Association of aortic stiffness and pressure pulsatility with noninvasive estimates of hepatic steatosis and fibrosis: the Framingham Heart Study

Abstract

Background:

Arterial stiffening may contribute to the pathogenesis of metabolic dysfunction-associated steatotic liver disease. We aimed to assess relations of vascular hemodynamic measures with measures of hepatic steatosis and fibrosis in the community.

Methods:

Our sample was drawn from the Framingham Offspring, New Offspring Spouse, Third Generation, and Omni-1 and Omni-2 cohorts (N=3875; mean age 56 years, 54% women). We used vibration controlled transient elastography to assess controlled attenuation parameter and liver stiffness measurements as measures of liver steatosis and liver fibrosis, respectively. We assessed noninvasive vascular hemodynamics using arterial tonometry. We assessed cross-sectional relations of vascular hemodynamic measures with continuous and dichotomous measures of hepatic steatosis and fibrosis using multivariable linear and logistic regression.

Results:

In multivariable models adjusting for cardiometabolic risk factors, higher carotid-femoral pulse wave velocity (estimated β per standard deviation=0.05; 95% confidence interval [CI]), 0.01—0.09; P=0.003), but not forward pressure wave amplitude and central pulse pressure, was associated with more liver steatosis (higher controlled attenuation parameter). Additionally, higher carotid-femoral pulse wave velocity (β=0.11; 95% CI, 0.07—0.15: P<0.001), forward pressure wave amplitude (β=0.05; 95% CI, 0.01—0.09: P=0.01), and central pulse pressure (β=0.05; 95% CI, 0.01—0.09; P=0.01) were associated with more hepatic fibrosis (higher liver stiffness measurement). Associations were more prominent among men and among participants with obesity, diabetes, and metabolic syndrome (interaction P values: <0.001 to 0.04). Higher carotid-femoral pulse wave velocity, but not forward pressure wave amplitude and central pulse pressure, was associated with higher odds of hepatic steatosis (odds ratio [OR]=1.16; 95% CI, 1.02—1.31; P=0.02) and fibrosis (OR=1.40; 95% CI, 1.19—1.64; P<0.001).

Conclusion:

Elevated aortic stiffness and pressure pulsatility may contribute to hepatic steatosis and fibrosis.

Graphical Abstract

Introduction

Metabolic dysfunction-associated steatotic liver disease (MASLD), formerly named non-alcoholic fatty liver disease,¹ is a spectrum of progressive conditions that contribute to the burden of chronic liver disease worldwide. Although the accumulation of adiposity, inflammation, and fibrosis contribute to liver disease progression, cardiovascular disease (CVD) is the most common cause of death among MASLD patients.^2–5 Indeed, MASLD shares common risk factors with CVD.⁶ A large body of evidence established that MASLD is associated with cardiometabolic disease risk factors and is the hepatic presentation of metabolic syndrome.⁷ For example, epidemiologic studies reported that liver steatosis⁸ and fibrosis^9,10 are both associated with the presence of cardiometabolic risk factors.

Elevated aortic stiffness is a strong CVD risk factor and reflects the cumulative burden of CVD risk factors on the vascular wall.^11–13 Recent studies have shown that higher aortic stiffness is a risk factor for development of^14–17 and is accelerated by presence of¹⁸ various elements of cardiometabolic disease, including insulin resistance, diabetes, hypertension, and lipid abnormalities. Multiple studies demonstrated that aortic stiffening is associated with MASLD,^19–24 suggesting a possible interplay in the pathogenesis of aortic stiffness, metabolic syndrome, and MASLD. A study in adolescent participants suggested that MASLD was only associated with increased aortic stiffness among participants with adverse metabolic profiles.²² We have shown that the foregoing associations of vascular measures with MASLD, when defined as decreased liver attenuation on multidetector computed tomography (MASLD-CT), was mostly attributable to coexisting cardiometabolic risk factors, although relations of microvascular dysfunction with MASLD-CT remained strong with consideration of established risk factors.²⁴ However, MASLD-CT is primarily a measure of liver steatosis that is insensitive to liver fibrosis.

Aortic stiffness increases the transmission of pressure pulsatility into the microcirculation, leading to microvascular damage in downstream target organs.^25,26 Aortic stiffness, in the presence of metabolic syndrome, may further contribute to the milieu of MASLD by increasing levels of hepatic steatosis, while also contributing to microvascular damage and repair, leading to inflammation and fibrosis. In a small case-control study, Sunbul et al. observed that higher estimated aortic stiffness was associated with higher liver fibrosis among patients with MASLD.¹⁹ Additionally, aortic stiffening may precede MASLD. For example, aortic stiffness preceded incident metabolic syndrome among adolescents in the Avon Longitudinal Study of Parents and Children.²⁷ However, relations in the community-based sample of aortic stiffness and vascular hemodynamic measures with measures of hepatic steatosis and more direct measures of fibrosis are unknown. Therefore, we aimed to assess relations of aortic stiffness and vascular hemodynamic measures with vibration controlled transient elastography (VCTE) measures of hepatic steatosis and fibrosis.

Methods

Our study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines.²⁸ The procedure for requesting data from the Framingham Heart Study (FHS) can be found at https://framinghamheartstudy.org/.

Participants and study sample

The cross-sectional sample was drawn from the Framingham Offspring, New Offspring Spouse, Third Generation, and Omni-1 and Omni-2 cohorts, which have been described.^29–31 Omni participants were individuals residing in the metro-west area of Massachusetts who were of African or Asian ancestry or Hispanic ethnicity. We included participants at examinations during which arterial tonometry assessment and VCTE assessment of liver steatosis and fibrosis were performed on a subset of participants. Framingham Offspring participants at examination 10 (N=1506), Omni-1 participants at examination 5 (N=198), and Third Generation (N=4095), New Offspring Spouse (N=103), and Omni-2 participants (N=410) at examination 3 were eligible for this investigation. Participants were excluded for missing or poor quality VCTE, clinical, laboratory, or tonometry data (Fig. 1). All protocols were approved by Boston University Medical Center’s Institutional Review Board, and participants provided written informed consent.

Figure 1. Flow diagram of the Framingham Heart Study analysis sample selection.

VCTE, vibration controlled transient elastography.

Vibration-controlled transient elastography

The VCTE method is a validated, non-invasive method to assess liver fibrosis^32–34 and steatosis.^35–37 An ultrasound transducer probe is mounted onto the axis of a vibrator, which enables the transducer to produce vibrations that give rise to an elastic shear wave that propagates through the target tissues. The progression and velocity of the shear wave is monitored through pulse wave ultrasound. The speed of the shear wave is directly associated with the liver stiffness measurement. Using the same radiofrequency, controlled attenuation parameter is based on the ultrasonic attenuation using a proprietary algorithm. The ultrasonic attenuation coefficient is an estimate of the total ultrasonic attenuation at 3.5 MHz. As previously described,⁹ VCTE (Fibroscan 502 Touch; Echosens, Paris, France) was performed by a certified operator to assess controlled attenuation parameter and liver stiffness measurement as measures of liver steatosis and liver fibrosis, respectively. Participants fasted for >3 hours prior to assessment and were examined in the supine position. The examiner placed the VCTE probe in the intercostal space over the right lobe of the liver. The M or XL probe was used based on the recommendations of the device and the manufacturer’s instructions. The examiner obtained a minimum of 10 measurements from each participant, and the device calculated the median values and interquartile ranges for controlled attenuation parameter and liver stiffness measurement. A qualified hepatologist evaluated all examinations for quality. Poor quality examinations (i.e., interquartile range/median ratio >0.30 when the median liver stiffness measurement is ≥7.1 kPa) were excluded.³⁸ As previously detailed,^9,39 we defined presence of clinically significant hepatic steatosis and fibrosis as controlled attenuation parameter ≥290 dB/m and liver stiffness measurement ≥8.2 kPa, respectively.

Vascular hemodynamic assessment

Hemodynamic assessment was performed as previously described.^40,41 We obtained noninvasive arterial tonometry with simultaneous electrocardiography from supine participants for the brachial, radial, femoral, and carotid arteries using a custom tonometer. We performed pulsed Doppler of the left ventricular outflow tract to assess aortic flow. We digitized and transferred tonometric data and Doppler audio to a core laboratory (Cardiovascular Engineering, Inc., Norwood, MA) for blinded analyses. We signal-averaged and synchronized tonometry waveforms using the electrocardiographic R-wave. The signal averaged brachial pressure waveform peak and trough were calibrated to cuff systolic and diastolic pressures, respectively. Mean arterial pressure was calculated as the integral of the calibrated brachial pressure tonometry waveform.⁴¹ All other tonometry waveforms were calibrated by using brachial mean and diastolic pressures. We calculated carotid-femoral pulse wave velocity (CFPWV) from tonometry waveforms and body surface measurements that adjusted for parallel transmission in the aortic arch and brachiocephalic artery as previously described.²⁵ We calculated central pulse pressure as the difference between carotid systolic and diastolic blood pressures. We defined forward pressure wave amplitude as the difference between pressure at the foot and at the peak of the forward pressure waveform by performing time domain wave separation analysis using central pressure and flow.⁴² We calculated augmentation index as the fraction of central pulse pressure attributable to late systolic pressure augmentation. Characteristic impedance was calculated in the time domain as the ratio of the pressure increase and the flow increase during the time interval between flow onset and 95% of peak flow.⁴²

Clinical evaluation and covariates

Medical history and physical examination were performed routinely at each research examination. Age, sex, use of antihypertensive and hyperlipidemia medications, alcohol consumption, and smoking status were assessed through questionnaires. Height (meters) and weight (kilograms) were assessed during the examination. Body mass index was calculated as weight in kilograms divided by height in meters squared. Waist circumference was assessed by a measuring tape at the horizontal level of the umbilicus. Blood pressures were assessed during tonometry. Serum cholesterol, C-reactive protein, alanine aminotransferase, and aspartate aminotransferase levels were measured from a fasting blood test. Hemoglobin A1c was measured by high-performance liquid chromatography assays.⁴³ Criteria for diabetes were a fasting glucose level of ≥126 mg/dL or treatment with insulin or an oral hypoglycemic agent. Physical activity index was assessed as a composite score constructed for each participant by weighting each hour in their typical day based on their activity level and summing up these weighted hours over a 24-hour period as previously described.⁴⁴ Participants reported physical activity using the physical activity index or the Physical Activity Scale for the Elderly. The physical activity index assessed the number of hours in a typical day spent sleeping (weighting factor [WF]=1) and in sedentary (WF=1.1), slight (WF=1.5), moderate (WF=2.4), and heavy activities (WF=5). The Physical Activity Scale for the Elderly assessed the number of times per week participants engaged in physical activities.^45,46 Metabolic syndrome was defined as meeting ≥3 of 5 criteria:^47,48 (1) high waist circumference (≥102 cm in men; ≥88 cm in women); (2) high fasting triglyceride (≥150 mg/dL/≥1.7mmol/L or treatment for elevated lipids); (3) high blood pressure (≥130 mm Hg systolic blood pressure, ≥85 mm Hg diastolic blood pressure, or treatment for hypertension); (4) low high-density lipoprotein (HDL) cholesterol (<40 mg/dL in men; <50 mg/dL in women); and (5) high fasting glucose (≥100 mg/dL or treatment for elevated glucose).

Statistical analysis

We tabulated characteristics for the sample. CFPWV rises nonlinearly and shows significant increases in variability as age progresses, leading to heteroscedasticity and a distribution that is heavily skewed to the right.⁴² Therefore, we inverted CFPWV to limit heteroscedasticity; then we multiplied it by −1000 to convert units to ms/m and rectify directionality of associations with aortic stiffness. We natural logarithm transformed triglycerides, fasting blood glucose, alcohol consumption (past year as drinks per week), liver aminotransferases, and liver stiffness measurement to normalize skewed distributions. We ranked physical activity index and Physical Activity Scale for the Elderly data individually into quintiles and used a combined quintile variable. We standardized (mean=0, standard deviation=1) all continuous measures for modeling.

We assessed relations of controlled attenuation parameter and liver stiffness measurement with various cardiometabolic risk factors using stepwise multivariable regression models that adjusted for age, sex, cohort, and mean arterial pressure; the threshold for inclusion and removal from the model was two-sided P<0.05. We selected covariates as potential confounders a priori based on literature review and included body mass index, waist circumference, triglycerides, fasting blood glucose, total cholesterol, diabetes, alcohol consumption (past year as drinks per week), smoking status, hypertension treatment, lipid lowering treatment, and physical activity index.

We used multivariable linear and logistic regression models to relate measures of vascular hemodynamics with (1) continuous measures of hepatic steatosis and fibrosis (controlled attenuation parameter and liver stiffness measurement) and (2) presence of hepatic steatosis or fibrosis, respectively. We included age, sex, cohort, and mean arterial pressure in minimally-adjusted models (Model 0). Expanded models further adjusted for all covariates that were retained in the foregoing stepwise models (Model 1). We further adjusted models where continuous liver stiffness measurement or presence of hepatic fibrosis were dependent variables for controlled attenuation parameter (Model 2). We excluded mean arterial pressure as a covariable in models where mean arterial pressure, systolic blood pressure, or diastolic blood pressure were independent variables.

For the continuous primary hemodynamic measures, we assessed potential effect modification (interaction) by median age, sex, body mass index (<30 vs ≥30), diabetes status, and metabolic syndrome status by incorporating corresponding interaction terms in the regression models. For significant interactions, we performed stratified analyses. Additionally, we performed a sensitivity analysis of the primary analysis by excluding participants using medications for hypertension, hyperlipidemia, or diabetes. We also assessed the associations of aortic stiffness and pressure pulsatility measures with levels of liver aminotransferases. Furthermore, we assessed the relations of inflammatory marker C-reactive protein with continuous measures of hepatic steatosis and fibrosis using multivariable linear regression. We performed all analyses with SAS version 9.4 for Windows (SAS Institute, Cary, NC) and considered two-sided P<0.05 as statistically significant.

Results

The sample included 3875 participants (2108 women [54%]). The mean (SD) age was 56 (10) years. We present characteristics of the sample in Table 1 and compare these characteristics between included and excluded participants in Table S1. The sample comprised relatively healthy middle-aged and older adults with similar prevalence of diabetes but lower prevalence of smoking and hypertension treatment (compared to the general global adult population). We present multivariable-adjusted cross-sectional relations of cardiometabolic risk factors with continuous measures of hepatic steatosis and fibrosis in Table S2. The relations are similar as previously reported in FHS participants.⁹

Table 1.

Clinical characteristics of the sample (N=3875).

Variable	Value*
Age, years	56±10
Women, n (%)	2108 (54)
Offspring Cohort, n (%)	630 (16)
Gen 3 Cohort, n (%)	2869 (74)
Omni-1 Cohort, n (%)	72 (2)
Omni-2 Cohort, n (%)	260 (7)
New Offspring Spouse Cohort, n (%)	44 (1)
Body mass index, kg/m2	28.3±5.5
Waist circumference, cm	39.3±5.8
Total/high-density lipoprotein cholesterol ratio	3.37±1.14
Triglycerides, mg/dL	91 [67, 132]
Fasting glucose, mg/dL	97 [90, 105]
Hemoglobin A1c, %	5.3 [5.1, 5.6]
Alcohol consumption, drinks per week (past year)	3 (1, 8)
Hypertension treatment, n (%)	1142 (29)
Lipid lowering treatment, n (%)	1121 (29)
Prevalent diabetes, n (%)	335 (9)
Prevalent metabolic syndrome, n (%)	1543 (40)
Current smoking, n (%)	222 (6)
Physical activity index, unitless†	32 [28, 35]
PASE activity index, no. times exercise/week‡	6 [5, 8]

^*Values are mean±standard deviation, number (%), or median [25^th, 75^th percentile].

^†Physical activity index was primarily recorded in 3^rd generation cohorts; it is a weighted combination of hours 1×sleeping + 1.1×sedentary + 1.5×low exercise + 2.4×moderate exercise + 5×high exercise.

^‡PASE (Physical Activity Scale for the Elderly) was recorded only in 2^nd generation cohorts.

We present associations of aortic stiffness and pressure pulsatility measures with continuous measures of hepatic steatosis and fibrosis in Table 3. In multivariable models, higher CFPWV (β per SD=0.05; 95% CI, 0.01—0.09; P=0.003), but not forward pressure wave amplitude (β per SD=0.01; 95% CI, −0.01—0.03; P=0.63) and central pulse pressure (β per SD=0.02; 95% CI, −0.02—0.06; P=0.28), was associated with more liver steatosis. In similar models that further adjusted for controlled attenuation parameter, higher CFPWV (β per SD=0.11; 95% CI, 0.07—0.15; P<0.001), forward pressure wave amplitude (β per SD=0.05; 95% CI, 0.01—0.09; P=0.01), and central pulse pressure (β per SD=0.05; 95% CI, 0.01—0.09; P=0.01) were associated with greater liver fibrosis. In the sensitivity analysis, we present similar results excluding participants using medications for hypertension, hyperlipidemia, or diabetes (Table S3). CFPWV, the reference standard for aortic stiffness, remained significant in a model that adjusts for multiple cardiometabolic risk factors, with an effect size (β) comparable to that of controlled attenuation parameter in the final model (Model 2, Table 3). We present similar results for models of CFPWV with continuous measures of hepatic steatosis and fibrosis that further adjust for C-reactive protein (Table S4). We observed significant associations of C-reactive protein with controlled attenuation parameter; however, adjusting for C-reactive protein did not explain the relations of CFPWV with controlled attenuation parameter. C-reactive protein was also related to liver stiffness measurement in minimally-adjusted models, but the association was completely attenuated in Models 1 and 2. Therefore, adjusting for C-reactive protein did not explain the relations of CFPWV with liver stiffness measurement. Relations of measures of aortic stiffness and pressure pulsatility with liver aminotransferases are presented in Table S5.

Table 3.

Relations of measures of aortic stiffness with continuous measures of hepatic steatosis and fibrosis (N=3875).

	Controlled attenuation parameter			Liver stiffness measurement#
Variable and Model	Est. β±SE (95% CI)	P	Adjusted R2	Est. β±SE (95% CI)	P	Adjusted R2
CFPWV
Model 0*	0.21±0.02 (0.17, 0.25)	<0.001	0.11	0.19±0.02 (0.15, 0.23)	<0.001	0.05
Model 1†	0.05±0.02 (0.01, 0.09)	0.003	0.45	0.12±0.02 (0.08, 0.16)	<0.001	0.13
Model 2‡	-- --	--	--	0.11±0.02 (0.07, 0.15)	<0.001	0.13
FWA
Model 0*	0.02±0.02 (−0.02, 0.06)	0.29	0.09	0.07±0.02 (0.03, 0.11)	<0.001	0.04
Model 1†	0.01±0.01 (−0.01, 0.03)	0.63	0.45	0.05±0.02 (0.01, 0.09)	0.01	0.12
Model 2‡	-- --	--	--	0.05±0.02 (0.01, 0.09)	0.01	0.13
CPP
Model 0*	−0.05±0.02 (−0.09, −0.01)	0.01	0.09	0.04±0.02 (0.00, 0.08)	0.03	0.03
Model 1†	0.02±0.02 (−0.02, 0.06)	0.28	0.45	0.06±0.02 (0.02, 0.10)	0.004	0.12
Model 2‡	-- --	--	--	0.05±0.02 (0.01, 0.09)	0.01	0.13

CFPWV, carotid-femoral pulse wave velocity. FWA, forward wave amplitude. CPP, central pulse pressure. All estimated β±standard error (SE) and 95% confidence interval (CI) represent standard deviation difference in the hepatic steatosis or fibrosis measure per standard deviation difference in stiffness variable.

^*Adjusted for age, sex, cohort, and mean arterial pressure.

^†Further adjusted for waist circumference, triglycerides, blood glucose, body mass index, total cholesterol, diabetes, alcohol consumption, smoking status, hypertension treatment and lipid treatment.

^‡Further adjusted for controlled attenuation parameter.

^#Liver stiffness measurement is natural log-transformed in all models.

We present associations of secondary vascular hemodynamic measures with continuous measures of hepatic steatosis and fibrosis in Table S6. In multivariable models, higher mean arterial pressure and systolic blood pressure, but not diastolic blood pressure, peripheral pulse pressure, characteristic impedance, or augmentation index, were associated with more liver steatosis. In similar models that further adjusted for controlled attenuation parameter, higher systolic blood pressure, peripheral pulse pressure, and characteristic impedance were associated with more liver fibrosis. Mean arterial pressure, diastolic blood pressure, and augmentation index were not associated with liver fibrosis.

We summarize interactions for associations of measures of aortic stiffness with continuous measures of hepatic steatosis and fibrosis in Table S7. Among younger participants (below median age≤57 years), higher CFPWV (β per SD=0.07; 95% CI, 0.01—0.13; P=0.01) and central pulse pressure (β per SD=0.05; 95% CI, 0.01—0.10; P=0.02) were associated with higher liver steatosis; however, CFPWV (β per SD=0.04; 95% CI, −0.01—0.08; P=0.10) and central pulse pressure (β per SD=0.00; 95% CI, −0.04—0.04; P=0.98) were not associated with liver steatosis among older participants (age>57 years). The association of higher CFPWV with higher liver fibrosis was stronger among participants with obesity, diabetes, and metabolic syndrome compared to participants without those conditions (Fig. 2). Additionally, we observed significant associations of higher forward pressure wave amplitude and central pulse pressure with higher liver fibrosis among men but not women and among participants with prevalent metabolic syndrome but not among participants without metabolic syndrome (Fig. 2).

Figure 2. Effect modification by obesity status, diabetes status, sex, and metabolic syndrome status on associations of vascular hemodynamic measures with liver stiffness measure.

Effect sizes (βs) and 95% CIs from linear regression models that assessed associations of tonometry measures with liver fibrosis stratified by BMI group (A), diabetes status (B), sex (C), and metabolic syndrome status (D). Liver stiffness measurement is natural log-transformed in all models. niCFPWV, negative inverse carotid-femoral pulse wave velocity. FWA, forward pressure wave amplitude. CPP, central pulse pressure. BMI, body mass index. MetS, metabolic syndrome. Models in panel A are adjusted for age, sex, cohort, mean arterial pressure, waist circumference, triglycerides, blood glucose, total cholesterol, diabetes, alcohol consumption, smoking status, hypertension treatment and lipid treatment, and controlled attenuation parameter. Models in panel B are adjusted for age, sex, cohort, mean arterial pressure, body mass index, waist circumference, triglycerides, blood glucose, total cholesterol, alcohol consumption, smoking status, hypertension treatment and lipid treatment, and controlled attenuation parameter. Models in panel C are adjusted for age, cohort, mean arterial pressure, body mass index, waist circumference, triglycerides, blood glucose, total cholesterol, diabetes, alcohol consumption, smoking status, hypertension treatment and lipid treatment, and controlled attenuation parameter. Models in panel D are adjusted for age, sex, cohort, mean arterial pressure, body mass index, waist circumference, triglycerides, blood glucose, total cholesterol, diabetes, alcohol consumption, smoking status, hypertension treatment and lipid treatment, and controlled attenuation parameter.

We present associations of measures of aortic stiffness with presence of hepatic steatosis and fibrosis (categorical dependent variables) in Table 4. Higher CFPWV, but not forward pressure wave amplitude and central pulse pressure, was associated with higher odds of hepatic steatosis (odds ratio [OR]=1.16; 95% CI, 1.02—1.31) and fibrosis (OR=1.40; 95% CI, 1.19—1.64). Among secondary vascular hemodynamic measures (Table S8), higher mean arterial pressure and diastolic blood pressure were associated with higher odds of hepatic steatosis. No secondary vascular hemodynamic measure was associated with presence of hepatic fibrosis.

Table 4.

Relations of measures of aortic stiffness with presence of hepatic steatosis and fibrosis (N=3875).

	Hepatic steatosis*		Hepatic fibrosis†
Variable and Model	OR (95% CI)	P	OR (95% CI)	P
CFPWV
Model 0‡	1.56 (1.41, 1.73)	<0.001	1.69 (1.46, 1.97)	<0.001
Model 1§	1.16 (1.02, 1.31)	0.02	1.43 (1.22, 1.68)	<0.001
Model 2#	--	--	1.40 (1.19, 1.64)	<0.001
Forward wave amplitude
Model 0‡	1.03 (0.95, 1.12)	0.47	1.10 (0.98, 1.24)	0.12
Model 1§	1.01 (0.91, 1.12)	0.86	1.03 (0.91, 1.16)	0.67
Model 2#	--	--	1.03 (0.91, 1.17)	0.63
Central pulse pressure
Model 0‡	0.90 (0.83, 0.99)	0.03	1.04 (0.91, 1.18)	0.60
Model 1§	1.03 (0.93, 1.15)	0.58	1.07 (0.94, 1.23)	0.31
Model 2#	--	--	1.07 (0.94, 1.23)	0.32

CFPWV, carotid-femoral pulse wave velocity. Odds ratios (OR) expressed per 1 standard deviation higher value. CI, confidence interval.

^*Controlled attenuation parameter≥290 dB/m; 1083 (28%) participants had prevalent hepatic steatosis.

^†Liver stiffness measurement≥8.2 kPa; 368 (10%) participants had prevalent hepatic fibrosis.

^‡Adjusted for age, sex, cohort, and mean arterial pressure.

^§Further adjusted waist circumference, triglycerides, blood glucose, body mass index, total cholesterol, diabetes, alcohol consumption, smoking status, hypertension treatment and lipid treatment.

^#Further adjusted for controlled attenuation parameter.

Discussion

In this cross-sectional analysis of middle-aged and older FHS participants, higher aortic stiffness was associated with higher levels of liver steatosis. Additionally, higher aortic stiffness and pressure pulsatility were associated with greater liver fibrosis, and associations were more prominent in men and among participants with obesity, diabetes, and metabolic syndrome. Higher aortic stiffness was also associated with prevalent hepatic steatosis and fibrosis. Our data are consistent with the hypothesis that abnormal central vascular hemodynamics may contribute to higher levels of hepatic steatosis and fibrosis, both through known effects of aortic stiffness on the pathogenesis of cardiometabolic abnormalities,^14,15,17,49 as well as through direct adverse effects of higher aortic stiffness and excessive pressure pulsatility on liver structure and function that are independent of various cardiometabolic risk factors.

An accumulating body of literature has shown a consistent association of MASLD with aortic stiffness.^19–24 Multiple clinical and community-based studies have suggested that MASLD precedes and contributes to arterial stiffness.^27,50–52 However, our study of participants in the community (most of whom were free of MASLD) demonstrated a continuous, graded relation between aortic stiffness and liver dysfunction, consistent with the hypothesis that higher aortic stiffness and pressure pulsatility may contribute to development of liver steatosis and fibrosis, possibly through mechanisms that involve liver small vessel damage and dysfunction. Prior studies have revealed that elevated aortic stiffness and pressure pulsatility were associated with downstream damage in high flow, low resistance organs, like the brain and kidneys, through mechanisms that included microvascular damage and dysfunction.^53–57 The liver also has a low-resistance vascular bed that receives approximately 25% of the cardiac output from a dual vascular supply via the portal vein and hepatic artery;⁵⁸ therefore, the liver microvasculature is vulnerable to direct damage associated with excessive pressure pulsatility entering into and dissipating in hepatic sinusoids. However, longitudinal, interventional, and basic science studies (including existing animal models of liver MASLD^59,60) are required to investigate mechanisms of the putative effects of elevated aortic stiffness and pressure pulsatility on hepatic microvascular and parenchymal injury and scarring.

Higher pressure pulsatility was associated with higher liver stiffness in men but not in women. These data are consistent with current trends in MASLD. For example, the prevalence of MASLD is higher in adult men^61–63 as well as in male children and adolescents.⁶⁴ Some researchers specifically observed that fatty liver^65,66 and liver fibrosis^67,68 were more prevalent among men while others observed no sex differences.⁶⁹ Sex-specific structural and functional differences in the liver or aorta as well as sex hormones may contribute to differences in the mechanisms by which abnormal aortic stiffness and pressure pulsatility may contribute to hepatic steatosis and fibrosis. Park et al. observed that menopause and presence of estrogen replacement therapy were independent risk factors for prevalent MASLD in women.⁷⁰ In a FHS sample, the incidence of hepatic steatosis was similar for men and postmenopausal women but lower among premenopausal women.⁷¹ Similarly, after puberty, measures of pressure pulsatility decrease in women and increase in men; however, the pressure pulsatility increases disproportionately in postmenopausal women compared to men of similar age. Thus, greater discontinuity (or mismatch) in the stiffness of the aorta relative to the hepatic arteries may explain the lower liver sensitivity to pulsatile stress in women. Sex differences of the relations of elevated aortic stiffness on liver structure and function warrants further investigation.

Due to hepatic arterial buffering, the hepatic artery can compensate for flow changes in response to changes in portal venous flow; however, changes of the hepatic arterial flow do not produce reciprocal changes in portal flow.^72,73 Furthermore, aortic stiffness contributes to hypertension,⁷⁴ left ventricular hypertrophy,^75–78 and diastolic dysfunction,^79–81 which can reduce cardiac output. For example, in a recent study in youths, aortic stiffening preceded worsening cardiac hypertrophy, which was partially mediated by hypertension and insulin resistance.⁷⁵ In the setting of elevated aortic stiffness and pressure pulsatility, the hepatic arterioles may remodel or become damaged, resulting in impaired vascular reactivity.^82,83 Although the aforementioned structural changes in the hepatic artery mitigate excessive pressure and flow pulsatility within the sinusoidal capillaries, the trade-off is a reduction in the proportion of blood flow from the highly-oxygenated hepatic artery with an increase in the blood flow from the poorly-oxygenated portal vein. Although a physiologic oxygen gradient exists in healthy livers, destruction of the oxygen gradient by hypoxia has been shown to accelerate hepatic steatosis.^84,85 Imbalances in relative blood supply to the liver may render hepatocytes more susceptible to hypoxia and metabolic dysfunction. Additionally, higher relative portal flow increases the burden to metabolize and eliminate microbes and toxins from the gastrointestinal tract. Thus, elevated aortic stiffness may indirectly contribute to hepatic steatosis and subsequent fibrosis by disrupting the metabolic functions of hepatic cells. Metabolic disorders, such as obesity, diabetes, and CVD, often coincide with MASLD. With advancing age, however, the presence of comorbidities and the cumulative effects of metabolic risk factors may attenuate or obscure the association of aortic stiffness on liver fat accumulation. Indeed, higher CFWPV was associated with greater hepatic steatosis predominantly in younger participants (Interaction P=0.01). Since aortic stiffness antedates and contributes to development of components of metabolic disease,^14–18 aortic stiffening may contribute to the burden of MASLD. Yet, the negative relation of CFPWV with AST/ALT ratio suggests that our study was detecting a gradient of subclinical liver damage in which fibrosis is not advanced (Table S5). Additionally, we observed that the association of higher aortic stiffness and pressure pulsatility with more hepatic fibrosis was more prominent among participants with obesity, diabetes, and metabolic syndrome. These data imply that complex interrelations exist among aortic stiffness, metabolic syndrome, and markers of MASLD; however, additional longitudinal studies that address the potential moderation and mediation of the observed associations are warranted.

Our study has limitations that should be considered. We employed a cross-sectional observational study design, and this restricts our ability to infer causal and temporal relations between vascular hemodynamic measures and liver dysfunction and damage. However, given that our observations pertain to subacute fibrosis occurring prior to changes in intrahepatic resistance, we anticipate that the extent of reverse causality would be limited. Our study is susceptible to type-1 error because we did not adjust for multiple testing; however, we observed strong associations particularly for liver stiffness measure outcomes that would survive multiple testing adjustments. Aortic stiffness was associated with hepatic steatosis in younger but not older participants; however, we cannot rule out potential confounding by generational and historic factors that may contribute to observed differences between age groups in our cross-sectional study. We define presence of hepatic steatosis and fibrosis based on controlled attenuation parameter and liver stiffness measurement thresholds, and without liver histology to confirm the diagnosis, we may have misclassified participants on hepatic steatosis or fibrosis status. Since this nondifferential misclassification would have biased our results to the null, it would not have contributed to the significant association of higher CFPWV with hepatic fibrosis in the present study. Although we included the multiethnic Omni cohorts, the sample was primarily middle aged to older White participants of European descent; therefore, our findings may not be generalizable to other ages or ethnic groups. We adjusted for known cardiometabolic risk factors, but the possibility of residual confounding by unmeasured or unknown factors remains.

In conclusion, in this cross-sectional study of middle-aged and older FHS participants, higher aortic stiffness was related to higher liver steatosis and higher aortic stiffness and pressure pulsatility were associated with higher liver stiffness. Associations were more prominent in men and among participants with obesity and cardiometabolic disease. These findings support the hypothesis that elevated aortic stiffness and pressure pulsatility may contribute to higher levels of hepatic steatosis and fibrosis, independent of various cardiometabolic risk factors. Thus, aortic stiffness may exert direct damage on the liver microvasculature through pulsatile hemodynamic mechanisms as well as indirectly by contributing to cardiometabolic disease and hepatic steatosis. Treatments for MASLD are limited to mostly behavioral changes, and transplantation at advanced stages is expensive; therefore, prevention is essential. A growing body of evidence suggests that targeted interventions, including exercise, weight loss, and pharmacological treatments such as renin-angiotensin-aldosterone system blockers, neutral endopeptidase inhibitors, sodium-glucose cotransporter-2 inhibitors, senolytics, and others medications, may offer a promising strategy to prevent or reverse aortic stiffness.^86–93 Further research is warranted to explore these interventions in large, diverse populations to solidify their efficacy on reducing aortic stiffness and prevention of its putative downstream effects on liver health.

Table 2.

Vascular hemodynamic measures and measures of hepatic steatosis and fibrosis (N=3875).

Variable	Value
Primary vascular measures
Carotid-femoral pulse wave velocity, m/s	8.5±2.6
Forward wave amplitude, mm Hg	51.5±15.2
Central pulse pressure, mm Hg	63.5±19.2
Secondary vascular measures
Systolic blood pressure, mm Hg	129.3±17.6
Diastolic blood pressure, mm Hg	67.0±9.2
Mean arterial pressure, mm Hg	93.5±11.4
Characteristic impedance, dyne × sec/cm5	226±90
Augmentation index, %	16.6±11.9
Measures of hepatic steatosis and fibrosis
Controlled attenuation parameter, dB/m	259±55
Prevalent hepatic steatosis, n (%)*	1083 (28)
Liver stiffness measurement, kPa	5.1 [4.2, 6.3]
Prevalent hepatic fibrosis, n (%)†	368 (10)

Data are presented as mean±standard deviation or median [25th, 75th percentile].

^*Controlled attenuation parameter steatosis≥290 dB/m.

^†Liver stiffness measurement≥8.2 kPa.

Highlights.

• We assessed the cross-sectional association of measures of aortic stiffness and pressure pulsatility with noninvasive estimates of hepatic steatosis and fibrosis in the community.

• Adjusting for cardiometabolic risk factors, higher aortic stiffness was associated with higher estimates of liver steatosis and fibrosis.

• Elevated aortic stiffness and pressure pulsatility may directly and indirectly contribute to liver dysfunction.

• Noninvasive hemodynamic assessment using arterial tonometry may have utility for indicating potential subclinical target organ damage to the liver.

Non-standard Abbreviations and Acronyms

CFPWV

carotid-femoral pulse wave velocity

CT

computed tomography

CVD

cardiovascular disease

FHS

Framingham Heart Study

MASLD

metabolic dysfunction-associated steatotic liver disease

VCTE

vibration controlled transient elastography