The association of obstructive sleep apnea with dyslipidemia in Korean children and adolescents: a single-center, cross-sectional study

Abstract

Study Objectives:

To evaluate whether obstructive sleep apnea (OSA) and its severity are related to dyslipidemia and alanine transaminase elevation as a marker of nonalcoholic fatty liver disease in children.

Methods:

The data collected from polysomnography, laboratory measurements (lipid profile and liver enzyme), and body mass index in children aged 0–18 years who visited the pediatric department between 2012 and 2018 were retrospectively analyzed.

Results:

There were a total of 273 participants in the study (ages 0–6 years, 7–12 years, and 13–18 years: 61.9%, 26.4%, and 11.7%, respectively). In the ages 7–12 and 13–18 years groups, obesity was strongly associated with OSA severity (Cramer’s V = 0.498, P < .001). High-density lipoprotein cholesterol levels were significantly lower in the OSA group than in the non-OSA group, irrespective of the presence of obesity. In addition, high-density lipoprotein cholesterol levels were significantly different between the OSA severity groups after adjusting for body mass index (P = .000). In participants who were obese, moderate and severe OSA were associated with alanine transaminase elevation (P = .023 and P = .045, respectively).

Conclusions:

This study suggests that OSA may be an independent risk factor for dyslipidemia and that OSA and obesity have a synergistic effect on alanine transaminase elevation. Early diagnosis and treatment of OSA from childhood, especially in obese children, will reduce metabolic complications.

Citation:

Kang EK, Jang MJ, Kim KD, Ahn YM. The association of obstructive sleep apnea with dyslipidemia in Korean children and adolescents: a single-center, cross-sectional study. J Clin Sleep Med. 2021;17(8):1599–1605.

BRIEF SUMMARY

Current Knowledge/Study Rationale: The association between obstructive sleep apnea (OSA) and metabolic syndrome has mainly been evaluated in adults, and there have been few studies in children.

Study Impact: High-density lipoprotein cholesterol was lower in the OSA group than in the non-OSA group and was significantly different between OSA severity groups after adjusting for body mass index. This result suggests that OSA may independently contribute to dyslipidemia.

INTRODUCTION

Obstructive sleep-disordered breathing (SDB) is a syndrome of upper airway dysfunction during sleep, and the most severe clinical entity is obstructive sleep apnea (OSA). ¹ The leading cause of pediatric OSA is adenotonsillar hypertrophy, and obesity is an independent risk factor. ¹ In a community-based study, ² 46.6% of obese children (mean age, 10.8 years) had OSA (obstructive apnea-hypopnea index [AHI] > 1 event/h), and 19% of obese children (mean age, 11.2 years) recruited from the obesity clinic ³ had OSA (obstructive AHI ≥ 2 events/h). The prevalence of obesity across all pediatric age groups has increased globally and substantially. This rising incidence impacts the prevalence of pediatric OSA. ⁴ In Korea, obesity in children increased from 6.8% in 1998–10.0% in 2013. ⁵ Obesity and OSA share common comorbidities, such as cardiometabolic/cardiovascular disease and nonalcoholic fatty liver disease (NAFLD). ⁶

Some experimental and clinical data suggest that dyslipidemia may be associated with OSA because of intermittent hypoxia and sleep fragmentation. ^7,8 The potential mechanism by which OSA may affect lipid metabolism (as suggested by animal models) is that OSA and intermittent hypoxia increase adipose tissue lipolysis and liver biosynthesis during the fasting state, whereas lipoprotein clearance is delayed in the postprandial state. ⁷ Increased sympathetic tone and oxidative stress can also have adverse effects on lipid metabolism.

The association between OSA, metabolic syndrome including dyslipidemia, ⁷ and cardiovascular consequences ⁹ has mainly been evaluated in adults, and there have been few studies in children and adolescents. ^10–12 Moreover, there are few studies on the relationship between OSA, obesity, and dyslipidemia in Korean children. The objectives of this study were to (1) analyze the effect of children’s age on obesity and OSA, (2) evaluate whether OSA and the severity of OSA are associated with dyslipidemia independent of obesity, and (3) determine the effect of obesity and OSA on alanine transaminase (ALT) elevation in children and adolescents.

METHODS

Study participants and data collection

The data collected from the polysomnography, laboratory measurements (lipid profile and liver enzyme), body mass index (BMI), and z score of the children ages 0–18 years who visited the pediatric department from 2012–2018 were retrospectively analyzed. Overnight polysomnography (Grael system) was scored according to the 2012 American Academy of Sleep Medicine pediatric criteria and adult criteria for patients aged 13–18 years. Participants were divided into 3 age groups: ages 0–6, 7–12, and 13–18 years. Obesity was defined by a BMI ≥ 95th percentile for age and sex (Centers for Disease Control and Prevention criteria). The BMI z score was calculated using the CDC growth chart. OSA was diagnosed if the AHI was ≥ 1 event/h. Participants with OSA were divided into 3 groups according to the degree of AHI: those with 1 ≤ AHI < 5 events/h were included in the mild OSA group, those with 5 ≤ AHI < 10 events/h were included in the moderate OSA group, and those with AHI ≥ 10 events/h were included in the severe OSA group. Blood samples were drawn in the morning after polysomnography and 12-hour fasting. A lipid profile including total cholesterol (TC), triglycerides (Tg), direct low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), and Tg/HDL-C was collected to evaluate the risk for metabolic syndrome. ¹³ ALT was assessed to evaluate the risk of NAFLD as a surrogate marker and its association with obesity and OSA in the group aged 7–18 years: ALT ≥ 26 U/L for boys and ALT ≥ 22 U/L for girls. ^14,15 We chose this age group because in Korea the ALT level for the diagnosis of NAFLD is assessed mainly in older children ^16,17 and the obesity rate is low among young children. The participants were otherwise healthy children. Children with familial hyperlipidemia/hepatic disorder, those taking medications that could predispose to dyslipidemia/transaminase elevation, and adolescents with a recent history of alcohol ingestion or who were pregnant were excluded from the study. The study protocol was approved by the institutional review board of the Eulji Medical Center (IRB number: 2020-04-005).

Statistical analyses

All data are presented as mean ± standard deviation. A Pearson χ² test was used to determine the differences between the categorical variables. Cramer’s V was obtained as the effect size for the relationship between 2 categorical values for tables bigger than a 2 × 2 tabulation. The effect size of a Cramer’s V higher than 0.25 was considered to be a strong association. ¹⁸ In the analysis of the relationship between OSA severity and obesity, the odds ratio of each severity group was expressed relative to the non-OSA group. Likewise, in the analysis of the relationship between OSA severity and ALT elevation, the odds ratio of each severity group was expressed relative to the non-OSA group. We performed a 2-way analysis of variance to evaluate the interaction between OSA and obesity on lipid profiles. A 1-way analysis of variance was conducted to compare the OSA severity groups to show the baseline characteristics and lipid profiles of the study population (n = 191). After this analysis, a multivariate analysis of covariance was used to determine the relationship between OSA severity and lipid profile, adjusting for age and obesity. In all of the analyses, a P value of < .05 was considered to be statistically significant. All of the statistics were rounded off to the second decimal point. All statistical analyses were conducted using IBM SPSS Statistics version 20.0 (IBM Corporation, Armonk, NY).

RESULTS

OSA status according to obesity in 3 age groups

A total of 273 children and adolescents ages 0–18 years were included in the study ( Table 1 ). Among them, 169 were in the group aged 0–6 years (61.9%), 72 were in the group aged 7–12 years (26.4%), and 32 were in the group aged 13–18 years (11.7%). AHI was significantly higher in the group aged 13–18 years (P = .001). The BMI z score was significantly different among the 3 age groups (P < .001). Eighty-four (30.8%) patients were placed in the group of those who were obese, and 189 (69.2%) were placed in the group of those who were not obese. The participants who were obese were older (9.30 ± 3.62 vs 5.54 ± 3.64, P < .001) and had a higher AHI (8.57 ± 17.68 vs 3.47 ± 8.74, P = .014) than those who were not obese ( Table 1 ).

Table 1.

Baseline characteristics of the study population (n = 273).

Parameters	Ages 0–6 y Group (n = 169)	Ages 7–12 y Group (n = 72)	Ages 13–18 y Group (n = 32)	P	Nonobese Group (n = 189)	Obese Group (n = 84)	P
Age, y	4.04 ± 1.45	9.31 ± 1.52	14.84 ± 1.53	< .001	5.54 ± 3.64	9.30 ± 3.62	< .001
Sex, % boys (n)	72.78 (123)	63.89 (46)	71.88 (23)	.376	68.78 (130)	73.81 (62)	.401
BMI, kg/m2	16.38 ± 2.16	21.47 ± 4.63	26.88 ± 6.92	< .001	16.43 ± 2.21	24.63 ± 5.53	< .001
BMI z score	0.14 ± 1.55	1.72 ± 1.26	1.73 ± 1.26	< .001	–0.12 ± 1.11	2.68 ± 0.69	< .001
AHI, events/h	3.53 ± 7.76	5.31 ± 14.42	12.40 ± 21.85	.001	3.47 ± 8.74	8.57 ± 17.68	.014
SaO2 nadir, %	85.8 ± 7.46	86.66 ± 6.73	86.89 ± 7.43	.589	86.55 ± 7.47	85.27 ± 6.73	.179

Data are presented as means ± standard deviations, percentage (number). AHI = apnea-hypopnea index, BMI = body mass index, SaO₂ nadir = oxygen saturation nadir.

The prevalence of OSA was not significantly higher in the participants aged 0–6 years who were obese (P > .99) but was significantly higher in participants aged 7–12 years who were obese (odds ratio, 11.10; 95% confidence interval, 3.39–36.33; P < .001) and those aged 13–18 years who were obese (odds ratio, 8.57; 95% confidence interval, 1.43–51.36; P = .021). The association between OSA severity and obesity was determined by using a Pearson χ² test and Cramer’s V. In the group aged 0–6 years, there was no significant association between obesity and the OSA severity groups (Cramer’s V = 0.203, P = .072). However, in participants aged 7–12 years (Cramer’s V = 0.533, P < .001) and 13–18 years (Cramer’s V = 0.518, P = .035), obesity was associated with OSA severity. Specifically, the severe OSA group in both age groups was associated with obesity (P < .05; Table 2 ).

Table 2.

Comparison of the effect of obesity on the OSA severity between the different age groups.

Age Groups With OSA Severity	Obesity
	OR (95% CI)	P
0–6 years
Non-OSA	—	—
Mild	0.74 (0.26–2.11)	.596
Moderate	3.33 (0.74–14.99)	.104
Severe	2.50 (0.50–12.46)	.253
7–12 years
Non-OSA	—	—
Mild	10.80 (3.02–38.59)	< .001
Moderate	4.80 (0.80–28.90)	.073
Severe	32.40 (3.28–320.36)	< .001
13–18 years
Non-OSA	—	—
Mild	4.80 (0.68–33.80)	.104
Moderate	NA*	—
Severe	12.00 (1.29–111.32)	.020

The OR of each OSA severity group was expressed relative to the non-OSA group. *There was no moderate OSA in participants aged 13–18 years who were not obese. CI = confidence interval, NA = not applicable, OR = odds ratio, OSA = obstructive sleep apnea.

The effect of OSA status on lipid profile (n = 191)

The effect of OSA and obesity status on lipid profile

Among the 191 participants, 69.1% (n = 132) had OSA and 30.9% (n = 59) had no OSA. Of the 132 participants with OSA, 42.4% (n = 56) were obese and 57.6% (n = 76) were not obese. Among the non-OSA group, 20.3% (n = 12) were obese and 79.7% (n = 47) were not obese. A 2-way analysis of variance was performed to evaluate the interaction between OSA and obesity on the lipid profile. There was a significant difference in the HDL-C level and no difference in total cholesterol, Tg, Tg/HDL-C ratio, or LDL-C, according to the status of OSA and obesity. The HDL-C level was lowest in participants with OSA who were obese (45.80 mg/dL) and was highest in participants in the non-OSA group who were not obese (55.78 mg/dL; Figure 1 ). There was no interaction between OSA and obesity status on HDL-C (P = .975). HDL-C levels were significantly different according to the presence of OSA (P = .002) but did not differ according to obesity status (P = .114).

Figure 1. HDL-C levels according to OSA and obesity status (plot by 2-way analysis of variance).

HDL-C = high-density lipoprotein cholesterol, OSA = obstructive sleep apnea.

The effect of OSA severity on lipid profile

The mean ages of the severity groups were different (7.98 ± 3.88 vs 6.01 ± 3.85 vs 8.60 ± 3.83 vs 9.06 ± 4.15 for non-OSA, mild, moderate, and severe OSA, respectively; P = .001). The BMI z scores were also different among the groups (0.58 ± 1.24 vs 0.71 ± 1.64 vs 2.10 ± 1.56 vs 2.30 ± 1.41 for non-OSA, mild, moderate, and severe OSA, respectively; P < .001). The moderate and severe OSA groups had higher BMI profiles than the non-OSA and mild OSA groups ( Table 3 ).

Table 3.

Baseline characteristics and lipid profile of participants according to OSA severity (n = 191).

Parameters	Non-OSA (n = 59)	Mild (n = 99)	Moderate (n = 15)	Severe (n = 18)	P*
Age, y	7.98 ± 3.88	6.01 ± 3.85	8.60 ± 3.83	9.06 ± 4.15	.001
Sex, % boys (n)	66.10 (39)	68.69 (68)	80 (12)	66.67 (12)	.775
BMI, kg/m2	18.17 ± 3.80	18.57 ± 4.69	23.37 ± 6.86	25.99 ± 8.67	< .001
BMI z score	0.58 ± 1.24	0.71 ± 1.64	2.10 ± 1.56	2.30 ± 1.41	< .001
AHI, events/h	0.40 ± 0.24	2.07 ± 1.00	7.32 ± 1.46	32.98 ± 29.49	< .001
SaO2 nadir, %	90.75 ± 2.54	85.74 ± 6.19	84.55 ± 3.42	77.35 ± 11.10	< .001
TC, mg/dL	167.32 ± 26.99	166.46 ± 27.98	167.73 ± 25.15	173.28 ± 21.82	.807
Tg, mg/dL	94.36 ± 60.84	90.72 ± 53.97	127.20 ± 99.47	133.00 ± 99.62	.026
LDL-C, mg/dL	95.19 ± 25.26	103.18 ± 24.71	100.47 ± 26.46	106.28 ± 15.90	.194
HDL-C, mg/dL	55.12 ± 12.46	47.64 ± 11.47	50.27 ± 10.13	46.44 ± 9.65	.001
Tg/HDL-C	1.95 ± 1.75	2.13 ± 1.57	2.82 ± 2.58	3.13 ± 2.68	.059

Data are presented as means ± standard deviations, percentage (number). *A 1-way analysis of variance was applied to compare the means among the severity groups. AHI = apnea-hypopnea index, BMI = body mass index, HDL-C = high-density lipoprotein cholesterol, LDL-C = low-density lipoprotein cholesterol, OSA = obstructive sleep apnea, SaO₂ nadir = oxygen saturation nadir, TC = total cholesterol, Tg = triglycerides.

The serum Tg (P = .026) and HDL-C levels (P = .001) were significantly different among the OSA severity groups. A posthoc analysis showed a significant difference of mean HDL-C between the non-OSA and mild OSA group and between the non-OSA and severe OSA group ( Figure 2 ). The total cholesterol and LDL-C levels and the Tg/HDL-C ratio showed no significant difference among the OSA severity groups ( Table 3 ).

Figure 2. HDL-C levels according to OSA severity group.

*A significant difference between groups was found after a post hoc test (Bonferroni) of a multivariate analysis of covariance, adjusted for age and body mass index. HDL-C = high-density lipoprotein cholesterol, OSA = obstructive sleep apnea.

Because age, BMI, and z score were found to be different among the OSA groups according to the 1-way analysis of variance test, a multivariate analysis of covariance was conducted to identify the relationship between the lipid profile (Tg and HDL-C) and OSA severity after adjusting for age, BMI, and z score. The HDL-C levels were still significantly different among groups after adjusting for age, BMI, and z score (P = .000). On the other hand, the Tg levels were not different among the groups after adjusting for age and BMI profiles (P = < .001; Table 4 )

Table 4.

A multivariate analysis of covariance summary for OSA severity groups and lipid profile, adjusted for age, BMI, and z score.

Source	df	Mean Squares	F	P	Partial η2
BMI
Tg	1	4,032.38	1.01	.316	0.005
HDL-C	1	554.63	4.49	.035	0.024
BMI z score
Tg	1	3,874.90	0.97	.326	0.005
HDL-C	1	410.28	3.32	.070	0.018
Age
Tg	1	50,645.11	12.69	.000	0.065
HDL-C	1	1,586.67	12.85	.000	0.065
OSA severity groups
Tg	3	7,962.09	1.99	.116	0.031
HDL-C	3	886.77	7.18	.000	0.105

BMI = body mass index, df = degrees of freedom, HDL-C = high-density lipoprotein cholesterol, OSA = obstructive sleep apnea, Tg = triglycerides.

The effect of OSA severity on ALT elevation according to obesity status in ages 7–18 years (n = 102)

As shown in Figure 3 , there was no significant association between OSA severity and ALT elevation in participants who were not obese (P > .05). However, ALT elevation showed a strong association with OSA severity in participants who were obese (Cramer’s V = 0.386, P = .032). Among them, the moderate OSA (P = .023) and the severe OSA groups (P = .045) were at higher risk of ALT elevation. We compared the risk of ALT elevation between non-OSA and mild OSA and moderate-to-severe OSA according to obesity. The moderate-to-severe OSA group had a 3.43-fold increase in the risk of ALT elevation relative to the non-OSA and mild OSA group among participants who were obese (95% confidence interval, 1.1–10.72; P = .031).

Figure 3. The effect of OSA severity on alanine transaminase in the group aged 7–18 years (n = 102).

The OR of each severity group was expressed relative to the non-OSA group. CI = confidence interval, OR = odds ratio, OSA = obstructive sleep apnea.

DISCUSSION

This study showed that OSA and its severity were associated with lower HDL-C levels after adjusting for obesity in Korean children and adolescents, suggesting that OSA may be an independent risk factor for dyslipidemia. It is speculated that obesity and SDB converge to disrupt normal lipid homeostasis. ⁸ Furthermore, there is evidence that OSA is independently associated with metabolic syndrome. ¹⁹ In children with OSA who were obese, there was a significant association between SDB and metabolic abnormalities after adjusting for BMI or adiposity. ^10,20 In another study, children who were not obese with moderate-to-severe OSA had lower levels of HDL-C than those with primary snoring/mild OSA. ¹² Similarly, our study supports this finding. HDL-C levels were lower in the OSA group than in the non-OSA group, independent of obesity. In addition, HDL-C levels were significantly different between the OSA severity groups after adjusting for BMI and z score.

If OSA contributes to metabolic syndrome, then the treatment of OSA will theoretically improve metabolic abnormalities. For example, OSA treatment changed metabolic markers significantly in a multicenter study of Spanish children aged 3–14 years who were obese. ²¹ The lipid profile improved after adenotonsillectomy in the subgroup of children with moderate to severe OSA whose SDB was completely resolved, with no decrease in BMI after treatment. Similarly, another study showed that adenotonsillectomy in children with OSA resulted in significant improvements in the lipid profile, C-reactive protein, and apolipoprotein B, suggesting a role for OSA in lipid homeostasis and systemic inflammation. ²² These results also indicates that OSA could be involved in metabolic dysfunction independent of obesity.

Although obesity increases the risk of OSA, there may be a difference in the association between obesity and OSA among children and adolescents. ²³ In 1 study, the risk of OSA among adolescents (ages ≥ 12 years) increased with an increase in BMI, whereas no significant increase with increasing BMI was seen among younger children. ²³ Similarly, in this study, the rate of OSA was significantly higher in the older participants who were obese (ages 7–12 and 13–18 years) and not higher in preschool children who were obese (ages 0–6 years). Furthermore, OSA severity was strongly associated with the older participants who were obese. Although the prevalence of OSA among preschool children who were obese (ages 0–6 years) evaluated for SDB was reported to be as high as 36.6%, ²⁴ the association between OSA and obesity seems to differ with age.

The estimated prevalence of NAFLD in Korean children and adolescents aged 10–18 years was 5.5% in 2010 and 7.1% in 2015 when using the same ALT cutoffs used in this study (BMI ≥ 85th percentile plus ALT > 25.8 U/L for boys and > 22.1 U/L for girls). ¹⁶ This increasing prevalence of NAFLD suggests an increasing level of obesity and a possible increase of OSA among children. Some observational studies have shown that pediatric patients with NAFLD and OSA/hypoxia had more advanced liver disease and fibrosis than those without OSA/hypoxia. ^25,26 Treating OSA and chronic intermittent hypoxia with continuous positive airway pressure in children with NAFLD reduced the ALT level. ²⁷ In our study, ALT elevation showed a strong association with OSA severity in participants who were obese, suggesting OSA severity as a risk factor for NAFLD in that group. There may be a synergistic effect between obesity and OSA severity on NAFLD. It would be significant to analyze the association of ALT elevation with OSA as a marker of NAFLD in young children.

There are some limitations in this study. First, prepubertal and pubertal children were mixed in the analysis of lipid profiles because pubertal status was not assessed. Nonetheless, there was an association between OSA and dyslipidemia after adjusting for obesity. Second, we used only the ALT level to assess the risk of NAFLD; therefore, we could not evaluate the association between OSA and NAFLD. Third, other metabolic markers, including glucose/insulin, hypertension, and central adiposity (waist circumference), were not evaluated. Therefore, additional research is needed to assess the interactions between all those factors in children. The presence of adenotonsillar hypertrophy or adenotonsillectomy was not accurately evaluated in this retrospective study. These factors would partially contribute to the development of OSA in children who are obese.

CONCLUSIONS

OSA and its severity were associated with lower HDL-C levels after adjusting for obesity. Higher-severity OSA showed a strong association with ALT elevation in children who were obese. These findings suggest that OSA and its severity may be independent risk factors for dyslipidemia and that OSA and obesity may have a synergistic effect on ALT elevation. Early diagnosis and treatment of OSA in childhood, especially in children who are obese, will reduce these metabolic complications.

DISCLOSURE STATEMENT

All authors have seen and approved the manuscript. Work for this study was performed at Nowon Eulji Medical Center, Seoul, Korea. The authors report no conflicts of interest.

ABBREVIATIONS

AHI

apnea-hypopnea index

ALT

alanine transaminase

BMI

body mass index

HDL-C

high-density lipoprotein cholesterol

LDL-C

low-density lipoprotein cholesterol

NAFLD

nonalcoholic fatty liver disease

OSA

obstructive sleep apnea

SDB

sleep-disordered breathing

Tg

triglycerides