PDF
ePub
Citation
Print
INTRODUCTION
Obstructive sleep apnea syndrome (OSAS) is characterized by the temporary cessation of breathing (apnea) or shallow breathing (hypopnea) with decreased hemoglobin oxygen saturation.1 On the basis of the etiology, severity, and collapse site in the upper airway (UA), various treatment modalities have been applied, such as weight reduction, continuous positive airway pressure therapy, intraoral appliance therapy, soft tissue procedures, and maxillomandibular advancement (MMA) surgery.123
Although conventional rapid palatal expansion (RPE) with tooth-borne anchorage has been suggested as a treatment option for OSAS, the subjects in most previous studies were children.456 Because conventional RPE is thought to be less effective for adult patients, use of surgery-assisted RPE (SARPE) is recommended for adult patients with OSAS.789 However, it has some disadvantages, including surgical invasiveness and a long retention period. Of late, miniscrew-assisted RPE (MARPE) has been receiving attention by clinicians because of increased orthopedic effects on expansion of the maxillary basal arch width when compared with conventional RPE.10 Moreover, it induces noninvasive expansion of the midpalatal suture in adult patients, as opposed to SARPE.111213 Although MARPE is thought to be helpful for relieving the symptoms of OSAS in adult patients,1415 scientific evidence regarding its effects is limited.
Three-dimensional (3D) computational fluid dynamics (CFD) has been used to investigate airflow dynamics in patients with OSAS.31617181920 However, because UA, particularly the pharynx, is a deformable structure, adoption of fluid-structure interaction (FSI), which is a computational method to simulate flow and structural changes, has been recommended.1721 The use of FSI simulation can overcome the limitation of the rigid wall design in CFD and provide more realistic results compared with CFD.1721
Despite the important anatomical and physiological roles of the nasal cavity (NC) in respiration, most numerical airflow simulation studies have omitted this structure during the fabrication of 3D airway models because of the complex structure and diverse shape as well as the computational cost of airflow simulation.1617181920 To investigate the effects of MARPE on the airflow dynamics in UA, NC must be included in 3D-computational modeling and FSI analysis.2223 From these perspectives, we conducted the present study to investigate the effects of MARPE on airflow in UA of an adult patient with OSAS using 3D-CFD/FSI analysis.
MATERIALS AND METHODS
Geometric modeling
For the fabrication of geometric models, a male patient with OSAS aged 18 years and 7 months was selected. The patient presented with a narrow maxillary basal arch and severe crowding and refused to use an intraoral appliance or undergo MMA surgery. Accordingly, MARPE was selected as the treatment method. Before the initiation of MARPE, he underwent an in-laboratory polysomnography study (age, 18 years and 7 months; body mass index [BMI], 25.0 kg/m2; apnea and hypopnea index [AHI], 49.5 events/hour; respiratory disturbance index [RDI], 52.2 events/hour; lowest O2 saturation rate [LSR], 85%). After the completion of MARPE at S-plant Dental Hospital in Seoul, Korea, his signs and symptoms of OSAS significantly resolved (age, 19 years and 1 month; BMI, 24.9 kg/m2; AHI, 2.2 events/hour; RDI, 20.2 events/hour; LSR, 95%). This study was reviewed and approved by the Institutional Review Board of the School of Dentistry Seoul National University, Seoul, Korea (S-D20170001).
Cone-beam computed tomography (CBCT) was performed (3D eXam; Imaging Sciences International, Hatfield, PA, USA) before (T0) and 6 months after (T1) MARPE using the following parameters: volume size, 23-cm diameter × 17-cm height; resolution, 0.3 voxel; exposure conditions, 37.1 mA and 120 kVp for 17.8 seconds. Although there is no standardized method for setting the head and tongue posture during 3D image acquisition, CBCT was performed using the natural head posture and the letter “N” tongue posture in the present study. For the letter “N” tongue posture, the patient was instructed to place the tongue on the roof of the mouth, directly behind the anterior teeth without pressing against them, and purse the lips together.24
Digital imaging and communications in medicine (DICOM) data obtained at the T0 and T1 stages were exported to ICEM-CFD (ver. 15; ANSYS Inc., Canonsburg, PA, USA) to fabricate the surface model (STL file). Then, 3D computational models of UA, including NC and the pharynx, were constructed. The meshes were generated with tetrahedron and prism elements. The total number of nodes and elements in each model is summarized in Table 1.
Computational FSI simulation
FSI simulation of UA was performed under the following conditions: boundary conditions, atmospheric pressure for the inlet and a regular respiration pattern (quantity/time) for the outlet; material properties, Young's modulus of 7,540 Pa and Poisson's ratio of 0.4525; seven planes in the NC area, with an interplane distance of 10 mm from the aperture of the nostril to the beginning of the pharynx (Figure 1); and nine planes in the pharynx area, with an interplane distance of 10 mm from the beginning of the pharynx to the beginning of the larynx (Figure 1). The respiratory cycle used during simulation is described in Figure 2.
For airflow simulation, the flow was assumed to be incompressible. The governing differential equations, the Reynolds-averaged Navier–Stokes equation for turbulence with the k - ε model, and constitutive equations for structure dynamics are shown in Table 2. Constant air density (1.185 kg/m3) and viscosity (1.831 × 10−5 kg/m·s) values were assumed. No-slip wall boundary conditions were imposed on the airway walls.
The cross-sectional airway area at each plane was assessed at T0 and T1. In addition, airflow pressure and velocity, node displacement in the soft tissue of the airway wall, and the average pressure, the flow rate, and the total resistance were measured at maximum inspiration (MI), rest, and maximum expiration (ME) at both time points. Changes in the total resistance of airflow were calculated on the basis of changes in each variable from T0 to T1.
RESULTS
Changes in the cross-sectional airway area (Table 3)
In NC, the cross-sectional area at planes 1 to 3 showed a significant increase (range of increase: 0.96 cm2 to 1.16 cm2, 38.5% to 53.3%) at T1 compared with that at T0, while the area at planes 4 to 6 showed a moderate increase (0.65 cm2 to 1.04 cm2, 27.5% to 38.8%) and the area at plane 7 showed a slight increase. In the pharynx, a strong increasing trend was observed at planes 1 to 4 (1.39 cm2 to 1.69 cm2, 40.4% to 67.3%); this trend weakened at planes 5 and disappeared at planes 6 to 9. These findings imply that MARPE primarily affected the anterior and middle parts of NC, the nasopharynx (planes 1 and 2) and the upper and middle part of the oropharynx (planes 3, 4, and 5). The amount of increase in the cross-sectional airway area was larger for the pharynx than for NC, even though the pharynx was located farther than NC from the MARPE appliance.
Changes in the pressure and velocity of airflow (Tables 4 and 5, Figures 3 and 4)
At planes 1 to 7 in NC and planes 1 to 5 in the pharynx, the absolute values for airflow pressure and velocity at MI, rest, and ME were lower at T1 than at T0.
In NC, airflow pressure and velocity at MI showed a significant decrease at planes 1 to 4 and planes 1, 2, 3, and 7, respectively, after MARPE (pressure: −29.3 Pa to −10.2 Pa, −52.9% to −33.3%; velocity: −0.98 m/s to −1.44 m/s; −29.9% to −37.9%). At rest, there was a slight decrease in pressure and no significant change in velocity at all seven planes. At ME, there was a decrease in pressure at planes 1 to 7 (26.2 Pa to 2.8 Pa, −53.6% to −17.5%) and a decrease in velocity at planes 1 to 3 (−0.87 m/s to −1.54 m/s; −35.9% to −42.6%).
In the pharynx, at MI, there was a considerable decrease in pressure at planes 1 to 7 (−9.6 Pa to −10.8 Pa; −34.1% to −37.0%) and a significant decrease in pressure at planes 8 (−12.6 Pa, −66.9%) and 9 (−13.5 Pa, −67.5%). Velocity showed a significant decrease at planes 1 to 4 (−0.61 m/s to −1.01 m/s; −31.1% to −41.3%), a moderate decrease at planes 5 to 7, and no significant change at planes 8 and 9. At rest, there were no significant changes in pressure and velocity at any of the nine planes. At ME, there was a steady pressure drop at planes 1 to 7 and a substantial drop at planes 8 (9.4 Pa, −73.9%) and 9 (8.9 Pa, −68.9%). Velocity also decreased at planes 1 through 5.
Changes in node displacement in the soft tissue of the airway wall (Table 6 and Figure 5)
In both NC and the pharynx, the node displacement value showed a decreasing pattern from MI to rest and an increasing pattern from rest to ME, particularly after MARPE. However, there was no clinically meaningful change in the absolute values, regardless of the respiration stage. These findings indicate that the changes in airflow pressure and velocity (Tables 4 and 5) did not affect node displacement in both NC and the pharynx (Table 6).
Changes in the average pressure, flow rate, and total resistance (Table 7)
Considering the significant decrease in the average pressure at MI (−55.5%), rest (−49.7%), and ME (−52.6%) from T0 to T1, the total resistance also exhibited a significant decrease at MI (−55.1%), rest (−35.9%), and ME (−33.9%).
DISCUSSION
In the present study, we investigated the effects of MARPE on airflow in UA of an adult patient with OSAS using 3D-CFD/FSI analysis. The findings are discussed in detail as follows.
Changes in the cross-sectional airway area
The amount of increase in the cross-sectional airway area after MARPE was significantly larger in the anterior NC than in the posterior NC (Table 3). The increase extended below to the oropharynx (planes 1, 2, and 3) and the upper part of the laryngopharynx (planes 4 and 5; Table 3). Because of the anatomical continuity of NC and the pharynx, transverse expansion of the hard palate through MARPE can directly and indirectly increase the cross-sectional airway area at the NC and pharynx, respectively. These results are in agreement with those of previous studies that reported enlargement of the pharyngeal airway after conventional RPE in children26 and SARPE in adults.9
The amount of increase in the cross-sectional area was larger for the pharynx than for NC (Table 3). We believe this was because of the complex structure of NC, which primarily comprises hard tissues such as the nasal septum, conchae, and palatal bone. On the other hand, the pharynx is a relatively simple tubing system made of soft tissue.
Changes in airflow pressure and velocity
After MARPE, we found a pressure drop in NC and the pharynx at MI and ME. A decrease in airflow velocity existed primarily in the anterior NC at both ME and MI and the oropharynx (planes 1 to 3) and upper part of the laryngopharynx (planes 4 and 5) at MI (Tables 4 and 5, Figures 3 and 4). Although NC exhibited a smaller increase in the cross-sectional airway area compared with the pharynx (Table 3), the former exhibited a greater decrease in both pressure and velocity compared with the pharynx. This was probably because the anterior NC plays an important role as the main entrance and exit for airflow and decreases the airflow resistance in UA (Tables 4 and 5). The impact of NC expansion may extend to the middle part of the oropharynx (Tables 4 and 5, Figures 3 and 4). These results are in accordance with those of Iwasaki et al.22 However, their study modeled only NC; therefore, their data cannot be generalized to the entire process of respiration.
Changes in node displacement in the soft tissue of the airway wall
The amount of node displacement in both NC and the pharynx was too small to be considered clinically meaningful (Table 6 and Figure 5). This finding was not concordant with that in a previous FSI study by Zhao et al.,17 who reported that the amount of node displacement in the pharynx appeared to be significant. This difference in findings may have resulted from the fact that we included NC in our 3D airway models, whereas Zhao et al.17 excluded the same. Therefore, the findings from our study suggest that the airway wall in NC and the pharynx do not move during the respiratory cycle to the extent reported by Zhao et al.17 In other words, the amounts of node displacement in the pharynx turned into insignificant when the nasal cavity was included in FSI analysis.
Changes in the average pressure, flow rate, and total resistance
In the present study, MARPE significantly decreased the total resistance in UA during the entire respiratory cycle (Table 7). This finding is in accordance with that from a previous review,27 which suggested that expansion of the NC floor can benefit patients with a constricted maxillary arch and nasal airflow problem.
Because we included NC in our 3D computational model and used FSI simulation, our results may reveal more realistic effects of MARPE on changes in the airflow of adult patients with OSAS compared with previous studies.9182327 However, further studies will be needed to use CT data of OSAS patients taken in the supine position and to find more realistic values of material properties of the upper airway for better CFD/FSI model. And it should also be endeavored to investigate other treatment modalities for OSAS (e.g., intraoral appliance, soft tissue procedures, and MMA) with a large sample size.
CONCLUSION
Three-dimensional CFD/FSI analysis with an appropriate model incorporating NC and the pharynx showed an improvement in airflow and a decrease in resistance in UA after MARPE.
MARPE can serve as an alternative treatment modality for adult patients with who have moderate to severe narrow basal arch and crowding and refuse to the intraoral appliance or MMA surgery.
XML Download