https://orcid.org/0000-0002-1388-8075
Figures
Abstract
Background & aims
Nonalcoholic fatty liver disease (NAFLD) is one of the most common health problems worldwide. Sleep apnea (SA) causes cardiovascular and metabolic problems, as well as a significant socioeconomic burden. Although several studies have found that SA causes NAFLD, there is no evidence that NAFLD causes SA. The goal of this study was to look at the relationship between NAFLD and SA in realworld data.
Methods
We evaluated 334,334 healthy individuals without comorbidities who underwent National Health checkups in the Republic of Korea from 2009 to 2014. NAFLD was defined by a surrogate marker, the fatty liver index (FLI). The association between FLI and SA was analyzed using multivariate Cox proportional hazards regression models.
Results
During a median followup of 5.3 years, 1,351 patients (0.4%) were newly diagnosed with SA. Subjects were categorized into quartile groups according to FLI (range: Q1, 0–4.9; Q2, 5.0–12.5; Q3, 12.6–31.0; Q4, >31.0). Subjects with higher FLIs had a significantly higher cumulative incidence of SA than those with lower FLIs (Q1, 119 [0.1%]; Q2, 210 [0.3%]; Q3, 339 [0.4%]; Q4, 683 [0.8%]; P < 0.001). Adjusted hazard ratios (HRs) revealed that a higher FLI was independently associated with an increased risk of SA (HR between Q4 and Q1, 4.03; 95% confidence interval, 3.22–5.05; P < 0.001). This association remained statistically significant after further adjustment for Body mass index (BMI) (HR between Q4 and Q1, 2.19; 95% confidence interval, 1.69–2.83; P < 0.001). FLI was significantly associated with an increased risk of new-onset SA regardless of baseline characteristics.
Citation: Kim N, Roh J-H, Lee H, Kim D, Heo SJ (2022) The impact of non-alcoholic fatty liver disease on sleep apnea in healthy adults: A nationwide study of Korea. PLoS ONE 17(7): e0271021. https://doi.org/10.1371/journal.pone.0271021
Editor: Giannicola Iannella, Sapienza University of Rome, ITALY
Received: March 15, 2022; Accepted: June 21, 2022; Published: July 20, 2022
Copyright: © 2022 Kim et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Data are available from the National Health Insurance Service (NHIS) of Republic of Korea. NHIS provide data for researchers who meet the criteria for access to confidential data. The data underlying the results presented in the study are available from from (https://nhiss.nhis.or.kr/bd/ab/bdaba000eng.do;jsessionid=Fl6wPgxuUcdGBZ6iwkSzEIZfKHpGVJBx1zzU7JSV45vTV17X7UTQ3CdwINI20hlc.primrose22_servlet_engine10) after review of study purpose and method.
Funding: This work was supported by the Korea Medical Device Development Fund grant funded by the Korea government (the Ministry of Science and ICT, the Ministry of Trade, Industry and Energy, the Ministry of Health & Welfare, the Ministry of Food and Drug Safety) (Project Number: 1711138161, KMDF_PR_20200901_0097) The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Nonalcoholic fatty liver disease (NAFLD) is one of the most common health problems worldwide. Because NAFLD is typically a silent disease, exact incidence rates are unknown; however, it is commonly estimated to be in the range of 27%–34% in the general population [1]. Obesity, diabetes mellitus, and metabolic syndrome are all associated with the prevalence of NAFLD. It is estimated that 75%–92% of morbidly obese patients and 60%–70% of type 2 diabetes patients are affected [1, 2]. Sleep apnea (SA) is a variety of respiratory diseases including obstructive sleep apnea (OSA) and central sleep apnea that cause nocturnal hypoxemia. OSA is a heterogeneous disorder characterized by repetitive narrowing or closure of the upper airway during sleep. It affects 2%–4% of the general population and 35%–45% of obese people [3, 4]. OSA can lead to systemic hypertension, cognitive decline, and even permanent brain damage [5, 6]. However, since the hypertension and decreased cognitive function can be improved with OSA treatment, caution is required for OSA [7].
Although several studies are showing a positive association between SA and NAFLD using fatty liver index (FLI) or histologic examination, they are small sample sizes and showed a simple association in several cross-sectional studies [8, 9]. Previous studies focused only on the occurrence of NAFLD due to hypoxic damage in SA. Liver inflammation is caused by the interaction of various inflammatory factors including hypoxic damage [10]. Thus, we examined a large, comprehensive cohort of healthy Korean adults who lacked traditional risk factors and comorbidities to determine the direct relationship between NAFLD and SA.
Methods
Data source
We analyzed the National Health Insurance Service–National Sample Cohort 2.0 (NHIS-NSC 2.0) data set which is the population-based sample cohort data of Korea. The NHIS-NSC 2.0 was made by extracting random samples of about 2% of the total Korean population in 2006 with retrospective and prospective followup data from 2002 to 2015. Koreans are required by law to obtain government-provided National Health insurance, and NHIS covers 97% of Koreans. The Korean government stratified insured population (n = 48,222,537) into 2,142 classes according to their by gender, age, region, eligibility status, and income level. The Korean government then randomly selected 2.1% from each stratum (n = 1,021,208). As a result, the NHIS-NSC 2.0 cohort can represent the entire Korean population [11].
The cohort contains four databases: the first is a sociodemographic information dataset of the beneficiaries; the second is a medical claims dataset with diagnoses based on the 10th revision of the International Classification of Disease (ICD-10) codes, admission, and treatment; the third is a National Health Screening dataset of the cohort members; and the fourth is a medical institutions dataset.
Korean adults over the age of 20 are recommended the National Health checkup biennially, which is composed of blood tests, chest X-rays, physical examinations, and questionnaires on medical history and health-related behaviors including smoking status and alcohol consumption. According to NHIS data from 2013, 72.1% of eligible beneficiaries received a National Health checkup [7]. The cohort also included mortality data such as the date and cause of death from the National statistical office’s death registration database, which is a government statistical agency.
All researchers whose protocols have been approved by the NIHS review board can use the NHIS-NSC 2.0. The Chungnam National University Hospital’s Institutional Review Board in Daejeon, Korea, approved this study and waived the requirement for informed consent.
Study population
This study included a population over 20 years of age who received National Health checkups at least once from 2009 to 2014. The index checkup was defined as the first National checkup, and the index year was the year of the index checkup. Subjects who met the predefined exclusion criterion were excluded.: (a) those suffering from congestive heart failure, hypertension, diabetes, cerebrovascular disease, ischemic heart disease, peripheral artery disease, or liver disease; (b) those with rheumatic or nonrheumatic valvular heart disease; (c) those who with increased baseline SBP ≥140 mmHg or DBP ≥90 mmHg, and fasting blood glucose ≥126 mg/dL at the index checkups; and (d) those found to have missing data in the index checkup. Each diagnosis was based on 1-year clams data before the index year or index checkups (the questionnaires for medical history, and measurement of blood pressure and fasting glucose levels). We defined each diagnosis in the claims data as the first occurrence on at least two different days of hospital visits (outpatient) or on the first admission. The ICD-10 codes are used to define each diagnosis.
Definition of data
Smoking status was categorized into ex- or current smokers, and nonsmokers. Alcohol consumption was assessed using standardized self-reporting questionnaires that included questions about how many days of the week alcohol is consumed and how much alcohol is consumed on each drinking day. The amount of alcohol consumption was calculated by multiplying them. Body mass index (BMI) was calculated by dividing weight (kg) by height squared (m2). The physical activity questionnaires included questions about how many days a week 30 minutes of light exercise, 30 minutes of moderate exercise, and 20 minutes of vigorous exercise are performed, respectively. The light exercise was assumed to be 2 metabolic equivalent of tasks (METs), moderate 3METs, and vigorous 6METs, which were multiplied by 20, 30, and 30 minutes and the respective number of days a week, and summed.
The primary outcome of interest was finding the incidence of SA. The first occurrence of SA during at least two different days of outpatient visits, admission for SA, or death with a diagnosis of SA was defined as the SA incidence. Data were censored at the time of SA occurrence, NHIS disqualification (death or immigration), or study completion (December 31th, 2015).
Diagnosis of steatosis
A well-validated, surrogate marker, FLI was used to identify patients with steatosis [12]. FLI was composed of four variables (triglycerides [TG], Body mass index [BMI], gamma-glutamyltransferase [GGT], and waist circumference [WC] and calculated as follows:
The original study suggested an FLI ≥ 60 as the cutoff for a diagnosis of hepatic steatosis with a positive likelihood ratio of 4.3 in the general population [12]. Although the FLI is simple to obtain and calculate for screening fatty liver disease, there is insufficient evidence to support the diagnosis of fatty liver disease in Asians with the FLI due to lower BMI and WC than other ethnic populations [13]. Resultantly, we categorized our study population into quartiles according to their FLI value and used the quartile group in the statistical analysis.
Statistical analysis
Data are presented as mean ± standard deviation for continuous variables and as number (percentage) for categorical variables. The differences between FLI quartiles were compared using oneway analysis of variance and chi-square tests. The incidence of SA was calculated by dividing the total number of newly diagnosed SA cases between 2009 and 2014 by the total population during the same period and expressing the result as cases per 100,000 person-years. Cumulative event rates were calculated and compared between FLI quartile-based groups using Kaplan–Meier estimates and the log-rank test. For SA incidence, Cox proportional hazards regression was used to estimating adjusted hazard ratios (HRs), and 95 percent confidence intervals (CIs) were used. Data were first adjusted for age and gender (model 1). Then, clinical characteristics of which association with new-onset SA demonstrates borderline statistical significance (P < 0.10), smoking, alcohol consumption, activity, systolic blood pressure, diastolic blood pressure, fasting glucose, and low-density lipoprotein cholesterol were incorporated into model 1 to obtain model 2. In model 3, multivariate analysis was performed, and BMI, which is an important risk factor for both NAFLD and SA, was added to model 2. The variance inflation factor was used to test the multicollinearity of the variables included in the analysis, and all of them were less than 10, indicating that they had no effect on the results.
For subgroup analyses, FLI was integrated into the statistical models as a continuous variable after log transformation. A two-sided P value of <0.05, was considered to indicate statistical significance. R software, version 3.3.3, was used for statistical analyses (R Foundation for Statistical Computing, Vienna, Austria; www.r-project.org).
Results
Baseline characteristics (of the participants)
Of the total 556,884 NHIS-NSC 2.0 sample cohort population, the exclusion of subjects who met the prespecified exclusion criteria left 334,334 subjects into the analysis. Fig 1 shows a list of subjects who meet each exclusion criterion. According to FLI quartile values, the study population was divided into four groups (first quartile [Q1], 0–4.9; second quartile [Q2], 5.0–12.5; third quartile [Q3], 12.6–31.0; and fourth quartile [Q4], >31.0). Table 1 summarizes the baseline clinical characteristics of the study population based on the FLI quartile. The subjects with higher FLI were older and more likely to be male and obese than those with a lower FLI. From Q1 to Q4, the proportion of current smokers and the amount of alcohol consumed were expected to rise. Subjects with higher FLIs had higher blood pressure, fasting blood glucose, and GGT levels than those with lower FLIs. With increasing FLI values, the lipid profile becomes increasingly unfavorable.
Association between FLI and the incidence of SA
The followup duration was a median of 5.3 years (interquartile range, 4.1–6.3), during which 1,351 subjects (0.4%) developed new-onset SA. Fig 2 depicts the cumulative incidence of new-onset SA according to FLI quartile groups. When comparing subjects with higher FLIs to those with lower FLIs, the incidence of SA was significantly higher in those with higher FLIs (Q1, 119 [0.1%]; Q2, 210 [0.3%]; Q3, 339 [0.4%]; Q4, 683[0.8%]). In multivariate models which were adjusted for age and gender (model 1), the HR (95% CI) was 3.67 (2.95–4.57) for the highest group of FLI compared with that of the lowest. When the model was further adjusted for clinical characteristics with P < 0.10 in the univariate analyses (model 2), the HR (95% CI) was 4.03 (3.22–5.05) for the highest group of FLI compared with that of the lowest. The relationship between FLI quartile and SA incidence remained statistically significant after adjusting for BMI in model 2 (Table 2).
In addition to the FLI groups, the analysis was carried out after categorizing the study subjects according to the various FLI cutoff criteria proposed in previous studies. Regardless of the cutoff-points used, the highest FLI goup had the highest risk for new onset SA.
Subgroup analysis (of clinical variable affecting SA incidence)
Adjusted HRs according to subgroups summarized in Fig 3. FLI was a significant determinant of new-onset SA in all subgroups (Fig 3).
HR, hazard ratio; CI, confidence interval.
Discussion
In this study, we looked at the relationship between NAFLD as measured by FLI, a well-validated surrogate biomarker, and the prevalence of SA in a large, realworld data set from the Korean population. Recent studies demonstrate an association between NAFLD and SA [8, 14, 15]. Krolow et al. analyzed 51 patients and reported significant associations between histologically proven liver fibrosis and moderate to severe OSA [8]. With 153 participants who had NAFLD confirmed by liver ultrasound, Zhang et al. demonstrated that OSA plays a role in liver injury among severely obese individuals [14]. Wei et al. reported that patients with a sleep disorder had a higher risk of developing NAFLD, including those sleep disorder patients who did not have SA [16]. Previous studies, however, had limitations in that the sample size was too small [5, 10], the subjects were limited to severely obese patients [10], or it was a meta-study [11]. Furthermore, there is limited evidence that fatty liver is directly related to the occurrence of SA. In this study, realworld data revealed that the incidence of SA was high in patients with fatty liver.
Several mechanisms potentially link NAFLD and SA beyond the shared risk factors. Firstly, the wider neck circumference in NAFLD patients than in the general population may cause SA. NAFLD patients have a wider neck circumference (NC) than healthy subjects, according to previous research [13]. Kim et al. demonstrated that the NC can predict both the presence and severity of OSA [17]. Secondly, patients with NAFLD have a high incidence of asthma [18], and asthma and allergic rhinitis are closely related [19]. Furthermore, given the high prevalence of SA in patients with allergic rhinitis [17], it may explain why patients with NAFLD have more SA. Finally, NAFLD may be associated with increased systemic inflammation [18], which may lead to SA. Excess triglyceride accumulation in hepatocytes has been linked to impaired fatty acid oxidation, increased oxidative stress, and local inflammation, all of which can potentially lead to the progression from simple steatosis to steatohepatitis [20]. Previous studies have found that the degree of hepatic steatosis and inflammation correlated with systemic levels of inflammatory biomarkers including C-reactive protein (CRP) [21]. According to Gaines et al., patients with elevated CRP are more likely to develop SA [22]. The oxidative stress associated with NAFLD is also closely related to SA [23]. As the prevalence of NAFLD, which is potentially associated with SA, is increasing, attention should be paid to SA as well as NAFLD [24].
Given the observed link between NAFLD and SA, as well as the mechanisms underlying this link, it is reasonable to assume that interventions for NAFLD management may have the therapeutic potential to reverse the natural course of SA. Obesity is widely recognized as a major risk factor for SA, and if achieved, significant and sustained weight loss is likely to be a useful therapeutic option in the management of SA [25]. Because NAFLD is also strongly linked with obesity and has been shown to improve with weight loss [26], the observed beneficial effect of weight loss on the natural progression of NAFLD may be partly attributed to the improvement of SA. Many therapeutic interventions other than weight loss are under intense investigation for the management of NAFLD. They include bariatric surgery, treatment with thiazolidinediones, glucagon-like peptide-1 receptor agonists and sodium–glucose cotransporter-2 inhibitors, all of which showed a promising effect against NAFLD [26]. The impact of these upstream interventions aimed at NAFLD on the course of SA disease must be studied further in the future. Despite significant correlations between hepatic steatosis and SA severity markers, continuous positive airway pressure alone did not improve hepatic steatosis or fibrosis. However, the additional role of weight reduction through lifestyle modification deserves further investigation [27].
Several limitations of this study need to be acknowledged. To begin, we estimated new-onset SA using the NHIS-NSC 2.0 database, which is based on ICD-10 disease codes; however, the accuracy of SA diagnosis was not validated in this study. More research is needed to validate SA patient data with health insurance data using hospital-based medical records. Second, steatosis was determined using a surrogate biochemical marker rather than histology or magnetic resonance spectroscopy. An invasive procedure or a complex, expensive imaging modality, on the other hand, is not feasible in large cohorts. FLI, on the other hand, is a widely accepted surrogate for hepatic steatosis that has been validated in the general population [12]. Interestingly, an increased risk for new-onset SA was observed in the subjects with FLI values of around 30, which was much lower than the cutoff point suggested by Bedogni et al [12]. These findings back up Yang et al.’s study, in which they proposed lower cutoff points for Asians with different anthropometric characteristics than Westerns [28]. Finally, the study population’s exclusivity to Koreans may limit the current study’s generalizability to other ethnic populations.
In conclusion, our study demonstrated that NAFLD, as measured by FLI, was independently associated with increased risk for SA in the healthy Korean population. Because management of SA remains a difficult clinical challenge, the identification of potentially modifiable risk factors may have important for clinicians. More research on the effect of NAFLD management on the incidence and severity of SA is required.
References
- 1. Fazel Y, Koenig AB, Sayiner M, Goodman ZD, Younossi ZM. Epidemiology and natural history of non-alcoholic fatty liver disease. Metabolism. 2016;65(8):1017–25. Epub 2016/03/22. pmid:26997539.
- 2. McPherson S, Hardy T, Henderson E, Burt AD, Day CP, Anstee QM. Evidence of NAFLD progression from steatosis to fibrosing-steatohepatitis using paired biopsies: implications for prognosis and clinical management. J Hepatol. 2015;62(5):1148–55. Epub 2014/12/06. pmid:25477264.
- 3. Kono M, Tatsumi K, Saibara T, Nakamura A, Tanabe N, Takiguchi Y, et al. Obstructive sleep apnea syndrome is associated with some components of metabolic syndrome. Chest. 2007;131(5):1387–92. Epub 2007/05/15. pmid:17494788.
- 4. West SD, Nicoll DJ, Stradling JR. Prevalence of obstructive sleep apnoea in men with type 2 diabetes. Thorax. 2006;61(11):945–50. Epub 2006/08/25. pmid:16928713; PubMed Central PMCID: PMC2121154.
- 5. Vanek J, Prasko J, Genzor S, Ociskova M, Kantor K, Holubova M, et al. Obstructive sleep apnea, depression and cognitive impairment. Sleep Med. 2020;72:50–8. Epub 2020/06/17. pmid:32544796.
- 6. Mashaqi S, Gozal D. Obstructive Sleep Apnea and Systemic Hypertension: Gut Dysbiosis as the Mediator? J Clin Sleep Med. 2019;15(10):1517–27. Epub 2019/10/10. pmid:31596218; PubMed Central PMCID: PMC6778338.
- 7. Pollicina I, Maniaci A, Lechien JR, Iannella G, Vicini C, Cammaroto G, et al. Neurocognitive Performance Improvement after Obstructive Sleep Apnea Treatment: State of the Art. Behav Sci (Basel). 2021;11(12). Epub 2021/12/24. pmid:34940115; PubMed Central PMCID: PMC8698492.
- 8. Krolow GK, Garcia E, Schoor F, Araujo FBS, Coral GP. Obstructive sleep apnea and severity of nonalcoholic fatty liver disease. Eur J Gastroenterol Hepatol. 2021;33(8):1104–9. Epub 2020/09/15. pmid:32925501.
- 9. Chen X, Lin X, Chen LD, Lin QC, Chen GP, Yu YH, et al. Obstructive sleep apnea is associated with fatty liver index, the index of nonalcoholic fatty liver disease. Eur J Gastroenterol Hepatol. 2016;28(6):650–5. Epub 2016/02/20. pmid:26894633.
- 10. Khan HA, Ahmad MZ, Khan JA, Arshad MI. Crosstalk of liver immune cells and cell death mechanisms in different murine models of liver injury and its clinical relevance. Hepatobiliary Pancreat Dis Int. 2017;16(3):245–56. Epub 2017/06/13. pmid:28603092; PubMed Central PMCID: PMC7172563.
- 11. Lee J, Lee JS, Park SH, Shin SA, Kim K. Cohort Profile: The National Health Insurance Service-National Sample Cohort (NHIS-NSC), South Korea. Int J Epidemiol. 2017;46(2):e15. Epub 2016/01/30. pmid:26822938.
- 12. Bedogni G, Bellentani S, Miglioli L, Masutti F, Passalacqua M, Castiglione A, et al. The Fatty Liver Index: a simple and accurate predictor of hepatic steatosis in the general population. BMC Gastroenterol. 2006;6:33. Epub 20061102. pmid:17081293; PubMed Central PMCID: PMC1636651.
- 13. Grundy SM, Cleeman JI, Daniels SR, Donato KA, Eckel RH, Franklin BA, et al. Diagnosis and management of the metabolic syndrome: an American Heart Association/National Heart, Lung, and Blood Institute Scientific Statement. Circulation. 2005;112(17):2735–52. Epub 20050912. pmid:16157765.
- 14. Zhang L, Zhang X, Meng H, Li Y, Han T, Wang C. Obstructive sleep apnea and liver injury in severely obese patients with nonalcoholic fatty liver disease. Sleep Breath. 2020;24(4):1515–21. Epub 2020/02/01. pmid:32002742.
- 15. Umbro I, Fabiani V, Fabiani M, Angelico F, Del Ben M. Association between non-alcoholic fatty liver disease and obstructive sleep apnea. World J Gastroenterol. 2020;26(20):2669–81. Epub 2020/06/12. pmid:32523319; PubMed Central PMCID: PMC7265151.
- 16. Wei YT, Lee PY, Lin CY, Chen HJ, Lin CC, Wu JS, et al. Non-alcoholic fatty liver disease among patients with sleep disorders: a Nationwide study of Taiwan. BMC Gastroenterol. 2020;20(1):32. Epub 2020/02/12. pmid:32041532; PubMed Central PMCID: PMC7011431.
- 17. Kim SE, Park BS, Park SH, Shin KJ, Ha SY, Park J, et al. Predictors for Presence and Severity of Obstructive Sleep Apnea in Snoring Patients: Significance of Neck Circumference. Journal of Sleep Medicine. 2015;12(2):34–8.
- 18. Roh JH, Lee H, Yun-Jeong B, Park CS, Kim HJ, Yoon SY. A nationwide survey of the association between nonalcoholic fatty liver disease and the incidence of asthma in Korean adults. PLoS One. 2022;17(1):e0262715. Epub 20220121. pmid:35061826; PubMed Central PMCID: PMC8782316.
- 19. Compalati E, Ridolo E, Passalacqua G, Braido F, Villa E, Canonica GW. The link between allergic rhinitis and asthma: the united airways disease. Expert Rev Clin Immunol. 2010;6(3):413–23. pmid:20441427.
- 20. Browning JD, Horton JD. Molecular mediators of hepatic steatosis and liver injury. J Clin Invest. 2004;114(2):147–52. pmid:15254578; PubMed Central PMCID: PMC449757.
- 21. Targher G. Relationship between high-sensitivity C-reactive protein levels and liver histology in subjects with non-alcoholic fatty liver disease. J Hepatol. 2006;45(6):879–81; author reply 81–2. Epub 20061004. pmid:17049665.
- 22. Gaines J, Vgontzas AN, Fernandez-Mendoza J, He F, Calhoun SL, Liao D, et al. Increased inflammation from childhood to adolescence predicts sleep apnea in boys: A preliminary study. Brain Behav Immun. 2017;64:259–65. Epub 20170418. pmid:28432036; PubMed Central PMCID: PMC7085276.
- 23. Maniaci A, Iannella G, Cocuzza S, Vicini C, Magliulo G, Ferlito S, et al. Oxidative Stress and Inflammation Biomarker Expression in Obstructive Sleep Apnea Patients. J Clin Med. 2021;10(2). Epub 2021/01/17. pmid:33451164; PubMed Central PMCID: PMC7828672.
- 24. Huang DQ, El-Serag HB, Loomba R. Global epidemiology of NAFLD-related HCC: trends, predictions, risk factors and prevention. Nat Rev Gastroenterol Hepatol. 2021;18(4):223–38. Epub 2020/12/23. pmid:33349658; PubMed Central PMCID: PMC8016738.
- 25. Cowan DC, Livingston E. Obstructive sleep apnoea syndrome and weight loss: review. Sleep Disord. 2012;2012:163296. Epub 20120123. pmid:23471145; PubMed Central PMCID: PMC3581237.
- 26. Stefan N, Häring H-U, Cusi K. Non-alcoholic fatty liver disease: causes, diagnosis, cardiometabolic consequences, and treatment strategies. The Lancet Diabetes & Endocrinology. 2019;7(4):313–24.
- 27. Ng SSS, Wong VWS, Wong GLH, Chu WCW, Chan TO, To KW, et al. Continuous Positive Airway Pressure Does Not Improve Nonalcoholic Fatty Liver Disease in Patients with Obstructive Sleep Apnea. A Randomized Clinical Trial. Am J Respir Crit Care Med. 2021;203(4):493–501. Epub 2020/09/15. pmid:32926803.
- 28. Yang BL, Wu WC, Fang KC, Wang YC, Huo TI, Huang YH, et al. External validation of fatty liver index for identifying ultrasonographic fatty liver in a large‐scale cross‐sectional study in Taiwan. PLoS ONE. 2015;10(3):e0120443. pmid:25781622.