Association Between Fatty Liver Index and Brachial-Ankle Pulse Wave Velocity in the General Population　― A Cross-Sectional Study ―

Abstract

Background: Although metabolic dysfunction-associated fatty liver disease is typically diagnosed using ultrasonography, the fatty liver index (FLI) offers a simple alternative. Brachial-ankle pulse wave velocity (baPWV) is an established marker of arterial stiffness and a predictor of cardiovascular events. This study investigated the association between FLI and baPWV in a general Japanese population.

Methods and Results: This cross-sectional study included participants aged ≥18 years, excluding those with atrial fibrillation, lower extremity artery disease, severe aortic stenosis, or missing data. The primary outcome was an elevated baPWV (≥1,800 cm/s). Multivariable logistic regression analysis was performed to assess the association between FLI and elevated baPWV, considering FLI both as a categorical variable (low risk: FLI <30; moderate risk: FLI ≥30 and <60; high risk: FLI ≥60) and as a continuous variable (per 10-unit increase). The analysis included 10,122 individuals (mean age 54.3 years; 55% male). In multivariable-adjusted models, the odds of elevated baPWV were significantly higher in the moderate-risk (odds ratio [OR] 1.47; 95% confidence interval [CI] 1.20–1.79) and high-risk (OR 1.78; 95% CI 1.33–2.38) groups, using the low-risk group as the reference category. Each 10-unit increase in FLI was associated with significantly higher odds of the outcome (OR 1.16; 95% CI 1.10–1.22).

Conclusions: FLI showed a significant association with arterial stiffness in the general Japanese population.

Central Figure

Fatty liver disease is one of the most prevalent chronic liver conditions globally, with its incidence steadily increasing.¹ Non-alcoholic fatty liver disease (NAFLD) has been shown to be associated with an increased risk of atherosclerosis and cardiovascular events through complex pathophysiological mechanisms.² In 2020, a new framework called ‘metabolic dysfunction-associated fatty liver disease’ (MAFLD) was proposed as a replacement for NAFLD.³ Unlike NAFLD, MAFLD adopts diagnostic criteria based on metabolic dysfunction, allowing coexistence with other liver conditions. Moreover, MAFLD, or the newer, more comprehensive term, metabolic dysfunction-associated steatotic liver disease (MASLD) may have similar cardiovascular risks to NAFLD,^4,5 and MAFLD has been identified as an independent risk factor for atrial fibrillation.⁶ Therefore, accurate diagnosis and risk stratification of MAFLD remain critical issues for improving the management of cardiovascular diseases associated with fatty liver.

Although liver ultrasonography is commonly used for diagnosis in fatty liver disease, the fatty liver index (FLI), a score derived from body mass index (BMI), triglyceride (TG), gamma-glutamyl transpeptidase (GGT), and waist circumference (WC), provides a simple, non-invasive, and cost-effective surrogate. An FLI <30 suggests the absence of fatty liver, whereas an FLI ≥60 indicates a high likelihood of its presence.⁷ Additionally, Asian populations require lower cut-off values compared with their European counterparts, with notable sex-specific thresholds: FLI <25 and ≥35 for men, and <10 and ≥20 for women to exclude or confirm fatty liver, respectively.⁸ The FLI demonstrated an outstanding discriminative ability in identifying MAFLD, with a c-statistic exceeding 0.90 in a large Japanese cohort.⁹

Brachial-ankle pulse wave velocity (baPWV) is a well-established, non-invasive marker of arterial stiffness and has been shown to predict cardiovascular events and mortality in large-scale cohort studies.^10–12 A cut-off value of ≥1,800 cm/s has been widely used to identify individuals at high cardiovascular risk. In a general Japanese population, individuals with baPWV ≥1,800 cm/s had a significantly increased risk of cardiovascular events compared with those with baPWV <1,800 cm/s.¹³

Several studies have reported significant associations between MAFLD or its surrogate marker, FLI, and baPWV, suggesting a link between hepatic steatosis and arterial stiffness.^14–18 Few studies have investigated these relationships in detail in the Japanese population.¹⁹ Moreover, despite accumulating evidence of the association between FLI and arterial stiffness, few studies have comprehensively investigated these relationships while accounting for sex differences or stratification by BMI or age. Therefore, we aimed to investigate the association between the FLI and baPWV in the general Japanese population with a particular focus on potential effect modifications by sex, age, and BMI. By providing insight into the interplay between MAFLD and arterial stiffness across different demographics and subgroups, this study may help improve cardiovascular risk stratification and help guide targeted preventive strategies.

Methods

Study Population

The JA Kagoshima Kouseiren Hospital Health Care Center is a prominent health check-up institution situated in Kagoshima Prefecture, Japan, spanning a demographic area of 9,187 km² and serving a population of 1.6 million people.²⁰ In this cross-sectional study, we used a database of individuals aged ≥18 years who underwent health check-ups, including baPWV measurement, between 2007 and 2019. We excluded participants with atrial fibrillation, lower-extremity artery disease (defined as ankle-brachial index [ABI] <0.9), severe aortic stenosis, or missing data.

This study adhered to the principles outlined in the Declaration of Helsinki. This study was approved by the Institutional Ethics Committee of the Graduate School of Medical and Dental Sciences of Kagoshima University and JA Kagoshima Kouseiren Hospital (no. 170130 [520]). The Ethics Committee agreed that because only existing anonymized data were used in this study, the requirement to obtain informed consent from each individual was waived.

Data Collection

The collected data included age, BMI, blood pressure, GGT, fasting blood glucose, uric acid, total cholesterol, TG, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, estimated glomerular filtration rate (eGFR), smoking status (current smoker), alcohol consumption, regular exercise (≥30 min/day), use of antihypertensive or lipid-lowering medications, history of diabetes, FLI, ABI, and baPWV.

Blood pressure was measured using a mercury sphygmomanometer after the participants were seated for 5 min. The baPWV and ABI were measured using an automated waveform analyzer (BP-203RPE III, Omron, Kyoto, Japan). The measurement procedure was as follows. The participants rested in the supine position for 5 min in a quiet, temperature-controlled room. Electrocardiogram electrodes and a phonocardiographic microphone were placed on both wrists and on the left sternal border. Subsequently, blood pressure cuffs equipped with flow and pressure sensors were attached to the arms and ankles. The path lengths from the suprasternal notch to the brachium (Lb) and to the ankle (La) were calculated, and the time interval between the upstrokes of the pulse waveforms (∆Tba) was automatically recorded. Finally, baPWV was calculated using the following formula:²¹ baPWV = (La − Lb) / ∆Tba.

Blood samples were collected after overnight fasting. Biochemical parameters, including TG, GGT, uric acid, fasting blood glucose, total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, and creatinine levels, were measured using the standard laboratory procedures. Medical history, lifestyle factors (i.e., current smoking status, alcohol consumption, and regular exercise), and medication use were obtained through self-administered questionnaires. A history of diabetes was defined as the presence of any of the following: current use of glucose-lowering medications, fasting blood glucose level of ≥110 mg/dL, or HbA1c level above the diagnostic threshold (≥6.1% according to the JDS value or ≥6.5% according to the NGSP value).

The eGFR was calculated using a formula validated for the Japanese population:²² eGFR (mL/min/1.73 m²) = 0.881 × 186 × (serum creatinine [mg/dL])^−1.154 × (age)^−0.203 × (0.742 [if female]).

The FLI was calculated using the following formula:⁷ FLI = (e^{0.953 * log(TG) + 0.139 * BMI + 0.718 * log(GGT) + 0.053 * WC − 15.745}) * 100 / (1 + e^{0.953 * log(TG) + 0.139 * BMI + 0.718 * log(GGT) + 0.053 * WC − 15.745}).

The primary outcome was defined as a binary variable indicating whether the average left and right baPWV was ≥1,800 cm/s.

Statistical Analysis

We conducted a complete case analysis by excluding participants with missing data for any of the variables included in the study. Continuous variables were expressed as mean±standard deviation (SD), and categorical variables were expressed as numbers and percentages. Univariate logistic regression was used to assess the association between each variable and baPWV ≥1,800 cm/s.

Participants were categorized into 3 FLI groups: (1) low-risk group (FLI <30); (2) moderate-risk group (FLI ≥30 and <60); and (3) high-risk group (FLI ≥60). Continuous variables were summarized as mean±SD, and categorical variables as frequency and percentage. Differences between groups were assessed using 1-way analysis of variance for continuous variables and the chi-square test for categorical variables, followed by Tukey’s post-hoc test. To investigate the association between FLI classification and elevated baPWV (≥1,800 cm/s), logistic regression analyses were performed using the low-risk group as the reference category. The association per 10-unit increase in the FLI was also evaluated. Following the STROBE guidelines, we reported the results from a crude model (unadjusted), minimally adjusted model (adjusted for age and sex), and fully adjusted model (adjusted for age, sex, BMI, systolic and diastolic blood pressure [SBP and DBP], total cholesterol, uric acid, eGFR, history of diabetes, current smoking, alcohol consumption, and regular exercise (≥30 min/day]).

Subgroup analyses were conducted to assess whether the association between the FLI and baPWV ≥1,800 cm/s differed across subgroups defined by age (<55 or ≥55 years), sex, BMI (<23 or ≥23 kg/m²), history of diabetes, current smoking, alcohol consumption, regular exercise, SBP (≥140 mmHg), and DBP (≥90 mmHg). Stratified multivariable logistic regression using the fully adjusted model was applied to each subgroup, and the interaction effects were evaluated using the likelihood ratio test.

To assess whether the association between the FLI and baPWV remained robust under different conditions, several sensitivity analyses were conducted. The analytical models used in each study are described below.

First, multiple linear regression analysis was conducted with baPWV as the dependent variable and FLI as the primary independent variable, using the same covariates as in the fully adjusted model. Second, additional analyses were conducted by including fasting blood glucose levels, ABI, use of antihypertensive medications, and use of lipid-lowering medications as covariates in the fully adjusted model. Third, the FLI-based risk stratification thresholds were modified as follows: for men – low-risk group (FLI <25), moderate-risk group (FLI ≥25 and <35), and high-risk group (FLI ≥35); and for women – low-risk group (FLI <10), moderate-risk group (FLI ≥10 and <20), and high-risk group (FLI ≥20). Based on this stratification, sex-specific logistic regression analyses were conducted using a fully adjusted model to assess the risk of baPWV ≥1,800 cm/s, with the low-risk group as the reference. Last, the baPWV threshold used to define the dependent variable was altered to ≥1,600 cm/s and ≥1,400 cm/s. The association with a 10-unit increase in FLI was reassessed using logistic regression with a fully adjusted model.

Odds ratios (ORs) with 95% confidence intervals (CIs) were reported for logistic regression, and regression coefficients (β) with 95% CIs for linear regression analysis. All statistical analyses were performed using JMP Pro v.17 for Windows (SAS Institute, Cary, NC, USA), except for likelihood ratio tests for interactions, which were conducted in R v.4.4.2 (R Core Team, Auckland, New Zealand). Statistical significance was set at a 2-sided P value <0.05.

Results

Selection of Study Participants

Among the 10,277 individuals initially screened, 155 were excluded based on the following criteria: atrial fibrillation (n=96); lower extremity artery disease based on an ABI of <0.9 (n=36); and missing data (n=25). The final analysis included 10,122 participants, including those who were receiving antihypertensive agents, statins, or antidiabetic medications. No severe aortic stenosis was observed in any participant (Figure 1).

Figure 1.

Flow chart of the participant selection process. Among the 10,277 individuals initially screened, 155 were excluded based on the following criteria: atrial fibrillation (n=96), lower extremity artery disease based on an ankle-brachial index of <0.9 (n=36), and missing data (n=25). The final analysis included 10,122 participants, including those who were receiving antihypertensive agents, statins, or antidiabetic medications. No severe aortic stenosis was observed in any participant. baPWV, brachial-ankle pulse wave velocity.

Demographic and Clinical Characteristics of the Participants

The baseline characteristics and results of the univariate analysis are presented in Table 1. The mean age of participants was 54.3±13.1 years, and 54.9% were male. The mean baPWV was 1484±334 cm/s, with 15.3% of the participants classified as having baPWV ≥1,800 cm/s. In the univariate analyses, older age, SBP, DBP, regular exercise, use of antihypertensive or lipid-lowering medications, history of diabetes, GGT, uric acid, total cholesterol, FLI, and ABI were significantly associated with elevated baPWV (a significant positive correlation was observed between FLI and PWV [Pearson’s r=0.134; 95% CI 0.115–0.153; P<0.001; Spearman’s ρ=0.239; P<0.001]). In contrast, BMI and low-density lipoprotein cholesterol levels were not significantly associated with baPWV, whereas high-density lipoprotein cholesterol levels, eGFR, current smoking, and alcohol consumption showed inverse associations.

Table 1.

Demographic, Clinical Characteristics, and Results of Univariate Analyses for Elevated baPWV (≥1,800 cm/s)

	Mean±SD	OR (95% CI)	P value
Sex
Male	5,557 (54.9)	1.317 (1.179–1.470)	<0.001
Female	4,565 (45.1)	Ref.
Age (years)	54.3±13.1	1.156 (1.147–1.165)	<0.001
BMI (kg/m2)	23.4±3.5	1.007 (0.992–1.022)	0.380
Systolic blood pressure (mmHg)	121.1±17.3	1.084 (1.080–1.089)	<0.001
Diastolic blood pressure (mmHg)	74.8±11.4	1.054 (1.049–1.060)	<0.001
Current smoker	2,102 (20.8)	0.422 (0.357–0.499)	<0.001
Regular exercise	2,642 (26.1)	1.933 (1.725–2.165)	<0.001
Alcohol consumption	6,901 (68.2)	0.765 (0.683–0.856)	<0.001
Antihypertensive medication	1,978 (19.5)	4.83 (4.30–5.43)	<0.001
Lipid-lowering medication	762 (7.5)	2.80 (2.38–3.30)	<0.001
History of diabetes	1,102 (10.9)	3.701 (3.224–4.248)	<0.001
GGT (U/L)	40.2±49.8	1.002 (1.001–1.003)	<0.001
Fasting blood glucose (mg/dL)	104.1±21.5	1.018 (1.016–1.020)	<0.001
Uric acid (U/L)	5.4±1.4	1.140 (1.099–1.183)	<0.001
Total cholesterol (mg/dL)	206.3±35.2	1.002 (1.000–1.003)	<0.001
Triglycerides (mg/dL)	117.3±102.5	1.001 (1.000–1.001)	<0.001
HDL cholesterol (mg/dL)	59.7±15.1	0.995 (0.991–0.999)	0.007
LDL cholesterol (mg/dL)	121.1±31.1	1.000 (0.998–1.002)	0.805
eGFR (mL/min/1.73 m2)	77.7±15.0	0.957 (0.953–0.961)	<0.001
FLI	28.8±25.3	1.005 (1.003–1.007)	<0.001
ABI	1.14±0.07	15.967 (6.928–36.800)	<0.001
baPWV (cm/s)	1,484±334
baPWV ≥1,800 cm/s	1,553 (15.3)

Continuous variables are expressed as mean±standard deviation, and categorical variables as n (%). Univariate logistic regression was used to assess the association between each variable and baPWV ≥1,800 cm/s. ABI, ankle-brachial index; baPWV, brachial-ankle pulse wave velocity; BMI, body mass index; eGFR, estimated glomerular filtration rate; FLI, fatty liver index; GGT, gamma-glutamyl transpeptidase; HDL, high-density lipoprotein; LDL, low-density lipoprotein; OR, odds ratio; Ref., reference group.

Participants were stratified into 3 groups based on FLI thresholds: low-risk group (FLI <30), moderate-risk group (FLI ≥30 and <60), and high-risk group (FLI ≥60). The baseline characteristics of the FLI group are shown in Table 2. As the FLI increased, the proportion of males, BMI, blood pressure (both SBP and DBP), prevalence of current smoking habits and alcohol consumption, history of diabetes, GGT levels, fasting blood glucose, uric acid, total cholesterol, and TG levels increased. In contrast, high-density lipoprotein cholesterol levels and the proportion of individuals who engaged in regular exercise were higher in those with lower FLI values. Age was highest in the moderate-risk group, followed by the low- and high-risk groups. Similarly, baPWV was the highest in the moderate-risk group, followed by the high-risk and low-risk groups. Low-density lipoprotein cholesterol and ABI were the lowest in the low-risk group; however, no significant differences were observed between the moderate- and high-risk groups. The eGFR was lowest in the moderate-risk group, whereas no significant differences were noted between the low- and high-risk groups. The peak baPWV observed in the moderate-risk group was largely attributable to the influence of age (Supplementary Tables 1,2).

Table 2.

Demographic and Clinical Characteristics Based on FLI Classification

	Low-risk group (FLI <30) [a]	Moderate-risk group (FLI ≥30 and <60) [b]	High-risk group (60≤ FLI) [c]	ANOVA or Chi-square test	P [a-b]	P [b-c]	P [c-a]
No. participants	6,309	2,320	1,493
Sex
Male	2,708 (42.9)	1,646 (70.9)	1,203 (80.6)	<0.001	<0.001	<0.001	<0.001
Female	3,601 (57.1)	674 (29.1)	290 (19.4)
Age (years)	54.3±13.6	56.0±11.9	51.3±11.9	<0.001	<0.001	<0.001	<0.001
BMI (kg/m2)	21.6±2.3	25.1±2.3	28.1±3.6	<0.001	<0.001	<0.001	<0.001
Systolic blood pressure (mmHg)	118.0±17.2	125.8±16.1	127.1±15.9	<0.001	<0.001	0.012	<0.001
Diastolic blood pressure (mmHg)	72.3±10.9	77.9±10.6	80.7±11.1	<0.001	<0.001	<0.001	<0.001
Current smoker (%)	16.2	24.7	33.8	<0.001	<0.001	<0.001	<0.001
Regular exercise (%)	27.9	25.4	20.0	<0.001	0.019	<0.001	<0.001
Alcohol consumption (%)	62.6	74.9	81.4	<0.001	<0.001	<0.001	<0.001
History of diabetes (%)	6.9	14.5	22.0	<0.001	<0.001	<0.001	<0.001
GGT (U/L)	23.6±16.3	50.1±42.8	95.0±92.5	<0.001	<0.001	<0.001	<0.001
Fasting blood glucose (mg/dL)	100.5±18.1	107.9±22.7	113.4±28.1	<0.001	<0.001	<0.001	<0.001
Uric acid (U/L)	5.0±1.3	5.9±1.3	6.5±1.4	<0.001	<0.001	<0.001	<0.001
Total cholesterol (mg/dL)	202.1±33.7	210.6±35.7	217.3±37.3	<0.001	<0.001	<0.001	<0.001
Triglycerides (mg/dL)	81.2±36.7	141.0±67.6	233.2±196.6	<0.001	<0.001	<0.001	<0.001
HDL cholesterol (mg/dL)	64.1±14.8	54.1±12.9	50.2±12.4	<0.001	<0.001	<0.001	<0.001
LDL cholesterol (mg/dL)	117.2±29.8	127.8±31.6	127.0±33.4	<0.001	<0.001	0.428	<0.001
eGFR (mL/min/1.73 m2)	78.2±14.9	75.8±14.7	78.4±15.4	<0.001	<0.001	<0.001	0.635
Fatty liver index	12.0±8.0	43.5±8.6	76.8±10.8	<0.001	<0.001	<0.001	<0.001
ABI	1.13±0.07	1.15±0.06	1.15±0.06	<0.001	<0.001	0.570	<0.001
baPWV (cm/s)	1,451±336	1,552±333	1,516±309	<0.001	<0.001	0.001	<0.001
baPWV ≥1,800 cm/s	872 (13.8)	456 (20.1)	225 (15.1)	<0.001	<0.001	<0.001	0.214

Unless indicated otherwise, data are presented as mean±standard deviation, or n (%). Differences between groups were assessed using 1-way analysis of variance for continuous variables and the chi-square test for categorical variables, followed by Tukey’s post-hoc test. ANOVA, one-way analysis of variance. Other abbreviations as in Table 1.

Association Between FLI and Elevated baPWV (≥1,800 cm/s) Across Different Logistic Regression Models

Table 3 presents the association between FLI and baPWV ≥1,800 cm/s across all models. In the stratified analysis, using the low-risk group as the reference, the moderate-risk group showed a significant association with elevated baPWV in all models: the crude model OR 1.53 (95% CI 1.35–1.73; P<0.001), the minimally adjusted model OR 1.59 (95% CI 1.37–1.85; P<0.001), and the fully adjusted model OR 1.47 (95% CI 1.20–1.79; P<0.001). In the high-risk group, no significant association was observed in the crude model (OR 1.10; 95% CI 0.94–1.73; P=0.212), using the low-risk group as the reference. However, significant associations were found in the minimally adjusted model (OR 1.98; 95% CI 1.64–2.41; P<0.001) and the fully adjusted model (OR 1.78; 95% CI 1.33–2.38; P<0.001). In the analysis using FLI as a continuous variable, defined by a per 10-unit increase in FLI, the odds of baPWV ≥1,800 cm/s were as follows: crude model OR 1.05 (95% CI 1.03–1.07; P<0.001), minimally adjusted model OR 1.14 (95% CI 1.11–1.17; P<0.001), and fully adjusted model OR 1.16 (95% CI 1.10–1.22; P<0.001), indicating a consistent and significant association. Using the low-risk group as the reference, the lack of significance for the high-risk group in the crude model, but the presence of significance in the minimally adjusted model, was largely explained by the influence of age (Supplementary Tables 1,2).

Table 3.

Association Between FLI and Elevated baPWV (≥1,800 cm/s) Across Different Logistic Regression Models

	Crude model, OR (95% CI); P	Minimally adjusted model, OR (95% CI); P	Fully adjusted model, OR (95% CI); P
Low-risk group (FLI <30)	Ref.	Ref.	Ref.
Moderate-risk group (30≤FLI<60)	1.53 (1.35–1.73); <0.001	1.59 (1.37–1.85); <0.001	1.47 (1.20–1.79); <0.001
High-risk group (FLI ≥60)	1.10 (0.94–1.73); 0.212	1.98 (1.64–2.41); <0.001	1.78 (1.33–2.38); <0.001
FLI per 10 units	1.05 (1.03–1.07); <0.001	1.14 (1.11–1.17); <0.001	1.16 (1.10–1.22); <0.001

Crude model: no covariates were adjusted. Minimally adjusted model: adjusted for age and sex. Fully adjusted model: adjusted for age, sex, BMI, systolic blood pressure, diastolic blood pressure, total cholesterol, uric acid, eGFR, history of diabetes, current smoking, alcohol consumption, and regular exercise (≥30 min/day). CI, confidence interval. Other abbreviations as in Table 1.

Subgroup Analyses of the Association Between a 10-Unit Increase in FLI and Elevated baPWV (≥1,800 cm/s) in Fully Adjusted Logistic Regression Models

Subgroup analyses evaluating the association between a 10-unit increase in FLI and baPWV ≥1,800 cm/s are presented in Table 4. Significant interactions were observed for age (<55 vs. ≥55 years), sex, and BMI (<23 vs. ≥23 kg/m²). Among participants aged <55 years, the OR was 1.17 (95% CI 1.01–1.38), compared with 1.15 (95% CI 1.09–1.22) in those aged ≥55 years. In women, the OR was 1.24 (95% CI 1.13–1.37), whereas in men the OR was 1.13 (95% CI 1.07–1.21). For participants with a BMI <23 kg/m², the OR was 1.37 (95% CI 1.15–1.51), compared with 1.09 (95% CI 1.02–1.16) for those with a BMI ≥23 kg/m².

Table 4.

Subgroup Analyses of the Association Between a 10-Unit Increase in FLI and Elevated baPWV (≥1,800 cm/s) in Fully Adjusted Logistic Regression Models

	No. participants	OR (95% CI)	P for interaction
Age (years)			0.007
<55	4,758	1.17 (1.01–1.38)
≥55	5,364	1.15 (1.09–1.22)
Sex			0.048
Male	5,557	1.13 (1.07–1.21)
Female	4,565	1.24 (1.13–1.37)
BMI (kg/m2)			<0.001
<23	4,943	1.37 (1.15–1.51)
≥23	5,179	1.09 (1.02–1.16)
History of diabetes			0.474
Yes	1,102	1.11 (0.99–1.24)
No	9,020	1.18 (1.11–1.25)
Current smoking			0.907
Yes	2,102	1.21 (1.07–1.36)
No	8,020	1.15 (1.09–1.22)
Regular exercise			0.604
Yes	2,642	1.18 (1.08–1.30)
No	7,480	1.15 (1.08–1.22)
Alcohol consumption			0.268
Yes	6,901	1.15 (1.09–1.22)
No	3,221	1.20 (1.08–1.33)
Systolic blood pressure (mmHg)			0.643
<140	8,764	1.14 (1.07–1.21)
≥140	1,358	1.20 (1.09–1.31)
Diastolic blood pressure (mmHg)			0.121
<90	9,168	1.14 (1.07–1.21)
≥90	954	1.23 (1.10–1.37)

The models were fully adjusted as described in Table 3. P for the interaction was derived from the likelihood ratio tests. Abbreviations as in Tables 1,3.

The P-values for interaction were <0.05 for age, sex, and BMI. No significant interactions were observed for diabetes status, smoking status, alcohol consumption, regular exercise, SBP (≥140 vs. <140 mmHg), or DBP (≥90 vs. <90 mmHg).

Sensitivity Analyses

To assess the robustness of the association between the FLI and elevated baPWV under different assumptions, the following sensitivity analyses were conducted.

Multiple Linear Regression Multiple linear regression analysis with baPWV as a continuous dependent variable and FLI, age, sex, BMI, SBP, DBP, total cholesterol, uric acid, eGFR, diabetes, smoking status, alcohol consumption, and regular exercise as independent variables yielded a regression coefficient of β=1.39 (95% CI 1.08–1.69; P<0.001), indicating a significant positive association.

Fully Adjusted Model Including Additional Covariates When fasting blood glucose, ABI, antihypertensive medication, and lipid-lowering medication were added to the fully adjusted model, the OR for baPWV ≥1,800 cm/s per 10-unit increase in FLI was 1.14 (95% CI 1.08–1.20; P<0.001).

Sex-Specific Stratification Using Alternative FLI Cut-Offs for Elevated baPWV (≥1,800 cm/s) For men (n=5,557), using FLI cut-off values of 25 and 35, participants were categorized into low-risk (FLI <25), moderate-risk (FLI ≥25 and <35), and high-risk (FLI ≥35) groups. Using the low-risk group as the reference category, the moderate-risk group had OR 1.69 (95% CI 1.25–2.29; P<0.001), and the high-risk group had OR 1.71 (95% CI 1.30–2.25; P<0.001), indicating a consistent positive association across groups.

For women (n=4,565), using FLI cut-off values of 10 and 20, participants were classified as low-risk (FLI<10), moderate-risk (FLI ≥10 and <20), and high-risk (FLI ≥20). Using the low-risk group as the reference category, the moderate-risk group had OR 1.28 (95% CI 0.92–1.77; P= 0.143), whereas the high-risk group showed a significant association with OR 2.00 (95% CI 1.37–2.93; P<0.001).

Analyses Using Alternative baPWV Thresholds When the outcome was defined as baPWV ≥1,600 cm/s, the OR per 10-unit increase in FLI was 1.18 (95% CI 1.13–1.23; P<0.001). When the cut-off was set at baPWV ≥1,400 cm/s, the OR was 1.19 (95% CI 1.14–1.24; P<0.001). These results consistently confirmed the association between a higher FLI and an elevated baPWV.

Discussion

In the present study, we demonstrated the association between FLI, which reflects MAFLD, and baPWV, an indicator of arterial stiffness, using logistic regression analysis (Figure 2). This relationship remained robust across sensitivity analyses using linear regression, alternative baPWV cut-offs (1,600 and 1,400 cm/s), and additional covariate adjustments (e.g., medication use, fasting blood glucose, and ABI). Stratified analyses using different FLI categories further supported the consistency of these findings.

Figure 2.

Central illustration showing the association between the fatty liver index (FLI; reflecting MAFLD) and elevated baPWV based on logistic regression analysis. The central illustration depicts the association between the FLI and arterial stiffness, as assessed using brachial-ankle pulse wave velocity (baPWV), in a large Japanese population. In this cross-sectional study of 10,122 participants, a higher FLI was significantly associated with an elevated baPWV (≥1,800 cm/s). Compared with the low-risk group (FLI <30), the odds of elevated baPWV were increased in the moderate-risk (FLI ≥30 and <60; OR 1.47; 95% CI 1.20–1.79) and high-risk (FLI ≥60; OR 1.78; 95% CI 1.33–2.38) groups. Each 10-unit increase in FLI was associated with higher odds of the outcome (OR 1.16; 95% CI 1.10–1.22). Mechanistically, metabolic dysfunction-associated fatty liver disease (MAFLD) contributes to arterial stiffness through endothelial dysfunction, oxidative stress, systemic inflammation, and metabolic abnormalities such as insulin resistance and dyslipidemia. Collectively, these pathways contribute to a proatherogenic state that promotes both atherosclerosis and arterial stiffness progression. BMI, body mass index; GGT, gamma-glutamyl transpeptidase; TG, triglycerides; WC, waist circumference.

A PubMed search using ‘brachial-ankle pulse wave velocity’ and ‘fatty liver index’ identified 4 studies published as of February 2025, with 2 relevant to our investigation.^18,19 Unlike previous studies that used linear regression, our study used logistic regression with defined baPWV thresholds to provide novel methodological insights. The use of a 1,800 cm/s cut-off may improve the interpretability of cardiovascular risk associations.¹³ Furthermore, the inclusion of both men and women from the general Japanese adult population, along with stratified and subgroup analyses, represents a novel aspect of this study.

Several mechanisms can explain the association between MAFLD and increased arterial stiffness. Endothelial dysfunction is triggered by increased asymmetric dimethylarginine, which impairs nitric oxide production and vascular relaxation.^23–25 Elevated homocysteine levels, resulting from altered methionine metabolism, further exacerbate endothelial dysfunction via oxidative stress.²⁶ Systemic inflammation plays a central role, characterized by elevated levels of inflammatory cytokines, chemokines, and soluble adhesion molecules, coupled with an imbalance in macrophage polarization.^23,27 Additionally, hepatic insulin resistance and altered lipid metabolism lead to dyslipidemia, which promotes proatherogenic lipid profiles and accelerates atherosclerosis and cardiovascular risk.^28–30

Interaction analysis revealed important sex-based differences. A stronger association in women than in men, suggesting that hepatic steatosis may have a greater impact on arterial stiffness in women. Previous studies have demonstrated a stronger association between MAFLD and baPWV in women than in men.¹⁶ Estrogen decline during menopause promotes visceral fat accumulation and ectopic lipid deposition, thereby exacerbating insulin resistance and metabolic dysfunction.^31,32 Furthermore, reduced estrogen-dependent antioxidant capacity³³ and heightened systemic inflammation³⁴ may synergistically enhance the association between insulin resistance or obesity and arterial stiffness, particularly in women.³⁵

We identified a significant interaction effect stratifying individuals by age, using 55 years as a cut-off, which was selected because it was close to the median age of the study population. This study demonstrated that the association between the FLI and arterial stiffness was more pronounced in younger individuals than in their older counterparts. Associations between NAFLD and arterial stiffness have been previously reported, tending to be greater in individuals aged <58 years,³⁶ particularly in women. These findings are consistent with those of the present study.

A significant interaction was also found for BMI, with a cut-off of 23 kg/m², which was selected because it approximated the median BMI in our cohort. Lean MAFLD, despite a lower BMI, has been reported to represent a metabolically unhealthy phenotype that is strongly associated with subclinical atherosclerosis.¹⁷

This study highlights the utility of the FLI in assessing arterial stiffness, particularly in younger, female, and non-obese populations overlooked by conventional measures, and underscores its potential for cardiovascular risk stratification.

Study Limitations

This study has some limitations that warrant consideration when interpreting the findings. First, its cross-sectional design precluded any causal inference between FLI and arterial stiffness. Second, although the FLI was derived from ultrasonographic assessments, ultrasonography was unable to distinguish between steatohepatitis and fibrosis and lacked the diagnostic precision of liver biopsy or advanced imaging techniques, which limits the accuracy and generalizability of the FLI.⁷ Third, the study population was derived from health check-up participants at a single center, which may limit the generalizability of the findings due to potential selection bias.^37–39 Furthermore, unexpected inverse associations between smoking and baPWV suggest the potential influence of unmeasured confounding factors or biases.^40–42 Fourth, the FLI, originally developed in populations with NAFLD,⁷ was applied in the present study without excluding participants with alcohol intake, viral hepatitis, or other liver diseases, thereby extrapolating its use beyond the original target population.

Future longitudinal studies are necessary to clarify the temporal relationships between FLI trajectories, arterial stiffness progression, and cardiovascular outcomes.

Conclusions

This study identified a significant association between the FLI and elevated baPWV, highlighting the potential role of the FLI in cardiovascular risk stratification in the Japanese population.

Sources of Funding

The authors did not receive any funding for this study.

Disclosures

M.O. is a member of Circulation Reports’ Editorial Team.

IRB Information

This study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Ethics Committees of Kagoshima University and JA Kagoshima Kouseiren Hospital (approval no. 170130 [520]). Informed consent was waived because only pre-existing anonymized data were used.

Data Availability

The deidentified participant data will not be shared.

Supplementary Files

Please find supplementary file(s);

https://doi.org/10.1253/circrep.CR-25-0218

References

• 1.   Ito T, Ishigami M, Zou B, Tanaka T, Takahashi H, Kurosaki M, et al. The epidemiology of NAFLD and lean NAFLD in Japan: A meta-analysis with individual and forecasting analysis, 1995–2040. Hepatol Int 2021; 15: 366–379.
• 2.   Stahl EP, Dhindsa DS, Lee SK, Sandesara PB, Chalasani NP, Sperling LS. Nonalcoholic fatty liver disease and the heart: JACC State-of-the-Art Review. J Am Coll Cardiol 2019; 73: 948–963.
• 3.   Eslam M, Newsome PN, Sarin SK, Anstee QM, Targher G, Romero-Gomez M, et al. A new definition for metabolic dysfunction-associated fatty liver disease: An international expert consensus statement. J Hepatol 2020; 73: 202–209.
• 4.   Zhou XD, Targher G, Byrne CD, Somers V, Kim SU, Chahal CAA, et al. An international multidisciplinary consensus statement on MAFLD and the risk of CVD. Hepatol Int 2023; 17: 773–791.
• 5.   Ogawa T, Sato T, Tanaka M, Akiyama Y, Nakata K, Kouzu H, et al. Metabolic dysfunction-associated steatotic liver disease (MASLD) is an independent risk factor for the development of ischemic heart disease: A 10-year cohort study. Circ Rep 2025; 7: 350–358.
• 6.   Takami A, Kato M, Kotake Y, Okamura A, Tomomori T, Kawatani S, et al. Impact of metabolic dysfunction-associated fatty liver disease on atrial fibrillation recurrence after ablation: A retrospective study in Japanese patients. Circ J 2025, doi:10.1253/circj.CJ-25-0169.
• 7.   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, doi:10.1186/1471-230X-6-33.
• 8.   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: e0120443, doi:10.1371/journal.pone.0120443.
• 9.   Okada A, Yamada G, Kimura T, Hagiwara Y, Yamaguchi S, Kurakawa KI, et al. Diagnostic ability using fatty liver and metabolic markers for metabolic-associated fatty liver disease stratified by metabolic/glycemic abnormalities. J Diabetes Investig 2023; 14: 463–478.
• 10.   Shi Y, Wu LD, Feng XH, Kan JY, Kong CH, Ling ZY, et al. Estimated pulse wave velocity predicts all-cause and cardiovascular-cause mortality in individuals with hypertension: Findings from a National Health and Nutrition Examination Study 1999–2018. Circ J 2024; 88: 417–424.
• 11.   Turin TC, Kita Y, Rumana N, Takashima N, Kadota A, Matsui K, et al. Brachial-ankle pulse wave velocity predicts all-cause mortality in the general population: Findings from the Takashima study, Japan. Hypertens Res 2010; 33: 922–925.
• 12.   Ninomiya T, Kojima I, Doi Y, Fukuhara M, Hirakawa Y, Hata J, et al. Brachial-ankle pulse wave velocity predicts the development of cardiovascular disease in a general Japanese population: The Hisayama Study. J Hypertens 2013; 31: 477–483, doi:10.1097/HJH.0b013e32835c5c23.
• 13.   Takashima N, Turin TC, Matsui K, Rumana N, Nakamura Y, Kadota A, et al. The relationship of brachial-ankle pulse wave velocity to future cardiovascular disease events in the general Japanese population: The Takashima Study. J Hum Hypertens 2014; 28: 323–327.
• 14.   Liu L, Zhou Y, Deng S, Yuan T, Yang S, Zhu X, et al. Arterial stiffness progression in metabolic dysfunction-associated fatty liver disease subtypes: A prospective cohort study. Nutr Metab Cardiovasc Dis 2024; 34: 1890–1900.
• 15.   Wang J, Liu S, Cao Q, Wu S, Niu J, Zheng R, et al. New definition of metabolic dysfunction-associated fatty liver disease with elevated brachial-ankle pulse wave velocity and albuminuria: A prospective cohort study. Front Med 2022; 16: 714–722.
• 16.   Yoo TK, Lee SW, Lee MY, Choi H, Sung KC. Influence of MAFLD and NAFLD on arterial stiffness: A longitudinal cohort study. Nutr Metab Cardiovasc Dis 2024; 34: 1769–1778.
• 17.   Liu L, Wang C, Deng S, Yuan T, Zhu X, Deng Y, et al. Transition patterns of metabolic dysfunction-associated fatty liver disease status in relation to arterial stiffness progression: A health check-up cohort study. Sci Rep 2023; 13: 9690, doi:10.1038/s41598-023-35733-0.
• 18.   Cicero AFG, Gitto S, Fogacci F, Rosticci M, Giovannini M, D’Addato S, et al. Fatty liver index is associated to pulse wave velocity in healthy subjects: Data from the Brisighella Heart Study. Eur J Intern Med 2018; 53: 29–33.
• 19.   Iwasaki Y, Shiina K, Matsumoto C, Nakano H, Fujii M, Yamashina A, et al. Correlation of the fatty liver index with the pathophysiological abnormalities associated with cardiovascular risk markers in Japanese men without any history of cardiovascular disease: Comparison with the fibrosis-4 score. J Atheroscler Thromb 2021; 28: 524–534.
• 20.   Ministry of Health, Labor, and Welfare of Japan. The vital statistics. https://www.e-stat.go.jp/stat-search/files?page=1&toukei=00450011 (accessed March 1, 2025).
• 21.   Yamashina A, Tomiyama H, Takeda K, Tsuda H, Arai T, Hirose K, et al. Validity, reproducibility, and clinical significance of noninvasive brachial-ankle pulse wave velocity measurement. Hypertens Res 2002; 25: 359–364.
• 22.   Imai E, Horio M, Nitta K, Yamagata K, Iseki K, Hara S, et al. Estimation of glomerular filtration rate by the MDRD study equation modified for Japanese patients with chronic kidney disease. Clin Exp Nephrol 2007; 11: 41–50.
• 23.   Francque SM, van der Graaff D, Kwanten WJ. Non-alcoholic fatty liver disease and cardiovascular risk: Pathophysiological mechanisms and implications. J Hepatol 2016; 65: 425–443.
• 24.   Deanfield JE, Halcox JP, Rabelink TJ. Endothelial function and dysfunction: Testing and clinical relevance. Circulation 2007; 115: 1285–1295.
• 25.   Kasumov T, Edmison JM, Dasarathy S, Bennett C, Lopez R, Kalhan SC. Plasma levels of asymmetric dimethylarginine in patients with biopsy-proven nonalcoholic fatty liver disease. Metabolism 2011; 60: 776–781.
• 26.   Pacana T, Cazanave S, Verdianelli A, Patel V, Min HK, Mirshahi F, et al. Dysregulated hepatic methionine metabolism drives homocysteine elevation in diet-induced nonalcoholic fatty liver disease. PLoS One 2015; 10: e0136822, doi:10.1371/journal.pone.0136822.
• 27.   Cusi K. Role of obesity and lipotoxicity in the development of nonalcoholic steatohepatitis: Pathophysiology and clinical implications. Gastroenterology 2012; 142: 711–725.e6.
• 28.   DeFilippis AP, Blaha MJ, Martin SS, Reed RM, Jones SR, Nasir K, et al. Nonalcoholic fatty liver disease and serum lipoproteins: The Multi-Ethnic Study of Atherosclerosis. Atherosclerosis 2013; 227: 429–436.
• 29.   Siddiqui MS, Fuchs M, Idowu MO, Luketic VA, Boyett S, Sargeant C, et al. Severity of nonalcoholic fatty liver disease and progression to cirrhosis are associated with atherogenic lipoprotein profile. Clin Gastroenterol Hepatol 2015; 13: 1000-8.e3.
• 30.   Sonmez A, Nikolic D, Dogru T, Ercin CN, Genc H, Cesur M, et al. Low- and high-density lipoprotein subclasses in subjects with nonalcoholic fatty liver disease. J Clin Lipidol 2015; 9: 576–582.
• 31.   Abildgaard J, Danielsen E, Dorph E, Thomsen C, Juul A, Ewertsen C, et al. Ectopic lipid deposition is associated with insulin resistance in postmenopausal women. J Clin Endocrinol Metab 2018; 103: 3394–3404.
• 32.   Ko S, Kim HS. Menopause-associated lipid metabolic disorders and foods beneficial for postmenopausal women. Nutrients 2020; 12: 202, doi:10.3390/nu12010202.
• 33.   Cervellati C, Bergamini CM. Oxidative damage and the pathogenesis of menopause related disturbances and diseases. Clin Chem Lab Med 2016; 54: 739–753.
• 34.   Abildgaard J, Tingstedt J, Zhao Y, Hartling HJ, Pedersen AT, Lindegaard B, et al. Increased systemic inflammation and altered distribution of T-cell subsets in postmenopausal women. PLoS One 2020; 15: e0235174, doi:10.1371/journal.pone.0235174.
• 35.   Nakagomi A, Sunami Y, Kawasaki Y, Fujisawa T, Kobayashi Y. Sex difference in the association between surrogate markers of insulin resistance and arterial stiffness. J Diabetes Complications 2020; 34: 107442.
• 36.   Lu Y, Pechlaner R, Cai J, Yuan H, Huang Z, Yang G, et al. Trajectories of age-related arterial stiffness in Chinese men and women. J Am Coll Cardiol 2020; 75: 870–880.
• 37.   Munafò MR, Tilling K, Taylor AE, Evans DM, Davey Smith G. Collider scope: When selection bias can substantially influence observed associations. Int J Epidemiol 2018; 47: 226–235.
• 38.   Nohr EA, Liew Z. How to investigate and adjust for selection bias in cohort studies. Acta Obstet Gynecol Scand 2018; 97: 407–416.
• 39.   Endo K, Tanaka M, Sato T, Inyaku M, Nakata K, Kawaharata W, et al. High level of estimated small dense low-density lipoprotein cholesterol as an independent risk factor for the development of ischemic heart disease regardless of low-density lipoprotein cholesterol level: A 10-year cohort study. Circ J 2025; 89: 1182–1189.
• 40.   Ambrose JA, Barua RS. The pathophysiology of cigarette smoking and cardiovascular disease: An update. J Am Coll Cardiol 2004; 43: 1731–1737.
• 41.   Chen G, Chen Y, Yao Y, Ding L, Wu S, Wu W. High cumulative non-high-density lipoprotein-cholesterol concentration increases the risk of new-onset arterial stiffness: A prospective cohort study. Circ J 2025; 89: 629–637.
• 42.   Fujiyoshi A, Kohsaka S, Hata J, Hara M, Kai H, Masuda D, et al. JCS 2023 Guideline on the primary prevention of coronary artery disease. Circ J 2024; 88: 763–842.