Our study demonstrated several anatomical characteristics differentiating obese children with OSAS and obese children without OSAS. These include the increased size of upper airway lymphoid tissues and increased size of the parapharyngeal fat pads and abdominal visceral fat, noted in children with OSAS. However, our results did not demonstrate that variation in our subjects' BMI Z-score was an effect modifier on the above anatomical characteristics or severity of OSAS in these subjects.
To maximize the statistical power of comparisons between the two groups our study used matched pairs. We matched each obese subject with OSAS by age, height, weight, BMI Z-score, sex, and ethnicity with an obese control. Such matching serves to minimize the effect of confounding factors related to demographics and anthropometrics, which can be highly variable in children during growth and development.
When considering the various mechanisms leading to OSAS in obese children one should initially consider whether these may be related to an augmented effect of known causative factors resulting from their obesity and/or a distinct underlying pathophysiology of the disorder distinguishing it from OSAS in nonobese children. Because obesity is characterized by extensive somatic growth, we postulated, as previously described in younger nonobese children with OSAS (21
), that older obese children with OSAS will have an excess of upper airway lymphoid tissues compared with obese subjects without OSAS.
In this study, in addition to comparing the size of the adenoid and tonsils, we have compared the difference in retropharyngeal nodes, which may also contribute to reduced airway volume. The retropharyngeal nodes are defined as lymph nodes located between the internal carotid arteries from the base of the skull to the hyoid bone. The afferents of these nodes drain the pharyngeal nasal cavity, adenoid, paranasal sinuses, and the auditory tubes and their efferents pass to the deep cervical lymph nodes. Because of the proximity of the retropharyngeal nodes to the nasopharynx and oropharynx, and because no rigid anatomic barrier lies between these nodes and the upper airway, the retropharyngeal nodes may contribute directly to reduction of the volume of the upper airway when they are enlarged. This is the first study to examine how these nodes contribute to airway restriction. Although we have focused in this study on obese children, these nodes may play a similar role in nonobese children with OSAS as well.
Several important relationships that enhance our understanding about the pathophysiology of OSAS in obese children have been observed in our study: first, the larger size of the adenoid (48%), tonsils (29%), and retropharyngeal nodes (54%) in the obese OSAS group, which is not offset by other soft tissue volumes being reduced, could explain the 28% reduction in oropharyngeal size, as compared with control subjects. It should be noted that the oropharynx loses 1.4 cm3
whereas lymphatic tissue gains 7.3 cm3
; however, not all the lymphatic tissue actually impinges into the oropharyngeal space. This difference, measured during wakefulness, may predispose to upper airway obstruction during sleep. The oropharyngeal airway (velopharynx) has been identified to be an area of vulnerability to airway collapse in both children (26
) and adults (28
). The velopharynx in children has also been described as the “overlap region” (27
), because it is where the lower portion of the adenoid and upper poles of the tonsils overlap with each other and the soft palate. Second, the size of each of the upper airway lymphoid tissues significantly correlates with severity of OSAS whereas BMI Z-score did not have an effect modifier on these tissues. This suggests that lymphoid proliferation as it relates to OSAS is independent of obesity and could possibly be linked to other conditions such as local or systemic inflammation described in nonobese children with OSAS (31
) as well as in obese children with the disorder (34
). Third, the soft palate, tongue, and mandible did not differ in size between groups or correlate with OSAS severity. However, the tongue and mandible did correlate with BMI Z-score. Thus, these tissues may not contribute to OSAS in obese children as has been reported in adults (16
). It is possible that these tissues develop as risk factors later in life and contribute to the adult form of OSAS.
Regarding body fat composition, we found significant differences between groups. On average, children with OSAS had a 28 and 30% increase in size of the parapharyngeal fat pads and abdominal visceral fat, respectively. However, we could not demonstrate a significant correlation between severity of OSAS and alterations in body fat composition, or establish BMI Z-score as an effect modifier on these tissues. These findings could reflect the nature of our design in that only obese subjects within a narrow range of BMI Z-score (median, 2.4; range, 1.7–3.1) were included. Thus, it would be important to confirm these findings in a larger scale study.
Nevertheless, the finding of increased parapharyngeal fat pad size contributing to airway restriction in subjects with OSAS in this study is similar to that noted in obese adults with OSAS (14
), but not in nonobese adults (15
) or children with the disorder (21
). Similarly, increased abdominal visceral fat in our subjects with OSAS parallels the observation in obese children with metabolic syndrome and OSAS that used an indirect anthropometric measure of waist circumference (6
). Thus, increased visceral fat in the neck and body has several important clinical implications in relation to OSAS. Visceral fat can directly affect chest wall and upper airway mechanics by reducing functional residual capacity, making such subjects more vulnerable to hypoxemia during sleep (20
). Such a reduction of lung volume could also predispose to airway collapse by reducing tracheal traction and airway stability (36
). In addition, visceral obesity is strongly associated with the proinflammatory and prothrombotic conditions that increase risk for insulin resistance, type 2 diabetes, atherosclerosis, and OSAS (37
). On the other hand, studies have demonstrated that OSAS can independently induce metabolic syndrome by decreasing insulin sensitivity in both animal and humans (34
Although we noted that one subject in the OSAS group had glucose intolerance and one had type 2 diabetes, and none of the control subjects had evidence of prediabetes or type 2 diabetes, our study was not designed to correlate between OSAS, anatomical measures, and the biomarkers of metabolic syndrome, such as insulin resistance, dyslipidemia, inflammation, and so on. Therefore, we are limited in drawing conclusions regarding this possible association. Nevertheless, we suggest that the two groups emerging from the present study with similar BMI Z-score and that were a priori
selected according to the existence or nonexistence of OSAS, present distinct phenotypes of childhood obesity: the first, with marked visceral adiposity, upper airway lymphoid hypertrophy, and OSAS, and a second with less profound visceral adiposity, smaller upper airway lymphoid tissues, and no evidence of OSAS. We speculate that the first phenotype is more prone to develop metabolic syndrome and its sequelae because OSAS was shown to independently lead to the disorder in both children and adults (6
In summary, this study noted generalized overgrowth of upper airway lymphoid tissues in obese children with OSAS, restricting their upper airway. However, because surgical modalities, particularly adenotonsillectomy, in obese children often do not completely ameliorate OSAS (8
), a prospective randomized control trial evaluating outcomes of this approach versus other treatment alternatives such as noninvasive ventilation and weight management would be helpful. Such a study should also include the metabolic outcomes linked to obesity and OSAS, such as insulin resistance and metabolic syndrome.