This study investigated correlations between airway size and age, sex, and skeletal patterns; identified airway change trends; and measured volumetric norms in children via cone-beam computed tomography.
Four hundred twenty nontreated white patients were stratified by age, sex, and anteroposterior skeletal pattern. The nasopharyngeal airway (NPA), oropharyngeal airway (OPA), and total pharyngeal airway (TPA) volumes were delineated on 3-dimensional digital cephalograms. SPSS (SPSS, Chicago, Ill) was used to run an analysis of variance and post-hoc analyses.
The Class III group had significantly larger OPA volumes than Class I and II groups. Male subjects had considerably larger NPA volumes than female subjects. Age was significantly associated with all 3 airway volumes ( P <0.05). The young cohort (ages, 9-10 years) had a mean TPA of 11,435.34 ± 484.45 mm 3 , the middle cohort (ages, 11-13 years) had a mean TPA of 14,152.07 ± 395.46 mm 3 , and the older cohort (ages, 14-15 years) had a mean TPA of 18,057.99 ± 484.25 mm 3 .
An effect of skeletal classification on OPA and a sex effect on NPA were observed. The annual change in the mean of TPA volume decreased in subjects aged 10-12 years, then rebounded until 14 years. TPA change peaked in female subjects 1 year before male subjects.
Pharyngeal airway volumes of 420 untreated white patients were examined.
There was a sex effect on the nasopharyngeal airway.
The Class III group showed significantly larger oropharyngeal airway.
No interaction effects were found among age, sex, and skeletal pattern.
Change of the total airway coincided with the onset of the pubertal growth spurt.
Abnormal breathing has been linked with adverse growth of the craniofacial complex. It has been posited that airway impairment can lead to mouth breathing and may contribute to the development of the adenoid facies phenotype, a term initially coined by Tomes. Predisposing risk factors to the increase in airway resistance include nasal obstruction, adenotonsillar hypertrophy, and allergic rhinitis. , As a result, considerable anomalies can arise in children, such as proclined maxillary incisors, increased lower anterior facial height, and narrow maxilla. Obstructive sleep apnea (OSA), one of the more serious complications, has been associated with neurocognitive impairments such as attention deficit hyperactivity disorder in children and systemic hypertension in adults. The obstruction can occur at single or multiple levels along the upper airway, from the tip of the nose to the larynx.
Traditionally, cephalograms have been used to identify airway obstruction, adenoid hypertrophy, and very constricted airways. However, the cephalogram is an image with incomplete information because the axial plane cannot be visualized. The recent development of cone-beam computed tomography (CBCT) has improved our capability of evaluating individual airways with less radiation than a medical computed tomography scan at a lower cost. Previous 3-dimensional (3D) studies have attempted to elucidate the normative values of airway volumes in children of different ages. , However, these studies lacked an equal distribution of ages among groups and consideration for other areas of the airway. ,
Some authors argued that the minimum cross-sectional area (MCA) has a high positive correlation with volume. However, volume and MCA measurements were found to have limited diagnostic value, showing poor sensitivity and specificity for assessing upper airway obstruction related to adenoid hypertrophy compared with nasoendoscopy. In addition, a recent systematic review found that the majority of the studies limited their airway assessment to intraexaminer reliability and did not consider interexaminer reliability.
The primary objective for using 3D pharyngeal airway measurements was to investigate any significant relationships between airway size and age, sex, and skeletal patterns. The secondary objective was to define the normal changes in airway volumes across different years in childhood. The tertiary objective was to define volumetric norms in children.
Material and methods
The present retrospective study was approved by the research ethics board at the University of Detroit Mercy. All patients signed an informed consent form allowing the use of their data for scientific purposes. CBCT scans used in the present study were part of diagnostic records collected to assess orthodontic treatment needs, and no patient was contacted, and no CBCTs were taken for the present study. All patient records between July 2006 and November 2017 from the University of Detroit Mercy Orthodontic Clinic were screened for study inclusion in this cross-sectional retrospective study. Eligibility criteria are listed in Table I . Patients were stratified on the basis of age, sex, and skeletal pattern. Skeletal pattern types were characterized by ANB angle: Class I (1° < ANB < 5°), Class II (ANB ≥5°), and Class III (ANB ≤1°). Because of the low proportion of patients with Class III skeletal malocclusion in our sample pool, this group constrained the total number of patients included in this study. Equal numbers of subjects were selected from each subgroup by using a random number generator, with the Class III group acting as sample size limiter. This selection process formed our final sample size of 420. The overall patient selection process is described in Figure 1 .
|Between the ages of 9 y and 15 y||Facial neoplasms|
|Visibly enlarged tonsils noted in oral examination records|
|Previous orthodontic treatment|
|Noticeable head extension and flexion|
Patients had a CBCT scan taken with an i-CAT unit (Imaging Sciences International, Hatfield, Pa), as part of their orthodontic examination and diagnostic record. Each image was acquired using the following settings: 120 kV; 5 mA; 0.3-mm voxel size; scan time, 8.9-seconds; and field of view no more than 16 cm in height × 23 cm in depth. The scans were stored in a Diagnostic Imaging and Communications in Medicine format file and loaded into Dolphin 3D (Dolphin Imaging and Management Solutions, Chatsworth, Calif) for viewing and screening. Each digitized i-CAT image was oriented parallel to the Frankfurt horizontal plane (FHP) ( Fig 2 ) for analysis of airway and cephalometric measurements. The airway sensitivity setting—which controls the program’s ability to find differences in gray scale resolution—was standardized at 45 to best recognize the airway. The oropharyngeal airway volume (OPA) was defined as the volume of the pharynx between a line parallel to FHP passing through the posterior nasal spine and another line parallel to FHP at the level of the tip of the epiglottis ( Fig 3 ). The nasopharyngeal airway volume (NPA) was calculated to be the volume between the superior limit of the OPA, a line from posterior nasal spine perpendicular to the upper limit of the OPA extending superiorly to intersect the posterior wall and the outline of the pharynx ( Fig 4 ). The total pharyngeal airway volume (TPA) was calculated as the sum of the volumes of NPA and OPA. Craniofacial features of the patients were examined by 3 angular measurements (SNA, SNB, and ANB). Three linear measurements denoting the skeletal bony dimensions were maxillary length (MxL)/Ar-A, mandibular corpus length (MnL)/Go-Pg, and mandibular ramus height (RH)/Ar-Go. MCA yielded low interrater and intrarater reliability, particularly with respect to MCA locations: kappa of 0.11 for examiner 1 (L.C.), kappa of 0.38 for examiner 2 (H.K.), and kappa values of 0.09 and 0.28 for interrater reliability. MCA was therefore dropped from this study.
Two examiners (orthodontist 1, L.C.; orthodontist 2, H.K.) measured and analyzed all airway and cephalometric tracings. To assess intrainvestigator and interinvestigator errors, we randomly selected 3D images (n = 25) that were traced and compared at baseline and then remeasured 2 weeks after the first evaluations. Pearson correlation ( r ) was used for intra- and interreliability tests. Statistical analyses of NPA, OPA, and TPA were calculated with SPSS software (version 13.0, SPSS, Chicago, Ill). Descriptive statistics were analyzed initially to establish sample means and standard deviations. To establish and quantify relationships between the calculated data, we performed an analysis of variance (ANOVA) analyses with significant P values set at 0.05 for the evaluation of interactions between age, sex, and skeletal classes. The post-hoc test (Fisher least significant difference) was conducted after ANOVA to compare each group with each other and to identify if significant differences occurred. Data was then consolidated to determine the norms of the pharyngeal airway volume according to age group. For this study, ANOVA with 60 patients in 7 age groups yields statistical power greater than 0.80 at the 5% significance level; for moderate effect sizes (0.25), the statistical power was 0.98. For a full factorial model of age × sex × skeletal class (7 × 2 × 3), the study sample of 420 patients has a power of 0.99 for large effect sizes (0.40) at the 5% significance level, and just under 0.80 or 0.72 for moderate effect sizes (0.25).
Intrarater reliability was extremely high, with the first examiner (L.C.) ranging from 0.94 to 0.99 and the second examiner (H.K.) ranging from 0.91 to 0.99, with RH being the outlier at 0.91. Interrater reliability was highest for SNA, SNB, and ANB, with r values consistently at 0.99. NPA, OPA, and TPA ranged from 0.94 to 0.99, with MxL, MnL, and RH ranking slightly lower at 0.93 to 0.99. Skeletal characteristics of the subjects were presented in Table II . The mean NPA, OPA, and TPA volumes for the entire sample of patients were shown in Figure 5 . The comparisons of NPA, OPA, and TPA volumes accounting for age, sex, and malocclusion type are shown in Table III .
|Age, y||Skeletal pattern||Sex||SNA||SNB||ANB||MxL||MnL||RH|