A longitudinal cohort study of sleep-disordered breathing in children: interrelationships with craniofacial morphology, malocclusion, and intervention outcomes

Abstract

Objective

This prospective longitudinal cohort study aimed to investigate the prevalence, risk factors, and craniofacial correlates of sleep-disordered breathing (SDB) in children, and to evaluate the efficacy of targeted interventions in a nested sub-study.

Methods

A total of 1,589 children (aged 6–9 years) were enrolled and followed for 36 months. Assessments included Portable polysomnography, cone-beam computed tomography, 3D facial imaging, and standardized clinical examinations. A sub-study of 275 children with moderate-to-severe SDB (obstructive apnea–hypopnea index, OAHI ≥ 5) were allocated to ENT intervention (adenotonsillectomy), orthodontic intervention (rapid maxillary expansion), or control groups.

Results

The baseline prevalence of SDB was 28.9%, with a significant male predominance. Multivariate analysis identified adenotonsillar hypertrophy (aOR = 3.42), paternal snoring (aOR = 2.38), obesity (aOR = 2.15), and allergic rhinitis (aOR = 1.86) as key independent risk factors. Children with SDB exhibited distinct craniofacial features, including reduced airway volume, mandibular retrusion, and a 38.2% smaller minimum cross-sectional area. Strong associations were found between SDB and specific malocclusions, including Class II malocclusion (48.6% vs. 27.3%), increased overjet, and posterior crossbite. A bidirectional relationship was observed, where SDB persistence worsened occlusal traits, while SDB remission facilitated spontaneous improvement. In the intervention sub-study, both ENT (68.4% resolution) and orthodontic (52.1% resolution) groups showed significant improvement in OAHI and quality of life scores compared to controls (15.7%). Cost-effectiveness analysis favored the ENT intervention while acknowledging the additional occlusal benefits of orthodontic treatment.

Conclusion

SDB is highly prevalent and dynamically interacts with craniofacial growth and occlusal development in children. An interdisciplinary approach, incorporating both ENT and orthodontic perspectives, is supported for effective management, leading to improved respiratory, dentofacial, and quality-of-life outcomes.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12903-026-07791-z.

Introduction

Sleep-disordered breathing (SDB) in children represents a spectrum of conditions characterized by abnormalities of respiratory pattern and gas exchange during sleep, ranging from primary snoring to obstructive sleep apnea [1]. The prevalence of pediatric SDB varies considerably across populations, with recent studies reporting rates ranging from 7% to over 27% in different geographical regions [2]. This condition has gained increasing attention due to its association with significant morbidity, including neurocognitive impairment, behavioral problems, cardiovascular complications, and growth disturbances. The profound impact of SDB on children's quality of life and long-term health outcomes underscores the importance of early identification and intervention [3].

The pathophysiology of pediatric SDB is multifactorial, involving complex interactions between anatomical, neuromuscular, and inflammatory factors [4]. Adenotonsillar hypertrophy remains the most recognized risk factor, particularly in younger children, as the lymphoid tissue occupies a substantial proportion of the upper airway space. However, the role of craniofacial anatomy has emerged as equally crucial in understanding SDB pathogenesis [5]. Certain craniofacial features, including mandibular retrusion, maxillary constriction, and altered vertical facial dimensions, can reduce upper airway patency and contribute to airway collapse during sleep [6, 7]. The relationship between craniofacial morphology and SDB has been explained through Moss's functional matrix theory [8], which proposes that alterations in soft tissue function and breathing patterns can influence craniofacial growth and development.

The potential interconnection between SDB and malocclusion presents a particularly compelling area of investigation. Numerous studies have attempted to elucidate this relationship, though with inconsistent findings. Some investigations have reported significant associations between SDB and specific occlusal features, including increased overjet, anterior open bite, posterior crossbite, and Class II malocclusion. These dental characteristics may reflect underlying craniofacial patterns that predispose to airway compromise. However, other studies have failed to establish consistent relationships, highlighting the methodological challenges and multifactorial nature of both conditions. The existing literature is further limited by the predominance of cross-sectional designs, which cannot establish temporal sequence or causal relationships.

The Chinese population presents a unique context for examining these relationships. Recent decades have witnessed remarkable changes in lifestyle, environmental exposures, and healthcare access in China, with parallel increases in both allergic diseases and malocclusion prevalence [9]. Studies specifically investigating SDB in Chinese children remain relatively limited, and existing research has primarily relied on questionnaire-based assessments rather than objective polysomnographic measures. The study by Li et al. [10] in Shanghai reported a 17.7% prevalence of SDB among 6–11-year-old children and identified allergic rhinitis, adenotonsillar hypertrophy, and parental snoring as significant risk factors. While this study provided valuable epidemiological data, it also highlighted the need for more comprehensive investigations incorporating objective SDB measures and detailed craniofacial assessments.

Several critical gaps persist in our current understanding of the SDB-malocclusion relationship. First, the temporal sequence and potential bidirectional nature of this association remain unclear—does SDB predispose to malocclusion through altered breathing patterns and oral posture, or do specific craniofacial features underlying malocclusion predispose to SDB? Second, the relative contribution of modifiable versus non-modifiable risk factors in different populations requires further elucidation. Third, there is a pressing need to identify which children with malocclusion would benefit most from SDB screening and which children with SDB might require orthodontic evaluation.

This study aims to address these knowledge gaps through a comprehensive longitudinal investigation of children in Zhengzhou, China. By employing Portable polysomnography, detailed craniofacial imaging, and standardized malocclusion assessments, we seek to determine the prevalence and progression of SDB, identify independent risk factors, and elucidate the temporal relationships between SDB development and craniofacial growth patterns. Furthermore, through a nested intervention sub-study, we aim to evaluate the effects of targeted treatments on both respiratory and dentofacial outcomes. The findings from this research have the potential to inform screening protocols, guide interdisciplinary management, and ultimately improve the health and well-being of children affected by SDB and related malocclusions.

Methods

Ethnic consideration

This study employed a prospective longitudinal cohort design with a nested intervention sub-study to comprehensively investigate the relationships between pediatric SDB, craniofacial development, and malocclusion. The study protocol received approval from the Institutional Review Board of The First Affiliated Hospital of Zhengzhou University (Approval No: 2019-XY-001). Participant recruitment occurred from January 2020 to December 2021 through multiple channels including school-based screening in 20 randomly selected primary schools from urban and suburban districts of Zhengzhou, clinical recruitment from pediatric dental and ENT clinics, and community health centers during regular well-child visits. Inclusion criteria encompassed children aged 6–9 years at baseline, parental written informed consent along with child assent, and residence in Zhengzhou with no plans to relocate within three years. Exclusion criteria included previous adenotonsillectomy or orthodontic treatment, syndromic craniofacial abnormalities, severe neurological disorders, and inability to complete study procedures.

Ethical considerations included detailed informed consent procedures in Mandarin Chinese, optimized cone-beam computed tomography protocols adhering to ALARA principles for radiation safety, encrypted data storage with limited access to ensure confidentiality, and appropriate compensation for participants including transportation reimbursement and summary reports for families. This comprehensive methodological approach ensured rigorous investigation of the complex relationships between SDB, craniofacial development, and malocclusion in children while maintaining high ethical standards throughout the research process. Informed consent to participate was obtained from the parents or legal guardians of any participant under the age of 16, and all the procedures were in compliance with the Helsinki Declaration.

Sample size calculation

The sample size calculation determined that 1,500 participants would provide 90% power to detect a minimum odds ratio of 1.5 for primary risk factors, accounting for 20% attrition over three years. This calculation was based on SDB prevalence of 17.7% and an assumed 15% progression rate to clinically significant SDB. Data collection followed a structured timeline with comprehensive assessments at baseline and annual follow-ups for three years, including demographics, physical examinations, questionnaires, Portable polysomnography, 3D imaging, and clinical oral evaluations.

Outcome measure

SDB assessment incorporated both objective and subjective measures. Portable polysomnography using Nox A1s devices provided objective data including EEG, EOG, chin EMG, ECG, nasal pressure, thoracoabdominal movements, pulse oximetry, and body position, scored according to American Academy of Sleep Medicine pediatric criteria (Version 2.4, 2017) [11]. The primary outcomes included Obstructive Apnea–Hypopnea Index (OAHI) and oxygen desaturation index (ODI), severity was categorized into three levels: mild SDB for an OAHI of 1–4.9, moderate SDB for an OAHI of 5–9.9, and severe SDB for an OAHI of 10 or higher. Subjective assessment utilized the Chinese validated version of the Pediatric Sleep Questionnaire and the Sleep Clinical Record to evaluate craniofacial features and physical examination findings.

The 3dMDface System (3dMD LLC, Atlanta, GA, USA) was used to capture three-dimensional facial images under standardized conditions: participants were seated with natural head position (self-balanced posture with visual focus on a distant eye-level target), relaxed lip posture, and neutral facial expression. Images were acquired in a controlled environment with consistent lighting and background. A total of 27 soft-tissue landmarks (Supplementary Table 1) were identified by a single trained examiner using 3dMDvultus software (version 3.1.0). These landmarks were used to derive measurements of facial convexity, mandibular retrusion (through sagittal jaw relationship indices), and vertical facial proportions (lower facial height ratio) [12, 13].

Cone-beam computed tomography (CBCT) was performed using i-CAT FLX devices (Imaging Sciences International, Hatfield, PA, USA) with a standardized protocol: 120 kVp, 5 mA, 8.9 s scan time, and 0.25 mm voxel size. Scans were acquired with participants in maximum intercuspation and lips at rest. DICOM files were imported into Dolphin Imaging software (version 11.9, Dolphin Imaging & Management Solutions, Chatsworth, CA, USA) for analysis. Airway volume measurements were segmented for the nasopharynx, oropharynx, and hypopharynx using threshold-based region growing. The minimum cross-sectional area was identified through sequential axial slicing. Cephalometric analysis included sagittal skeletal relationships (SNA, SNB, ANB), vertical dimensions (MP-SN, FMA), direct pharyngeal linear dimensions (PAS, SPAS), hyoid bone position (MP-H), and craniocervical inclination.

Malocclusion assessment combined clinical examinations with digital analysis. Five calibrated orthodontists conducted clinical oral examinations with inter-class correlation coefficients exceeding 0.85 for all measurements, evaluating Angle's classification, overjet, overbite, crossbite patterns, open bite, crowding and spacing, and maxillary constriction. Intraoral scanning using TRIOS 4 devices created digital dental models analyzed through 3Shape Ortho Analyzer software for comprehensive tooth size-arch length analysis, Bolton analysis, and arch form assessment.

Risk factor assessment included detailed medical history and physical examination components. Allergic rhinitis diagnosis combined ISAAC questionnaire data with ENT confirmation, while adenotonsillar hypertrophy was graded using the Brodsky scale. Additional medical factors included physician-diagnosed asthma, anthropometric measurements using WHO standards, and neck circumference measurement at thyroid cartilage level. Family history and environmental factors encompassed parental SDB assessment using STOP-Bang questionnaires, tobacco exposure evaluation through urinary cotinine levels, sleep environment characterization, and retrospective feeding history documentation.

Quality of life and behavioral assessment incorporated validated instruments including the Chinese version of Pediatric Quality of Life Inventory for physical, emotional, social, and school functioning domains. Neurobehavioral function was evaluated using Conners 3-Parent Scale for attention and executive function, Pediatric Daytime Sleepiness Scale, and standardized school performance metrics with appropriate permissions.

The intervention sub-study allocated children diagnosed with moderate-to-severe SDB to one of three groups, taking into account both parental preference and clinical recommendation. The ENT intervention group received evaluation within one month and adenotonsillectomy when indicated, with post-operative Portable polysomnography at six months. The orthodontic intervention group underwent rapid maxillary expansion using tooth-borne hyrax expanders, suitable for the mixed dentition stage. The protocol included an active expansion phase followed by a 6-month retention period. For participants with significant Class II malocclusion and mandibular retrusion, functional appliance therapy could be incorporated based on individual clinical assessment and phenotypic traits. In conjunction, all participants in this group received standardized counseling on nasal breathing promotion, lip seal, and basic tongue posture exercises. However, a formalized, intensive myofunctional therapy program was not implemented as a core protocol component. The control group practiced watchful waiting with sleep hygiene education and repeated Portable polysomnography at six months, with cross-over options if no improvement occurred.

Quality control measures included extensive examiner training with an eight-hour workshop using standardized patients, monthly calibration sessions targeting inter-examiner reliability above 0.80, and blinding of examiners to participants' SDB status and group assignment. Data management utilized Research Electronic Data Capture system with double data entry and random verification processes, while an independent committee conducted data quality reviews every six months. Portable polysomnography quality assurance involved pre-study device calibration, bi-annual maintenance, manual scoring by two certified sleep technologists, and third-party resolution of scoring discrepancies.

Statistic analysis

Statistical analysis plans incorporated linear mixed models to examine SDB progression and craniofacial growth trajectories, structural equation modeling to test causal pathways between risk factors, SDB, and malocclusion, and generalized estimating equations to account for within-subject correlations. Secondary analyses included machine learning approaches using random forests and XGBoost for SDB prediction models, mediation analysis to examine airway dimensions as mediators between SDB and malocclusion, and latent class analysis to identify SDB endophenotypes. Intervention analysis followed intention-to-treat principles including all participants, with per-protocol analysis as a secondary approach and cost-effectiveness analysis based on quality-adjusted life years gained. All analyses utilized R version 4.2.0 with significance level set at α = 0.05 using two-sided tests.

Results

Baseline data

A total of 2,814 children were screened between January 2020 and December 2021, with 1,589 enrolled after exclusions (Fig. 1). Follow-up rates were 89.7% at 12 months, 86.9% at 24 months, and 84.0% at 36 months; attrition (16.0%) was mainly due to relocation, missing data were handled using multiple imputation in longitudinal analyses. The cohort consisted of 846 boys and 743 girls (mean age 7.5 ± 1.2 years). Most were urban residents (58.3%). Key baseline measures included mean BMI (16.4 ± 2.8 kg/m²) and mean neck circumference (27.3 ± 2.4 cm). Common medical histories included allergic rhinitis (28.7%) and adenotonsillar hypertrophy (12.9%). Portable polysomnography showed a mean OAHI of 3.2 ± 4.1 events/hour. Craniofacial features were mostly upright facial profile (62.4%) and proper mandibular development (68.3%). No significant baseline differences were found between study completers and those lost to follow-up, except for a slightly higher proportion of urban residents among completers. Intervention groups were balanced at baseline (Supplementary Table 2).

Fig. 1.

Study definitions, severity thresholds, and participant flow

Prevalence and progression of SDB

Overall, SDB occurred in 459 (28.9%) out of 1,589 children. When stratified by severity, mild SDB was present in 184 children (11.6%), moderate SDB in 153 children (9.6%), and severe SDB in 122 children (7.7%). Boys had a significantly higher prevalence across all severity categories than girls (p < 0.001) (Supplementary Table 3). For a subsequent intervention sub-study, eligibility was limited to children with moderate-to-severe SDB. A total of 275 children met this criterion and were enrolled in the sub-study.

Over the 36-month follow-up, the SDB status of the cohort showed dynamic changes. Among children with no SDB at baseline, 12.4% developed incident SDB. Of the children with SDB at baseline, 48.6% experienced complete remission, while 36.7% saw their condition progress. A fluctuating course was common, observed in 41.5% of baseline SDB cases. Longitudinal analysis of Portable polysomnography parameters showed significant improvement over time. The mean OAHI decreased from 3.2 to 2.4 events/hour (p < 0.001). Parallel improvements were seen in the ODI and minimum SpO₂. These improvements were most substantial in children who had moderate or severe SDB at baseline (Table 1).

Table 1.

Longitudinal changes in SDB status over 36 months (n = 1,335)

Baseline Status	n	Complete Remission	Persistent SDB	Progressed	Fluctuating Course
No SDB (OAHI < 1)	954	836 (87.6)	-	118 (12.4)	-
Mild SDB	153	104 (68.0)	11 (7.2)	38 (24.8)	42 (27.5)
Moderate SDB	130	56 (43.1)	18 (13.8)	56 (43.1)	62 (47.7)
Severe SDB	98	25 (25.5)	27 (27.6)	46 (46.9)	54 (55.1)
All SDB	381	185 (48.6)	56 (14.7)	140 (36.7)	158 (41.5)

Percentages in mild, moderate, and severe SDB groups may sum to > 100% due to overlapping categories in fluctuating course

Risk factors for SDB

Univariate analysis identified multiple factors significantly associated with SDB, with adenotonsillar hypertrophy (OR = 3.52), allergic rhinitis (OR = 3.32), and obesity (OR = 2.41) showing the strongest associations. Paternal snoring (OR = 2.64) and tobacco exposure (OR = 1.96) were also significant risk factors (Supplementary Table 4). Multivariable analysis confirmed several independent risk factors. Adenotonsillar hypertrophy remained the strongest predictor (aOR = 3.42), followed by paternal snoring (aOR = 2.38), obesity (aOR = 2.15), and allergic rhinitis (aOR = 1.86). Male sex was also an independent risk factor (aOR = 1.64) (Table 2). Risk factor combinations demonstrated synergistic effects. The presence of all three major factors—adenotonsillar hypertrophy, paternal snoring, and obesity—resulted in a markedly elevated odds ratio of 8.92. A risk score incorporating the four strongest predictors showed excellent discriminatory power for SDB (AUC = 0.82) (Table 3).

Table 2.

Portable polysomnography Parameter Changes Over Time (Mean ± SD)

Parameter	Baseline (n = 1,589)	12 Months (n = 1,426)	24 Months (n = 1,381)	36 Months (n = 1,335)	p-value*
OAHI (events/hour)	3.2 ± 4.1	2.9 ± 3.7	2.6 ± 3.4	2.4 ± 3.3	< 0.001
- No SDB	0.4 ± 0.3	0.5 ± 0.4	0.5 ± 0.4	0.6 ± 0.5	0.124
- Mild SDB	2.8 ± 1.2	2.2 ± 1.4	1.8 ± 1.3	1.6 ± 1.2	< 0.001
- Moderate SDB	6.8 ± 1.4	5.9 ± 2.1	5.2 ± 2.3	4.7 ± 2.4	< 0.001
- Severe SDB	14.2 ± 3.8	11.6 ± 4.2	9.8 ± 4.5	8.4 ± 4.7	< 0.001
ODI	2.8 ± 3.6	2.5 ± 3.2	2.3 ± 2.9	2.1 ± 2.8	< 0.001
Minimum SpO₂	90.2 ± 3.8	90.6 ± 3.5	91.0 ± 3.3	91.5 ± 3.2	< 0.001
Arousal Index	8.4 ± 3.2	7.9 ± 3.0	7.6 ± 2.8	7.3 ± 2.7	< 0.001

^*p-value for linear trend across all time points using mixed-effects models

Table 3.

Multivariate Logistic Regression Analysis of Independent Risk Factors for SDB

Risk Factor	Adjusted OR	95% CI	p-value
Medical Factors
Adenotonsillar hypertrophy	3.42	2.68–4.36	< 0.001
Allergic rhinitis	1.86	1.51–2.29	< 0.001
Asthma	1.28	0.89–1.84	0.184
Tympanitis	1.45	1.02–2.06	0.038
Family History
Paternal snoring	2.38	1.92–2.95	< 0.001
Maternal snoring	1.72	1.36–2.17	< 0.001
Environmental Factors
Tobacco exposure	1.41	1.12–1.78	0.003
Prone sleep position	1.58	1.15–2.17	0.005
Anthropometric Factors
Obesity (BMI ≥ 95th %)	2.15	1.68–2.75	< 0.001
Neck circumference (per 1 cm increase)	1.18	1.12–1.24	< 0.001
Demographic Factors
Male sex	1.64	1.33–2.02	< 0.001
Age (per year increase)	0.92	0.85–0.99	0.032

^*Model adjusted for all variables listed; Hosmer–Lemeshow goodness-of-fit test: χ² = 7.23, p = 0.512; Nagelkerke R² = 0.28*

Craniofacial and airway morphology

CBCT and craniofacial analyses revealed significant morphological differences in children with SDB. The SDB group had a 38.2% smaller total airway volume and a significantly reduced minimum cross-sectional area (48.3 mm² vs. 82.7 mm², p < 0.001) compared to controls (Supplementary Table 5). Cephalometric analysis showed the SDB group had characteristics of mandibular retrusion (smaller SNB angle), a steeper mandibular plane, larger overjet, and constricted dental arches. The hyoid bone was also positioned more inferiorly (Supplementary Table 6). These morphological parameters were strongly correlated with SDB severity. Minimum cross-sectional area showed the strongest inverse correlation with the OAHI (r = −0.723, p < 0.001). A model combining airway and skeletal factors explained 58.7% of the variance in disease severity (Supplementary Table 7).

Malocclusion traits and their association with SDB

Children with SDB showed a significantly higher prevalence of specific malocclusion traits compared to those without SDB. Class II malocclusion (48.6% vs. 27.3%), overjet > 4 mm (52.7% vs. 18.9%), posterior crossbite (34.2% vs. 14.7%), and anterior open bite (15.9% vs. 5.2%) were all strongly associated with SDB. These malocclusion traits were also correlated with SDB severity (Supplementary Table 8). A composite malocclusion score showed a strong positive correlation with the Obstructive Apnea–Hypopnea Index (OAHI) (r = 0.512, p < 0.001). Longitudinally, a bidirectional relationship was observed (Supplementary Table 9). Children with persistent SDB experienced worsening occlusal features, such as increasing overjet. Conversely, children whose SDB went into remission showed spontaneous improvement in malocclusion traits, including overjet reduction and posterior crossbite correction (Supplementary Table 10).

Age-stratified analysis revealed that younger children (6–7 years) with SDB had higher potential for occlusal improvement following SDB resolution compared to older children (8–9 years), suggesting a critical period for intervention. The interaction between SDB duration and occlusal worsening followed a dose–response relationship, with longer SDB duration associated with greater malocclusion progression. Intervention sub-analysis indicated that children receiving early orthodontic or ENT treatment showed significantly better occlusal outcomes compared to the watchful waiting group, highlighting the potential for targeted interventions to modify both respiratory and occlusal trajectories (Table 4).

Table 4.

Risk Factor Combinations and Synergistic Effects

Risk Factor Combination	Prevalence in SDB Group	Prevalence in Non-SDB Group	Synergistic OR (95% CI)
Adenotonsillar hypertrophy + Allergic rhinitis	38.1%	4.2%	5.89 (4.12–8.42)
Paternal snoring + Obesity	24.6%	3.8%	4.76 (3.45–6.57)
Adenotonsillar hypertrophy + Paternal snoring	29.4%	2.9%	6.34 (4.52–8.89)
All three major factors*	18.3%	1.2%	8.92 (5.78–13.76)

^*Adenotonsillar hypertrophy + Paternal snoring + Obesity

Intervention sub-study outcomes

Baseline characteristics of the three intervention groups were well-balanced across key demographic, clinical, and polysomnographic variables (Supplementary Table 11). All 275 participants completed the 6-month follow-up assessment, with no dropouts recorded. In the control group, 18 children (20.2%) opted to cross over to active treatment after the 6-month evaluation due to insufficient improvement. Treatment fidelity was high in both intervention arms: all 92 ENT-group participants underwent adenotonsillectomy, while 91 of 94 orthodontic-group participants (96.8%) completed rapid maxillary expansion per protocol, achieving a mean total expansion of 8.2 ± 1.3 mm. All orthodontic participants completed the 6-month retention phase (Supplementary Table 12). The intervention sub-study for children with moderate-to-severe SDB demonstrated significantly different outcomes among the three groups at the 6-month follow-up. The ENT intervention group achieved the highest rate of complete SDB resolution (68.4%), followed by the orthodontic intervention group at 52.1%, while only 15.7% of the control group experienced resolution (Supplementary Table 13). Portable polysomnography data confirmed the clinical results. The ENT group showed a dramatic 74.2% reduction in the OAHI, and the orthodontic group achieved a substantial 58.7% reduction. In contrast, the control group showed minimal improvement. Parallel improvements were seen in oxygen saturation, which significantly increased in both active treatment groups. The mechanisms of improvement differed: the ENT group showed a targeted increase in retropalatal airway volume, while the orthodontic group achieved a more generalized airway expansion and corrected maxillary constriction (Table 5). A cost-effectiveness analysis revealed that both interventions were cost-effective compared to watchful waiting. The ENT intervention demonstrated a lower cost per Quality-Adjusted Life Year gained compared to the orthodontic intervention, though the orthodontic approach provided the added benefit of simultaneous occlusal correction.

Table 5.

Portable polysomnography Parameter Changes Following Intervention

Parameter	ENT Group	Orthodontic Group	Control Group
OAHI (events/hour)
Baseline	12.4 ± 4.2	10.8 ± 3.9	11.2 ± 4.1
6-month	3.2 ± 2.8*	4.5 ± 2.6*	9.8 ± 3.7
ODI (events/hour)
Baseline	10.8 ± 3.7	9.6 ± 3.4	10.1 ± 3.6
6-month	2.7 ± 2.1*	3.9 ± 2.3*	8.9 ± 3.2
Minimum SpO₂ (%)
Baseline	85.4 ± 4.2	86.2 ± 3.9	85.8 ± 4.1
6-month	91.8 ± 3.1*	90.3 ± 2.8*	86.5 ± 3.9

^*p < 0.001 vs. baseline

Quality of life and behavioral correlates

Children with SDB had significantly worse quality of life and more behavioral problems at baseline compared to their non-SDB peers. Their scores were markedly lower across all domains, including physical, psychosocial, and school functioning. They also scored significantly higher on scales for inattention, hyperactivity, and daytime sleepiness (Supplementary Table 14). Following treatment, both the ENT and orthodontic intervention groups showed substantial improvements in these areas. Quality of life scores increased significantly, while symptoms of inattention and daytime sleepiness decreased. The improvements were strongly correlated with the reduction in OAHI, indicating that better sleep breathing directly led to better daytime function. The control group showed minimal change (Supplementary Table 15).

Discussion

This prospective longitudinal cohort study provides compelling evidence for a significant and bidirectional relationship between pediatric SDB, craniofacial growth, and occlusal development. We observed a high baseline prevalence of SDB in our cohort, underscoring the importance of Portable polysomnography for accurate diagnosis. Our findings confirm established risk factors including adenotonsillar hypertrophy and paternal snoring as potent predictors, but more critically, we identified a distinct craniofacial phenotype in children with SDB characterized by mandibular retrusion, maxillary constriction, and reduced pharyngeal airway dimensions that is strongly correlated with disease severity. Crucially, our longitudinal design revealed a dynamic, two-way interaction: persistent SDB actively exacerbates malocclusion, while SDB remission facilitates spontaneous occlusal improvement. This bidirectional interplay suggests not only a vicious cycle that can negatively reinforce both respiratory and dentofacial development but also a vital window of opportunity in which early intervention may disrupt this cycle and positively influence craniofacial growth trajectories.

In the present study, the independent risk factors identified for pediatric SDB including adenotonsillar hypertrophy, paternal snoring, obesity, and allergic rhinitis resonate with findings from prior epidemiological work [10] while also highlighting the multifactorial nature of its etiology. Our results confirm adenotonsillar hypertrophy as the most potent predictor, consistent with its well-established role in physically obstructing the pediatric airway. The significant association with paternal snoring underscores a substantial hereditary predisposition, likely involving inherited craniofacial features that predispose the airway to collapse. Furthermore, the roles of obesity, which can impose extrinsic load on the upper airway, and allergic rhinitis, which contributes through inflammatory nasal congestion and increased airway resistance, illustrate the complex interplay between anatomical, inflammatory, and metabolic pathways. This constellation of risk factors affirms that SDB pathogenesis in children is not attributable to a single cause but arises from a convergence of genetic, anatomical, and environmental influences, necessitating a comprehensive clinical assessment for effective risk stratification and management.

Our craniofacial findings indicated by notably the smaller SNB angle, larger ANB angle, and constricted maxillary and mandibular arches in children with SDB provide strong support for Moss’s functional matrix theory. This theory posits that the growth of craniofacial skeletal structures is directed by the functional demands of adjacent soft tissues and spaces. The chronic upper airway obstruction inherent in SDB necessitates a shift from nasal to oronasal or oral breathing, which in turn alters the postural equilibrium of the tongue, mandible, and perioral muscles. The resultant low tongue posture and mouth-open stance are believed to disrupt the balanced forces required for normal transverse maxillary development, leading to the high prevalence of palatal crossbite and narrow arches we observed. Simultaneously, the posteriorly rotated mandible and increased anterior facial height, reflected in our steeper mandibular plane angles, can be interpreted as an adaptive mechanism to maintain airway patency, yet this adaptation perpetuates a skeletal framework that is inherently vulnerable to airway collapse. When contextualized within the existing literature, our results are highly consistent with the patterns of mandibular retrusion and maxillary constriction reported in preschool children with obstructive sleep apnea by Marino et al. [14] and the narrowed airway dimensions associated with specific skeletal patterns in nonsnoring children by Zhong et al. [15]. However, our longitudinal data add a critical temporal dimension, suggesting that these features are not merely static risk factors but are dynamically influenced by the presence and persistence of SDB itself. This contrasts with some cross-sectional studies, such as the orthodontic clinic-based study by Katyal et al. [16], which found no significant sagittal skeletal predictors, a discrepancy potentially explained by their sample's positioning at the milder end of the SDB spectrum. The high prevalence of analogous craniofacial features—dolichofacial pattern, lip incompetence, and atresic palate—in the community-based sample of healthy mouth-breathing children described by Pacheco et al. [17] further underscores that this SDB-associated craniofacial phenotype is widespread and detectable even outside clinical sleep cohorts, reinforcing the profound impact of obstructed breathing on facial growth.

Of particular significance in our study is the compelling evidence for a bidirectional relationship between SDB and malocclusion. On one hand, persistent SDB can actively drive the development or worsening of malocclusion. Chronic nasal obstruction leads to mouth breathing, which results in low tongue posture and an anterior open bite, and disrupts the equilibrium of orofacial muscles. This imbalance increases buccal pressure on the maxillary arches, contributing to a narrow, high-vaulted palate and posterior crossbite—features we observed significantly more often in children with SDB. On the other hand, a pre-existing skeletal predisposition, such as a Class II malocclusion characterized by mandibular retrognathia, can itself be a primary risk factor for SDB by reducing posterior airway space and facilitating pharyngeal collapse during sleep. Our data demonstrate that this is not a one-way relationship: children with persistent SDB experienced worsening occlusal traits such as overjet, whereas those who achieved SDB remission showed spontaneous improvement in these same malocclusion features. This suggests a vicious cycle where airway obstruction and dentofacial development negatively reinforce each other, but also a window of opportunity where early SDB intervention can positively influence craniofacial growth. When compared to the existing literature, our findings on this bidirectional interplay provide crucial longitudinal validation for what previous cross-sectional and review studies could only hypothesize. The systematic review by Hansen et al. [18] strongly associated malocclusions like anterior open bite and large overjet with SDB but could not establish temporal sequence. Similarly, clinical guidelines by Pacheco et al. [19] highlighted the common co-occurrence of mouth breathing with atresic palates and anterior open bites, aligning with our proposed pathway from obstruction to malocclusion. Furthermore, the study by Schütz et al. [20] demonstrated that Herbst appliance therapy for Class II correction not only advanced the mandible but also increased posterior airway space and improved nocturnal breathing, directly supporting the concept that correcting a retrognathic skeletal base can alleviate SDB. However, these studies lacked the longitudinal follow-up to show that the relationship is dynamic and reciprocal. Our results bridge this gap, confirming that the association is not merely correlational but causal and bidirectional over time. This underscores the limitation of previous research, such as the cross-sectional nature of Kim & Guilleminault's work [21], which, while identifying craniofacial risk factors for residual SDB post-adenotonsillectomy, could not track how SDB and occlusion evolve together. Our longitudinal evidence solidifies the need for an interdisciplinary approach that breaks this cycle, positioning orthodontic intervention not only as a means to correct dentition but also as a potential strategy to modify airway patency and SDB trajectory.

The efficacy of both ENT and orthodontic interventions in managing pediatric obstructive sleep apnea underscores the multifactorial nature of the condition and supports a phenotype-driven approach to treatment. Adenotonsillectomy, as demonstrated in the studies by Yang et al. [22] and the longitudinal cohort, primarily addresses soft-tissue obstruction by physically removing hypertrophic lymphoid tissue, thereby directly enlarging the retropalatal and retroglossal airway spaces. In contrast, dentofacial orthopedic treatment, particularly rapid maxillary expansion and mandibular advancement using devices such as the Twin-Block, targets skeletal constraints—maxillary constriction and mandibular retrognathia—that contribute to airway narrowing [23]. By expanding the maxilla and advancing the mandible, dentofacial orthopedic treatment not only increases nasal and pharyngeal volumes but also repositions associated soft tissues, such as the tongue and soft palate, thereby reducing airway resistance and collapsibility. This mechanistic distinction is further illustrated by the computational fluid dynamics findings of Yang et al. [22], in which both interventions reduced airflow resistance and wall shear stress, albeit through different anatomical pathways. When compared with existing literature, these findings align with studies such as those by Pavoni et al. [24] and He et al. [25], which reported significant OAHI reductions following both adenotonsillectomy and functional appliance therapy. However, the present studies add nuance by revealing that dentofacial orthopedic treatment may offer more generalized and sustained improvements in craniofacial form and airway patency, whereas adenotonsillectomy provides rapid relief from lymphoid obstruction but may not address underlying skeletal risk factors for residual obstructive sleep apnea. This supports the growing consensus that treatment should be tailored to the dominant phenotypic traits—whether soft-tissue or skeletal—in each child, reinforcing the need for interdisciplinary evaluation and personalized intervention strategies.

Consistent with the findings of Yu et al. [26], which demonstrated a weak correlation between caregiver- and child-reported global quality of life in children with SDB, our study reinforces the critical importance of directly assessing the child’s perspective in both research and clinical practice. In our cohort, children with SDB exhibited significantly lower quality of life scores across physical, emotional, social, and school functioning domains compared to their non-SDB peers. Notably, following intervention—whether ENT or orthodontic—significant improvements in quality of life were observed, paralleling objective respiratory and occlusal outcomes. This suggests that successful treatment not only alleviates physiological symptoms but also enhances the child’s subjective well-being. However, consistent with Yu et al.’s [26] caution that caregiver reports may overestimate quality of life, our results underscore the need to incorporate child-reported quality of life measures, particularly in younger children, to avoid underestimating the true burden of SDB. Future studies should prioritize the development and validation of disease-specific, child-friendly quality of life instruments to more accurately capture the impact of SDB and its treatment from the child’s own viewpoint.

Several limitations of this study should be considered. First, while our cohort was substantial and drawn from a major urban center in China, the findings may not be fully generalizable to rural populations or other ethnic groups with differing craniofacial characteristics and lifestyle factors. Second, the allocation to intervention groups in the sub-study, while incorporating clinical recommendations, was not fully randomized, potentially introducing selection bias, as parents with stronger preferences or children with more specific clinical presentations may have been directed toward a particular treatment. This limits the ability to make definitive causal inferences about the comparative efficacy of the interventions. Third, the 36-month follow-up period may still be insufficient to capture the full long-term trajectory of craniofacial growth and the potential for late recurrence of SDB. Fourth, Portable polysomnography is validated for pediatric use and recommended when in-lab polysomnography is impractical, though home-based device may reduce sensitivity for arousal-based hypopnea. Finally, our airway assessment was limited to static anatomical imaging. While we inferred the role of tongue posture from associated craniofacial features, we could not directly assess dynamic airway collapse or functional changes during sleep. Future studies using sleep MRI would better visualize sites of obstruction and soft tissue dynamics, providing a more functional understanding of SDB pathophysiology in children.

In conclusion, this longitudinal cohort study demonstrates that pediatric SDB is a highly prevalent condition characterized by a complex, clinically significant, and bidirectional relationship with craniofacial growth and occlusal development. We have identified a distinct SDB-associated craniofacial phenotype and established a critical two-way interaction: persistent SDB actively worsens malocclusion, while SDB remission promotes spontaneous occlusal improvement. This evidence underscores that SDB is not merely a comorbidity but a modifiable factor in dentofacial development. The demonstrated efficacy of both ENT and orthodontic interventions, albeit through distinct mechanisms, reinforces the multifactorial nature of SDB and affirms the necessity of an interdisciplinary diagnostic and therapeutic approach. Our findings advocate for the integration of SDB screening into routine pediatric dental and orthodontic assessments and, conversely, for craniofacial evaluation in children diagnosed with SDB. Early, phenotype-targeted intervention can disrupt this detrimental cycle, thereby improving respiratory health, optimizing dentofacial development, and enhancing the quality of life for affected children.

Abbreviations

SDB

Sleep-disordered breathing

OAHI

Obstructive apnea–hypopnea index

OR

Odds ratio

ENT

Ear Nose Throat

ALARA

As Low As Reasonably Achievable

ODI

Oxygen desaturation index

ISAAC

The International Study of Asthma and Allergies in Childhood

Authors’ contributions

Study design: FZ, GZ, SS Manuscript writing: FZ, GZ, SS Studies selecting: FZ, GZ, SS Data analysis: FZ, GZ, SS Study quality evaluating: FZ, GZ, SS Manuscript revising: FZ, GZ, SS The final manuscript was read and approved.

Funding

None.

Data availability

Availability of data and materials All data generated or analyzed during this study are included in this published article. And the primary data could be achieved from the corresponding author.

Declarations

Ethics approval and consent to participate

This study was approved by The First Affiliated Hospital of Zhengzhou University Institutional Research Committee, and written informed consent for medical research was obtained from all patients prior to initial treatment. Informed consent to participate was obtained from the parents or legal guardians of any participant under the age of 16, and all the procedures were in compliance with the Helsinki Declaration.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.