Home » Associations of physical activity domains and muscle strength exercise with non-alcoholic fatty liver disease: a nation-wide cohort study – Scientific Reports

Associations of physical activity domains and muscle strength exercise with non-alcoholic fatty liver disease: a nation-wide cohort study – Scientific Reports

Study design, setting, and participants

The KNHANES is a nation-wide surveillance system to monitor the health and nutritional status of the general population of South Korea19. Each year, representative samples of approximately 10,000 people are selected. Health examination, health interview, and nutritional survey are then conducted.

We screened a total of 28,194 adult men or women aged 20–79 years who participated in the KNHANES from January 2014 to December 2018. Among them, we excluded 4446 participants who met the following exclusion criteria to include participants without chronic viral hepatitis, liver cirrhosis, heavy alcohol use, or malignancy: (1) chronic hepatitis B (n = 893, determined by the presence of hepatitis B surface antigen); (2) chronic hepatitis C (n = 73, determined by the presence of hepatitis C virus RNA test or history of chronic hepatitis C); (3) liver cirrhosis (n = 42, determined by a history of liver cirrhosis); 4) heavy alcohol intake (n = 2096, 30 g or more for a day for men and 20 g or more for a day for women)6; (5) history of malignancy (n = 1200); and (6) pregnant women (n = 142). Of these participants, we further excluded 2723 participants missing key variables for assessing NAFLD [n = 1714: missing values for alanine aminotransferase (n = 1126), heights (n = 44), body weights (n = 2), and alcohol intake (n = 542)] or missing key information on physical activity (n = 1019). Finally, a total of 21,025 participants were analyzed (Fig. 1). The survey was conducted after receiving written informed consent from all study participants. The study protocol was reviewed and approved by the Institutional Review Board of the Korea Disease Control and Prevention Agency (No: 2013-12EXP-03-5C, 2018-01-03-P-A) and the Samsung Medical Center (No: 2021-01-013). The study was performed in accordance with the Declaration of Helsinki.

Figure 1

Flowchart showing the selection of study participants.

Study outcomes, variables, and measurements

The diagnosis of NAFLD was made using hepatic steatosis index (HSI)20. HSI consists of aspartate aminotransferase, alanine aminotransferase, sex, body mass index (BMI), and diabetes mellitus. Participants with HSI of 36 or higher were considered to have NAFLD. The health interview including physical activity was conducted by trained surveyors consisting of nurses and epidemiologists. To gather comprehensive physical activity information, the Global Physical Activity Questionnaire (GPAQ) and frequency of muscle strength exercise were collected. The level of physical activity was interviewed using the Korean version of GPAQ21. The GPAQ was originally developed by WHO to monitor physical activity in numerous countries. It is grouped into three domains of physical activity: recreation, travel, and work activities. The recreation domain includes sports, fitness, and leisure activities. The travel domain includes transport to and from places. The work domain has paid or unpaid work, study/training, household chores, harvesting food/crops, fishing or hunting for food, and seeking employment. The GPAQ provides information on the frequency (times in a week) and duration (minutes at a day) of each domain of physical activity. In recreation and work domains, the intensity of physical activity was also provided (moderate or vigorous). Results were analyzed as suggested by the WHO: (1) the duration of vigorous physical activity was doubled and added to the duration of moderate physical activity, (2) three domains of physical activity were summed to calculate the duration of total physical activity. Since the WHO guideline states that all adults should do at least 150 min/week of moderate-intensity aerobic physical activity, total and each domain of physical activity were divided into < 150 min/week and ≥ 150 min/week. To investigate additional benefits of ≥ 300 min/week of physical activity, we performed further categorization: 0 min/week, 1–149 min/week, 150–299 min/week, and ≥ 300 min/week. The frequency of muscle strength exercise was determined by the number of muscle strength exercise in a week. Since the WHO 2020 guideline recommends that all adults should perform muscle strength exercise at least twice a week9, muscle strength exercise was categorized as < 2/week and ≥ 2/week.

Variables collected were age, sex, height, body weight, waist circumference, triglyceride, high density cholesterol, systolic blood pressure, diastolic blood pressure, fasting serum glucose, hepatitis B surface antigen, anti-HCV antibody, history of chronic hepatitis C, history of malignancy, history of liver cirrhosis, use of antihypertensive medications, antidiabetic medications, antidyslipidemic medications, alcohol use behavior, smoking status, pregnancy, household income information, and education level. BMI was calculated as weight in kilograms/height in square meters (kg/m2). Household income information was classified into quartiles: low, low-intermediate, intermediate-high, and high. Education levels were stratified into four categories: elementary school or lower, middle school, high school, college or higher. Alcohol intake was categorized into < 10 g/day and ≥ 10 g/day. Metabolic syndrome was defined for participants having three or more of the followings: (1) elevated waist circumference: ≥ 90 cm for men and ≥ 85 cm for women, (2) elevated triglycerides: ≥ 150 mg/dL or use of antidyslipidemic medications, (3) low high-density HDL-C: < 40 mg/dL for men,< 50 mg/dL for women, (4) elevated blood pressure: ≥ 130/85 mmHg or use of blood pressure lowering agents, (5) elevated fasting glucose: ≥ 100 mg/dL or on treatment for elevated glucose.

Statistical analysis

Descriptive statistics for continuous variables are presented as median and interquartile range (IQR). Categorical variables are presented as numbers and proportions (%). Comparison of variables between groups was performed using Student’s t-test or Wilcox rank-sum test for continuous variables and Chi-square test for categorical variables. Generalized logistic regression was performed to determine whether the prevalence of NAFLD was different depending on physical activity after adjusting for potential confounding or mediating factors. When adjusting for age and sex, we used age per year as a continuous variable. In the fully adjusted model, we further adjusted for BMI (continuous), metabolic syndrome (yes vs. no), income levels (low, low-intermediate, intermediate-high, high), education levels (elementary or lower, middle school, high school, college or higher), smoking (current, ex-smoker, and never smoker), alcohol intake (< 10 g/day vs. ≥ 10 g/day), total physical activity (< 150 min/week vs. ≥ 150 min/week), and muscle strength exercise (< 2/week vs. ≥ 2/week). When specific domains of physical activity were assessed, other specific domains were adjusted. For recreation, travel (< 150 min/week vs. ≥ 150 min/week) and work activity (< 150 min/week vs. ≥ 150 min/week) were adjusted. For travel, recreation (< 150 min/week vs. ≥ 150 min/week) and work activity (< 150 min/week vs. ≥ 150 min/week) were adjusted. For work, recreation (< 150 min/week vs. ≥ 150 min/week) and travel activity (< 150 min/week vs. ≥ 150 min/week) were adjusted.

Subgroup analysis was performed to evaluate the relationship between physical activity or muscle strength exercise and NAFLD within each subgroup. Subgroups were predefined as follows: by age (< 65 years vs. ≥ 65 years), sex (male vs. female), BMI (< 25 kg/m2 vs. ≥ 25 kg/m2), metabolic syndrome (yes vs. no), muscle strength exercise (< 2/week vs. ≥ 2/week), and total physical activity (< 150 min/week vs. ≥ 150 min/week). All variables with a p value < 0.05 were considered statistically significant. All statistical analyses were performed using R version 3.6.3 (The R Foundation for Statistical Computing, Vienna, Austria).