February 19, 2023

What are the dental and craniofacial characteristics of children with sleep apnoea (OSA)?

Obstructive Sleep Apnoea is a problem that may be related to the dentofacial characteristics of children. Therefore, it is important to identify these features as they may be potential treatment targets. This new systematic review identifies the characteristics of children with OSA.

The authors of this interesting paper provided us with a well-constructed literature review. They highlighted:

  • The prevalence of childhood OSA ranges from 1-5% and peaks at the age of 2-8.
  • Early detection and possible treatment are necessary
  • The AAO white paper on OSA strongly recommended that orthodontists be familiar with the signs and symptoms of OSA and refer where necessary.
  • Children with OSA may have clear signs of malocclusion that are reflected in an increased ANB, increased lower facial height and narrowed maxilla.
  • There is limited evidence to suggest that malocclusion is associated with OSA.

A team from Wuhan, China, did this systematic review. The EJO published the paper.

The dental and craniofacial characteristics among children with obstructive sleep apnoea: a systematic review and meta-analysis.

Yanxiaoxue Liu et al. EJO advanced access: https://doi.org/10.1093/ejo/cjac074

What did they ask?

They undertook the review to answer the following;

“Is there an association between paediatric OSA and children’s dental and craniofacial characteristics”?

What did they do?

They conducted a systematic review that they prospectively registered with PROSPERO and wrote following PRISMA guidelines.

The inclusion criteria were;

Children less than 18 years old. The study group was children with OSA, who they compared with healthy children. OSA had to be diagnosed with overnight polysomnography. The outcomes were any craniofacial or dental characteristic detected by any method.

The authors decided to concentrate on ANB and overbite as the primary outcomes. However, they also included some secondary outcomes derived from a limited set of cephalometric and relevant dental measurements. For example, arch width, overjet and presence of crossbites.

They included cross-sectional studies in the review.

They followed the standard steps of electronic and hand searching—identification and filtering of papers followed by data extraction and meta-analysis. Two examiners then measured the risk of bias with the Joanna Briggs Institute critical appraisal checklist for cross-sectional studies. Finally, they classified the certainty of evidence using the GRADE evaluations tool.

What did they find?

They identified a final sample of 16 studies and included 14 in the meta-analysis. Interestingly, they excluded two papers because they were multiple publications of the same study.

When they looked at the risk of bias, most studies had an unclear risk of bias. Notably, the GRADE analysis revealed that there was a very low to low certainty for the outcomes. So, things aren’t looking too good at this point, but we shall press on.

The team presented a large amount of data on many outcome measures and relevant meta-analyses. They did an excellent job confining the amount of cephalometric data, which was great. However, I do not have enough space to analyse all this data. I thought that the main points were that children with OSA had greater;

  • Overjet (mm) compared to the control group MD = 0.86, 95% CI: 0.20 to 1.51; P = 0.01)
  • ANB angle; MD = 1.78, 95% CI: 1.04 to 2.52; P < 0.00001
  • FH-MP and SN-MP angles; FH-MP angle; MD = 3.65, 95% CI: 2.45 to 4.85; P < 0.00001), (SN-MP angle; MD = 2.02, 95% CI: 0.91 to 3.13; P = 0.0004.

The OSA group had smaller inter-molar width than the control group; MD = −1.86, 95% CI: −3.52 to −0.20; P = 0.03).

Their overall conclusions were:

“Children with OSA had a decreased upper arch width, increased overjet, decreased mandibular growth, short hard palate and increased vertical growth”.

However, they also pointed out that the overall level of evidence was low. Most importantly, the statistically significant differences were minor.

What did I think?

Some readers will remember that we decided only to include interesting and scientifically sound systematic reviews in our posts. This was because we are under an onslaught of systematic reviews that add little to our knowledge. I decided to include this review because it was well written. Importantly, it attempted to answer a relevant and interesting clinical question. When I started reading the paper, I was concerned that I was heading for a “cephalometric festival of superfluous findings adding nothing” (CFSFAN). However, it was great that the authors confined their analysis to a few relevant measurements. As a result, I thought that the review was a good addition to knowledge.

When I looked at the results, I saw that they reinforced our clinical impression that children with OSA have increased overjets and mandibular deficiency with high angle and narrow maxilla. Nevertheless, I must be cautious in interpreting this data because the effect sizes were tiny, even though these were statistically significant differences. Therefore, these results represent a tendency rather than clear findings.

So, we must be cautious before we all identify OSA with cephalograms and expand maxillae as a first line of treatment. We also need to consider that association does not mean causation. I am sorry to be Professor misery!

In summary, this paper provides information that may point the direction of future clinical research into the orthodontic treatment of this important group of children.


Related Posts

Have your say!

  1. surely it is time for 3D facial analysis instead of cephs?
    the technology and analysis software has been around for years and seems to be ignored by orthodontics. It has been shown that 3D imaging can be related to ceph data and a BIG plus is no invasive radiation (and yes )

    • Here’s a new paper addressing this:

      Sutherland K, Kim S, Veatch OJ, Keenan BT, Bittencourt L, Chen NH, Gislason T, Han F, Jafari N, Li QY, Lim DC, Maislin G, Magalang U, Mazzotti DR, McArdle N, Mindel J, Pack AI, Penzel T, Singh B, Wiemken A, Xu L, Sun Y, Tufik S, Schwab RJ, Cistulli PA. Facial and Intraoral Photographic Traits Related to Sleep Apnea in a Clinical Sample with Genetic Ancestry Analysis. Ann Am Thorac Soc. 2023 Feb 13. doi: 10.1513/AnnalsATS.202207-577OC. Epub ahead of print. PMID: 36780658.

      • Thanks, I will have a look at this one.

      • https://eurjrhinol.org/Content/files/sayilar/98/EJRA_20220078_nlm_new_indd.pdf

        And another one on the mechanism of how the entities mouth-breathing, facial structure, and obstructive sleep apnea may be related.

        • Thanks, I could not get the link to work, can you post it again?

        • Here it is again. I think the site was down that day. https://eurjrhinol.org/en/the-mechanics-of-mouth-breathing-and-its-role-in-nasal-and-sleep-disorders-16286

          To summarize, the article distinguishes between mouth-breathing and OSA as a two step process: Mouth-breathing is the first step as one descends into sleep, and in deeper sleep, there can be actual over-relaxation of the tongue and palate as well, partially or completely due to deeper loss of tone temporarily in some cases. Mouth-breathing and tongue collapse are related like the loading and then firing of a gun in that mouth-breathing is required to permit tongue collapse in most, but tongue collapse does not always result, thus in children with favorable tone, mouth-breathing may be present but not OSA on a sleep study. Further, models were used to find the characteristics of jaw and occlusive structures that predisposed to mouth-breathing. Strain at the lips from anterior occlusion and obtuse gonial angle created the most mechanical predisposition to mouth-breathing. Thus, orthodontia in improving lip strain by improving the anterior harmony (decreasing their separation that creates strain on the orbicularis) between maxillary and mandibular teeth or improving gonial angle obtuseness that predisposes to mouth-breathing would be expected to be the most useful invasive form of mouth-breathing, and subsequently OSA treatment.

          In contrast, A mechanical predisposition to mouth-breathing in children with normal tone, nonobese tended towards development of an adenoid to compensate for the dead space of nasal underuse. These children have mouth-breathing but not OSA on study. Removal of the adenoid does not help because the mechanical cause, has not been removed, and is downstream to the cause. Tonsils form from partial tongue and palate collapse during sleep where negative pressure draws pharyngeal tissue into the lumen and creates secondary inflammation. These are children with poor tone, obesity, AND a partial predisposition to mouth-breathing. Here, the solution is to improve tone AND to reduce tendency to mouth-breathe. Finally, in adults with complete tongue and palate collapse, the pharyngeal strain is more distal, pulling on the base of tongue, uvula, and epiglottis, where these structures remodel to the point where tone is very poor at baseline, and nothing helps but blasting the airway with CPAP during collapse. So, to me, orthodontia may be the key to prevention of mouthbreathing, which itself is the key to permitting the next stage of osa collapse, that has more to do with overall tone, and tone of the tongue muscle. I personally think T+A simply remove the visual manifestation and downstream consequences of the problem.

          • Thanks for the clarification. While the paper raises some interesting points. I think that it is a bit of a quantum leap to suggest that orthodontics can prevent mouthbreathing. Have you any evidence that this is the case?

          • Nothing prevents mouth breathing. But one can reduce the strain upon the lips that are responsible, and one can support the jaw with devices like knee braces designed for the head that keep the jaw closed. The only thing that “works” to prevent mouth-breathing are using things that keep the mouth closed. Tape doesn’t work unless its enough tape to keep the mouth closed. Chin straps don’t work unless they provide enough support to keep the mouth closed. Clinical data isn’t required, just the fact of whether the mouth stays closed or not in an individual which varies from person to person, and differs at different tones, or sleep stages, with or without sedatives, and at different stages of life and even season. This may or may not require devices. One can observe with their own eyes whether the mouth stays closed by watching someone sleep with the. device. All of the other symptoms are just noise. It is obvious without data that people with obvious skeletal problems have more mouth breathing than people with ideal facial skeletons. It just takes observation. Tom Brady’s jaw structure is less inclined to mouth-breathing than one with the Hapsburg jaw. But with enough force, any jaw can be held shut even in low tone sleep. We have all read the literature. We have all published. The literature will never answer the questions. Orthodontics of course doesn’t fix mouth breathing. But, it is obvious that if the teeth themselves physically interfered with lip closure by having very unbalanced occlusion (lip strain), improvement in the harmony or distance between upper and lower bite would reduce lip strain and thus of course make it easier to keep the mouth closed. There are no single solutions, just mechanical ones. If we spent less time looking for the elusive data and just had discussion about what is in front of our eyes, there would be more productivity. That being said, of course, I of course respect people who do studies and collect data. That is their choice, and I participate as well. But, for my own kids, or if people asked me how to really make clinical choices, I would say by using one’s eyes and brain. I would not use the “data” to make clinical decisions for my kids.

  2. Without looking at tongue resting position, mouth breathing, history of breast feeding, head position, side plumb line, etc. study is severely lacking. Study supports results one is looking for at the beginning.

  3. Thanks, Kevin. The current line of thought is that only a subgroup of children with specific systemic lifestyle choices and craniofacial characteristics may benefit from an orthodontic intervention – clinical phenotyping has not yet been entirely determined. In other words, not all children that have a subset of altered craniofacial and have an OSA diagnosis necessarily face significant benefits long-term after an orthodontic intervention. A blanket approach is unnecessary and counterproductive (it gives the impression/hope to the families that the disease is being taken care of). We need to individualize health care (precision medicine) and not treat a lot just for a few that may benefit. In this conversation, orthodontic management is likely indicated in those children, but the OSA impact is likely nil or short-lived for some. Nevertheless, a similar systematic review with meta-analysis was published in 2022 in the Journal of Clinical Sleep Medicine (J Clin Sleep Med. 2022;18(7):1865–1875). The difference is that the control group was only those with a negative OSA diagnosis (PSG used) – so a more stringent criterion. No differences were observed in the considered cephalometric angles. So, in this previous review, even the cephalometric values assessed as having statistical significance were not. Long story short, the expected set of craniofacial problems is not supported by currently available data. There is little more clarity regarding dental malocclusion characteristics. As you mention in your blog, with the available very low level of evidence, I would paraphrase the conclusion of the 2022 review “Due to the very low to moderate level of certainty, neither an association nor a lack thereof between craniofacial morphology and pediatric OSA can be supported by these data.” It could be argued that the absence of evidence is not evidence of absence. This story has not yet a final chapter.

  4. Thank you Prof O’Brien for posting your review of this systematic review.

    I’d like to commend you sir for your astute observation that each and all of us here, as qualified dental professionals, ‘….must be cautious before we all identify OSA with cephalograms and expand maxillae as a first line of treatment….’. This concept I think should be obvious to most, but unfortunately sometimes is not. As you’d mentioned, OSA can only be definitively diagnosed (i.e., ‘identified’) by an overnight/in-lab PSG ‘sleep study’ performed by a fully qualified Sleep Medicine physician-guided team; dental professionals are only qualified to definitively diagnose and treat so-called dental diseases (e.g., maxillary transverse deficiency, etc.), which might often be co-morbid with OSA symptoms, and when appropriately treated, might often result in OSA symptom resolution.

    As I have mentioned here before, what you’d reported from this review as being ‘trending’ skeletal malocclusion traits often co-morbid with pediatric OSA (“Children with OSA had a decreased upper arch width, increased overjet, decreased mandibular growth, short hard palate and increased vertical growth”.), is nothing new professor; medical and dental literature dating from recently and back to the late 19th-Century, based mostly upon controlled observational studies (as EBM did not exist until after the Nazi war criminal trials at Nuremberg), have repeatedly reported on the phenomenon of Early Childhood (under 72 months) Malocclusion (ECM) and compromised naso-respiratory ability. It is unfortunate I think that most of the papers selected for this particular, and other published systematic reviews on this topic, did not much consider the most vulnerable (to Dx:OSA) 2-8 year old cohort; but did include so-called ‘children’ under 18 yrs old who would soon be eligible to enlist in the US military and vote in our elections.

    I’d like to propose the following hypothesis:

    The specific skeletal malocclusion phenotypes cited in this systematic review (e.g., decreased upper arch width, increased overjet, decreased mandibular growth, short hard palate and increased vertical growth) as being frequently co-morbid with OSA, are usually initially evident between 30-72 months of age, and they will seldom, if ever, self-correct, will usually persist beyong 72 months of age, and often become more complex without therapeutic intervention, either already are, or will reliably become co-morbid with compromised respiratory health during wakefulness and sleep, and finally, now confirmed by published evidence to be often associated with decreased Quality of Life and poor general health.

    According to Karl Popper, in order for a scientific hypothesis to be considered valid and thus worthy of exploration, it must not only be supportable, but also refutable. As support for the aforementioned speculation seems now to be abundantly gathering, being mostly a question of balance, I would appreciate refutation from anyone on this esteemed platform who might be willing and able to offer it.

    Thanks for considering my request for input.

    • Thanks for the comments. We agree over the treatment concept etc. However, with respect to the last section of your comments, I am not so sure. Firstly, we need to consider whether their findings of narrow maxilla etc occurred by chance. I wonder if this could be the case because our confidence in this data is low. Furthermore, we do not know if the morphology caused the OSA or were a consequence of it. Finally, we also need to recongnise that no orthodontic appliances can substantially alter skeletal morphology. So in many ways we are in a catch 22 position and it is clear that we need high quality trials in this important area.

      • Dear Prof O’Brien

        Thank you for your input. Will you please elaborate upon your implication that, in the absence of evidence for supporting a causal relationship between OSA and maxillary transverse deficiency(MTD), ‘….we need to consider whether their findings of narrow maxilla etc occurred by chance.’? As it seems fairly well understood now that skeletal MTD is usually first recognizable within the first 6 years of life (Moyers, R. et al), how might you explain how maybe ‘…a narrow maxilla etc occurred by chance.’?

        Thank you for considering my question.

        • Thanks, I am not sure if I made myself clear. By chance I meant statistical chance of a certain feature or event being detected as signficant.

          • Thanks for the clarification.

            However, I think it is important Prof. O’Brien, that even in the absence of statistical elimination of a ‘chance’ (apparent causal) relationship between two co-morbid disease entities, say for instance Dx:maxillary skeletal transverse deficiency and Dx: conductive hearing loss in a 5 year old child, why would one not attempt to at least begin therapeutic resolution of the MTD given that it might very well indeed (references upon request) coincide with improved hearing acuity?

          • are you suggesting that orthodontics to correct MTD will influence hearing?

  5. Is morphology conducive to OSA or is it the other way around. Chicken and egg.

    • Dear Jim and Robert

      In 2018 Howard Stupak discussed in detail this very ‘chicken-egg’ sequencing issue relative to ATH and associated ‘nasal disuse’; also described as Guilleminault’s ‘alternative hypothesis’:

      Gravitational forces, negative pressure and facial structure in the genesis of airway dysfunction during sleep: a review of the paradigm
      Howard D Stupak et al. Sleep Med. 2018 Nov.


      The recent and distant literature has extensive discussion of how sleep apnea, adeno-tonsillar growth, and facial structural deformity are related. Conventionally, the order of cause and effect is as follows: (1) Inflammatory/infectious process→tonsillar/adenoid tissue growth→(2) airway obstruction and mouth breathing/Obstructive Sleep Apnea (OSA)→(3) altered facial structure (adenoid facies). Using this same reasoning, adenotonsillectomy is the first line of treatment in the prevention of structural abnormalities. However, through a lifetime of clinical research Christian Guilleminault and his colleagues have challenged this paradigm. Through multiple articles and studies, Guilleminault et al., teach that even slight (subclinical) facial structure/muscle tone variations may be the inciting event triggering mouth-breathing and the eventual adenotonsillar growth in most patients. Essentially, this is the reverse of the conventional paradigms. Initial treatments therefore shift from simplified removal of inflammatory tissue to limiting mouth-breathing via musculo-skeletal modification. The purpose of this paper is to synthesize and analyze the recent (and distant) relevant literature to provide support for, and provide a potential anatomic mechanism for Guilleminault et al.’s paradigm-questioning clinical observations.

      Keywords: Adeno-tonsillary hypertrophy; Facial structure; Mouth-breathing; Nasal obstruction; Nasal valve; Obstructive sleep apnea.

    • I agree, we do not know which came first and again throws doubt on orhtodontic treatmet of SDB

  6. I read the list of associated cephalometric and clinical factors and thought that they might also describe what we would expect to see in a description of Adenoid Facies.

    Certainly there is a correlation between SDB and hyperplastic/hypertrophic adenoids and tonsils in pediatric patients. I wonder if the cephalometric findings are more the egg than the chicken.

    Also, this being more the symptom than the cause would be one more factor in favor of orthodontists screening for SDB, making the appropriate referral to a physician board certified in sleep medicine, and then participating in the treatment of that SDB patient only if that physician indicates this is warranted.

    I also appreciated one of the final comments “We also need to consider that association does not mean causation.” That is right out of the writings of the great Scottish Philosopher David Hume. We cannot take a constant conjunction of events and infer a causal relationship. In this case, it could be dangerously misinterpreted that the cephalometric and clinical measurements were somehow responsible for the SDB, and that their correction would be the cure. More, it would seem that both the SDB and cephalometric and clinical measurements might be different effects of the same cause.

    Thanks for posting and for remembering how a Philosopher’s words from the 18th century still apply today!

  7. Dear Robert

    As you might be aware, there are myriad peer-reviewed papers on the positive effects of RPE in young children upon QOL and general systemic health(McNamara, J. et al) when Dx:’MTD-SDB co-morbidity is present’, and also RPE’s effect upon normalization of adenotonsillar hypertrophy (Yoon, A. et al). If, after your having made ‘the appropriate referral to a physician board certified in (Pediatric?) sleep medicine’, might you still correct the maxillary transverse deficiency in an 8 yr old even when you might not have been recruited by said Sleep Medicine physician for collaborative care of this symptomatic child?

    Quite possibly the SDB symptoms will not resolve after RPE, but of course, also the skeletal MTD will not resolve (self-correct) without your competent intervention. Most orthodontic care providers (Pediatric Dentists, Orthodontists and GDP’s) for children with ‘malocclusion-SDB co-morbidity’ in the US do not have ready access to qualified Pediatric Sleep Medicine physicians as there are now so few of them, so, I’m curious, how might you advise when/if to treat MTD w/SDB symptom co-morbidity when, say, you might be in your own?

Leave a Reply

Your email address will not be published. Required fields are marked *