2021/12/14 by Laurens A. van Kleef, Amy Hofman, Trudy Voortman +1 · 37 citations
Medicine · #Confidence interval #Diabetes, Cardiovascular Risks, and Lipoproteins #Disease #Environmental health #Fatty liver #Homeostatic model assessment #Insulin resistance #Internal medicine #Liver Disease Diagnosis and Treatment #Liver Disease and Transplantation #Medicine #Nonalcoholic fatty liver disease #Obesity #Odds ratio #Population #Steatosis #Waist
paper · doi:10.14309/ajg.0000000000001584
published in The American Journal of Gastroenterology 117(2), 311-318 (Lippincott Williams & Wilkins)
openalex publication_date 2021/12/14 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/26
INTRODUCTION: The disease burden of nonalcoholic fatty liver disease (NAFLD) increases rapidly, in line with the obesity pandemic. Physical activity has been linked to a lower risk of NAFLD. However, the impact of different intensities of activity and sedentary behavior and whether their effects on NAFLD are explained by metabolic health remain unclear. METHODS: We performed cross-sectional analyses within the population-based Rotterdam Study cohort. Abdominal ultrasound and accelerometry data were collected between 2009 and 2014. NAFLD was defined as hepatic steatosis diagnosed by ultrasound, in the absence of secondary causes for steatosis: viral hepatitis, steatogenic drugs, and excessive alcohol. We categorized accelerometry data into sedentary time and light, moderate, and vigorous physical activities. RESULTS: We included 667 participants (aged 63.3 ± 6.3 years, 53% female individuals), and 34.3% had NAFLD. Total physical activity was associated with lower NAFLD prevalence adjusted for demographic, lifestyle, and socioeconomic factors (odds ratio: 0.958 per 10 min/d, 95% confidence interval [CI]: 0.929-0.986). More intensive physical activity was more strongly associated with lower NAFLD prevalence: odds ratios for light, moderate, and vigorous physical activities were 0.931 (95% CI: 0.882-0.982), 0.891 (95% CI: 0.820-0.967), and 0.740 (95% CI: 0.600-0.906) per 10 min/d, respectively. These associations were explained by metabolic health, particularly homeostatic model assessment of insulin resistance (proportion mediated: 0.59, P < 0.001) and waist circumference (proportion mediated: 1.08, P < 0.001). Beyond this indirect effect, no direct effect could be demonstrated (P = 0.282-0.827). DISCUSSION: Physical activity at each intensity is inversely associated with NAFLD prevalence, with larger effects for higher intensities of physical activity. This association is mediated by better metabolic health, mainly lower insulin resistance and waist circumference. Physical activity should therefore be incorporated into NAFLD disease management and prevention programs.