Do enlarged tonsils really influence dentofacial development?
Adenotonsillar hypertrophy can obstruct the upper airway, which may result in children breathing through their mouths. It is a common belief that this may influence their craniofacial growth and dental occlusion.
This concept tends to be based on various interpretations of the classic papers by Linder-Aronson and Harvold. However, these two papers are not the only sources of information on the relationship between enlarged tonsils and craniofacial morphology. We should consider other studies. This new systematic review, which incorporates data from many studies, is of both clinical and scientific importance.
A team from Wuhan, Guangzhou, China; West Virginia, USA; and Manchester, North of England, conducted this study. The AJO-DDO published the paper.
Tingting Zhao, Min Wang, Peter Ngan, Zhendong Tao, Xueqian Yu, Fang Hua, and Hong He.
AJO-DDO advance access: https://doi.org/10.1016/j.ajodo.2025.04.024
What did they ask?
They did the study to ask
“What is the evidence regarding the association between adenotonsillar hypertrophy and dentofacial characteristics in children”?
What did they do?
The team did an extensive systematic review of the current literature. The details of the review were:
Types of study
This included cross-sectional studies comparing the dental/craniofacial characteristics between children with and without adenoid or tonsillar hypertrophy.
Participants
Children aged less than 18 years with Adenoid and or tonsillar hypertrophy. These were the hypertrophy group (HG). Children without hypertrophy were in the control group (CG).
The authors of the included paper diagnosed hypertrophy using cephalometric assessments, oral examination, or nasopharyngeal endoscopy. Importantly, there was no universally applied measure for this variable.
They derived the outcomes of craniofacial pattern or dental features from cephalometric measurements, photographs and dental casts. Again, there was a large amount of variation in these measures.
Outcomes
The primary outcomes for the review were SN-MP and overjet for dental occlusion. They included several other cephalometric and dental measurements as secondary outcomes.
The team followed standard systematic review methods for their study. This was an experienced and well-regarded research team.
Importantly, they used the Newcastle-Ottawa scale of cross-sectional studies to evaluate any bias and the GRADE assessment to evaluate the certainty of evidence.
What did they find?
Following the initial filter of papers, they identified a final sample of 31 studies. These included data on 8,132 children; 3,845 were in the hypertrophy group, and 4,287 were in the control group.
Upon examining the quality of the studies, nine were unsatisfactory, 14 were satisfactory, 11 were good, and two were classified as of very good quality.
When they looked at the primary outcome measures. The following measurements were statistically significant. The SN-MP angle mean difference between the HG and CG was 2.2 degrees (95% CI=1.47-2.92). There were no differences in ANB between the groups. They reported minor differences for SNA and SNB angles.
Evaluation of dental features revealed a higher relative risk for Angle Class II and III malocclusion (1.29) and open bite (1.65) among children in the HG. Finally, the intermaxillary molar width was smaller in the HG group by 1.3 mm (95% CI=2.12-0.56).
The GRADE assessment revealed that for 16 of the outcome measures, the certainty of evidence was graded as very low.
Their conclusion was
“Based on evidence of very low or low certainty, children with ATH hypertrophy had different craniofacial characteristics than children without ATH”.
However, another important conclusion was
“The statistically significant differences identified are relatively small, raising questions about their clinical significance”
What did I think?
This was a thorough review conducted to a high standard by an experienced team. They followed standardised protocols and reported on a large amount of data. I have only highlighted the data I considered most important.
I found it interesting that they identified differences between the groups in their analyses. Importantly, these differences were statistically significant. However, we need to consider whether the effect sizes were clinically significant. To do this, we need to examine both the effect sizes and the 95% confidence intervals of the differences. This indicates that these findings are not clinically significant.
The authors of the review wrote a balanced and clear discussion. I particularly thought that their discussion of the shortcomings was very clear. They pointed out that it was difficult to analyse the data because of the marked differences in measurement of both hypertrophy and dentofacial form. They felt that this reflected the real-world diversity in study populations and methods. I would also like to add that some potential biases would be influenced by the large sample size of both publications and children included in the review.
Final comments
We now need to consider how this review enhances our understanding of this issue. This is for you as a practising clinician to decide. My impression is that this research suggests adenotonsillar hypertrophy is associated with statistically significant effects on morphology but these were not clinically significant. This is an important distinction.
So, can we be more sensible about expounding incorrect theories on the role of breathing and facial growth? Or at least have a rational discussion?
Key points
- Common Belief about Tonsils and Facial Growth: It is widely believed that enlarged tonsils and adenoids can negatively influence facial development, a notion often supported by classic studies cited by orthodontists and dentists.
- Limitations of Classic Studies: The foundational papers by Linder-Aronson and Harvold are not the only sources on this topic, and relying solely on them overlooks newer, comprehensive research.
- Purpose of the New Systematic Review: A recent systematic review aimed to investigate the evidence linking adenotonsillar hypertrophy with craniofacial characteristics in children.
- Methodology of the Review:** The study analyzed cross-sectional studies comparing children with and without hypertrophy, using rigorous tools like the Newcastle-Ottawa scale and GRADE for bias and evidence quality.
- Key Findings and Clinical Significance:** The review included data from 31 studies with over 8,000 children; it found statistically significant but clinically small differences in facial measurements, with very low certainty of evidence.
Emeritus Professor of Orthodontics, University of Manchester, UK.
Hi Kevin:
The study uses good ol’ fashioned 2D cephalometric analyses. However, Moyer and Bookstein [1] explained the inappropriateness of cephalometric analysis and lament that the newer techniques they proposed have failed to be implemented [2]. So, do most orthodontists still consider cephalometric analysis as the gold standard in maxillofacial analysis, even while acknowledging that one known weakness of 2D landmark data is superimposition; so the assumption that the location of a landmark in 2D is the same as that in 3D space is unwarranted? For example in the Results it talks about the mandibular plane angle (SN-MP). The SN plane is mid-sagittal while the MP is not. The third dimension has been lost, so the metric cannot fully account for the claim being made unless you believe in a (physically non-existent) cephalometric space. Perhaps the findings need to be taken with a grain of salt?
References
1. Moyers, R.E., Bookstein, F.L., 1979. The inappropriateness of conventional cephalometrics. Am. J. Orthod. 75, 599-617.
2. Bookstein, F.L., 2016. Reconsidering “The inappropriateness of conventional cephalometrics”. Am. J. Orthod. Dentofacial Orthop. 149, 784-97
And how does change the fact the children have large tonsils that needs to be assessed by an ENT familiar with paediatric sleep disordered breathing?
This recurring question might be of academic interest but it’s clinically futile as it does not change the fact these children need help.
Thanks for the comment. Your point is a good one and you are correct in the need for ENT involvement. But this was not the subject of the study.
Hi Kevin
As an orthodontist approaching my Twighlight, even after studying L-A and Harvold’s monkeys; I feel that this is an area of orthodontics that I am perhaps least prepared. Growth, development and environmental influence remain elusive in their prediction and prognosis. Personally, I would appreciate a set of diagnostic criteria to use as a guideline for ENT weighted referral scale. For example, perceived history of “mouthbreathing”, cepahlometric indication (appreciating the 2D flaws)- and from this paper enlarged tonsils may not be weighted as highly as we may hav done previously; OPG indications such as deviated septum, foggy antrums or suggestion of polys (again acknowledging limitations of this data alone) and other – facial feature and malocclusion type, snoring etc. I am a strong proponent that as orthos we are on the front line of detection of such potential pathology; however, I feel that ENT specialists are the clinicians best equipped to further diagnose and treat where necessary. So if that maybe the case, development of a structured and diagnostic collective criteria scale that we may teach and utilize clinically may be of value. If anyone knows of one in existence, would appreciate the reference/s. Meantime I err on the conservative side and suggest ENT referrals for many patients who may or may not benefit from intervention.
Back to the paper in question – duration, intensity and frequency; form and function echo – do you believe that the sample age of up to 18 years may confound detection of any effect that enlarged tonsils may have on dentofacial growth due to post adolescent atrophy and subsequent “normalization” of the environment in many individuals and do we know the incidence of associated facial and malocclusion types in populations without adenotonsillar hypertrophy, suggesting other environmental and genetic aetiologies? (another method to test the significance of this variable, rather than mining for characteristics within a solely “affected” cohort). How many dolichofacial, bilateral crossbite and anterior open bite patients exist who never experienced adenotonsillar hyertrophy?
Thankyou to the researchers and for your post
Thank you for hightlighting this article. I feel that Orthodontists/Paediatric Dentists or all dentists who see children should take interest in this topic. Ultimately we should still screen young children for adenotonsillar hypertrohpy & refer appropriately just like how we screen for OSA & refer as it’s a medical condition that needs to be addressed; regardless of whether there’s a link between adenotonsillar hypertrophy & dentofacial development. There should be more collaboration between the ENT, respiratory physicians & dentists in this area too. Perhaps somebody can develop or come out with a standard screening tool (maybe there’s already one which i’m unaware of) to screen for adenotonsillar hypertrophy similar to the STOP-BANG for OSA. What I do currently is ask patient/parents, check intra-orally for enlarged tonsils & look at the lat ceph x-rays for s/s. I also cover 1 nostril alternately & ask px to blow/breathe to check if they have any partial nasal obstruction.
Not sure if you are aware of the Paediatric Sleep Questionnaire (PSQ) or not Jason [but just in case 🙂 ] which screens for sleep disordered breathing in children. Many references but one here – https://pmc.ncbi.nlm.nih.gov/articles/PMC9497855/ If they answer Yes to 8 or more questions then quite high sensitivity and even better specificity at detecting SDB. Not perfect but pretty good.
As to craniofacial morphology, these differences are small (so you wouldn’t treat solely on the chance it alters facial growth but there are many other valid reasons for treatment) and cause and effect (chicken-egg argument) cannot be demonstrated by cross-sectional studies. One prospective study from Sweden (https://angle-orthodontist.kglmeridian.com/view/journals/angl/85/5/article-p728.xml ) that followed individuals at ages 3, 7 and 11 found “Habits, allergies, or breathing disturbances found at 3 years of age had no associations with malocclusions at 11.5 years of age”.
Hi Dr Peter Miles,
Thank you so much for the references.
My point is that regardless of whether there’s a link between adenotonsillar hypertrophy & dentofacial development, we as Orthodontists should screen for adenotonsillar hypertrophy as we can see it intra-orally & on radiographs. I spoke to ENTs from where i’m from (Singapore) & it seems like our population is not open to surgical management of adenotonsillar hypertrophy. Also there’s very low awareness on this topic among patients & parents as everybody assumed this is normal due to either poor air quality in a city state or allergy to dust, etc hence no diagnosis or treatment is sought for it usually.