyzed.

Results:

Among 117 children (mean age 3.5 years, SD 2.2), 63.2% were male, 54.7% were Black, and 76.1% had Medicaid insurance. Synchronous airway lesions were present in 44 (37.6%). Laryngomalacia-related abnormalities were most common (19.7%), followed by mucosal cobblestoning (5.1%). Children with synchronous airway lesions were more likely to have asthma (P=0.02), but no associations were observed with age, race, insurance status, or prematurity. Notably, no patient required intervention based on MLB findings, and surgical management was unchanged. Preoperative polysomnographic parameters—including apnea-hypopnea index (30 vs. 26.7 events/hour, P=0.39), oxygen saturation nadir (81.0% vs. 78.6%, P=0.22), and sleep efficiency (77.5% vs. 74.5%, P=0.32)—did not differ between groups.

Conclusion:

Secondary airway lesions were common but did not correlate with OSA severity or alter management. These findings suggest many are incidental, supporting a selective, individualized approach to MLB use in pediatric OSA pending evidence-based guidelines.

Introduction

Obstructive sleep apnea (OSA) affects 1–5% of children and is associated with recurrent upper airway obstruction during sleep, leading to fragmented sleep and intermittent hypoxemia [1,2]. Untreated pediatric OSA is associated with neurocognitive morbidity [3], behavioral problems [4], and poor quality of life [5].

Adenotonsillar hypertrophy is the predominant anatomic contributor to pediatric OSA, accounting for the majority of cases in otherwise healthy children [6]. This pathophysiology supports adenotonsillectomy (AT) as the first-line surgical intervention, with success rates of 65–80% in achieving OSA resolution or significant improvement [7]. Despite the general effectiveness of AT, treatment failure or incomplete resolution occurs in 20–35% of otherwise healthy children, with significantly higher rates in children with obesity, craniofacial abnormalities, or genetic syndromes [8]. Recognition that multiple anatomic sites may contribute to upper airway obstruction has led to increased utilization of drug-induced sleep endoscopy (DISE) to evaluate dynamic upper airway collapse patterns [9,10]. Importantly, current DISE protocols and scoring systems focus primarily on upper airway evaluation from the nasopharynx to the supraglottis. However, the potential contribution of lower airway pathology, including subglottic, tracheal, and bronchial abnormalities, to pediatric OSA remains less well characterized.

Microlaryngoscopy and bronchoscopy (MLB) comprehensively evaluates both upper and lower airway structures under controlled conditions, potentially identifying synchronous lesions that may influence OSA severity or treatment outcomes. Few studies examining the role of comprehensive airway endoscopy in pediatric OSA demonstrate wide variability in synchronous airway lesions in pediatric OSA [11–14]. Only one study thus far has examined children assessed for OSA severity with respiratory polygraphy; however the sample size (n=16 with polygraphy data) with comorbidities such as Down syndrome precluded meaningful stratification of airway findings in typically developing children [15]. Thus, while conditions such as tracheomalacia and subglottic stenosis can theoretically contribute to OSA, the relationship between MLB findings and OSA severity measures has not been well established raising questions about the routine incorporation of MLBs into the evaluation of pediatric OSA in otherwise healthy children.

To address this gap in knowledge, we aimed to evaluate the prevalence, characteristics, and clinical significance of secondary airway lesions identified during MLB in children undergoing surgical management for OSA. Specifically, we sought to determine: (1) the frequency of upper and lower airway abnormalities identified on an MLB, (2) the relationship between endoscopic findings and polysomnographic measures of OSA severity, and (3) the clinical impact of MLB results on surgical management decisions. We hypothesize that while synchronous airway lesions may be identified in a significant proportion of children undergoing MLB, most represent incidental findings that do not correlate with OSA severity measures or require therapeutic intervention, suggesting limited clinical utility for routine lower airway evaluation in pediatric OSA management.

Materials and Methods

We conducted a retrospective cohort study of children younger than 18 years who underwent an MLB between January 1, 2019, and December 31, 2023, at the University of Maryland Children’s Hospital, a tertiary care academic medical center in Baltimore, Maryland. This study was approved by the University of Maryland Baltimore Institutional Review Board with waiver of informed consent due to the retrospective nature.

Inclusion criteria comprised age less than 18 years at time of surgery, MLB performed during surgical management for OSA, available preoperative polysomnographic data within 12 months of surgery, and complete medical records including operative notes. Children were excluded if they had craniofacial abnormalities or genetic syndromes such as Down syndrome, tracheostomy dependence, MLB performed for indications other than OSA evaluation, or previous airway surgery.

The decision to perform MLB was made by the attending pediatric otolaryngologist based on clinical judgment. Common indications included severe OSA defined as apnea-hypopnea index ≥30 events per hour, age less than 2 years, or clinical suspicion of synchronous airway lesions. All MLB procedures were performed concurrently with planned AT.

Sample size estimation was performed using Cochran’s formula for prevalence studies based on an expected prevalence of secondary airway lesions of 40% and desired precision of 10%.[16] This yielded a minimum required sample size of 93 patients. To account for potential exclusions and missing data, we aimed to include at least 120 patients.

Three fellowship-trained pediatric otolaryngologists of experience performed all MLB procedures using a standardized protocol. After induction of general anesthesia, patients were positioned supine. Age-appropriate laryngoscopes, either Parsons or Phillips models, were used for exposure, and systematic evaluation was performed using 2.7 mm or 4.0 mm rigid telescopes (Karl Storz, Tuttlingen, Germany). Topical 2% lidocaine solution was applied to the vocal cords prior to instrumentation. Patients were maintained on spontaneous ventilation with supraglottic oxygen insufflation to assess dynamic airway behavior. The systematic evaluation protocol included assessment of the supraglottis for epiglottic morphology, aryepiglottic folds, and supraglottic collapse patterns, followed by examination of the glottis for vocal cord mobility, appearance, and lesions. The subglottis was assessed circumferentially for stenosis, cysts, or other abnormalities, and the trachea was evaluated for malacia, compression, or intrinsic lesions. Finally, bilateral mainstem bronchi were examined for proximal bronchial anatomy and patency.

Following endoscopic evaluation, patients were intubated with age-appropriate endotracheal tubes for concurrent surgical procedures. All endoscopic findings were documented by the operating surgeon immediately post-procedure. Demographics and clinical characteristics collected included age at surgery, sex assigned at birth, race and ethnicity as self-reported by caregivers, age-adjusted body mass index (BMI) percentile, gestational age at birth with full-term defined as ≥37 weeks and preterm as <37 weeks, and insurance status. Medical history variables included asthma diagnosis, previous intubation, and other comorbidities. Physical examination findings documented were tonsil size grading using the Brodsky scale [17] and adenoid hypertrophy grading based on the severity of nasopharyngeal airspace obstruction [18].

Polysomnographic parameters recorded in an accredited sleep lab included total sleep time in minutes, sleep efficiency percentage, sleep stages as percentages of N1, N2, N3, and rapid eye movement (REM) sleep, apnea hypopnea index (AHI) as events per hour, oxygen saturation nadir (SpO₂) percentage, time with SpO₂ <90% in minutes, arousal index as events per hour, peak end-tidal CO₂ (EtCO₂) in mmHg, and time with EtCO₂>50 mmHg in minutes. All of these measures followed standard pediatric OSA criteria [19].

Synchronous airway lesions were defined as any structural abnormality beyond adenotonsillar hypertrophy that could potentially contribute to OSA. Findings were categorized as supraglottic including laryngomalacia, glottic lesions such as vocal cord paralysis, subglottic abnormalities including stenosis, tracheal findings such as malacia, and anatomic variants including accessory bronchi.

Statistical analyses were performed using GraphPad Prism version 10.0.0 (GraphPad Software, San Diego, CA). Descriptive statistics included means with 95% confidence intervals and frequencies with percentages for categorical variables. The primary outcome was presence of secondary airway lesions identified on MLB as a binary variable. Continuous variables were compared using independent samples t-tests for normally distributed data and Mann-Whitney U tests for non-normally distributed data, while categorical variables were analyzed using chi-square tests or Fisher’s exact tests when expected cell counts were less than 5. We evaluated associations between MLB findings and patient demographics including age, race, and insurance status, clinical characteristics, comorbidities, polysomnographic parameters including AHI, oxygen saturation nadir, and sleep efficiency. 95% Confidence intervals were calculated for all point estimates using appropriate methods for the data distribution. Complete case analysis was performed for all comparisons. A two-tailed P value <0.05 was considered statistically significant, with no adjustments made for multiple comparisons given the exploratory nature of this study.

Results

During the study period from 2019 to 2023, a total of 726 pediatric MLBs were performed, of which 139 were indicated for OSA. After applying exclusion criteria, 117 patients comprised the final study cohort. Six cases were excluded due to lack of preoperative polysomnographic data, eight patients were excluded for genetic syndromes (six with Down syndrome, one with Prader-Willi syndrome, and one with hypochondroplasia), and an additional eight patients were excluded for tracheostomy dependence.

Patient demographics are summarized in Table 1. The mean age was 3.5 years (SD 2.2), with 74 (63.2%) boys. Most children identified as Black (64, 54.7%) or White (37, 31.6%). Most patients (89, 76.1%) were Medicaid beneficiaries. Birth history showed that 79 (67.5%) children were born full-term. The most common comorbidity was asthma, affecting 36 (30.8%) children, followed by history of prior intubation in 27 (23.1%).

Table 1.

Demographic characteristics of children included in the current study. Most variables are categorical (%) except age.

Variable	N
Sex
Male	74 (63.2)
Female	43 (36.8)
Mean age in years (standard deviation)	3.5 (2.2)
Race and ethnicity
White	37 (31.6)
Black or African American	64 (54.7)
Latino or Hispanic	11 (9.4)
Asian	1 (0.85)
Other or mixed race	4 (3.4)
Insurance class
Private	25 (21.3)
Medicaid	89 (76)
Uninsured	2 (1.7)
Birth history
Full term	79 (67.5)
Preterm	34 (29.1)
Unknown	4 (3.4)
Medical history
Asthma	36 (30.8)
Prior intubation	27 (23.1)

Synchronous airway lesions were identified in 44 of 117 (37.6%) (Table 2). Since all patients were diagnosed with OSA, adenotonsillar hypertrophy was not considered abnormal. The most common abnormalities were findings consistent with laryngomalacia or related supraglottic abnormalities, identified in 23 (19.7%). These findings included shortened aryepiglottic folds, omega-shaped epiglottis, aryepiglottic fold tethering, and arytenoid prolapse on inspiration. The second most frequent finding was mucosal cobblestoning, present in 6 (5.1%) patients. Other identified lesions included tracheomalacia in 3 (2.6%) patients, vocal nodules in 3 (2.6%) patients, and vascular compression in 2 (1.7%) patients. Uncommon findings included laryngotracheomalacia, anterior glottic web, type I laryngeal cleft, an accessory bronchus, grade 1 subglottic stenosis, subglottic cyst, and vocal fold irregularities, each occurring in 1 (0.9%) patient.

Table 2.

Airway findings on microlaryngoscopy and bronchoscopy listed as number (proportion) of the cohort. All findings are mutually exclusive unless specified.

Airway abnormality	N (%)
Laryngomalacia	23 (19.7)
Mucosal cobblestoning	6 (5.1)
Tracheomalacia	3 (2.6)
Vocal nodules	3 (2.6)
Vascular compression	2 (1.7)
Aryepiglottic fold tethering	1 (0.85)
Arytenoid prolapse on inspiration	1 (0.85)
Anterior glottic web	1 (0.85)
U-shaped epiglottis	1 (0.85)
Tubular-shaped epiglottis	1 (0.85)
Laryngeal cleft, type 1	1 (0.85)
Laryngomalacia and tracheomalacia	1 (0.85)
Right accessory bronchus	1 (0.85)
Subglottic stenosis, grade 1	1 (0.85)
Subglottic cyst	1 (0.85)
Vocal cord atrophy	1 (0.85)

Analysis of associations between patient characteristics and endoscopic findings revealed that children with positive MLB findings were more likely to have asthma (χ²=5.56, P=0.02). No significant associations were found between airway findings and patient age, race, insurance status, history of prematurity, or history of intubation. No patient required surgical intervention based on MLB findings, and the results did not alter planned surgical procedures or management in any child.

The mean pre-operative AHI for the cohort was 28.7 events per hour (95% CI, 25.1–32.4), with 51 children (43.6%) demonstrating very severe OSA defined as AHI >30 events per hour.[20] Additional pre-operative sleep study parameters included a mean total sleep time of 341.6 minutes (326.5–356.7), sleep efficiency of 76.4% (73.5–79.3), and SpO₂ nadir of 80.0% (78.2–81.6).

Physical examination findings showed considerable variation in tonsil size and adenoid hypertrophy. Among the 110 patients with documented tonsil grading, the most common tonsil size was 3+ in 51 patients (46.4%). For adenoid hypertrophy assessment, the most frequent degree was 50% hypertrophy in 27 patients (25.2%).

Comparison of preoperative polysomnographic parameters between children with and without identified airway lesions showed no significant differences (Table 3). Postoperative polysomnographic data were available for 18 patients (15.3% of the cohort). Among these patients, eight had airway abnormalities during MLB. Preoperatively, children with positive endoscopic findings had AHI of 30 events/hour (95% CI, 25.0–35.0) compared to 26.7 events/hour (21.3–32.1) in those without findings (P=0.39). Children with MLB findings suggestive of laryngomalacia had a mean AHI of 32.3 (25.9–38.7) and the mean SpO₂ nadir of 78.3% (74.4–82.2). Children with and without airway anomalies on the MLB showed similar improvements in AHI of 9.2 events/hour (−1.1–19.3) vs. 16.1 events/hour (4.4–28.6), although limited in sample size. This trend extended to other parameters, such as SpO₂ nadir, which also showed similar improvements after surgical intervention (88.8% [82.4–95.2] vs. 83.6% [70.8–96.3]).

Table 3.

Polysomnography parameters of patients with and without airway abnormalities during microlaryngoscopy and bronchoscopy (MLB). Abbreviations: EtCO₂, End-tidal carbon dioxide SpO₂, Oxygen saturation.

Polysomnographic parameter	Airway anomaly on MLB (n=44)	No findings on MLB (n=73)
Total sleep time, min	328.9 (302.8–355)	349.5 (330.8–368.2)
Sleep efficiency, %	74.5 (69.6–79.3)	77.5 (73.8–81.2)
Rapid eye movement sleep duration, %	10.1 (7.3–12.2)	10.1 (7.6–12.6)
Sleep stage N1 duration, %	1.3 (0.7–2)	1.5 (0.8–2.3)
Sleep stage N2 duration, %	45.2 (40.1–50)	50 (45.3–54.7)
Sleep stage N3 duration, %	24.6 (18.6–29.1)	24.6 (21.4–27.8)
Arousal index	10.4 (7.1–13.8)	12.7 (9.5–16)
Apnea hypopnea index	26.7 (21.3–32)	30 (25–35)
Sleep time with SpO2 <90%, min	14.5 (4.3–24.8)	11 (5–17.1)
SpO2 nadir, %	78.6 (75.5–81.8)	80.9 (78.7–83.1)
Mean EtCO2, mmHg	39.3 (36.4–42.3)	40.1 (37.8–42.5)
Peak EtCO2, mmHg	51.8 (48–55.6)	52.6 (50.2–55.1)
Time spent with EtCO2 >50 mm Hg, min	19 (1.5–36.6)	19.2 (4.7 to 33.7)

Discussion

In our retrospective cohort of children with OSA undergoing MLB, synchronous airway lesions were identified in 37.6% of patients, most commonly related to laryngomalacia. However, these lesions did not correlate with polysomnographic measures or alter management. These observations raise questions about the clinical utility of routine comprehensive airway endoscopy in pediatric OSA and support emerging evidence questioning the routine use of comprehensive airway evaluation in pediatric OSA.

Our findings are consistent with recent literature reporting variable prevalence (5–77%) but overall similar clinical insignificance of synchronous airway lesions in pediatric OSA. Bliss et al. observed that concurrent MLB with drug-induced sleep endoscopy (DISE) provided additional positive findings in patients with history of intubation, prematurity, or genetic/neurologic comorbidities. The age-dependent patterns observed in our study also align with findings from Mandell and Yellon [13], Dritsoula et al. [15], and Argyriou et al. [21] who reported high rates of airway abnormalities in young children (59%–77%) but found that these lesions seldom affected treatment outcomes or the need for intervention. Collectively, these studies and our observations indicate that the high prevalence of secondary airway findings represent incidental findings without clear pathophysiologic relevance rather than clinically significant OSA risk. This is further supported by the lack of differences between polysomnographic parameters in children with and without airway lesions in our cohort. Even when considering laryngomalacia, which can theoretically contribute to sleep-disordered breathing through dynamic airway collapse, the relationship between static endoscopic findings and functional airway obstruction during sleep is complex [22]. Although tracheomalacia and other intrathoracic lesions may only contribute to overall respiratory physiology and sleep-related breathing dynamics rather than OSA, these findings are likely to be incidental in nature. The absence of functional assessment represents a fundamental limitation of rigid endoscopy in evaluating sleep-related airway obstruction, as static anatomic findings may not predict dynamic collapse patterns that contribute to OSA.

The socioeconomic and demographic characteristics of our study population reflect the community served by our urban academic medical center, which includes a high proportion of Medicaid beneficiaries (76.1%) and Black children (54.7%). While these demographics align with prior literature describing socioeconomic factors associated with pediatric sleep disorders [26] our single-center, retrospective design limits the ability to draw conclusions about broader healthcare disparities.

This study had some limitations. The retrospective design introduces inherent bias, and the tertiary care setting may have resulted in a cohort with more severe disease than typically encountered in community practice. Our mean preoperative AHI of 28.7 events per hour represents severe OSA, limiting generalizability to children with milder disease. Selection bias resulting from the tertiary care referral pattern and surgeon-dependent decision-making further constrains the generalizability of these findings. Despite standardized surgical protocols, the subjective nature of endoscopic interpretation could introduce variability, particularly for subtle abnormalities such as mild laryngomalacia or mucosal changes. The absence of dynamic evaluation during the endoscopic assessment represents a significant limitation, as static findings may not accurately reflect the pathologic collapsibility patterns of OSA. The lack of a standardized approach to selecting candidates for MLB means our findings represent an exploratory cohort rather than a systematically evaluated population, limiting the strength of conclusions regarding clinical utility. Additionally, limited postoperative polysomnography precludes evaluation of treatment outcomes. Despite these constraints, the absence of clinically actionable findings across a diverse cohort provides valuable insight into optimizing the role of MLB in pediatric OSA workup.

The resource implications of routine MLB must also be considered in the context of our findings. These procedures require additional anesthesia time, specialized equipment, and fellowship-trained personnel, representing significant healthcare costs. In an era of increasing focus on value-based care and resource optimization, the absence of actionable findings in our cohort raises questions about cost-effectiveness. Future investigations should focus on developing evidence-based criteria for patient selection, potentially incorporating factors such as age, OSA severity, previous treatment failure, or specific clinical indicators that may predict higher likelihood of clinically significant findings.

Based on our findings, we propose that MLB be reserved for children with specific indicators, such as clinical suspicion for lower airway abnormalities or children with complex medical histories. This individualized approach requires prospective multicenter studies to validate appropriate selection criteria, which could include combining techniques that leverage the strengths of both dynamic upper airway evaluation, such as DISE, and comprehensive anatomic assessment. Ultimately, this will optimize resource utilization while ensuring that children most likely to benefit from comprehensive airway evaluation receive appropriate assessment.