Introduction: Nonalcoholic fatty liver disease (NAFLD) is associated with vascular dysfunction, one of the signs of which is arterial stiffness. Carotid-femoral pulse wave velocity (PWV), which is considered the gold standard measure of arterial stiffness, can be estimated using two commonly assessed clinical variables: age and blood pressure. This study aimed to evaluate the association between estimated PWV (ePWV) and the prevalence and incidence of NAFLD among Korean adults. Methods: This study used data from the Ansan-Ansung cohort study, a subset of the Korean Genome and Epidemiology Study, and included 8,336 adult participants with and without NAFLD at baseline. The participants were subdivided into three tertile groups according to ePWV. Results: At baseline, the prevalence of NAFLD was 10.5, 27.5, and 35.0% in the first (lowest), second, and third (highest) tertiles of ePWV, respectively. During the 18-year follow-up period, 2,467 (42.9%) incident cases of NAFLD were identified among 5,755 participants who did not have NAFLD at baseline. After adjustment for clinically relevant variables, participants in the second (adjusted hazard ratio [HR], 1.25; 95% confidence interval [CI], 1.12–1.40) and third (adjusted HR, 1.42; 95% CI, 1.24–1.64) tertiles of ePWV had a significantly higher risk of incident NAFLD than those in the first tertile. Conclusion: Higher ePWV is independently associated with an elevated risk of NAFLD in the general population.

Nonalcoholic fatty liver disease (NAFLD) is the predominant chronic liver disorder worldwide, with an estimated prevalence of approximately 30% [1]. The incidence of NAFLD continues to increase in both men and women [2]. In approximately one in every four to five individuals with NAFLD, the condition progresses to nonalcoholic steatohepatitis, which is the second leading cause of liver disease necessitating transplantation [3, 4]. However, the leading cause of death in patients with NAFLD is not liver disease but cardiovascular disease (CVD) [5]. This is because many risk factors for NAFLD, including diabetes mellitus, hypertension, dyslipidemia, obesity, and metabolic syndrome, are also risk factors for atherosclerotic CVD [3, 5]. Moreover, increasing evidence suggests that NAFLD is independently associated with subclinical and clinical CVD [6, 7].

Arterial stiffness is a surrogate marker of atherosclerotic changes in the arterial wall and a predictor of cardiovascular events and mortality [8‒10]. Previous research has revealed an association between arterial stiffness, measured based on carotid-femoral pulse wave velocity (cfPWV) or brachial-ankle PWV (baPWV), and NAFLD [11]. cfPWV and baPWV are the two most commonly used measures of arterial stiffness, and current hypertension guidelines recommend them as markers of hypertension-mediated organ damage [12‒14]. However, measurement of cfPWV and baPWV requires specialized equipment and trained personnel, thereby limiting their widespread use [15]. Recent studies have suggested estimated PWV (ePWV) as a simple and easily obtainable estimator of cfPWV. ePWV is calculated from age and mean blood pressure (MBP) using the equation published by the Reference Values for Arterial Stiffness Collaboration and has been shown to have a predictive value comparable to that of cfPWV [14, 16]. Although some reports have explored the association between cfPWV or baPWV and NAFLD, no evidence supporting the utility of ePWV in predicting NAFLD currently exists. Therefore, this study investigated the association between ePWV and incident NAFLD in middle-aged Korean adults.

Study Population

This study used data from the Ansung-Ansan cohort, which comprised 10,030 South Koreans aged 40–69 years residing in Ansung and Ansan. The Ansung-Ansan cohort study was initiated in 2001 and integrated into the Korean Genome and Epidemiology Study, a government-funded population-based cohort study that aimed to investigate genetic and environmental factors underlying prevalent cardiovascular and metabolic diseases in South Korea. The detailed protocol of the Ansan-Ansung cohort study has been previously published [17]. Briefly, the participants underwent comprehensive health examinations during which demographic, social, and past medical information was collected. In addition, the participants underwent physical examination, laboratory tests (urine and blood), electrocardiography, and chest radiology at a tertiary hospital in the region. This study encompassed eight sequential follow-up evaluations that fulfilled the cohort protocol. Follow-up assessments after the baseline evaluation were conducted through scheduled biennial hospital visits until 2018.

From the initial 10,030 participants in the baseline survey, we sequentially excluded individuals who met the following exclusion criteria: (1) history of hepatitis (n = 423); (2) alcohol consumption >30 g/day in men or 20 g/day in women (n = 964); and (3) missing data required for the calculation of the NAFLD liver fat score (n = 307). The remaining 8,336 participants were included in the analysis of the association between ePWV and the prevalence of NAFLD. Additionally, to investigate the association between ePWV and the incidence of NAFLD, we analyzed data from a subset comprising 5,775 participants who did not have NAFLD at baseline. This subset was obtained by excluding (1) individuals with NAFLD at baseline (n = 2,030) and (2) those without follow-up assessments after the baseline survey (n = 531) (shown in Fig. 1).

Fig. 1.

Study flowchart. NAFLD, nonalcoholic fatty liver disease; ePWV, estimated pulse wave velocity.

Fig. 1.

Study flowchart. NAFLD, nonalcoholic fatty liver disease; ePWV, estimated pulse wave velocity.

Close modal

Data Collection

Demographic, social, and past medical information, including smoking status, alcohol consumption, education, income level, marital status, medical conditions (e.g., hypertension, diabetes mellitus, dyslipidemia, and chronic kidney disease), and history of CVD (defined as a composite of ischemic heart disease, hemorrhagic stroke, ischemic stroke, heart failure, and peripheral artery disease), was collected using a questionnaire administered by trained investigators at the tertiary hospital. The questionnaire also gathered data on physical activity, including regular exercise, types of exercise, weekly exercise frequency and duration, daily physical activities, and duration of physical activities. The total physical activity per week was calculated as the sum of the metabolic equivalent task scores of exercise and routine physical activities per week.

Hypertension was defined as a systolic blood pressure (SBP) of ≥140 mm Hg, a diastolic blood pressure (DBP) of ≥90 mm Hg, or the use of antihypertensive medications [12]. Diabetes mellitus was defined as a fasting blood glucose level of ≥126 mg/dL, a hemoglobin A1c level of ≥6.5%, or the use of antidiabetic medications [18]. Dyslipidemia was defined as a total cholesterol level of ≥240 mg/dL, a low-density lipoprotein cholesterol level of ≥160 mg/dL, a triglyceride level of ≥200 mg/dL, a high-density lipoprotein (HDL) cholesterol level of <40 mg/dL in men or <50 mg/dL in women, or the use of lipid-lowering medications [19].

Trained examiners measured the participants’ blood pressure (BP) using a mercury sphygmomanometer, at the heart level in a sitting position after at least 5 min of quiet rest. At least two BP readings with an interval of at least 1 min were obtained, and the average was recorded. Blood samples were collected after an overnight fast and analyzed using an automated analyzer (Hitachi Automatic Analyzer 7600; Hitachi, Tokyo, Japan). Fasting blood glucose, serum insulin, HDL cholesterol, triglyceride, aspartate aminotransferase (AST), alanine aminotransferase, and hemoglobin A1c levels were measured. Low-density lipoprotein cholesterol level was determined using the Friedewald equation when the triglyceride level was <400 mg/dL [20]. The estimated glomerular filtration rate was determined using the Modification of Diet in Renal Disease equation [21].

NAFLD Definition

NAFLD was defined using the NAFLD liver fat score, which was calculated using the following formula: −2.89 + 1.18 × metabolic syndrome (1 if yes, 0 if no) + 0.45 × diabetes mellitus (2 if yes, 0 if no) + 0.15 × insulin (mU/L) + 0.04 × AST (U/L) − 0.94 × AST/alanine aminotransferase. The cutoff value of the NAFLD liver fat score for NAFLD was >−0.640 [22].

Metabolic syndrome components were defined in accordance with the harmonizing worldwide consensus criteria [23]. Metabolic syndrome was identified when three or more of the following components were present: (1) abdominal obesity determined through waist circumference measurements, with a threshold of ≥90 cm for men or ≥85 cm for women, following Korean-specific cutoff values established by the Korean Society of Obesity [24]; (2) hypertriglyceridemia, defined as a serum triglyceride level of ≥150 mg/dL or specific treatment for hypertriglyceridemia; (3) low HDL cholesterol, defined as a serum HDL cholesterol level of <40 mg/dL for men or <50 mg/dL for women or specific treatment for low HDL cholesterol; (4) high BP, defined as an SBP of ≥130 mm Hg and a DBP of ≥85 mm Hg or the use of antihypertensive medications; and (5) elevated fasting blood glucose level, defined as a fasting blood glucose level of ≥100 mg/dL or the use of antidiabetic medications.

ePWV Calculation

We calculated ePWV using the method described by Greve et al. [16], applying the following two formulas based on the presence or absence of cardiovascular risk factors. We calculated ePWV for individuals with one or more cardiovascular risk factors using the following equation: 9.58748315543126 − 0.402467539733184 × age + 4.56020798207263 × 10−3 × age2 − 2.6207705511664 × 10−5 × age2 × MBP + 3.1762450559276 × 10−3 × age × MBP − 1.83215068503821 × 10−2 × MBP.

Additionally, for participants without cardiovascular risk factors, defined as nonsmokers without any metabolic syndrome component or a history of CVD, we calculated ePWV using the following equation: 4.62 − 0.13 × age + 0.0018 × age2 + 0.0006 × age × MBP + 0.0284 × MBP. MBP was calculated as DBP + 0.4 × (SBP − DBP). We categorized ePWV into tertiles (<7.69, 7.69–9.39, and >9.39 m/s) to analyze the prevalence of NAFLD among participants with and without NAFLD at baseline. To analyze the incidence of NAFLD in participants without NAFLD at baseline, we recategorized ePWV into the following tertile categories: <7.46, 7.46–9.07, and >9.07 m/s.

Statistical Analysis

The baseline characteristics of the 8,336 participants with and without NAFLD were compared based on ePWV tertiles. Additionally, the baseline characteristics of participants were compared according to the presence of NAFLD. Moreover, the baseline characteristics of the 5,775 participants without NAFLD at baseline were compared according to ePWV tertiles.

Categorical variables are presented as numbers and percentages, whereas continuous variables are expressed as means with standard deviations or medians with interquartile ranges, depending on their distribution. Continuous variables were compared among the three groups based on tertiles of ePWV using either one-way analysis of variance or the Kruskal-Wallis test, followed by Tukey’s post hoc test or Dunn’s multiple-comparison test, as appropriate. For the comparison between two groups according to the presence of NAFLD, Student’s t test or the Mann-Whitney U test was employed. The χ2 or Fisher’s exact test was used to compare categorical variables. The distribution of continuous variables was assessed using the Kolmogorov-Smirnov test and through visual inspection of Q-Q plots.

For participants with and without NAFLD at baseline, we assessed the predictive ability of ePWV for NAFLD prevalence using a receiver operating characteristic (ROC) curve analysis. The optimal ePWV cutoff value for predicting the prevalence of NAFLD was determined using Youden’s index. Additionally, logistic regression analyses were conducted to assess the association between NAFLD prevalence and ePWV while adjusting for potential confounding variables. We selected clinically relevant metabolic risk factors as confounding variables and constructed the adjusted models in the following order: model 1 was adjusted for age (per 1 year) and sex. Model 2 included the variables from model 1 along with sociodemographic factors, such as SBP, body mass index (BMI), smoking status, alcohol consumption, physical activity (per 1 metabolic equivalent task-h/week), income level, and education. Furthermore, model 3 incorporated comorbid conditions associated with NAFLD (e.g., hypertension, diabetes mellitus, dyslipidemia, and chronic kidney disease) in addition to the variables from model 2. The analysis was performed using cutoff values obtained from ROC curve analysis, ePWV tertiles, and ePWV as a continuous variable.

For participants without NAFLD at baseline, we assessed the predictive ability of ePWV for NAFLD incidence using a time-dependent ROC curve analysis during the follow-up period. The optimal cutoff value for predicting incident NAFLD was determined using Youden’s index. A Kaplan-Meier curve analysis was conducted to estimate the cumulative incidence of NAFLD based on ePWV tertiles. The independent predictive value of ePWV for incident NAFLD during the follow-up period was determined using a Cox proportional hazards regression model, considering potential confounding variables. The same models as described above were applied in the multivariable analysis. The analysis was performed using both the cutoff values derived from the time-dependent ROC curve analysis and the tertiles of ePWV. Furthermore, subgroup analyses were performed based on age (<50 or ≥50 years), sex, BMI (<23, 23–24.9, or ≥25 kg/m2), presence or absence of metabolic syndrome, presence or absence of insulin resistance (homeostatic model assessment for insulin resistance >2.5 or ≤2.5), presence of hypertension, and presence of diabetes mellitus.

All statistical analyses were conducted using the open-source statistical software R (version 4.3.1, www.R-project.org) and RStudio (version 2023.06.2, www.rstudio.com), as well as statistical packages including tableone, pROC, timeROC, rms, survminer, and survival. A p value of <0.05 was considered statistically significant.

Baseline Characteristics

A total of 8,336 participants with and without NAFLD at baseline (men, 41.7%; mean age, 52.37 ± 8.91 years) were analyzed. Table 1 lists the baseline characteristics of all participants with and without NAFLD at baseline categorized by ePWV tertiles. Age showed an increasing trend from the first tertile (44.27 ± 3.71 years) to the third tertile (61.59 ± 5.73 years) (p < 0.001). The proportion of men was higher in the second tertile (47.1%) than in the first (37.7%) and third (40.3%) tertiles (p < 0.001). BMI was higher in the second and third tertiles than in the first tertile. Waist circumference, physical activity, SBP, and DBP gradually increased across the tertiles (p < 0.001 for all). Income level, education, smoking status, and alcohol consumption demonstrated significant differences across the tertiles (p < 0.001 for all). A history of medical conditions, including hypertension, diabetes mellitus, dyslipidemia, and CVD, was significantly more prevalent in the third tertile (p < 0.001 for all except dyslipidemia [p = 0.005]). The results of laboratory tests indicated lower estimated glomerular filtration rate, higher fasting blood glucose levels, and worse lipid profile in the higher ePWV tertiles (p < 0.001 for all except HDL cholesterol [p = 0.02]). The prevalence of metabolic syndrome was significantly higher in the third tertile than in the first and second tertiles (p < 0.001), and NAFLD liver fat scores were significantly different across the tertiles (p < 0.001).

Table 1.

Baseline characteristics of participants with and without NAFLD

ePWVp value
first tertile (<7.69 m/s), n = 2,779second tertile (7.69–9.39 m/s), n = 2,781third tertile (>9.39 m/s), n = 2,776
Age, years 44.27±3.71ab 51.26±6.28* 61.59±5.73 <0.001 
Men, n (%) 1,047 (37.7) 1,311 (47.1) 1,120 (40.3) <0.001 
BMI, kg/m2 24.02±2.87ab 24.85±3.14 24.76±3.34 <0.001 
Waist circumference, cm 79.17±8.28ab 83.46±8.37* 85.37±8.88 <0.001 
Income level, n (%)    <0.001 
 ≥Median 1,919 (70.0) 1,343 (49.4) 623 (23.0)  
Education, n (%)    <0.001 
 Lower than middle school 326 (11.8) 869 (31.5) 1,674 (61.2)  
 Middle school 665 (24.1) 723 (26.2) 508 (18.6)  
 High school 1,255 (45.4) 820 (29.7) 397 (14.5)  
 University and college 519 (18.8) 348 (12.6) 157 (5.7)  
Smoking status, n (%)    <0.001 
 Current smoker 622 (22.7) 655 (23.8) 533 (19.5)  
 Previous smoker 307 (11.2) 429 (15.6) 405 (14.8)  
 Never smoker 1,810 (66.1) 1,663 (60.5) 1,792 (65.6)  
Alcohol consumption, n (%)    <0.001 
 Current drinker 1,253 (45.4) 1,214 (44.1) 951 (34.7)  
 Previous drinker 159 (5.8) 199 (7.2) 222 (8.1)  
 Never drinker 1,348 (48.8) 1,342 (48.7) 1,569 (57.2)  
Physical activity, MET-h/week 150.50±90.58ab 171.14±104.60* 190.76±112.15 <0.001 
SBP, mm Hg 108.54±9.52ab 123.35±11.80* 140.63±17.78 <0.001 
DBP, mm Hg 72.18±7.64ab 82.45±9.19* 89.24±11.01 <0.001 
Medical history, n (%) 
 Hypertension 32 (1.2) 252 (9.1) 670 (24.1) <0.001 
 Diabetes mellitus 30 (1.1) 63 (2.3) 95 (3.4) <0.001 
 Dyslipidemia 7 (0.3) 12 (0.4) 24 (0.9) 0.005 
 Chronic kidney disease 65 (2.3) 83 (3.0) 92 (3.3) 0.086 
 CVD 34 (1.2) 65 (2.3) 167 (6.0) <0.001 
Laboratory data 
 eGFR, mL/min/1.73 m2 93.3±20.1ab 89.4±19.8* 86.6±20.4 <0.001 
 Fasting glucose, mg/dL 81 (76–87)ab 82 (77–90)* 83 (78–92) <0.001c 
 Hemoglobin A1c, % 5.5 (5.3–5.7)ab 5.6 (5.4–5.9)* 5.7 (5.5–6.0) <0.001c 
 Total cholesterol, mg/dL 182 (161–206)ab 191 (169–217) 193 (171–218) <0.001c 
 Triglyceride, mg/dL 115 (89–160)ab 140 (102–195)* 145 (109–199) <0.001c 
 HDL cholesterol, mg/dL 44 (38–50)a 43 (37–50) 43 (37–50) 0.02c 
 LDL cholesterol, mg/dL 116 (97–138)ab 123 (103–146) 125 (105–148) <0.001c 
Metabolic syndrome, n (%) 265 (9.5) 898 (32.3) 1,382 (49.8) <0.001 
HOMA-IR 1.38 (1.02–1.90)ab 1.48 (1.05–2.06)* 1.50 (1.07–2.18) <0.001 
NAFLD liver fat score −1.94 (−2.36 to −1.38)ab −1.54 (−2.18 to −0.52)* −1.13 (−2.00 to −0.30) <0.001c 
ePWVp value
first tertile (<7.69 m/s), n = 2,779second tertile (7.69–9.39 m/s), n = 2,781third tertile (>9.39 m/s), n = 2,776
Age, years 44.27±3.71ab 51.26±6.28* 61.59±5.73 <0.001 
Men, n (%) 1,047 (37.7) 1,311 (47.1) 1,120 (40.3) <0.001 
BMI, kg/m2 24.02±2.87ab 24.85±3.14 24.76±3.34 <0.001 
Waist circumference, cm 79.17±8.28ab 83.46±8.37* 85.37±8.88 <0.001 
Income level, n (%)    <0.001 
 ≥Median 1,919 (70.0) 1,343 (49.4) 623 (23.0)  
Education, n (%)    <0.001 
 Lower than middle school 326 (11.8) 869 (31.5) 1,674 (61.2)  
 Middle school 665 (24.1) 723 (26.2) 508 (18.6)  
 High school 1,255 (45.4) 820 (29.7) 397 (14.5)  
 University and college 519 (18.8) 348 (12.6) 157 (5.7)  
Smoking status, n (%)    <0.001 
 Current smoker 622 (22.7) 655 (23.8) 533 (19.5)  
 Previous smoker 307 (11.2) 429 (15.6) 405 (14.8)  
 Never smoker 1,810 (66.1) 1,663 (60.5) 1,792 (65.6)  
Alcohol consumption, n (%)    <0.001 
 Current drinker 1,253 (45.4) 1,214 (44.1) 951 (34.7)  
 Previous drinker 159 (5.8) 199 (7.2) 222 (8.1)  
 Never drinker 1,348 (48.8) 1,342 (48.7) 1,569 (57.2)  
Physical activity, MET-h/week 150.50±90.58ab 171.14±104.60* 190.76±112.15 <0.001 
SBP, mm Hg 108.54±9.52ab 123.35±11.80* 140.63±17.78 <0.001 
DBP, mm Hg 72.18±7.64ab 82.45±9.19* 89.24±11.01 <0.001 
Medical history, n (%) 
 Hypertension 32 (1.2) 252 (9.1) 670 (24.1) <0.001 
 Diabetes mellitus 30 (1.1) 63 (2.3) 95 (3.4) <0.001 
 Dyslipidemia 7 (0.3) 12 (0.4) 24 (0.9) 0.005 
 Chronic kidney disease 65 (2.3) 83 (3.0) 92 (3.3) 0.086 
 CVD 34 (1.2) 65 (2.3) 167 (6.0) <0.001 
Laboratory data 
 eGFR, mL/min/1.73 m2 93.3±20.1ab 89.4±19.8* 86.6±20.4 <0.001 
 Fasting glucose, mg/dL 81 (76–87)ab 82 (77–90)* 83 (78–92) <0.001c 
 Hemoglobin A1c, % 5.5 (5.3–5.7)ab 5.6 (5.4–5.9)* 5.7 (5.5–6.0) <0.001c 
 Total cholesterol, mg/dL 182 (161–206)ab 191 (169–217) 193 (171–218) <0.001c 
 Triglyceride, mg/dL 115 (89–160)ab 140 (102–195)* 145 (109–199) <0.001c 
 HDL cholesterol, mg/dL 44 (38–50)a 43 (37–50) 43 (37–50) 0.02c 
 LDL cholesterol, mg/dL 116 (97–138)ab 123 (103–146) 125 (105–148) <0.001c 
Metabolic syndrome, n (%) 265 (9.5) 898 (32.3) 1,382 (49.8) <0.001 
HOMA-IR 1.38 (1.02–1.90)ab 1.48 (1.05–2.06)* 1.50 (1.07–2.18) <0.001 
NAFLD liver fat score −1.94 (−2.36 to −1.38)ab −1.54 (−2.18 to −0.52)* −1.13 (−2.00 to −0.30) <0.001c 

Categorical variables are presented as number (%), and continuous variables are presented as mean ± standard deviation or median (interquartile range), as appropriate.

NAFLD, nonalcoholic fatty liver disease; BMI, body mass index; eGFR, estimated glomerular filtration rate; HDL, high-density lipoprotein; LDL, low-density lipoprotein; HOMA-IR, homeostatic model assessment for insulin resistance; MET, metabolic equivalent task.

*Post hoc p: second tertile versus third tertile, statistically significant (p < 0.05).

aPost hoc p: first tertile versus second tertile, statistically significant (p < 0.05).

bPost hoc p: first tertile versus third tertile, statistically significant (p < 0.05).

cAssessed using a nonparametric test.

Among the baseline characteristics of participants without NAFLD at baseline (n = 5,775) categorized by ePWV tertiles (online suppl. Table S1; for all online suppl. material, see https://doi.org/10.1159/000535580), age, BMI, waist circumference, physical activity, SBP, DBP, income level, education, smoking status, alcohol consumption, and laboratory test results showed a similar pattern to the characteristics of participants with and without NAFLD at baseline categorized by ePWV tertiles. However, in terms of medical history, differences between the groups were observed only for hypertension and CVD. Baseline characteristics of participants with or without NAFLD at baseline are also provided in online supplementary Table S2. Participants with NAFLD at baseline were older and had a higher prevalence of cardiometabolic risk factors such as higher BMI, hypertension, diabetes mellitus, and dyslipidemia.

Predictive Power of ePWV for NAFLD Prevalence

The area under the ROC curve of ePWV for predicting the prevalence of NAFLD was 0.667. The optimal cutoff value of ePWV for predicting NAFLD prevalence was 8.056 m/s, with a sensitivity of 0.781 and a specificity of 0.480 (shown in online suppl. Fig. 1). Table 2 shows the prevalence of NAFLD at baseline and the corresponding odds ratios (ORs) with 95% confidence intervals (CIs) based on the ePWV cutoff values and tertiles. The overall prevalence of NAFLD at baseline was 24.4%. Comparatively, participants with an ePWV of >8.056 m/s exhibited a higher prevalence of NAFLD (32.6%) than those with an ePWV of ≤8.056 m/s (12.8%). Additionally, a graded increase in NAFLD prevalence was observed with increasing ePWV tertile, ranging from 10.5% in the first tertile to 35.0% in the third tertile. After adjusting for confounding variables, participants with an ePWV of >8.056 m/s exhibited significantly higher odds of NAFLD (adjusted OR, 1.99; 95% CI, 1.64–2.42) than those with an ePWV of ≤8.056 m/s. Additionally, both the third (adjusted OR, 2.43; 95% CI, 1.73–3.42) and second (adjusted OR, 2.13; 95% CI, 1.74–2.62) tertiles of ePWV demonstrated significantly greater associations with NAFLD prevalence than the first tertile.

Table 2.

Prevalence of NAFLD and ORs for NAFLD prevalence according to ePWV tertiles

NAFLD prevalence, n (%)Unadjusted OR (95% CI)Model 1a (95% CI)Model 2b (95% CI)Model 3c (95% CI)
Overall 2,030 (24.4)     
 ePWV (per 1 m/s)  1.39 (1.35–1.43) 1.89 (1.79–2.00) 1.16 (1.09–1.22) 1.27 (1.07–1.51) 
ePWV cutoff value, m/s 
 ≤8.056 445 (12.8) Reference Reference Reference Reference 
 >8.056 1,585 (32.6) 3.28 (2.92–3.69) 4.28 (3.68–4.98) 1.94 (1.66–2.27) 1.99 (1.64–2.42) 
ePWV tertiles 
 First tertile 293 (10.5) Reference Reference Reference Reference 
 Second tertile 766 (27.5) 3.23 (2.79–3.74) 4.40 (3.74–5.16) 2.05 (1.72–2.44) 2.13 (1.74–2.62) 
 Third tertile 971 (35.0) 4.56 (3.95–5.27) 10.05 (8.09–12.49) 2.24 (1.81–2.77) 2.43 (1.73–3.42) 
NAFLD prevalence, n (%)Unadjusted OR (95% CI)Model 1a (95% CI)Model 2b (95% CI)Model 3c (95% CI)
Overall 2,030 (24.4)     
 ePWV (per 1 m/s)  1.39 (1.35–1.43) 1.89 (1.79–2.00) 1.16 (1.09–1.22) 1.27 (1.07–1.51) 
ePWV cutoff value, m/s 
 ≤8.056 445 (12.8) Reference Reference Reference Reference 
 >8.056 1,585 (32.6) 3.28 (2.92–3.69) 4.28 (3.68–4.98) 1.94 (1.66–2.27) 1.99 (1.64–2.42) 
ePWV tertiles 
 First tertile 293 (10.5) Reference Reference Reference Reference 
 Second tertile 766 (27.5) 3.23 (2.79–3.74) 4.40 (3.74–5.16) 2.05 (1.72–2.44) 2.13 (1.74–2.62) 
 Third tertile 971 (35.0) 4.56 (3.95–5.27) 10.05 (8.09–12.49) 2.24 (1.81–2.77) 2.43 (1.73–3.42) 

OR, odds ratio; CI, confidence interval; ePWV, estimated pulse wave velocity; NAFLD, nonalcoholic fatty liver disease; MET, metabolic equivalent task.

aModel 1: adjusted for age (per 1 year) and sex.

bModel 2: adjusted for age, sex, SBP (per 1 mm Hg), BMI, smoking status, alcohol consumption, physical activity (per 1 MET-h/week), income level, and education.

cModel 3: adjusted for age, sex, SBP (per 1 mm Hg), BMI, smoking status, alcohol status, physical activity (per 1 MET-h/week), income level, education, hypertension, diabetes mellitus, dyslipidemia, and chronic kidney disease.

Predictive Power of ePWV for NAFLD Incidence

During a median follow-up period of 187 months (interquartile range, 163–190 months), 2,467 (42.9%) incident cases of NAFLD were observed. The optimal cutoff value of ePWV for predicting NAFLD incidence was 7.354 m/s (area under the ROC curve, 0.596; sensitivity, 0.771; specificity, 0.394) (shown in Fig. 2). During the follow-up period, the cumulative incidence of NAFLD was the highest in the third tertile (50.4%) and the lowest in the first tertile (32.6%) of ePWV (shown in Fig. 3 and Table 3). In addition, the incidence of NAFLD was higher in participants with an ePWV of >7.354 m/s (49.1%) than in those with an ePWV of ≤7.354 m/s (36.0%). After adjustments for confounding variables, participants with an ePWV of >7.354 m/s showed a significantly higher risk of developing NAFLD (adjusted hazard ratio [HR], 1.21; 95% CI, 1.09–1.35) than those with an ePWV of ≤7.354 m/s in the multivariable Cox proportional hazards regression model. Furthermore, both the third (adjusted HR, 1.42; 95% CI, 1.24–1.64) and second (adjusted HR, 1.25; 95% CI, 1.12–1.40) tertiles of ePWV showed a significantly higher risk of NAFLD development than the first tertile.

Fig. 2.

Predictive ability of ePWV for incident NAFLD using a time-dependent ROC curve analysis during the follow-up period in participants without NAFLD at baseline. ePWV, estimated pulse wave velocity; AUC, area under the ROC curve; NAFLD, nonalcoholic fatty liver disease.

Fig. 2.

Predictive ability of ePWV for incident NAFLD using a time-dependent ROC curve analysis during the follow-up period in participants without NAFLD at baseline. ePWV, estimated pulse wave velocity; AUC, area under the ROC curve; NAFLD, nonalcoholic fatty liver disease.

Close modal
Fig. 3.

Kaplan-Meier curves comparing the cumulative incidence of NAFLD according to tertiles of ePWV in participants without NAFLD at baseline. NAFLD, nonalcoholic fatty liver disease.

Fig. 3.

Kaplan-Meier curves comparing the cumulative incidence of NAFLD according to tertiles of ePWV in participants without NAFLD at baseline. NAFLD, nonalcoholic fatty liver disease.

Close modal
Table 3.

Incidence of NAFLD and HRs for NAFLD incidence according to ePWV tertiles in participants without NAFLD at baseline

NAFLD incidence, n (%)Unadjusted HR (95% CI)Model 1a (95% CI)Model 2b (95% CI)Model 3c (95% CI)
Overall 2467 (42.9)     
 ePWV (per 1 m/s)  1.19 (1.16–1.21) 1.38 (1.32–1.44) 1.08 (1.04–1.12) 1.05 (1.01–1.10) 
ePWV cutoff value, m/s 
 ≤7.354 991 (36.0) Reference Reference Reference Reference 
 >7.354 1,476 (49.1) 1.85 (1.68–2.03) 1.42 (1.25–1.60) 1.26 (1.14–1.40) 1.21 (1.09–1.35) 
ePWV tertiles 
 First tertile 625 (32.6) Reference Reference Reference Reference 
 Second tertile 875 (45.6) 1.59 (1.43–1.76) 1.85 (1.66–2.07) 1.25 (1.12–1.40) 1.25 (1.12–1.40) 
 Third tertile 967 (50.4) 2.06 (1.86–2.27) 3.07 (2.62–3.60) 1.45 (1.26–1.67) 1.42 (1.24–1.64) 
NAFLD incidence, n (%)Unadjusted HR (95% CI)Model 1a (95% CI)Model 2b (95% CI)Model 3c (95% CI)
Overall 2467 (42.9)     
 ePWV (per 1 m/s)  1.19 (1.16–1.21) 1.38 (1.32–1.44) 1.08 (1.04–1.12) 1.05 (1.01–1.10) 
ePWV cutoff value, m/s 
 ≤7.354 991 (36.0) Reference Reference Reference Reference 
 >7.354 1,476 (49.1) 1.85 (1.68–2.03) 1.42 (1.25–1.60) 1.26 (1.14–1.40) 1.21 (1.09–1.35) 
ePWV tertiles 
 First tertile 625 (32.6) Reference Reference Reference Reference 
 Second tertile 875 (45.6) 1.59 (1.43–1.76) 1.85 (1.66–2.07) 1.25 (1.12–1.40) 1.25 (1.12–1.40) 
 Third tertile 967 (50.4) 2.06 (1.86–2.27) 3.07 (2.62–3.60) 1.45 (1.26–1.67) 1.42 (1.24–1.64) 

HR, hazard ratio; CI, confidence interval; ePWV, estimated pulse wave velocity; NAFLD, nonalcoholic fatty liver disease; MET, metabolic equivalent task.

aModel 1: adjusted for age (per 1 year) and sex.

bModel 2: adjusted for age, sex, SBP (per 1 mm Hg), BMI, smoking status, alcohol consumption, physical activity (per 1 MET-h/week), income level, and education.

cModel 3: adjusted for age, sex, SBP (per 1 mm Hg), BMI, smoking status, alcohol status, physical activity (per 1 MET-h/week), income level, education, hypertension, diabetes mellitus, dyslipidemia, and chronic kidney disease.

We performed a subgroup analysis stratified by the following covariates: age (<50 or ≥50 years), sex, BMI (<23, 23–24.9, or ≥25 kg/m2), presence or absence of metabolic syndrome, presence or absence of insulin resistance (homeostatic model assessment for insulin resistance >2.5 or ≤2.5), presence of hypertension, and presence of diabetes mellitus. The significance of ePWV in relation to NAFLD incidence was more pronounced in older, female, and overweight groups; it was also more prominent in participants without metabolic syndrome, insulin resistance, hypertension, or diabetes mellitus. Significant interactions were observed in subgroups of age, sex, BMI, presence of metabolic syndrome, and presence of insulin resistance (shown in Fig. 4).

Fig. 4.

Subgroup analysis of the risk of incident NAFLD according to ePWV tertiles. NAFLD, nonalcoholic fatty liver disease; ePWV, estimated pulse wave velocity; HR, hazard ratio; CI, confidence interval; BMI, body mass index; HOMA-IR, homeostatic model assessment for insulin resistance.

Fig. 4.

Subgroup analysis of the risk of incident NAFLD according to ePWV tertiles. NAFLD, nonalcoholic fatty liver disease; ePWV, estimated pulse wave velocity; HR, hazard ratio; CI, confidence interval; BMI, body mass index; HOMA-IR, homeostatic model assessment for insulin resistance.

Close modal

This large-scale prospective cohort study investigated the association between ePWV and NAFLD in Korean adults. The key findings of this study are as follows: (1) higher ePWV correlated with advanced age and a higher prevalence of medical conditions such as hypertension, diabetes mellitus, and CVD. (2) ePWV exhibited predictive capability for NAFLD prevalence. (3) Higher ePWV was also linked to an increased risk of future development of NAFLD. (4) The association of ePWV with the risk of NAFLD was more prominent in older participants and those without cardiometabolic risk factors such as metabolic syndrome and insulin resistance.

Arterial stiffening is a sensitive indicator of vascular pathologies, including atherosclerosis, vascular calcification, inflammation, and aging [25]. Arterial wall stiffening is caused by multiple mechanisms, including alterations in the elastin-to-collagen ratio and phenotypic changes in vascular smooth muscle cells driven by mechanotransduction, oxidative stress, genetic factors, or metabolic factors [26]. In terms of metabolic factors, arterial stiffness is strongly influenced by lipid and glucose metabolism as well as insulin resistance. Lipid markers and lipid ratios are closely associated with arterial stiffness due to factors such as cholesterol accumulation, oxidative stress from excess lipids, and arterial wall inflammation [27]. Furthermore, increasing evidence indicates that NAFLD is a systemic metabolic disorder affecting various extrahepatic systems, including the heart and vascular system [7]. Several studies have established an association between NAFLD and subclinical markers of CVD such as carotid intima-media thickness, carotid plaque formation, coronary artery calcification, and arterial stiffness [28, 29]. A previous meta-analysis demonstrated a significant association between NAFLD and increased arterial stiffness evaluated using cfPWV and baPWV [11]. The association between NAFLD and arterial stiffness may be bidirectional [30]. In this study, using a large sample cohort, we identified ePWV as an independent factor associated with NAFLD. In addition, ePWV was found to be an independent predictor of NAFLD incidence. Unlike measured PWV, ePWV is calculated using only BP and age. Therefore, it is a valuable tool for screening large-scale populations and a practical tool for clinical use.

The possible mechanisms underlying the relationship between NAFLD and increased arterial stiffness remain largely unknown; however, several plausible explanations can be proposed. First, the pivotal role of insulin resistance in the development of NAFLD needs to be acknowledged. Insulin resistance is a potential mechanism associated with increased arterial stiffness. Chronic hyperglycemia and hyperinsulinemia have been demonstrated to increase the activity of the renin-angiotensin-aldosterone system, thereby contributing to vascular wall hypertrophy and fibrosis [31, 32]. Second, oxidative stress and inflammation, which are known to have crucial roles in hepatic injury and NAFLD progression, seem to also contribute to the development of arterial stiffness. Oxidative injury can trigger increased vascular inflammation and cellular proliferation, potentially impairing arterial elasticity [33, 34]. Lastly, the levels of adipose cytokines, particularly adiponectin, may explain the relationship between NAFLD and arterial stiffness [35]. Additional longitudinal studies are required to define the precise mechanisms that link NAFLD to arterial stiffness.

An interesting finding of this study is that the association between ePWV and the incidence of NAFLD was more pronounced in older participants and those without cardiometabolic risk. This suggests that ePWV may be a useful marker for predicting the incidence of NAFLD in lower risk populations. However, in some subgroups, the sample size was insufficient to detect statistical significance; therefore, this result should be interpreted with caution.

This study is clinically relevant because it highlights the importance of assessing arterial stiffness as a predictor of NAFLD, a prevalent liver disease with implications for CVD risk. In addition to its clinical relevance as an indicator of hypertension-mediated organ damage in patients with hypertension [12, 13], it may also be valuable as a marker for predicting CVD associated with metabolic conditions such as NAFLD in the general population. Furthermore, ePWV can be obtained using simple indicators (age and BP), making it a useful marker that can be easily incorporated into routine clinical practice.

The strengths of the current study include its large sample size and long-term follow-up (18 years), which allowed for a comprehensive assessment of the research questions. Additionally, the inclusion of a community-based general population, commonly encountered in clinical practice, further enhanced the value of this study, as it represented a broad demographic rather than a specific population group. Furthermore, the quality of the study was enhanced by meticulous adherence to a standardized protocol that included face-to-face interviews during each examination.

This study had several limitations. First, this was an observational study, which implies that any observed association between ePWV and NAFLD incidence cannot be definitively interpreted as a causal relationship. Despite comprehensive adjustments for various confounding factors, the possibility of unmeasured confounders remains. Second, the study population was limited to residents of two cities in Korea, which could have introduced a selection bias. Third, data on income level, education, physical activity, and medical history were collected through questionnaires, potentially introducing a recall bias. Finally, during the 18-year follow-up period, 38.6% of the participants were lost to follow-up.

In conclusion, this study identified a significant association between an elevated ePWV and the prevalence and incidence of NAFLD in a large-scale prospective cohort in Korea. Individuals with higher ePWV have an increased risk of developing NAFLD in the future. These findings underscore the importance of evaluating arterial stiffness as a predictor of NAFLD, which has implications for both liver and cardiovascular outcomes.

All participants in this study were enrolled voluntarily, and written informed consent was obtained from each participant at every visit. The study protocol followed the principles outlined in the Declaration of Helsinki and was approved by the Korean National Research Institute of Health and the Institutional Review Board of Hanyang University Guri Hospital (IRB no. GURI-2023-06-020).

The authors declare no conflicts of interest.

The study was not supported by any sponsor or funder.

Byung Sik Kim: data curation, methodology, validation, and writing – original draft; Hyun-Jin Kim: formal analysis, investigation, and writing – original draft; Jeong-Hun Shin: conceptualization and writing – review and editing.

Additional Information

Byung Sik Kim and Hyun-Jin Kim contributed equally to this work.

Korean Genome and Epidemiology Study Ansung-Ansan cohort data are available upon reasonable request through the Korea Center for Disease Control and Prevention website (https://www.nih.go.kr/ko/main/contents.do?menuNo=300566).

1.
Younossi
ZM
,
Golabi
P
,
Paik
JM
,
Henry
A
,
Van Dongen
C
,
Henry
L
.
The global epidemiology of nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH): a systematic review
.
Hepatology
.
2023
;
77
(
4
):
1335
47
.
2.
Riazi
K
,
Azhari
H
,
Charette
JH
,
Underwood
FE
,
King
JA
,
Afshar
EE
et al
.
The prevalence and incidence of NAFLD worldwide: a systematic review and meta-analysis
.
Lancet Gastroenterol Hepatol
.
2022
;
7
(
9
):
851
61
.
3.
Francque
SM
,
Marchesini
G
,
Kautz
A
,
Walmsley
M
,
Dorner
R
,
Lazarus
JV
et al
.
Non-alcoholic fatty liver disease: a patient guideline
.
JHEP Rep
.
2021
;
3
(
5
):
100322
.
4.
Cotter
TG
,
Charlton
M
.
Nonalcoholic steatohepatitis after liver transplantation
.
Liver Transpl
.
2020
;
26
(
1
):
141
59
.
5.
Loomba
R
,
Friedman
SL
,
Shulman
GI
.
Mechanisms and disease consequences of nonalcoholic fatty liver disease
.
Cell
.
2021
;
184
(
10
):
2537
64
.
6.
Kasper
P
,
Martin
A
,
Lang
S
,
Kütting
F
,
Goeser
T
,
Demir
M
et al
.
NAFLD and cardiovascular diseases: a clinical review
.
Clin Res Cardiol
.
2021
;
110
(
7
):
921
37
.
7.
Duell
PB
,
Welty
FK
,
Miller
M
,
Chait
A
,
Hammond
G
,
Ahmad
Z
et al
.
Nonalcoholic fatty liver disease and cardiovascular risk: a scientific statement from the American Heart Association
.
Arterioscler Thromb Vasc Biol
.
2022
;
42
(
6
):
e168
85
.
8.
Vlachopoulos
C
,
Aznaouridis
K
,
Stefanadis
C
.
Prediction of cardiovascular events and all-cause mortality with arterial stiffness: a systematic review and meta-analysis
.
J Am Coll Cardiol
.
2010
;
55
(
13
):
1318
27
.
9.
Kim
BS
,
Lee
Y
,
Park
JK
,
Lim
YH
,
Shin
JH
.
Association of the estimated pulse wave velocity with cardio-vascular disease outcomes among men and women aged 40–69 years in the Korean population: an 18-year follow-up report on the Ansung-Ansan cohort in the Korean genome environment study
.
J Pers Med
.
2022
;
12
(
10
):
1611
.
10.
Park
JB
,
Avolio
A
.
Arteriosclerosis and atherosclerosis assessment in clinical practice: methods and significance
.
Pulse
.
2023
;
11
(
1
):
1
8
.
11.
Jaruvongvanich
V
,
Chenbhanich
J
,
Sanguankeo
A
,
Rattanawong
P
,
Wijarnpreecha
K
,
Upala
S
.
Increased arterial stiffness in nonalcoholic fatty liver disease: a systematic review and meta-analysis
.
Eur J Gastroenterol Hepatol
.
2017
;
29
(
9
):
e28
35
.
12.
Kim
HL
,
Lee
EM
,
Ahn
SY
,
Kim
KI
,
Kim
HC
,
Kim
JH
et al
.
The 2022 focused update of the 2018 Korean hypertension Society Guidelines for the management of hypertension
.
Clin Hypertens
.
2023
;
29
(
1
):
11
.
13.
Brunström
M
,
Burnier
M
,
Grassi
G
,
Januszewicz
A
et al
Mancia chairperson G, kreutz co-chair R. J hypertens. 2023 ESH guidelines for the management of arterial hypertension the task force for the management of arterial hypertension of the European society of hypertension endorsed by the European renal association (ERA) and the international society of hypertension
.
ISH
.
2023
.
14.
Park
JB
,
Sharman
JE
,
Li
Y
,
Munakata
M
,
Shirai
K
,
Chen
CH
et al
.
Expert consensus on the clinical use of pulse wave velocity in Asia
.
Pulse
.
2022
10
1–4
1
18
.
15.
Vlachopoulos
C
,
Terentes-Printzios
D
,
Laurent
S
,
Nilsson
PM
,
Protogerou
AD
,
Aznaouridis
K
et al
.
Association of estimated pulse wave velocity with survival: a secondary analysis of Sprint
.
JAMA Netw Open
.
2019
;
2
(
10
):
e1912831
.
16.
Greve
SV
,
Blicher
MK
,
Kruger
R
,
Sehestedt
T
,
Gram-Kampmann
E
,
Rasmussen
S
et al
.
Estimated carotid-femoral pulse wave velocity has similar predictive value as measured carotid-femoral pulse wave velocity
.
J Hypertens
.
2016
;
34
(
7
):
1279
89
.
17.
Kim
Y
,
Han
BG
KoGES group
.
Cohort profile: the Korean genome and epidemiology study (KoGES) consortium
.
Int J Epidemiol
.
2017
;
46
(
4
):
1350
.
18.
Hur
KY
,
Moon
MK
,
Park
JS
,
Kim
SK
,
Lee
SH
,
Yun
JS
et al
.
2021 clinical practice guidelines for diabetes mellitus of the Korean diabetes association
.
Diabetes Metab J
.
2021
;
45
(
4
):
461
81
.
19.
Jin
ES
,
Shim
JS
,
Kim
SE
,
Bae
JH
,
Kang
S
,
Won
JC
et al
.
Dyslipidemia fact sheet in South Korea, 2022
.
Diabetes Metab J
.
2023
;
12
(
3
):
237
51
.
20.
Friedewald
WT
,
Levy
RI
,
Fredrickson
DS
.
Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge
.
Clin Chem
.
1972
;
18
(
6
):
499
502
.
21.
Levey
AS
,
Coresh
J
,
Greene
T
,
Stevens
LA
,
Zhang
YL
,
Hendriksen
S
et al
.
Using standardized serum creatinine values in the modification of diet in renal disease study equation for estimating glomerular filtration rate
.
Ann Intern Med
.
2006
;
145
(
4
):
247
54
.
22.
Kotronen
A
,
Peltonen
M
,
Hakkarainen
A
,
Sevastianova
K
,
Bergholm
R
,
Johansson
LM
et al
.
Prediction of non-alcoholic fatty liver disease and liver fat using metabolic and genetic factors
.
Gastroenterology
.
2009
;
137
(
3
):
865
72
.
23.
Alberti
KG
,
Eckel
RH
,
Grundy
SM
,
Zimmet
PZ
,
Cleeman
JI
,
Donato
KA
et al
.
Harmonizing the metabolic syndrome: a joint interim statement of the international diabetes federation task force on epidemiology and prevention; national heart, lung, and blood Institute; American heart association; world heart federation; international atherosclerosis society; and international association for the study of obesity
.
Circulation
.
2009
;
120
(
16
):
1640
5
.
24.
Lee
SY
,
Park
HS
,
Kim
DJ
,
Han
JH
,
Kim
SM
,
Cho
GJ
et al
.
Appropriate waist circumference cutoff points for central obesity in Korean adults
.
Diabetes Res Clin Pract
.
2007
;
75
(
1
):
72
80
.
25.
Lu
Y
,
Kiechl
SJ
,
Wang
J
,
Xu
Q
,
Kiechl
S
,
Pechlaner
R
et al
.
Global distributions of age- and sex-related arterial stiffness: systematic review and meta-analysis of 167 studies with 509,743 participants
.
EBioMedicine
.
2023
;
92
:
104619
.
26.
Lacolley
P
,
Regnault
V
,
Laurent
S
.
Mechanisms of arterial stiffening: from mechanotransduction to epigenetics
.
Arterioscler Thromb Vasc Biol
.
2020
;
40
(
5
):
1055
62
.
27.
Baba
M
,
Maris
M
,
Jianu
D
,
Luca
CT
,
Stoian
D
,
Mozos
I
.
The impact of the blood lipids levels on arterial stiffness
.
J Cardiovasc Dev Dis
.
2023
;
10
(
3
):
127
.
28.
Ampuero
J
,
Gallego-Durán
R
,
Romero-Gómez
M
.
Association of NAFLD with subclinical atherosclerosis and coronary-artery disease: meta-analysis
.
Rev Esp Enferm Dig
.
2015
;
107
(
1
):
10
6
.
29.
Fan
Y
,
Wei
F
,
Zhou
Y
,
Zhang
H
.
Association of non-alcoholic fatty liver disease with impaired endothelial function by flow-mediated dilation: a meta-analysis
.
Hepatol Res
.
2016
46
3
E165
73
.
30.
Villela-Nogueira
CA
,
Leite
NC
,
Cardoso
CR
,
Salles
GF
.
NAFLD and increased aortic stiffness: parallel or common physiopathological mechanisms
.
Int J Mol Sci
.
2016
;
17
(
4
):
460
.
31.
Fang
FS
,
Liu
MY
,
Cheng
XL
,
Zhong
WW
,
Miao
XY
,
Li
J
et al
.
Insulin resistance correlates with the arterial stiffness before glucose intolerance
.
Intern Med
.
2014
;
53
(
3
):
189
94
.
32.
Nickenig
G
,
Röling
J
,
Strehlow
K
,
Schnabel
P
,
Böhm
M
.
Insulin induces upregulation of vascular AT1 receptor gene expression by posttranscriptional mechanisms
.
Circulation
.
1998
;
98
(
22
):
2453
60
.
33.
Delli Bovi
AP
,
Marciano
F
,
Mandato
C
,
Siano
MA
,
Savoia
M
,
Vajro
P
.
Oxidative stress in non-alcoholic fatty liver disease. An updated mini review
.
Front Med
.
2021
;
8
:
595371
.
34.
Polimeni
L
,
Del Ben
M
,
Baratta
F
,
Perri
L
,
Albanese
F
,
Pastori
D
et al
.
Oxidative stress: new insights on the association of non-alcoholic fatty liver disease and atherosclerosis
.
World J Hepatol
.
2015
;
7
(
10
):
1325
36
.
35.
Omelchenko
E
,
Gavish
D
,
Shargorodsky
M
.
Adiponectin is better predictor of subclinical atherosclerosis than liver function tests in patients with nonalcoholic fatty liver disease
.
J Am Soc Hypertens
.
2014
;
8
(
6
):
376
80
.