Neck fat and obstructive sleep apnea in obese adolescents

Abstract

Study Objectives

Increased neck circumference, a surrogate for the neck fat that can narrow the upper airway in obese individuals, is a risk factor for obstructive sleep apnea syndrome (OSAS) in adults, but the association between neck fat and OSAS in adolescent males and females is unknown. We hypothesized that obese adolescents with OSAS have more neck fat than controls, females more neck fat than males, and that neck fat correlates with obesity and OSAS severity.

Methods

Obese adolescents with OSAS and obese and normal-weight controls underwent upper airway magnetic resonance imaging, polysomnography, and anthropometrics, including neck circumference measurement. Intra-neck and subcutaneous neck fat measurements were manually segmented and compared among the three groups using ANOVA and between males and females using t-tests. The relationship between polysomnographic parameters and neck fat measurements was assessed in adolescents with OSAS using Pearson correlations.

Results

One-hundred nineteen adolescents (38 females) were studied: 39 obese with OSAS, 34 obese controls, and 46 normal-weight controls. Neck fat was not greater in adolescents with OSAS compared to obese controls (p=0.35), and neck fat volume was not related to OSAS severity (p = 0.36). However, obese adolescents had more neck fat than normal-weight controls (p < 0.001), and neck fat volume correlated with neck circumference (r = 0.53, p < 0.001). Females had significantly greater cross-sectional neck fat than males (p < 0.001).

Conclusions

While neck fat is associated with obesity and neck circumference in adolescents and is greater in females versus males, it does not appear to correlate with presence and severity of OSAS.

Statement of Significance.

Obesity is associated with obstructive sleep apnea syndrome (OSAS) in adolescents, but differences in regional fat distribution may have a better correlation with OSAS status than general measures of obesity. Neck circumference is associated with OSAS in adults, but its association with neck fat and OSAS in obese male and female adolescents is poorly understood.

This study recruited a large cohort of obese adolescents with and without OSAS and normal-weight controls, who underwent polysomnography and upper airway MRI for evaluation of neck fat volumes and cross-sectional area. While neck fat was strongly associated with neck circumference and obesity, there was no association between neck fat and OSAS status or severity. Females had significantly more neck fat than males.

Introduction

Obstructive sleep apnea syndrome (OSAS) is a disorder of breathing during sleep characterized by upper airway obstruction that causes disruption of gas exchange and sleep patterns in children and adolescents [1]. Obesity is a risk factor for OSAS in both adults [2] and children [3]. However, differences in regional fat distribution in obese individuals may correlate better with OSAS status or severity than more general measures of obesity such as body mass index (BMI). Neck circumference correlates well with obesity in adults [4] and is a moderate predictor of obesity in children [5]. In adults, increased neck circumference has been associated with both snoring and OSAS [6, 7], especially when corrected for height [8]. In children, the association between neck size and OSAS is less clear [9–11], particularly in obese children [12]. One study that developed pediatric reference ranges for neck circumference found that in adolescent males, relative risk for OSAS was 3.3-times higher for those with neck circumference ≥95th percentile compared to those with neck circumference <95th percentile, but the association was weaker for females [13]. Pubertal females have greater fat mass than males [14], but whether sex-related differences in neck fat exist for adolescents is unknown.

Adenotonsillectomy is recommended as the first-line treatment for adolescents with OSAS, including those who are obese [3]. However, adenotonsillectomy is less effective in obese children, in whom OSAS resolves in only 24%–67% of cases [15–18]. Work by our group and others has shown that OSAS in obese adolescents is mediated by increased adenotonsillar volume that reduces nasopharyngeal airway volume, as well as by neuromotor factors [19–21]. It is unknown whether the distribution of fat in the neck is associated with OSAS in obese adolescents. In such patients, neck fat may contribute to greater neck circumference and reduce airway patency, increasing OSAS risk and its persistence despite adenotonsillectomy. Understanding the contribution of neck fat to OSAS could have implications for treatments in adolescents, including weight loss, adenotonsillectomy, and potential pharmacotherapies. Results may also extend our understanding of OSAS-related complications to metabolic syndrome and the evolution of OSAS into adulthood.

Thus, in this study we compared measures of the amount of fat in the neck among obese adolescent males and females with OSAS and obese and non-obese adolescents without OSAS. We hypothesized that obese adolescents with OSAS would have greater amount of neck fat than adolescents without OSAS and that neck fat would correlate with both obesity and clinically-measured neck circumference. In addition, we hypothesized that females would have greater neck fat than males since, in general, adolescent females accumulate more fat than males. Within the OSAS group, we hypothesized that neck fat would correlate with OSAS severity.

Methods

Study design and participants

This study compared obese adolescents aged 12–16 years with OSAS (OSAS group), obese adolescents without OSAS (obese control group), and normal-weight adolescents without OSAS (normal-weight control group) recruited for a comprehensive study examining the pathophysiology of OSAS in adolescents. Some components of this study have been published previously, including the role of both upper airway collapsibility and structural contributors (e.g. the tonsils and adenoids) in OSAS among these patients [19, 20, 22, 23]. Obesity was defined as BMI Z-score greater than 1.64 and normal-weight as BMI Z-score between −1.64 and 1.04, based on sex-specific growth curves. OSAS was defined as obstructive apnea hypopnea index (OAHI) ≥5 events/h and no OSAS was defined as OAHI <1.5 events/h. Participants with OAHI between 1.5 and 5 events/h were excluded to allow for a clear separation between groups. Adolescents with OSAS were recruited from the Children’s Hospital of Philadelphia Sleep Center. Control participants were recruited from the community through advertisement and through the Healthy Weight program, a weight management program for obese patients. Snoring or any other symptoms of OSAS were exclusion criteria for control participants. Adolescents with a history of major chronic medical illness including uncontrolled asthma, as well as a history of craniofacial condition or genetic syndrome were excluded; previous adenotonsillectomy was an exclusion criterion. Informed consent from parents or guardians and assent from participants were obtained. The study was approved by the Children’s Hospital of Philadelphia institutional review board.

Assessments

Height and weight were measured in triplicate and averaged; BMI was calculated. Neck circumference measurements just inferior to the laryngeal prominence were made in triplicate using non-stretchable fiberglass tape and averaged [24]. The ratio of neck circumference to height was calculated. Sex-specific height-for-age Z-scores, weight-for-age Z-scores, and BMI-for-age Z-scores (BMI-Z) were calculated using the Centers for Disease Control and Prevention (CDC) 2000 growth charts [25]. Tanner staging of breast/genitalia and pubic hair development to assess sexual maturity in these pubertal adolescents was obtained based upon a validated self-assessment form [26]. Participants underwent in-laboratory attended, overnight polysomnography performed using standard pediatric recording and scoring techniques [27].

Upper airway magnetic resonance imaging (MRI) was performed using a 3 T scanner (MAGNETOM Sonata; Siemens, Malvern, PA) equipped with a prototype enhanced gradient system. Using Amira 4.1.2 image analysis software (Visage Imaging, San Diego, CA), fat regions were segmented manually with automated thresholding tools. Fat regions on each 5 mm slice of the sagittal upper airway T1-weighted MRI scans were analyzed and volumetric analysis was performed. Total neck length was calculated as the number of MRI slices times slice thickness (5 mm). Measures of cross-sectional area were quantified by dividing volume by total neck length. The neck region was delineated superiorly by C3–C4 and inferiorly by C7–T1, and multiple fat areas were identified by pixel valuation within this region (see Figure 1, panel A), including the intra-neck fat (all fat deep to the superficial fascia) and the subcutaneous fat both anterior and posterior to the visceral structures. Submental fat was segmented separately and was defined as extending from borders that included the mandible superiorly, the gonion posteriorly, and atlas-hyoid line inferiorly (see Figure 1, panel B). In addition, the overall region was sub-divided into superior and inferior regions at the midpoint (C5–C6) for additional analysis. The investigator performing MRI segmentation (CC) was blinded to the results of the polysomnogram and participant group. The investigator was validated against a gold standard set and segmentations were evaluated by a separate analyst. Re-analyses were performed in a subsample of n = 30 to confirm results for all segmented regions were within 5% of a gold standard.

Figure 1.

Panel A: Mid-sagittal, axial, and three-dimensional cutaway views showing segmentation of subcutaneous, internal, and submental fat in an obese adolescent male with obstructive sleep apnea syndrome. Subcutaneous and internal fat are segmented between C3 and C7; submental fat is segmented in its entirety. Mandible shown in three-dimensional view for reference. Panel B: Mid-sagittal view and three-dimensional reconstruction showing segmentation of submental fat and boundaries. The inferior edge of the mandible forms the superior border throughout the region. Posterior boundaries at the gonion and hyoid-atlas line extend laterally throughout the region.

Data analysis

Categorical variables were summarized as frequencies and percentages and compared among groups using chi-squared tests. For demographics and clinical measures, including neck circumference, continuous variables were summarized as means and standard deviations (SDs) and compared among groups using ANOVA and between groups using Student’s t-test. There were no meaningful differences in p-values when examined using complementary non-parametric tests. Comparisons of neck fat measures were adjusted for race and Tanner stage; comparisons between groups were also adjusted for sex and comparisons between males and females were adjusted for group and BMI. Associations between neck fat and demographic and clinical variables were performed using Pearson correlations, controlled for sex, race, Tanner staging, and OAHI. Correlations between neck fat and polysomnographic variables (OAHI, SpO₂ nadir, and percent sleep time with SpO₂ < 90%) were performed using Pearson correlations, controlled for sex, race, and Tanner staging. Statistical significance for associations with neck fat volume or cross-sectional area was based on a p < 0.01, Bonferroni-corrected for five individual fat measures within each domain (total, anterior subcutaneous, posterior subcutaneous, intra-neck and submental) and a p < 0.0167 in pairwise comparisons between patient groups (corrected for three pairwise comparisons). Similarly, statistical significance for superior and inferior neck fat volume and cross-sectional area was based on a p < 0.0125, Bonferroni-corrected for four individual fat measures (total, anterior subcutaneous, posterior subcutaneous, and intra-neck). A p < 0.05 was considered nominal evidence in these analyses, and statistically significant in all other comparisons. The available sample of 39 obese OSAS, 34 obese controls and 46 normal-weight controls provided between 81%–87% power to detect a moderately large [28] standardized mean difference of 0.675 between groups at a nominal α = 0.05. For correlation analyses, we had 92% power for a moderate correlation of 0.30 in analyses including all adolescents (n = 119) and 91% power for a correlation of 0.37 in analyses restricted to OSAS (n = 73). Finally, when comparing females (n = 38) and males (n = 81), we had 91% power for a standardized difference of 0.65 at a nominal α = 0.05. Non-significant results for smaller effect sizes should be interpreted more cautiously due to lower statistical power. Statistical analyses were performed using Stata versions 14.0 and 15.0 (College Station, TX).

Results

Study sample

A total of 119 adolescents participated in the study: 39 obese adolescents with OSAS (mean ± SD OAHI of 20.6 ± 27.9 events/h), 34 obese adolescents without OSAS (0.5 ± 0.3 events/h), and 46 normal-weight adolescents without OSAS (0.4 ± 0.4 events/h) (see Table 1). Obstructive apnea index was <1 event/h in all control participants. No differences in age, sex, race, ethnicity, or Tanner staging, were found among groups. Tanner staging was not completed in 14 participants, but no age difference was found in this group (14.8 ± 2.0 years) compared to the group who had Tanner staging (14.6 ± 1.5 years, p = 0.39). No difference in OAHI was found between the normal-weight (0.4 ± 0.4/h) and obese control groups (0.5 ± 0.3/h, p = 0.98). There was no difference between obese adolescents with OSAS compared to obese controls in terms of weight (100.8 ± 24.8 kg vs. 97.3 ± 26.6 kg, respectively, p = 0.49) and BMI (36.8 ± 8.2 kg/m² vs. 34.3 ± 5.9 kg/m², respectively, p = 0.08).

Table 1.

Comparison of demographics and clinical measures for each of the three study groups

						Pairwise comparisons‡
Variable	Overall (n = 119)	Obese OSAS (n = 39)	Obese control (n = 34)	Normal-weight control (n = 46)	Overall p-value†	Obese OSAS vs. obese control	Obese OSAS vs. NW control	Obese control vs. NW control
Age, years	14.6 ± 1.4	14.6 ± 1.4	14.4 ± 1.5	14.8 ± 1.5	0.54	–	–	–
Female, n (%)	38 (31.9)	11 (28.2)	14 (41.2)	13 (28.3)	0.39	–	–	–
Race, n (%)					0.27	–	–	–
Black	90 (75.6)	29 (74.4)	29 (85.3)	32 (69.6)
White	24 (20.2)	7 (18.0)	4 (11.8)	13 (28.3)
Multiple races	5 (4.2)	3 (7.7)	1 (2.9)	1 (2.2)
Hispanic ethnicity, n (%)	13 (10.9)	3 (7.7)	3 (8.8)	7 (15.2)	0.49	–	–	–
Weight (kg)	82.1 ± 30.3	100.8 ± 24.8	97.3 ± 26.6	55.0 ± 11.9	<0.001	0.49	<0.001	<0.001
Weight, Z-score	1.51 ± 1.50	2.6 ± 0.7	2.5 ± 0.6	0.0 ± 0.9	<0.001	0.49	<0.001	<0.001
Height (cm)	165.3 ± 10.0	165.6 ± 8.5	166.5 ± 10.2	164.1 ± 11.1	0.56	–	–	–
Height Z-score	0.18 ± 1.14	0.3 ± 0.9	0.6 ± 1.2	0.0 ± 1.0	0.03	0.17	0.15	0.01
BMI (kg/m2)	29.7 ± 9.6	36.8 ± 8.2	34.3 ± 5.9	20.2 ± 2.6	<0.001	0.08	<0.001	<0.001
BMI, Z-score	1.5 ± 1.2	2.4 ± 0.4	2.3 ± 0.3	0.2 ± 0.9	<0.001	0.47	<0.001	<0.001
Tanner stages, n (%)§					0.95	–	–	–
Early (Tanner 1–3)	28 (23.5)	8 (25.0)	7 (25.9)	13 (28.3)
Late (Tanner 4–5)	77 (64.7)	24 (75.0)	20 (74.1)	33 (71.7)
OAHI (events/h)	7.0 ± 18.5	20.6 ± 27.9	0.5 ± 0.3	0.4 ± 0.4	<0.001	<0.001	<0.001	0.98
Neck circumference (cm)	37.2 ± 3.9	39.3 ± 3.0	38.3 ± 3.5	34.5 ± 3.2	<0.001	0.18	<0.001	<0.001
Neck:height ratio	0.225 ± 0.018	0.238 ± 0.014	0.230 ± 0.016	0.210 ± 0.011	<0.001	0.020	<0.001	<0.001

^† p-value from ANOVA comparing values among obese OSAS, obese control and normal-weight controls.

^‡Pairwise comparisons between groups performed when overall p-value <0.05.

^§Tanner staging not done in n = 7 OSAS and n = 7 obese controls.

Abbreviations: OAHI, obstructive apnea hypopnea index; BMI, body mass index; NW, normal weight. Results shown are n (%) or mean ± standard deviation.

Comparisons of demographics and clinical data between females and males are shown in Supplemental Table 1. Females were shorter than males (161.0 ± 6.0 cm vs. 167.3 ± 10.9 cm, respectively, p < 0.001) but had comparable weight 84.3 ± 33.1 kg vs. 81.1 ± 29.1 kg, respectively, p = 0.61), resulting in greater average BMI (32.4 ± 10.8 kg/m² vs. 28.4 ± 8.8 kg/m², respectively, p = 0.054). However, between females and males there were no differences in height Z-score (0.0 ± 0.9 vs. 0.3 ± 1.1, respectively, p = 0.11) or BMI Z-score (1.6 ± 1.2 vs. 1.5 ± 1.2, respectively, p = 0.54). Among those with OSAS, there were no differences in OAHI between females and males (6.8 ± 23.5 events/h vs. 7.2 ± 15.8 events/h, respectively, p = 0.82).

Associations with clinical measures of neck size

Neck circumference (p < 0.001) and the ratio of neck circumference to height (p < 0.001) both differed among patient groups in unadjusted comparisons (see Table 1). Obese adolescents with or without OSAS had significantly greater neck circumference than normal-weight controls. Between obese adolescents with and without OSAS, the ratio of neck circumference to height differed (0.238 ± 0.014 vs. 0.230 ± 0.016, respectively, p = 0.02), but not neck circumference alone (39.3 ± 3.0 cm vs. 38.3 ± 3.5 cm, respectively, p = 0.18). Differences remained similar after controlling for sex, race, and Tanner stage (see Table 2).

Table 2.

Adjusted comparisons of neck circumference and neck fat measures for each of the three study groups

						Pairwise comparisons†
Variable	Overall (n = 119)	Obese OSAS (n = 39)	Obese control (n = 34)	Normal- weight control (n = 46)	Overall p-value†	Obese OSAS vs. obese control	Obese OSAS vs. NW control	Obese control vs. NW control
Neck circumference (cm)	37.2 (36.6, 37.7)	39.1 (38.2, 40.1)	38.5 (37.5, 39.5)	34.5 (33.6, 35.4)	<0.001	0.35	<0.001	<0.001
Neck:height ratio	0.225 (0.222, 0.227)	0.237 (0.233, 0.242)	0.230 (0.226, 0.235)	0.210 (0.206, 0.214)	<0.001	0.03	<0.001	<0.001
Volume (cm3)
Total	247.2 (219.2, 275.1)	373.8 (324.5, 423.1)	339.9 (286.5, 393.2)	71.2 (24.9, 117.6)	<0.001	0.35	<0.001	<0.001
Anterior subcutaneous	64.0 (54.4, 73.5)	97.6 (80.7, 114.4)	86.0 (67.8, 104.2)	19.3 (3.4, 35.1)	<0.001	0.36	<0.001	<0.001
Posterior subcutaneous	112.6 (99.4, 125.9)	178.2 (154.9, 201.6)	150.5 (125.3, 175.8)	29.0 (7.1, 50.9)	<0.001	0.11	<0.001	<0.001
Intra-neck	58.6 (51.5, 65.6)	81.3 (68.9, 93.7)	87.0 (73.6, 100.4)	18.3 (6.7, 29.9)	<0.001	0.54	<0.001	<0.001
Submental	39.8 (36, 43.6)	63.0 (56.3, 69.8)	53.5 (46.2, 60.8)	10.0 (3.6, 16.3)	<0.001	0.06	<0.001	<0.001
Cross-sectional area (cm2)
Total	25.5 (22.2, 28.7)	36.1 (30.4, 41.7)	37.7 (31.3, 44.2)	7.6 (2.2, 13.0)	<0.001	0.70	<0.001	<0.001
Anterior subcutaneous	6.5 (5.5, 7.5)	9.3 (7.6, 11.0)	9.5 (7.5, 11.4)	2.0 (0.4, 3.7)	<0.001	0.90	<0.001	<0.001
Posterior subcutaneous	11.6 (10.1, 13.1)	17.2 (14.5, 19.8)	16.7 (13.7, 19.7)	3.1 (0.6, 5.7)	<0.001	0.82	<0.001	<0.001
Intra-neck	6.0 (5.2, 6.9)	7.9 (6.5, 9.4)	9.6 (7.9, 11.2)	1.9 (0.5, 3.3)	<0.001	0.14	<0.001	<0.001
Submental	4.1 (3.6, 4.6)	6.2 (5.3, 7.0)	5.8 (4.9, 6.7)	1.1 (0.4, 1.9)	<0.001	0.57	<0.001	<0.001
Total neck length (cm)	10.25 (10.02, 10.48)	10.50 (10.11, 10.90)	10.07 (9.61, 10.52)	10.16 (9.78, 10.54)	0.293	–	–	–

Statistically significant associations after Bonferroni correction shown in bold.

^† p-value comparing values among obese OSAS, obese control and normal-weight controls, adjusted for sex, race, and Tanner staging.

^‡Pairwise comparisons between groups performed when overall p-value <0.05.

Abbreviations: OAHI, obstructive apnea hypopnea index; BMI, body mass index; NW, normal weight. Results shown are model estimated adjusted mean and 95% confidence interval.

Males had greater neck circumference than females (37.7 ± 3.8 cm vs. 36.0 ± 3.8 cm, respectively, p = 0.02), but ratio of neck circumference to height was not different by sex (0.225 ± 0.017 cm vs. 0.224 ± 0.021 cm, respectively, p = 0.65) in unadjusted analyses (see Supplemental Table 1). After adjusting for patient group, race, Tanner staging and BMI, differences in both neck circumference (mean [95% CI] difference = 3.34 [2.34, 4.34]; p < 0.0001) and the ratio of neck circumference to height (0.007 [0.003, 0.012]; p=0.001) became statistically significant, with larger values in males than females (see Table 3).

Table 3.

Adjusted comparisons of neck circumference and neck fat measures between female and male participants

Variable	Female (n = 38)	Male (n = 81)	p †
Neck circumference (cm)	34.9 (34.1, 35.7)	38.2 (37.7, 38.8)	<0.001
Neck:height ratio	0.220 (0.216, 0.223)	0.227 (0.225, 0.230)	0.001
Volume (cm3)
Total	290.0 (250.6, 329.4)	227.1 (200.7, 253.4)	0.012
Anterior subcutaneous	74.0 (60.3, 87.7)	59.3 (50.1, 68.4)	0.09
Posterior subcutaneous	136.7 (116.5, 156.9)	101.4 (87.9, 114.9)	0.006
Intra-neck	65.5 (54.4, 76.5)	55.3 (47.9, 62.7)	0.14
Submental	43.4 (38.1, 48.8)	38.1 (34.5, 41.7)	0.11
Cross-sectional area (cm2)
Total	33.3 (28.3, 38.3)	22.0 (18.8, 25.2)	<0.001
Anterior subcutaneous	8.4 (7.0, 9.9)	5.7 (4.8, 6.6)	0.003
Posterior subcutaneous	15.5 (12.9, 18.0)	9.9 (8.2, 11.5)	0.0005
Intra-neck	7.7 (6.4, 9.1)	5.3 (4.4, 6.2)	0.005
Submental	4.9 (4.2, 5.7)	3.7 (3.3, 4.2)	0.0097
Total neck length (cm)	9.42 (8.97, 9.86)	10.62 (10.34, 10.91)	<0.001

Statistically significant associations after domain-specific Bonferroni correction shown in bold.

^† p-value comparing female and male participants, adjusted for group (normal-weight control, obese control, obese OSAS), race, Tanner staging and BMI. Results shown are model estimated adjusted mean and 95% confidence interval.

As illustrated in Supplemental Table 2, there were strong positive correlations between amount of neck fat and both neck circumference and neck-to-height ratio. Neck circumference and ratio of neck circumference to height both strongly correlated with total neck fat volume (Pearson’s rho = 0.53 and 0.63, respectively) and average total neck fat volume per slice (0.48 and 0.62, respectively), controlling for sex, race, Tanner stage, and OAHI (all p < 0.0001).

Associations with MRI-measured neck fat volumes

Neck fat measurements were significantly lower in the normal-weight control group compared to both the obese OSAS and obese control groups (see Table 2). However, no differences in neck fat volumes were found between the OSAS group and obese control group (see Supplemental Figure 1; only submental fat volume approached significance {p = 0.06}). When sub-dividing into superior and inferior neck fat, there were still no differences in the neck fat volumes between the OSAS and obese control groups (see Supplemental Table 3) In the OSAS group, neck fat volumes were generally not associated with measures of OSAS severity, including OAHI, saturation nadir, or log-transformed time with saturation below 90% (see Supplemental Table 4); moderate correlations were observed between lower SpO₂ nadir and increases in both total submental fat volume (rho = −0.34, p = 0.07) and average submental volume per slice (rho = −0.40, p = 0.04). Thus, overall, neck fat does not appear to significantly associate with risk of OSAS or OSAS severity among obese adolescents.

When comparing overall neck fat volumes based on sex, females had significantly more total and posterior subcutaneous neck fat than males (see Table 3 and Supplemental Figure 2); differences in anterior subcutaneous, intra-neck and submental neck fat were in the same direction but did not reach statistical significance. Moreover, females had significantly larger cross-sectional areas (these data control for neck length) for each neck fat measurement. Females also demonstrated significantly shorter neck length (p < 0.0001). Similar results were found when sub-dividing into superior and inferior neck fat (see Supplemental Table 5). Thus, results are consistent with increased neck fat in female adolescents, independent of patient group, race, Tanner staging, and BMI (see Figure 2).

Figure 2.

Mid-sagittal T1 MR image comparing subcutaneous neck fat thickness and volume in an obese male with OSAS compared with an obese female with OSAS who has a similar neck circumference.

Discussion

This study compared neck fat in obese and normal-weight adolescents and assessed its relationship with neck circumference, OSAS and sex. We found that measures of neck fat were strongly associated with obesity in these adolescents, but not with OSAS status or severity. Results also showed that adolescent females had significantly more neck fat than males, despite a smaller neck circumference and a shorter neck.

Overall, our results suggest that in obese adolescents, factors other than neck fat contribute to OSAS risk. We previously showed that reduced neuromotor tone and increased tonsil and adenoid size are associated with OSAS status and severity [19, 20]. While measures of neck fat were not associated with OSAS, some relationship between ratio of neck circumference to height and OSAS may exist. In our cohort, obese adolescents with OSAS had similar neck circumferences compared to obese controls, but a greater ratio of neck circumference to height. This finding is consistent with that reported by Narang et al. [9], which found that the ratio of neck circumference to height to be the only anthropometric measure that differed between obese children with and without OSAS. Similarly, in a study by Katz et al. [13], no difference in neck circumference between children with OSAS and controls was present, although the proportion of participants with neck circumference above the 95th percentile was greater in the OSAS group compared to controls.

This study also identified the novel finding of greater neck fat in adolescent females compared to males. Our data show that in adolescents, neck fat primarily appears to correlate with sex-related differences or obesity. This difference is not surprising, as females have a greater fat mass index compared to males independent of obesity status, and an age-dependent increase in fat is well-recognized in females during adolescence that does not occur in males [29]. In our study, the greatest differences in neck fat were found in the posterior subcutaneous region, which is unlikely to directly affect the airway and, thus, may not impact OSAS. Interestingly, despite having a greater neck circumference and longer necks, males had less neck fat, likely due to the greater volume of other structures, including muscle mass [30].

Although neck fat was not associated with OSAS in obese adolescents, it may be a risk factor for OSAS during adulthood, when tonsils and adenoids are smaller in relation to the airway volume [31]. Robust data confirm fat and increased soft tissue volume in the airway can contribute to OSAS in adults [32, 33]. Specifically, adults with OSAS have larger soft tissue structures surrounding the upper airway, including the volume of the lateral pharyngeal walls and tongue, and greater tongue fat volume compared to adults without OSAS. However, MRI studies of adolescents have shown that OSAS status and severity correlates with adenotonsillar size rather than airway fat or other soft tissue structures [20, 21], suggesting similarities with children rather than adults. Additional studies assessing the relationship between neck fat, neck circumference, and OSAS in adults are needed.

Neck fat may be associated with other important health outcomes beyond OSAS. An imaging study of 303 adults found that posterior and subcutaneous neck adipose tissue was associated with increased triglycerides and risk for metabolic syndrome, especially in females [34]. In that study, neck fat was associated with obesity, and subcutaneous neck fat was greater in adult women than men. Data from the Framingham Heart Study have demonstrated that neck circumference is a risk factor for cardiovascular disease in adults [35], independent of BMI and visceral adiposity. Additional studies in children and adolescents are needed to determine if neck fat, which could be a manifestation of ectopic fat distribution, contributes to other medical problems in youth.

There were several important limitations to this study. The cohort was primarily African-American, and relationships could differ depending on race or ethnicity, as both female and male African-Americans have greater fat-free mass compared to Caucasians [36]. Relatively fewer females were included in this cohort which also precluded specific analysis of females based on OSAS and obesity status. The cross-sectional study design precluded analysis of the role of neck fat or other anatomic factors on clinical outcomes. Moreover, the study design did not recruit normal-weight adolescents with OSAS, so comparison of neck fat with that group could not be made. Future studies are needed to determine the structural contributors to OSAS in non-obese adolescents.

In summary, neck fat is a marker of obesity and correlates with neck circumference, but is not associated with OSAS status or severity in obese adolescents. Confirming previous studies, obese adolescents with OSAS had greater ratio of neck circumference to height than obese controls. Adolescent females demonstrated more neck fat than males, regardless of OSAS status. More research is needed to determine whether regional fat distribution, including neck fat, poses health risks to adolescents and whether neck fat is related to OSAS in adults. In addition, future studies should assess the contribution of tongue size to OSAS in children and adolescents, including enlargement from fat deposition or congenital conditions.

Funding

This work was supported by the National Institutes of Health (NIH R01 HL058585, NIH K23 HL135346).

Disclosure Statement

None declared.