Associations between MASLD, atrial fibrillation, cardiovascular events, mortality and aspirin use in older adults

Abstract

The impact of metabolic dysfunction-associated steatotic liver disease (MASLD), the preferred nomenclature for NAFLD, on cardiovascular health and mortality among older adults is uncertain. As such, we aimed to identify whether MASLD increases the risk of Major Adverse Cardiovascular Events (MACE) (a composite of fatal coronary heart disease [excluding heart failure], nonfatal myocardial infarction, or fatal or nonfatal ischemic stroke), Atrial Fibrillation (AF), or all-cause mortality in older adults, and whether aspirin attenuates these risks in individuals with MASLD. This is a non-prespecified post-hoc analysis of the ASPREE (ASPirin in Reducing Events in the Elderly) randomized trial. Participants were community dwelling well adults aged ≥ 70 years without a history of atherosclerotic cardiovascular disease or AF. Fatty Liver Index (FLI) was used to identify MASLD at baseline. FLI is a composite of anthropometric and biochemical markers used in epidemiologic studies to rule in and rule out hepatic steatosis. MACE and cause of death were adjudicated by clinical experts; AF was assessed by previously defined algorithm in ASPREE. 9,097 participants were stratified into groups according to FLI. In univariate analysis, prevalent MASLD (FLI ≥ 60 with evidence of metabolic dysfunction; n = 2,998 [33.0%]) was associated with an increased risk of MACE (HR 1.47 [95% CI 1.22–1.78]) and AF (HR 1.50 [95% CI 1.19–1.88] but not all-cause mortality (HR 1.04 [95% CI 0.91–1.19]). After adjusting for cardiovascular disease risk factors, only the association between MASLD and AF remained significant (HR 1.46 [95% CI 1.11–1.93]). Aspirin did not reduce the risk of MACE, death, or AF in the MASLD group. MASLD was associated with an increased hazard of incident AF, but not of MACE or all-cause mortality, in community dwelling older adults. Primary prevention with aspirin does not ameliorate these risks in older adults with MASLD.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11357-024-01435-2.

Introduction

Steatotic liver disease (SLD) is characterized by ≥ 5% of hepatocytes containing fat irrespective of the underlying etiology [1]. Non-alcoholic fatty liver disease (NAFLD), recently re-classified by an international Delphi process to metabolic dysfunction-associated steatotic liver disease (MASLD) [2], is defined by the presence of SLD alongside at least one cardiometabolic comorbidity in the absence of excess alcohol use and other concomitant liver diseases. Previous work has linked MASLD with atherosclerotic cardiovascular disease (ASCVD) [3, 4], atrial fibrillation (AF) [5], and in some studies all-cause mortality [6], which is of particular importance given the significant and growing prevalence of MASLD in the United States [7] and globally [8].

Rising in parallel with the increasing prevalence of SLD is the aging of the global population [9]. Cardiovascular disease (CVD) remains the single most common cause of mortality globally [10], accounting for one third of all deaths [11]; age is a known important contributor to cardiovascular health and disease [12]. Similarly, the burden of AF increases with increasing age and is associated with several important diseases of aging, including heart failure and stroke, as well as overall mortality [13]. However, the link between MASLD and these clinical outcomes has been less well studied in older adults and while a strong association between SLD and CVD events has been reported [14], recent work has reported some discrepant findings. Shang et al. reported decreased life expectancy in adults with MASLD, and while statistically significant, the influence was attenuated with increasing age over 70 years [15]. Another study of hepatic steatosis in an older community cohort reported no significant association between the presence of SLD and mortality [16]. Similarly, in a recent analysis of the NHANES-III dataset, MASLD was associated with all-cause and CVD-related mortality in those aged 60–74 years but not in those over 74 years of age [17]. It is possible that these discrepancies in outcomes are due to a survival bias where older adults with SLD have a more benign course than middle-aged adults with SLD.

It is also unclear whether MASLD adds significantly to an association beyond that of the CVD risk factors that contribute to its definition and presence. It has been shown that MASLD is associated with increased rates of incident chronic kidney disease (CKD) [18] and incident type 2 diabetes mellitus (T2DM) [19], and MASLD has also been shown to have a complex bi-directional relationship with hypertension [20], all of which are known risk factors for ASCVD. It is plausible that the impact of steatotic liver disease on CVD events is at least partly mediated by these effects, and that models adjusting for them may impact the interpretation of the relationship between MASLD and important incident clinical outcomes.

Despite the growing burden of MASLD, there are no clear preventive therapies to reduce the rate of death, disability, or major adverse cardiovascular events (MACE) in adults with steatotic liver disease. Aspirin, an irreversible COX inhibitor, has an antiplatelet action via its effect on the thromboxane A₂ pathway [21]. While its use is recommended for secondary prevention of CVD [22], the use of aspirin in the primary prevention setting is generally recommended only on a case-by-case basis depending on the balance between an individual’s risk profile for MACE, and for bleeding and non-vascular mortality, especially in older adults [22, 23]. Publications describing major endpoints of the ASPREE (ASPirin in Reducing Events in the Elderly) trial, which randomized participants to low-dose daily aspirin or placebo, previously reported no significant benefit for the use of aspirin in the primary prevention setting for MACE [24] and all-cause mortality [25]. However, given the potential independent link between MASLD and MACE reported in some studies [3, 26], and given that MASLD has been associated with endothelial dysfunction [26] and increased mean platelet volume (linked to an elevated risk of CVD both generally and in MASLD [27]), evaluation of aspirin as a primary prevention strategy in this presumed higher-risk group has been called for to help guide clinical decision making [28]. As such, the present study aimed to evaluate whether MASLD confers an increased risk of MACE, incident AF, and all-cause mortality, and whether aspirin as a potential primary preventive therapy ameliorated these outcomes in ASPREE participants with MASLD.

Methods

Study population

The study design and findings from the main ASPREE trial have been previously reported in detail [24, 25, 29, 30]. We performed an analysis of Australian participants who provided serum to allow calculation of the Fatty Liver Index (FLI) [31]. Participants from the USA were unable to be analyzed as they did not provide serum for baseline steatotic liver disease case identification. Further details of the methods are provided in the supplementary file. In summary, between 2010 and 2014, ASPREE recruited 16,703 Australian participants via primary care who were aged ≥ 70 years without significant cognitive impairment, established or previous CVD events, AF, an inability or significant difficulty in independently performing any one or more of six basic activities of daily living [30], or a life expectancy of less than five years.

Participants were randomly assigned 1:1 to 100 mg of enteric coated aspirin or matched placebo stratified by age (70–79 years or ≥ 80 years) and trial center. Annual follow-up and medical record review occurred. The primary outcome of the trial was disability-free survival. Cardiovascular events were among the secondary endpoints. The initial ASPREE trial was approved by the Monash University Human Research Ethics Committee (MUHREC) (IRB00002519; ethics #2006/745MC) and other allied institution ethics committees. In Australia, the Alfred Hospital Ethics Committee (ethics #HREC/17/Alfred/198) oversees the ASPREE-XT project as the primary site approver. The ASPREE trial and ASPREE-XT cohort study are registered on ClinicalTrials.gov (NCT01038583) and the International Standard Randomised Controlled Trial Number Registry (ISRCTN83772183) and were conducted in accordance with the Declaration of Helsinki.

Participant assessment & laboratory data

At baseline and during follow-up, in-person/phone call interviews and assessments by blinded research staff collected the following data: (a) self-reported information on medical history and lifestyle (including alcohol use); (b) anthropometry and markers of physical function (including BMI, abdominal circumference, blood pressure, and heart rate and rhythm); and (c) laboratory parameters. Further details of these assessments can be seen in the supplementary methods. During follow-up, annual pill counts from ASPREE trial bottles were performed to assess for treatment adherence, with a random selection of residual tablets chemically analyzed to confirm that the contents were the trial medication.

Initial baseline data included fasting triglycerides and other lipids and fasting glucose, repeated annually at local pathology laboratories. In addition, as part of the ASPREE Healthy Ageing Biobank sub-study, 11,914 Australian participants provided blood samples for storage and later analysis. These samples were collected, where possible, pre-commencement of the trial medication (66%), and were subsequently used for biochemical analysis to determine gamma-glutamyltransferase (GGT) amongst other measures. The analysis of the Healthy Ageing Biobank serum was performed centrally at Alfred Health Pathology using an Abbott Alinity ci analyzer and Abbott reagents (Abbott Diagnostics, Macquarie Park, NSW, Australia).

Identifying MASLD

As previously described in the ASPREE cohort [32], the FLI [31] was calculated using BMI, abdominal circumference, triglycerides, and GGT in participants in whom complete data were available and who had blood collected within 90 days of their baseline visit. Possible FLI scores range from 0 – 100; a score of ≥ 60 is considered to represent hepatic steatosis [31]. Though the FLI was originally used and validated for identifying NAFLD, recent work from our group and others showed near total concordance between NAFLD and the newer MASLD definition [32, 33]. To identify a MASLD subgroup, participants with a FLI ≥ 60 were excluded if they drank excess alcohol (males more than 210 g/week, females more than 140 g/week) or were on steatogenic medications (methotrexate, amiodarone, glucocorticoids, and/or tamoxifen) [34]. Subsequently, those with a FLI ≥ 60 meeting these criteria were classified as MASLD if they had one or more requisite cardiometabolic comorbidities². Individuals with a FLI of < 30 were considered a no-MASLD comparator, irrespective of cardiometabolic comorbidity, medication use, or alcohol consumption (Fig. 1). Those with a FLI between 30 and 60 were excluded from the analyses as indeterminate as those scores are unable to accurately either rule-in or rule-out MASLD [31, 32].

Fig. 1.

Baseline participant flow diagram. * Cardiometabolic features include those previously described²: BMI ≥ 25 kg/m² (or BMI ≥ 23 kg/m² in Asians); elevated abdominal circumference (> 94cm in males, > 80cm in females); fasting blood glucose ≥ 100mg/dL; self-described T2DM; drug treatment for T2DM; blood pressure ≥ 130/85 mmHg; treatment with an antihypertensive; plasma triglycerides ≥ 150 mg/dL; low HDL cholesterol (≤ 40mg/dL in males, ≤ 50mg/dL in females); treatment with a fibrate or HMG-CoA reductase inhibitor. EtOH is ethanol, MASLD is metabolic dysfunction-associated steatotic liver disease

Baseline characteristics and cardiometabolic comorbidities

For the purposes of this study (as is standard in the SLD literature), obesity was defined as BMI ≥ 25 kg/m² for Asians and ≥ 30 kg/m² for non-Asians [35]; and an elevated abdominal circumference [36] was defined as ≥ 90 cm for Asian males, ≥ 102 cm for non-Asian males, ≥ 80 cm for Asian females, and ≥ 88 cm for non-Asian females. Hypertension was defined as a systolic blood pressure ≥ 140 mmHg, a diastolic blood pressure ≥ 90 mmHg, and/or the use of antihypertensive medication(s) [37]. T2DM was defined as one or more of self- or physician-reported T2DM, a fasting glucose ≥ 126 mg/dL, and/or the prescription of hypoglycemic medication(s) [38]. No HbA1c data are available in ASPREE. Dyslipidemia was defined as a total cholesterol of ≥ 212 mg/dL and/or LDL cholesterol ≥ 160 mg/dL and/or the prescription of an HMG-CoA reductase inhibitor. CKD was defined as one or both of eGFR < 60 ml/min/1.73 m² or a urinary albumin-creatinine ratio of ≥ 25 mg/g (male) or ≥ 35 mg/g (female) [39]. The cardiometabolic abnormalities required for a diagnosis of MASLD were used as previously published, and include BMI ≥ 25 kg/m² (or BMI ≥ 23 kg/m² in Asians); elevated abdominal circumference (> 94 cm in males, > 80 cm in females); fasting blood glucose ≥ 100 mg/dL; self-described T2DM; drug treatment for T2DM; blood pressure ≥ 130/85 mmHg; treatment with an antihypertensive; triglycerides ≥ 150 mg/dL; low HDL cholesterol (≤ 40 mg/dL in males, ≤ 50 mg/dL in females); and/or treatment with a fibrate or HMG-CoA reductase inhibitor².

Defining outcomes

The study adjudication process for validating MACE, death, and cause of death in the overall ASPREE study has been previously described [29]. Briefly, each case was independently reviewed by two clinical experts, blinded to randomized treatment, with a third adjudicator resolving discordance. MACE was defined (as per the ASPREE publication [24] and based on World Health Organization criteria) as a composite of fatal coronary heart disease (excluding death from heart failure), nonfatal acute myocardial infarction (AMI), or fatal or nonfatal ischemic stroke [24]. Hospitalization for heart failure (HHF) was also an adjudicated outcome. The follow-up and adjudication of these endpoints has continued following cessation of the randomized interventional trial for a median of 8.4 (IQR 7.3 – 9.5) years. To evaluate the safety of aspirin in the MASLD subgroup, major hemorrhage during the trial period [24] was also adjudicated with the group stratified by aspirin or placebo allocation.

Similarly, the ascertainment and adjudication of death data in ASPREE has been previously described [25]. Death was usually identified either during a routine trial activity (which involved patient contact and concurrent clinical record examination) or through a relative/next-of-kin notifying researchers that the trial participant had died. Regardless of the initial method of notification, all deaths required confirmation from two independent sources (e.g., family, primary care physician, and/or a public death notice). In addition, staff performed linkages with the Ryerson Index, a community-maintained register compiled by volunteers who monitor death notices and obituaries. Once a death was confirmed, clinical details were sought from relevant facilities and practitioners (including hospitals, hospices, and treating clinicians), including progress notes, discharge summaries, and death certificates. Adjudication occurred via collated case presentations to at least two adjudicators (one USA-based and one Australian-based); discordance was resolved via consensus.

The identification of incident AF (subclassified as probable or possible) occurred as part of a separate ASPREE sub-study [40]. For the trial period only, the overall ASPREE cohort was reviewed by blinded investigators for triggers likely to be associated with AF. Participants were considered to have probable incident AF if there were multiple self-reports of AF or multiple reports of an irregular heart rate during multiple annual study contacts or were prescribed a medication used for AF management. These triggers prompted review of doctor correspondence and medical records where available to confirm AF diagnosis. Medication triggers included new prescription of anticoagulants (e.g., warfarin, direct oral anticoagulant) without diagnosis of incident thrombosis, or those commonly used for AF rate control. ECG records were noted where available. Participants with suggestive but lesser evidence of incident AF were classified as possible AF (e.g., irregular heart rate on a single study visit but no report AF); this subgroup (n = 391) was excluded from the AF-related analyses. Including them in the ‘not AF’ arm did not significantly alter results (data not shown). The identification and follow-up of AF cases had a median follow-up of 4.6 (IQR 3.5 – 5.5) years.

Statistical analysis

Baseline data were compared using a one-way ANOVA or two-tailed Student’s t-test (for continuous variables), or a Chi-squared test (for categorical variables), both for the MASLD vs no-MASLD group and for the aspirin vs placebo stratified MASLD sub-groups. The co-primary outcomes were whether MASLD increases the hazard of all-cause mortality, MACE, or AF. Secondary outcomes were whether aspirin has a protective effect in the MASLD group compared to placebo for each of the primary outcomes of interest. Tertiary outcomes were whether MASLD was associated with an increased hazard of individual sub-groups of MACE (incident AMI or stroke), or HHF. An exploratory analysis was also performed to determine whether any association between prevalent MASLD and incident AF remained after excluding those who had an AMI and/or HHF prior to their documented AF. Due to not all of the Australian participants having a calculable FLI, a crude comparison of rates of incident AF, MACE, and mortality was undertaken in the calculable FLI group vs others.

In the initial analysis, unadjusted Cox proportional hazard regression models were utilized to evaluate these outcomes. The model was subsequently adjusted for age, sex, and cardiometabolic comorbidities known to be associated with MACE and used for adjustment in other MASLD-related cardiovascular outcome studies³. For an ASPREE adjusted model, we used the variables known to be associated with cardiovascular disease in the overall ASPREE cohort [41] as well as treatment with aspirin/placebo. These variables include age, sex, creatinine, HDL and non-HDL cholesterol, systolic blood pressure, the prescription of an antihypertensive(s), T2DM, and being a current smoker. No models included BMI or triglycerides due to their inclusion in the FLI. Sensitivity analyses were performed by sub-analyzing the group with and without T2DM separately. Additional sensitivity analyses were performed on the individual components of the FLI. A two-tailed p value of < 0.05 was considered statistically significant. Statistical analyses were performed using Stata software v17.0 (StataCorp LLC, College Station, TX, USA). Patients or the public were not involved in the study design or reporting of this research.

Results

Study populations & cardiometabolic risk factors

Of the 16,703 Australian ASPREE participants, 11,914 provided blood for the biobank, and 9,847 participants had a calculable FLI using laboratory results from blood samples taken within 90 days post-recruitment. The group with a calculable FLI was similar to the group without, particularly in terms of atherosclerotic cardiovascular disease risk factors; however, they were slightly younger (75.0 vs 75.7 years, p < 0.001) and more likely to be male (46.8% vs 42.5%, p < 0.001) (Supplementary Table 1). Rates of MACE and AF were similar between groups with a calculable FLI and those without, though there were higher rates of death in those without a calculable FLI (19.2% vs 14.4%, p < 0.001) (Supplementary Table 2).

Of the group with a calculable FLI, 750 participants with a FLI ≥ 60 were excluded due to steatogenic medication and/or excess alcohol consumption, defined as > 140 g of alcohol per week for females and > 210 g of alcohol per week for males, leaving 9,097 with up to 76,860 person-years follow up time for the final MASLD analysis (Fig. 1). All the FLI ≥ 60 group (n = 2,998) had at least one cardiometabolic criterion present² with 449 (14.98%) having all five. The mean age of all participants was 75.1 ± 4.3 years, 55% were female, and 99% were Caucasian. The mean ± SD cohort BMI was 27.8 ± 4.4 kg/m²; 73% were overweight or obese. General characteristics of the participants stratified by FLI can be seen in Table 1 (comparing no-MASLD [FLI < 30], an indeterminate FLI [30–60], and MASLD [FLI ≥ 60]). In the final analysis group, 33% had a diagnosis of MASLD.

Table 1.

Baseline characteristics of Australian ASPREE participants

Characteristic	All participants	Non-MASLD FLI < 30	FLI 30–60	MASLD FLI ≥ 60	p-value
Number of participants	9097	2968 (32.6%)	3131 (34.4%)	2998 (33.0%)
FLI Values (median, IQR)		17 (10 – 23)	45 (37 – 53)	79 (69 – 89)
Sex (male)	4093 (45.0%)	921 (31.0%)	1624 (51.9%)	1548 (51.6%)	< 0.001a
Age (years) (mean ± SD)	75.06 ± 4.25	75.28 ± 4.42	75.23 ± 4.36	74.66 ± 3.91	< 0.001b
Age Category					< 0.001a
70–72 years (n, %)	3776 (41.5%)	1184 (39.9%)	1277 (40.8%)	1315 (43.9%)
73–75 years (n, %)	2349 (25.8%)	764 (25.7%)	782 (25.0%)	803 (26.8%)
76–78 years (n, %)	1363 (15.0%)	436 (14.7%)	484 (15.5%)	443 (14.8%)
79–81 years (n, %)	856 (9.4%)	300 (10.1%)	305 (9.7%)	251 (8.4%)
82–84 years (n, %)	449 (4.9%)	156 (5.3%)	171 (5.5%)	122 (4.1%)
≥ 85 years (n, %)	304 (3.3%)	128 (4.3)	112 (3.6%)	64 (2.1%)
Alcohol Use					0.027a
Current Drinker (n, %)	7148 (78.6%)	2359 (79.5%)	2489 (79.5%)	2300 (76.7%)
Former Drinker (n, %)	437 (4.8%)	129 (4.3%)	140 (4.5%)	168 (5.6%)
Never Drinker (n, %)	1512 (16.6%)	480 (16.2%)	502 (16.0%)	530 (17.7%)
Ethnic Background					0.050a
White/Caucasian§	8961 (98.6%)	2907 (98.0%)	3090 (98.8%)	2964 (99.0%)
Asian	74 (0.8%)	37 (1.2%)	21 (0.6%)	16 (0.5%)
More than one race	34 (0.4%)	17 (0.6%)	8 (0.3%)	9 (0.3%)
Other	22 (0.2%)	6 (0.2%)	10 (0.3%)	6 (0.2%)
Weight (kg) (mean ± SD)	76.12 ± 14.03	64.07 ± 8.69	75.75 ± 8.68	88.44 ± 12.16	< 0.001b
BMI (kg/m2) (mean ± SD)	27.75 ± 4.43	23.90 ± 2.37	27.33 ± 2.23	31.99 ± 4.01	< 0.001b
BMI Category35*					< 0.001a
Underweight	43 (0.5%)	43 (1.4%)	0 (0.0%)	0 (0.0%)
Healthy weight	2422 (26.6%)	1971 (66.4%)	419 (13.4%)	32 (1.1%)
Overweight	4224 (46.4%)	929 (31.3%)	2328 (74.4%)	967 (32.3%)
Obese	2408 (26.5%)	25 (0.8%)	384 (12.3%)	1999 (66.7%)
Waist Circumference (cm) (mean ± SD)	96.07 ± 12.29	84.27 ± 7.84	96.22 ± 6.38	107.59 ± 9.18	< 0.001b
Elevated abdominal circumference36 (n, %)†	4971 (54.6%)	499 (16.8%)	1743 (55.7%)	2729 (91.0%)	< 0.001a
Hypertension (n, %)¶	5360 (58.9%)	1507 (50.8%)	1873 (59.8%)	1980 (66.1%)	< 0.001a
Diabetes Mellitus (n, %)§§	839 (9.2%)	109 (3.7%)	227 (7.3%)	503 (16.8%)	< 0.001a
Insulin requiring (n, % of diabetes)	67 (8.0%)	11 (10.1%)	19 (8.4%)	37 (7.4%)
Dyslipidemia (n, %)**	6134 (67.4%)	1930 (65.0%)	2102 (67.1%)	2102 (70.1%)	< 0.001a
Metabolic Syndrome (n, %)††	2116 (23.3%)	209 (7.0%)	584 (18.7%)	1323 (44.1%)	< 0.001a
Use of HMG-CoA Reductase Medication	3082 (33.9%)	774 (26.1%)	1067 (34.1%)	1241 (41.4%)	< 0.001a
Use of Fibrate Medication	76 (0.8%)	12 (0.4%)	23 (0.7%)	41 (1.4%)	< 0.001a
Smoking History (n, %)					< 0.001a
Never-smoker	5215 (57.3%)	1863 (62.8%)	1735 (55.4%)	1617 (53.9%)
Former smoker	3604 (39.6%)	1000 (33.7%)	1293 (41.3%)	1311 (43.7%)
Current smoker	278 (3.1%)	105 (3.5%)	103 (3.3%)	70 (2.3%)
Pack-Year History (mean ± SD)	21.45 ± 23.70	17.62 ± 20.86	21.51 ± 23.05	24.45 ± 25.95	< 0.001b
Laboratory Values (mean ± SD)¶¶
GGT (U/L)	27.54 ± 29.92	18.67 ± 10.94	24.93 ± 15.94	39.04 ± 46.01	< 0.001b
ALT (U/L)	20.12 ± 10.94	17.15 ± 6.77	19.42 ± 7.91	23.77 ± 15.15	< 0.001b
AST (U/L)	21.76 ± 7.73	21.31 ± 5.72	21.20 ± 6.04	22.79 ± 10.46	< 0.001b
TSH (mIU/L) (mean ± SD)	1.55 ± 2.03	1.49 ± 1.14	1.54 ± 2.25	1.63 ± 2.44	< 0.001b
Total Cholesterol (mg/dL)	202.9 ± 37.4	206.6 ± 36.3	203.1 ± 37.4	198.9 ± 38.2	< 0.001b
Non-HDL-C (mg/dL)	142.0 ± 35.8	136.2 ± 33.9	143.4 ± 35.4	146.5 ± 37.1	< 0.001b
HDL-C (mg/dL)	60.9 ± 17.6	70.4 ± 17.8	60.0 ± 15.9	52.6 ± 14.2	< 0.001b
Fasting Glucose ≥ 100 mg/dL (n, %)	3129 (34.7%)	604 (20.6%)	1066 (34.4%)	1459 (49.2%)	< 0.001a
Triglycerides ≥ 150 mg/dL (n, %)	1965 (21.6%)	119 (4.0%)	507 (16.2%)	1339 (44.7%)	< 0.001a
Renal Function
eGFR (mL/min/1.73 m2) (mean ± SD)	72.76 ± 13.36	74.71 ± 12.74	72.61 ± 13.08	70.99 ± 13.99	< 0.001b
Creatinine (mg/dL)	0.90 ± 0.22	0.85 ± 0.19	0.92 ± 0.22	0.95 ± 0.23	< 0.001b
Chronic Kidney Disease§§§ (n, %)	1622 (17.8%)	413 (13.9%)	538 (17.2%)	672 (22.4%)	< 0.001a

^§ Where ethnicity wasn’t recorded/reported, White/Caucasian was assumed

^* Underweight = BMI < 18.5 kg/m², Healthy weight = BMI 18.5–22.9 kg/m² (Asians) or 18.5–24.9 kg/m² (non-Asian), Overweight = BMI 23–24.9 kg/m² (Asians) or 25–29.9 kg/m² (non-Asian), Obese = BMI ≥ 25 kg/m² (Asians) or ≥ 30 kg/m² (non-Asian)

^† Elevated abdominal circumference = If Asian, ≥ 88cm (males) and ≥ 80cm (females); if non-Asian, ≥ 102cm (males) and ≥ 90cm (females)

^¶ Defined as Systolic Blood Pressure ≥ 140mmHg and/or Diastolic Blood Pressure ≥ 90 mmHg and/or prescription of at least one antihypertensive medication at baseline

^§§ Defined as one or more of: (a) self-reported diabetes mellitus, (b) prescription of at least one glucose lowering therapy at baseline, (c) fasting blood glucose of ≥ 126 mg/dL

^** Defined as one or more of: total cholesterol ≥ 212 mg/dL, LDL cholesterol ≥ 160 mg/dL, or an HMG-CoA reductase inhibitor

^†† Defined as 3 or more of: elevated abdominal circumference; triglycerides ≥ 150 mg/dL and/or being on a fibrate; HDL < 1.03 mmol/L (males) or < 1.24 mmol/L (females) and/or being on a HMG-CoA reductase inhibitor; systolic blood pressure ≥ 130mmHg and/or diastolic blood pressure ≥ 85mmHg and/or antihypertensive medication(s); fasting glucose ≥ 100 mg/dL and/or a self-reported diagnosis of diabetes mellitus and/or prescription of diabetes medication(s)

^¶¶ GGT is gamma-glutamyltransferase, ALT is alanine aminotransferase, AST is aspartate aminotransferase, TSH is thyroid stimulating hormone

^§§§ Defined as eGFR < 60 mL/min/1.73 m² and/or a urine albumin-creatinine ratio of > 25 mg/g (males) or > 35 mg/g (females)

^a Chi-Square p value

^b One-way ANOVA p value

There were significantly higher rates of baseline cardiometabolic comorbidities in the MASLD group vs the no-MASLD group, including obesity (66.7% vs 0.8%), T2DM (16.8% vs 3.7%), hypertension (82.7% vs 69.6%), dyslipidemia (70.1% vs 65.0%), CKD (30.3% vs 22.4%), and ever-smoking status (46.1% vs 37.6%) (Table 1). In the aspirin vs placebo treated subgroup of MASLD there was a slightly lower proportion with baseline hypertension in the aspirin-treated group (64.1% vs 67.9%, p = 0.03) with no other differences in any clinical, biochemical, or demographic factors (Supplementary Table 3).

Major adverse cardiovascular events

There were 680 cases of MACE in the total population group (7.5%) over a median (IQR) follow-up period of 8.3 (7.0 – 9.5) years, of which a median (IQR) 4.6 (3.3 – 5.5) years were under exposure to aspirin, corresponding to an event rate of 9.6 per 1000 person-years. As expected, the group with MACE had a higher prevalence of known risk factors for CVD (Table 2). On unadjusted analysis and when adjusted for age and sex, MASLD was significantly associated with incident MACE (HR 1.42, 95% CI 1.16 – 1.72) (Fig. 2, Table 3). Similarly, when adjusting for known CVD risk factors (dyslipidemia, T2DM, hypertension, and smoking status), MASLD remained significantly associated with incident MACE (HR 1.36, 95% CI 1.11 – 1.67). However, when adjusting for factors shown to be associated with MACE in the more comprehensive ASPREE MACE risk model [41], this association was attenuated (HR 1.12, 95% CI 0.89 – 1.42) (Table 3). In our exploratory analyses of sub-categories of MACE or HHF (Supplementary Table 4), there was an association between MASLD and both AMI and HHF in minimally adjusted models, though this association was no longer apparent when the model was fully adjusted for all known CVD risk factors in the ASPREE population. There was no association between MASLD and stroke (Supplementary Table 4). Aspirin use during the trial period (median [IQR] 4.5 [3.4 – 5.5] years) was not protective against MACE in the MASLD group either during the trial period alone (HR 1.11, p = 0.581) or during the overall trial and post-trial follow-up period (HR 1.05, p = 0.692) (Supplementary Table 5). However, aspirin also did not lead to an excess of major hemorrhage in the MASLD group during the trial period (3.37% vs 2.63%, p = 0.26). The absence of baseline T2DM did not change these results (Supplementary Table 6), and each individual component of the FLI was weakly associated with MACE when adjusted for age and sex (Supplementary Table 7).

Table 2.

Relationship between baseline characteristics and MACE, AF, and all-cause mortality in Australian ASPREE participants

Outcome	No MACE	MACE	p-value	No AF	Incident AF	p-value	No death	Death	p-value
Number of participants	8417	680		8229	477		7807	1290
Sex (male)	3699 (43.9%)	394 (57.9%)	< 0.001a	3652 (44.4%)	247 (51.8%)	0.002a	3397 (43.5%)	696 (54.0%)	< 0.001a
MASLD Category (3 groups)			0.001a			0.010a			0.917a
No MASLD (FLI < 30) (n, %)	2790 (33.1%)	178 (26.2%)		2731 (33.2%)	129 (27.4%)		2541 (32.5%)	427 (33.1%)
Indeterminate MASLD (FLI 30 – 60) (n, %)	2880 (34.2%)	251 (36.9%)		2821 (34.3%)	167 (35.0%)		2692 (34.5%)	439 (34.0%)
MASLD (FLI ≥ 60) (n, %)	2747 (32.6%)	251 (36.9%)		2677 (32.5%)	181 (37.9%)		2573 (33.0%)	424 (32.9%)
MASLD Category (2 groups)			< 0.001a			0.002a			0.788a
No MASLD (FLI < 30) (n, %)	2790 (50.4%)	178 (41.5%)		2731 (50.5%)	129 (41.6%)		2541 (49.7%)	427 (50.2%)
MASLD (FLI ≥ 60) (n, %)	2747 (49.6%)	251 (58.5%)		2677 (49.5%)	181 (58.4%)		2573 (50.3%)	424 (49.8%)
Age (years) (mean ± SD)	74.92 ± 4.17	76.72 ± 4.75	< 0.001b	74.94 ± 4.18	76.56 ± 4.78	< 0.001b	74.57 ± 3.85	78.00 ± 5.24	< 0.001b
Age Category			< 0.001a			< 0.001a			< 0.001a
70–72 years (n, %)	3589 (42.6%)	187 (27.5%)		3503 (42.6%)	142 (29.8%)		3501 (44.8%)	275 (21.3%)
73–75 years (n, %)	2189 (26.0%)	160 (23.5%)		2125 (25.8%)	113 (23.7%)		2096 (26.8%)	253 (19.6%)
76–78 years (n, %)	1243 (14.8%)	120 (17.6%)		1219 (14.8%)	82 (17.2%)		1123 (14.4%)	240 (18.6%)
79–81 years (n, %)	748 (8.9%)	108 (15.9%)		750 (9.1%)	64 (13.4%)		633 (8.1%)	223 (17.3%)
82–84 years (n, %)	385 (4.6%)	64 (9.4%)		373 (4.5%)	49 (10.3%)		306 (3.9%)	143 (11.1%)
≥ 85 years (n, %)	263 (3.1%)	41 (6.0%)		259 (3.1%)	27 (5.7%)		148 (1.9%)	156 (12.1%)
Alcohol Use			0.047a			0.889a			0.058a
Current Drinker (n, %)	6635 (78.8%)	513 (75.4%)		6463 (78.5%)	374 (78.4%)		6164 (79.0%)	984 (76.3%)
Former Drinker (n, %)	393 (4.7%)	44 (6.5%)		393 (4.8%)	25 (5.2%)		362 (4.6%)	75 (5.8%)
Never Drinker (n, %)	1389 (16.5%)	123 (18.1%)		1373 (16.7%)	78 (16.4%)		1281 (16.4%)	231 (17.9%)
Ethnic Background			0.431a			0.635a			0.412a
White/Caucasian§	8299 (98.6%)	668 (98.2%)		8106 (98.5%)	472 (99.0%)		7693 (98.5%)	1274 (98.8%)
Asian	65 (0.8%)	9 (1.3%)		68 (0.8%)	4 (0.8%)		68 (0.9%)	6 (0.5%)
More than one race	32 (0.4%)	2 (0.4%)		22 (0.3%)	0 (0.0%)		18 (0.2%)	4 (0.3%)
Other	21 (0.2%)	1 (0.2%)		33 (0.4%)	1 (0.2%)		28 (0.4%)	6 (0.5%)
Weight (kg) (mean ± SD)	76.00 ± 14.01	77.59 ± 14.20	0.005b	75.86 ± 13.89	79.74 ± 15.55	< 0.001b	76.18 ± 13.87	75.77 ± 14.95	0.337b
BMI (kg/m2) (mean ± SD)	27.73 ± 4.43	27.90 ± 4.47	0.351b	27.70 ± 4.38	28.42 ± 5.25	0.001b	27.81 ± 4.37	27.36 ± 4.75	0.001b
BMI Category35*			0.631a			0.099a			< 0.001a
Underweight	38 (0.5%)	5 (0.7%)		37 (0.4%)	3 (0.6%)		25 (0.3%)	18 (1.4%)
Healthy weight	2250 (26.7%)	172 (25.3%)		2225 (27.0%)	113 (23.7%)		2047 (26.2%)	375 (29.1%)
Overweight	3906 (46.4%)	318 (46.8%)		3821 (46.4%)	214 (44.9%)		3636 (46.6%)	588 (45.6%)
Obese	2223 (26.4%)	185 (27.2%)		2146 (26.1%)	147 (30.8%)		2099 (26.9%)	309 (24.0%)
Waist Circumference (cm) (mean ± SD)	95.90 ± 12.27	98.10 ± 12.28	< 0.001b	95.87 ± 12.16	98.68 ± 13.83	< 0.001b	95.97 ± 12.28	97.07 ± 12.34	0.002b
Elevated abdominal circumference36 (n, %)†	4582 (54.4%)	389 (57.2%)	0.163a	4453 (54.1%)	287 (60.2%)	0.010a	4279 (54.8%)	692 (53.6%)	0.436a
Hypertension (n, %)¶	4916 (58.4%)	444 (65.3%)	< 0.001a	4817 (58.5%)	305 (63.9%)	0.020a	4549 (58.3%)	811 (62.9%)	0.002a
Diabetes Mellitus (n, %)§§	757 (9.0%)	82 (12.1%)	0.008a	741 (9.0%)	53 (11.1%)	0.120a	677 (8.7%)	162 (12.6%)	< 0.001a
Insulin requiring (n, % of diabetes)	63 (0.8%)	4 (0.6%)	0.638a	56 (0.7%)	6 (1.3%)	0.145a	49 (0.6%)	18 (1.4%)	0.003a
Dyslipidemia (n, %)**	5692 (67.6%)	442 (65.0%)	0.160a	5583 (67.8%)	292 (61.2%)	0.003a	5354 (68.6%)	780 (60.5%)	< 0.001a
Metabolic Syndrome (n, %)††	1919 (22.8%)	197 (29.0%)	< 0.001a	1915 (23.3%)	107 (22.4%)	0.673a	1779 (22.8%)	337 (26.1%)	0.009a
Use of HMG-CoA Reductase Medication	2872 (34.1%)	210 (30.9%)	0.086a	2783 (33.8%)	159 (33.3%)	0.827a	2676 (34.3%)	406 (31.5%)	0.049a
Use of Fibrate Medication	70 (0.8%)	6 (0.9%)	0.889a	72 (0.9%)	1 (0.2%)	0.121a	66 (0.8%)	10 (0.8%)	0.797a
Smoking History			0.001a			0.011a			< 0.001a
Never-smoker (n, %)	4871 (57.9%)	344 (50.6%)		4760 (57.8%)	246 (51.6%)		4559 (58.4%)	656 (50.9%)
Former smoker (n, %)	3295 (39.1%)	309 (45.4%)		3213 (39.0%)	219 (45.9%)		3052 (39.1%)	552 (42.8%)
Current smoker (n, %)	251 (3.0%)	27 (4.0%)		256 (3.1%)	12 (2.5%)		196 (2.5%)	82 (6.4%)
Pack-Year History (mean ± SD)	21.26 ± 23.67	23.45 ± 23.93	0.105b	21.10 ± 23.54	24.31 ± 23.38	0.045b	20.53 ± 23.28	26.19 ± 25.23	< 0.001b
Laboratory Values¶¶
GGT (U/L) (mean ± SD)	27.44 ± 30.17	28.71 ± 26.67	0.286b	27.51 ± 30.63	29.92 ± 24.60	0.214b	27.31 ± 29.39	28.90 ± 32.97	0.077b
ALT (U/L) (mean ± SD)	20.15 ± 10.95	19.69 ± 10.85	0.289b	20.19 ± 11.14	19.54 ± 8.75	0.211b	20.31 ± 11.04	18.92 ± 10.26	< 0.001b
AST (U/L) (mean ± SD)	21.80 ± 7.74	21.28 ± 7.62	0.090b	21.78 ± 7.78	22.09 ± 7.77	0.393b	21.81 ± 7.74	21.44 ± 7.69	0.109b
Fasting Glucose ≥ 100 mg/dL (n, %)	2937 (34.4%)	192 (41.3%)	0.001a	2820 (34.6%)	164 (35.0%)	0.851a	2636 (34.1%)	493 (38.5%)	0.003a
Non-HDL-C (mg/dL) (mean ± SD)	141.7 ± 35.7	146.2 ± 36.7	0.002b	142.3 ± 37.6	137.1 ± 34.7	0.002b	142.6 ± 35.6	138.4 ± 36.6	< 0.001b
HDL Cholesterol (mg/dL) (mean ± SD)	61.3 ± 17.7	56.5 ± 15.9	< 0.001b	61.1 ± 17.6	60.1 ± 17.7	0.245b	61.1 ± 17.5	59.8 ± 17.9	0.016b
Triglycerides ≥ 150 mg/dL (n, %)	1798 (21.4%)	167 (24.6%)	0.051a	1782 (21.7%)	92 (19.3%)	0.221a	1691 (21.6%)	275 (21.3%)	0.790a
Renal Function
eGFR (mL/min/1.73 m2) (mean ± SD)	73.01 ± 13.25	69.75 ± 14.41	< 0.001b	72.84 ± 13.35	72.34 ± 13.02	0.393b	73.29 ± 13.14	69.58 ± 14.26	< 0.001b
Serum creatinine (mg/dL)	0.90 ± 0.21	0.97 ± 0.27	< 0.001b	0.90 ± 0.22	0.91 ± 0.21	0.358b	0.90 ± 0.21	0.95 ± 0.25	< 0.001b
Chronic Kidney Disease§§§ (n, %)	1438 (17.1%)	184 (27.1%)	< 0.001a	1443 (17.5%)	87 (18.2%)	0.019a	1285 (16.5%)	337 (26.1%)	< 0.001a

^§ Where ethnicity wasn’t recorded/reported, White/Caucasian was assumed

^* Underweight = BMI < 18.5 kg/m², Healthy weight = BMI 18.5–22.9 kg/m² (Asians) or 18.5–24.9 kg/m² (non-Asian), Overweight = BMI 23–24.9 kg/m² (Asians) or 25–29.9 kg/m² (non-Asian), Obese = BMI ≥ 25 kg/m² (Asians) or ≥ 30 kg/m² (non-Asian)

^† Elevated abdominal circumference = If Asian, ≥ 88cm (males) and ≥ 80cm (females); if non-Asian, ≥ 102cm (males) and ≥ 90cm (females)

^¶ Defined as Systolic Blood Pressure ≥ 140mmHg and/or Diastolic Blood Pressure ≥ 90 mmHg and/or prescription of at least one antihypertensive medication at baseline

^§§ Defined as one or more of: (a) self-reported diabetes mellitus, (b) prescription of at least one glucose lowering therapy at baseline, (c) fasting blood glucose of ≥ 126 mg/dL

^** Defined as one or more of: total cholesterol ≥ 212 mg/dL, LDL cholesterol ≥ 160 mg/dL, or an HMG-CoA reductase inhibitor

^†† Defined as 3 or more of: elevated abdominal circumference; triglycerides ≥ 150 mg/dL and/or being on a fibrate; HDL < 1.03 mmol/L (males) or < 1.24 mmol/L (females) and/or being on a HMG-CoA reductase inhibitor; systolic blood pressure ≥ 130mmHg and/or diastolic blood pressure ≥ 85mmHg and/or antihypertensive medication(s); fasting glucose ≥ 100 mg/dL and/or a self-reported diagnosis of diabetes mellitus and/or prescription of diabetes medication(s)

^¶¶ GGT is gamma-glutamyltransferase, ALT is alanine aminotransferase, AST is aspartate aminotransferase

^§§§ Defined as eGFR < 60 mL/min/1.73 m² and/or a urine albumin-creatinine ratio of > 25 mg/g (males) or > 35 mg/g (females)

^a Chi-Square p value

^b One-way ANOVA p value

Fig. 2.

Unadjusted Kaplan–Meier survival curves of MASLD vs no-MASLD for the co-primary outcomes

Table 3.

Association between MASLD and MACE, AF, and all-cause mortality in Australian ASPREE participants

	MACE Median Follow-Up 8.4 (7.3 – 9.5) Years		Incident AF Median Follow-Up 4.4 (3.3 – 5.4) Years		Death Median Follow-Up 8.5 (7.4 – 9.5) Years
MASLD (FLI ≥ 60) vs no-MASLD (FLI < 30)	Hazard Ratio (95% C.I)	p-value	Hazard Ratio (95% C.I)	p-value	Hazard Ratio (95% C.I)	p-value
Unadjusted/Crude Model	1.473 (1.22 – 1.78)	< 0.001	1.50 (1.19 – 1.88)	< 0.001	1.04 (0.91 – 1.19)	0.603
Age and Sex Adjusted	1.42 (1.16 – 1.72)	0.001	1.49 (1.18 – 1.88)	0.001	1.10 (0.96 – 1.27)	0.173
Adjusted for Common Risk Factors*	1.36 (1.11 – 1.67)	0.003	1.47 (1.15 – 1.87)	0.002	1.04 (0.90 – 1.20)	0.575
ASPREE adjusted model**	1.12 (0.89 – 1.42)	0.327	1.46 (1.11 – 1.93)	0.007	0.97 (0.82 – 1.14)	0.708

^* Common risk factors defined as age, sex, current smoking status, hypertension (SBP ≥ 140 mmHg and/or DBP ≥ 90 mmHg and/or antihypertensive medication), dyslipidemia (total cholesterol ≥ 212 mg/dL and/or LDL cholesterol ≥ 160 mg/dL and/or use of an HMG-CoA reductase inhibitor), and T2DM (self-described T2DM and/or a fasting serum glucose ≥ 126 mg/dL and/or hypoglycemic medication(s))

^** ASPREE adjusted model⁴¹ includes: age, sex, creatinine, HDL-C, non-HDL-C, systolic blood pressure, prescription of an antihypertensive medication(s), T2DM (self-described T2DM and/or a fasting blood glucose ≥ 126 mg/dL and/or hypoglycemic medication(s)), current smoking status, and trial treatment with aspirin or placebo

Incident atrial fibrillation

There were 477 cases of incident AF in the study population over a median (IQR) follow-up of 4.4 (3.3 – 5.5) years representing an incident AF rate of 12.1 per 1000 person-years (Fig. 2, Table 3) during the on-treatment phase of the ASPREE clinical trial. Like the baseline univariate associations seen in the MACE vs no-MACE groups, the incident AF group had higher prevalence of known cardiometabolic comorbidities, higher baseline BMI, and larger abdominal circumference (Table 2). MASLD was associated with AF in the unadjusted (Fig. 2) and adjusted models, including in an ASPREE adjusted model (HR 1.46, 95% CI 1.11 – 1.93) (Table 3) and in the group without T2DM at baseline (Supplementary Table 6). This risk was not reduced by aspirin in the MASLD sub-group (Supplementary Table 5). MASLD was associated with 4.5 additional cases of AF per 1000 person years. Of the 477 cases of incident AF, 163 (34.2%) occurred within the first 2 years, and 269 (56.4%) occurred between years 2 and 5 post-enrolment. When excluding those who were hospitalized for heart failure and/or had an AMI prior to the development of their incident AF, the association between MASLD and AF remained in all the models (Supplementary Table 8). Additionally, we examined the effect of sex on the association between MASLD and AF by including an interaction term in the fully adjusted model. The interaction between sex and MASLD was not statistically significant (p = 0.541), indicating that the relationship between MASLD and AF does not differ by sex.

All-cause mortality

In our subcohort of 9,097 participants, there were 1,290 deaths (17.0 per 1000 person-years). There were lower rates of cancer death and higher rates of ‘other’ (non-cancer non-CVD) deaths in the no-MASLD group (Table 4). Cancer deaths were the largest proportion overall (562, 43.6%) with cardiovascular deaths the third most common (298, 23.1%). MASLD had no influence on the hazard for all-cause mortality in this cohort in either crude or adjusted analyses (HR 1.04 [95% CI 0.91 – 1.19] and HR 0.97 [95% CI 0.82 – 1.14] respectively) (Table 3, Fig. 2). Aspirin had no impact on mortality in the MASLD subgroup (Supplementary Table 5). There were 4 (0.3%) liver-related deaths (primarily determined by review of clinical documentation and death certificates). All 4 occurred in the MASLD group.

Table 4.

Causes of Death by MASLD status. CVD-Related death includes ischemic events, strokes, and arrhythmias

Characteristic	All participants	FLI < 30	FLI 30–60	FLI ≥ 60	p value
Deaths	1290	427	439	424
Death Type					0.008a
CVD-Related (n, %)	298 (23.1%)	107 (25.1%)	92 (21.0%)	99 (23.3%)
Cancer (n, %)	562 (43.6%)	153 (35.8%)	203 (46.2%)	206 (48.6%)
Bleeding (n, %)	17 (1.3%)	6 (1.4%)	4 (0.9%)	7 (1.7%)
Not Adjudicated (n, %)	2 (0.2%)	1 (0.2%)	1 (0.2%)	0 (0.0%)
Other (n, %)	411 (31.9%)	160 (37.5%)	139 (31.7%)	112 (26.4%)
Liver-Related (n, % of Other)	4 (1.0%)	0 (0.0%)	0 (0.0%)	4 (3.6%)

^a Chi-Squared p value

Discussion

Steatotic liver disease has been associated with both liver- and non-liver related morbidity and mortality in adults [42]. However, a multitude of studies have subsequently shown no [16] or attenuated risk [15, 17] with SLD in individuals as they age. Given these discrepant findings, and that many guidelines discuss or recommend screening for SLD in adults with known risk factor(s) for this [26, 43, 44], we evaluated whether community-dwelling older adults with MASLD had an increased risk of incident major CVD events and whether preventive therapy with aspirin improved outcomes in those with MASLD. The main findings were that in older adults MASLD was associated with AF but not MACE or all-cause mortality in a fully adjusted model, and that aspirin as a primary preventive agent did not benefit older adults with MASLD.

Mortality in older adults with MASLD

In keeping with the results recently shown by van Kleef and colleagues [16], we have shown that MASLD is not associated with all-cause mortality in a large population of relatively healthy community-dwelling older Australian adults. Furthermore, unlike data from predominantly middle-aged MASLD groups where CVD deaths are often the largest contributor to mortality [26, 45], CVD-deaths were only the third most commonly categorized cause of death in this population overall and in the MASLD sub-group (with cancer-related mortality the most common cause of death). Hepatic causes accounted for very few deaths, only 0.8% of the MASLD sub-group. While this population is not representative of all older adults (due to the lack of previous CVD events, cognitive dysfunction, and functional decline), it is a group commonly seen in ambulatory care settings worldwide. These data lend support to the notion that actively screening for MASLD in at-risk older adults is unlikely to yield meaningful improvements in mortality.

MACE

The risk of incident MACE was higher in the MASLD group compared to the non-MASLD group, driven by increased rates of AMI, even after adjusting for key categorical risk factors for incident CVD, including T2DM, hypertension, and dyslipidemia. Interestingly, this association was lost after undertaking a more comprehensive adjustment for the established CVD risk factors in the ASPREE group [41]. The primary difference between these models was the inclusion of HDL cholesterol, non-HDL cholesterol, systolic blood pressure, and creatinine as continuous variables. There are multiple possible reasons for this finding. One interpretation of these data is that a proportion of the association between MASLD and CVD is driven by known consequences of MASLD including CKD [18] and insulin resistance/T2DM [19]. Mediation – or confounding – by these factors may help explain the association with MASLD. Secondly, it’s possible that other reports linking SLD to increased rates of CVD have not adequately adjusted for known risk factors for CVD (including by not utilizing all of the variables commonly used in CVD risk calculators [41, 46]) thus overestimating the strength of the association between MASLD and CVD. Finally, it’s also possible that the enrolment criteria of this study of relatively healthy older community dwelling adults pre-selected a MASLD group with a relatively more benign phenotype (i.e., healthier) such that their steatotic liver disease did not impact on their CVD risk as strongly as in a younger population or in older adults with pre-existing dementia or independence-limiting physical disability.

An additional important and novel finding from this study is that treatment with aspirin was not effective at reducing the risk of MACE or mortality in the MASLD population either during or after the randomized controlled trial. Previous work has suggested that aspirin may have important benefits for liver health in terms of reducing the risk of both hepatocellular carcinoma [47] and fibrosis progression [48], reducing liver fat content [49], and also reducing atherosclerosis [50]. Moreover, given that previous data have highlighted the strong association of SLD with CVD events, the evaluation of the merits of primary prophylaxis of a potentially high-risk population is an important clinical consideration. However, data from this study do not support the use of aspirin as primary prophylaxis for MACE or death in older adults with MASLD.

Atrial fibrillation

We have also shown an association between MASLD and incident AF in older adults. This link, unlike the association between MASLD and MACE, remains significant even in models adjusted for known risk-factors for incident CVD. There have been conflicting data on this topic previously. Biologically, the link between MASLD and AF is plausible due to the known association between overweight/obesity, epicardial adipose tissue, and AF [51]. Supporting this, a recent review [52] has shown a link between MASLD and both prevalent and incident AF. However, in older populations, this relationship is less clear. The Rotterdam Study has shown an association between elevated liver stiffness but not ultrasound-determined liver steatosis with both incident and prevalent AF [53] in a group with a mean age of 69.5 years. Our data conflict with these, possibly due to our use of the FLI to define MASLD rather than ultrasound, though the FLI has been validated in the Rotterdam Study population previously [54], and possibly due to a longer follow-up period for our study (median 4.4 years vs 2.1 years). Supporting this is our finding that only 34.2% of the AF cases occurred within 2 years; the risk of AF appears to accumulate over time. These have important implications for screening and counselling older adults with MASLD.

Strengths and weaknesses

Our study has several strengths, including its size, the rigorous protocol-driven prospective data capture at enrolment and during follow-up, and robust ascertainment and adjudication of key endpoints. However, some limitations should be discussed. FLI was used to determine MASLD, rather than biopsy or imaging. While FLI has been previously validated as a population-based marker of hepatic steatosis when compared to abdominal ultrasound diagnosed SLD in a similar population to ours [54], abdominal ultrasound itself is not the most sensitive modality for diagnosing hepatic steatosis. Though there are a lack of data directly correlating the FLI to biopsy-proven SLD, these results broadly support our categorization of MASLD vs no-MASLD. It should be noted that our AF allocation is based on clinical grounds (including self-reports, recorded irregular heart rates, prescription medications, and clinical notes) and did not always have ECG confirmation. Additionally, the ASPREE population was a relatively healthy community-dwelling cohort and care should be taken in extrapolating this data to other older populations. However, because of this, our study population has a presumed longer life expectancy such that some may believe that screening for MASLD could be considered worthwhile despite their biological age; understanding the utility of this is important for determining age-appropriate recommendations in clinically relevant populations. Additionally, our study relied on self-reported alcohol intake as well as the identification of AF by algorithm rather than objective tests. Finally, this population was not ethnically diverse, thereby limiting the capacity to generalize the findings to other ethnicities.

Conclusion

In this sub-study of a very large randomized-controlled trial involving relatively healthy community-dwelling adults aged ≥ 70 years, CVD was only the third most commonly categorized cause of death amongst the MASLD cohort and liver-related deaths accounted for < 1%. Furthermore, MASLD was not associated with all-cause mortality but was associated with MACE prior to adjustment for key cardiometabolic risk factors; age, sex, dyslipidemia, T2DM, hypertension, and smoking status. Additionally, we found an independent association of MASLD with incident AF. While the association with MACE was lost after adjustment was performed for CVD risk factors previously established in the ASPREE cohort, the association with AF remained strong. Aspirin as a potential primary preventive therapy did not affect any of these outcomes. These data have important clinical implications when counselling older adults and determining age-appropriate screening guidelines.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

DCC: conceptualization, formal analysis, investigation, writing (original draft), writing (review & editing). SKR: conceptualization, writing (review & editing). AM: conceptualization, writing (review & editing). RLW: conceptualization, writing (review & editing), funding acquisition, investigation, project administration. AMT: writing (review & editing), investigation. MRN: writing (review & editing), funding acquisition, investigation, project administration. ATC: writing (review & editing), investigation, funding acquisition, project administration. JR: writing (review & editing), funding acquisition, investigation. CT: writing (review & editing). ADH: writing (review & editing). JSL: writing (review & editing). HGS: writing (review & editing), funding acquisition, investigation. AB: writing (review & editing), investigation. SMF: writing (review & editing), investigation. SGO: writing (review & editing), investigation. JJM: writing (review & editing), investigation, funding acquisition, project administration. WWK: writing (original draft), writing (review & editing), supervision.

Funding

The ASPREE clinical trial was supported by the National Institute on Aging and the National Cancer Institute at the National Institutes of Health (U01AG029824, U19AG062682); the NHMRC (334047, 1127060); Monash University; and the Victorian Cancer Agency. The ASPREE Biobank was supported by research grants from the Australian Government’s CSIRO (Commonwealth Scientific and Industrial Research Organisation; Preventative Health Flagship 2009) and the National Cancer Institute/National Institutes of Health (5U01AG029824-02). Abbott Diagnostics provided a grant for the measurement of laboratory parameters. Dr Chan is an American Cancer Society Research Professor. No funding sources were involved in the design or conduct of the study; collection, management, or analysis of the data; interpretation of the results; preparation, review, or approval of the manuscript; or decision to submit it for publication.

National Institutes of Health and Australian National Health and Medical Research Council, reagent grant from Abbott Diagnostics.

Data availability

The datasets used and/or analysed for this publication are available via the ASPREE Principal Investigators. Requests for data access can be directed to aspree.ams@monash.edu.