Cardiovascular Health Trajectories and Prevalent Metabolic Dysfunction‐Associated Steatotic Liver Disease in Midlife: The CARDIA Study

Abstract

Background

Metabolic‐dysfunction associated steatotic liver disease (MASLD) is associated with prevalent cardiovascular disease. More favorable cardiovascular health (CVH) profiles are associated with a lower prevalence of MASLD in cross‐sectional studies. The relationship between long‐term CVH patterns and MASLD prevalence in midlife remains unknown.

Methods and Results

Participants (aged 18–30 years at baseline) of the CARDIA (Coronary Artery Risk Development in Young Adults) study who had individual CVH components measured at 7 examinations over 20 years and liver fat assessed by noncontrast computed tomography at year 25 follow‐up were included. CVH score was defined using published American Heart Association definitions. Group‐based trajectory modeling was used to identify CVH trajectories. MASLD was defined as liver attenuation of ≤51 Hounsfield units with at least 1 metabolic risk factor after excluding other causes of liver fat. Logistic regression was used to examine associations of CVH trajectory groups and MASLD prevalence. At baseline, 39% of 2529 participants had high and 5% had low CVH, respectively. MASLD prevalence at year 25 was 23% (n=587). Five distinct CVH trajectories were identified. Between the 2 groups that started at similar CVH scores, those whose CVH declined over time had a higher prevalence of MASLD at year 25 (7.0% in high‐stable versus 23.0% high‐decreasing; 24.4% in moderate‐stable versus 35.7% in moderate‐decreasing). Lower and decreasing trajectories were associated with higher year‐25 MASLD prevalence compared with the high‐stable trajectory.

Conclusions

Achieving and maintaining high CVH scores starting in young adulthood lowers the risk of prevalent MASLD in midlife.

Nonstandard Abbreviations and Acronyms

AHA

American Heart Association

CARDIA

Coronary Artery Risk Development in Young Adults

CVH

cardiovascular health

HU

Hounsfield units

LA

liver attenuation

MASLD

metabolic dysfunction‐associated steatotic liver disease

NAFLD

nonalcoholic fatty liver disease

Clinical Perspective.

What Is New?

• A Life's Simple 7 cardiovascular health (CVH) score has meaningful value and is associated with metabolic dysfunction‐associated steatotic liver disease (MASLD) risk in a time‐dependent manner, and we found that achieving high CVH scores starting in young adulthood and maintaining levels with aging can help achieve lower risk of prevalent MASLD in midlife.

• These findings suggest that assessing one's overall CVH throughout their lifetime, starting in young adulthood, and increasing efforts to improve or maintain CVH throughout adulthood are critical for the primary prevention of MASLD.

What Are the Clinical Implications?

• The relationship between CVH trajectories and liver diseases should be further investigated based on new metabolic dysfunction and alcohol‐associated liver disease (MetALD) and MASLD criteria, which accounts for metabolic abnormalities and recognized overlap between alcoholic liver disease and MASLD.

Metabolic dysfunction‐associated steatotic liver disease (MASLD) is characterized by macrovascular steatosis of the liver in the setting of ≥1 metabolic risk factors and in the absence of other causes, including excess alcohol intake. Formally termed nonalcoholic fatty liver disease (NAFLD), MASLD has become the most common cause of liver disease in high‐income countries, with global prevalence of ≈30%. ¹ With a continuous increase in obesity and diabetes in the US population, the modeling analysis forecasts significant increases in prevalence and progression of MASLD‐related liver disease and mortality. ² MASLD is a wide spectrum of disease ranging from more benign metabolic dysfunction‐associated steatosis (isolated fatty liver) to more severe metabolic dysfunction‐associated steatohepatitis, which is present in up to 20% of adults with MASLD and 3% to 6% of the US population. ³ With its potential to progress to cirrhosis and hepatocellular carcinoma, MASLD is associated with significant morbidity and mortality but particularly with a high prevalence of cardiovascular disease (CVD) and its risk factors, including hypertension, dyslipidemia, type 2 diabetes, obesity, and metabolic syndrome. ⁴ , ⁵ , ⁶ , ⁷ , ⁸ Patients with MASLD are more likely to die from complications of CVD than from liver‐related causes. ⁹ Thus, addressing the risk factors related to CVD may be critical in reducing both MASLD as well as the complications and mortality associated with MASLD.

The American Heart Association (AHA) identified Life's Simple 7 (LS7), which comprise 7 components (cholesterol, blood pressure, blood glucose level, body mass index, nonsmoking status, diet, and physical activity patterns), to be important for an individual's cardiovascular health (CVH) profile. ¹⁰ , ¹¹ Multiple studies have suggested that having an ideal CVH profile is linked with a longer life with significantly fewer cardiovascular events and better quality of life, especially in adults who maintain ideal CVH profiles into their midlife. ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹

To date, no studies have explored the relationship between longitudinal CVH profiles throughout adulthood and MASLD. ²⁰ , ²¹ , ²² , ²³ Establishing the relationship between CVH profiles through adulthood and MASLD in midlife is critical to inform future public health interventions targeting MASLD prevention in early adulthood. The goal of this proposal is to understand whether the AHA's LS7 CVH scores in young adulthood are associated with the prevalence of MASLD in midlife and whether different trajectories of CVH scores throughout adulthood are associated with prevalent midlife MASLD. It is hypothesized that participants who have higher CVH scores in young adulthood, and who maintain higher levels of CVH throughout adulthood, will be less likely to have MASLD in midlife.

METHODS

CARDIA (Coronary Artery Risk Development in Young Adults) study data are available upon reasonable request from the CARDIA Coordinating Center. CARDIA investigators are eager to collaborate with investigators interested in using CARDIA data. Please see the CARDIA website (http://www.cardia.dopm.uab.edu/publications‐2/publications‐documents) for publication policies and for a list of CARDIA investigators. CARDIA data are also publicly available on the National Institutes of Health‐supported BioLINCC and dbGaP platforms.

Study Population

The CARDIA study is a multicenter prospective cohort study of 5115 Black and White men and women, aged 18 to 30 years at baseline exam (year 0; 1985–1986), recruited from 4 study sites: Birmingham, Alabama; Minneapolis, Minnesota; Chicago, Illinois; and Oakland, California. ²⁴ In this study, data from the follow‐up examinations of participants were extracted at years 2, 5, 7, 10, 15, 20, and 25, with participant retention rates of 91%, 86%, 81%, 79%, 74%, 72%, and 72%, respectively. The CARDIA study is ongoing, and study details are described elsewhere. ²⁴ All research was conducted in accordance with both the Declarations of Helsinki and Istanbul. CARDIA was approved by the institutional review board at each field center, and all participants provided written consent.

Study Sample

All nonpregnant women and men in CARDIA who had CVH scores at baseline and liver fat measurement at year 25 were included. Participants who were missing CVH scores at baseline or not examined at year 25 (n=777) were excluded. Participants who identified as transgender (n=2) and a participant withdrawing consent (n=1) were excluded. Participants who were pregnant at year 25 (n=2) and had missing liver fat measurement at year 25 (n=1358) were excluded. Participants with self‐reported cirrhosis or viral hepatitis at year 25 (n=51), risk factors for other chronic liver disease (eg, intravenous drug use [n=77] or causes of secondary hepatic steatosis: alcohol consumption ≥7 drinks per week in women and ≥14 drinks per week in men [n=269], HIV [n=22]) and medications known to cause hepatic steatosis (eg, valproic acid, methotrexate, tamoxifen, amiodarone [n=27]) at year 25 were excluded. The remaining participants were included in the MASLD‐eligible sample population. The final analytical sample included 2529 participants (Figure 1).

Figure 1. Flow diagram for inclusion and exclusion of participants.

CVH indicates cardiovascular health; and MASLD, metabolic dysfunction‐associated steatotic liver disease.

Measures

The primary exposure of this study was CVH measured at baseline exam and the trajectory of CVH scores based on individual components of CVH measured throughout follow‐up. The CVH profiles are based on the 7 metrics (diet, physical activity, smoking, body mass index, blood pressure, total cholesterol, and glucose) included in the AHA's LS7 criteria and defined as described in Table 1. ²⁵ AHA categorized each of the 7 metrics into 3 levels, which are scored as either ideal (2 points), intermediate (1 point), or poor (0 points). Ideal CVH in LS7 was defined to represent favorable health behaviors and health factors, based on numerous studies showing the association of each ideal CVH construct with longevity, disease‐free survival, and quality of life. ²⁶ The values for the factors were summed for an overall LS7 CVH score (0–14), which are categorized into either high (12–14 points; reference), moderate (8–11 points), or low (0–7 points) CVH, with the higher scores indicating more optimal CVH. Hence, the summed CVH scores represent a comprehensive assessment of one's CVH and overall management of cardiovascular risk factors.

Table 1.

LS7 Definitions of Ideal, Intermediate, and Poor CVH as Defined in the CARDIA Study

CVH metrics	LS7 definitions of CVH profiles
	Ideal CVH (score=2)	Intermediate CVH (score=1)	Poor CVH (score=0)
Blood pressure, mm Hg	SBP <120 and DBP <80	SBP 120–139 or DBP 80–89, or treated to SBP <140/DBP <90	SBP ≥140 or DBP ≥90
Body mass index, kg/m2	<25	25–29.9	≥30
Current smoking	Never smoker or former smoker who quit smoking >12 mo ago	Former smoker who quit smoking ≤12 mo ago	Yes
Diet (no. of healthy components)*	4 or 5	2 or 3	0 or 1
Fasting plasma glucose, mg/dL	<100	100–125 or treated to <126	≥126
Physical activity, EU†	≥300	100–299	<100
Total cholesterol, mg/dL	<200	200–239 or treated to <240	≥240

CVH indicates cardiovascular health; DBP, diastolic blood pressure; EU, exercise units; LS7, Life's Simple 7; and SBP, systolic blood pressure.

^*The 5 dietary score components include: (1) ≥4.5 servings per day of fruits/vegetables, (2) ≥7 oz of fish per week, (3) ≥3 servings of whole grains per day, (4) <1500 mg per day of sodium, and (5) <32 oz per week of sugar‐sweetened beverages. Dietary values are scaled according to 2000 kcal per day. Each dietary component is given a score of 1.

^† 300 EU are equivalent to 30 minutes of moderate physical activity 5 times per week.

A diet history was obtained at years 0, 7, and 20 through a trained interviewer‐administered questionnaire. ²⁷ , ²⁸ Physical activity was assessed using the CARDIA Physical Activity History questionnaire, which determines the frequency of participating in moderate to vigorous intensity physical activity over the past 12 months and calculated as exercise units. ²⁹ Smoking status was assessed through a self‐reported responses to a tobacco use survey. ³⁰ Former smokers were categorized based on abstinence time <1 or >1 year. Due to missing year 7 abstinence information, participants were assumed to have abstinence time <1 year if they were self‐reported as current smokers or the data were missing at the prior visit.

Each participant's height (nearest 0.5 cm) and weight (nearest 0.2 kg) measured at each examination was used to calculate body mass index (BMI), or weight (in kilograms) divided by height squared (meters squared). Blood pressure (in millimeters of mercury) was measured after 5 minutes of sitting using a random‐zero sphygmomanometer (W.A. Baum, Copiague, NY) at the years 0, 2, 5, 7, 10, and 15 visits and an Omron digital BP monitor (Omron model HEM907XL; Omron Healthcare, Lake Forest, IL) at the year 20 visit. The average of the second and third blood pressure readings were used for calculation of systolic and diastolic blood pressure. Fasting blood draws were obtained according to standard protocol. ²⁴ Serum total cholesterol (milligrams per deciliter) was measured by the enzymatic methods using the ABA Bichromatic instrument. ³¹ Fasting blood glucose levels (milligrams per deciliter) were acquired from venous blood draws and measured as described elsewhere. ³²

The primary outcome was the prevalence of computed tomography‐defined hepatic steatosis at year 25 or in midlife. Details of the CARDIA year 25 liver computed tomography exam and measurement of liver attenuation (LA) are detailed elsewhere. ³³ In this study, hepatic steatosis is defined as computed tomography‐derived LA cutoff of ≤51 Hounsfield units (HU). Low levels of LA are equivalent to high levels of liver fat. For example, a liver‐to‐spleen ratio <1.0, which is often used in clinical practice for a diagnosis of liver fat, is comparable to an LA cutoff of ≤51 HU. ³⁴ , ³⁵ , ³⁶ A lower LA cutoff of ≤40 HU was used to define moderate–severe hepatic steatosis. This lower LA cutoff, which correlates with liver‐to‐spleen ratio <0.8, is most specific for diagnosis of moderate–severe liver fat compared with the gold standard of liver biopsy but can result in exclusion of individuals with mild liver fat in epidemiologic studies. ³⁷ , ³⁸

There are multiple causes of hepatic steatosis (eg, heavy alcohol use, certain medications, hepatitis C, HIV, metabolic dysfunction). The nomenclature and diagnostic criteria for defining metabolic dysfunction‐associated hepatic steatosis changed during the conduct of the current study. NAFLD previously was a diagnosis of exclusion. In contrast, MASLD is a positive diagnosis that includes the presence of ≥1 metabolic risk factors in addition to exclusion of alternate causes of hepatic steatosis. ³⁹ For primary analysis, the new diagnostic criteria for MASLD were applied (Table 2). In sensitivity analysis, the historic criteria for NAFLD were used.

Table 2.

New Diagnostic Criteria for Metabolic Dysfunction‐Associated Steatotic Liver Disease

Presence of steatotic liver disease*
At least 1 out of 5
1.	BMI ≥25 kg/m2 or waist circumference >94 cm (for men) and >80 cm (for women)
2.	Fasting serum glucose ≥5.6 mmol/L (100 mg/dL) or 2‐hour postload glucose levels ≥7.8 mmol/L (≥140 mg/dL) or hemoglobin A1c ≥5.7% (39 mmol/L) or type 2 diabetes or treatment for type 2 diabetes
3.	Blood pressure ≥130/85 mm Hg or on antihypertensive drug treatment
4.	Plasma triglycerides ≥1.70 mmol/L (150 mg/dL) or lipid‐lowering treatment
5.	Plasma HDL cholesterol ≤1.0 mmol/L (40 mg/dL) (for men) and ≤1.3 mmol/L (50 mg/dL) (for women) or lipid‐lowering treatment

Exclusions for other causes of liver fat (medications, excessive alcohol use, HIV, viral hepatitis). BMI indicates body mass index; and HDL, high‐density lipoprotein.

^*Defined in the current study by computed tomography liver attenuation ≤51 Hounsfield units.

All covariates were measured at baseline and at each follow‐up exam. Demographic variables included age (years), sex (men or women), race (non‐Hispanic White or non‐Hispanic Black), and maximum educational attainment at year 7 (years). Other covariates include field center, cumulative drinking years across all exam years, year 25 BMI, year 25 menopause status, ⁴⁰ baseline CVH scores, and year 20 CVH scores. Difficulty paying for basics (eg, ability to pay for food, heating, medical care) was assessed at the baseline exam year as a marker for socioeconomic status or financial burden, as previously published. ⁴¹ , ⁴²

Statistical Analysis

The participant characteristics at baseline were summarized by MASLD status in midlife and by CVH trajectory groups. Continuous variables were reported as mean±SD, and categorical variables were reported as counts percent. For the primary hypotheses, binary logistic regression was used to examine the association of baseline CVH scores, with the MASLD prevalence in midlife. Odds ratios (ORs) and 95% CIs were calculated to quantify the odds of participants having MASLD in midlife for a 1‐SD unit change in CVH scores at baseline and for having low versus high CVH and moderate versus high CVH at baseline. Covariates were sequentially added to each model. Model 1 adjusted for age, sex, race, maximum educational attainment at year 7, and field center at baseline. Model 2 further adjusted for cumulative drinking years across all exam years. Model 3 (fully adjusted model) is Model 2 further adjusted for year 25 BMI. Model 4 is Model 3 further adjusted for baseline CVH scores. Model 5 is Model 3 further adjusted for year 20 CVH scores.

To examine whether the changes in CVH scores from young adulthood to late adulthood (year 0–year 20) were associated with prevalent MASLD in midlife, latent class modeling was used to evaluate the risk differences across participants in different trajectories of CVH scores. The best model was identified using Bayesian information criteria. Participants with at least 3 measurements of CVH were included in the latent trajectory modeling to identify groups of participants who share similar CVH trajectories. Diet data from year 0 were carried forward to determine the diet component of CVH scores for years 2 and 5, whereas year 7 data were carried forward for years 10 and 15. Sensitivity analysis was performed using the MASLD criteria with the lower LA cutoff of ≤40 HU, which reliably represent >30% of liver fat content and correlate with liver‐to‐spleen ratio <0.8, for the diagnosis of moderate–severe liver fat. ³⁷ , ³⁸ Historically, steatosis is graded by the pathologist based on percentage of its severity (eg, grade 0=≤5%; grade 1 (mild)=5%–33%; grade 2 (moderate)=34%–66%; and grade 3 (severe)=≥67% of hepatocytes affected by liver fat). ⁴³ Sensitivity analysis was performed for the historic criteria for NAFLD using the primary LA cutoff of ≤51 HU for any steatosis. Another sensitivity analysis was performed to further adjust for the year 25 menopause status in analyses performed with the new inclusion definition of MASLD.

A 2‐sided P value of <0.05 were considered statistically significant. All analyses were performed using SAS (version 9.4; SAS Institute, Cary, NC). One author had full access to all the data in the study and takes responsibility for its integrity and the data analysis.

RESULTS

Of the 2529 CARDIA participants included in the study (58.4% women, 47.6% Black race), 39% (n=989) had high, 56% (n=1413) had moderate, and 5% (n=127) had low CVH, respectively, at baseline or young adulthood (mean age 25 years; Table 3). At year 25, or in midlife (mean±SD age was 50±3.6 years), 23.2% (n=594) of participants had MASLD (Table 3). Year 25 MASLD prevalence was highest in participants with moderate CVH scores at baseline and low CVH scores at year 20 exam (Figure 2). The participants with MASLD in midlife were mostly men (54.3%), non‐Hispanic White (55.0%), and a current alcohol drinker (57.6%) in young adulthood. Of 1476 female participants, the majority had undergone menopause at year 25 (n=1105, 74.9%), and the majority of female participants with MASLD in midlife had undergone menopause at year 25 (n=215, 80.2%). Among female participants who had undergone menopause, participants who had low CVH scores at baseline had the highest percentage of participants with MASLD in midlife (n=28, 40.0%).

Table 3.

Participant Characteristics at Baseline (1985–1986) by Year 25 MASLD Status Using the New Diagnostic Criteria for MASLD (N=2529)

Baseline participant characteristics	Year 25 MASLD* status
	No MASLD	MASLD
No. (%)	1942 (76.8%)	587 (23.2%)
CVH score at baseline, mean±SD	10.9±1.8	10.3±1.9
Low, n (%)	81 (4.2%)	46 (7.8%)
Moderate, n (%)	1039 (53.5%)	374 (63.7%)
High, n (%)	822 (42.3%)	167 (28.4%)
Demographic variables
Age, y, mean±SD	24.9±3.6	25.3±3.6
Sex, n (%)
Women	1208 (62.2%)	268 (45.7%)
Men	734 (37.8%)	319 (54.3%)
Race, n (%)
Non‐Hispanic Black	940 (48.4%)	264 (45.0%)
Non‐Hispanic White	1002 (51.6%)	323 (55.0%)
Highest level of education, y, mean±SD	16.0±3.7	15.6±2.5
Field centers at baseline, n (%)
Birmingham, Alabama	509 (26.2%)	100 (17.0%)
Chicago, Illinoi	413 (21.3%)	138 (23.5%)
Minneapolis, Minnesota	446 (23.0%)	219 (37.3%)
Oakland, California	574 (29.6%)	130 (22.1%)
Individual‐level characteristics
Body mass index, kg/m2, mean±SD	23.8±4.4	26.8±5.3
Systolic blood pressure, mm Hg, mean±SD	108.9±10.4	112.7±10.7
Diastolic blood pressure, mm Hg, mean±SD	68.0±9.1	69.6±10.0
Total cholesterol, mg/dL, mean±SD	176.4±32.8	179.0±32.8
HDL cholesterol, mg/dL, mean±SD	53.8±12.2	49.4±12.4
LDL cholesterol, mg/dL, mean±SD	109.2±30.4	112.8±30.7
Serum glucose, mg/dL, mean±SD	81.3±10.8	84.0±12.8
Physical activity intensity score, mean±SD	399.8±288.4	429.2±305.9
Healthy diet score, 0–5, mean±SD	2.2±1.1	2.1±1.0
Current alcohol drinker, n (%)	1097 (56.8%)	337 (57.6%)
Current smoking, n (%)	412 (21.2%)	156 (26.6%)
Former smoking, n (%)	247 (12.7%)	78 (13.3%)
Hypertension treatment, n (%)	34 (1.8%)	18 (3.1%)
Diabetes treatment, n (%)	2 (0.1%)	1 (0.2%)

CVH indicates cardiovascular health; HDL, high‐density lipoprotein; LDL, low‐density lipoprotein; and MASLD, metabolic dysfunction‐associated steatotic liver disease.

^*Computed tomography‐defined MASLD status based on new MASLD criteria and defined as liver attenuation cutoff of ≤51 Hounsfield units.

Figure 2. Year 25 MASLD prevalence (2010–2011) with CVH scores at baseline (1985–1986), CVH trajectory groups, and year 20 (2005–2006) CVH scores.

Each of the 7 CVH metrics (diet, physical activity, smoking, body mass index, blood pressure, total cholesterol, and glucose) was scored as ideal (2 points), intermediate (1 point), or poor (0 points) and summed for a range of CVH scores (0–14). The summed CVH scores were categorized into either high (12–14 points, reference), moderate (8–11 points), or low (0–7 points) CVH. Participants with at least 3 measurements of CVH were included in the study. Diet data from year 0 were carried forward to determine the diet component of CVH scores for year 2 and 5, whereas year 7 data were carried forward for year 10 and year 15. The best model was identified using Bayesian information criteria. All 5 groups had average posterior probabilities >0.75. Computed tomography‐defined MASLD status based on new MASLD criteria and as liver attenuation cutoff of ≤51 Hounsfield units and reported as n (%). A, Year 25 MASLD prevalence with CVH scores at baseline or year 0. B, Year 25 MASLD prevalence with CVH trajectories from year 0 to year 20. C, Year 25 MASLD prevalence with year 20 CVH scores. CVH indicates cardiovascular health; and MASLD, metabolic dysfunction‐associated steatotic liver disease.

After adjustment for demographic variables and field centers, participants with low CVH scores in young adulthood were (Model 1: OR, 3.44 [95% CI, 2.24–5.28]) more likely to have prevalent MASLD in midlife compared with participants with high CVH scores (Table 4). These findings remained significant after further adjustment for cumulative drinking years across all exam years (Model 2: OR, 3.38 [95% CI, 2.20–5.19]). Similarly, participants with moderate CVH scores in young adulthood (Model 1: OR, 1.91 [95% CI, 1.53–2.37]; Model 2: OR, 1.91 [95% CI, 1.54–2.38]) were more likely to have prevalent MASLD in midlife compared with participants with high CVH scores (Table 4). However, the association of baseline CVH scores and year 25 MASLD prevalence did not remain statistically significant after further adjustment for year 25 BMI in Model 3 (Table 4). Every 1‐SD unit increase in CVH scores at baseline was associated with lower odds of prevalent MASLD in midlife (Model 1: OR, 0.80 [95% CI, 0.76–0.85]; Model 2: OR, 0.80 [95% CI, 0.76–0.85]), but these findings were not statistically significant after further adjustment for year 25 BMI in Model 3 (Table 4). Similar findings were observed in analyses performed with the historic criteria for NAFLD (Table S1) and analyses that further adjusted for the menopause status (Table S2).

Table 4.

Adjusted Odds Ratios for Associations of Baseline Cardiovascular Health Scores (1985–1986) With Y 25 Metabolic Dysfunction‐Associated Steatotic Liver Disease Prevalence (2010–2011)

CVH* scores at baseline	Participants with MASLD†, n (%)	Odds ratio (95% CI) of prevalent MASLD† at y 25
		Model 1‡	Model 2§	Model 3‖
CVH score (continuous)	587 (23.3)	0.80 (0.76–0.85)¶	0.80 (0.76–0.85)¶	0.95 (0.90–1.02)
High CVH (reference)	167 (28.8)	1 [reference]	1 [reference]	1 [reference]
Moderate CVH	374 (63.5)	1.91 (1.53–2.37)¶	1.91 (1.54–2.38)¶	1.18 (0.93–1.51)
Low CVH	46 (7.7)	3.44 (2.24–5.28)¶	3.38 (2.20–5.19)¶	1.00 (0.60–1.66)

CVH indicates cardiovascular health; and MASLD, metabolic dysfunction‐associated steatotic liver disease.

^*Each of the 7 CVH metrics (diet, physical activity, smoking, body mass index, blood pressure, total cholesterol, and glucose) was scored as ideal (2 points), intermediate (1 point), or poor (0 points) and summed for a range of CVH scores (0–14). The summed CVH scores were categorized into either high (12–14 points, reference), moderate (8–11 points), or low (0–7 points) CVH.

^† Computed tomography‐defined MASLD status based on new MASLD criteria and defined as liver attenuation cutoff of ≤51 Hounsfield units.

^‡ Model 1: Dichotomous logistic regression models adjusted for age, sex, race, field center, and maximum educational attainment at y 7 (y).

^§ Model 2: Model 1 additionally adjusted for cumulative drinking y across all exam y.

^‖ Model 3: Model 2 additionally adjusted for body mass index at y 25.

^¶ Statistical significance (P < 0.05).

Among 2529 CARDIA participants with CVH measurements at 3 or more examinations between year 0 and year 20, 5 distinct CVH trajectories were identified (Figure 3). All 5 groups had average posterior probabilities >0.75; 24.8% (n=627) of included participants maintained high CVH scores throughout (high‐stable group); 33.7% (n=852) started at high CVH scores and decreased (high‐decreasing group); 13.0% (n=328) maintained moderate CVH scores throughout (moderate‐stable group); 23.4% (n=592) started at moderate CVH scores and decreased (moderate‐decreasing group); and 5.1% (n=130) started at low CVH scores and decreased (low‐decreasing group). Most CVH trajectory groups showed a decrease in CVH scores over the 20‐year follow‐up period, with the moderate‐decreasing group undergoing the largest decrease (Figure 3). However, moderate‐stable and high‐stable groups maintained relatively the same CVH throughout follow‐up (Figure 3 and Table S3). The high‐stable group was mostly White participants and had less difficulty paying for basics, had lower BMI, and drank alcohol at baseline compared with other CVH trajectory groups (Table S3). Worsening of each CVH component at baseline was observed when moving from higher to low CVH trajectory groups (Table S3).

Figure 3. Trajectories of CVH score from year 0 (1985–1986) to year 20 (2005–2006) among participants in the Coronary Artery Risk Development in Young Adults study (total N=2529).

Each of the 7 CVH metrics (diet, physical activity, smoking, body mass index, blood pressure, total cholesterol, and glucose) was scored as ideal (2 points), intermediate (1 point), or poor (0 points) and summed for a range of CVH scores (0–14). The summed CVH scores were categorized into either high (12–14 points; reference), moderate (8–11 points), or low (0–7 points) CVH. Participants with at least 3 measurements of CVH were included in the study. Diet data from year 0 were carried forward to determine the diet component of CVH scores for year 2 and year 5, whereas year 7 data were carried forward for year 10 and year 15. The best model was identified using Bayesian information criteria. All 5 groups had average posterior probabilities >0.75. CVH indicates cardiovascular health.

Among the 5 CVH trajectories identified, the low‐decreasing group had the highest percentage of participants with MASLD (43.9%, n=57) in midlife (Figure 2). Between the 2 groups that started at similar CVH scores (high‐stable versus high‐decreasing, moderate‐stable versus moderate‐decreasing), the high‐decreasing group (23.0% versus 7.0%) and moderate‐decreasing group (35.7% versus 24.4%) had a higher prevalence of MASLD in midlife compared with high‐stable and moderate‐stable groups, respectively (Figure 2). Although the percentage of female participants with positive menopause status was lowest in the low‐decreasing CVH trajectory group (n=63, 5.7%), the MASLD prevalence among female participants with positive menopause status was highest in the low‐decreasing CVH trajectory group (44.4%, n=28).

Apart from the trajectory data, there were 331 participants (≈13% of participants) whose CVH scores were higher at the year 20 follow‐up exam relative to the baseline exam, whereas the majority (87%) of participants had lower CVH scores at the year 20 follow‐up. Approximately 65% of those 13% participants whose CVH was higher at year 20 relative to the baseline exam were categorized into being in the moderate‐ or high‐stable groups. MASLD prevalence in midlife was 12.1% in those whom CVH was higher at year 20 compared with those who had lower year 20 CVH (27%). This finding is consistent with the primary results.

In Model 1, the adjusted odds ratios for prevalent MASLD in midlife ranged from 4.84 (95% CI, 3.4–6.9) for the high‐decreasing group to 17.8 (95% CI, 10.8–27.4) for the low‐decreasing group in comparison with the high‐stable (reference) group. These findings remained statistically significant in Model 2, which further adjusted for cumulative drinking years, and Model 3, a fully adjusted model which further adjusted for year 25 BMI (Table 5). Although the adjusted odds ratios were much lower in Model 3, the findings remained statistically significant. In Model 4, a fully adjusted model additionally adjusted for the baseline CVH scores, the adjusted odds ratios for prevalent MASLD in midlife remained statistically significant and ranged from 2.98 (95% CI, 2.0–4.4) for the high‐decreasing group to 8.94 (95% CI, 4.2–19.2) for the low‐decreasing group in comparison with the reference group (Table 6). In Model 5, a fully adjusted model that additionally adjusted for year 20 CVH scores, the findings were similar but attenuated compared with Model 4 and remained statistically significant only for the high‐decreasing group (OR, 1.63 [95% CI, 1.04–2.6]) and moderate‐decreasing group (OR, 1.93 [95% CI, 1.1–3.5]) in comparison with the reference group (Table 6). Similar findings were observed in analyses performed with the historic criteria for NAFLD for all models (Tables S4 and S5). In a sensitivity analysis further adjusting for the menopause status, similar findings were observed, but the adjusted odds ratios were slightly higher for moderate‐decreasing, low‐decreasing, and high‐decreasing groups for Model 1, 2, 3, and 4 (Tables S6 and S7). Although results were similar in Model 5, the adjusted odds ratio was no longer significant in the high‐decreasing group after further adjustment for menopause status (Table S7).

Table 5.

Y 25 MASLD Prevalence (2010–2011) and Adjusted OR for Associations of CVH Score Trajectory Groups With Y 25 MASLD Prevalence (2010–2011)

CVH trajectory group*	Participants with MASLD† at y 25, n (%)	Model 1‡	Model 2§	Model 3‖
		OR (95% CI)	OR (95% CI)	OR (95% CI)
High‐stable	44 (7.0)	1 [reference]	1 [reference]	1 [reference]
High‐decreasing	195 (23.0)	4.84 (3.4–6.9)¶	4.78 (3.3–6.8)¶	2.53 (1.7–3.7)¶
Moderate‐stable	80 (24.4)	5.49 (3.6–8.3)¶	5.39 (3.6–8.2)¶	2.22 (1.4–3.5)¶
Moderate‐decreasing	211 (35.7)	12.44 (8.5–18.3)¶	12.29 (8.4–18.0)¶	4.0 (2.6–6.1)¶
Low‐decreasing	57 (43.9)	17.8 (10.8–29.4)¶	17.28 (10.4–28.6)¶	4.15 (2.4–7.3)¶

CVH indicates cardiovascular health; MASLD, metabolic dysfunction‐associated steatotic liver disease; and OR, odds ratio.

^*Each of the 7 CVH metrics (diet, physical activity, smoking, body mass index, blood pressure, total cholesterol, and glucose) was scored as ideal (2 points), intermediate (1 point), or poor (0 points) and summed for a range of CVH scores (0–14). The summed CVH scores were categorized into either high (12–14 points, reference), moderate (8–11 points), or low (0–7 points) CVH. Participants with at least 3 measurements of CVH were included in the study. Diet data from y 0 were carried forward to determine diet component of CVH scores for y 2 and y 5, whereas y 7 data were carried forward for y 10 and y 15. The best model was identified using Bayesian information criteria. All 5 groups had average posterior probabilities >0.75.

^† Computed tomography‐defined MASLD status based on new MASLD criteria and as liver attenuation cutoff of ≤51 Hounsfield units and reported as n (%).

^‡ Model 1: Dichotomous logistic regression models adjusted for age, sex, race, field center, and maximum educational attainment at y 7 (y).

^§ Model 2: Model 1 additionally adjusted for cumulative drinking y across all exam y.

^‖ Model 3: Model 2 additionally adjusted for body mass index at y 25.

^¶ Statistical significance (P<0.05).

Table 6.

Y 25 MASLD Prevalence (2010–2011) and Adjusted OR for Associations of CVH Score Trajectory Groups With Y 25 MASLD Prevalence in Fully Adjusted Models, Additionally Adjusted for Either Baseline or Y 20 CVH Scores

CVH trajectory group*	Participants with MASLD† at y 25, n (%)	Model 4‡	Model 5§
		OR (95% CI)	OR (95% CI)
High‐stable	44 (7.0)	1 [reference]	1 [reference]
High‐decreasing	195 (23.0)	2.98 (2.01–4.44)‖	1.63 (1.04–2.55)‖
Moderate‐stable	80 (24.4)	3.48 (2.02–6.00)‖	1.54 (0.90–2.61)
Moderate‐decreasing	211 (35.7)	5.98 (3.62–9.86)‖	1.93 (1.07–3.47)‖
Low‐decreasing	57 (43.9)	8.94 (4.16–19.22)‖	1.72 (0.77–3.8)

CVH indicates cardiovascular health; MASLD, metabolic dysfunction‐associated steatotic liver disease; and OR, odds ratio.

^*Each of the 7 CVH metrics (diet, physical activity, smoking, body mass index, blood pressure, total cholesterol, and glucose) was scored as ideal (2 points), intermediate (1 point), or poor (0 points) and summed for a range of CVH scores (0–14). The summed CVH scores were categorized into either high (12–14 points, reference), moderate (8–11 points), or low (0–7 points) CVH. Participants with at least 3 measurements of CVH were included in the study. Diet data from y 0 were carried forward to determine diet component of CVH scores for y 2 and y 5, whereas y 7 data were carried forward for y 10 and y 15. The best model was identified using Bayesian information criteria. All 5 groups had average posterior probabilities >0.75.

^† Computed tomography‐defined MASLD status based on new MASLD criteria and as liver attenuation cutoff of ≤51 Hounsfield units and reported as n (%).

^‡ Model 4: Dichotomous logistic regression models with full adjustment (age, sex, race, maximum educational attainment at y 7 [y], field center, body mass index at y 25, cumulative drinking y across all exam y), additionally adjusted for baseline CVH scores.

^§ Model 5: Dichotomous logistic regression models with full adjustment, additionally adjusted for y 20 CVH scores.

^‖ Statistical significance (P<0.05).

In a sensitivity analysis performed using the lower LA cutoff of ≤40 HU, 9.9% (n=251) of participants had prevalent moderate–severe hepatic steatosis in midlife. Between the CVH trajectory groups that started at similar CVH scores (high‐stable versus high‐decreasing, moderate‐stable versus moderate‐decreasing), the high‐decreasing group (10.1% versus 2.6%) and moderate‐decreasing group (15.9% versus 10.1%) had a higher prevalence of moderate–severe hepatic steatosis compared with the high‐stable and moderate‐stable groups, respectively, in midlife (Table S8). The adjusted ORs for moderate–severe hepatic steatosis in midlife ranged from 5.14 (95% CI, 3.0–8.9) for the high‐decreasing group to 12.13 (95% CI, 6.0–24.5) for the low‐decreasing group in comparison with the reference group in Model 1 (Table S8). These findings remained statistically significant in Model 2 and Model 3 (Table S8).

DISCUSSION

In a large, prospective, biracial community‐based study of adults followed for 25 years in the CARDIA study, we have demonstrated that the majority of individuals failed to maintain high CVH scores over time, and that failure to maintain high CVH is associated with greater odds of midlife MASLD. We identified 5 distinct trajectories of CVH score from young adulthood to midlife in the CARDIA study. Most importantly, our study highlights that among those who started with similar CVH scores, those who had declining CVH over time had higher prevalence of MASLD in midlife. This was also apparent in a sensitivity analysis with lower LA cutoff for moderate–severe hepatic steatosis. Hence, understanding CVH trajectory patterns is critical in determining risk for prevalent MASLD rather than assessing CVH at 1 time point in young adulthood or in midlife. Similar to findings from prior cross‐sectional studies, we have demonstrated that individuals with lower CVH scores in young adulthood are more likely to have higher odds of prevalent MASLD in midlife compared with those with higher CVH scores. However, the association between MASLD prevalence in midlife and CVH scores in young adulthood was no longer significant after adjusting for BMI in midlife, suggesting that assessing overall CVH at 1 time point in young adulthood may not have meaningful value in comparison with assessing one's overall CVH throughout their lifetime, especially in those with high or unhealthy BMI in midlife. Furthermore, having high or low CVH scores early in life may not necessarily mean that individuals will continue their health promoting or degrading behaviors throughout their adulthood. Hence, our findings highlight the importance of achieving high CVH scores in early adulthood and maintaining these levels throughout adulthood for the primary prevention of MASLD.

The findings from this study provide a unique insight into long‐term patterns of CVH change and highlight that within the CARDIA population there are heterogenous patterns in CVH trajectories. Furthermore, latent class modeling has allowed identification of different patterns of CVH change as separate trajectory groups, thus providing understanding of lifetime trends in CVH and how they associate with MASLD. Only 25% of CARDIA participants were able to achieve and maintain high CVH scores throughout the 20‐year follow‐up period, with a decrease in CVH seen in most trajectory groups over time. As expected, the individuals who started with low CVH and had decreasing CVH over time had the highest prevalence of MASLD in midlife and were 4 times more likely to have prevalent MASLD compared with those who had high CVH scores all throughout their young adulthood into midlife. In Model 2, which did not adjust for year 25 BMI, these individuals were 17 times more likely to have prevalent MASLD compared with those who had high CVH scores starting in young adulthood into midlife, highlighting the impact of obesity. Furthermore, all CVH trajectory groups were significantly associated with MASLD prevalence in midlife independent of baseline CVH. The associations between CVH trajectories and year 25 MASLD prevalence were stronger after further adjustment for baseline CVH scores, because the individuals in the low‐decreasing group were 9 times more likely to have prevalent MASLD compared with those in the high‐stable group. This suggests that individuals who start out at lower CVH scores in young adulthood can still lower their odds of having MASLD later in life if they maintain their baseline CVH scores without further decline. Hence, maintaining CVH throughout one's lifetime is as important as achieving high CVH in young adulthood for MASLD prevention. Although the association of the CVH trajectory groups and MASLD prevalence in midlife was attenuated after adjusting for CVH scores in midlife, the findings were similar and remained significant for high‐decreasing and moderate‐decreasing trajectory groups in comparison with the high‐stable group. Thus, our study highlights that understanding lifetime trends in CVH is as important as assessing one's concurrent CVH profile in midlife for the primary prevention of steatotic liver disease. Furthermore, these findings persisted in analyses that used the historic criteria for NAFLD. This unique relationship observed between the CVH trajectories and MASLD in our study may be able to inform public health interventions targeting young adults.

Few studies to date, mostly cross‐sectional, have investigated the relationship between overall CVH and MASLD. Large cross‐sectional studies in China ²⁰ and in the Multiethnic Study of Atherosclerosis cohort, ²¹ Framingham Heart Study, ²² and National Health and Nutrition Examination Survey ²³ showed that the more favorable CVH profiles are associated with a lower MASLD prevalence. A prospective cohort study in China with up to a 5‐year follow‐up demonstrated that the risk of developing MASLD decreased in a dose–response manner with the increase in the number of ideal CVH metrics. ⁴⁴ Furthermore, improvement in CVH scores during the follow‐up period reduced the risk of developing MASLD by 71% compared with worsening CVH, ⁴⁴ which is consistent with our study showing the lowest MASLD prevalence in individuals who maintained high and stable CVH trajectory. However, studies in China did not include or measure all CVH metrics according to the AHA guideline. A cohort study of 93 500 Korean participants with a median follow‐up period of 3.7 years (interquartile range of 2.0–4.8 years) reported that higher CVH scores were associated with regression of existing MASLD and lower incidence of MASLD diagnosed using abdominal ultrasound. ⁴⁵ A longitudinal analysis of a middle‐aged Korean population with a follow‐up period of 5 years showed that individuals with poor CVH scores had a higher odds of developing new‐onset MASLD than those with intermediate CVH. ⁴⁶ However, the participants were already in their midlife at the baseline examination, and different criteria were used for CVH metrics to adjust for the Korean population. The recent CARDIA study explored the relationship between intensity‐based physical activity trajectories from young to middle adulthood and MASLD prevalence in midlife. ⁴⁷ In this study, individuals with higher levels of vigorous‐intensity physical activity throughout adulthood had a lower odds of MASLD in midlife compared with those with lower levels of vigorous‐intensity physical activity, highlighting the importance of investigating long‐term patterns of different factors that contribute to one's overall health. ⁴⁷ Although these studies suggest the close relationship between CVH and MASLD, most studies were cross‐sectional and if prospective, only followed the cohort for a short period of time. Also, most studies did not account for changes in CVH over time and used the old NAFLD criteria for outcome. Hence, our study is the first to explore the association of CVH and MASLD longitudinally using CVH trajectories starting in young adulthood.

Studying the relative contribution of individual metrics of CVH throughout young adulthood on MASLD prevalence is an area that requires more study, because it may provide more insight into which CVH metrics are driving the association with liver disease. However, investigating the trajectories of composite CVH scores is more compelling than that of individual CVH metrics, because it still has a meaningful value even if all metrics of CVH may not meet the ideal definition of LS7. In addition to studies showing that achieving a greater number of ideal CVH metrics is associated with a lower risk of developing MASLD, a higher LS7 CVH score is also associated with lower incident CVD, ⁴⁸ , ⁴⁹ , ⁵⁰ coronary heart disease, ⁵¹ stroke, ⁵¹ , ⁵² heart failure, ⁵³ , ⁵⁴ , ⁵⁵ atrial fibrillation, ⁵⁵ , ⁵⁶ and peripheral arterial disease, ⁵⁷ highlighting the value of composite CVH scores. Furthermore, some studies have assessed the association of individual metrics and composite CVH scores with liver diseases. For example, in a study that used controlled attenuation parameter and liver stiffness measurement to define MASLD, advanced liver fibrosis, and cirrhosis, individuals with poor BMI, poor blood pressure, poor total cholesterol, and poor hemoglobin A1c had a higher controlled attenuation parameter in comparison with individuals in ideal categories, but no significant associations were observed between controlled attenuation parameter and diet quality, physical activity, and smoking status. ⁵⁸ Similar findings were seen in the association between liver stiffness measurement and each component of the CVH metric, but liver stiffness measurement was higher in those with poor physical activity and lower in those with poor total cholesterol. ⁵⁸ Although this study was helpful in delineating which CVH metric may be driving the association with fatty liver disease, it also showed that CVH score <8 is a key threshold for increased risk of MASLD, advanced liver fibrosis, and cirrhosis. ⁵⁸ Also, this study showed that the risk of MASLD, advanced liver fibrosis, and cirrhosis were 7, 10, and 6 times higher, respectively, in the low CVH group than in the high CVH group. ⁵⁸ Furthermore, another study using AHA's Life's Essential 8 score showed that blood glucose and BMI contributed the most to the risk of MASLD, while still highlighting that there is a nonlinear dose–response relationship between Life's Essential 8 score and MASLD. ⁵⁹ Collectively, these studies suggest that the composite CVH score offers a more robust and comprehensive assessment in risk prediction for MASLD than individual components of the score. Lastly, in clinical practice, a global focus on cardiovascular health that increases awareness of all metrics that contribute to MASLD risk through collective interventions such as global lifestyle interventions may be more effective for MASLD risk reduction than a focus only on individual metrics. ⁶⁰ , ⁶¹

Our study accounted for the potential confounding factors, including cumulative drinking years and BMI in midlife, to better delineate the role of CVH from other factors that may be associated with prevalent MASLD. Multiple studies have suggested the independent role of alcohol consumption and obesity as well as their synergistic negative effects on liver disease. ⁶² , ⁶³ , ⁶⁴ , ⁶⁵ , ⁶⁶ , ⁶⁷ For instance, a recent study suggested that frequent episodic drinking in young adulthood is associated with higher rates of alcohol dependence later in life, highlighting the potential role of cumulative or chronic alcohol consumption. ⁶⁸ The new nomenclature recognizes this important overlap proposing that a term such as metabolic and alcohol‐associated liver disease (MASLD and at risk alcohol use defined as weekly alcohol intake 140–350 g in women and 210–420 g in men) be used to describe these patients. ³⁹ A recent meta‐analysis suggests that although obesity is associated with a worse long‐term prognosis in patients with MASLD, obesity alone does not account for development of metabolic dysfunction‐associated steatohepatitis or advanced fibrosis in MASLD and should not be used as a solitary criterion for MASLD screening. ⁶⁹ This is reinforced by another meta‐analysis, which highlighted that ≈40% of the global MASLD population was nonobese and ≈20% were lean, but both groups had consequential long‐term liver and nonliver comorbidities. ⁷⁰ Although both obesity and chronic alcohol use merit attention, our findings continue to show strong association between individual CVH trajectories and MASLD prevalence in midlife even after further adjustment for cumulative drinking and BMI in midlife. Thus, CVH score has meaningful and statistically significant associations with MASLD risk.

Interestingly, MASLD was more prevalent among men in our study, but majority of women with MASLD were postmenopausal, with MASLD prevalence being highest in the low‐decreasing CVH trajectory group. This is likely related to menopause, because it is well established that premenopausal women have lower risk of MASLD and metabolic dysfunction‐associated steatohepatitis than postmenopausal women due to the protective effects of estrogen on hepatic triglyceride handling and insulin sensitivity. ⁷¹ , ⁷² However, associations of the CVH trajectory groups with midlife MASLD mostly remained the same and significant even after further adjustment for menopause status, except for the high‐decreasing CVH trajectory group, which was no longer significant in a fully adjusted model that additionally adjusted for CVH scores in midlife. This suggests that the clinical role of menopause status may be more evident in ones who failed to maintain high CVH scores throughout life rather than ones who started with lower CVH scores earlier in life.

Limitations

MASLD was not assessed at baseline nor before year 25, so it is unknown when during follow‐up hepatic steatosis may have developed. Considering that the baseline exam occurred in 1985 to 1986 before the onset of the obesity and diabetes epidemics, we hypothesize that prevalent MASLD at baseline was also low. Our study used self‐reported alcohol use, which could potentially result in some misclassification of alcohol‐related liver disease in the MASLD group. Although self‐reported alcohol use is a standard method of assessing alcohol intake, it is likely to be underreported at high levels due to social stigma and possibly a differential dropping out of heavy alcohol users, who are less likely to show up at a study exam because of other social problems related to excessive intake. Hence, the relationship between CVH trajectories and liver diseases should be further investigated based on new metabolic dysfunction‐associated fatty liver disease and metabolic and alcohol‐associated liver disease diagnostic criteria, which accounts for metabolic abnormalities and recognized overlap between alcoholic liver disease and MASLD. The diet information was carried forward for calculation of CVH for exam year 2, 5, 10, and 15, because diet was measured only at year 0, 7, and 20. However, this is a standard procedure in other published works that assessed diet data in CARDIA longitudinally. ⁷³ Although there is the AHA's Life's Essential 8 that is comprehensive and additionally includes sleep health, the present analysis used LS7 because sleep health was not assessed in the full cohort across examinations, and the large body of research described LS7 score in relation to CVD. ¹⁰ , ¹¹ , ¹³ , ¹⁴ , ¹⁹ Although BMI is part of AHA's LS7 and is a valuable clinical screening tool and has usefulness in population‐level studies, it has limitations in predicting an individual's risk of chronic diseases and measuring body fat distribution, not accounting for more harmful abdominal adiposity. ⁷⁴ , ⁷⁵ , ⁷⁶ Multiple studies have indicated that a BMI cutoff may lead to misclassification in some individuals of different body habitus, age, sex, race, or ethnicity. ⁷⁴ , ⁷⁵ , ⁷⁶ , ⁷⁷ , ⁷⁸ , ⁷⁹ For example, a recent study demonstrated that among individuals with the same BMI, a substantial heterogeneity in body mass and body composition proportion was observed across groups differing in ages and race or ethnicity. ⁸⁰ Although there are attempts to redefine BMI threshold by sex and race and ethnicity based on association with metabolic disease, ⁷⁸ there is no consensus for BMI cutoffs in varying populations, because defining set BMI cutoffs specific to race and ethnicity solely based on adiposity would be overly complicated. ⁷⁴ Our study used the standard BMI cutoffs, which are higher than proposed cutoffs for men and White women, but are deemed sufficient, because they are viewed as a global measure of health along with other metrics in LS7 definitions. Although the CARDIA study is a multisite prospective cohort study, this study was not intended to be nationally representative or representative of the recruitment sites.

CONCLUSIONS

The CVH trajectories throughout early adulthood are associated with MASLD risk in a time‐dependent manner. Our findings highlight the importance of achieving high CVH scores starting in young adulthood and maintaining levels with aging to achieve a lower risk of prevalent MASLD in midlife. Public health interventions targeting young adults throughout their adulthood are needed to achieve this goal. Future research can further explore the relationship between the long‐term CVH maintenance and the progression of MASLD to advanced liver diseases, such as metabolic dysfunction‐associated steatohepatitis and cirrhosis, and their complications.

Sources of Funding

The CARDIA study is conducted and supported by the National Heart, Lung, and Blood Institute in collaboration with the University of Alabama at Birmingham (75N92023D00002 and 75N92023D00005), Northwestern University (75N92023D00004), University of Minnesota (75N92023D00006), and Kaiser Foundation Research Institute (75N92023D00003). L.B.V.W. received support from the National Clinical and Translational Sciences Institute (KL2TR001424).

Disclosures

L.B.V.W. consults for Madrigal Pharmaceuticals, Gerson Lehrman Group, Slingshot Insights and Alpha Sights, receives research support from W.L. Gore & Associates and Madrigal Pharmaceuticals, and provides medicolegal services outside of the submitted work. The remaining authors have no disclosures to report.

Supporting information

Tables S1–S8