The aim of this retrospective three dimensional (3D) computed tomographic analysis was to investigate the morphological airway changes in 17 obstructive sleep apnea (OSA) patients following bimaxillary rotation advancement procedures. Morphological changes of the nasal cavity and naso-, oro- and hypopharynx were analysed separately, as were the total airway changes using nine parameters of airway size and four of shape. The Wilcoxon test was used to compare airway changes and the intraclass correlation coefficient to qualify inter-observer reliability. Following bimaxillary advancement and anti-clockwise maxillary rotation, the total airway volume and the lateral dimension of the cross-sectional airway increased significantly. The total length of the airway became shorter ( p < 0.05). Remarkable changes were seen in the oropharynx: the length, volume, cross-sectional area (CSA), antero-posterior and medio-lateral distance changed ( p < 0.05). This combined with a significant 3D change in the shape of the airway from round to elliptical. The average cross-sectional oropharyngeal area was nearly doubled, the minimal CSA increased 40%, and the hyoid bone was located more anterior and superior. Inter-examiner reliabilities were high (0.89). 3D airway analysis aids the understanding of postoperative pathophysiological changes in OSA patients. The airway became shorter, more voluminous, medio-laterally wider, and more compact and elliptical.
Obstructive sleep apnea (OSA) is a disorder characterized by temporary cessation of breathing (apnea) or shallow breathing (hypopnea). The signs and symptoms are repeated apneic and hypopneic episodes during sleep, excessive daytime sleepiness, night sweats, confusion, headaches and reduced attention span. Recent studies have suggested an association between OSA and certain cardiovascular sequelae, such as hypertension and coronary artery disease.
The primary goal of a surgical approach to severe cases of OSA is to resolve or significantly improve the clinical situation, thus avoiding the use of nasal-continuous positive airway pressure (N-CPAP), which is frequently badly tolerated or refused. Empirical studies have suggested that rates for CPAP use range from 30 to 60%. Surgical techniques involving the soft tissues, such as uvulopalatopharyngoplasty, hyoid suspension, partial glossectomy and lingual suspension have given partial but not long-lasting results.
Therefore orthognathic procedures have gained ground. Initially, mandibular advancement alone and more recently, bimaxillary maxillo-mandibular advancement (MMA) and rotation advancement (RA) procedures have been employed. With MMA, all the soft-tissue structures making up the pharyngeal walls are tightened at once; this stops them from collapsing, or reduces this occurrence, by acting on the suprahyoid and palatal muscles and on the lateral musculature of the pharynx. The result is a significant increase in airway space and the resolution of the syndrome in a high (95%) percentage of cases, as well as improved quality of life. While 2D imaging, orthopantomograms and lateral cephalograms have traditionally served as the radiologic standard for airway assessment in OSA and cephalometric measurements are useful for analysing airway size in the sagital plane, they do not depict the three-dimensional (3D) airway anatomy accurately. The most physiologically relevant information is obtained from axial images, perpendicular to the direction of airflow: the axial plane is not visualized on lateral cephalograms. 3D computed tomography (CT)-studies have been used to characterize the airway anatomy and the morphological changes in patients with OSA. A systematic evaluation of OSA patients compared to non-OSA controls, using normative 3D CT upper airway parameters, has been published recently. The results indicate that the presence of OSA is associated with an increase in airway length. Airways that were more elliptical in shape and mediolaterally oriented had a decreased tendency towards obstruction. Only a small number of 3D CT studies with a limited number of patients are available evaluating the morphological changes of the airway in OSA using patients treated with orthognathic surgery procedures. The authors hypothesize that 3D CT analysis of airway size and shape in patients with OSA will provide reliable and clinically useful information to supplement or finally replace that provided by 2D cephalograms.
The purpose of the present study was to investigate the 3D morphological and pathophysiological airway changes in OSA patients following the bimaxillary rotation advancement procedure.
Materials and methods
This is a retrospective study of 17 patients (10 male and 7 female) with OSA who were treated by the third author (HFS). Patients included in this study had severe clinical OSA symptoms, and the diagnosis was confirmed by overnight polysomnogram findings. 11 patients had previously undergone other surgery, such as uvuloplasties, septoplasties, tonsillectomies and adenotomies ( Table 1 ), at different institutions with no long-lasting effect. All patients had been undergoing treatment with N-CPAP for some time but had not tolerated it well.
|No.||Age (years)||Gender||Height (cm)||Surgical history||Polysomnogram||Rotation advancement (mm)||Adjunctive surgeries|
|Pre-AHI/h||Post-AHI/h||Mandible antero-posteriorly||Maxilla counter-clockwise|
|1||56||M||180||UP, SP, TA||60||6||9.0||3.5||G, RT|
|4||29||F||190||TA, UP||36||4||11.0||3.5||G, SP|
|5||51||M||177||UP, TA||40||3||15.0||4||G, SP, RT|
|10||33||M||178||UP, T||45||9||12.8||4||G, RT|
|14||25||F||170||UP, T||39||11||10.5||4.4||G, SP|
|16||32||F||172||UP, TA||70||7||10.6||4||G, SP, RT|
|Mean||38.64 ± 10.75||173.9 ± 8.64||47.94 ± 15.64||5.64 ± 2.09||11.84 ± 1.82||4.10 ± 0.60|
The bimaxillary RA procedure was performed in all patients to create harmonic antefacial physiognomies ( Figs. 1 and 2 ). This technique represents a further modification of the MMA procedure. Following a Le Fort I osteotomy and a long sagittal splitting of the mandible, the maxilla was advanced forward and mainly rotated anti-clockwise to create a new inclination angle of the occlusional and maxillary plane. The mean antero-posterior advancement of the mandible was 11.84 ± 1.82 mm, measured perpendicular to the sagittal plane, and the mean anti-clockwise rotation of the maxilla was 4.13 ± 1.01 mm, creating a gap in the posterior part of the zygomatic buttress ( Table 1 and Fig. 3 ). The anterior advancement of the maxilla was 2.1 ± 0.53 mm measured parallel to the maxillary plane. The maxilla was fixed with two to three plates (2.0) on each side.
Following the new anti-clockwise rotated occlusional plane, the mandible could be positioned farther anteriorly, as opposed to the conventional MMA procedure, where the maxilla would be positioned too far forward (>10 mm), thus decreasing the nasolabial angle too much and resulting in an unnatural appearance.
The mandible was fixed with 2–4 bicortical screws on each side. In 13 patients an associated genioplasty was performed, following the technique described by Trauner and Obwegeser. Additional septoplasty was carried out in 8 patients and a reduction of the inferior turbinates in 11 patients ( Table 1 ). Maxillo-mandibular fixation was not necessary but patients used maxillo-mandibular bands for 3–4 weeks. Approximately 3–4 months after the RA procedure, all patients underwent an overnight polysomnogram study.
Image acquisition and analysis
Helical maxillofacial noncontrast CT scans (Evolution 6, Siemens, Munich, Germany) consisting of 2.5 mm axial tomograms, with reconstructions in the axial, coronal and sagittal planes, were used. The patients were in the supine position and were instructed to remain still, to not swallow, to place the tongue against the incisor teeth and to hold their breath at the end of exhalation. The mandible was positioned centrally and the lips were relaxed. A reference line of 50 mm was used to calibrate the measurements for each image. Postoperative scans were completed 3–6 months following RA.
The scans were imported into the analysing software ZIB-Amira ® (Zuse Institute Berlin). Digital 3D model reconstructions of the airways were made using a semiautomatic region growing method with a fixed Hounsfield threshold value ( Figs. 3–5 ). The superior upper airway boundary was defined at the level of the ethmoid cells and the inferior boundary 2 cm below the base of the epiglottis, consistent with methods described in previous studies. A reliable coordinate system consisting of the Frankfort horizontal plane or basal skull plane (BSP) and a median plane was set up ( Fig. 4 ). The boundaries of the compartments were defined parallel through the BSP, indicating the nasal cavity (compartment A) intersected inferior with the posterior nasal spine, the nasopharyngeal compartment (B) with the tip of the uvula as its inferior boundary, the oropharyngeal compartment (C) intersected inferior with the tip of the epiglottis and the hypopharyngeal compartment (D) with its inferior border 2 cm below the base of the epiglottis ( Figs. 4 and 5 ).
Once the 3D digital models were constructed, the airway was systematically analysed using metric, 2D and 3D parameters. The following airway parameters ( Tables 2 and 3 ) were used to analyse the anatomical compartments of the airway: volume (VOL), surface area (SA), length (L), CSA, antero-posterior dimension (AP), and medio-lateral dimension (LAT). The LAT/AP ratio and sphericity describes the 3D shape of the airway. The metric and morphological changes of the position of the hyoidal bone following RA surgery were described, measuring the distance of the hyoidal bone to the BSP (L hyo), as well as the distance between the hyoidal bone and the mandible (L mh) and the angle between the hyoid and the mandible ( Figs. 5 and 6 ). All parameters were measured in a reproducible way according to the BSP ( Fig. 4 ). To determine inter-observer variability, a second operator performed the analysis in 7 randomly selected patients in the study group, independent of the first operator.
|Nasal cavity||A||From ethmoid cells to posterior nasal spine|
|Nasopharynx||B||From posterior nasal spine to tip of the uvula|
|Oropharynx||C||From tip of uvula to tip of epiglottis|
|Hypopharynx||D||From tip of epiglottis to 2 cm below the base of the epiglottis|
|Length and size of the airway|
|Nasal cavity||VOL A|
|Surface area||SA||cm 2||Surface area of airway|
|Nasal cavity||SA A|
|Length||L||1D||mm||Length of the airway, measured perpendicular to the basal skull plane (BSP)|
|Nasal cavity||L A|
|Cross-sectional area||CSA||2D||mm 2||Cross sectional area of the airway, measured parallel to the basal skull plane|
|Medio-lateral distance||LAT||1D||mm||Medio-lateral distance of the cross sections of the airways, measured in the axial planes parallel to the basal skull plane|
|Antero-posterior dimension||AP||1D||mm||Antero-posterior distance of the cross sections of the airways, measured in the axial planes parallel to the basal skull plane|
|Shape of the airway|
|LAT/AP ratio||LAT/AP||Ratio||–||Ratios of the medio-lateral and antero-posterior distances, indicating the geometrical shape of the cross-sectional area of the airway in 2D|
|Sphericity ( Ψ )/compactness||Ψ||Formula||–||Mathematical measure of the sphericity ( Ψ ) or compactness|
|Total||Ψ tot||( Ψ ) = [ π 1/3(6 × VOL) 2/3 ]/SA|
|Nasopharynx||Ψ A||Indicating the 3D shape of the airway, a flat and less compact object has a Ψ of 0 and a sphere and compact object a Ψ of 1|
|Position of the hyoid bone|
|Distance of the mandible to hyoid bone||L mh||1D||mm||Distance of the genial tubercle of the mandible to the hyoid bone|
|Mandible–hyoid angle||α||Angle between the genial tubercle of the mandible and the hyoid bone|
|Length of the hyoid to the skull base||L hyo||Length of the hyoid bone to basal skull plane|