Obesity and craniofacial structure as risk factors for obstructive sleep apnoea: Impact of ethnicity



    1. Centre for Sleep Health and Research, Department of Respiratory Medicine, Royal North Shore Hospital, Sydney
    2. NHMRC Centre for Sleep Health (CIRUS), Woolcock Institute of Medical Research, University of Sydney, Sydney
    Search for more papers by this author

    1. Centre for Sleep Health and Research, Department of Respiratory Medicine, Royal North Shore Hospital, Sydney
    2. NHMRC Centre for Sleep Health (CIRUS), Woolcock Institute of Medical Research, University of Sydney, Sydney
    3. Department of Respiratory Medicine, Gosford Hospital, Gosford
    4. School of Medicine and Public Health, University of Newcastle, Newcastle, New South Wales, Australia
    Search for more papers by this author

    Corresponding author
    1. Centre for Sleep Health and Research, Department of Respiratory Medicine, Royal North Shore Hospital, Sydney
    2. NHMRC Centre for Sleep Health (CIRUS), Woolcock Institute of Medical Research, University of Sydney, Sydney
    Search for more papers by this author

  • The Authors: Kate Sutherland, PhD, is a Postdoctoral Research Fellow at the Woolcock Institute of Medical Research and Royal North Shore Hospital, with research interests in craniofacial and upper airway imaging in OSA. Richard Wai Wing Lee, MD PhD, is a Research Fellow at the Woolcock Institute of Medical Research and a Staff Specialist at Gosford Hospital, with research interests in craniofacial and upper airway imaging in OSA. Peter Cistulli, MD PhD, is a Professor of Respiratory Medicine and Head of the Discipline of Sleep Medicine at the University of Sydney, and Head of Respiratory and Sleep Medicine at Royal North Shore Hospital, with research interests in the pathophysiology of OSA and novel treatments for OSA.


Peter Cistulli, Centre for Sleep Health and Research, Department of Respiratory Medicine, Level 8, Main Block, Royal North Shore Hospital, Pacific Highway, St Leonards, NSW 2065, Australia. Email: peter.cistulli@sydney.edu.au


OSA is the result of structural and functional abnormalities that promote the repetitive collapse of the upper airway during sleep. This common disorder is estimated to occur in approximately 4% of men and 2% of women, with prevalence studies from North America, Australia, Europe and Asia indicating that occurrence is relatively similar across the globe. Anatomical factors, such as obesity and craniofacial morphology, are key determinants of the predisposition to airway collapse; however, their relative importance for OSA risk likely varies between ethnicities. Direct inter-ethnic studies comparing craniofacial phenotypes in OSA are limited. However, available data suggest that Asian OSA populations primarily display features of craniofacial skeletal restriction, African Americans display more obesity and enlarged upper airway soft tissues, while Caucasians show evidence of both bony and soft tissue abnormalities. Our recent comparison of Chinese and Caucasian OSA patients found for the same degree of OSA severity. Caucasians were more obese, and Chinese had more skeletal restriction. However, the ratio of obesity to craniofacial bony size (or anatomical balance, an important determinant of upper airway volume and OSA risk) was similar between Caucasians and Chinese OSA patients. Ethnicity appears to influence OSA craniofacial phenotype but furthermore the relative contribution of the anatomical factors underlying OSA risk. The skeletal restriction craniofacial phenotype may be particularly vulnerable to increasing obesity rates. Better understanding of craniofacial phenotypes encompassing ethnicity may help improve OSA recognition and treatment; however, further studies are needed to elucidate ethnic differences in OSA anatomical risk factors.


OSA is a disorder characterized by the repeated partial or complete collapse of the upper airway during sleep. These nocturnal complete (apnoea) or partial (hypopnoea) airway closures preclude or reduce airflow, leading to oxygen desaturation and sleep fragmentation.1 The condition is associated with a multitude of adverse consequences, including excessive daytime sleepiness, neurocognitive impairment, increased motor vehicle accident risk, diabetes, increased cardiovascular morbidity and all-cause mortality.2–4 As such, this common disorder poses a significant public health burden. OSA occurs as a result of anatomical and functional abnormalities of the upper airway that compromise airway space and increase airway collapsibility during sleep. Although alterations in neuromuscular and ventilatory control mechanisms can contribute to the reduced airway patency underlying OSA,5 anatomical abnormalities play a key role in the development of OSA.6 Anatomical risk factors for OSA include obesity, excess regional adipose tissue, enlarged upper airway soft tissues and craniofacial abnormalities.7 Obesity and craniofacial factors are reported to explain two-thirds of the variance OSA severity measured by AHI.8 The interaction between these anatomical factors, together with patient demographics and symptoms, are the main determinants of the likelihood and severity of OSA in the majority of patients.9,10 However, as the pathogenesis of OSA is multifactorial, the relative contribution of the various anatomical OSA risk factors is likely to differ in individuals and different ethnic populations.


Large population studies to investigate OSA prevalence in adults that incorporate objective overnight monitoring now exist across multiple continents. Caucasian populations in Europe,11 North America12–14 and Australia15 all show consistency in their estimates of adult OSA prevalence. More recent studies from Asia of Korean,16 Hong Kong Chinese17,18 and Indian19,20 populations have shown OSA prevalence rates that closely mirror those from Western populations. Overall, regardless of the population that was surveyed, these studies report very similar rates of OSA prevalence in the range of 3–7.5% in males and 2–3% in females. Within North America, studies comparing Caucasians and African Americans,21,22 and Caucasians, African Americans and Hispanics23 have found no significant difference in OSA prevalence between ethnicities. However, there are some suggestions that African Americans may have more severe OSA than Caucasian counterparts.21 Therefore, population studies show similar prevalence of OSA across different countries and ethnicities despite different demographics, craniofacial structure and obesity in the various populations. Hence, it is likely that the predominant underlying pathophysiology differs between ethnic groups, with differential contributions of craniofacial and obesity risk factors.



Obesity is the most common and well-recognized risk factor for OSA. Its contribution to sleep-disordered breathing appears to outweigh other established factors that predispose to the condition, such as genetics, upper airway abnormalities and craniofacial phenotype.10 In sleep clinic populations, the greater proportion of OSA patients are overweight.24 And conversely, OSA appears to be over-represented in clinical patients who are obese.25 For example, in a series of severely obese bariatric surgery candidates, overnight sleep monitoring has indicated the presence of OSA (AHI > 10/h) in as high as 95.7% of men and 65.9% of women.26 Population-based prevalence studies of OSA all indicate a clear association between AHI ≥ 5/h and increased body habitus,16–18,20,27,28 and even modest increases in weight have been associated with increases in OSA severity.29 Estimates have been given that in 41% of adults who have mild or worse OSA, sleep-disordered breathing can be attributed to excess weight. In moderate or worse OSA, the proportion is higher, with 58% of adults having sleep-disordered breathing attributable to overweight or obesity.30 Obesity is unique among other factors that predispose to OSA (such as age, male gender and postmenopausal status in women); in that, it is a modifiable risk factor. Weight loss through lifestyle, surgical and pharmacological interventions lead to demonstrable improvements in OSA severity.31–33 Furthermore, randomized controlled studies of weight loss have now reported observations at 1 and 2 years following the weight loss intervention, suggesting that benefits in terms of AHI reduction are sustained long term.32,34,35

Despite the clear association between OSA and obesity, and weight loss and OSA amelioration, we do not yet fully understand the mechanisms behind this relationship. Although increased overall body weight or BMI are clearly linked to the occurrence of OSA, particular patterns of fat distribution and fat type may be specifically relevant and underlie the pathophysiological mechanisms. Neck fat deposition (neck circumference), central obesity (waist circumference) and particularly abdominal visceral adipose tissue have been specifically related to OSA.28,36–38 Although obesity may alter upper airway functional control through neurochemical means via its metabolic and inflammatory consequences,39 anatomical loads from excess fat presumably alters upper airway mechanical function through direct effects on structure. Increases in local fat deposition around the pharynx and in surrounding soft tissues (reflected in larger neck circumferences) compromise airway space resulting in a narrower, more collapsible airway. The parapharyngeal fat pads are adipose tissue structures that laterally border the airway, and their enlargement in obesity could readily impact on airway size. However, a direct relationship between the size of the fat pads and OSA severity has not been demonstrated.40,41 Upper airway structure in obesity and weight loss in OSA, and their effects on upper airway function require further exploration. Central obesity is associated with reduced lung volume from fat deposition around the chest and abdomen which reduces caudal traction on the pharynx, thereby increasing pharyngeal collapsibility.26 Therefore, obesity has the potential to impact on pharyngeal patency and collapsibility via mechanical loads on both the upper airway and respiratory system. Different patterns in relative fat distribution may partially explain differences in OSA prevalence between genders. Recent evidence suggests that OSA in women appears to be associated with increased neck fat implying a direct impact of adipose tissue on the upper airway, whereas increased abdominal fat was the primary obesity factor in men that contributed to OSA.42

Craniofacial morphology

Although obesity is considered the major risk factor for OSA, craniofacial morphology is increasingly acknowledged as an important interacting factor in the pathogenesis of OSA. Craniofacial characteristics associated with OSA include aspects of skeletal morphology pertaining to the mandible, maxilla, cranial base, hyoid and head position, as well as soft tissue morphology, relating to size of upper airway soft tissues.

Aspects of skeletal craniofacial morphology relating to OSA have been explored in many studies primarily using cephalometric analyses of lateral radiographs. Across all the studies that have been performed, key characteristics have been consistently found to relate to the presence of OSA. Particularly prominent are differences in maxillary-mandibular morphology and relationship. In terms of the maxilla, a shorter length,8,43 and also narrower and more tapered maxillary arch44 are noted. Mandibular deficiency has also been associated with OSA, and three-dimensional imaging techniques confirm a smaller mandibular enclosure area in OSA subjects.45–47 Findings of greater mandibular retroposition suggest the relative positioning of the maxilla and mandible is also important in the development of OSA.48,49 Another common cephalometric feature to be reported is inferior displacement of the hyoid bone (or increased mandibular plane to hyoid distance).50–53 The lowered hyoid position may be a reflection of the greater pressure exerted by excess pharyngeal tissues in the craniofacial bony compartment of apneic subjects,9 as differences in hyoid position between apneics and control subjects are no longer present when controlling for tongue volume.45 It has also been postulated that the hyoid position may reflect adaptive changes to maintain airway patency.54 Cephalometric studies also often report a smaller cranial base (length or angle)51,55,56 that may relate to OSA occurrence through an influence on craniofacial growth, and the development of maxillary deficiencies and relative mandible positioning. Other reported skeletal features include an increased anterior facial height57 and increased craniocervical angle (extended natural head position).58,59 However, rather than craniofacial risk factors, these aspects are more likely to represent compensatory phenomena to maintain pharyngeal airway patency secondary to upper airway obstruction.60

OSA patients also have demonstrated morphological differences in upper airway soft tissues compared with non-OSA controls. These generally refer to enlargement of the tongue, soft palate and uvula.43,52,55 Enlarged upper airway soft tissues in apneic patients have also been identified using MRI.6 These confirm a larger tongue volume in OSA patients, and additionally show larger lateral pharyngeal walls and increased volume of the parapharyngeal fat pads compared with non-OSA controls.

Imaging studies also show reduced upper airway dimensions in patients with OSA,52,55,61 a result of the impact of craniofacial morphology in terms of both enlarged upper airway soft tissue and/or maxillary/mandibular deficiency on upper airway space.62 The craniofacial characteristics commonly associated with OSA are summarized in Table 1.

Table 1.  Craniofacial features commonly associated with OSA Thumbnail image of

Interaction of obesity and craniofacial morphology

It is clear that both obesity and craniofacial morphology are key anatomical risk factors that predispose to the development of OSA. Cephalometric studies of craniofacial structure suggest divergent pathogenesis of sleep-disordered breathing between obese and non-obese OSA patients.48,56,63–65 Non-obese patients tend to display more craniofacial skeletal abnormalities, such as a small mandible and maxilla, suggesting that they arrive at sleep-disordered breathing through the effects of skeletal restriction on airway patency, whereas obese patients show comparatively less bony restriction and instead have a larger soft palate and tongue, and associated antero-inferior positioning of the hyoid bone. This suggests that crowding of the airway space through enlargement of soft tissues is the main catalyst for upper airway collapse in obese patients. However, there is overlap between the occurrence of obesity and craniofacial skeletal abnormalities resulting in intermediate phenotypes, as indicated by cephalometric studies comparing different degrees of obesity.50,66

A synergistic contribution of obesity and craniofacial factors to upper airway collapsibility has been demonstrated, and furthermore, the relative contribution between obesity and craniofacial factors in an individual appear to determine the region and type of pharyngeal collapse (velopharyngeal and/or oropharyngeal collapse).9 A combination of mandibular enclosure area and degree of obesity explains a large proportion of the variance in OSA severity.47 The relationship between the limiting size of the craniofacial skeletal enclosure and the amount of soft tissue within has been termed ‘anatomical balance’.67 The ratio of tongue volume to oral cavity volume is higher in OSA than in non-OSA subjects.68 Cephalometric analysis of craniofacial size and tongue cross-sectional area shows that when OSA patients and non-OSA controls are matched on maxillomandibular dimensions, OSA patients have a significantly larger tongue.67 The larger tongue size in OSA patients is also apparent when non-OSA controls are matched on both maxillomandibular dimensions and obesity. Therefore, the interaction between craniofacial factors, upper airway soft tissues and obesity, and not just individual anatomical factors alone, appear to be key determinants of upper airway structure and collapsibility. Imbalances between these anatomical determinants may arise through genetic and environmental factors, and represent an important mechanism in the pathogenesis of OSA.

Additionally, regional deposition of fat in the craniofacial region may be important in the development of OSA. Buccal fat deposition (or ‘chubby cheeks’) has been shown to correlate in amount with visceral abdominal fat, and additionally, may be equally metabolically active.69,70 Surface analysis of facial soft tissue show excessive soft tissue in the bucco-submandibular regions of the faces of OSA subjects compared with those without OSA.71 In this way, craniofacial obesity patterns may be an indicator of the presence and severity of OSA.72,73


Ethnicity incorporates genetic, cultural and environmental factors, which will individually or in combination influence other recognized risk factors for OSA, such as obesity and craniofacial morphology.74 The relative importance of these OSA risk factors may, therefore, vary with ethnicity.


It is well recognized that the incidence of overweight and obesity has increased dramatically from past decades. Rapid urbanization across the globe, with associated decreased energy expenditure and increased energy availability, make the growing obesity trend a worldwide phenomenon, and not just in Western countries.75 Furthermore, worldwide increases in early childhood obesity suggest that the trend will not be reversible for some generations.76 Obesity is, therefore, a significant public health issue across the world, and sleep-disordered breathing is clearly one of the adverse health consequences associated with its increasing prevalence.

All prevalence studies of OSA across Europe, Asia, Australia and the USA show a positive association between excess weight and the occurrence of sleep-disordered breathing. However, the Asian populations studied are generally less obese compared with their Caucasian counterparts in communities from Australia and the USA.15–18,27 Despite the relatively lower BMI in these Asian community samples, the relationship between increased BMI and OSA appears to be comparable, with one SD increments in BMI showing similar odds ratios for the presence of OSA in both Western and Eastern nations.30 Optimal BMI is lower in Asians, and there is a lower cut-off to describe obesity based on body fat composition and associated health outcomes.77 The OSA risk attributable to being overweight would appear to be similar despite different definitions of obesity in Asian populations. The limited number of studies from USA that allow inter-ethnic comparisons indicate that the association between BMI and OSA exists in African Americans and that the effect is of a similar magnitude to that in Caucasians.27,29

Overall, there are ethnic differences in obesity prevalence and fat distribution.78 Obesity and particularly extreme obesity in the USA tends to be over-represented in African Americans.79,80 Prevalence of obesity in Asian countries, although still less than in Western countries, is also rising and corresponds to large numbers of affected people.76,80–82 Patterns of body fat distribution also vary with ethnicity, with Asian populations more prone to developing central or abdominal fat.77,83,84 This central adiposity distribution, which predisposes to OSA, occurs at a lower overall BMI in these populations. Although different ethnicities appear to be equally vulnerable to OSA through excess weight, there may be specific ethnic variations in obesity and fat distribution patterns, and hence the underlying anatomical mechanisms that predispose to OSA. In this way, obesity and other anatomical risk factors like craniofacial morphology may be more or less relevant for OSA risk in different ethnicities.

Craniofacial morphology

Assessments of craniofacial morphology in OSA have now been conducted within other ethnic populations besides Caucasians, primarily Chinese and Japanese studies. Intra-ethnic study designs are inherently problematic to extrapolate inter-ethnic differences in OSA craniofacial morphology, as there are issues around variation in measurements, definitions of OSA, and ethnicity and control groups. However, studies within Asian populations tend to identify similar features associated with craniofacial morphology in OSA as Caucasian studies. Predominantly noted are the inferior positioning of the hyoid bone,56,63,85–87 retropositioning of the mandible,85–87 a smaller cranial base56,66,88 and an increase in the craniocervical extension angle.56,66

Existing studies that have made direct inter-ethnic comparisons are somewhat more limited in number. However, data from such studies do suggest that there is ethnic variation in the craniofacial characteristics associated with OSA. These comparisons have generally been made between Caucasian OSA subjects and other ethnicities, particularly Asians,89 Polynesians,90 Hispanics91 and African Americans.92

Craniofacial comparisons between Caucasian and African Americans in OSA have been made in the Cleveland Family Study. Anthropometric assessment of cranial form found that brachycephaly (head shape with wider lateral and shorter anterior–posterior dimensions) was associated with a higher AHI in Caucasians but not in African Americans.92 Cephalometric analyses showed that both craniofacial hard and soft (tongue, soft palate) tissue measurements were associated with OSA in Caucasians.27 However, African Americans with OSA predominantly displayed enlargement of soft tissues and not skeletal abnormalities. This suggests that OSA craniofacial phenotypes differ between these two ethnicities, with upper airway soft tissues largely predisposing to OSA in African Americans but a contribution of both skeletal and soft tissue characteristics in Caucasians.

In comparison with Caucasian subjects, Asians appear to show greater restriction in skeletal measurements, including a smaller maxilla, smaller and retropositioned mandible, and a shorter, steeper anterior cranial base.89,93,94 The soft tissue differences between Caucasian and Asian OSA subjects appeared to be minimal, although hyoid position was reported to be more inferior in Caucasian subjects.89,93 Overall, Asian subjects appear to have more severe OSA at lower levels of obesity, suggesting that craniofacial factors and particularly skeletal restriction may be more important for OSA risk in Asians.93,94

Relative contribution of obesity and craniofacial factors

It is likely that the balance and interaction between obesity and craniofacial risk factors are strongly influenced by ethnicity. Anatomical imbalance predisposing to upper airway collapse is likely to be driven by different components (soft tissue vs skeletal) depending on ethnicity. Asian patients with OSA are reported to have more severe disease but be less obese, and have more craniofacial skeletal restriction.89 In a direct inter-ethnic comparison study, we have shown that Chinese patients display more craniofacial restrictions compared with Caucasian OSA patients, after matching for OSA severity.94 In particular, Chinese OSA patients showed shorter measures of cranial base, midface length, maxillary length and mandibular corpus length. These measures of craniofacial bony restriction were still apparent when controlled for height (Fig. 1a). In contrast, Caucasian patients showed greater obesity (Fig. 1b), being more overweight with greater neck circumference. However, when comparing the ratio of BMI with height-adjusted mandible or maxilla size, there was no difference between Caucasians and OSA severity-matched Chinese patients (Fig. 1c). The similar ratios of obesity to mandibular and maxillary dimensions suggest that the degree of anatomical imbalance that predisposes to upper airway obstruction may be similar between Caucasian and Chinese OSA patients. However, the relative contributions of obesity versus craniofacial bony restriction appear to be different between these ethnicities (Fig. 2a). The smaller craniofacial bony enclosure identified in Chinese OSA patients may make them particularly vulnerable to obesity by further exacerbating anatomical imbalance (Fig. 2b). Therefore, weight gain would have a more negative impact in this population and lead to worse outcomes in terms of OSA severity.

Figure 1.

Comparison of severity-matched Caucasian and Chinese OSA patients (n = 52, Mean ± SEM).94 (a) Height-adjusted maxillary and mandibular dimensions from cephalometry. Chinese patients display significantly more craniofacial skeletal restriction than Caucasians at the same level of OSA severity. (b) BMI. Caucasian patients were significantly more obese than their Chinese counterparts. (c) The ratio of obesity to mandibular and maxillary dimensions between Chinese and Caucasian OSA patients were not significantly different. (inline image) Caucasian and (inline image) Chinese. NS, not significant.

Figure 2.

Anatomical balance in OSA. The balance between the relative size of the craniofacial bony compartment and degree of obesity is an important determinant of upper airway obstruction. (a) Influence of ethnicity on anatomical balance. Degree of anatomical imbalance between Caucasian and Chinese OSA patients may be equivalent; however, the relative portions of obesity versus craniofacial bony restriction contributing to anatomic imbalance appear to be different between these ethnicities. (b) In this scenario, Chinese OSA patients with a restricted bony compartment may be more vulnerable to weight gain, as excess obesity will further exacerbate anatomical imbalance.

In another inter-ethnic comparison between Caucasians and African Americans, anatomical imbalance or a higher ratio of tongue area to intermaxillary length was found in Caucasians with OSA, suggesting a larger tongue for a defined skeletal enclosure area as a contributory factor to OSA. However, African Americans showed no association of the craniofacial soft-to-hard tissue ratio with OSA and only displayed increased soft tissue dimensions.27 Three-dimensional analysis of upper airway soft tissues have identified greater tissue volumes (tongue, soft palate, lateral pharyngeal walls) in OSA patients compared with controls.6 Similar analysis in a Japanese sample, however, did not find enlarged soft tissues compared with control patients, although mandibular dimensions were restricted in OSA patients.46 Therefore, there may be differential contributions of upper airway soft tissues, craniofacial bony dimensions and obesity to OSA between ethnic groups. Furthermore, particular ethnicities may be more or less vulnerable to changes in this relationship, such as through weight gain, based on their anatomical substrate. More studies are needed to understand the complexity and interaction of OSA craniofacial phenotypes with obesity, and furthermore, the impact of ethnicity on these relationships.


Overall, the limited studies available do provide evidence for ethnic differences in craniofacial morphology in OSA, but their relatively small number highlights the need for larger and broader studies. Although studies of craniofacial morphology are required to further understand the influence of ethnicity on OSA anatomical risk factors, there are methodological challenges in addressing this issue. Detailed craniofacial characteristics can be investigated using quantitative methods for various modalities of craniofacial and upper airway imaging. Different imaging modalities have unique strengths and limitations in their ability to comprehensively assess bony and/or soft tissue structures, and also applicability to large-scale studies.

Our current knowledge of craniofacial morphology in OSA has predominantly been derived from studies using lateral cephalometry. Cephalograms are standardized lateral radiographs of the head and neck, and are primarily useful for measuring skeletal structures, although upper airway and soft tissue areas can also be assessed to a certain extent. Cephalometry has been used extensively, as it is widely available, comparatively inexpensive and relatively safe albeit with some radiation exposure. However, it is limited by the two-dimensional nature of the image produced, which precludes more detailed volumetric analyses and limited resolution of soft tissue structures.

More detailed craniofacial assessments are possible using three-dimensional imaging modalities. CT scanning provides more detailed cross-sectional images, especially suited in capturing bony and upper airway features. Three-dimensional volumetric reconstructions and analyses can be performed, although at a greater radiation cost and expense than cephalometry. MRI has been used for detailed assessment of upper airway and surrounding soft tissues in OSA research.6 MRI allows three-dimensional reconstruction of specific soft tissues, and although bony resolution is not as good, craniofacial skeletal assessments have been examined using detailed analysis techniques.45,46,73 MRI is also useful for assessing upper airway fat, such as the parapharyngeal fat pads and regional facial fat, which may be relevant to OSA.41 However, in-depth assessments using MRI are limited by cost and availability.

Although these techniques are able to comprehensively assess craniofacial morphology in OSA and would be invaluable for assessing craniofacial differences between ethnicities, there are definite limitations in applying them to large-scale and multicentre studies. MRI and CT have feasibility issues regarding cost and availability as well as standardization issues between centres, and CT has the additional consideration of radiation exposure. A recently developed method using photogrammetry for quantitative craniofacial assessment in OSA95 may be a potential solution to the need for a simplified imaging technique for detailing OSA craniofacial risk factors. The technique has been used to demonstrate phenotypic differences between patients with OSA and controls across all craniofacial regions, including the face, mandible, maxilla, eyes, nose, head and neck.72 Furthermore, relationships between surface facial measurements and upper airway soft tissue structures suggest that facial dimensions may provide phenotypic information about upper airway soft tissues.73 Surface facial dimensions were found to relate to upper airway soft tissue volumes, particularly tongue volume. A combination of facial measurements, including midface width, was a better indicator of tongue volume (r2 = 0.69, P < 0.001) than BMI (r2 = 0.11, P = 0.005) in OSA patients. Both bony and soft tissue measures of midface width correlated with tongue volume. Surface facial measures are, therefore, likely to provide a composite picture of skeletal, upper airway and adiposity characteristics. The simplicity of this facial photogrammetry technique makes it particularly suited to large-scale phenotyping and genetic studies.


Understanding craniofacial features in OSA is important not only in the recognition of the disorder but also for treatment strategies. As evidence suggests, there are likely to be ethnic differences in obesity and craniofacial morphology as risk factors for OSA, hence understanding ethnicity-specific craniofacial risk factors is particularly important in the context of recognition and diagnosis strategies for OSA.

Gold standard diagnosis of OSA by overnight polysomnography is problematic due to the costly and limited nature of this resource, and consequently, the majority OSA is left undiagnosed and untreated. Simpler recognition strategies are, therefore, needed in the diagnostic clinical pathway for OSA, and much work has been done on developing clinical methods for prediction of OSA risk. Combinations of various parameters, such as age, BMI and/or neck circumference, in addition to patient symptoms, have been shown to be useful in OSA risk assessment.96–98 In addition to assessment of obesity, identification of craniofacial factors is an important component in clinical assessment of OSA. Quantitative methods using craniofacial and upper airway imaging in routine screening for OSA may not be feasible using techniques generally applied for detailed craniofacial assessment. However, our recent work using simple facial photographic analysis technique, which may lend itself more to clinical application, is able to detect morphological differences between patients with OSA and controls,95 and furthermore, these photographic craniofacial measurements were also useful in the risk stratification of OSA.72 However, facial photogrammetry to assess craniofacial characteristics has only been applied in Caucasians with OSA. Due to the probable differences in and relative contributions of skeletal characteristics and obesity in OSA between ethnicities, this technique may be more or less useful for stratifying OSA risk in different ethnic groups. Simple clinical OSA recognition techniques, such as photographic craniofacial phenotyping, have a promise, but more work needs to be done on understanding the influence of ethnicity to make it widely applicable.

Craniofacial characteristics are also relevant for OSA treatment strategies directed at altering dentofacial structures to reduce upper airway collapse during sleep. Oral appliances are a treatment alternative to CPAP often favoured by patients. Most commonly used is a mandibular advancement splint to hold the jaw in a forward position during sleep. The protrusion of the jaw in this manner mechanically alters upper airway structure, particularly by increasing the lateral diameter of the velopharynx such that the propensity to collapse is mitigated.99,100 Although oral appliances are able to improve or resolve OSA in a majority of patients, prospective prediction of treatment outcome would facilitate their clinical use. There are anthropometric and cephalometric parameters101 that have been related to positive response to oral appliance therapy, but further research is needed to define the factors that allow appropriate patient selection for this form of therapy. Craniofacial structure is an influential factor in treatment response to oral appliances, and may shape the upper airway structural and functional response to an oral appliance. As ethnicity has an influence of craniofacial risk factors for OSA, the specific characteristics and relationship to oral appliance treatment success, and the mechanisms by which this is achieved are likely to differ with ethnicity. As predictive models of oral appliance treatment often incorporate craniofacial cephalometric characteristics,102 ethnicity may be an important consideration when evaluating their performance in other populations.

Similarly, outcomes of surgical treatment approaches that aim to favourably alter upper airway bony (e.g. maxillary-mandibular advancement) or soft tissue (e.g. uvulopalatopharyngoplasty) structures are related to craniofacial morphology. While a key component of selecting patients for surgical treatment is to determine the site of pharyngeal collapse using a combination of cephalometry and nasopharyngoscopy, further research is required to define the patient phenotypes that are best suited to this treatment modality, and this will include consideration of ethnicity-specific craniofacial risk factors. This form of non-reversible treatment may be more suitable to certain craniofacial phenotypes.

Because of the apparent interaction between craniofacial morphology and obesity in determining upper airway obstruction, consideration of these aspects together may be used to predict outcomes from other treatment approaches such as weight loss. Craniofacial phenotyping and fat distribution may be able to indicate which patients are likely to have a significant response in terms of sleep apnoea severity with weight loss. Both these anatomical risk factors show ethnic influence, and therefore must be undertaken in consideration of ethnicity. Hence, craniofacial phenotyping would be fundamental aspect of simple OSA detection methods and the selection of patients for various treatment strategies. However, accurate phenotyping of craniofacial form and understanding of the influence of ethnicity would seem to be a vital element for the advancement of such approaches.


Anatomical characteristics, such as obesity and craniofacial morphology, are important risk factors for the development of OSA. Skeletal restriction, enlarged upper airway soft tissues and increased adiposity, separately and in combination, have the potential to influence upper airway patency and the predisposition to upper airway collapse. Ethnicity, through genetic and environmental influences, is likely to differentially affect one of more of these factors, leading to differences in the pathogenic mechanisms of OSA. Evidence suggests that there are ethnic differences in OSA craniofacial phenotypes; however, direct inter-ethnic comparative studies are limited, and larger and broader studies are needed. Understanding of ethnic differences in OSA craniofacial phenotypes is important in the recognition, treatment and prevention of OSA.