Maxillary Expansion, Pharyngeal Volume and Sleep-Disordered Breathing in Children with Cleft Lip and Palate — A Comprehensive Evidence Review
Prepared: 2026-11-04 Audience: Physicians and clinicians working in cleft/craniofacial teams, paediatric sleep medicine, otolaryngology and orthodontics Citation style: Vancouver numbering, generated mechanically from atomic source tokens (PubMed identifiers, ClinicalTrials.gov identifiers, and non-indexed source keys). Every PubMed identifier in this review was retrieved from NCBI E-utilities during preparation; every numerical claim is traceable to an abstract or record read during this run.
Abstract
A 2026 randomised trial in Sleep and Breathing compared a clear-aligner-based expansion screw with a nickel-titanium palatal expander in 24 children with unilateral cleft lip and palate, transverse maxillary deficiency and elevated sleep questionnaire scores, and reported that both appliances “significantly improved pharyngeal airway volume and sleep parameters” [1]. The trial has been received as evidence that maxillary expansion is a sleep intervention. This review argues that it is not, and that the gap between what the trial measured and what a family wants to know is the central clinical problem in this literature.
Four findings organise the review. First, the trial has no untreated arm: both groups were expanded, so every within-group change is confounded by growth, adenotonsillar involution and regression to the mean in children selected for high symptom scores — a bias that is not hypothetical, because 42% of children with OSA randomised to watchful waiting in the Childhood Adenotonsillectomy Trial no longer met polysomnographic criteria seven months later [2]. Second, static upper-airway volume is a weak proxy for nocturnal obstruction. Pooled analyses show that rapid maxillary expansion reliably enlarges the nasal cavity (about 1.2 cm³) but not consistently the nasopharynx or oropharynx, and that volumetric change does not track apnoea change within the same children [3] [4] [5]. Third, when controlled comparisons exist, the apnoea signal disappears: a network meta-analysis estimated the effect of rapid maxillary expansion alone on the apnoea-hypopnoea index at 0.02 events/hour (95% CI −1.72 to 1.75) [6], and the only randomised comparison against watchful waiting found no significant difference [7]. Fourth, adenotonsillectomy remains first-line for the ordinary child with adenotonsillar hypertrophy and OSA, on the strength of a 464-child randomised trial showing improved polysomnography, behaviour and quality of life but no benefit on attention or executive function [8] [9].
For children with cleft lip and palate the picture is different in emphasis but not in evidentiary quality. Sleep-disordered breathing is common — pooled prevalence 16% (95% CI 12–21%), and 32% in syndromic cleft populations [10] — and midface hypoplasia gives expansion a stronger anatomical rationale than it has in the non-cleft child. But the cleft expansion literature is small, single-centre, unblinded, and dominated by surrogate endpoints. The defensible clinical position is that expansion is indicated for transverse maxillary deficiency on orthodontic and dentofacial grounds, that it may improve nasal breathing, and that it is not established as a treatment for paediatric OSA — a position now stated explicitly by the most recent systematic review in the field [11]. What would change that position is a trial with an untreated or delayed-treatment arm, blinded polysomnographic scoring, stratified adenotonsillar status, and apnoea-hypopnoea index as a pre-specified primary outcome with a declared minimal important difference. No such trial exists.
1. Introduction: Why a Palate-Expansion Trial Landed in a Sleep Journal
1.1 The clinical question behind the paper
Every cleft team is asked some version of the same question: “if you widen his palate, will he breathe better at night?” The question is reasonable. The child has a narrow maxilla, a small nasal cavity, and noisy or obstructed breathing; the operation widens the nasal floor; therefore the airway should improve. The syllogism is old, intuitive, and partly true. It is also the kind of syllogism that clinical research exists to test rather than to confirm.
The trial by Bansal and colleagues is the most recent attempt to answer part of it [1]. It randomised 24 children with unilateral cleft lip and palate (UCLP) in the mixed dentition, all with transverse maxillary deficiency and elevated sleep questionnaire scores, to one of two active expansion systems: a clear-aligner-based expansion screw or a nickel-titanium palatal expander. Cone-beam computed tomography (CBCT) and polysomnography (PSG) were performed before and after expansion. The authors report that both appliances improved pharyngeal airway volume and sleep parameters, that the aligner-based system produced greater intercanine and intermolar width gains, and that anterior transverse widening correlated with airway volume gains while basal skeletal widening and molar inclination did not. They conclude that the two appliances are comparably effective and that the overall expansion effect, rather than its mechanistic subtype, may drive functional benefit [1].
That is a coherent paper. It is not, however, a paper that can answer the question a parent asks, for a reason that is structural rather than statistical: both arms were treated. When every participant receives an active intervention, the trial can compare two ways of expanding a palate. It cannot compare expanding with not expanding, and therefore it cannot attribute the within-group improvements in airway volume or sleep parameters to the appliance. This is not a subtle methodological quibble. It is the difference between “this appliance is as good as that appliance” and “this treatment works.”
1.2 Why the question is harder than it looks
Three separate propositions are routinely conflated in this literature, and separating them is the organising principle of this review.
Proposition one: expansion enlarges the measured airway. This is the most defensible of the three, with important qualifications. Rapid maxillary expansion (RME) increases nasal cavity volume with reasonable consistency across studies [3] [4]. Effects on the nasopharynx and oropharynx are smaller, less consistent, and in several well-conducted studies absent [12] [13] [14].
Proposition two: an enlarged measured airway reduces apnoea and improves gas exchange. This is where the literature becomes uncomfortable. Static awake airway volume explains only a modest fraction of the variance in the apnoea-hypopnoea index (AHI), and change in volume does not reliably track change in AHI even within the same patients [15] [5] [16]. The airway that matters is a dynamic, sleep-state-dependent, neuromuscular tube, not a rigid pipe measured in the upright position while awake.
Proposition three: any of this changes what a child feels, how they behave, how they grow, or their blood pressure. Here the evidence is thinnest of all. Even adenotonsillectomy — a far more studied intervention with a far larger effect on the AHI — produced no measurable benefit on attention or executive function in the largest randomised trial, and only small, selective cognitive effects on secondary analysis [8] [17].
The Bansal trial speaks mainly to proposition one, partly to proposition two, and not at all to proposition three [1]. A review that presents it as settling the matter would be doing the opposite of what the evidence supports.
1.3 What this review does
This review has three aims. First, to place the trial in the epidemiological and anatomical context that makes its question intelligible — paediatric OSA as a clinical problem, sleep-disordered breathing in cleft populations, and the anatomical rationale for expansion. Second, to interrogate each link in the causal chain with the best available evidence, including the measurement science of CBCT volumetry and the growth confounder that makes single-arm paediatric studies so difficult to interpret. Third, to state plainly what a clinician can and cannot say to a family today, and what kind of trial would actually change practice.
The review is deliberately sceptical, but not nihilistic. Maxillary expansion is a real treatment with real effects on the nasal airway and real benefits for nasal breathing in selected children. The argument is not that expansion does nothing. The argument is that the leap from “widens the nasal cavity” to “treats sleep apnoea” is unsupported, that the leap is being made routinely in abstracts and reviews, and that the cost of that leap is borne by children who may be treated for years for an indication that has not been demonstrated.
2. Paediatric Obstructive Sleep Apnoea as a Clinical Problem
2.1 Epidemiology and the shape of the population
Paediatric obstructive sleep apnoea is common enough to matter and uncommon enough to be missed. A widely cited synthesis of the epidemiological literature estimated parent-reported always-snoring at 1.5–6%, parent-reported apnoeic events at 0.2–4%, sleep-disordered breathing by questionnaire at 4–11%, and OSA diagnosed by varying polysomnographic criteria at 1–4%; the meta-analytic prevalence of parent-reported snoring by any definition was 7.45% (95% CI 5.75–9.61) [18]. A more recent systematic review of preschool children found prevalence estimates of 1.8% and 6.4% in studies that diagnosed OSA by polysomnography using an AHI threshold of 5 events/hour with the 4% desaturation rule, while questionnaire-only studies published between 2016 and 2023 reported 12.8–20.4% [19]. The spread is not noise; it reflects different case definitions applied to different populations with different ascertainment. Any statement that “paediatric OSA affects X% of children” is therefore a statement about a definition as much as about a disease.
The clinically important consequence is that the population arriving in a sleep clinic is heterogeneous. The classic phenotype is a normal-weight child aged 2–8 years with adenotonsillar hypertrophy, mouth breathing, snoring and witnessed pauses. But a contemporary series of 491 children with paediatric OSA found that only 15.6% were obese, that snoring was not the most prevalent nocturnal symptom (restlessness 91.6%, chronic insomnia 86.1%, leg kicking 82.0%, frequent awakening 78.4%), and that neurocognitive or behavioural concerns were present in 93.7%, most commonly mood swings (86.5%) and attention-deficit/hyperactivity problems (51.9%) [20]. The child who “does not look like OSA” is common. This matters for cleft children in particular, because their facial difference makes clinicians attend to the airway while their symptom profile may not resemble the textbook.
2.2 Diagnosis and the AHI problem
The reference standard is attended polysomnography with scoring according to American Academy of Sleep Medicine rules, and the conventional severity metric is the AHI — obstructive and mixed apnoeas plus hypopnoeas per hour of sleep. The diagnostic threshold of an AHI of 1 event/hour in children is itself historically contingent: a critical review notes that it derives from old statistical comparisons without clinical validation, and that the Pediatric Adenotonsillectomy Trial for Snoring (PATS) found no significant clinical difference between primary snorers (AHI <1) and children with mild OSA (AHI 1–5) [21]. The same review observes that correlations between the AHI and the two most-used symptom instruments are very weak — on the order of r² ≈ 2% [21]. If the physiological measure and the symptom measure share almost no variance, then “OSA severity” is not one thing, and a treatment that moves one need not move the other.
The AHI is also unstable within a child. A study of 27 children who underwent two polysomnograms a median of 48 days apart found that 7 of 27 (26%) with moderate-to-severe OSA on the first night had mild or no OSA on the second; the intraclass correlation coefficient for the obstructive AHI across nights was 0.27 (95% CI −0.04 to 0.51), and for the oxygen desaturation index 0.10 (95% CI −0.26 to 0.53) [22]. A diagnostic instrument with an ICC of 0.27 in the very range where treatment decisions are made is a warning against over-reading small changes in small samples — which is precisely the setting of every expansion trial discussed below.
None of this means the AHI is useless. It predicts outcomes in some contexts, it is the endpoint on which the major trials were powered, and it is the only measure regulators and guideline panels currently accept. It means that a single AHI value, in a single child, on a single night, carries more uncertainty than the decimal places suggest.
2.3 Consequences: what justifies treatment, and what treatment actually changes
The case for treating paediatric OSA rests on associations with neurocognitive impairment, behavioural problems, impaired quality of life, cardiovascular and metabolic risk, and — in the cleft population — academic and psychosocial difficulty [23]. The case for expecting treatment to reverse all of these is weaker than the associations.
The Childhood Adenotonsillectomy Trial (CHAT) randomised 464 children aged 5–9 years with OSA to early adenotonsillectomy or watchful waiting with supportive care [8]. The primary outcome, the attention and executive function score of the Developmental Neuropsychological Assessment, improved by 7.1 ± 13.9 points with surgery and 5.1 ± 13.4 with watchful waiting (P = 0.16) — no significant difference, in a population whose baseline scores were already close to the population mean. Behaviour, quality of life, symptoms and polysomnographic findings improved significantly more with surgery, and polysomnographic normalisation occurred in 79% versus 46% [8]. Secondary analyses found small, selective cognitive effects (Cohen’s d 0.20–0.24 on nonverbal reasoning, fine motor skills and selective attention) and clear improvements in parent-reported behaviour [17] [24]. The overall pattern — substantial improvement in sleep and symptoms, modest and domain-specific effects on cognition — has been summarised as the central lesson of CHAT [25].
Cardiovascular outcomes are similarly nuanced. In an exploratory analysis of the PATS trial, adenotonsillectomy was associated with a 9.02-point greater reduction in systolic blood pressure percentile (95% CI −14.77 to −3.28) and a 6.53-point greater reduction in diastolic percentile (95% CI −11.02 to −2.03) at 12 months in children with mild sleep-disordered breathing [26]. A separate randomised study of surgery versus watchful waiting found no overall difference in ambulatory blood pressure, but a significant improvement in night-time diastolic z-score (−0.43 ± 1.01, P = 0.027) confined to children with severe preoperative OSA (OAHI ≥10/hour) [27]. The reasonable inference is that blood pressure benefit, if it exists, is concentrated in more severe disease — which is exactly the group least likely to be enrolled in a 24-child appliance trial.
Two further findings complicate the “treat early, treat aggressively” narrative. First, adenotonsillectomy was associated with a 32% reduction in total health care encounters and a 48% reduction in prescriptions in the PATS cohort, a utilitarian argument that has nothing to do with apnoea counts [28]. Second, surgery is associated with greater gains in height and weight percentiles at 12 months (height mean difference 2.74, 95% CI 0.33 to 5.15) [29], and concerns about postoperative weight gain have been partly but not fully allayed: in a PATS analysis, adenotonsillectomy was not independently associated with undesirable weight gain, although one-third of all children gained undesirable weight during the study [30].
2.4 Why natural history matters more than any of this
The single most important fact for interpreting every expansion study in this review is that paediatric OSA improves spontaneously. In the CHAT watchful-waiting arm, 194 children were followed for seven months; 82 (42%) no longer met polysomnographic criteria for OSA, although only 25 of 167 (15%) with baseline PSQ scores ≥0.33 experienced symptomatic resolution [2]. In the PATS control arm, 13% of children progressed to an OAHI ≥3 at 12 months while 57% had persistent or progressive symptoms [31]. In the Karolinska Adenotonsillectomy trial of children aged 2–4 years, the change in OAHI did not differ between surgery and watchful waiting (Cohen’s d 0.05, 95% CI −0.7 to 0.6), and 13 of 31 children in the watchful-waiting arm (42%) crossed over to surgery [32].
A 42% polysomnographic remission rate over seven months in untreated children is the benchmark against which any single-arm improvement must be measured. A pre-post study reporting that the AHI fell from 8.9 to 2.7 events/hour looks impressive in isolation [33]. It looks considerably less impressive when the expected untreated trajectory includes spontaneous normalisation in two of five children. This is not a reason to withhold treatment from a child who needs it. It is a reason to demand a control group before attributing improvement to an appliance.
3. Sleep-Disordered Breathing in Cleft Lip and Palate: A Population with a Different Airway
3.1 Prevalence
Children with cleft lip and/or palate carry a higher burden of sleep-disordered breathing than their peers, though the size of the excess depends heavily on how it is measured. A 2026 systematic review and meta-analysis of seven observational studies involving 1,567 syndromic and non-syndromic children with cleft lip and/or palate estimated an overall prevalence of sleep-disordered breathing of 16% (95% CI 12–21%), with no significant gender difference [10]. The estimate was highest in syndromic cleft populations at 32%, and lower in non-syndromic cases, where study-level estimates ranged from 8.8% to 18% [10]. All seven studies used the Pediatric Sleep Questionnaire, albeit with different cut-offs, and only two included polysomnography; among children who screened positive and underwent PSG, over 75% were diagnosed with OSA [10].
Individual studies illustrate both the signal and its fragility. A prospective cross-sectional study of 155 cleft and 155 non-cleft children found a positive OSA screen in 12.2% versus 4.5% (P = 0.020), with mean PSQ scores of 0.18 ± 0.12 versus 0.13 ± 0.1 (P < 0.001) [34]. A study of 239 patients found positive PSQ screens in 18%, with class III dental occlusion and maxillary retrusion associated with positive screening (OR 2.65, 95% CI 1.2–5.8); of 25 who underwent PSG, 21 (84%) had OSA [35]. A further comparative study reported positive OSA risk in 57.6% of children with cleft lip/palate versus 28.8% without, with a relative risk of 1.7 [36]. In a Thai UCLP cohort of 102 children aged 7–15 years, 8.8% screened high-risk, and of the nine who underwent polysomnography, seven had mild and two moderate OSA, with a mean respiratory event index of 3.47 ± 1.65 events/hour [37].
The oldest and largest single-institution series remains instructive. Among 459 patients with non-syndromic cleft palate, 172 (37.5%) had symptoms of sleep-disordered breathing and 39 (8.5%) had polysomnography-diagnosed OSA; 46 patients underwent 49 studies, of which 83.1% were positive; and 89 patients (51.7%) underwent surgery for sleep-disordered breathing, most commonly adenotonsillectomy [38]. In the same report, 35 of 48 patients with Pierre Robin sequence (72.9%) had symptoms of sleep-disordered breathing or OSA [38].
3.2 The anatomical substrate
The cleft airway is not simply a narrow version of the normal airway. Unilateral and bilateral cleft lip and palate are associated with midface hypoplasia, reduced nasal cavity and nasopharyngeal dimensions, septal deviation, altered tongue posture, and — after palatoplasty — velopharyngeal dysfunction that can itself obstruct the airway [39] [40]. A cephalometric study across growth stages found posterior airway space significantly narrower at all stages in males with bilateral cleft lip and palate and after stage 1 in females, with age- and sex-dependent differences in skeletal morphology and upper-airway widths compared with controls [41].
Cross-sectional imaging studies in UCLP children show the same pattern from a different angle. Among 40 UCLP children, those with OSA had significantly greater relative tongue volume and tongue-volume-to-mandibular-volume ratios, and the AHI correlated negatively with minimum oxygen saturation — but showed no significant association with the anatomical variables measured [42]. A separate CBCT study of 40 Thai children with non-syndromic UCLP found that palatal surface area, volume, height and buccal alveolar crest width were significantly reduced in those with OSA, and that palatal volume was the only variable independently associated with AHI (β = −0.631, P < 0.001) [43]. In adults with cleft lip and palate and Class III malocclusion, 6 of 21 (29%) had OSA, and total upper-airway and oropharyngeal volumes were significantly smaller in those affected [44].
The tension between these studies is itself the point. One finds that anatomy does not explain the AHI; another finds that a single anatomical measure explains a substantial share of it. Both are small, single-centre, and cross-sectional. The honest summary is that cleft anatomy plausibly contributes to OSA risk, that the contribution is not captured reliably by any single measurement, and that a narrow maxilla is neither necessary nor sufficient for a high AHI.
3.3 The longitudinal course
Cleft-related sleep-disordered breathing is not a fixed trait; it evolves with growth and with surgery. In a cohort of infants with cleft palate, those with micrognathia had a markedly higher baseline AHI (median 20.1 events/hour, range 0.8–54.7) than infants with isolated cleft palate (3.2, range 0.3–30.7) or controls (3.1, range 0.5–23.3; P = 0.001); OSA improved with growth in isolated cleft palate and postoperatively in micrognathia, and cephalometric measures of both midface and mandibular hypoplasia correlated with severity [45].
Palatoplasty, the operation that closes the palate, has an ambivalent relationship with sleep. In 64 children (mean age 2.0 ± 2.8 years), the mean obstructive AHI was 3.4 ± 3.9 events/hour before repair and 5.9 ± 14.5 after (P = 0.30); 34.4% worsened by more than one obstructive event/hour and 18.9% by five or more, and the presence of a syndrome was a risk factor for postoperative OSA (OR 4.2, 95% CI 1.1–15.8) [46]. A longitudinal polysomnographic cohort found that only 19–27% of children with cleft palate had normal studies at any time point, with mild OSA the most frequent finding in 30–35%; more frequent desaturations were associated with special education receipt and family-reported academic concerns [23].
Two practical conclusions follow. First, a cleft child’s sleep study result at age four predicts very little about age eight, which is precisely why a pre-post expansion study in the mixed dentition cannot use the child as their own control. Second, the population is enriched for exactly the confounders that make expansion studies uninterpretable: growth, adenotonsillar change, and repeated airway surgery.
3.4 How cleft teams currently screen
Contemporary cleft and craniofacial team guidance emphasises routine screening for obstructive sleep-disordered breathing, interpretation of sleep testing, and awareness that standard management may shift in children with craniofacial conditions, including 22q11.2 deletion syndrome and Pierre Robin sequence [39]. In practice, screening is questionnaire-based and confirmatory testing is inconsistent. In one cleft cohort, one third of children with positive screens did not complete recommended testing [35]. In a Saudi screening study, most children with cleft were underweight (70%) and no demographic or clinical variable distinguished screen-positive from screen-negative children except the risk score itself [47] — a finding that should temper confidence in questionnaires as discriminative tools in this population.
This is the context in which the Bansal trial recruited. Its entry criterion was “elevated sleep questionnaire scores,” not polysomnographically confirmed OSA, and its population was a mixed-dentition UCLP cohort already embedded in orthodontic care [1]. That is a legitimate population for an orthodontic trial. It is not a population in which a change in PSG parameters can be attributed to the appliance rather than to the passage of time in a growing child with a remitting condition.
4. The Anatomical Rationale for Expansion: Nasal Floor, Nasal Resistance and Pharyngeal Volume
4.1 Why the rationale is plausible
The mechanism proposed for maxillary expansion in sleep-disordered breathing has three links. Widening the maxilla separates the palatal shelves at the midpalatal suture, lowering and widening the nasal floor. A wider nasal floor increases nasal cavity volume and reduces nasal resistance. Lower nasal resistance reduces the negative inspiratory pressure generated during sleep, which reduces pharyngeal collapse. There is a fourth, less discussed link: expansion may alter tongue posture and the position of the soft palate, and may reduce adenotonsillar size by an incompletely characterised mechanism [48].
Each link is testable, and each has been tested — with decreasing success as one moves down the chain.
4.2 Link one: the nasal cavity does enlarge
The most consistent finding in the entire expansion literature is that the nasal cavity gets bigger. A systematic review and meta-analysis of 20 studies found that total airway volume increased by 1,218.3 mm³ from baseline after RME (95% CI 702.0 to 1,734.6) and remained increased by 1,143.9 mm³ after retention (95% CI 696.9 to 1,590.9) [3]. A second meta-analysis of 12 studies concluded that nasal cavity volume increased significantly and was maintained after the retention phase, whereas nasopharyngeal and oropharyngeal volumes did not change significantly in most studies [4]. A third systematic review of 41 studies found consistent imaging evidence of nasal cavity and nasopharyngeal enlargement, with smaller or variable effects in the oropharynx, hypopharynx and maxillary sinuses [49].
The regional specificity matters clinically. The nasal cavity is the segment where expansion has its most reliable anatomical effect; the oropharynx — the segment most relevant to obstructive apnoea in children with adenotonsillar hypertrophy — is where the effect is least reliable. A 2026 CBCT study of RME in children with UCLP reported an increase in nasopharyngeal volume of 707 mm³ (P = 0.027) and in minimum cross-sectional area of 29.98 mm² (P < 0.001) [50]. A CBCT comparison of four expanders in cleft patients found that only the Hyrax and inverted mini-hyrax designs significantly increased nasal passage volume, and that none of the expanders produced significant changes in oropharyngeal measurements [12]. A comparison of bonded Hyrax and fan-type expanders in cleft patients found that airway volume increased but not significantly [13].
4.3 Link two: nasal resistance falls, but the evidence is thinner and device-dependent
Objective nasal function is measured by rhinomanometry (resistance), acoustic rhinometry (cross-sectional area and volume), and peak nasal inspiratory flow (PNIF). A systematic review and meta-analysis of 10 studies and 257 adult patients found a statistically significant reduction in nasal resistance of 0.27 Pa/cm³/s after palatal expansion (95% CI 0.15 to 0.39) and a mean reduction in the NOSE score of 40.08 points (95% CI 36.28 to 43.89), while cautioning that the evidence was too limited to recommend expansion as a primary treatment for nasal breathing [51].
In children, the picture is consistent in direction but heterogeneous in magnitude. A randomised, prospective trial in children found that PNIF improved significantly at the end of expansion (P = 0.003) and further at six months (P = 0.0002), while active anterior rhinomanometry showed no significant change over the same period — a discordance between subjective and objective measures that recurs throughout this literature [52]. Another randomised trial comparing tooth-bone-borne with tooth-borne RME found significantly higher post-expansion nasal airflow with the tooth-bone-borne appliance (mean difference 51.0 cm³/s, P = 0.018) and a similar pattern for nasal resistance [53]. A further randomised trial using acoustic rhinometry found significant increases in minimal nasal cross-sectional areas and nasal volume in all groups, with no difference between appliance designs [54].
Computational fluid dynamics studies have generally been less impressive than geometric ones. A systematic review noted that while peak nasal inspiratory flow and nasopharyngeal obstruction improved in several studies, computational fluid dynamics analyses “generally showed non-significant trends toward lower resistance” [49]. This is an important asymmetry: the airway gets geometrically larger, but the modelled resistance does not always fall proportionately, because resistance is dominated by the narrowest segment and by flow regime rather than by total volume.
4.4 Link three: skeletal change does not predict airway change
The most damaging evidence for a simple mechanical model comes from studies that examined the correlation between the amount of skeletal expansion and the amount of airway change. A study of bone-borne versus tooth-borne RME examined 480 pairings of skeletal and nasal airway measurements; only nine showed statistically significant positive correlations, the correlation data were highly variable in all categories, and the authors concluded that no realistic correlation could be demonstrated between changes in skeletal dimensions and changes in the nasal airway [55].
This is not an isolated result. A CBCT comparison of bone-borne and hybrid MARPE appliances in young adults found significant skeletal expansion with both systems but minimal airway change and no significant intergroup differences in nasopharyngeal, oropharyngeal or hypopharyngeal dimensions [14]. A prospective CBCT study of tooth-borne versus tooth-bone-borne RME found increases in anterior and middle airway cross-sectional areas, nasal cavity volume and total upper-airway volume in both groups, with greater improvement in the tooth-bone-borne group, but no significant change in nasopharyngeal volume [56]. A retrospective comparison of bone-borne and tooth-borne RME found significant increases only in nasal cavity and nasopharynx volume, with no change in oropharynx or maxillary sinus volume, and no difference between designs except for greater buccal tipping with tooth-borne appliances [57].
The clinical implication is uncomfortable for the mechanistic story that expansion “opens the pharynx.” What expansion reliably does is enlarge the nasal cavity and the nasopharynx. What it does to the oropharynx — the collapsible segment that determines whether a child obstructs — is small, inconsistent, and poorly predicted by how much the maxilla widened.
4.5 The adenotonsillar confounder inside the mechanism
One of the more intriguing findings in this literature is that RME may reduce adenoid and tonsil size. A controlled CBCT study reported a statistically significant decrease in both adenoid and tonsil volume in the expansion group compared with controls, with 90.0% and 97.5% of patients experiencing significant reduction respectively; mean volume decreases were 16.8% for adenoids and 38.5% for tonsils, and PSQ scores fell from 5.81 ± 3.31 to 3.75 ± 2.38 (P < 0.001) [48].
If true, this would be a more plausible mechanism for any sleep benefit than pharyngeal widening: less adenotonsillar tissue means less obstruction. But it also creates a serious interpretive problem. In a twelve-month CBCT study of RME, part of the upper-airway volume increase — more marked in the nasopharynx — appeared to result from reduced adenotonsillar tissue, and the authors concluded that clinicians “should not expect changes in the upper airway dimensions to be solely related to treatment effects of RME,” since normal craniofacial growth and spontaneous regression of adenotonsillar tissue could be the most significant factors [58]. In other words, the mechanism that might make expansion work is also a mechanism that operates without expansion, in every growing child, over the same time frame.
5. Expansion Technique Taxonomy: What “Clear Aligner-Based Expansion” Actually Is
5.1 The appliance families
Maxillary expansion is not one intervention. The clinically meaningful taxonomy runs along two axes: how force is delivered to the maxilla, and how much of the force reaches the skeleton rather than the teeth.
Tooth-borne appliances. The Hyrax is a banded or bonded framework with an expansion screw, anchored on the posterior teeth. The Haas appliance adds an acrylic pad against the palate, making it tooth-tissue-borne and increasing the skeletal component. These are the workhorses of paediatric practice: cheap, robust, and effective at separating the midpalatal suture in growing children.
Bone-borne and mini-screw-assisted appliances. Miniscrew-assisted rapid palatal expansion (MARPE) anchors the expansion force directly to the palatal bone via miniscrews, and hybrid designs combine screws with tooth anchorage. The rationale is to deliver more of the force to the suture and less to the teeth, reducing buccal tipping and improving skeletal efficiency, particularly in adolescents and young adults whose midpalatal suture is closing. A ten-year follow-up study found that MARPE produced more stable skeletal expansion and fewer dental side effects than conventional RPE, with an 80% relapse in buccal maxillary width at the premolar after RPE versus 30% after MARPE [59]. A CBCT comparison of bone-borne and hybrid MARPE in young adults confirmed significant skeletal expansion with both, with greater posterior buccal width increase for the bone-borne appliance but similar airway effects [14].
Surgically assisted expansion. Surgically assisted rapid maxillary expansion (SARME) and distraction osteogenesis maxillary expansion (DOME) are reserved for skeletally mature patients or severe constriction, and are outside the paediatric scope of this review except as a reminder that the same anatomical logic is applied at the other end of the age spectrum [60].
Nickel-titanium palatal expanders. NiTi appliances use shape-memory alloy springs to deliver continuous, low-magnitude force rather than the intermittent, higher-magnitude force of a manually activated screw. The claimed advantages are more physiological force delivery, less patient discomfort, and fewer visits. The evidence for superiority over screw-based appliances is limited, and the trials that exist tend to be small and single-centre.
Clear aligner-based expansion. This is the newest and least standardised category, and it is what the Bansal trial tested [1]. The term covers at least three distinct things: (i) a conventional expansion screw embedded in or attached to a clear aligner appliance; (ii) a series of aligners with progressive transverse arch form widening; and (iii) proprietary “palatal expander” aligner systems marketed for mixed dentition. These are not equivalent. A screw embedded in an aligner delivers force through the aligner body to the teeth and palate; a sequence of widening aligners delivers primarily orthodontic tipping and arch-form change.
5.2 What clear aligners can and cannot do
The systematic review evidence is reasonably clear that clear aligners can widen the maxillary arch, and reasonably clear that they do so less skeletally than conventional expanders. A 2026 systematic review and meta-analysis of 15 studies (seven in meta-analysis) found that aligners produced significantly less transverse expansion than conventional expanders: intermolar width difference −1.77 mm (95% CI −2.57 to −0.97), total palatal volume difference −460.63 mm³ (95% CI −738.32 to −182.94), and arch perimeter difference −1.75 mm (95% CI −2.71 to −0.80) [61]. The authors concluded that aligners produce expansion “mainly through dentoalveolar changes that are more evident in the anterior region and at the level of deciduous teeth” [61]. A second systematic review reached a more permissive conclusion, describing clear aligner therapy as “a reasonable alternative to traditional maxillary expanders for limited transverse maxillary deficiency” in the mixed dentition [62].
The two conclusions are compatible if read carefully: aligners can produce useful arch expansion for mild deficiency, and they produce less skeletal expansion than screw-based appliances. The clinical corollary is that an aligner-based system is appropriate when the goal is arch coordination and mild transverse gain, and less appropriate when the goal is genuine skeletal expansion of a narrow maxilla — which is precisely the situation in which an airway benefit would be hypothesised.
5.3 The comparison the Bansal trial actually made
This is the crux of interpreting the seed trial. Bansal and colleagues randomised children to a clear-aligner-based expansion screw or a nickel-titanium palatal expander [1]. Both are expansion devices. The trial therefore compared two delivery systems, not expansion versus no expansion, and not even expansion versus a sham. The finding that the aligner-based system produced greater intercanine and intermolar width gains [1] is a finding about dental change, and it sits somewhat awkwardly with the systematic review evidence that aligners produce less skeletal expansion than conventional appliances [61] — a discrepancy that could reflect the specific device, the activation protocol, the age of the children, or measurement differences. None of these can be resolved from the abstract.
The clinically relevant point is simpler. A trial comparing two active expansion systems cannot tell a clinician whether to expand. It can only inform the choice of hardware once the decision to expand has been made on other grounds — and even then, a 24-patient trial is a weak basis for that choice.
5.4 Skeletal versus dental: the split that determines everything
Every expansion appliance produces a mixture of skeletal widening and dental tipping. The proportion determines whether the nasal floor actually descends and whether the result is stable. Buccal tipping of the maxillary molars is the classic dental side effect: the crowns move laterally while the roots stay put, producing the appearance of expansion without skeletal gain.
CBCT studies consistently show that tooth-borne appliances produce more buccal tipping than bone-borne designs. In the bone-borne versus tooth-borne comparison, both groups showed significant increases in intermolar and maxillary width, but buccal inclination of the maxillary molars increased significantly only in the tooth-borne group [57]. A meta-analysis of root resorption found that tooth-borne RME resulted in significantly greater root resorption volume (−20.7 mm³) and length (−0.40 mm) than baseline, and recommended bone-borne expanders for high-risk patients [63].
The Bansal trial’s own correlation analysis is interesting in this light. It reports that anterior transverse widening (intercanine and intermolar width) correlated with airway volume gains, whereas basal skeletal widening (J-J’ width, basal bone width) and molar inclination did not [1]. If accurate, this inverts the usual mechanical logic: the dental change correlates with the airway change, and the skeletal change does not. That is biologically odd, and it is one of several reasons to treat the trial’s correlation findings as hypothesis-generating rather than mechanistic.
5.5 Activation protocols and the meaning of “rapid”
Rapid maxillary expansion conventionally means one or two turns per day of a screw with roughly 0.2–0.25 mm of expansion per turn, delivering several hundred grams of force and separating the suture within weeks. Slow expansion uses fewer turns per week and relies more on orthodontic tooth movement and sutural remodelling. Alternative protocols — alternate rapid maxillary expansion and constriction (Alt-RAMEC), and various hybrid protocols — aim to mobilise the suture more gently before protraction.
A randomised trial of miniscrew-supported versus conventional expansion in adolescents found that a slow activation protocol produced a better overall patient experience than a rapid one, with mild-to-moderate discomfort and functional limitation anticipated during activation [64]. A prospective observational study found that pain was greatest in the first days of treatment and was significantly greater with the Hyrax than the Haas appliance (P < 0.001), greater with anchorage on first permanent molars than second primary molars (P < 0.001), and modulated by age (P < 0.001) [65].
The trial by Bansal and colleagues does not report its activation protocol in the abstract [1]. This matters because the two appliances being compared plausibly differ in force magnitude, force continuity and rate of delivery, and because activation protocol is one of the few modifiable determinants of both skeletal efficiency and patient experience.
5.6 Retention and relapse
Expansion is not finished when the screw stops turning. Relapse is common, particularly dental relapse in tooth-borne cases, and retention protocols vary widely. The ten-year MARPE versus RPE comparison found 80% relapse of buccal maxillary width at the premolar after RPE versus 30% after MARPE [59]. A survey of cleft teams found that the most common complications of expansion in cleft patients were asymmetric expansion, relapse and fistula formation, and that protocols and materials vary widely between centres with a persistent lack of consensus [66].
For a sleep endpoint, relapse has a specific consequence: if the airway benefit depends on maintained skeletal widening, then a treatment that relapses by 80% at ten years cannot be expected to deliver a durable sleep benefit. No expansion trial in this literature has followed airway or sleep outcomes for long enough to answer that question.
6. CBCT Upper-Airway Volumetry: What It Measures, How Reliable It Is, and Why Static Volume Is a Weak OSA Surrogate
6.1 What the measurement actually is
A CBCT airway volume is not a direct measurement of anything physiological. It is a segmentation product: the scan is thresholded to separate air-density voxels from soft tissue, the resulting region is cropped by anatomical boundaries chosen by the operator (usually the posterior nasal spine, the hard palate, the epiglottis, the posterior pharyngeal wall), and the enclosed volume is computed. Every step involves a decision.
The boundaries are not standardised. A study comparing different upper and lower boundaries for airway assessment found high correlation between boundary definitions (r = 0.896–0.999 for volume), which is reassuring for some purposes but also means that different studies are not necessarily measuring the same anatomical compartment even when they use the same words [67]. “Total pharyngeal airway volume” in one paper may exclude the nasopharynx that another includes.
6.2 Reliability: moderate to excellent within a protocol, uncertain across protocols
The most comprehensive reliability assessment is a systematic review of 42 studies representing CBCT scans of 956 patients. It found that under specific restricted conditions there is moderate to excellent intra- and inter-examiner reliability, and that airway volume has greater reliability than minimum cross-sectional area. But only five of the 42 studies were deemed high quality, and the review concluded that reliability “has not been adequately established” because of methodological limitations, most importantly the lack of manual image orientation and of a stated threshold sensitivity in study protocols [68].
Newer segmentation methods perform well against manual reference standards. A validation study found excellent reliability for AI-automated segmentation (ICC 0.945) and semiautomated segmentation (ICC 0.992) for total volume, with excellent intrareader repeatability, and a substantial time advantage for the AI method (under 30 seconds versus 161.4 seconds semiautomated and 336.6 seconds manual) [69]. A comparison of six software programs found all were reliable but all had errors: with an interactive threshold protocol, four programs had less than 2% volume error while two had more than 5%; with a fixed threshold protocol, all programs underestimated volume by 11.1–11.7% [70]. A further comparison found significant differences in accuracy among software tools, with one program showing systematic bias and wider limits of agreement [71].
The practical translation is that a 10–20% change in airway volume is at the edge of what different software, thresholds and operators can reliably distinguish. A study designed to detect such a change needs a pre-specified, locked segmentation protocol on the same software by the same operator, blinded to time point. The Bansal abstract does not state whether any of this was done [1].
6.3 The longitudinal measurement problem
The most damaging study for the use of CBCT volumetry as an outcome measure is not about reliability between examiners but about error within the same patient over time. A study of successive measurement errors in consecutive CT scans found almost perfect intra-examiner reliability for upper-airway analysis, yet average successive measurement errors were notably high, particularly for intraoral and oropharyngeal airway volumes; eight of 26 common measurements had errors greater than one unit, and the oropharyngeal volume error on CBCT was more than three times that on MDCT. The authors concluded that the findings “do not support the quantitative use of CT for analyzing changes in airway dimensions for research purposes” [72].
That is an extreme position, and it is about CT generally rather than CBCT specifically. But it is a serious challenge to the assumption that a statistically significant volumetric change in a small sample represents a real anatomical change rather than accumulated measurement noise plus positional variation.
Positional and physiological variation compounds the problem. A longitudinal reliability study in untreated adults found good to excellent reliability for all parameters (ICCs > 0.80), yet significant differences between time points for nasopharyngeal volume (6,577.79 ± 2,061.56 mm³ versus 6,942.93 ± 1,836.83 mm³, P = 0.013), total airway height (P = 0.024) and hypopharyngeal height (P = 0.023), with no change in cranio-cervical angles [73]. In other words, the same untreated people measured twice had a statistically significant change in nasopharyngeal volume. Head position, tongue posture, swallowing phase and respiratory phase all influence pharyngeal dimensions, and none of them are held constant in routine clinical scanning.
6.4 Does static airway volume predict OSA severity?
Only weakly, and less in children than the field would like. A study relating three-dimensional airway measurements to the AHI in 99 paediatric CBCT scans (median age 11 years) found that nasopharyngeal volume correlated with AHI at approximately rho = −0.4 (P < 0.05), and proposed age-specific cut-offs (2,400 mm³ and 1,600 mm³ for AHI ≥5 and ≥10 in 7–11-year-olds; 3,500 mm³ and 2,700 mm³ in 12–17-year-olds) [15]. A correlation of −0.4 means roughly 16% of variance explained. That is not nothing, and the authors argue — against the adult literature — that nasopharyngeal volume has screening promise in children. But a variable explaining 16% of variance cannot carry a treatment indication.
The paediatric physiology literature explains why. In 7–12-year-old children, the ratio of retropalatal airway cross-sectional area to soft palate cross-sectional area correlated inversely with the obstructive AHI at r² = 0.49, and the high-AHI group had a narrower retropalatal airway in exactly the region where soft palate, adenoids and tonsils overlap [74]. That is a much stronger anatomical signal than raw volume — but it is a ratio that captures the balance between soft tissue and airway space, not the size of the airway alone. It is the difference between measuring a room and measuring how much furniture is in it.
Two further observations sharpen the point. First, in children undergoing adenotonsillectomy — an intervention that produces large, real changes in both airway volume and AHI — no significant correlation was found between PSG parameters and airway dimensions before surgery, or between changes in PSG parameters and changes in airway dimensions twelve months after surgery, despite airway volume roughly doubling [16]. If volume change and apnoea change do not correlate even when the intervention is removal of the obstructing tissue, the case for volume as a surrogate is very weak. Second, in a study of RME in individuals with sleep-disordered breathing, oropharyngeal volume increased significantly while oxygen saturation and AHI did not change significantly, and no correlation was found between total airway volume, oxygen saturation and AHI changes [5].
6.5 What a defensible imaging claim looks like
None of this means CBCT volumetry is worthless. It means that it measures anatomy, and should be reported as anatomy. A defensible claim is: “expansion increased nasopharyngeal volume by X mm³ using a locked segmentation protocol with an ICC of Y.” An indefensible claim is: “expansion improved the airway,” where “improved” silently imports a functional meaning that the measurement does not carry.
The Bansal trial reports that both modalities “significantly improved pharyngeal airway volume” without reporting the volumes, the segmentation method, the reliability data, or the regional breakdown beyond the correlation analysis [1]. The word “improved” is doing work that the data, as reported, cannot support.
7. Does the Measured Airway Change? The Imaging Literature and Its Quality
7.1 The shape of the evidence
There are now dozens of studies reporting airway changes after maxillary expansion, and their aggregate message is more modest than their individual abstracts suggest. Three systematic reviews with meta-analysis converge on a consistent pattern: nasal cavity volume increases, nasopharyngeal volume may increase modestly, oropharyngeal and hypopharyngeal volumes change inconsistently, and the overall quality of evidence is low to very low [3] [4] [49].
The largest pooled estimate comes from 20 studies (17 in quantitative synthesis, including 3 controlled clinical studies and 14 cohort studies). Total airway volume increased by 1,218.3 mm³ from baseline (95% CI 702.0 to 1,734.6) and by 1,143.9 mm³ after retention (95% CI 696.9 to 1,590.9). The authors judged the overall quality of evidence “very low” because of methodological limitations, absence of untreated control groups, and inconsistency among studies [3].
The second meta-analysis, of 12 studies (2 randomised trials, 9 non-randomised clinical trials, 1 non-randomised study), found a significant increase in nasal cavity volume maintained after retention, but no significant change in nasopharyngeal and oropharyngeal volume in the majority of studies. Its conclusion is worth quoting because it anticipates the entire argument of this review: “This increase in volume may not be considered as an equivalent for enhancement of airway and function unless proven so” [4].
7.2 Where the studies disagree
Disagreement is not random; it clusters around region, appliance and control group.
Region. Studies that measure the nasal cavity find effects; studies that measure the oropharynx often do not. A four-expander CBCT comparison in cleft patients found that only Hyrax and inverted mini-hyrax increased nasal passage volume, and that “none of the expanders caused significant changes in the oropharyngeal measurements” [12]. A bonded Hyrax versus fan-type comparison in cleft patients found that airway volume increased but not significantly [13]. A bone-borne versus hybrid MARPE comparison in young adults found minimal airway change in both groups with no intergroup differences in nasopharyngeal, oropharyngeal or hypopharyngeal dimensions [14].
Appliance. Some studies find greater airway gain with bone-borne or tooth-bone-borne designs [56], others find no difference between designs [14] [57], and at least one found no difference in nasal outcomes despite differences in skeletal outcome [54]. A short-term study of tooth-bone-borne versus bone-borne expansion reported significant increases in total nasopharyngeal volume (0.59 ± 1.42 cm³), total oropharyngeal volume (3.83 ± 7.53 cm³) and minimum oropharyngeal cross-section (53.23 ± 126.46 mm²) across all cases, with no difference between MARPE and bone-anchored maxillary expansion except for minimum nasal cross-sectional area [75]. The standard deviations in that study are larger than the means — a detail that deserves more attention than it usually receives.
Control group. This is the decisive variable. A controlled prospective study in children with OSA and posterior crossbite, with 39 untreated controls, found a significant increase in nasopharyngeal width in the treated group and no significant change in controls [76]. That is one of the few studies in the literature designed in a way that allows the change to be attributed to treatment rather than to time, and its outcome is narrower than the phrase “airway improvement” implies.
7.3 The problem of the growing control
The absence of untreated controls is not a neutral omission. A twelve-month CBCT study found that while all pharyngeal sub-regions increased significantly after RME, part of the increase — more marked in the nasopharynx — appeared to result from reduced adenotonsillar tissue, and concluded that “normal craniofacial growth changes and spontaneous regression of the adenotonsillar tissue could represent the most significant factors influencing” upper-airway changes [58].
The same study found that adenoid and tonsil volumes decreased by 16.8% and 38.5% respectively in the expansion group [48] — a finding that could be a genuine treatment effect, a growth effect, or a measurement artefact of soft-tissue segmentation on CBCT. Without an untreated comparison group scanned at the same intervals with the same protocol, there is no way to distinguish these.
7.4 What the imaging literature supports
A fair reading is that maxillary expansion produces a real, modest, regionally specific increase in nasal cavity and nasopharyngeal dimensions, of the order of 1 cm³ in pooled estimates, with uncertain effects beyond the nasopharynx, and with a measurement error that may be of similar magnitude to the effect in individual patients. The literature does not support the broader claim that expansion “increases the pharyngeal airway” as a general statement, and it does not support using airway volume as a surrogate for sleep outcomes.
8. Does the Apnoea Change? Expansion for Paediatric OSA and the Missing Adenotonsillectomy Comparator
8.1 Two literatures that reach opposite conclusions
The expansion-OSA literature contains two bodies of evidence that are rarely presented side by side. Uncontrolled pre-post studies and their meta-analyses report large AHI reductions. Controlled comparisons report little or no effect. The difference between them is not the quality of the statistical methods; it is the presence of a comparator.
The uncontrolled estimates. The most cited meta-analysis pooled 17 studies and 314 children (mean age 7.6 ± 2.0 years) with high-arched or narrow hard palates and OSA. For follow-up of three years or less, the AHI fell from 8.9 ± 7.0 to 2.7 ± 3.3 events/hour, a 70% reduction; the cure rate (AHI <1/hour) among the 90 patients for whom it could be calculated was 25.6%; lowest oxygen saturation rose from 87.0 ± 9.1% to 96.0 ± 2.7%; and the random-effects standardised mean difference for AHI was −1.54, a large effect. AHI improved more in children with previous adenotonsillectomy or small tonsils (73–95% reduction) than in children with large tonsils (61% reduction) [33].
A second meta-analysis of 10 articles and 215 children (mean age 6.7 years, 58.6% male) reported a mean AHI change of −6.86 (P < 0.0001) [77]. A 2026 meta-analysis reported an AHI mean difference of −4.04 (95% CI −6.39 to −1.70, P = 0.0007) within six months of completing active RME, with I² = 93%, and −6.15 (95% CI −11.66 to −0.64, P = 0.03) within twelve months, with I² = 97% [78]. Heterogeneity of 93–97% means that the pooled estimate describes a distribution so wide that it barely constrains what a clinician should expect for an individual child.
The controlled estimates. A systematic review that restricted inclusion to controlled studies found only one randomised trial comparing RME with watchful waiting, plus four studies comparing RME with adenotonsillectomy, three of them non-randomised. The randomised trial showed no statistically significant difference in AHI improvement or secondary outcomes between RME and watchful waiting. The authors concluded that “in the absence of solid evidence with RCT, RME should not be recommended for the treatment of pediatric OSA” [7].
A 2026 update of the same review searched for controlled prospective studies published to February 2026 and identified three new randomised trials in addition to the five original studies. Critically, none of the new trials included a watchful-waiting or supportive-care control arm, leaving the sole available comparison against watchful waiting unchanged. Of the new trials, two found marginal AHI improvement with RME and one found no significant AHI reduction. The authors’ conclusion: “The observed AHI improvements in studies lacking an untreated comparator cannot be distinguished from the known potential for spontaneous improvement in growing children. Extreme caution is warranted before recommending RME as a treatment for pediatric OSA” [11].
A network meta-analysis provides the cleanest quantitative expression of the contrast. It included 11 studies and 595 patients and compared mandibular advancement appliances, RME, and myofunctional therapy against control. Mandibular advancement appliances reduced the AHI by −2.18/hour (95% CI −3.48 to −0.89); combined RME plus adenotonsillectomy reduced it by −5.13/hour (95% CI −7.50 to −2.76); RME plus mandibular advancement reduced it by −3.79/hour (95% CI −5.21 to −2.37). RME alone was not associated with significant AHI reduction: 0.02/hour (95% CI −1.72 to 1.75, P = 0.985). The network heterogeneity was I² = 32.6%, far lower than the pre-post meta-analyses, which is what one would expect if the pre-post heterogeneity were driven by uncontrolled variation in natural history [6].
A second network meta-analysis of 14 trials and 1,064 children with adenotonsillar hypertrophy found that surgical approaches were the most effective interventions for AHI reduction, that RME was one of the most effective interventions for improving lowest oxygen saturation, and that no comparison showed statistical significance for the oxygen desaturation index [79]. That is a more favourable reading of RME than the other network meta-analysis, and the difference is instructive: RME may have a real effect on nadir oxygen saturation even where its effect on AHI is not demonstrable against control. Oxygen saturation and apnoea counts are not the same outcome, and a treatment can move one without the other.
8.2 The individual trials
The pre-post literature includes several studies that are frequently cited as stronger evidence than they are. Villa and colleagues reported that in 14 children who completed follow-up, the AHI decreased significantly (P = 0.005), as did the hypopnoea obstructive index (P = 0.002) and arousal index (P = 0.001) [80]. A 36-month follow-up reported that AHI decreased and symptoms resolved by the end of treatment, with no significant change 24 months later [81]. A larger retrospective series of 40 children with moderate adenotonsillar hypertrophy found that 34 (85%) had an AHI decrease greater than 20% (mean ΔAHI 67.45% ± 25.73%) and were classified as responders, while 6 (15%) were non-responders; 57.5% had residual OSA (AHI >1/hour) after treatment [82]. A study of 52 children compared adenotonsillectomy (25 children), RME (22 children) and both (5 children), finding improvements in all groups but with baseline differences that make the groups non-comparable — the RME group was older, had longer disease duration, higher BMI, lower AHI and lower arousal index at baseline [83].
An earlier controlled comparison in preschool children with OSA and retrognathic jaws reported that the intervention group showed significantly greater increases in SNA (P = 0.004) and SNB (P = 0.003) than the comparison group, without a between-group difference in skeletal divergence or total facial height; the abstract reports cephalometric outcomes rather than a between-group AHI comparison [84]. A study of sleep microstructure found that RME almost normalised sleep architecture and improved respiratory disturbances, but that sleep microstructure and respiratory parameters did not completely recover, with persistent elevation of cyclic alternating pattern rate in slow-wave sleep [85].
These are real findings from real children. The problem is that none of them can separate the appliance from the passage of time in a population with a high spontaneous remission rate.
8.3 The adenotonsillectomy comparator that is usually missing
The single most important omission in the expansion literature is not a placebo — it is adenotonsillectomy. Adenotonsillectomy is the first-line treatment for paediatric OSA with adenotonsillar hypertrophy, and it has been tested in randomised trials of a size no expansion study approaches [9].
CHAT. 464 children aged 5–9 years, randomised to early adenotonsillectomy or watchful waiting. Polysomnographic normalisation occurred in 79% versus 46%; behaviour, quality of life and symptoms improved more with surgery; attention and executive function did not differ significantly (7.1 ± 13.9 versus 5.1 ± 13.4, P = 0.16) [8]. Cognitive effects on secondary analysis were small and selective (Cohen’s d 0.20–0.24) [17]. Parent-reported behaviour improved even when sleep-specific items were excluded [24].
KATE. 60 children aged 2–4 years with an obstructive AHI ≥2 and <10, randomised to adenotonsillectomy (n = 29) or watchful waiting (n = 31) with three-year follow-up. The change in OAHI did not differ (Cohen’s d 0.05, 95% CI −0.7 to 0.6), while the OSA-18 total score favoured surgery with a medium effect size (d 0.54, 95% CI −1.3 to 0.1). Thirteen of 31 children in the watchful-waiting arm (42%) crossed over to surgery [32].
PATS. 458 children with mild sleep-disordered breathing; adenotonsillectomy reduced blood pressure percentiles relative to watchful waiting (systolic difference −9.02, 95% CI −14.77 to −3.28) [26]; 13% of controls progressed to an OAHI ≥3 at 12 months while 57% had persistent symptoms [31]; surgery reduced health care encounters by 32% and prescriptions by 48% [28].
Head-to-head expansion versus adenotonsillectomy. One randomised trial has compared them directly. Twenty-four children with paediatric OSA, significant adenotonsillar hypertrophy and transverse maxillary deficiency were randomised to adenotonsillectomy or RME, with type I polysomnography at baseline and six months. Median baseline AHI was 7.0 (5.25–9.9) in the adenotonsillectomy group and 6.85 (5.6–8.05) in the RME group. Both improved significantly. There was no significant difference in PSG parameters (AHI, lowest saturation, mean saturation, REM sleep time) or cure rate between groups; post-treatment AHI was 1.4 (0.7–1.85) with adenotonsillectomy and 2.3 (1.15–5.7) with RME. However, PSQ and OSA-18 scores were significantly higher (worse) in the RME group [86].
That trial is the most informative single study in this literature for a clinician, and it is worth being precise about what it shows. In 24 children with both adenotonsillar hypertrophy and transverse deficiency, RME produced apnoea outcomes statistically indistinguishable from surgery, but children felt worse on symptom and quality-of-life measures. A plausible interpretation is that removing the obstructing lymphoid tissue relieves symptoms faster and more completely than widening the skeleton, even when the apnoea count ends up similar. Another interpretation is that the trial was too small to detect a difference in PSG parameters and that the quality-of-life difference is the more reliable signal. Both interpretations argue against expansion as a substitute for adenotonsillectomy.
Sequential and combination therapy. An early randomised trial of treatment sequence enrolled 32 prepubertal children with moderate OSA judged to require both adenotonsillectomy and orthodontic expansion, randomising the order of the two treatments. Two children treated with orthodontics first did not require subsequent adenotonsillectomy; 30 underwent both; two remained symptomatic with abnormal polysomnography after both; and in the remaining 28, results after both treatments were significantly better than at entry (P = 0.001) and better than after single therapy (P = 0.01), regardless of order [87]. A retrospective analysis of adenotonsillectomy and RPE found that initial AHI severity and adenotonsillectomy as the first intervention explained 94.9% of AHI improvement, and that RPE had a marginal effect on AHI and minimum oxygen saturation when adjusted for confounders [88]. A 2026 study of maxillary arch expansion and adenotonsillectomy in prepubertal children found that combined therapy produced substantial reductions in the respiratory disturbance index compared with individual interventions, with treatment order and BMI associated with different final outcomes [89]. A randomised comparison of dentofacial orthopaedic treatment with adenotonsillectomy in mild-to-moderate OSA found a mean AHI reduction of 1.6 events/hour with orthopaedic treatment and 1.3 with adenotonsillectomy, with no significant difference (−0.43, 95% CI −1.48 to 0.61, P = 0.40) [90].
The pattern across these studies is consistent: adenotonsillectomy does most of the work; expansion adds skeletal correction and may help a subset; combination therapy is plausible; and no study demonstrates that expansion alone substitutes for surgery in a child with adenotonsillar hypertrophy.
8.4 Where the reviews land
The review literature has moved from optimism to scepticism within a decade. A 2017 meta-analysis concluded that improvement in AHI and lowest oxygen saturation “has consistently been seen” and called for randomised trials [33]. A 2017 evidence-based dentistry appraisal concluded that RME reduces AHI and is an appropriate alternative [91]. A 2023 umbrella review of seven systematic reviews with meta-analysis found “no consistent evidence favoring RME for long-term treatment of OSA in children,” noted considerable heterogeneity due to variability in age and follow-up, and concluded that “RME is not recommended for treating OSA in children” [92]. A 2023 narrative review in Current Opinion in Pediatrics described RPE as showing promise as an additional treatment while noting that long-term efficacy and safety require further investigation [93]. A 2025 systematic review of maxillary transverse deficiency and respiratory problems found that polysomnography indicated reductions in AHI and improved oxygenation in some paediatric OSA cohorts, “although other RCTs reported null PSG effects” [49].
The gradient is unmistakable: the more controlled the design, the smaller the effect. That gradient is the single most important finding in this literature, and it is not an artefact of statistical technique. It is what happens when you stop comparing children to themselves.
9. Reading the Seed Trial: Design, Size, Entry Criterion, Blinding, Multiplicity
9.1 What the trial did
The trial randomised 24 patients with UCLP, transverse maxillary deficiency and elevated sleep questionnaire scores to a clear aligner-based expansion screw or a nickel-titanium palatal expander. CBCT and polysomnography were performed before and after expansion. Outcomes included airway volume, PSG parameters, CBCT transverse measurements (intercanine width, intermolar width, J-J’ width, basal bone width, buccal tipping of molar), and correlations between expansion characteristics and airway or PSG outcomes [1].
The setting was a single orthodontic department in Faridabad, Haryana, India. The article is indexed as a randomised controlled trial and comparative study. The authors declare no conflicts of interest. The trial is registered as CTRI/2024/04/066366 [1] [94]. The registry record itself could not be retrieved during preparation — the host resolved but connections were refused — so the registration number is reported as it appears in the article and has not been independently verified against the registry entry.
9.2 What the abstract reports, and what it does not
This distinction is essential, because the abstract is the version of the trial that most clinicians will ever read, and it is the version that will be cited.
| Domain | Reported in the abstract? | Detail |
|---|---|---|
| Design | Yes | Randomised clinical trial, two active arms, single centre |
| Sample size | Yes | 24 patients with UCLP |
| Per-arm sample size | No | Not stated; ~12 per arm can be inferred but is not reported |
| Losses to follow-up | No | Not stated |
| Baseline AHI | No | No numerical AHI values at any time point |
| Post-treatment AHI | No | Reported only as “significantly improved” |
| Airway volumes | No | No numerical volumes; only direction and significance |
| Effect sizes | No | No mean differences, no confidence intervals |
| p-values | No | No numerical p-values |
| Scorer blinding | No | Not stated for PSG or CBCT segmentation |
| PSG scoring rules | No | AASM version, hypopnoea definition, same-scorer status not stated |
| Randomisation method | No | Sequence generation, allocation concealment not stated |
| Activation protocol | No | Not stated |
| Multiplicity correction | No | Not described |
| Correlations | Yes, qualitatively | Positive correlations for intercanine/intermolar width with airway volume; negative correlation between skeletal expansion and Δminimum SpO₂ in the NiTi group |
| Adverse events | No | Not reported |
This table is not a criticism of the authors, who were writing within a journal’s word limit. It is a description of the evidentiary surface available to a reader who does not have the full text. Every statement in this review about the trial is confined to what the abstract contains, and where a number would normally be expected, this review says it was not reported rather than supplying one.
9.3 The four structural problems
No untreated arm. Both groups received active expansion. Within-group improvement therefore cannot be attributed to the appliance rather than to growth, time, maturation of the pharyngeal airway, adenotonsillar involution, or regression to the mean in children selected for elevated sleep scores [1]. The magnitude of the last factor is not speculative: 42% of children with OSA randomised to watchful waiting in CHAT no longer met polysomnographic criteria at seven months [2], and in PATS 13% of controls progressed while 57% had persistent symptoms [31].
Sample size. Twenty-four patients across two arms is underpowered for between-group inference and, more importantly, for the correlation analyses reported. A correlation matrix across roughly six transverse variables and several airway and PSG outcomes, in 24 children, invites multiplicity. With 24 participants, the 95% confidence interval around a correlation of 0.5 spans roughly 0.1 to 0.75. A “significant” correlation in this setting is a signal, not a finding. The abstract does not describe any correction for multiple comparisons [1].
Unblinded, single-centre conduct. CBCT segmentation and PSG scoring can be blinded in principle, and should be in any trial where the outcome is a measurement subject to expectation. The abstract does not state that they were [1]. Single-centre conduct limits generalisability, particularly in cleft care, where surgical and orthodontic protocols vary widely between centres — a point made explicitly by a survey of cleft teams that found “a lack of consensus” on expansion protocols and materials [66].
Highly selected population. Mixed dentition, UCLP, transverse maxillary deficiency, and elevated sleep questionnaire scores is a narrow intersection. Generalisation to ordinary paediatric OSA, to non-cleft transverse deficiency, or to bilateral cleft lip and palate is not automatic [1].
9.4 The PSG problem in 24 children
Paediatric polysomnography is technically demanding, and its measurement properties in small samples are poor. The abstract does not report the scoring rules (AASM version), the definition of hypopnoea used, or whether the same scorer read both nights [1]. These are not administrative details. The AHI is sensitive to the hypopnoea definition, and the night-to-night intraclass correlation coefficient for the obstructive AHI in children has been reported at 0.27 (95% CI −0.04 to 0.51), with 26% of children with moderate-to-severe OSA reclassified as mild or none on a second night [22]. A trial that measures the AHI twice in 24 children, without reporting scorer blinding or reliability, is measuring a variable whose within-child variability may exceed the between-group difference it hopes to detect.
There is also a subtler issue: polysomnography in a child with a cleft palate and a repaired velopharyngeal mechanism may be affected by the mechanics of the repair itself, by nasal obstruction, and by the arousal threshold. None of these are reported, and all of them bear on whether the measured change reflects the appliance.
9.5 The correlation findings, read correctly
The trial reports positive correlations between Δintercanine width and Δnasopharyngeal, Δoropharyngeal and Δtotal pharyngeal airway volume, and between Δintermolar width and Δoropharyngeal volume; no significant correlations for J-J’ width, basal bone width, or buccal tipping with airway or PSG parameters; and a significant negative correlation between skeletal expansion and Δminimum SpO₂ in the NiTi group [1].
Three observations.
First, the positive correlations are biologically counter-intuitive in the direction reported. Intercanine width is the most dental of the transverse measures; basal bone width and J-J’ width are the most skeletal. If expansion worked by widening the nasal floor, one would expect the skeletal measures to correlate with airway change and the dental measures not to. The trial finds the opposite, and the authors interpret this as evidence that “the overall expansion effect rather than its mechanistic subtype may be the primary driver of functional benefit” [1]. That is a reasonable attempt to make sense of an awkward result, but it is a post hoc interpretation of a correlation matrix in 24 children, not a mechanism.
Second, the negative correlation between skeletal expansion and Δminimum SpO₂ in the NiTi group is a single finding in a subgroup of roughly twelve children, reported without an effect size or confidence interval. A negative correlation means more skeletal expansion was associated with less improvement in nadir oxygen saturation. That is the opposite of the hypothesised mechanism. The abstract does not offer an explanation and does not report whether the finding survived any adjustment [1]. It should be reported as hypothesis-generating, not as a mechanism — and it is a useful reminder that in a 24-child trial with many comparisons, one counter-intuitive “significant” correlation is roughly what one expects by chance.
Third, the absence of correlation between basal skeletal widening and airway or PSG outcomes is consistent with the wider literature, which has repeatedly failed to find a relationship between the magnitude of skeletal expansion and the magnitude of airway change. The 480-pairing study found only nine significant correlations and concluded that no realistic relationship could be demonstrated [55]. The Bansal trial’s null findings for J-J’ and basal bone width are therefore not anomalous; they are the norm. It is the positive dental correlations that require explanation.
9.6 What the trial can and cannot support
It can support the statement that, in 24 children with UCLP and transverse deficiency, two expansion systems produced measurable transverse dental and skeletal changes, with greater anterior width gain in the aligner-based group, and that airway volume and sleep parameters improved in both groups over the study period.
It cannot support the statement that either appliance improves sleep or airway outcomes relative to no treatment, because there was no untreated arm. It cannot support a claim of equivalence between the appliances, because 24 patients cannot exclude a clinically meaningful difference. It cannot support a mechanistic claim about which component of expansion drives airway change, because the correlation analysis is underpowered and uncorrected for multiplicity. And it cannot support generalisation to non-cleft children, to other cleft types, or to older or younger age groups.
10. The Growth Confounder: What Happens to the Paediatric Airway Without Intervention
10.1 The airway grows
The pharyngeal airway is not a fixed structure that is either obstructed or not. It grows, changes shape, and changes its soft-tissue contents throughout childhood. A cephalometric study of bilateral cleft lip and palate patients across four growth stages found that posterior airway space was narrower at all stages in males and after stage 1 in females compared with controls, while middle airway space was wider after stage 1 — a pattern of region-specific, sex-specific and stage-specific change that no single cross-sectional measurement can capture [41]. A study of adolescents with different craniofacial growth patterns found significant year-by-growth-pattern interactions in nasopharyngeal and lower airway dimensions, with the 2020 sample showing significantly smaller nasopharyngeal widths than the 2005 sample in the vertical growth pattern group (PNS-UPW −3.10 mm, PNS-AD2 −2.75 mm) [95]. Secular change in airway dimensions is real, and it is of the same order of magnitude as the treatment effects being claimed.
10.2 The adenotonsillar clock
Adenoid and tonsillar tissue grows in early childhood and involutes during later childhood and adolescence. This is the physiological basis of the spontaneous remission seen in CHAT and PATS, and it is the single largest confounder in any paediatric airway study without a control group. A longitudinal cohort of children with sleep-disordered breathing found a baseline prevalence of 28.9%, with adenotonsillar hypertrophy (adjusted OR 3.42), paternal snoring (aOR 2.38), obesity (aOR 2.15) and allergic rhinitis (aOR 1.86) as independent risk factors; children with SDB had reduced airway volume, mandibular retrusion and a 38.2% smaller minimum cross-sectional area. In an intervention sub-study, ENT treatment produced 68.4% resolution and orthodontic treatment 52.1% [96].
That last comparison is instructive and rarely cited. In an observational cohort, orthodontic treatment was associated with a 52.1% resolution rate — impressive in isolation, and almost certainly inflated by the same natural history that produces 42% polysomnographic remission in untreated children [2]. The two numbers are not directly comparable, but their proximity is a warning.
10.3 The arch-growth literature makes the same point
A two-centre longitudinal study of 4,271 children contributing 4,907 temporally ordered radiograph pairs found a reproducible breakpoint in the adenoid-nasopharyngeal ratio at approximately 0.62–0.69, with positive dose-response relationships between adenoid size and craniofacial angle change, on the order of 10–25° per year per unit increase in A/N ratio for some measures [97]. The direction of causality in that relationship is contested — does adenoid hypertrophy alter craniofacial growth, or does craniofacial growth alter the A/N ratio? — but either way, the airway and the face are changing together throughout the period in which expansion trials are conducted.
10.4 Why “the child as their own control” fails
The logic of the pre-post design is that each child serves as their own control, so between-child variation is removed. That logic is sound for a stable trait measured with error. It fails completely for a trait that is changing systematically over time, because the “control” condition is not the same child at the same time — it is the same child at a different time, and time itself is the intervention.
Three specific failures follow.
Failure one: regression to the mean. Children are enrolled because they have elevated sleep questionnaire scores. Some of that elevation is true signal; some is measurement error and transient fluctuation. On average, the second measurement will be closer to the population mean regardless of treatment. A trial that enrols on the basis of a high score and reports improvement at follow-up has demonstrated regression to the mean unless it has a control group. The PSQ’s diagnostic performance makes this worse: a meta-analysis of diagnostic test accuracy found pooled sensitivity of 0.70 and specificity of 0.43 at an AHI cut-off of ≥1, with a pooled correlation between AHI and PSQ of 0.28 (95% CI 0.15 to 0.40) [98]. A screening instrument with 43% specificity selects a population in which a large fraction do not have the condition, and in whom the index measurement is most likely to fall on retesting.
Failure two: maturation of the pharyngeal airway. The pharyngeal airway matures and its soft-tissue contents change throughout the mixed dentition. A child treated at age 9 and measured at age 10 has a different airway for reasons that have nothing to do with the appliance.
Failure three: concurrent care. Cleft children are embedded in multidisciplinary care. Between two time points they may undergo alveolar bone grafting, adenotonsillectomy, speech therapy, or a change in nasal medication. The Bansal abstract does not report whether concurrent treatments were controlled [1], and in a 24-patient trial even a few such events can dominate the result.
10.5 What a growth-adjusted design would require
The minimum requirement is an untreated or delayed-treatment comparison group scanned and sleep-studied on the same schedule with the same protocol. A stronger design adds a pre-specified growth covariate (skeletal maturation stage, cervical vertebral maturation, or age-sex-specific normative data), blinded central scoring, and a primary outcome that is either a change score adjusted for the control group’s change or a between-group difference at a fixed time point.
A controlled prospective study in children with OSA and posterior crossbite, with 39 untreated controls, shows that this is feasible in orthodontic research [76]. The fact that it is rare is a choice, not a technical necessity.
11. Harms, Tolerability, Patient Experience and the Family Burden of Expansion
11.1 Pain is the dominant early adverse effect
Expansion hurts. A prospective observational study of pain during RME found that pain was greatest in the first days of treatment, that the Hyrax expander was associated with significantly greater initial pain than the Haas appliance (P < 0.001), that tooth-borne anchorage on first permanent molars produced greater initial pain than anchorage on second primary molars (P < 0.001), that children with unilateral posterior crossbite and anterior crossbite reported higher average pain (P < 0.001), and that age modulated pain (P < 0.001) with no sex difference (P = 0.287) [65].
The magnitude is clinically meaningful rather than trivial. A randomised comparison of miniscrew-supported and conventional expansion in adolescents found median scores in the bottom quartiles of the numerical rating scale for all reported outcomes, but the conventional RME group reported significantly higher scores for all measured variables except headache and dizziness [64]. A study of MARPE with and without micro-osteoperforation recorded moderate pain of 5.11 ± 0.30 in the minimally invasive group and 6.87 ± 0.40 in the standard group during the first two weeks, falling to 2.77 ± 0.39 and 5.11 ± 0.32 respectively in the following two weeks [99].
In a trial of 24 children, a difference of this size between appliances would matter to families even if it did not reach statistical significance in the between-group comparison. The Bansal abstract does not report pain, discomfort, analgesic use, sleep disturbance during activation, or school absence [1].
11.2 Dental and periodontal harm
Root resorption is the best-quantified dental harm. A three-dimensional volumetric meta-analysis found that RME was associated with greater root length loss (weighted mean difference −0.14 mm, 95% CI −0.21 to −0.08, 95% prediction interval −0.20 to −0.08, I² = 95.6%), and that tooth-borne RME produced significantly greater root resorption volume (−20.7 mm³) and length (−0.40 mm) than baseline, with bone-borne devices associated with reduced resorption risk [63]. The heterogeneity is very high, which limits precision, but the direction is consistent across studies and the mechanism — heavy force transmitted through the periodontal ligament — is plausible.
Other harms reported in the literature include buccal tipping of molars (significant only with tooth-borne appliances in controlled comparisons [57]), mild gingivitis from plaque accumulation around appliance components [54], asymmetric expansion, relapse and fistula formation in cleft patients [66], and anchorage loss. In cleft patients, expansion also interacts with the alveolar bone graft: a survey of cleft teams found that secondary alveolar bone grafting was typically planned 5–10 months after expansion, and that protocols varied widely between centres [66].
11.3 The family burden
Expansion is a family-level intervention, not a child-level one. It requires daily activation by a parent, regular visits, dietary modification during discomfort, and management of a fixed appliance in a child who may already be negotiating the social consequences of a facial difference. Adherence to activation schedules is rarely measured and almost never reported in trials. The Bansal abstract does not report adherence, appliance breakage, or the number of unscheduled visits [1].
This matters for the risk-benefit calculation in a specific way. If expansion were a definitive treatment for OSA, a year of discomfort and family logistics would be a reasonable price. If expansion is an orthodontic treatment with a plausible but unproven sleep benefit, the price is being paid for the orthodontic indication, and the sleep benefit should not be used to justify it.
11.4 The comparator’s harms
A balanced account must acknowledge that the alternative is not harmless. Adenotonsillectomy carries anaesthetic risk, postoperative pain, bleeding risk, dehydration, and a period of recovery; guidelines recommend inpatient monitoring for high-risk children postoperatively [9]. It is also associated with greater gains in height and weight percentile at 12 months [29], and one-third of children in a PATS analysis gained an undesirable amount of weight over the study period, although surgery was not independently associated with that gain [30].
Continuous positive airway pressure — the fallback when surgery is not indicated or fails — has its own burden. In a contemporary series of 491 children with paediatric OSA, positive airway pressure was prescribed for 240 (49.1%), of whom 176 (73%) had enough usage to generate a machine-estimated AHI [20]. Paediatric PAP adherence is a well-recognised limiting factor in real-world care, and a treatment that requires nightly tolerance in a child is not a low-burden alternative.
The honest framing is that all three options — surgery, PAP, expansion — impose burdens, and the choice should be driven by the indication, not by which burden is most convenient to measure.
12. Guidelines and Current Practice: Where Expansion Sits
12.1 The first-line position
The American Academy of Sleep Medicine practice parameters for childhood OSA, published in 2012, recommend adenotonsillectomy as the first-line treatment of patients with adenotonsillar hypertrophy, recommend postoperative reevaluation to determine whether further treatment is required, recommend objective testing in high-risk patients or those with persistent symptoms after therapy, recommend continuous positive airway pressure if adenotonsillectomy is not performed or OSA persists, and recommend weight loss in overweight or obese children [9]. That position has not been superseded by a randomised trial showing another intervention to be superior for the general paediatric OSA population.
The 2023 update of the AASM quality measures for paediatric OSA focuses on diagnosis and management performance measures captured in the Sleep Clinical Data Registry, rather than on endorsing specific second-line therapies [100]. It is a quality-improvement instrument, not a treatment recommendation, and it does not endorse expansion for OSA.
12.2 The cleft and craniofacial position
Contemporary cleft and craniofacial team guidance emphasises routine screening for obstructive sleep-disordered breathing, appropriate interpretation of sleep testing, and awareness that standard management may need to shift in children with craniofacial conditions, including 22q11.2 deletion syndrome and Pierre Robin sequence; it also discusses how velopharyngeal insufficiency surgery can affect OSA [39]. A craniofacial consensus review notes that children with OSA refractory to adenotonsillectomy may benefit from positive airway pressure, medical therapy, orthodontics, craniofacial surgery or combined interventions, and that for children with impairment of facial skeletal growth or craniofacial anomalies, rapid maxillary expansion, midface distraction and mandibular distraction “have all been demonstrated to have therapeutic value” and may significantly improve respiratory status [101].
That last statement is the strongest guideline-adjacent endorsement of expansion in this literature, and it is carefully bounded: it applies to children with impaired facial skeletal growth or craniofacial anomalies, not to all children with OSA. It is a statement about a selected population with a mechanical problem, which is exactly the population in which the anatomical rationale is strongest.
12.3 The explicit non-recommendation
Against that, two systematic reviews state the opposite position in unambiguous terms. The 2022 review of controlled evidence concluded that “in the absence of solid evidence with RCT, RME should not be recommended for the treatment of pediatric OSA” [7]. The 2026 update, after identifying three additional randomised trials, concluded that “extreme caution is warranted before recommending RME as a treatment for pediatric OSA” [11]. A 2023 umbrella review concluded that “RME is not recommended for treating OSA in children” [92].
| Source | Position on expansion for paediatric OSA | Population |
|---|---|---|
| AASM practice parameters, 2012 [9] | Adenotonsillectomy first-line for adenotonsillar hypertrophy; no endorsement of expansion | General paediatric OSA |
| AASM quality measures, 2023 [100] | Quality measures for diagnosis and management; no treatment endorsement | General paediatric OSA |
| Craniofacial consensus review [101] | RME “demonstrated to have therapeutic value” in children with impaired facial skeletal growth or craniofacial anomalies | Selected craniofacial population |
| Cleft/craniofacial oSDB primer, 2026 [39] | Routine SDB screening in cleft teams; management may shift in craniofacial conditions | Cleft and craniofacial |
| Systematic review of controlled evidence, 2022 [7] | “Should not be recommended” for paediatric OSA | General paediatric OSA |
| Systematic review update, 2026 [11] | “Extreme caution is warranted” | General paediatric OSA |
| Umbrella review, 2023 [92] | “RME is not recommended for treating OSA in children” | General paediatric OSA |
The apparent contradiction resolves once the population is specified. In a child with a craniofacially constrained airway and true transverse deficiency, expansion has a mechanical rationale and some supporting data. In an unselected child with OSA and adenotonsillar hypertrophy, it does not. The error in the literature is not that expansion is studied; it is that results from the first population are generalised to the second.
12.4 What cleft teams actually do
Practice varies more than guidelines imply. A survey of cleft teams on RME found a 40.5% response rate, a typical age for expansion of 9–10 years, secondary alveolar bone grafting planned 5–10 months after expansion, and a lack of consensus on protocols and materials, with asymmetric expansion, relapse and fistula formation the most common complications [66]. A survey of 231 cleft surgeons (21% response rate) found that with increasing years of practice, surgeons were less likely to refer patients for preoperative and postoperative sleep studies (P = .00 and P = .001), screen patients for sleep apnoea (P = .008), or change management on the basis of a sleep study (P = .001) [102].
For a clinician, this means that the answer to “what does the team recommend?” depends heavily on which team is asked. That is a signal that the evidence base is not constraining practice — which is precisely the situation in which a well-designed trial would be most valuable.
12.5 What is currently being studied
The registered trial landscape confirms that the field knows what question it wants to answer and has not yet designed the study to answer it. A search of ClinicalTrials.gov on 2026-11-04 returned 185 studies for “maxillary expansion,” 117 for “rapid maxillary expansion,” 98 for “rapid palatal expansion,” 15 for “maxillary expansion AND sleep apnea,” and 6 for “palatal expansion obstructive sleep apnea child.”
The studies most relevant to the question in this review illustrate both the activity and the design problem.
- A recruiting study of 102 children with obstructive sleep apnoea-hypopnoea syndrome and transverse orthodontic deficiency is testing whether children with transverse deficiency have a higher risk of OSAHS than those without — a case-control question, not a treatment question [103].
- A recruiting study of 40 participants is measuring MARPE-associated changes in sleep and airway, with the primary outcome being change in OSA risk on the teen-STOP-BANG questionnaire — a screening-instrument endpoint rather than a polysomnographic one [104].
- A not-yet-recruiting study of 24 participants will examine miniscrew-assisted rapid palatal expansion for OSA [105].
- A recruiting study of 40 participants is comparing a new clear expander with Hyrax, with CBCT evaluation as the primary outcome — a device comparison of exactly the kind the Bansal trial performed [106].
- A completed study of 51 children compared polysomnography and rhinomanometry changes after rapid maxillary expansion or adenotonsillectomy, which is the right comparison even if the sample is small [107].
- Two completed studies are directly relevant to the residual-disease question: a 14-participant study of maxillary expansion versus adenotonsillectomy with AHI as the primary outcome [108], and a 16-participant study of rapid maxillary expansion for residual paediatric OSA against standard clinical practice [109].
The pattern is a field conducting many small device comparisons and few adequately powered, controlled tests of whether expansion changes apnoea. The registration of CTRI/2024/04/066366, the trial under review here, follows the same pattern [94].
13. Radiation, Cost and Access
13.1 Radiation
Children with cleft lip and palate receive substantially more diagnostic radiation than their peers, and the excess is largely from computed tomography. A cohort study estimated cumulative dose to cleft palate patients at 1,963 µSv by age 20, compared with 597 µSv in non-cleft patients who received orthodontic treatment and 383 µSv in non-cleft patients who did not, with a three- to five-fold higher lifetime attributable risk of cancer, primarily because of the higher frequency of CT scanning [110].
The implication for research and practice is not that CBCT should be avoided — it has real diagnostic value in cleft care — but that every additional scan must be justified. Dose can be reduced substantially by protocol optimisation: in a study of CBCT for unilateral alveolar cleft assessment, a field of view of 50 × 40 mm reduced dose by 41% compared with the original field of view while still depicting the cleft area, central incisors, nasal floor and both canines, and a 35 × 40 mm field reduced dose by 56% [111]. A phantom study found that shielded low-dose CBCT protocols can achieve lower effective doses than conventional panoramic and lateral cephalometric radiographs combined, while standard CBCT protocols remain higher [112].
The critical question for an airway trial is whether the airway volume measurement requires a scan that would not otherwise be performed. If the child is already having a CBCT for orthodontic or surgical planning, adding a segmentation is nearly free in dose terms. If the scan is being done for research, the dose must be justified against a benefit that has not yet been demonstrated. In a population with a three- to five-fold excess lifetime attributable risk, that is a serious ethical consideration, not a footnote.
13.2 Cost and access
The economics of paediatric sleep medicine are dominated by the cost and scarcity of polysomnography. A cost analysis found that the total cost of adenotonsillectomy was 17.7 times the cost of polysomnography testing at average institutional costs, and that using symptom counts to diagnose OSA instead of polysomnography yielded low specificity [113]. That comparison cuts both ways: polysomnography is cheap relative to surgery, which argues for testing before operating; but it is expensive relative to a clinic visit, which is why access is limited.
The downstream economic case for treating sleep-disordered breathing is not trivial. In the PATS trial, adenotonsillectomy was associated with a 32% reduction in total health care encounters (mean difference −1.25 per participant per year, 95% CI −1.96 to −0.53) and a 48% reduction in prescriptions (mean difference −2.53 per participant per year, 95% CI −4.12 to −0.94), driven mainly by fewer office visits and outpatient procedures [28]. No comparable data exist for expansion.
Access to cleft team orthodontics varies enormously by health system. In systems where cleft care is centrally funded and multidisciplinary, expansion is available and integrated. In systems where it is not, the cost falls on families, and the decision to expand may be driven by what is reimbursable rather than by what is indicated. An appliance-based treatment that requires a year of activation and multiple visits is not equitably distributable, and a review that recommends it for a sleep indication without evidence of benefit is recommending an intervention whose access profile is already unequal.
13.3 The alternative-therapy access problem
It is worth stating the counterfactual explicitly. If expansion is not the answer, what is available to the child in whom adenotonsillectomy has failed or is not indicated? Positive airway pressure works but requires adherence: in one series, PAP was prescribed for 49.1% of children and only 73% of those had sufficient usage to generate a machine-estimated AHI [20]. Mandibular advancement appliances have randomised evidence of modest benefit in children — a network meta-analysis estimated an AHI reduction of −2.18/hour (95% CI −3.48 to −0.89) [6] — but require dental development and compliance, and an umbrella review of 12 systematic reviews found them promising as adjunctive therapy while judging the methodological quality of the included reviews predominantly low or critically low [114].
Expansion occupies a genuine gap: a child with a narrow maxilla, nasal obstruction, and mild residual OSA after surgery, in whom neither PAP nor mandibular advancement is attractive. The argument of this review is not that such a child should go untreated. It is that the treatment should be given for the reason that is supported — transverse deficiency and nasal obstruction — and that the sleep outcome should be measured rather than assumed.
14. The Trial That Would Answer the Question, and What to Tell a Family Today
14.1 The design
The trial that would settle whether maxillary expansion reduces OSA severity in children with transverse maxillary deficiency needs six features.
An untreated or delayed-treatment control. Randomisation to immediate expansion versus expansion after a defined delay (for example, 12 months) preserves the option of treatment for every child while creating a concurrent control. This is ethically preferable to a no-treatment arm in a population where expansion is often indicated on orthodontic grounds, and it solves the natural-history problem directly.
Adenotonsillar status stratified at randomisation. Children with large tonsils and children with small tonsils have different disease and different responses. The existing literature suggests that AHI improvement after RME is greater in children with previous adenotonsillectomy or small tonsils (73–95% reduction) than in children with large tonsils (61% reduction) [33] — a stratification that has never been prospectively applied.
AHI as the primary outcome, with a pre-specified minimal important difference. The primary analysis should be a between-group difference in change in AHI at a fixed time point, with the analysis population and the handling of missing data pre-specified. Given the night-to-night variability of the AHI (ICC 0.27) [22], two baseline nights and two follow-up nights would be preferable to one of each.
Blinded central scoring of PSG and CBCT. Scoring should be performed by individuals unaware of allocation and time point, with a pre-specified hypopnoea definition and a reported inter-scorer reliability coefficient.
A growth covariate and a reported maturation stage. At minimum, age, sex, skeletal maturation stage and baseline airway volume should be pre-specified covariates, with a sensitivity analysis adjusting for them.
Pre-registration with a published protocol and a reported sample size calculation. The trial should be registered before enrolment, with the primary outcome and analysis plan fixed, and the final report should reconcile any deviation from the registry entry. The observed association between trial registration and reduced spin in orthodontic abstracts (OR 0.336, 95% CI 0.117–0.962) suggests that registration has a measurable effect on reporting quality [115].
14.2 Sample size
The sample size question is not incidental. As an illustrative arithmetic exercise rather than a cited finding: to detect a between-group difference of 2 events/hour in AHI change with a standard deviation of 4 events/hour, at 80% power and a two-sided alpha of 0.05, would require approximately 64 children per arm — roughly five times the size of the Bansal trial, and that is before allowing for dropout or for the inflation needed to handle the night-to-night variability of the AHI. If the true effect is smaller than 2 events/hour, the required sample grows accordingly. No expansion trial to date has been sized for this question.
14.3 What a family should be told today
A clinician speaking to a parent of a child with a repaired cleft palate, a narrow maxilla and noisy breathing can say the following with confidence.
Expansion is a real treatment for the narrow maxilla. It widens the upper jaw, improves arch form and crossbite, and usually enlarges the nasal cavity. These are orthodontic and nasal-airway benefits, and they are worth having on their own terms [3] [51].
It has not been shown to cure or treat sleep apnoea. The best-controlled evidence available — a randomised comparison against watchful waiting, a network meta-analysis with a control arm, and a 2026 systematic review of all controlled prospective studies — does not show a reliable reduction in apnoea counts attributable to expansion [7] [6] [11].
A sleep study before and after is reasonable if the child has symptoms, but the result should be interpreted cautiously. Night-to-night variability is substantial, and a single improved study does not prove a treatment effect [22].
If the child has large tonsils and adenoids, the first conversation is about adenotonsillectomy. That is the first-line treatment, supported by a 464-child randomised trial showing improved polysomnography, behaviour and quality of life [8] [9], with the caveat that attention and executive function did not improve and that a substantial minority have residual disease.
The honest framing of the evidence is “this may help his breathing, and it will definitely help his bite.” That is a different sentence from “this will fix his sleep apnoea,” and it is the sentence the evidence supports.
14.4 What would constitute real progress
Progress in this field would look like a shift in the primary question. Instead of asking “does expansion change airway volume?”, the field should ask “in which children, defined before treatment, does expansion change apnoea, symptoms or blood pressure by an amount that matters?” That question requires phenotyping — adenotonsillar status, nasal resistance, skeletal maturity, craniofacial pattern — and it requires outcomes that patients and families care about.
A second form of progress would be the routine reporting of harms and patient-reported outcomes alongside efficacy. A trial that reports an AHI change but not pain, sleep disruption, school absence and family burden is reporting half the equation.
A third would be the publication of null results. The literature is small enough that unpublished null trials would materially change the pooled estimates, and the observed prevalence of spin in orthodontic and sleep medicine abstracts — 62.2% and 78.1% respectively, with “focusing on significant within-group comparison for primary outcomes” the most common strategy in orthodontics [115] [116] — suggests that the published record is systematically more positive than the underlying data.
15. Bottom Line for Clinicians
-
The Bansal trial is a comparison of two expansion systems, not a trial of expansion. Both arms were treated, in 24 children, at one centre, without blinding reported and without numerical effect sizes in the abstract [1]. It cannot show that expansion improves sleep, and it cannot establish equivalence of the appliances.
-
The abstract does not report AHI values, airway volumes, per-arm sample sizes, p-values or loss to follow-up. Any review or commentary that supplies those numbers is supplying them from somewhere other than the abstract, and the numbers should be checked against the full text before being repeated.
-
Expansion reliably enlarges the nasal cavity and inconsistently enlarges the nasopharynx and oropharynx. The pooled increase in total airway volume is about 1.2 cm³, with very low quality of evidence and no untreated controls in most studies [3] [4].
-
Static airway volume is a weak surrogate for OSA. Nasopharyngeal volume correlates with AHI at roughly rho = −0.4 [15]; airway volume change does not correlate with AHI change even after adenotonsillectomy [16]; and successive CBCT measurements carry errors large enough that one research group concluded CT should not be used quantitatively for airway change [72].
-
When controlled, the apnoea effect disappears. RME alone versus control: 0.02 events/hour (95% CI −1.72 to 1.75) [6]. The only randomised comparison with watchful waiting: no significant difference [7]. The 2026 update: no new trial with a watchful-waiting arm, and “extreme caution is warranted” [11].
-
The uncontrolled literature is not wrong, it is uncontrolled. A 70% mean AHI reduction in pre-post series [33] sits alongside a 42% spontaneous polysomnographic remission rate in untreated children over seven months [2]. Both numbers are real; only one of them is a treatment effect.
-
Adenotonsillectomy remains first-line for the child with adenotonsillar hypertrophy and OSA [9], with improved polysomnography, behaviour and quality of life but no demonstrated benefit on attention or executive function [8]. The one randomised head-to-head comparison found RME and adenotonsillectomy produced similar PSG outcomes in 24 children, but quality-of-life scores were significantly worse with RME [86].
-
In cleft populations, sleep-disordered breathing is common (pooled 16%, 32% in syndromic cleft) and under-tested. Screening is questionnaire-based and a substantial fraction of screen-positive children never get a sleep study [10] [35]. Improving ascertainment is a more defensible clinical priority than expanding more palates.
-
Expansion remains indicated for transverse maxillary deficiency on orthodontic and dentofacial grounds, and it may improve nasal breathing. That is a sufficient indication. It does not need a sleep justification, and it should not be given one that the evidence does not support.
-
The trial that would change practice has not been done. It requires an untreated or delayed-treatment control, stratified adenotonsillar status, blinded scoring, AHI as a pre-specified primary outcome with a declared minimal important difference, and roughly an order of magnitude more participants than any expansion trial to date.
Limitations
Of the seed trial. The review’s central object is a 24-patient, single-centre, two-active-arm randomised trial with no untreated control. Both arms received expansion, so within-group change cannot be attributed to the appliance. The abstract reports no numerical effect sizes, no AHI values, no airway volumes, no per-arm sample sizes, no p-values and no loss to follow-up; this review has confined every statement about the trial to those abstract-level facts and has explicitly marked the gaps rather than filling them. No attempt was made to obtain the full text, and no full-text-derived numbers appear anywhere in this review.
Of the evidence base. The expansion-airway-OSA literature is dominated by small, unblinded, single-centre studies using heterogeneous airway outcomes, with few untreated controls. The pooled estimates carry heterogeneity of I² = 93–97% for AHI outcomes [78], and the quality of evidence has been rated very low in the most careful meta-analysis [3]. The literature is also subject to substantial spin: 62.2% of orthodontic randomised trial abstracts contain spin, most commonly “focusing on significant within-group comparison for primary outcomes” [115], and 78.1% of sleep medicine randomised trial abstracts contain at least one spin strategy [116]. A review that reads only abstracts — as most clinicians must — will systematically overestimate treatment effects.
Of the measurement literature. CBCT airway volumetry has moderate-to-excellent reliability within a locked protocol but has not been adequately established across protocols, software and operators [68] [70] [71], and successive measurement error in the same patient can be large enough to exceed the treatment effect [72]. Polysomnography, the reference standard, has poor night-to-night reliability in children in exactly the severity range where treatment decisions are made [22].
Of this review’s method. Searches were run in PubMed via NCBI E-utilities, ClinicalTrials.gov via the v2 API, and Europe PMC and Crossref were available but not systematically used. The Clinical Trials Registry – India record for CTRI/2024/04/066366 could not be retrieved: the host resolved but connections were refused, so the registration details in this review are those reported in the article and not independently verified against the registry. No formal risk-of-bias assessment, GRADE assessment, or meta-analysis was performed by this review; where such assessments are cited, they are the assessments of the cited authors. The review is restricted to English-language publications and to records with retrievable abstracts. Grey literature, conference abstracts and unpublished trials were not searched, which is a particular limitation in a field where null results are likely to be under-published. Numbers attributed to abstracts are abstract-level only, and several are reported without the confidence intervals that the full texts may contain.
Of scope. This review addresses maxillary expansion and sleep-disordered breathing in children, with an emphasis on cleft lip and palate. It does not systematically address mandibular advancement, midface distraction, tongue reduction, or the surgical management of Pierre Robin sequence, each of which has its own literature and its own evidentiary problems. Where those interventions appear, they appear only as comparators or context.
Of inference. Several statements in this review are inferences rather than direct findings, and are marked as such in the text. The most important is the interpretation that the divergence between uncontrolled and controlled estimates reflects confounding by natural history rather than a methodological artefact of network meta-analysis. That interpretation is supported by the observed spontaneous remission rates in CHAT and PATS [2] [31], by the null randomised comparison against watchful waiting [7], and by the absence of any new controlled trial in the 2026 update [11] — but it remains an inference from convergent evidence rather than the result of a single definitive experiment. The sample-size arithmetic in Section 14.2 is an illustrative calculation, not a cited finding, and depends on assumptions about effect size and variance that no existing trial has established.
References
- Bansal R, Batra P, Mohelay N, Talwar A, Singh AK. Effects of clear aligner-based expansion screw and nickel-titanium palatal expander on airway dimensions and sleep parameters in unilateral cleft lip and palate patients: a randomized clinical trial. Sleep Breath. 2026;30(5). doi:10.1007/s11325-026-03806-9. PMID: 42799961.
- Chervin RD, Ellenberg SS, Hou X, Marcus CL, Garetz SL, Katz ES, et al. Prognosis for Spontaneous Resolution of OSA in Children. Chest. 2015;148(5):1204-1213. doi:10.1378/chest.14-2873. PMID: 25811889.
- Buck LM, Dalci O, Darendeliler MA, Papageorgiou SN, Papadopoulou AK. Volumetric upper airway changes after rapid maxillary expansion: a systematic review and meta-analysis. Eur J Orthod. 2017;39(5):463-473. doi:10.1093/ejo/cjw048. PMID: 27440774.
- Balasubramanian S, Kalaskar R, Kalaskar A. Rapid Maxillary Expansion and Upper Airway Volume: Systematic Review and Meta-analysis on the Role of Rapid Maxillary Expansion in Mouth Breathing. Int J Clin Pediatr Dent. 2022;15(5):617-630. doi:10.5005/jp-journals-10005-2421. PMID: 36865716.
- Bariani RCB, Bigliazzi R, Badreddine FR, Yamamoto LH, Tufik S, Moreira G, et al. A clinical trial on 3D CT scan and polysomnographyc changes after rapid maxillary expansion in children with snoring. Braz J Otorhinolaryngol. 2022;88 Suppl 5(Suppl 5):S162-S170. doi:10.1016/j.bjorl.2022.04.004. PMID: 35780010.
- Yu M, Ma Y, Xu Y, Bai J, Lu Y, Han F, et al. Orthodontic appliances for the treatment of pediatric obstructive sleep apnea: A systematic review and network meta-analysis. Sleep Med Rev. 2023;72:101855. doi:10.1016/j.smrv.2023.101855. PMID: 37820534.
- Fernández-Barriales M, Lafuente-Ibáñez de Mendoza I, Alonso-Fernández Pacheco JJ, Aguirre-Urizar JM. Rapid maxillary expansion versus watchful waiting in pediatric OSA: A systematic review. Sleep Med Rev. 2022;62:101609. doi:10.1016/j.smrv.2022.101609. PMID: 35286895.
- Marcus CL, Moore RH, Rosen CL, Giordani B, Garetz SL, Taylor HG, et al. A randomized trial of adenotonsillectomy for childhood sleep apnea. N Engl J Med. 2013;368(25):2366-76. doi:10.1056/NEJMoa1215881. PMID: 23692173.
- Marcus CL, Brooks LJ, Draper KA, Gozal D, Halbower AC, Jones J, et al. Diagnosis and management of childhood obstructive sleep apnea syndrome. Pediatrics. 2012;130(3):576-84. doi:10.1542/peds.2012-1671. PMID: 22926173.
- Elagib T, Fan Y, Cornelis MA, Cattaneo PM. Prevalence of Sleep-Disordered Breathing in Cleft-Affected Children: A Systematic Review and Meta-Analysis. Cleft Palate Craniofac J. 2026:10556656261465477. doi:10.1177/10556656261465477. PMID: 42429555.
- Fernández-Barriales M, López de Luzuriaga M, Lafuente-Ibáñez de Mendoza I, Alonso Fernández-Pacheco JJ, Álvarez Ruiz de Larrinaga A, Aguirre Urizar JM. Controlled Prospective Evidence of Rapid Maxillary Expansion Efficacy in Pediatric Obstructive Sleep Apnea: A Systematic Review Update. J Clin Med. 2026;15(8). doi:10.3390/jcm15082976. PMID: 42074778.
- Mordente CM, Palomo JM, Horta MC, Souki BQ, Oliveira DD, Andrade I Jr. Upper airway assessment using four different maxillary expanders in cleft patients: A cone-beam computed tomography study. Angle Orthod. 2016;86(4):617-24. doi:10.2319/032015-174.1. PMID: 26595658.
- Haritha SH, Krishna KR, Yelchuru SH, Kumar DK, Saveri K, Sonika B. Assessment of Dentoskeletal and Airway Changes Using Hyrax and Fan-Type Expanders in Cleft Patients: An In Vivo CBCT Study. Cleft Palate Craniofac J. 2026:10556656261446603. doi:10.1177/10556656261446603. PMID: 42377098.
- Demirsoy KK, Buyuk SK, Alev Y, Bicer T, Abay F. Comparative Evaluation of Skeletal, Dentoalveolar, and Airway Effects of Bone-Borne and Hybrid Expanders in Young Adults. Orthod Craniofac Res. 2026;29(2):375-384. doi:10.1111/ocr.70091. PMID: 41472450.
- Masoud AI, Alwadei AH, Gowharji LF, Park CG, Carley DW. Relating three-dimensional airway measurements to the apnea-hypopnea index in pediatric sleep apnea patients. Orthod Craniofac Res. 2021;24(1):137-146. doi:10.1111/ocr.12417. PMID: 32757439.
- Pinheiro de Magalhães Bertoz A, Souki BQ, Lione R, Theresa Webber SA, Bigliazzi R, Oliveira PM, et al. Three-dimensional airway changes after adenotonsillectomy in children with obstructive apnea: Do expectations meet reality?. Am J Orthod Dentofacial Orthop. 2019;155(6):791-800. doi:10.1016/j.ajodo.2018.06.019. PMID: 31153499.
- Taylor HG, Bowen SR, Beebe DW, Hodges E, Amin R, Arens R, et al. Cognitive Effects of Adenotonsillectomy for Obstructive Sleep Apnea. Pediatrics. 2016;138(2). doi:10.1542/peds.2015-4458. PMID: 27464674.
- Lumeng JC, Chervin RD. Epidemiology of pediatric obstructive sleep apnea. Proc Am Thorac Soc. 2008;5(2):242-52. doi:10.1513/pats.200708-135MG. PMID: 18250218.
- Magnusdottir S, Hill EA. Prevalence of obstructive sleep apnea (OSA) among preschool aged children in the general population: A systematic review. Sleep Med Rev. 2024;73:101871. doi:10.1016/j.smrv.2023.101871. PMID: 37976758.
- Hassinger AB, Granger H, Ibrahim A, Li S, Varol M, Danis A. Beyond Snoring, Sleepiness and Surgery: Update on the Prevalent Symptoms of and Treatment Effects in all Age Groups With Pediatric Obstructive Sleep Apnea. Pediatr Pulmonol. 2026;61(10):e71865. doi:10.1002/ppul.71865. PMID: 42829879.
- Couloigner V, Luscan R, Fauroux B. Limitations of the Apnea-Hypopnea Index in pediatric obstructive sleep apnea. A critical review: Toward a multiparametric approach. Eur Ann Otorhinolaryngol Head Neck Dis. 2026;143(5):403-408. doi:10.1016/j.anorl.2026.08.005. PMID: 42686450.
- Xiao L, Terrance M, Au CT, Massicotte C, Narang I. High night-to-night variability in childhood obstructive sleep apnea severity. Sleep Med. 2026;140:108784. doi:10.1016/j.sleep.2026.108784. PMID: 41548262.
- Riklin E, Choi DG, Trotter C, Lasky S, Kato RM, Davidson Ward SL, et al. Polysomnogram Findings and Psychosocial and Academic Concerns in Children With Cleft Palate With or Without Cleft Lip. J Craniofac Surg. 2025;36(3):820-825. doi:10.1097/SCS.0000000000010504. PMID: 39141811.
- Thomas NH, Xanthopoulos MS, Kim JY, Shults J, Escobar E, Giordani B, et al. Effects of Adenotonsillectomy on Parent-Reported Behavior in Children With Obstructive Sleep Apnea. Sleep. 2017;40(4). doi:10.1093/sleep/zsx018. PMID: 28199697.
- Decuzzi J, Redline S, Isaiah A. Secondary Analyses of the Childhood Adenotonsillectomy Trial: A Narrative Review. JAMA Otolaryngol Head Neck Surg. 2022;148(8):779-784. doi:10.1001/jamaoto.2022.1330. PMID: 35737366.
- Cabrera AJ, Li D, Redline S, Chervin RD, Ishman SL, Baldassari CM, et al. Adenotonsillectomy and Blood Pressure in Children With Mild Obstructive Sleep-Disordered Breathing: An Exploratory Analysis of the PATS Randomized Clinical Trial. JAMA Otolaryngol Head Neck Surg. 2025;151(10):957-966. doi:10.1001/jamaoto.2025.2555. PMID: 40875240.
- Au CT, Chan KC, Lee DLY, Leung NMW, Chow SMW, Kwok KL, et al. Effect of surgical intervention for childhood OSA on blood pressure: A randomized controlled study. Sleep Med. 2023;107:9-17. doi:10.1016/j.sleep.2023.04.006. PMID: 37094489.
- Bakker JP, Zhang F, Amin R, Baldassari CM, Chervin RD, Garetz SL, et al. Adenotonsillectomy and Health Care Utilization in Children With Snoring and Mild Sleep Apnea: A Randomized Clinical Trial. JAMA Pediatr. 2025;179(6):600-609. doi:10.1001/jamapediatrics.2025.0023. PMID: 40094698.
- Gong W, Huang K, Psychogios I, Li S, Chen L, Zhang J, et al. Growth and sleep outcomes after adenotonsillectomy in pediatric mild sleep-disordered breathing. Sci Rep. 2025;16(1):688. doi:10.1038/s41598-025-30271-3. PMID: 41350327.
- Kirkham EM, Ishman S, Baldassari CM, Mitchell RB, Naqvi SK, Tapia IE, et al. Weight Gain After Adenotonsillectomy in Children With Mild Obstructive Sleep-Disordered Breathing: Exploratory Analysis of the PATS Randomized Clinical Trial. JAMA Otolaryngol Head Neck Surg. 2024;150(10):859-867. doi:10.1001/jamaoto.2024.2554. PMID: 39172463.
- Kirkham EM, Ishman S, Garetz S, Baldassari CM, Mitchell RB, Liu C, et al. Progression of mild sleep-disordered breathing in children managed with watchful waiting. J Clin Sleep Med. 2026;22(1). doi:10.1007/s44470-026-00082-y. PMID: 42414790.
- Sjölander I, Fehrm J, Borgström A, Nerfeldt P, Friberg D. Adenotonsillectomy vs Watchful Waiting in Pediatric Mild to Moderate Obstructive Sleep Apnea: The KATE Randomized Clinical Trial. JAMA Otolaryngol Head Neck Surg. 2025;151(11):1038-1045. doi:10.1001/jamaoto.2025.2831. PMID: 40965870.
- Camacho M, Chang ET, Song SA, Abdullatif J, Zaghi S, Pirelli P, et al. Rapid maxillary expansion for pediatric obstructive sleep apnea: A systematic review and meta-analysis. Laryngoscope. 2017;127(7):1712-1719. doi:10.1002/lary.26352. PMID: 27796040.
- Gorucu-Coskuner H, Saglam-Aydinatay B, Aksu M, Ozgur FF, Taner T. Comparison of Positive Screening for Obstructive Sleep Apnea in Patients With and Without Cleft Lip and Palate. Cleft Palate Craniofac J. 2020;57(3):364-370. doi:10.1177/1055665619875321. PMID: 31529989.
- Fisher AH, Stanisce L, Nelson ZJ, Cohen MA, Matthews MS. Risk Assessment of Sleep Disordered Breathing in Cleft Lip and/or Palate. Cleft Palate Craniofac J. 2024;61(12):2074-2078. doi:10.1177/10556656231193552. PMID: 37545192.
- Budihardja AS, Nathania S. Association of obstructive sleep apnea in children with and without cleft lip palate. BMC Oral Health. 2025;25(1):883. doi:10.1186/s12903-025-06281-y. PMID: 40457294.
- Theprungsirikul W, Manosudprasit A, Ruangsri S, Laohasiriwong S, Chaichit R, Manosudprasit A. Prevalence and Risk Factors of Obstructive Sleep Apnea in Patients With Complete Cleft Lip and Palate in the Northeastern Thailand. Cleft Palate Craniofac J. 2026;63(5):1197-1206. doi:10.1177/10556656251327881. PMID: 40116723.
- Robison JG, Otteson TD. Increased prevalence of obstructive sleep apnea in patients with cleft palate. Arch Otolaryngol Head Neck Surg. 2011;137(3):269-74. doi:10.1001/archoto.2011.8. PMID: 21422312.
- Maksimoski M, Sitzman TJ, Tse RW, Baylis AL, Perry JL, Ishman SL. Management of Obstructive Sleep Apnea in Children With Cleft Palate and/or Velopharyngeal Insufficiency: A Primer on Screening, Testing, and Treatment. Cleft Palate Craniofac J. 2026;63(9):3082-3092. doi:10.1177/10556656251387799. PMID: 41212739.
- Akarsu-Guven B, Karakaya J, Ozgur F, Aksu M. Upper airway features of unilateral cleft lip and palate patients in different growth stages. Angle Orthod. 2019;89(4):575-582. doi:10.2319/022518-155.1. PMID: 30694706.
- Akarsu-Guven B, Karakaya J, Ozgur F, Aksu M. Growth-related changes of skeletal and upper-airway features in bilateral cleft lip and palate patients. Am J Orthod Dentofacial Orthop. 2015;148(4):576-86. doi:10.1016/j.ajodo.2015.02.031. PMID: 26432313.
- Laowong O, Sonsuwan N, Chaiworawitkul M. Investigating anatomical determinants of OSA in children with UCLP: insights into tongue, airway, and hyoid bone. Sleep Breath. 2026;30(1). doi:10.1007/s11325-026-03639-6. PMID: 41811608.
- Posiri C, Sonsuwan N, Chaiworawitkul M. Three-Dimensional Palatal Morphology and Obstructive Sleep Apnea Severity in Children with Unilateral Cleft Lip and Palate: A CBCT Study. Children (Basel). 2026;13(7). doi:10.3390/children13070909. PMID: 42509934.
- Campos LD, Trindade IEK, Yatabe M, Trindade SHK, Pimenta LA, Kimbell J, et al. Reduced pharyngeal dimensions and obstructive sleep apnea in adults with cleft lip/palate and Class III malocclusion. Cranio. 2021;39(6):484-490. doi:10.1080/08869634.2019.1668997. PMID: 31526316.
- Cielo CM, Taylor JA, Vossough A, Radcliffe J, Thomas A, Bradford R, et al. Evolution of Obstructive Sleep Apnea in Infants with Cleft Palate and Micrognathia. J Clin Sleep Med. 2016;12(7):979-87. doi:10.5664/jcsm.5930. PMID: 27092700.
- Bergeron M, Cohen AP, Maby A, Babiker HE, Pan BS, Ishman SL. The Effect of Cleft Palate Repair on Polysomnography Results. J Clin Sleep Med. 2019;15(11):1581-1586. doi:10.5664/jcsm.8016. PMID: 31739847.
- Alassaf MS, Abdulhameed FD, Othman AA, Khaja H, Habeeb AT, Alolayan AB, et al. Screening for Obstructive Sleep Apnoea in Patients With Cleft Lip and/or Palate in Saudi Arabia. Cleft Palate Craniofac J. 2026;63(11):3729-3736. doi:10.1177/10556656261425656. PMID: 41757857.
- Yoon A, Abdelwahab M, Bockow R, Vakili A, Lovell K, Chang I, et al. Impact of rapid palatal expansion on the size of adenoids and tonsils in children. Sleep Med. 2022;92:96-102. doi:10.1016/j.sleep.2022.02.011. PMID: 35390750.
- Ierardo G, Nicita F, Vozza I, Polimeni A, Luzzi V. Relationship Between Maxillary Transverse Deficiency and Respiratory Problems: A Systematic Review of the Effectiveness of Devices over the Past Decade. J Clin Med. 2025;14(24). doi:10.3390/jcm14248861. PMID: 41464765.
- Alforaidi SA. Effects of Rapid Maxillary Expansion on Upper Airway Parameters in Children With Unilateral Cleft Lip and Palate: A Cone-Beam Computed Tomography (CBCT) Study. Cleft Palate Craniofac J. 2026;63(11):3638-3645. doi:10.1177/10556656261420487. PMID: 41649946.
- Calvo-Henriquez C, Megias-Barrera J, Chiesa-Estomba C, Lechien JR, Maldonado Alvarado B, Ibrahim B, et al. The Impact of Maxillary Expansion on Adults’ Nasal Breathing: A Systematic Review and Meta-Analysis. Am J Rhinol Allergy. 2021;35(6):923-934. doi:10.1177/1945892421995350. PMID: 33583193.
- Ottaviano G, Maculan P, Borghetto G, Favero V, Galletti B, Savietto E, et al. Nasal function before and after rapid maxillary expansion in children: A randomized, prospective, controlled study. Int J Pediatr Otorhinolaryngol. 2018;115:133-138. doi:10.1016/j.ijporl.2018.09.029. PMID: 30368373.
- Bazargani F, Magnuson A, Ludwig B. Effects on nasal airflow and resistance using two different RME appliances: a randomized controlled trial. Eur J Orthod. 2018;40(3):281-284. doi:10.1093/ejo/cjx081. PMID: 29069383.
- Gokce G, Gode S, Ozturk A, Kirazlı T, Veli I. Evaluation of the effects of different rapid maxillary expansion appliances on airway by acoustic rhinometry: A randomized clinical trial. Int J Pediatr Otorhinolaryngol. 2022;155:111074. doi:10.1016/j.ijporl.2022.111074. PMID: 35189449.
- Kabalan O, Gordon J, Heo G, Lagravère MO. Nasal airway changes in bone-borne and tooth-borne rapid maxillary expansion treatments. Int Orthod. 2015;13(1):1-15. doi:10.1016/j.ortho.2014.12.011. PMID: 25665908.
- Giap HV, Alomaym M, Chung CJ, Choi SH, Chun JH, Alkhamees A, et al. Skeletal displacements and upper airway changes following tooth-borne and tooth-bone-borne rapid palatal expansion in adolescents: secondary analysis of a randomized clinical trial. BMC Oral Health. 2026;26(1). doi:10.1186/s12903-026-09784-4. PMID: 42816853.
- Kavand G, Lagravère M, Kula K, Stewart K, Ghoneima A. Retrospective CBCT analysis of airway volume changes after bone-borne vs tooth-borne rapid maxillary expansion. Angle Orthod. 2019;89(4):566-574. doi:10.2319/070818-507.1. PMID: 30768911.
- Ronsivalle V, Leonardi R, Lagravere M, Flores-Mir C, Grippaudo C, Alessandri Bonetti G, et al. Medium-term effects of rapid maxillary expansion on nasal cavity and pharyngeal airway volumes considering age as a factor: A retrospective study. J Dent. 2024;144:104934. doi:10.1016/j.jdent.2024.104934. PMID: 38461886.
- Firinciogullari EC, Erdinc AE, Celenk-Koca T, Walma DC, Akyalcin S. MARPE Improves Long-Term Stability of Maxillary Expansion in Adolescents. Orthod Craniofac Res. 2025;28 Suppl 1:S194-S202. doi:10.1111/ocr.12922. PMID: 40130401.
- Alrumaih R, Alterki A, Qali M. Maxillary Expansion in the Management of Obstructive Sleep Apnea: A Comprehensive Review. Dent J (Basel). 2025;13(9). doi:10.3390/dj13090410. PMID: 41002683.
- Fonseca Planells L, Bellot-Arcís C, Montiel-Company JM, Paredes-Gallardo V, Zamora-Martínez N. Effectiveness of clear aligners for maxillary expansion in growing patients: a systematic review and meta-analysis. Prog Orthod. 2026;27(1). doi:10.1186/s40510-026-00620-7. PMID: 42060239.
- Nagpal DI, Dorlikar JR, Rathi P, Gurharikar A, Khondalay S, Lamba G. Clinical outcome of maxillary arch expansion with clear aligners in mixed dentition: A systematic review. J Indian Soc Pedod Prev Dent. 2026;44(1):24-35. doi:10.4103/jisppd.jisppd_421_25. PMID: 41755433.
- Xue C, Wu J, Liu L, Qiu Y, Li L. Three-Dimensional Volumetric Assessment of Root Resorption in Rapid Maxillary Expansion: A Meta-Analysis Comparing Tooth-Borne and Bone-Borne Devices. Orthod Craniofac Res. 2026;29(3):506-520. doi:10.1111/ocr.70096. PMID: 41546473.
- ElNaghy R, Al-Qawasmi R, Hasanin M. Do patient-reported outcomes of miniscrew-supported maxillary expansion in adolescent patients differ between slow and rapid activation protocol?. Evid Based Dent. 2023;24(1):28-29. doi:10.1038/s41432-023-00858-8. PMID: 36890253.
- De Stefani A, Boutarbouche A, Barone M, Gracco A, Bruno G. Pain Experience During Rapid Maxillary Expansion: A Prospective Observational Study. Children (Basel). 2026;13(3). doi:10.3390/children13030361. PMID: 41897074.
- Khdairi N, Halilah T, Khandakji M, Bartzela T. Rapid Maxillary Expansion Treatment in Patients with Cleft Lip and Palate: A Survey on Clinical Experience in the European Cleft Centers. J Clin Med. 2023;12(9). doi:10.3390/jcm12093159. PMID: 37176600.
- Aksoz G, El H, Palomo JM. Correlation between different boundaries used in upper airway assessment. BMC Oral Health. 2025;25(1):19. doi:10.1186/s12903-024-05402-3. PMID: 39755673.
- Zimmerman JN, Lee J, Pliska BT. Reliability of upper pharyngeal airway assessment using dental CBCT: a systematic review. Eur J Orthod. 2017;39(5):489-496. doi:10.1093/ejo/cjw079. PMID: 27999121.
- Chwarścianek N, Kazimierczak N, Serafin Z, Kazimierczak W. Reliability of AI-Automated and Semiautomated Upper Airway Volume Segmentation. Diagnostics (Basel). 2026;16(7). doi:10.3390/diagnostics16071105. PMID: 41975816.
- Weissheimer A, Menezes LM, Sameshima GT, Enciso R, Pham J, Grauer D. Imaging software accuracy for 3-dimensional analysis of the upper airway. Am J Orthod Dentofacial Orthop. 2012;142(6):801-13. doi:10.1016/j.ajodo.2012.07.015. PMID: 23195366.
- Naralan ME, Köse TE, Gonca M, Beşer Gül B, Günaçar DN. Accuracy of upper airway volume measurements using different software products: a comparative analysis. Dentomaxillofac Radiol. 2025;54(5):350-356. doi:10.1093/dmfr/twaf023. PMID: 40084997.
- Sun JS, Lin SY, Hsieh CY, Hung MC, Tai HC, Chang JZ. Successive measurement errors of consecutive computed tomography for airway-related craniofacial dimensional measurements. J Dent Sci. 2024;19(4):1961-1971. doi:10.1016/j.jds.2024.07.033. PMID: 39347062.
- Meng C, Li Y, Li W, Niu L, Lu W, Su H, et al. Longitudinal reliability of upper airway assessment using cone-beam computed tomography: a retrospective study. BMC Oral Health. 2026;26(1). doi:10.1186/s12903-026-07769-x. PMID: 41634698.
- Fregosi RF, Quan SF, Kaemingk KL, Morgan WJ, Goodwin JL, Cabrera R, et al. Sleep-disordered breathing, pharyngeal size and soft tissue anatomy in children. J Appl Physiol (1985). 2003;95(5):2030-8. doi:10.1152/japplphysiol.00293.2003. PMID: 12897029.
- Echarri-Nicolás J, González-Olmo MJ, Echarri-Labiondo P, Romero M. Short-term outcomes in the upper airway with tooth-bone-borne vs bone-borne rapid maxillary expanders. BMC Oral Health. 2023;23(1):714. doi:10.1186/s12903-023-03461-6. PMID: 37794400.
- Galeotti A, Gatto R, Caruso S, Piga S, Maldonato W, Sitzia E, et al. Effects of Rapid Palatal Expansion on the Upper Airway Space in Children with Obstructive Sleep Apnea (OSA): A Case-Control Study. Children (Basel). 2023;10(2). doi:10.3390/children10020244. PMID: 36832373.
- Machado-Júnior AJ, Zancanella E, Crespo AN. Rapid maxillary expansion and obstructive sleep apnea: A review and meta-analysis. Med Oral Patol Oral Cir Bucal. 2016;21(4):e465-9. doi:10.4317/medoral.21073. PMID: 27031063.
- Yu K, Li Y, Ngan P, Li L, Huang J, Hua F, et al. The effect of rapid maxillary expansion on children with obstructive sleep apnea: a systematic review and meta-analysis. Sleep Breath. 2026;30(1). doi:10.1007/s11325-026-03606-1. PMID: 41781690.
- Lin SY, Su YX, Wu YC, Chang JZ, Tu YK. Management of paediatric obstructive sleep apnoea: A systematic review and network meta-analysis. Int J Paediatr Dent. 2020;30(2):156-170. doi:10.1111/ipd.12593. PMID: 31680340.
- Villa MP, Malagola C, Pagani J, Montesano M, Rizzoli A, Guilleminault C, et al. Rapid maxillary expansion in children with obstructive sleep apnea syndrome: 12-month follow-up. Sleep Med. 2007;8(2):128-34. doi:10.1016/j.sleep.2006.06.009. PMID: 17239661.
- Villa MP, Rizzoli A, Miano S, Malagola C. Efficacy of rapid maxillary expansion in children with obstructive sleep apnea syndrome: 36 months of follow-up. Sleep Breath. 2011;15(2):179-84. doi:10.1007/s11325-011-0505-1. PMID: 21437777.
- Villa MP, Rizzoli A, Rabasco J, Vitelli O, Pietropaoli N, Cecili M, et al. Rapid maxillary expansion outcomes in treatment of obstructive sleep apnea in children. Sleep Med. 2015;16(6):709-16. doi:10.1016/j.sleep.2014.11.019. PMID: 25934539.
- Villa MP, Castaldo R, Miano S, Paolino MC, Vitelli O, Tabarrini A, et al. Adenotonsillectomy and orthodontic therapy in pediatric obstructive sleep apnea. Sleep Breath. 2014;18(3):533-9. doi:10.1007/s11325-013-0915-3. PMID: 24277354.
- Marino A, Ranieri R, Chiarotti F, Villa MP, Malagola C. Rapid maxillary expansion in children with Obstructive Sleep Apnoea Syndrome (OSAS). Eur J Paediatr Dent. 2012;13(1):57-63. PMID: 22455530.
- Miano S, Rizzoli A, Evangelisti M, Bruni O, Ferri R, Pagani J, et al. NREM sleep instability changes following rapid maxillary expansion in children with obstructive apnea sleep syndrome. Sleep Med. 2009;10(4):471-8. doi:10.1016/j.sleep.2008.04.003. PMID: 18753006.
- Aksilp C, Pechpongsai P, Intakorn P, Chaweewannakorn C, Boonpratham S, Satravaha Y, et al. A randomized controlled trial comparing treatment efficacy between rapid maxillary expansion and adenotonsillectomy in pediatric obstructive sleep apnea. Sleep Breath. 2025;29(4):256. doi:10.1007/s11325-025-03427-8. PMID: 40739069.
- Guilleminault C, Quo S, Huynh NT, Li K. Orthodontic expansion treatment and adenotonsillectomy in the treatment of obstructive sleep apnea in prepubertal children. Sleep. 2008;31(7):953-7. PMID: 18652090.
- Magalhães MC, Normando D, Soares CJ, Araujo E, Novaes RMO, Teodoro VV, et al. Impact of adenotonsillectomy and palatal expansion on the apnea-hypopnea index and minimum oxygen saturation in nonobese pediatric obstructive sleep apnea with balanced maxillomandibular relationship: A cross-over randomized controlled trial. Pediatr Pulmonol. 2024;59(12):3507-3517. doi:10.1002/ppul.27239. PMID: 39239911.
- Mahony D, Bhattarai N, Petocz P. Maxillary Arch Expansion and Adenotonsillectomy in Prepubertal Children Diagnosed with Paediatric Obstructive Sleep Apnoea: An Interventional Study. J Clin Med. 2026;15(8). doi:10.3390/jcm15082981. PMID: 42074783.
- Li Y, Wang P, Liu A, Zhang C, Zhao L, Yu L, et al. Dentofacial Orthopedic Treatment vs. Adenotonsillectomy in Children with Mild to Moderate OSA and Mandibular Retrognathia: A Randomized Controlled Trial. Dent J (Basel). 2026;14(8). doi:10.3390/dj14080511. PMID: 42645485.
- Vale F, Albergaria M, Carrilho E, Francisco I, Guimarães A, Caramelo F, et al. Efficacy of Rapid Maxillary Expansion in the Treatment of Obstructive Sleep Apnea Syndrome: A Systematic Review With Meta-analysis. J Evid Based Dent Pract. 2017;17(3):159-168. doi:10.1016/j.jebdp.2017.02.001. PMID: 28865812.
- Barbosa DF, Bana LF, Michel MCB, Meira E Cruz M, Zancanella E, Machado Júnior AJ. Rapid maxillary expansion in pediatric patients with obstructive sleep apnea: an umbrella review. Braz J Otorhinolaryngol. 2023;89(3):494-502. doi:10.1016/j.bjorl.2023.02.004. PMID: 36894478.
- Ahsan M, Narala B, Ednick M, Kier C. Something to consider: rapid palatal expansion for treatment of obstructive sleep apnea in pediatric patients. Curr Opin Pediatr. 2023;35(6):710-715. doi:10.1097/MOP.0000000000001287. PMID: 37678405.
- Clinical Trials Registry – India. CTRI/2024/04/066366: Effects of clear aligner-based expansion screw and nickel-titanium palatal expander on airway dimensions and sleep parameters in unilateral cleft lip and palate patients: a randomized clinical trial. Registered April 2024. Registry record not retrievable at preparation (host resolved; connection refused); registration number as reported in the trial publication (PMID 42799961). https://ctri.nic.in/Clinicaltrials/pmaindet2.php?trialid=066366
- Xu L, Lu X, Wang S, Zheng F, Chen X. Upper airway dimensions in adolescents with different craniofacial growth patterns: a retrospective cross-sectional comparison of independent samples from 2005 and 2020. BMC Oral Health. 2026;26(1). doi:10.1186/s12903-026-08055-6. PMID: 41826971.
- Zhu F, Zhao G, Sun S. A longitudinal cohort study of sleep-disordered breathing in children: interrelationships with craniofacial morphology, malocclusion, and intervention outcomes. BMC Oral Health. 2026;26(1). doi:10.1186/s12903-026-07791-z. PMID: 41735993.
- Duan B, Zhang H, Liang J, Chen W. Adenoid-Nasopharyngeal Ratio and craniofacial growth in Children: A two-center longitudinal study with threshold and causal analyses. Eur Arch Otorhinolaryngol. 2026;283(9):5791-5802. doi:10.1007/s00405-026-10387-6. PMID: 42380517.
- Lacki OA, Iftikhar IH. Meta-analysis of diagnostic test accuracy of pediatric sleep questionnaire and the OSA-18 questionnaire in the screening of pediatric obstructive sleep apnea. Sleep Med. 2026;144:108953. doi:10.1016/j.sleep.2026.108953. PMID: 41955725.
- Elshehaby M, Albelasy NF, Elbialy MA, Hafez AM, Abdelnaby YL. Evaluation of pain intensity and airway changes in non-growing patients treated by MARPE with and without micro-osteoperforation: a randomized clinical trial. BMC Oral Health. 2024;24(1):1411. doi:10.1186/s12903-024-05196-4. PMID: 39563319.
- Lloyd RM, Crawford T, Donald R, Gray DD, Healy WJ, Junna MR, et al. Quality measures for the care of pediatric patients with obstructive sleep apnea: 2023 update after measure maintenance. J Clin Sleep Med. 2024;20(1):127-134. doi:10.5664/jcsm.10836. PMID: 37772707.
- Garg RK, Afifi AM, Garland CB, Sanchez R, Mount DL. Pediatric Obstructive Sleep Apnea: Consensus, Controversy, and Craniofacial Considerations. Plast Reconstr Surg. 2017;140(5):987-997. doi:10.1097/PRS.0000000000003752. PMID: 29068938.
- Bennett KG, Robinson AB, Kasten SJ, Buchman SR, Vercler CJ. Velopharyngeal Dysfunction and Sleep Apnea-A Survey to Ascertain Surgical Practice Patterns. Cleft Palate Craniofac J. 2017;54(1):13-18. doi:10.1597/15-250. PMID: 26752134.
- ClinicalTrials.gov. NCT07086612: registered study record. National Library of Medicine. https://clinicaltrials.gov/study/NCT07086612
- ClinicalTrials.gov. NCT07031349: registered study record. National Library of Medicine. https://clinicaltrials.gov/study/NCT07031349
- ClinicalTrials.gov. NCT07734688: registered study record. National Library of Medicine. https://clinicaltrials.gov/study/NCT07734688
- ClinicalTrials.gov. NCT07362160: registered study record. National Library of Medicine. https://clinicaltrials.gov/study/NCT07362160
- ClinicalTrials.gov. NCT07069543: registered study record. National Library of Medicine. https://clinicaltrials.gov/study/NCT07069543
- ClinicalTrials.gov. NCT01837914: registered study record. National Library of Medicine. https://clinicaltrials.gov/study/NCT01837914
- ClinicalTrials.gov. NCT02947464: registered study record. National Library of Medicine. https://clinicaltrials.gov/study/NCT02947464
- Jacobs R, Pauwels R, Scarfe WC, De Cock C, Dula K, Willems G, et al. Pediatric cleft palate patients show a 3- to 5-fold increase in cumulative radiation exposure from dental radiology compared with an age- and gender-matched population: a retrospective cohort study. Clin Oral Investig. 2018;22(4):1783-1793. doi:10.1007/s00784-017-2274-0. PMID: 29188451.
- Regnstrand T, Top SK, Karsten A, Jacobs R, Benchimol D. Dose Reduction by Field of View Optimisation in Cone-Beam Computed Tomography of Unilateral Alveolar Clefts. Orthod Craniofac Res. 2025;28(5):820-825. doi:10.1111/ocr.12938. PMID: 40261632.
- Ting S, Attaia D, Johnson KB, Kossa SS, Friedland B, Allareddy V, et al. Can modifying shielding, field of view, and exposure settings make the effective dose of a cone-beam computed tomography comparable to traditional radiographs used for orthodontic diagnosis?. Angle Orthod. 2020;90(5):655-664. doi:10.2319/072819-496.1. PMID: 33378479.
- Mitchell M, Werkhaven JA. Cost-effectiveness of polysomnography in the management of pediatric obstructive sleep apnea. Int J Pediatr Otorhinolaryngol. 2020;133:109943. doi:10.1016/j.ijporl.2020.109943. PMID: 32086039.
- Li Y, Zhao T, Lin L, Zhang J, Liu S, Wu Y, et al. The effects of mandibular advancement devices on pediatric obstructive sleep apnea: An umbrella review. Sleep Med Rev. 2025;84:102166. doi:10.1016/j.smrv.2025.102166. PMID: 41005116.
- Guo F, Fang X, Li C, Qin D, Hua F, He H. The presence and characteristics of ‘spin’ among randomized controlled trial abstracts in orthodontics. Eur J Orthod. 2021;43(5):576-582. doi:10.1093/ejo/cjab044. PMID: 34397084.
- Guo F, Zhao T, Zhai Q, Fang X, Yue H, Hua F, et al. ‘Spin’ among abstracts of randomised controlled trials in sleep medicine: A research-on-research study. Sleep. 2023. doi:10.1093/sleep/zsad041. PMID: 36861330.