Sex- and age-specific associations between abdominal fat and non-alcoholic fatty liver disease: a prospective cohort study

Figures

• 

  Figure 1. Cumulative incidence rates of NAFLD. (A and B) Cumulative incidence rates of NAFLD according to the tertiles of VAT and SAT areas in males (A) and females (B). The tertiles of the VAT area were 72.1 and 111.0 cm2 for males and 78.7 and 107.0 cm2 for females, respectively. The tertiles of the SAT area were 84.2 and 121.0 cm2 for males and 120.0 and 166.0 cm2 for females, respectively. T1–T3, tertiles 1–3.

• 

  Figure 2. aRRs (95% CIs) of VAT and SAT areas for incident NAFLD (per standard deviation increment). Model 1 was adjusted for age, education levels, smoking status, drinking status, and physical activity. Model 2 was adjusted for the variables in Model 1 plus BMI and VAT or SAT area. Model 3 was adjusted for the variables in Model 2 plus diabetes status, hypertension status, total cholesterol, triglyceride, and high-density lipoprotein cholesterol. The black boxes represent males, and the white boxes represent females. The standard deviations of the VAT area were 42.3 and 33.9 cm2 for males and females, respectively, and the standard deviations of the SAT area were 38.8 and 53.2 cm2 for males and females, respectively. The incidence rates of NAFLD for males and females were 16.6% (n = 74) and 31.0% (n = 196) in participants aged < 60 years and 24.5% (n = 186) and 29.6% (n = 294) in participants aged ≥60 years, respectively.

• 

  Figure 3. The non-linear association between VAT or SAT areas and incident NAFLD. (A and B) The non-linear association between VAT or SAT areas and incident NAFLD in males (A) and females (B). The spline analysis was performed after excluding the highest 2.5% and lowest 2.5% of VAT and SAT levels with adjustment for age, education levels, smoking status, drinking status, physical activity, diabetes status, hypertension status, BMI, VAT or SAT area, total cholesterol, triglyceride, and high-density lipoprotein cholesterol.

• 

  Figure 4. Exploratory mediation analyses. (A and B) Exploratory mediation analyses were performed in males (A) and females (B), respectively. Statistically significant and non-significant paths are presented with regression coefficients in black and gray, respectively. Mediators highlighted in gray boxes met the assumptions for mediation analysis.

• 

  Figure 5. Flowchart of the study population and analyses.