The relationship between metabolic syndrome and obstructive sleep apnea syndrome: a nationwide population-based study

There has been a need for research on the association between metabolic syndrome (MetS) and obstructive sleep apnea syndrome (OSAS) using large data such as nationwide population-based data that adjusts important confounding factors. Therefore, we investigated the relationship between MetS and OSAS. The data source we used was the National Health Insurance Service claims database managed by the Republic of Korea government, in which 10,113,560 individuals were enrolled in 2009 and followed up until 2018. The independent association of MetS with the risk of OSAS was determined using a Cox proportional hazards model with adjustment for age, sex, smoking status, alcohol consumption, regular physical exercise, and body mass index. Our results showed that MetS were strongly associated to OSAS which was adjusted for several confounding factors. Also, we found men, increased waist circumference and increased triglyceride are important risk factors for OSAS.

Clinical parameters and diagnostic criteria. Baseline data were collected on the following variables associated with the risk of OSAS: age (years), alcohol consumption (in 1 week and on a single occasion; 30 g/ day = heavy drinkers 19 ), income level, sex, and smoking status. The regular exercise has been defined as vigorous exercise at least 3 days a week or moderate exercise at least 5 days a week. BMI was calculated by dividing the subject's weight (kilograms) by the square of the subject's height (meters). According to the criteria for the Asia-Pacific region, participants with a BMI ≥ 25 kg/m 2 were considered obese 20 . Diabetes mellitus (DM) was defined as International Classification of Disease 10th Revision (ICD-10) codes E11-14, plus at least one prescription of antidiabetic medication per year or a fasting glucose level ≥ 7 mmol/L (data obtained from the health database). Hypertension was indicated by a prescription (at least once a year) of a antihypertensive agent under ICD-10 codes I10-I15, or systolic blood pressure /diastolic blood pressure [SBP/DBP] ≥ 140/90 mmHg 21 . Dyslipidemia was indicated by ICD-10 code E78 along with at least one prescription of lipid-lowering agents per year, or total cholesterol ≥ 240 mg/dL 22 . OSAS was defined as G47.30 in ICD-10 code as previously reported 23 . The OSAS group comprised those with OSAS diagnosed between 2009 and 2018 24,25 . Abdominal obesity was classified based on the International Obesity Task Force Asia-Pacific region and the cutoffs for Korean adults proposed by the Korean Society for the Study of Obesity 26 .

Statistical analyses.
To calculate the incidence of OSAS, the number of obstruction events was divided by the person-time at risk. The independent association of MetS with the risk of OSAS was determined using a Cox proportional hazards model adjusted for age, sex, smoking status, alcohol consumption, regular physical exer-

Results
The subjects were divided into two groups according to the presence or absence of MetS. The basic characteristics of the participants are shown in Table 1. The OSAS group was younger, male than female, less income, and higher rates of smoking and heavy drinkers. In addition, the OSAS group showed higher BMI, hypertension, dyslipidemia, higher total cholesterol, higher triglyceride, higher waist circumference, higher LDL cholesterol, and lower HDL cholesterol, indicating association with MetS. However, they exercised more regularly and had a lower incidence of DM. During 2009, 74,660 patients were newly diagnosed with OSAS (OSAS group). The OSAS group participants were younger, more likely to be male, had a lower income, and were more likely to be a current smoker, heavy drinker, and engage in regular exercise compared to the non-OSAS group. The OSAS group participants were more likely to have hypertension and dyslipidemia, and had a higher BMI, larger waist circumference, and higher levels of total cholesterol, low-density lipoprotein cholesterol (LDL-C) and TG. However, the non-OSAS group patients were more likely to have DM, and also had a lower HDL-C level. To further examine the relationship between OSAS and MetS, demographic data such as age, sex, and BMI, as well as smoking, drinking, and exercise status, were adjusted for in multivariable analysis.
Multivariable-adjusted analysis of the association between MetS and OSAS. The incidence rate of OSAS in the non-MetS and MetS groups was 0.82 and 1.16, respectively ( Table 2). The incidence rate of OSAS peaked in men in their 30 s and in women in their 50 s, which is when menopause begins (Fig. 2). The incidence rate of OSAS was higher in males, and the incidence probability of OSAS between the MetS and non-MetS population showed a significant difference between groups (log-rank test, p < 0.0001; Fig. 3 Table 2). The probability of OSAS increased as the number of MetS components increased, and with the passage of time (log-rank test, p < 0.0001; Fig. 4).

OSAS risk according to combinations of metabolic syndrome components.
We analyzed the incidence rates and multivariable-adjusted HRs for OSAS according to combinations of MetS components. All five components were individually associated with an increased risk of OSAS. Increases in waist circumference (HR: 1.99; 95% CI 1.90-2.08) and TG level (HR: 1.32; 95% CI 1.28-1.37) were associated with a marked increase in the risk of OSAS (Table 3). For the waist circumference component, there was a remarkable difference between

Discussion
The results of this nationwide Korean population-based study showed that MetS and the OSAS are closely related. The association was stronger in men than in women. Even when the criteria for MetS were not met, all individual MetS components were risk factors for OSAS, and the risk increased with the number of components. A large waist circumference and high TG level showed the strongest correlations with OSAS among the MetS components. Among patients with three or more MetS components, those with an elevated HDL-C and/or hypertension, accompanied by a large waist circumference and high TG level, showed a higher likelihood of OSAS. Obesity is strongly associated with MetS and a well-known risk factor for OSAS 27 . As the proportion of obese people continues to rise, the link between OSAS and MetS has become increasingly apparent 28 . Simple obesity (based on BMI) should be adjusted for in analyses of the relationship between OSAS and MetS, as a major potential confounder 14 . Age, sex, income, smoking status, and alcohol consumption are other important confounders that should be adjusted for 14,29 , but many studies did not do this 14 . In this study, we tried to minimize confounding effects by adjusting for BMI, age, sex, smoking status, alcohol consumption, and exercise status. www.nature.com/scientificreports/ Abdominal obesity is different from simple obesity. Waist circumference mainly depends on the fatty tissue in the peritoneum (i.e., in the areas between the stomach, liver, kidneys, intestines and other organs). High levels of visceral fat were also observed in obese and metabolically obese normal weight people 30 . Visceral fat is now known to be a metabolically active tissue type, in which a large amount of proinflammatory substances and vasoactivators are produced 31 ; these can cause metabolic dysregulation and atherogenesis. Also, serum leptin (adipocytokines) levels are considerably higher in patients with OSAS, suggesting that OSAS might be associated with leptin resistance 32 . Leptin resistance increases the likelihood of developing OSAS 33 . In addition to leptin resistance, abdominal obesity can cause systemic inflammation and metabolic dysfunction, leading to OSAS 34 . Compared to peripheral obesity, abdominal obesity has a greater effect on upper airway function 35 . Therefore, waist circumference should be measured as an important risk factor for OSAS. Visceral fat accounts for 5-8% of total body fat in women, and increases in menopause, while in men it accounts for 10-20% of total  www.nature.com/scientificreports/ body fat 36 . Since the prevalence of abdominal obesity is higher in men than women, waist circumference can also be considered an important factor in the difference in incidence of OSAS between genders, and might also explain why the OSAS incidence is higher in menopausal women. Recent studies have consistently demonstrated an independent association between adult OSAS and insulin resistance [37][38][39][40] . Insulin resistance sustains a low-grade inflammatory state, which can lead to upper airway narrowing, respiratory muscle fatigue, and decreased dilator muscle contraction 34 . The low success rate of continuous positive airway pressure treatment and upper airway surgery, for improving MetS or achieving weight loss in OSAS patients, may be explained by insulin resistance [41][42][43] .
There is considerable evidence that OSAS can lead to MetS 44 . Repeated respiratory obstruction events in OSAS result in intermittent hypoxia and frequent arousal. Temporarily ischemic tissue can release free radicals that cause oxidative stress, cytokine release and systematic inflammation 45,46 . Also, frequent arousal causes imbalances in the sympathetic nervous system and circadian misalignment. Consequently, intermittent hypoxia and the resultant oxidative stress, sympathetic activation, and sleep fragmentation have been suggested to underlie the pathogenic links between OSAS and glucose intolerance 47 , insulin resistance 48 , hypercholesterolemia 49 and hyperlipidemia 50 , all of which are MetS components. As shown in the literature, Mets and OSAS share various pathologic mechanisms and together constitute a potentially pathologic vicious cycle.
Our research had both strength and limitations. This is the largest population-based study to date, with the NHIS database covering the entire Korean population. However, any study that uses claims data is associated with a risk of misclassification. Therefore, potential confounding factors may exist and affect the quality of data. However, these could be compensated with large population data and high HR. Also, although causality could www.nature.com/scientificreports/ not be inferred in this study, a high risk of OSAS was observed among MetS patients according to our analysis of a nationwide population-based dataset covering a 10-year period.

Conclusions
OSAS is considered a major cause of MetS, and MetS can likewise trigger the development of OSAS. In this nationwide population-based analysis adjusted for several confounding factors, we confirmed the association of MetS components with OSAS. This is important because the coexistence of two pathologies within the same patient increases the levels of biomarkers, which directly contribute to, or increase the potential for, complications. Among the MetS components, particular attention should be paid to abdominal obesity in relation to OSAS, especially in men.

Data availability
The data that support the findings of this study are available from the Health Insurance Review & Assessment Service (HIRA). Restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available due to personal information protection. Data are available at https:// opend ata. hira. or. kr/ with the permission of the HIRA. www.nature.com/scientificreports/