INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The obstructive sleep apnea (OSA) syndrome results from repeated occlusion of the upper airway during sleep; however, it is still unclear why an individual's airway is obstructed during sleep. Previous studies have focused on anatomic factors as predictive factors in apnea density. Compared with control subjects and snorers, cephalographic evaluation in OSA patients has shown deviations in the position of the hyoid bone, in the dimensions of the posterior airway space, and in the diameter and length of the soft palate and the tongue (1). However, there is no consistent evidence that these differences in the upper airway anatomy have a substantial impact on the incidence of OSA. A considerable overlap in airway caliber is present between normals, snorers, and OSA patients (4) and facial bone and soft tissue abnormalities are unlikely to be major factors in obese groups (5, 6). Thus, the determinants of upper airway collapse are multiple, the airway anatomy being only one such determinant.

Recently, the process of pharyngeal collapse has been explained by applying the Starling resistor concept to upper airway flow mechanisms. According to this model, the suction pressure applied to the downstream end of the collapsible tube can increase the flow only to a limited extent and the flow mechanism depends essentially on the dynamic behavior of the collapsible segment, that is, the segment between soft palate and hypopharynx. Thus the air flows through the pharynx only when the upstream pressure exceeds a critical pressure. Many methods have been introduced to measure upper airway collapsibility, the most sensitive being the pharyngeal critical pressure (Pcrit). Positive values of Pcrit differentiate OSA patients from snorers (7) and Pcrit influences the density of the sleep-related breathing disorders (8).

Although there are studies supporting the effect of neuromuscular control on pharyngeal collapsibility (9), we have no detailed data on the role of anatomic factors. Schwartz and colleagues have demonstrated in OSA patients treated by uvulopalatopharyngoplasty (UPPP) (10) a significant decrease in the apnea-hypopnea index (AHI) only in patients in whom Pcrit becomes more negative. This suggests that anatomic alterations may be a contributing factor to upper airway collapsibility.

To determine the relative influence on pharyngeal collapsibility of the anatomic factors, we measured Pcrit in a group of patients with OSA having cephalometry and we studied the possible relationship of Pcrit with some craniofacial and soft tissue variables.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Among the consecutive patients referred to the sleep laboratory from March 1997 to March 1998 for assessment of possible sleep-disordered breathing, 130 were identified from the polysomnographic study register. The exclusion criteria were AHI < 10 (17 subjects), female sex (20 subjects), previous treatment for sleep apnea with nasal continuous positive airway pressure (nCPAP) (five patients), corrective upper airway surgery or cleft palate repair (five patients), and presence of craniofacial malformation (four patients). Patients with cephalograms with an intra- or interexaminer error of more than 1 mm or 1° (10 subjects) and patients showing mask and mouth leaks (12 patients) during the titration night were also excluded. Of the original sample, 57 male Caucasian patients met the selection criteria. The subjects were divided into two groups, obese (n = 30) and nonobese (n = 27), using a body mass index (BMI)  30 kg/m² as a cutoff score. Written informed consent was obtained before entry into the study.

All patients had a full evaluation including polysomnography, blood gas analysis, respiratory function tests, and cephalometry. Neck circumference was measured in all subjects at the level of the superior border of the cricothyroid membrane in the upright position.

Cephalometric radiographs, obtained according to a method described in our previous work (3, 11), were recorded in the sitting natural head posture (mirror technique) at the end-expiration phase and without swallowing. The radiographs were traced twice by two different investigators blinded to the subject's apnea frequency and Pcrit. The mean value of measurements for each patient was used for the calculation and the variables were assessed as angular (degree) or linear (millimeters) measures. The cephalometric variables were classified into four categories: soft tissue data, hyoid bone position data, airway size data (Figure 1A), and craniofacial data (Figure 1B). The following soft tissue variables were measured: the length of the soft palate (SPl) from the posterior nasal spine of the maxilla to the lower edge of the uvula; the soft palate width (Spw); the tongue length (Tl) measured from the lowest part of the epiglottis to the tip of the tongue; and the tongue width (Tw). Hyoid position variables included the distance of the hyoid bone to the mandibular plane (H-MP); the shortest distance of the hyoid bone to the posterior pharyngeal wall (H-Ph); and the distance of the hyoid bone from the posterior nasal spine (H-Pns) representing pharyngeal elongation. Upper airway size data were defined as the superior (SPAS) and inferior (IPAS) posterior airway space measured respectively as the shortest distance from the posterior pharyngeal wall to the posterior soft palate (SPAS) and to the dorsum of the tongue (IPAS). Bony measurements (Figure 1B) included: the length of the anterior cranial base from nasion to sella (SN); the angulation of the cranial base (BaSN); the sagittal dimension of the bony pharynx (BaPns); the maxillary position (SNA); the mandibular position (SNB); the difference in prognathism (ANB); the lower face height; the total face height; the angulation of the mandibular plane with the cranial base (SN-MP); the maxillary unit length (Ans-Pns) calculated as the distance between the medial condylar point of the maxilla to the tip of the posterior nasal spine; and the mandibular unit length (Cd-Gn) measured as the distance from the medial condylar of the mandible to gnathion, i.e., the most anterior- inferior point of the mandibular symphysis.

View larger version (28K): [in this window] [in a new window]   View larger version (29K): [in this window] [in a new window]   Figure 1.   (A) Cephalometric landmarks and soft tissue, hyoid position, and airway size variables used in the study. Landmarks: Pns: posterior nasal spine; H: hyoid bone; MP: mandibular plane (tangent line from the symphysis to the inferior border of the mandibular angle). Variables: 1) SPl: the length of the soft palate; 2) Spw: the width of the soft palate; 3) Tl: the length of the base of the tongue; 4) Tw: the width of the tongue; 5) H-Ph: the distance from the hyoid bone to the posterior wall of the pharynx; 6) H-Pns: the distance from the hyoid bone to Pns; 7) H-MP: the distance from the hyoid bone to the mandibular plane; 8) SPAS: the upper posterior pharyngeal space; 9) IPAS: the lower posterior pharyngeal space. (B) Anatomic reference points and variables for craniofacial skeletal measurements. A: the most posterior midline point in the concavity between the anterior nasal spine and the lowest point on the alveolar bone overlying the maxillary incisors; B: the most posterior midline point on the anterior concavity of the mandibular symphysis; Ans: anterior nasal spine: the anterior tip of the sharp bony process of the maxilla at the lower margin of the anterior nasal opening; Ba: basion; Cd: condylion, i.e., the most posterior superior point on the curvature of the condylar head; Gn: gnation, i.e., the most anterior-inferior point on the contour of the bony chin symphysis; Me: menton, i.e., the lowest point on the symphyseal shadow of the mandible; N: nasion; S: geometric center of the pituitary fossa. Skeletal variables: 10) SNA: maxillary protrusion; 11) SNB: mandibular prognathism; 12) ANB: maxillo-mandibular discrepancy; 13) BaSN: cranial base angle; 14) SN-MP: angle of facial divergence; 15) BaPns: upper pharynx bony opening; 16) total face height; 17) lower face height; 18) Ans-Pns: sagittal dimension of the maxilla; 19) Cd-Gn: overall length of the mandible.

Polysomnographic Study

Nocturnal recording was carried out for two nights, the first without nCPAP and the second with nCPAP, as part of the diagnostic workup. Polysomnography included an electroencephalogram, electro-oculogram, and electromyogram of chin muscles for conventional sleep staging. Breathing was analyzed with a pneumotachograph attached to a face mask. Oxygen saturation was measured continuously with an ear oximeter (Biox III; Ohmeda, Boulder, CO). Esophageal pressure was measured with a 10-cm latex balloon placed in the lower third of the esophagus, inflated with 1 ml of air and connected to a pressure transducer (Validyne MP 45; Validyne, Northridge, CA). During the titration night the nasal pressure (Pn) was measured through a side port in the nasal mask and a thermocouple was placed over the lips to ascertain that mouth breathing did not occur during the measurements. nCPAP was increased from the initial value of 2 cm H₂O with 1 cm H₂O increments up to the efficacious level.

Sleep was scored using the criteria of Rechtschaffen and Kales (12) for 10-s epochs and respiratory events were scored using standard criteria. Hypopneas were defined as a 50% or greater reduction in tidal volume from the baseline value lasting at least 10 s. The apnea index (AI) and the AHI were established as the ratio of the number of apneas and apneas + hypopneas per hour of sleep. Apnea time was expressed as the sleep time spent in apnea. As indices of nocturnal hypoxemia we considered the mean Sa_O2 (Sa_O2 mean) during sleep and the minimal value recorded during sleep (Sa_O2 min).

During the nCPAP titration night, we measured Pcrit according to described methods (8) by relating changes in maximal inspiratory airflow (I_max) to varying levels of nCPAP (Pn) by least square regression analysis. The Pcrit was defined as the extrapolated pressure at zero flow. Measurements were made during light sleep (Stages 1-2 of non-rapid eye movement [NREM] sleep) in the supine position. Recurrent obstructive apneas and hypopneas were always allowed to occur before initial changes in nasal pressure. Mask pressure was then raised in 1 cm H₂O increments and at least 10 min of light sleep were observed before a subsequent rise in nasal pressure. During periods with hypopneas, the relationship between maximal airflow and nasal pressure was obtained by averaging the maximal flow for the last four breaths of an hypopnea. During periods free of apnea and hypopneas but with flow limitation, defined as a reduced and plateaued inspiratory airflow while esophageal pressure was increasing, maximal inspiratory airflow was measured at each level of nCPAP for several breaths. In all patients the r of the regression analysis was above 0.90.

Statistical Analysis

Data were expressed as means ± SD. Bivariate correlation analysis using Pearson's correlation coefficient was used to find the variables correlated with Pcrit and multivariate regression analysis was done to define the contribution of anthropometric and cephalometric variables in explaining pharyngeal collapsibility. Obese and nonobese patients were compared using one-way analysis of variance. When a significant difference was found, individual means were compared using the Student-Newman-Keuls test. The statistical significance level selected for all analyses was p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Clinical, anthropometric, and cephalometric variables of the study group are shown in Table 1.

The patients, age 52.0 ± 9.0 yr with a mean BMI of 31.6 ± 5.7 kg/m², had a mean AHI of 72.6 ± 31.8 ranging from 19 to 152 and a mean AI of 45.3 ± 34.5. The mean time of sleep spent in apnea was 74.5 ± 68.4 min ranging from 26 to 228. A wide range of nocturnal hypoxemia severity was present with a mean Sa_O2 of 93.5 ± 2.5% and a minimal Sa_O2 of 74.9 ± 13.9%. The mean neck circumference was 43.4 ± 3.8 cm.

As a group, the patients had normal blood gases; eight patients were hypoxemic, defined as having a Pa_O2  65 mm Hg (mean Pa_O2: 61.9 ± 3 mm Hg) and 15 were hypercapnic, with a Pa_CO2  45 mm Hg (Pa_CO2: 46.5 ± 2 mm Hg).

In each subject there was a significant linear relationship between the values of I_max and the corresponding Pn, with a correlation coefficient ranging from 0.90 to 0.98. Calculated Pcrit values ranged from 0.3 to 4.7 with an average value of 2.4 ± 0.97 cm H₂O.

Table 2 reports the correlation coefficients of craniofacial variables, diurnal and nocturnal parameters, and Pcrit. As expected, Pcrit was highly correlated with AHI (r = 0.52, p < 0.0001) and apnea time (0.68, p < 0.001). Regarding anthropomorphic variables, Pcrit was strongly correlated with the neck circumference (p = 0.38, p = 0.004) and with BMI (r = 0.29, p = 0.02), whereas no correlations were found with age (r = 0.14, p = ns).

Correlation of Pcrit with the cephalometric variables indicated a higher linear relationship with the hyoid bone position (H-Ph: r = 0.29, p = 0.03; H-Pns: 0.32, p = 0.02), and the SPl (r = 0.27, p = 0.04) and a near significant relationship with the H-MP (0.26, p = 0.06). No relationships were found in our study sample between Pcrit, tongue length, or craniofacial data. With all variables in the regression analysis no cephalometric variable independently predicted Pcrit, neck circumference alone explaining the 14% of the variance in Pcrit.

Because cephalometric cranial and soft tissue variables may differ in patients with and without obesity, the patients were classified as obese and nonobese according to a BMI > 30. Table 3 shows the clinical, polygraphic, and cephalometric data of the patients according to BMI. Of the study group, 27 were nonobese and 30 had a BMI > 30. Significant differences between groups were seen for AHI, apnea time, neck circumference, and Pcrit, with higher values in obese patients. Regarding craniofacial, hyoid bone position, and upper airway size data, there were no significant differences between groups except for ANB, which was greater in obese patients (4.1 ± 0.5 mm versus 2.0 ± 0.6 mm, p = 0.009). No differences in pharyngeal soft tissue variables were seen between the two groups but obese patients had longer H-Ph (38.7 ± 0.8 mm versus 35.0 ± 0.7 mm). Correlation analysis (Table 2) in the nonobese patients showed that Pcrit was related to H-MP. In the obese group Pcrit was significantly related to neck circumference (r = 0.38, p = 0.04) and SPl (r = 0.39, p = 0.03). A near significant relationship was also present with H-Ph (r = 0.34, p = 0.07) and H-Pns (r = 0.33, p = 0.06) (Table 2).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the present study we demonstrate significant correlation between Pcrit and anatomic variables assessed by cephalometry. Elongated soft palate and lower hyoid bone position increase the individual tendency to pharyngeal collapse, i.e., Pcrit, which in turn affects the frequency of sleep-related breathing disorders. Although the correlations between cephalometric variables and Pcrit are in the moderate range, they are consistent with the hypothesis of an influence of soft tissue configuration and anatomic narrowing on the pharyngeal collapsibility.

Whereas airway compliance, pharyngeal muscular force, and anatomic variables may all influence upper airway configuration, the critical factors responsible for control of pharyngeal patency remain controversial. Two hypotheses have been proposed to explain the tendency of patients with OSA to collapse: a neural hypothesis implying reduced dilator muscle activity and an anatomic theory suggesting an anatomic narrowing of the upper airway. According to the neural hypothesis, the loss of wakefulness stimulus and changes in neuromuscular control (9) may account for the greater pharyngeal collapsibility in patients with OSA. It is generally assumed that upper airway muscle activity is modulated by chemical stimuli and by changes in upper airway pressure. Breathing through a narrow pharynx generates greater collapsing forces, and pharyngeal collapse would depend on the force of contraction of the upper airway muscles. The more forceful the pharyngeal muscle contraction, the less likely that collapse will occur. However, abnormalities in bony structure or fat deposition in the neck may contribute to low pharyngeal cross-sectional area. Increased upper airway compliance and collapsibility have been found in OSA patients (13), partially related to regional neck obesity (14). Moreover, anatomic abnormalities may influence the greater collapsibility in patients with OSA because closing pressure is elevated in micrognathic infants (15) and in patients with enlarged uvula (16). Furthermore, patients who responded to UPPP were those with enlarged and hypertrophic tonsils in whom Pcrit greatly decreased after surgery (10). Thus, it could be quite correctly argued that changes in soft tissue and body facial anatomy may interfere with upper airway collapsibility.

In our patients with OSA, Pcrit was related to the length of the soft palate and the position of the hyoid arch, two structures that contribute to pharyngeal aperture and stability of the upper airway. Although this study does not provide definitive evidence concerning the mechanisms by which the anatomic variables interfere with upper airway collapsibility, our results provide interesting preliminary information that may be useful in future studies. The statistical comparison between the two groups showed that they did not differ for craniofacial architecture, and the pharyngeal soft tissue variables were also very similar. However, pharyngeal soft tissues, neck circumference, and hyoid bone compensation seemed to be involved in a different way in obese and nonobese patients.

Because the site of the collapse in 80% of patients with OSA is at the upper pharyngeal level, we can postulate that the Pcrit would become more positive when the tongue impinges on an elongated soft palate. Obesity, through neck and soft tissue fat deposition, and increased pressure on the neck of submental adipose tissue, may predispose to upper airway obstruction (17). Thus the correlation between Pcrit and SPl in obese patients is a reasonable finding. This supports the data of Schwartz and colleagues (18) who demonstrated changes in Pcrit in patients treated with weight loss.

Assuming that hyoid bone position is crucial to pharyngeal patency, several hypotheses can be advanced to explain which factors influence the position of the hyoid bone and consequently Pcrit. An imbalance between suprahyoid and infrahyoid muscles may influence the hyoid bone position. Chronic snoring trauma and/or repetition of chemical and mechanical stimuli may produce morphological alterations in the pharyngeal dilator muscles (19, 20) or pharyngeal neuropathy (21, 22). The consequent changes in contractile properties (23) and the progressive loss in the compensatory muscular mechanisms maintaining upper airway patency (24) and influencing hyoid position may determine a greater pharyngeal collapsibility.

Although no significant differences were found between obese and nonobese patients, nonobese patients had smaller H-Ph distance and smaller ANB. We know that mandibular protrusion increases upper airway caliber at the oropharynx and hypopharynx (25), mostly in nonobese patients, suggesting that the pharyngeal occlusion will more likely occur when the mandible is smaller or receded. From these data we can suggest that alteration in mandibular position may affect hyoid bone position and, thus, Pcrit. It remains also possible that other factors, besides anatomic and neuromuscular variables, may alter the position of the hyoid bone. In our group larger H-MP and larger H-Pns were found in obese patients having greater Pcrit. Thus obesity, by neck fat deposition, may shift the hyoid bone caudally altering the pharyngeal patency and leading to greater upper airway collapsibility.

Although significant relationships do exist between the soft palate length, hyoid bone position, and Pcrit in our patients, the correlations are mostly in the moderate to lower range. It is likely that our failure to observe a clear relationship between Pcrit and cephalometric variables may be the result of other factors that we were unable to control. An important consideration in the present study is our ability to accurately assess upper airway collapsibility. Many methods have been used to measure pharyngeal collapsibility and the one most widely used is Pcrit. One problem in using Pcrit is the interference of anatomic and neuromuscular factors on this variable: mandibular advancement (26), sleep posture (27), tongue position, neck flexion and enlarged uvulas (16), may all affect the pharyngeal collapsibility. Thus, Pcrit may not adequately reflect the true individual pharyngeal collapsibility. The closing pressure (Pclose) has been introduced to improve the assessment of pharyngeal collapsibility. This is an entirely objective measure of pharyngeal properties obtained under general anesthesia (28) or during nCPAP treatment (29) in which muscular influences are removed. By using Pclose these investigators demonstrated that the anatomic narrowing related to mandibular position (26) or to enlarged adenoids and tonsils (30) induces a higher collapsibility of the pharynx. Moreover, adenotonsillectomy (30) and mandibular advancement (26) may be therapeutic only when Pclose becomes more negative. These observations are in agreement with our results supporting the hypothesis that an anatomic narrowing predisposes to pharyngeal collapse.

There are several other possible explanations why the airway of the awake patients did not strongly affect Pcrit in our sample. First, cephalometric data were obtained when the subject was awake and in the sitting position, and thus did not reflect the changes that could have occurred during sleep when the awake activation of pharyngeal muscles is removed and when the patient is supine. Even though tongue and soft palate thickness increases in supine position, a close relationship between upright and supine craniofacial and soft tissue measurements (31) is present. Second, cephalometry provides information for anteroposterior but not lateral pharyngeal structures that are implicated in the pharyngeal narrowing. Studies using magnetic resonance imaging techniques have demonstrated that differences between snorers and patients with OSA in airway caliber are more related to lateral pharyngeal wall size and thickness than to anteroposterior dimension (6). This finding is supported by previous studies showing that nCPAP dilates the upper airway more laterally than in the anteroposterior direction (32) and that the fat deposition is greater in the lateral upper airways walls (33). Third, it is possible that the shape of the upper airway (34) and its elliptical orientation may predispose to pharyngeal collapse during sleep more than the airway size itself. Finally, we collected data in patients with severe OSA, and snorers or hypopneic patients were not analyzed. Because Pcrit is more frequently subatmospheric in individuals with snoring, it seems reasonable to expect a greater relationship when patients having negative values of Pcrit are considered.

In summary, in analyzing the relationship between Pcrit and cephalometric variables we wondered whether patients with OSA had a more collapsible pharynx, dependent on anatomic variables. Overall, a relationship between Pcrit, hyoid bone position, and soft palate length was found in patients with severe OSA. Although these results should not be overstated, it could be quite correctly argued that changes in soft tissue and body facial anatomy may interfere with upper airway collapsibility. Further studies in sleeping condition would provide more information on the influence of the anatomic variables on the upper airway collapsibility.