Craniofacial morphology/phenotypes influence on mandibular range of movement in the design of a mandibular advancement device

The mandibular opening path movements have different directions according to the craniofacial morphology of the patient but always downward and backward, therefore increasing the collapse of the upper airway. The aim of this work is to determine if there is a relationship between the craniofacial morphology and the mandibular movement to help understand the impact on the mandibular position. 52 students with full permanent dentition aged 19 to 23 years (mean 21.3 SD 1.7; 29 females and 23 males), participated in the study. Each subject had a lateral cephalometric radiograph taken. The opening angle was determined for two levels of vertical openings at 5 and 10 mm. The opening angle showed a greater variability between subjects ranging from 63.15 to 77.08 for 5 mm angle and from for 61.65 to 75.72 for the 10 mm angle. Differences of facial phenotypes was evident when comparing the individual dissoccluding angle of the low angle horizontal pattern and high angle vertical pattern. The opening angle is related to craniofacial morphology with higher vertical anterior and shorter anteroposterior faces having a more horizontal path of mandibular movement than shorter vertical anterior and longer anteroposterior subjects who have a more vertical path.


Background
Obstructive sleep apnea (OSA) is an important health problem, which has the presence of repeated episodes of a partial (hypopnea) or complete (apnea) collapse of the upper airway [1]. The consequences are fragmentation of sleep structure, a decrease in oxygen saturation and an increase in blood pressure [2]. The arousals and the nocturnal hypoxemia can lead to excessive daytime sleepiness, loss of concentration and hypertension [3,4].
One treatment option for OSA is the use of a mandibular advancement device (MAD) [1,5]. MADs keep the mandible in a protruded position during sleep increasing the width of the airway and reducing its collapsibility [6]. Even though there are conflicting reports on the success rate of these appliances [1,7], MADs have been reported to be an alternative treatment to CPAP in moderate to severe OSA cases [6,8].
The human jaw/mandible has a specific motion that depends on the temporomandibular joints, which are the most sophisticated joints in the human body with three rotational and three translational degrees of freedom. These movements can be followed by observing the lower central incisor. The border movements of the incisal edge of the lower incisor describe an area or envelope called border movement area or Posselt diagram [9,10]. Within this area, we can find all the possible positions of this structure in the mandible. Little attention has been given to the large variation of craniofacial morphology and its impact on the mandibular motion characteristics (magnitude and direction) [9,11,12] and in particular to the use of MAD [13]. The mandible always retrudes by postero-rotation during aperture, but depending on the craniofacial morphology, it will present different degrees of retrusion [9]. The mandible in a more retruded position in a sleep apnea patient might increase the severity of the apnea [14].
It is important to know the position of the mandible at two different moments with MAD: Starting Position (SP) and mandibular movements when the patient is asleep. A MAD is constructed with a SP placing the mandible with a determined range of advancement (50-75% of maximum protrusion) and vertical openings (2-6 mm) [15]. After the MAD is in place at the SP, we must consider the mandibular movements allowed by the appliance; lateral, anteroposterior and vertical (opening). The opening path movements have different directions according to the craniofacial morphology of the patient [9] but they are always downward and backward, therefore increasing the collapse of the upper airway [16]. All these movements can be visualized within the border movement area in order to ensure the desired advance of the mandible with the MAD. Thus, the aim of this study is to determine if there is a relationship between the different craniofacial morphologies/phenotypes and mandibular movement.

Methods
Fifty-two dentistry students, at the Universidad Alfonso X Madrid, 29 females and 23 males, 19 to 23 years old (mean age 21.3 SD 1.7) agreed to participate in this study. Sample size was selected according to a similar previous study [9]. All subjects were asymptomatic for temporomandibular disorders, according to the Research Diagnostic Criteria/ Temporomandibular Disorders RDC/ TMD, RDC /TMD [17]. The ethical review board of Universidad Alfonso X Madrid UAX approved this study UAX-2016-021. All patients had full permanent dentition up to the second molar with no previous maxillofacial surgery or TMJ symptomology.
Lateral cephalometric radiograph were done for each participant at the start of the study with profile and frontal extraoral photos. The Kinovea software (Kinovea, France) was used to trace landmarks on the lateral cephalograms. For each landmark, X and Y coordinates were determined with the reference at the posterior nasal spine (PNS) and using the plane anterior nasal spine (ANS) to PNS. Angles and distances were measured and placed in an excel worksheet. The landmarks with their respective definitions are listed in Table 1. Figure 1 illustrates a lateral cephalometric radiograph with landmarks. Distances and angles measured on each radiograph are listed in Table 2. Radiographs and measurement of maximum retrusión and protrusion were taken with participants' head with Frankfurt plane parallel to the floor.
When taking measurements patients were asked to sit straight in the dental chair. The absolute range of maximal mandibular protrusion and retrusion was measured (in mm) using the George Gauge (Great Lakes Orthodontics, Ltd., New York, USA) with a 2 mm interincisal vertical opening bite fork [13]. The principal investigator asked the patient to protrude and retrude for three times and took measurements of the maximum protrusion and maximum retrusion (Fig. 2). A simplified kinematic border movement model of the mandible in the sagittal plane was used to determine the Posselt diagram. Using the cephalometric radiograph a simplification of the Posselt diagram was calculated [18].
The protrusion upper border was determined on the radiograph with the maximum retrusion ( Fig. 2 point 1) and maximum protrusion ( Fig. 2 point 2) measurements taken with the George Gauge. Once the position of the incisors is found, the condyle position at maximum advancement and retrusion is calculated considering its initial position in the radiograph and the morphology of the glenoid fossa (Fig. 3).
The next step was to calculate the curved posterior border in the first phase of the condyle rotation up to 25 mm opening ( Fig. 2 points 1-3) [10]. The rotational movement of the mandible in its hinge in the center of the condyles is represented as an arch. The center of the arch is located in the condyle (landmark Co) and having the curve pass through the lower incisor through the line that connects the condyle and the lower incisor (Co-II).
Once landmark 3 ( Fig. 2 point 3) is reached, the condyle moves through the glenoid fossa until the lower central incisors reach the maximum opening of 50 mm ( Fig. 2 point 4).
The anterior curve of the diagram was drawn by an arch passing by points 2 and 4 ( Fig. 2) with the radius at the condyle at maximum advancement. Figure 2 shows the resulting Posselt diagram where the shadowed area represents the region where the central lower incisors can be placed.
Once the diagrams were obtained, the disocclusion angle was calculated. (Fig. 4) The disocclusion angle has its vertex at the border of the lower central incisors and one side is parallel to the occlusal plane. The other side is formed from the border of the central lower incisors to a point that crosses the rotational curve of the mandible at the 5 mm opening. The same was done for the 10 mm opening. The disocclusion angle allows us to compare/ obtain the differences in the direction of the mandibular movement between subjects at the curved posterior border in the first phase of condyle rotation of the Posselt diagram (Fig. 5).
To calculate reliability and measurement error of the structures and measurements obtained, ten radiographs were randomly selected and landmarked three times leaving one week in between trials. Interclass Correlation Coefficient was used to calculate the reliability as well as the measurement error for each landmark in each coordinate ( Table 3). The statistical analysis included descriptive statistics, ANOVA and paired t-test using SPSS (version 24, IBM, Armonk, NY, USA). Paired sample test  and Pearson correlation were used to compare findings with respect to each disocclusion angle of 5 and 10 mm (more significant results shown in Table 4). Descriptive Statistics and Multiple Comparisons analysis was used to verify difference between the three different groups (hypodivergent, normodivergent and hyperdivergent). (Table 5).

Results
Reliability of cephalometric landmarks was measured on 10 random patients for three times for both x and y axis ( Table 3). All landmarks presented excellent reliability values with the lowest landmark being 0.97 (CI 95% 0.94; 0.99) in the A point y-axis. It is worth noting the landmark PNS presented very low reliability but when reviewing the measurement error, these were on average 0.04 mm in both axes. The highest measurement error was found in landmark Me with 1.94 mm ± 1.05 in the x-axis. The majority of landmarks had an average of < 1 mm error in all coordinates. Once reliability was determined, all the cephalograms were landmarked and measured. The cephalometric analysis principal vertical and horizontal measurements are shown in Table 4, and the Descriptive Statistics and Multiple Comparisons analysis between the three different groups (hypodivergent, normodivergent and hyperdivergent) in Table 5.
The range of the mandibular disoccluding angle or opening angle was determined for the two degrees of vertical openings at 5 and 10 mm. The disoccluding angle showed a greater interindividual variability from 63.15 to 77.08 with a mean value of 70.42 and standard deviation 3.06 for the angle at 5 mm and for 61.65 to 75.72 with a mean value of 68.90 and standard deviation 3.07 for the angle at 10 mm. The amount of retrusion for a 63º angle with an opening of 5 mm is 2.55 mm and for an angle of 77º is 1.15 mm. The same happens when the opening is 10 mm where the retrusion with an angle of 75º is 2.68 mm and with 61º is 5.54 mm. The opening angle was similar (p > 0.05) between males (70.04) and females (70.71) for the angle at 5 mm and males (68.55) and females (69. 16) for the angle at 10 mm. The sample was grouped based on their measurements into hypodivergent (disoccluding angle < 68; 10 subjects), normodivergent (68 ≤ and < 71; 25 subjects) and hyperdivergent (≥ 71; 17 subjects).
The value of normodivergent, from 68 to 71 for the disoccluding angle, was taken and a retrusion of 2 mm (1.80-2.2 mm) at a vertical increase of 5 mm was considered normal. The disoccluding angle showed a high Fig. 1 Cephalometric landmarks. Shows cephalometric radiography with landmarks as described in Table 1

Table 2 Measurements
Detailed description of the measurements used in the study.  significant correlation (p < 0.001) with ANB angle, Frankfort-Maxillary occlusal angle and Bmi-Maxillary occlusal angle. It was found that shorter mandibles and condyles in a higher position had smaller angle values and therefore a greater horizontal opening pattern. (Table 5). Differences of facial morphologies/phenotypes were evident when analyzing the individual opening angle of the high horizontal pattern and the low vertical pattern. (Fig. 5).

Discussion
The physiological position of the mandible during sleep is slightly opened (1-5) mm and in patients with OSA is opened more than 5 mm and can reach up to 10 or 15 mm [19,20]. The opening of the mandible induces mandibular retrusion, which is associated with an increase in collapsibility of the upper airway [21] and the reduction of the efficacy of the MAD [22]. Following this though, this study wanted to analyze the different opening movement paths and their impact on the mandibular position according to the craniofacial morphology of the patient.
During normal sleep, the mouth is in a position known as the mandibular rest position or freeway space. The freeway space is described as the space between the maxillary and mandibular occlusal surfaces when the mandible is in the rest position and should be 1-5 mm [23], an opening of up to 5 mm for 88.9% of total sleep time [19]. Increasing oscillating lowering movements of the mandible in response to the airway collapse during obstructive apnea have been described [24][25][26]. In patients with OSA, the mouth opening is greater than 5 mm for 69.3% of total sleep time [26]. A common pattern characterized by a gradual opening followed by a quick closure of the mouth, generally after an arousal, has been described in normal patients and patients with OSA [19,24,26]. For this reason the disoccluding angle of 5 mm that measures the normal opening in healthy patients and 10 mm that measures the opening in patients with OSA were obtained. The mandibular position and related structures are influenced by and participate in patency of the pharynx and the complex mechanisms that lead to obstruction of the upper airway [16]. Mandible opening   [27] and is associated with a reduced cross-sectional area of the lumen [16], reduced mechanical efficiency of the pharyngeal dilator muscles [24] and increased resistance and collapsibility of the upper airway [14,21,28]. All of which may contribute to sleep-related breathing abnormalities.
The findings of the present study show/suggest that the slope or angle of the mandibular movements are related to the craniofacial morphology with higher vertical anterior and shorter anteroposterior faces with a more horizontal path of mandibular movements than shorter vertical anterior and longer anteroposterior subjects who have a more vertical path (Fig. 5). For an opening of 5 mm with an angle of 77º, the mandible retrudes 1.15 mm and with an angle of 63º is 2.55 mm. This is double for the same amount of opening at 5 mm. A similar scenario is present at the 10 mm opening where the retrusion at an angle of 75º is of 2.68 mm and at 61º is 5.54 mm which is close to a 3 mm difference.
Our results are similar to L'Estrange et al. [29] where they found a smaller effect on the oropharynx in subjects who had a reduced lower facial height. They found that in these cases, maximal mandibular protrusion had a minimal increase of vertical opening and the mandibular symphysis was further forward in relation to the posterior wall of the pharynx. The path of the mandible during aperture started in a position with the lower incisor much further forward and much closer to the cranium than in other subjects. Therefore, the mandible has a longer path to travel before it reaches the point where the airway begins to occlude. Subjects with higher vertical anterior faces and shorter anteroposteriors have an opening path beginning further backwards and downwards. As found in our study, any increase in the vertical dimension would quickly retrude the mandible [29]. MADs place the mandible in a determined anteroposterior and vertical position to improve the upper airway cross-sectional area, due to a combination of both their effect of the protrusion of the mandible, and for their capability to stabilize the mandible [30][31][32] (Fig. 6). There are different MADs available in which control of the mandibular position with respect to the potential of mouth opening, especially in the supine position, depends on the MAD design [33]. A constantly larger mouth opening during sleep will reduce the efficacy of this treatment [8,22,28]. Two-piece appliance designs, which allow uncontrolled opening of the mandible, have shown a lower response rate in positional OSA [34], while appliances with limitation of mandibular opening are more effective in decreasing AHI [35]. Therefore, as shown in recent studies [18], the devices should incorporate vertical control in the design to not allow the jaw to move backwards at any time while opening the mouth, and as suggested by our study, consider the anatomic and kinematic of the mandible. Our results enhance recent studies that have identified the importance of phenotypic characteristics on treatment response, and suggest the relevance of MAD design features considering the kinematic behavior of the mandible as part of a personalized approach to treatment [18,30].  Limitations of this study are the inclusion of young adult population and not OSA patients. The use of young adult population allows us to determine normal range of craniofacial morphologies and their movements that can be applied in future studies with OSA patients. Extrapolation to an OSA population can only be valid if these results are confirmed in a, typically older, more obese, OSA population compared to the young, healthy study population used in the current study. Another limitation is the use of lateral cephalometric radiograph for the analysis of mandibular movements. The radiograph was taken with the patient awake and in an upright position. The range of motion and the position of the mandible may be altered when the patient is asleep and should be consider when analyzing the changes in upper airway if measured with lateral cephalometric radiograph. Table 5 Descriptive statistics and multiple comparisons *Correlation is significant at the 0.05 level (2-tailed) **Correlation is significant at the 0.01 level (2-tailed) Descriptive statistics and multiple comparisons for the three opening patterns horizontal (hypodivergent), normal (normodivergent) and vertical (hyperdivergent) for the more relevant cephalometric measurements. Descriptive statistics include mean and standard deviation and multiple comparisons shows Dependent Variable Bonferoni with significance.