© 2002 American Thoracic Society
Utility of Noninvasive Pharyngometry in Epidemiologic Studies of Childhood Sleep-disordered BreathingDepartments of Internal Medicine and Pediatrics, Case Western Reserve University, University Hospitals of Cleveland; and Rainbow Babies and Children's Hospital, Cleveland, Ohio Correspondence and requests for reprints should be addressed to Susan Redline, Department of Pediatrics, Rainbow Babies and Children's Hospital, 11100 Euclid Avenue, Cleveland, OH 44106. E-mail: sxr15{at}po.cwru.edu
Measurement of pharyngeal dimensions may contribute to the characterization of anatomic risk factors for sleep-disordered breathing (SDB) in children. Acoustic pharyngometry, a noninvasive method, has been used successfully in adults, but application in children has been limited. We sought to evaluate the feasibility and utility of this technique in children, including assessment of the variation of pharyngeal measurements with height, sex, ethnicity, prematurity, and indices of SDB. Subjects were drawn from a large, community-based cohort of children of age 8–11 years. Demographic, morphologic, and sleep-related information were collected via standard questionnaires, direct measurement, and home cardiorespiratory monitoring during sleep. Pharyngeal dimensions were assessed in 203 children using acoustic pharyngometry performed with an optimized mouthpiece. In this sample, the coefficient of variation of minimum pharyngeal cross-sectional area (CSA) and mean CSA were similar to those in adults (8.0 and 11.1%, respectively). The minimum CSA, but not mean CSA, was significantly reduced in preterm children, habitual snorers, and children with SDB relative to unaffected children. Thus, minimum CSA is a useful measure for evaluating SDB risk factors in preadolescent children.
Key Words: acoustic pharyngometry • habitual snoring • sleep-disordered breathing
Sleep-disordered breathing (SDB) is a complex, chronic disease that is expressed after a given threshold level of susceptibility is exceeded. Susceptibility relates to the propensity for repetitive upper airway collapse. In any given individual, anatomic and neuromuscular factors that influence upper airway size and/or function determine this propensity. In adults, the strongest risk factors for SDB are obesity and male sex (1). In children, adenotonsillar hypertrophy and major craniofacial anomalies are considered to be important risk factors for SDB (2, 3). However, the relative contribution of craniofacial morphologic (hard and soft tissue) features to SDB susceptibility has not been well-defined in children. Reduced pharyngeal cross-sectional area (CSA) may be a final common pathway by which genetic and other factors impact the development of SDB. Acoustic pharyngometry, a noninvasive technique, has been used to assess pharyngeal size and compliance in adult snorers with and without SDB (4–7). These studies showed that snorers had smaller CSA, regardless of apnea–hypopnea index (AHI), and that those with SDB (defined by an elevated AHI) exhibited higher pharyngeal compliance (as measured by fractional change of pharyngeal area with varying mouth pressure), a marker of susceptibility to collapse (4, 5). These studies also suggested that the acoustic pharyngometry technique is reasonably precise. In adults, intrasubject variability of pharyngometric measures was reported as less than 10% (5, 6). Similar data in children are not available, and the contribution of reduced pharyngeal CSA to SDB has not been quantified in the pediatric population. The paucity of data that address risk factors, outcomes, and the natural history of SDB in children has led to a recent interest in epidemiologic studies of pediatric SDB. Studies of large samples of children require reliable, noninvasive tools that provide discriminative or predictive information. As part of the Cleveland Children's Sleep and Health Study (CCSHS), we collected pharyngometry data, using the acoustic reflection technique, on a well-defined cohort of children born in the Cleveland area between 1988 and 1993. Using concurrently collected demographic, upper airway, and home sleep study data, we aimed to (1) evaluate the utility of using acoustic pharyngometry to characterize upper airway dimensions in children with and without SDB and (2) characterize the variation of pharyngeal dimensions with demographic, anthropometric, and sleep-related variables.
Details on definitions, measurements, and scoring of respiratory events can be found in the online data supplement.
Study Sample and Data Acquisition Demographic and medical data were assessed with the Children's Health Questionnaire, a pediatric modification of a validated instrument (9). Height and weight were measured by trained research personnel. Tonsillar size was noted on a 5-point scale (10). Overnight cardiorespiratory monitoring during sleep included recording of thoracic and abdominal excursions, estimated tidal volume (by inductance plethysmography) (11, 12), pulse oximetry, heart rate (bipolar electrocardiography leads), and body position (PT-2 system; SensorMedics, Yorba Linda, CA).
Pharyngometry Measurements To adapt the device to children, a customized pediatric mouthpiece was introduced partway through the study and was used for testing the last 60% of the sample (n = 374). The use of the new mouthpiece resulted in an approximate 2-fold increase in the proportion of subjects with high quality data. Analyses were further restricted to the 203 children who had used the new mouthpiece and who produced three high quality curves (54%).
Scoring of Respiratory Events
Cephalometry
Analyses The protocol was approved by University Hospitals' Institutional Review Board. Informed consent was obtained from the parents/legal guardian of each child, and assent was obtained from the child.
Sample Characteristics The sample consisted of 203 children with a mean age (± standard deviation) of 9.35 ± 0.82 years (described in Table 1). There was a slight female predominance (53%), and one-half of the participants were black. More than half (53%) of the subjects were born prematurely (48 born between 20 and 32 weeks, 41 born after 32 weeks, and 18 of uncertain gestational age younger than 36 weeks). The average height percentile was 56 ± 27% (slightly above the 50th percentile for age and sex). Mean body mass index (BMI) was 18.2 kg/m2 ± 4.0. The median AHI was 0.2 (range 0–42). The prevalence of SDB, defined by an AHI value of 5 events/hour or more, was 4.5%, and the prevalence of snoring with an AHI value less than 5 was 17.4%. As expected, snoring and AHI were significantly associated, although there was considerable overlap between these entities (Spearman's r = 0.22, p = 0.002).
None of the characteristics in the subsample defined by use of a customized mouthpiece and ability to produce high quality pharyngometry data varied substantially from those of the overall CCSHS cohort (n = 657 at the time of analysis), with the exception that the analyzed sample consisted of a larger proportion of black children than in the overall cohort. This difference is attributable to preferential recruitment of blacks later in the study, which coincided temporally with the introduction of the optimized pediatric pharyngometry mouthpiece.
Quality and Reproducibility of Pharyngeal Measurements
Pharyngeal Dimensions as a Function of Subject Characteristics
Although we did not aim to present a predictive model for SDB (but rather to identify the feasibility of the technique and the its correlates of pharyngeal dimensions), we explored the extent to which subject characteristics were related independently to pharyngeal CSA. In a multiple regression model, assessing the joint effects of age, height, sex, and prematurity, only height remained significantly related to minimal CSA (p = 0.003).
Variation of Pharyngeal Measurements by Sleep-related Measures
We also evaluated the association between tonsil size and pharyngeal measurements. For this analysis, tonsil size was analyzed both by comparing those with and without moderate–large tonsils and by comparing the two groups at each extreme for tonsil size (< 25% of pharyngeal space, n = 55 versus > 75% of pharyngeal space, n = 15). In both cases, no significant differences in mean or minimum CSA with tonsil size were apparent (p > 0.2). In contrast, nocturnal snoring, defined on the scale described previously was significantly increased in subjects with larger tonsils (r = 0.25, p < 0.005).
Association Between Pharyngometric and Cephalometric Measures
Utility of Pharyngometry in Studies of Pediatric SDB The usefulness of any diagnostic test depends on numerous factors, including its feasibility, predictive ability (regarding predicting clinical endpoints or discriminating subgroups of the population), precision (or reliability), accuracy, and cost. We have demonstrated that acoustic pharyngometry can be adapted for use in children, and trained personnel can successfully obtain highly reproducible data even in nonclinical settings (such as the home). In such settings, however, only 78% of children of age 8–11 years were able to produce minimally acceptable data; and high quality data could only be achieved reproducibly for 54% of the sample studied with the improved mouthpiece. In the subset able to produce three acceptable curves, reliability was excellent and comparable to that reported by reports for adults studied in more controlled office or laboratory settings (4, 6). Furthermore, as discussed subsequently, these measurements provided data that discriminated children with and without snoring and SDB. Measurements also varied in an expected manner with indices of body size and growth (height and age). In addition, pharyngeal dimensions were reduced in children who had been born prematurely—a subgroup we have shown in preliminary work to be at increased risk for SDB. The failure rate reported in the current study should be balanced against the portability and brevity of the measurements, which generally took less than 15 minutes and were conducted in the homes of study participants. Alternatives, such as magnetic resonance imaging (MRI) or endoscopy, although considerably more accurate and comprehensive, are also much more expensive and less accessible. The variation of acoustic pharyngometry measurements across snoring/SDB groups suggests the utility of this technique for the clinical assessment of children at risk for SDB or for phenotyping in research studies, provided that increased technical experience reduces the failure rates. This study did not directly compare acoustic pharyngometry with MRI. In our epidemiologic study, such assessments were not feasible and would have required comparing measurements made supine (MRI) with those made upright (pharyngometry). However, a previous study of adults assessed with both MRI and acoustic pharyngometry showed that the two techniques provided estimates of mean oropharyngeal area and pharyngeal volume within approximately 20 and 10% of each other, respectively (6). Radiographic comparisons have shown that pharyngometry allows accurate reconstructions of the geometry of other airway structures including the larynx, trachea, and nasal airways (6). Airway models that have been used to assess the accuracy of the technique further indicate that accuracy is greater with smaller mouth areas (6). These findings, together with the qualitatively similar curves obtained in children and adults, suggest that acoustic reflectometry in children who can reliably perform the maneuver should provide assessments of airway size at least as accurate as what has been reported for adults with larger mouths. We did assess the relationship of pharyngometric and cephalometric measurements (both made in the sitting position) in a sample of the cohort who participated in a second-stage laboratory study. A previous study of adults had reported that subjects with severe sleep apnea had both reduced mean pharyngeal CSA and mandibular length (7). Our finding of a correlation between mandibular length, which is generally considered to be an important skeletal determinant of pharyngeal size, and mean CSA supports the overall validity of the latter noninvasive measure for estimating anatomic risk factors for SDB in children. The lack of significant relationships between minimal CSA and mandibular length also is not surprising because minimal and mean CSA were only moderately correlated (r = -0.62), and minimal CSA may be influenced more by discrete areas of anatomic compromise (soft tissue) than by a general reduction in area. Unlike in pediatrics, a standard curve has been established for acoustic pharyngometry in adults. In 350 normal volunteers (77% male), the mean CSA was 3.19 ± 0.311 cm2 in males and 2.81 ± 0.108 cm2 in females, with a CV of 5–7% (14). Clearly, our results cannot be compared directly because of the disparate ages of subjects in the two studies. However, both estimates of reproducibility are of the same order of magnitude as in previous studies (4, 6). Thus, it appears that this tool can be a precise instrument when used in pediatric subjects, provided that data are screened for quality and that a proper mouthpiece is used.
Relationship of Pharyngeal Dimensions to Snoring and SDB In the case of airflow through the oropharynx, previous investigators have introduced the concept of a "critical pressure" (Pcrit) to provide a mechanical basis for upper airway collapsibility. In this model, pharyngeal airflow can only occur when upstream pressure exceeds the intrinsic pressure in the oropharynx, Pcrit, extrapolated from a plot of pharyngeal flow versus nasal pressure as the applied pressure at zero flow (16–18). In adults, Pcrit is negative (airways stay patent) in normal subjects (19), less negative in snorers (increased susceptibility to collapse), and positive (airway collapse leading to obstruction) during sleep in subjects with OSA (16). Similarly, in children, a positive critical pressure was found in subjects with OSA and a negative Pcrit was found in primary snorers and in a subset of OSA subjects after undergoing tonsillectomy and adenoidectomy (20). Our findings of a relationship between indices of SDB and a reduced minimum CSA may be interpreted within this physiologic framework: a minimal pharyngeal CSA may lead to decreased intraluminal pressure in the pharynx distal to the obstruction, resulting in an increased susceptibility to collapse. Other studies using pharyngometry have identified a reduced mean CSA in adult snorers and subjects with OSA (4, 7). However, those findings included individuals with a greater degree of SDB than what was observed in our study.
Relationship of Pharyngeal Area to Subject Demographics In contrast to adult data (14), we did not observe significant sex differences in pharyngeal dimensions in our sample of prepubertal children. This result may simply reflect similarities in overall body size, as there were no sex differences in height percentile or BMI in our sample (male 54 ± 27%, 18.3 ± 3.7 kg/m2 versus female 58 ± 28%, 18.1 ± 4.2 kg/m2; p = 0.41, p = 0.69, respectively). Longitudinal follow-up of children and young adults with simple tools such as acoustic reflectometry may shed further light on how pharyngeal dimensions change after puberty and with growth and maturation and hormonal changes.
Reduced Pharyngeal Area and Prematurity
Relationship of Tonsil Size, Habitual Snoring, and Pharyngeal Dimensions The propensity for habitual snoring and SDB is likely influenced by multiple factors including airway dimensions, tonsil size, soft tissue mass, facial anatomy, and neuromuscular function—combinations of abnormalities in some or all of these areas may produce the SDB phenotype. Tonsil size, as assessed by physical examination, may be too crude an index of anatomic compromise of the airway to differentiate children in the community with and without SDB. In contrast, our data suggest that pharyngometric measurements of minimum CSA could contribute to the prediction of SDB in large samples of unselected children.
Conclusions and Limitations We also provide new data that suggest that preterm children have smaller pharyngeal dimensions and that this may be one explanation for the increased risk of SDB with prematurity that we have observed. Future work is needed to further assess the link between prematurity, pharyngeal measurements, and subsequent development of SDB and to model additional physiologic and anatomic risk factors for SDB.
The authors wish to express their gratitude for the invaluable assistance of Jean Arnold, Sarah Bivins, Judy Emancipator, Najla Golebiewski, Heather Rosebrock, Susan Surovec, and Dina Tell, who have recruited the patients, collected the data, and worked diligently to implement the use of the new pharyngometry mouthpiece. In addition, they thank Gary Glass of E. Benson Hood Laboratories for customizing the pharyngometry software and reviewing the manuscript. Finally, they are most indebted to the members of the cohort who so generously continue to invite them into their homes. Supported by National Institutes of Health grants NHLBI RO1HL60957, NHLBI 04426, RO1 NR02707, and MO RR 00080.
This article has an online data supplement, which is accessible from this issue's table of contents online at www.atsjournals.org Received in original form November 5, 2001; accepted in final form February 28, 2002
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
INTRODUCTION
5). Habitual snorers showed a highly significant reduction (10.3%, p = 0.006) in minimum CSA compared with nonsnorers. Likewise, children with an AHI value of 5 or more had a significantly smaller minimum CSA (10.4%, p < 0.05) than those with an AHI value less than 5. Furthermore, minimal CSA appeared to vary in a "dose–effect" fashion with degree of SDB (1.17 ± 0.26, 1.07 ± 0.23, and 0.96 ± 0.14 in nonsnorers [n = 157], snorers with an AHI value less than 5 [n = 35], and children with an AHI value of 5 or more [n = 9]). Mean CSA did not vary significantly across these categories. Minimum CSA also correlated inversely with the degree of nocturnal snoring (assessed on a 0–5 point scale, 0 being no snoring and 5 being very frequent snoring over the past month) (r = -0.15, p < 0.03). 
