Effect of Lateral Positioning on Upper Airway Size and Morphology in Sedated Children

Children aged 2–12 yr requiring MR examination of the head or neck region using deep sedation with propofol were eligible for inclusion. The protocol was approved by The Children’s Hospital of Philadelphia Department of Radiology Research Review Board and the institutional review board. Written consent was obtained from all parents, and verbal assent was obtained from children when appropriate. Exclusion criteria included acute or chronic lung disease, upper airway disease or dysfunction of any kind, heart disease, obesity (body weight > 90th percentile for age), known sleep apnea syndrome, central nervous system disease likely to influence muscle tone, any type of abnormality in neck anatomy, or the requirement for any type of airway adjunct device (e.g. , oral airway, endotracheal tube).

Children were anesthetized before the MR scan with 8% sevoflurane in 70% N²O or with intravenously administered propofol if an indwelling intravenous catheter was present. Maintenance of deep sedation was accomplished using propofol at a starting dose of 200 μg · kg⁻¹· min⁻¹and titrated based on maintenance of unconsciousness (e.g. , lack of spontaneous limb movement) and normal hemodynamic and ventilatory parameters (e.g. , within 20% of baseline values). All children breathed spontaneously without assistance throughout the entire anesthetic. Supplemental oxygen was administered by nasal cannula with a capnograph attachment (2 or 3 l/min) in all cases. After completion of the clinically indicated scans, the study protocol was performed while the child remained in the supine position. The head and neck rested in the neutral position on a small blanket, without obvious flexion, extension, lateral flexion, or lateral rotation. The degree of mouth opening was not controlled in either position.

The study protocol consisted of T1 axial scans of the upper airway (based on a sagittal scout scan) in 3-mm slices without a gap. We used a 1.5-T Vision System MR (Siemens, Iselin, NJ), and images were acquired using an anterior–posterior volume head coil. The cephalad border consisted of the top of the nasal cavity, and the caudad border consisted of the subglottic region below the vocal cords. The time to completion of the study scan was approximately 5 min. At the completion of the airway scan in the supine position, each subject was turned laterally (based on a randomization scheme for each episode) on the MR gurney. Extreme care was taken to ensure that the head and neck were maintained in the neutral position, without flexion, extension, lateral flexion, or lateral rotation. All other body parts were appropriately protected, while the child’s body was maintained in a plane at 90° to the axis of the table. When the child was secured in this manner, the upper airway scans were repeated using the same protocol as for the supine scans.

After the study scans, the MR images were saved and transferred to a UNIX-based Sun workstation (Sun Microsystems, Santa Clara, CA) where the images were processed and segmented automatically to render a three-dimensional reconstruction of the upper airway, including its surface description, centerline, and volume. This was accomplished using a software program (3D-VIEWNIX; Medical Imaging Processing Group, University of Pennsylvania, Philadelphia, PA) that uses fuzzy connectedness segmentation.10,11A centerline through the airway is computed that passes through all points maximally distant from the perimeter of the airway at sequential planes orthogonal to the airway axis. The total upper airway volume was computed as a product of the centerline length and the mean cross-sectional area. Cross-sectional areas at planes orthogonal to the centerline were computed every 0.2 mm after interpolation, filtering, and thresholding of the original axial slices. We defined the cephalad border of the upper airway by the posterior edge of the vomer and used the vocal cords as a lower landmark with which to divide the upper airway into 10 consecutive (10%) equidistant intervals along the centerline to facilitate analysis and comparison between body positions. Only airflow-conducting portions of the upper airway are used for the analysis of volume and cross-sectional area.

Data from previous studies indicated that in normal children, the mean cross-sectional area of the upper airway is 47.1 ± 18.2 mm²(mean ± SD).12We considered a 30% change in upper airway cross-sectional area to represent a clinically important change. Based on this premise, a sample size of 16 would have 80% power to detect a difference in means of 14 mm²using a paired t  test with a 0.05 two-sided significance level. We determined differences in total airway volume and cross-sectional areas at multiple levels of the pharynx using two-sided paired t  tests.

Nineteen children enrolled into the study; however, the data on one child was accidentally lost in the transfer process, and another child did not have study data collected because of copious secretions and frequent coughing during the clinically indicated scan. Therefore, adequate data were collected on 17 children (10 in the left lateral decubitus position, 7 in the right). There were 9 boys and 8 girls. Their ages ranged from 2 to 11 yr (mean ± SD, 5.4 ± 2.2 yr), and their weights ranged from 11.1 to 52.0 kg (mean ± SD, 20.1 ± 10.1 kg). Table 1lists each patient and his or her diagnosis. All subjects except two received sevoflurane for induction of general anesthesia. There were no complications or adverse events during the clinical or study scans. No child demonstrated obvious upper airway obstruction, and no child experienced oxyhemoglobin desaturation (< 95% oxygen saturation measured by pulse oximetry) at any time during the scans.

The separate results from the right and left lateral positions were similar and were therefore combined for the final analysis. Sixteen of the 17 subjects had increases in total upper airway volume when placed in the lateral position. As a group, the total airway volume was 6.0 ± 2.9 ml³in the supine position (mean ± SD) and 8.7 ± 2.5 ml³(mean ± SD) in the lateral position (P < 0.001; fig. 1). Figure 2illustrates the reconstruction of the total upper airway in both body positions in one subject using 3D-VIEWNIX.

All noncartilaginous cross-sectional areas of the upper airway increased in area in the lateral compared with the supine position (fig. 3). The region between the tip of the epiglottis and vocal cords demonstrated the greatest relative percent change.

Using MR analysis of upper airway morphology, we demonstrated that placement of a sedated, spontaneously breathing child in the lateral position enlarges the upper airway. Although all noncartilaginous areas of the upper airway became significantly larger, the level at or below the tip of the epiglottis demonstrated the largest relative percent change. Isono et al.  7studied this phenomenon in paralyzed adult patients and reasoned that in the supine position, gravity caused the upper airway to be more constrained by surrounding anatomical structures when compared with the lateral position. More detailed examination of the MR images (fig. 4) revealed that the base of the tongue seemed to be in direct contact with the anterior surface of the epiglottis and thus may have been responsible for posterior displacement of the epiglottis with resultant narrowing of the airway in this region. In some children sedated with propofol, the region of most narrowing within the upper airway lies at the level of the epiglottis.13Lateral positioning decreased this area of narrowing, presumably by a gravitational effect. However, our methodology was not able to determine the cause of the airway narrowing in the supine position, and there may be other factors that influence airway size, such as changes in tissue and airway compliance at any level of the upper airway that occur with the change in position. Our findings provide a possible explanatory mechanism for the alleviation of upper airway obstruction by application of manual anterior mandibular advancement (jaw thrust) in sedated children.14Furthermore, our results provide further justification of lateral positioning in sleeping or sedated children with congenital laryngomalacia.

Clinically, these results indicate that less upper airway narrowing may occur when spontaneously breathing children undergo deep sedation in the lateral position. In pediatric anesthesia textbooks, the lateral position is often advocated for children after emergence from general anesthesia, based on clinical experience.15Although a randomized clinical trial comparing supine and lateral positions has not yet been performed, some patients with sleep apnea syndrome demonstrate less upper airway obstruction when in the lateral position.4,5Therefore, it is reasonable to conclude that semiconscious patients in the lateral position may demonstrate less upper airway obstruction.

Two of our subjects did not demonstrate an overall increase in upper airway volume after assumption of the lateral position. One was a 4-yr-old, 15-kg girl who was placed on her left side. Her diagnosis was a neck skin lesion that did not extend into the soft tissue of the neck. The reason her upper airway did not widen in the lateral position is otherwise unclear. The other patient was an 8-yr-old, 23-kg boy who was placed on his right side. This patient demonstrated larger than usual volumes in both supine and lateral positions when compared with the remainder of the study group (uppermost subject in fig. 1). Therefore, in retrospect, this subject had an unusually large upper airway of unknown etiology that may have influenced the results.

A possible limitation of this study is the lack of randomization of the order of the positions. In the context of the clinical situation, it was not practical to perform the clinically indicated MR scan initially supine, then turn half of the patients lateral and back to supine for the study scans. Because each patient served as his or her own control, we did not believe there were important disadvantages to performing the initial scan supine in every subject. We cannot think of an anatomical or physiologic reason why the second scan performed in the study sequence would consistently result in a larger upper airway, as long as the depth of propofol sedation was maintained at a constant level. Another limitation of this study is the lack of external validity to other types of patient populations (e.g. , age less than 2 yr, sleep apnea) and different types of sedatives and the depth of sedation. These remain to be seen in future studies. Patients presenting for MR imaging of the head and neck were inherently predisposed to abnormalities in that region, and we were careful during subject selection to exclude patients with possible airway abnormalities or muscle weakness secondary to a brain tumor or neurofibromas. However, subclinical abnormalities cannot be ruled out.

In conclusion, using three-dimensional MR image analysis, we demonstrated that the upper airway of sedated, spontaneously breathing children widens significantly in the lateral position compared with the supine position. This widening occurred at all noncartilaginous areas of the upper airway and was most pronounced in the region at and below the tip of the epiglottis. Our empirical findings confirm the widespread clinical experience that sedated children experience less upper airway obstruction in the lateral position when compared with the supine position.

The authors thank Rosetta Chiavacci, B.S.N. (Department of Anesthesiology and Critical Care, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania), without whose expert assistance this work would not have been possible, and the magnetic resonance imaging technicians of The Children’s Hospital of Philadelphia, whose dedication to the care of children is unparalleled.