Chronic Obstructive Pulmonary Disease Effect of Nonapnea Sleep Disorder on the Risk of Obesity: A Nationwide Population-Based Case–Control Study

Objectives: To investigate whether chronic obstructive pulmonary disease (COPD) affects nonapnea sleep disorder (NASD) on the risk of obesity. Materials and Methods: From 1 January 2000 to 31 December 2015, a total of 24,363 patients with obesity from the 2005 Longitudinal Health Insurance Database were identified; 97,452 patients without obesity were also identified from the same database. Multiple logistic regression was used to analyze the previous exposure risk of patients with obesity and NASD. A p value of <0.05 was considered significant. Results: The risk of developing obesity in patients with COPD is 3.05 times higher than that in patients without COPD. Patients with COPD with NASD had a 1.606-fold higher risk of developing obesity than those without NASD. Patients with obesity were more likely to be exposed to NASD than did those without obesity (adjusted odds ratio, 1.693; 95% confidence interval, 1.575–1.821, p < 0.001). Furthermore, the closeness of the exposure period to the index time was positively associated with the severity of obesity, with a dose–response effect. The exposure duration of NASD in patients with obesity was 1.693 times than that in those without obesity. Longer exposure durations were associated with more severe obesity, also with a dose–response effect. Conclusions: The COPD effect of NASD increases the subsequent risk of obesity, and the risk of obesity was determined to be significantly higher in patients with NASD in this case–control study. Longer exposure to NASD was associated with a higher likelihood of obesity, also with a dose–response effect.


Introduction
Obesity is a worldwide epidemic, and its prevalence is increasing in most Western societies and developing countries. At the current growth rate, the global obesity rate for 2 of 12 men and women will reach 18% and over 21%, respectively, by 2025 [1]. Furthermore, obesity may gradually cause or aggravate various complications, including type 2 diabetes, hypertension (HTN), dyslipidemia, cardiovascular disease (CVD), nonalcoholic fatty liver disease (NAFLD), reproductive dysfunction, abnormal breathing, and mental illness. It may even increase the risk of certain types of cancer depending on the degree, duration, and distribution of excess body weight and fat tissue [1]. The World Health Organization (WHO) indicated that obesity is a chronic disease and highlighted the health hazards of obesity. The 2016 WHO data revealed that over 1.9 billion adults (aged ≥18 years) are overweight. Among them, over 650 million were obese [2]. The top ten causes of death in Taiwan in 2017 were cancer, heart disease, CVD, diabetes, hypertensive disease, nephritis, renal syndrome, nephropathy, chronic liver disease, and cirrhosis-seven of which are related to obesity [3]. In the latest survey, the National Health Administration of the Ministry of Health and Welfare reported that the rate of obesity among adults (aged ≥18 years) has increased from 38% in 2009 to 43.9% in 2018 [4]. Compared with people with a healthy weight, obese people have more than three times the risk of diabetes, metabolic syndrome, and dyslipidemia and two times the risk of HTN, CVD, knee arthritis, and gout [4].
Sleep disorders (SDs) are a group of diseases characterized by disturbances in the amount, quality, or timing of sleep or sleep-related physiological conditions [5]. In Taiwan, the prevalence of SDs in the general population is 4.2% [6], and approximately 20% of women aged 25-44 years develop SDs [7]. According to the International Classification of Sleep Disorders (American Academy of Sleep Medicine, 2014), SDs are divided into seven categories: insomnia, sleep-related breathing disorders, central disorders of hypersomnolence, circadian rhythm SDs, parasomnias, sleep-related movement disorders, and other SDs [5]. Nonapnea SDs (NASDs) refer to SDs other than sleep apnea. Studies have extensively investigated the association between SDs and chronic diseases. NASD has recently been associated with an increased risk of other comorbidities, including HTN, diabetes mellitus (DM), chronic kidney disease (CKD), CVD, chronic obstructive pulmonary disease (COPD), and stroke [8]. However, the relationship between SDs and obesity may be a critical mediating factor that links SDs with chronic diseases, including CVD, COPD, and DM, in all age groups [9,10].
COPD is a common respiratory disease characterized by persistent expiratory airflow obstruction, which is progressive and accompanied by chronic airway inflammation [11]. SDs is one of the most common symptoms reported by patients with COPD, occurring in approximately 40% of patients in one large study [11]. These patients have problems initiating or maintaining sleep and have mild increases in sleep, decreases in rapid eye movement (REM) sleep, and frequent sleep stage transitions and microarousals. Sleep efficiency is low in most patient populations [12]. SDs may contribute to the nonspecific daytime symptoms of chronic fatigue, lethargy, and impaired overall quality of life described in 50-70% of patients with COPD [12]. Nocturnal symptoms in patients with COPD are often overlooked by physicians and/or not reported by patients themselves [13]. Understanding this connection may help in developing effective therapeutic interventions for SDs and obesity [14,15]. Longitudinal observational studies investigating the relationship between SDs and obesity are limited. Therefore, we hypothesized that COPD affects NASDs on the risk of obesity. We used the National Health Insurance Research Database (NHIRD) of the Ministry of Health and Welfare to investigate whether the COPD effect of NASDs increases the subsequent risk of obesity.

Data Source
Taiwan's National Health Insurance launched the single-payer system on 1 March 1995. As of 2017, 99.9% of Taiwan's population is enrolled in this program. Data for this study were collected from the 2005 Longitudinal Health Insurance Database (LHID2005), which is part of the NHIRD, and 2,000,000 people were randomly selected from the entire population. The National Institutes of Health encrypt all personal information before releasing the LHID2005 to protect the privacy of patients. In the LHID2005, the disease diagnosis code is based on the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) criteria [16]. Figure 1 shows the flowchart of the study design (case-control study) from the NHIRD in Taiwan. All methods were performed in accordance with relevant guidelines and regulations. This study was approved by the Ethical Review Board of the Tri-Service General Hospital of the National Defense Medical Center (TSGHIRB No. B-109-39).

Data Source
Taiwan's National Health Insurance launched the single-payer system on 1 March 1995. As of 2017, 99.9% of Taiwan's population is enrolled in this program. Data for this study were collected from the 2005 Longitudinal Health Insurance Database (LHID2005), which is part of the NHIRD, and 2,000,000 people were randomly selected from the entire population. The National Institutes of Health encrypt all personal information before releasing the LHID2005 to protect the privacy of patients. In the LHID2005, the disease diagnosis code is based on the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) criteria [16]. Figure 1 shows the flowchart of the study design (case-control study) from the NHIRD in Taiwan. All methods were performed in accordance with relevant guidelines and regulations. This study was approved by the Ethical Review Board of the Tri-Service General Hospital of the National Defense Medical Center (TSGHIRB No. B-109-39).

Determining Cases and Controls
This study applied a population-based case-control design. A case-control study is used to determine the causes of a health outcome by comparing two different groups: one group of patients with a particular disease or condition as cases and the other group without a certain disease or condition. Populations with a particular disease or condition are controls. Controls and cases should have similar background characteristics, especially characteristics that affect the disease (or condition) in question (e.g., choosing the same age or the same place of residence), which also means that researchers can use it to discover differences in lifestyle and possible causes of disease. Patients diagnosed with obesity (ICD-9-CM code 278) were defined as an obesity case group. The control group consisted of patients without obesity. Patients in both the case and control groups were matched by the index date, sex, and age at a ratio of 1:4.

Determining Cases and Controls
This study applied a population-based case-control design. A case-control study is used to determine the causes of a health outcome by comparing two different groups: one group of patients with a particular disease or condition as cases and the other group without a certain disease or condition. Populations with a particular disease or condition are controls. Controls and cases should have similar background characteristics, especially characteristics that affect the disease (or condition) in question (e.g., choosing the same age or the same place of residence), which also means that researchers can use it to discover differences in lifestyle and possible causes of disease. Patients diagnosed with obesity (ICD-9-CM code 278) were defined as an obesity case group. The control group consisted of patients without obesity. Patients in both the case and control groups were matched by the index date, sex, and age at a ratio of 1:4.

Statistical Analysis
Descriptive data are presented as percentages, means, and standard deviations. Chisquare test and t-test were used to evaluate the distribution of categorical and continuous variables between cases and controls. The Charlson comorbidity index (CCI) assesses comorbidity level by considering both the number and severity of 19 predefined comorbid conditions. It provides a weighted score of a client's comorbidities that can be used to predict short-and long-term outcomes, such as function, hospital length of stay, and mortality rates. After controlling the main reasons for admission and severity, survival analysis was used to explore the relationship between comorbidity and death within 1 year, and one category of comorbidity was weighted according to the adjusted relative risk. The 10-year survival of the patient was verified. Table 1 presents the comorbidity categories and weights. If the relative risk was >1.2 and <1.5, the weight was 1; if the relative risk was >1.5 and <2.5, the weight was 2; if the relative risk was >2.5 and <3.5, the weight was 3. The relative risk of type 2 comorbidity was >6 and was given a weight of 6. The comorbidity weights of patients were aggregated. Conditional logistic regression analyses were performed to evaluate the effect of NASD on the risk of obesity after adjusting for age, sex, education, insured premium, comorbidities, CCI, season, location, urbanization level, and level of care. The effect of the first to last NASD exposure before obesity diagnosis on the factors of obesity was examined using conditional logistic regression. All analyses were performed using SPSS version 22 (IBM, Armonk, NY, USA). A p value of <0.05 was considered significant.

Demographic Data
As presented in Table 1, the mean age of 121,815 patients was 44.30 ± 15.64 years, among whom 42.77% were men and 57.23% were women. A total of 24,363 patients with obesity (cases) and 97,452 patients without obesity (controls) were recruited. Patients in the case group had a higher prevalence of COPD and liver cirrhosis than did those in the control group. In the case group, the CCI, season, location, urbanization level, and level of care were significant.

Local hospital Reference
Adjusted OR, adjusted odds ratio; CI, confidence interval; CHF, chronic heart failure; COPD, chronic obstructive pulmonary disease.

Logistic Regression to Stratify the Obesity Factors of the Listed Variables
As presented in Table 3, the risk of obesity was higher in patients with NASDs than in controls (AOR, 1.693; 95% CI, 1.575-1.821). Women with NASDs had 1.926 times the risk of obesity compared with women without NASDs (AOR, 1.926; 95% CI, 1.791-2.071). The risk of obesity with NASDs was significantly higher in patients aged 20-44 years than in controls (AOR, 1.959; 95% CI, 1.822-2.107). Patients with COPD with NASDs had a 1.606-fold higher risk of developing obesity than those without NASDs (AOR, 1.606; 95% CI, 1.494-1.728). The conditional logistic regression analysis results reveal that patients with NASDs had a significantly higher risk of obesity in spring than controls (AOR, 1.976; 95% CI, 1.838-2.125).  Adjusted OR, adjusted odds ratio (adjusted for the variables listed in Table 3); CI, confidence interval; CHF, chronic heart failure; COPD, chronic obstructive pulmonary disease.

Logistic Regression to Analyze Obesity Factors between Different Periods of NASD Exposure
As illustrated in Figure 2, obese patients were more likely to have experienced NASDs than nonobese patients (AOR, 1.693; 95% CI, 1.575-1.821). Furthermore, the closeness of the exposure duration to the time of the study was positively associated with obesity severity in a dose-response manner (NASDs exposure of <1 year, AOR, 2.386; NASDs exposure of ≥1 and <5 years, AOR, 1.725; NASDs exposure of ≥5 years, AOR, 1.422). Figure 2, obese patients were more likely to have exper NASDs than nonobese patients (AOR, 1.693; 95% CI, 1.575-1.821). Furthermore, th ness of the exposure duration to the time of the study was positively associated wi sity severity in a dose-response manner (NASDs exposure of <1 year, AOR, 2.386; N exposure of ≥1 and <5 years, AOR, 1.725; NASDs exposure of ≥5 years, AOR, 1.422 Figure 2. The last nonapnea sleep disorder exposure before the first obesity diagnosis. Figure 2. The last nonapnea sleep disorder exposure before the first obesity diagnosis.

As illustrated in
As shown in Figure 3, the mean exposure duration of NASDs in patients with obesity was 1.693 times that in patients without obesity (AOR, 1.693; 95% CI, 1.575-1.821). Furthermore, a longer exposure duration was associated with more severe obesity, with a dose-response effect (NASDs exposure of <1 year, AOR, 1.420; NASDs exposure of ≥1 to <5 years, AOR, 2.240; NASDs exposure of ≥5 years, AOR, 2.863).

Discussion
The results of this study reveal that male patients had a significantly lower r obesity than did female patients, which is consistent with the findings of a study ducted by Kanter et al. [17]. The prevalence of overweight and obesity among me women varies greatly within and between countries, and more women are obese men overall [17]. Patients aged 45-64 or >65 years had a significantly lower risk of o than did patients aged 20-44 years; this may be because of weight loss or weight c in middle-aged and elderly people due to the disease itself. However, unknown f may affect this result [18]. Furthermore, the risk of obesity was significantly lower tumn and winter than in spring, according to the result reported by Ma et al. [19 total daily intake in spring is higher than in autumn, with a daily difference in the intake of 222 kcal/day [19]. The risk of NASDs comorbidities, such as HTN, diabetes, CVD, and stroke, also increased in spring. Lin et al. reported similar results [20]. Alth the physiological mechanism of the association between NASDs and obesity remain clear, we inferred the underlying mechanism from previous studies, which may pr some insights for our observations. Research articles (mainly cross-sectional and ob tional) have not clarified whether SDs cause obesity or obesity causes SDs [21]. Addi research with larger sample sizes and controlling for confounding factors is warr [21].
Our findings also show that the risk of developing obesity in patients with CO 3.05 times higher than that in patients without COPD. Patients with COPD with N had a 1.606 times higher risk of developing obesity than without NASDs. Alth weight loss is one way to treat COPD, previous studies have shown that approxim 65% of patients with COPD are overweight or obese and that obesity in patients COPD is associated with other several disease sequelae, such as increased sympto dyspnea, higher prescription rates of inhaled drugs, and higher use of healthca sources [22]. Obesity is independently associated with several CVD risk factors, incl

Discussion
The results of this study reveal that male patients had a significantly lower risk of obesity than did female patients, which is consistent with the findings of a study conducted by Kanter et al. [17]. The prevalence of overweight and obesity among men and women varies greatly within and between countries, and more women are obese than men overall [17]. Patients aged 45-64 or >65 years had a significantly lower risk of obesity than did patients aged 20-44 years; this may be because of weight loss or weight control in middle-aged and elderly people due to the disease itself. However, unknown factors may affect this result [18]. Furthermore, the risk of obesity was significantly lower in autumn and winter than in spring, according to the result reported by Ma et al. [19]. The total daily intake in spring is higher than in autumn, with a daily difference in the total intake of 222 kcal/day [19]. The risk of NASDs comorbidities, such as HTN, diabetes, CKD, CVD, and stroke, also increased in spring. Lin et al. reported similar results [20]. Although the physiological mechanism of the association between NASDs and obesity remains unclear, we inferred the underlying mechanism from previous studies, which may provide some insights for our observations. Research articles (mainly cross-sectional and observational) have not clarified whether SDs cause obesity or obesity causes SDs [21]. Additional research with larger sample sizes and controlling for confounding factors is warranted [21].
Our findings also show that the risk of developing obesity in patients with COPD is 3.05 times higher than that in patients without COPD. Patients with COPD with NASDs had a 1.606 times higher risk of developing obesity than without NASDs. Although weight loss is one way to treat COPD, previous studies have shown that approximately 65% of patients with COPD are overweight or obese and that obesity in patients with COPD is associated with other several disease sequelae, such as increased symptoms of dyspnea, higher prescription rates of inhaled drugs, and higher use of healthcare resources [22]. Obesity is independently associated with several CVD risk factors, including diabetes, HTN, dyslipidemia, cancer, sleep apnea, and other major CVDs; patients with COPD are also associated with obesity [23]. In addition, sleep disturbance is one of the most common symptoms reported by patients with COPD. In a large study, sleep disturbance occurred in approximately 40% of patients with COPD [11]. These patients with COPD have problems with maintaining sleep and have mild increases in sleep and decreased REM sleep, frequent changes in sleep stages, and microarousals. COPD, which is common in the elderly with respiratory diseases, not only affects a variety of coexisting diseases but also directly affects the patient's health and family care willingness [24].
Previous studies have pointed to a bidirectional link between poor sleep and COPD severity scores, i.e., COPD symptoms such as cough and dyspnea may be responsible for poor sleep quality; sleep disorder can lead to adverse outcomes associated with COPD [25]. Sleep efficiency is low in most COPD patient groups; sleep disturbance may lead these patients with COPD to describe chronic fatigue, lethargy, and overall impairment of quality of life. In nonspecific daytime symptoms, nocturnal symptoms in patients with COPD are often overlooked by physicians and/or not reported by patients themselves [26]. Sleep has several effects on breathing, including changes in central respiratory control, lung mechanics, and muscle contractions, which do not adversely affect healthy individuals but may lead to hypoxia in patients with COPD, especially during REM sleep [27]. Therefore, based on previous studies and the findings of this study, we speculate that NASDs may cause the development of COPD and that COPD may lead to the occurrence of obesity.
These pathophysiological factors may explain the association between the COPD effect of NASDs and obesity demonstrated in this study. Our research results reveal that the risk of obesity in patients with COPD was 3.05 times that of patients without COPD. Patients with COPD with NASDs had a 1.606 times higher risk of developing obesity than patients without NASDs. The prevalence of NASD among patients with obesity was 1.693 times than that in those without obesity. Furthermore, the NASDs duration and closeness to the time of the study were both positively related to the severity of obesity. Therefore, the relationship between the occurrence and duration of COPD effect of NASDs and obesity warrants consideration.
This study has several limitations. First, the NHIRD does not provide detailed information, such as that related to alcohol consumption, smoking, eating, and physical activity behaviors, which may affect our findings. Second, body mass index (BMI) was not a variable in our study. Third, although this study was carefully designed and controlled for confounding factors, biases may still exist due to unmeasured or unknown confounding factors (e.g., the onset of depression, the stage of obesity at the time of diagnosis, and drugs that may affect the outcome). A prospective cohort study is recommended to evaluate the relationship between NASDs, COPD, and obesity.

Conclusions
This study revealed that the relationship between the occurrence and duration of the effect of COPD on NASDs and obesity warrants consideration and that patients with obesity experienced more severe NASDs than did those without obesity. Furthermore, the closeness to the time of the study and the exposure duration were both positively related to the severity of obesity, with a dose-response effect. NASDs may be a risk factor for obesity. Healthcare providers should pay close attention to the relationship between NASDs, COPD, and obesity. Future prospective and experimental studies need to be conducted in order to better determine a cause-and-effect relationship between NASDs, COPD, and obesity. The research results will be important evidence for healthcare management when promoting the prevention of SDs and obesity.