Is there an association between facial pattern and obstructive sleep apnoea?

I am going to discuss a new study on childhood obstructive sleep apnea (OSA). As we are all aware, this remains a highly controversial topic in orthodontics. Over the past few years, I have outlined the relationship between orthodontics, facial growth, and obstructive sleep apnoea several times. In my posts, I have asked those advocating orthodontics for OSA to provide evidence supporting this link. It is, therefore, encouraging to see this new paper addressing an important question.

The lead author is Derek Mahoney, a specialist practitioner based in Australia. On his personal website, he states that he is 

“a world-renowned specialist orthodontist who feels that early orthodontic intervention can minimise the extraction of teeth, with a focus on improving the airway and the facial profile”.  

A multinational-based team did this research. Orthodontics and Craniofacial Research published their paper.

They wrote a well-balanced introduction. Importantly, they point out that there is a higher prevalence of OSA in a referred paediatric orthodontic population compared to a healthy population. They also outline that some studies report a higher prevalence of malocclusion in children with OSA than in controls without OSA, while others do not support this association.

What did they ask?

They did this study to

“explore whether vertical and sagittal craniofacial patterns are associated with the presence and severity of OSA, as confirmed by polysomnography (PSG), among children attending orthodontic consultations”.

What did they do?

They carried out a retrospective cross-sectional study. They did this in the following stages.

• Identification of a cohort of 4672 children who presented for consultation between February 2007 and December 2022. A specialist orthodontist saw the patients in one of nine private orthodontic clinics.
• During the initial consultation, the orthodontist screened all patients for Sleep-Disordered Breathing symptoms using the BEARS questionnaire. If this identified a child as being at risk, they referred them to a sleep physician for further assessment, which included a PSG conducted in a sleep laboratory.
• They then excluded all those who had not complete cephalometric and PSG tests. This reduced the sample to 3671 children.
• They analysed the cephalograms with the Sassouni analysis. From this, they classified the craniofacial pattern based on facial height as normal, short, or long. They defined the sagittal malocclusion categories as Class I, Class II, Class III, bimaxillary protrusion, or bimaxillary retrusion.
• They also classified the AHI severity as No Obstructive Sleep Apnoea (AHI=0), Mild (AHI=1- 5), Moderate AHI=6- 10), and Severe OSA as AHI greater than 11.
• The statistician analysed the data with Analysis of Variance to evaluate any association between facial pattern and AHI.

What did they find?

This paper is open access, making it easy for interested readers to access the data presented. I will concentrate on the significant findings. They found

• 97% of the included children with severe obstructive sleep apnoea had a long face pattern. However, those with mild and moderate OSA had a normal facial height.

• The children with severe OSA tended to have a Class II skeletal pattern.

Their conclusion was

“Class II profile, maxillomandibular retrusion and increased anterior facial height are highly present in children with severe OSA. Therefore, orthodontists are encouraged to consider the possibility of breathing issues in children presenting with such features”.

What did I think?

Previously, I was critical of some research presented by those advocating orthodontic treatment for breathing disorders. This was mainly because the research was of low quality and lacked acknowledgement of its limitations. However, it was encouraging to see that these authors discussed the limitations of this study. These were:

The study was retrospective and may have been subject to some selection bias. Notably, the sample was drawn from patients attending specialist practice consultations who were interested in treating SDB. Many of these patients were referred for such issues. Furthermore, they excluded 22% of the original sample because of incomplete records. As a result, we cannot generalise these findings to the broader population. If we wanted to do this, we would have to examine a cross-section of the general population.

We also need to remember that association does not imply causation. This is important because we cannot claim that correcting the facial patterns detected will cure SDB. Notably, the authors of this paper did not make this claim.

They also noted that they could not consider other risk factors for Obstructive Sleep Apnoea due to the retrospective nature of the study.

I also have two final concerns. I felt it was a shame that they used the Sassouni analysis, which is somewhat obscure and only produces categorical data. A more conventional analysis might provide more information. I was also unsure about the presentation of the data. It was somewhat vague, as the main table contained quite basic data, mainly in the form of proportions. This was relevant if the data were analysed with a Chi-squared test, yet they used ANOVA. This meant that I could not really work out their statistical analysis. This is a refereeing issue.

Final comments

This research must have involved significant effort, and I believe it adds to our knowledge. Importantly, the results and conclusions were clearly outlined, with a clear section on the study’s limitations.

I feel that we can conclude from this study that in a sample of referred orthodontic patients who attended a specialist practitioner with an interest in childhood breathing disorders, a high proportion with OSA had increased anterior facial height. We cannot establish a cause-and-effect relationship; therefore, orthodontic treatment directed at correcting the facial pattern will not necessarily correct the OSA.

It is good to see that the authors did not make this claim. However, the cynic in me feels that others will do this. Or perhaps this is just my concern about the whole breathing issue?