Objective Recent evidence suggests that alcoholic fatty liver disease (AFLD) and non-alcoholic fatty liver disease (NAFLD) may differentially affect risk of cardiovascular mortality. To investigate whether early liver disease due to AFLD or NAFLD have similar or dissimilar effects on risk of early coronary artery atherosclerosis, we have investigated the associations between AFLD and NAFLD and coronary artery calcium (CAC).
Design A cross-sectional study was performed in 105 328 Korean adults who attended a health check-up programme. CAC score was assessed using CT, daily alcohol intake was recorded as grams/day and liver fat by ultrasound. Logistic regression model was used to calculate ORs with 95% CIs for prevalent CAC.
Results Both NAFLD and AFLD were positively associated with CAC score. After adjusting for potential confounders, multivariable-adjusted OR (95% CIs) for CAC >0 comparing NAFLD and AFLD to the reference (absence of both excessive alcohol use and fatty liver disease) were 1.10 (95% CI 1.05 to 1.16) and 1.20 (95% CI 1.11 to 1.30), respectively. In post hoc analysis, OR (95% CI) for detectable CAC comparing AFLD to NAFLD was 1.09 (95% CI 1.01 to 1.17). Associations of NAFLD and AFLD with CAC scores were similar in both non-obese and obese individuals without significant interaction by obesity (p for interaction=0.088). After adjusting for homeostasis model assessment of insulin resistance and high-sensitivity C reactive protein, the associations between fatty liver disease and CAC scores remained statistically significant.
Conclusion In this large sample of young and middle-aged individuals, early liver disease due to NAFLD and AFLD were both significantly associated with the presence of coronary artery calcification.
- alcoholic liver disease
- fatty liver
- nonalcoholic steatohepatitis
- cardiovascular disease
Statistics from Altmetric.com
Significance of this study
What is already known on this subject?
Previous studies have reported the association of non-alcoholic fatty liver disease (NAFLD) with increased risk of clinical and subclinical cardiovascular disease (CVD), but the impact of alcoholic fatty liver disease (AFLD) on CVD has received little attention.
A recent study has reported that alcoholic liver disease requiring hospital admission was associated with a greater risk of CVD mortality than NAFLD.
The impact of AFLD on early coronary atherosclerosis is largely unknown.
What are the new findings?
In this large-scale study of 105 328 young and middle-aged adults, an increased risk of prevalent subclinical atherosclerosis was found in NAFLD and in AFLD.
These associations were observed in non-obese and obese individuals and with both low and intermediate/high fibrosis scores.
The association of AFLD and NAFLD with prevalent coronary artery calcium remained significant after adjustment for CVD risk factors.
How might it impact on clinical practice in the foreseeable future?
AFLD and NAFLD are histologically similar liver diseases, and clinicians need to be aware that both liver diseases are similarly associated with increased risk of subclinical early coronary atherosclerosis.
Preventive measures are required to ameliorate CVD risk in both NAFLD and AFLD.
Alcoholic fatty liver disease (AFLD) and non-alcoholic fatty liver disease (NAFLD) are two major types of fatty liver disease (FLD) with similar histological features.1 FLD ranges from simple steatosis to steatohepatitis that can progress to fibrosis, cirrhosis, liver failure or hepatocellular carcinoma. With the global increase in obesity and type 2 diabetes, FLD is becoming one of the most common liver disorders worldwide.1–3 While NAFLD and AFLD are each associated with significant morbidity, impaired health-related quality of life and use of healthcare resources,4 most recent studies have focused on NAFLD and have excluded participants with AFLD.
While many studies have reported the association of NAFLD with increased risk of clinical and subclinical cardiovascular disease (CVD),5 6 the impact of AFLD on CVD as an extrahepatic complication has received little attention,4 7 and there are few studies comparing the association of NAFLD and AFLD with CVD risk.8 9 Recent evidence has suggested that in patients with severe AFLD or severe NAFLD, that necessitated hospital admission or was identified as the specific cause of death, there was a greater risk of CVD mortality with ALD than with NAFLD.10
Coronary artery calcium (CAC) scoring using CT is a useful and reliable marker of early coronary atherosclerosis, and CAC correlates well with total coronary atherosclerotic burden.11 12 CAC scores reflect the long-term impact of CVD risk factors and CAC scores predict future CVD events.11 13
To investigate whether subjects with early liver disease from AFLD and NAFLD have similar (or dissimilar) risk of early coronary atherosclerosis, we have investigated the associations between AFLD and NAFLD, identified in subjects in a large Korean occupational cohort, and the presence of CAC, measured by high-resolution CT. Since it has been shown that even very modest alcohol consumption interacts with obesity to markedly increase the risk of cirrhosis,14 15 we have also evaluated whether the association between FLD and CAC differs by the presence of obesity, severity of hepatic steatosis (assessed by ultrasonography) and degree of hepatic fibrosis (using non-invasive biomarkers for liver fibrosis). For comparison, we have also investigated associations between excessive alcohol consumption (EAC) and CAC scores in the absence of FLD.
The Kangbuk Samsung Health Study (KSHS) is a cohort study of Korean men and women aged 18 years or over who underwent a comprehensive health examination annually or biennially at Kangbuk Samsung Hospital Total Healthcare Centers in Seoul and Suwon, South Korea.8 This study population consisted of a subset of KSHS participants who underwent cardiac CT to measure CAC scores as part of a comprehensive health exam from 2011 to 2017 (n=123 776). CAC scoring has become a common CVD screening test in Korea. Over 80% of participants were employees of various companies and local governmental organisations and their spouses. In South Korea, the Industrial Safety and Health Law requires annual or biennial health screening exams of all employees offered free of charge. The remaining participants were people voluntarily taking screening exams.
For the current cross-sectional study, we excluded 18 448 subjects for the following criteria: missing information on ultrasonography, alcohol consumption and important covariates including body mass index (BMI), glucose, blood pressures (BPs), high-density lipoprotein cholesterol (HDL-C), triglycerides (TG), homeostasis model assessment of insulin resistance (HOMA-IR) and high-sensitivity C reactive protein (hsCRP) (n=9035), history of CVD (n=1605), history of malignancy (n=3261), known liver disease or current use of medications for liver disease or positive serological markers for HBV or HCV (n=5421), history of liver cirrhosis or findings of liver cirrhosis on ultrasound (US) (n=61) and use of steatogenic medications within the past year, such as valproate, amiodarone, methotrexate, tamoxifen or corticosteroids (n=612).2 Some participants met more than one exclusion criteria, leaving 105 328 participants included in the final analysis (figure 1).
Data on demographic characteristics, lifestyle factors, education level, medical history and family history of CVD were collected by standardised, self-administered questionnaires.8 The questionnaire asked about the frequency of alcohol drinking and the amount of alcohol consumed per drinking day recorded in standard units.16 Average alcohol consumption per day was calculated using the frequency and amount of beverages consumed per drinking day. EAC was defined as average alcohol intake ≥30 g/day for men and ≥20 g/day for women.2 Smoking status was categorised as never, former or current smoker. Physical activity was assessed using the validated Korean version of the International Physical Activity Questionnaire short form.17 Participants were classified into inactive, minimally active or health-enhancing physical activity (HEPA). HEPA was defined as physical activity that meets either of two criteria: (1) vigorous intensity activity on three or more days per week accumulating ≥1500 metabolic equivalent (MET) min/week or (2) 7 days with any combination of walking, moderate intensity or vigorous intensity activities achieving at least 3000 MET min/week. History of CVD was defined as participants who reported physician-diagnosed CVD including angina/myocardial infarction and stroke (ischaemic or haemorrhagic). Typical dietary consumption was assessed using a 103-item self-administered food frequency questionnaire designed and validated for use in Korea.18
Height and weight were measured by trained nurses. Obesity was defined as BMI ≥25 kg/m2 according to Asian-specific criteria.19 Waist circumference was measured by trained personnel to the nearest 0.1 cm at the midpoint between the bottom of the rib cage and the top of the iliac crest with the subjects standing, their weight equally distributed on both feet, their arms at their sides and head facing straight forward. We had waist circumference measurements in about 95% (n=99 729) of participants (because one of the two study centres did not start measuring waist circumference until after 2012). BP was measured using an automated oscillometric device (53000, Welch Allyn, New York, USA) by trained nurses while examinees were in a sitting position with the arm supported at heart level. Three readings were recorded, and the average of the second and third readings was used in analysis. Hypertension was defined as systolic BP ≥140 mm Hg, diastolic BP ≥90 mm Hg or the use of antihypertensive medications.
Blood specimens were sampled from the antecubital vein after at least 10 hours of fasting. Blood tests included total cholesterol, low-density lipoprotein cholesterol (LDL-C), HDL-C, TG, aspartate aminotransferase (AST), alanine transaminase (ALT), gamma-glutamyl transferase (GGT), serum albumin, platelet count, glucose, insulin and hsCRP. The HOMA-IR was calculated as fasting insulin (mg/dL) * fasting glucose (mg/dL)/405. Diabetes mellitus was defined as fasting serum glucose ≥126 mg/dL, A1c ≥6.5% (48 mmol/mol) or use of blood glucose-lowering agents.
Ascertainment of FLD and non-invasive fibrosis indices
The diagnosis of fatty liver was based on abdominal US operated by experienced radiologists who were blinded to the aim of the present study. Ultrasonographic diagnosis of fatty liver was determined based on standard criteria, including a diffuse increase of fine echoes in the liver parenchyma compared with kidney or spleen parenchyma, deep beam attenuation and bright vessel walls.20 Interobserver and intraobserver reliability for fatty liver diagnosis was substantial (kappa statistic of 0.74) and excellent (kappa statistic of 0.94), respectively.21 Severity of hepatic steatosis was also recorded as mild, moderate or severe steatosis on sonography. Degree of hepatic steatosis was categorised into mild and moderate/severe steatosis since the number of severe steatosis was small and combined with moderate steatosis. Of 42 701 participants with a diagnosis of fatty liver, 1.6% (n=685) did not have information available on severity of hepatic steatosis.
NAFLD was defined as the presence of fatty liver in the absence of EAC. AFLD was defined as the presence of FLD in the presence of EAC. Other identifiable causes of secondary hepatic steatosis other than alcohol were excluded, as described earlier in the exclusion criteria.
For further FLD categorisation, two fibrosis scoring indices were used. The fibrosis-4 (FIB-4) index was calculated by the following formula: FIB-4=(age (years)×AST (U/L))/(platelet count (×109/L)×ALT (U/L)1/2).22 Cut-off values for low, intermediate and high probability of advanced fibrosis were <1.30, 1.30–<2.67 and ≥2.67, respectively.23 The FIB-4 index has been validated for use in assessing fibrosis stage in patients with both alcoholic liver disease and NAFLD.7 22 For sensitivity analysis, the aspartate transaminase-to-platelet ratio index (APRI) was used as a non-invasive fibrosis index and was calculated by the following formula: APRI=100 × (AST/upper limit of normal)/platelet count (×109/L). Cut-offs for low and high probability of advanced fibrosis were 0.5 and 1.5, respectively.24 25
Measurement of CAC by multidetector CT
CAC was detected with a Lightspeed VCT XTe-64 slice MDCT scanner (GE Healthcare, Tokyo, Japan) in both Seoul and Suwon centres using the same standard scanning protocol of 2.5 mm thickness, 400 ms rotation time, 120 kV tube voltage and 124 mAS (310 mA × 0.4 s) tube current under ECG-gated dose modulation. CAC scores were calculated as previously described by Agatston et al.26 The interobserver reliability and intraobserver reliability for CAC scores were both excellent (intraclass correlation coefficient of 0.99).8 CAC scores were categorised as 0, 1–100 and >100.27
Participants were categorised into four groups: (1) no EAC and no FLD (reference category); (2) EAC and no FLD; (3) NAFLD; and (4) AFLD. Descriptive statistics were used to summarise the characteristics of participants by FLD categories.
To assess the relationship of the presence of CAC with FLD categories, a logistic regression model was used to estimate the ORs with a 95% CIs for the presence of CAC comparing EAC and no FLD, NAFLD and AFLD to the reference category (no EAC and no FLD). We used three models with progressive adjustments: model 1 was initially adjusted for age and sex and then model 2 was further adjusted for study centre (Seoul or Suwon), year of screening examination (1 year categories), BMI, smoking status (never, past, current or unknown), physical activity (inactive, minimally active, HEPA or unknown), educational level (high school graduate or less, community college or university graduate, graduate school or higher and unknown), total calorie intake (in quintile or missing), family history of CVD (yes, no or unknown), diabetes, hypertension, LDL-C and medication for dyslipidaemia (yes, no or unknown). To assess whether the relationship between FLD categories and the presence of CAC is mediated by inflammation or insulin resistance, model 3 was further adjusted for hsCRP and HOMA-IR in addition to the variables included in models 1 and 2. We evaluated whether the associations between FLD categories and the presence of CAC differ by the presence of obesity since the prognostic implications of non-obese FLD remains unclear.28
Additionally, NAFLD and AFLD were further categorised into low and intermediate/high FIB-4 scores according to the degree of fibrosis based on FIB-4 index because fibrosis is the most important histological predictor of liver and non-liver related mortality.29 30 Since few subjects were identified with FLD and high probability of advanced fibrosis, intermediate and high probability of advanced fibrosis were combined. The association of NAFLD and AFLD with the presence of CAC according to degree of fibrosis based on FIB-4 index was evaluated compared with the reference category. The association between fibrosis severity based on APRI and the presence of CAC was also evaluated. We also performed analysis on the association of FLD categories with presence of CAC by degree of hepatic steatosis on ultrasonography. Degree of hepatic steatosis was categorised into mild and moderate/severe steatosis since the number of severe steatosis was small and combined with moderate steatosis. Information on alcohol intake and physician-diagnosed CVD was collected before US and CAC measurements. When we categorised FLD into AFLD and NAFLD, we assumed that persons with already recognised CVD may be more likely to abstain from alcohol as a result of their illness. Thus, in the main analysis, we excluded individuals who reported CVD. We also performed a further analysis in including participants with a history of CVD.
In sensitivity analyses, we also estimated the prevalence ratios and 95% CIs for CAC score 1–100 and >100 for EAC and no FLD, NAFLD and AFLD compared with the reference category (no EAC and no FLD) using participants with CAC 0 as the reference group in multinomial logistic regression models. In another sensitivity analysis, to evaluate the association between FLD categories and CAC as a continuous variable, we used a Tobit regression model for natural log (CAC score +1) with Huber-White estimation of SEs.8 31 Tobit models were used to estimate ratios and 95% CI of CAC score +1 by comparing EAC and no FLD, NAFLD and AFLD to the reference category (no EAC and no FLD). Estimates of the Tobit models were presented as exponentiated Tobit regression coefficients (CAC score ratios) approximately representing the relative CAC score increment comparing EAC and no FLD, NAFLD and AFLD to the reference category (no EAC and no FLD). For example, a CAC ratio of 1.50 is interpreted as a 50% increase in the CAC score for a specific category compared with the reference category.
Subgroup analyses were conducted according to age group (<40 vs ≥40 years of age), sex (men vs women), current smoking (no vs yes), physical activity (no HEPA vs HEPA), HOMA-IR (<2.5 vs ≥2.5) and hs-CRP (<1.0 vs ≥1.0 mg/L). Interactions by subgroups were tested using likelihood ratio tests comparing models with and without multiplicative interaction terms.
Finally, we evaluated a prospective association of NAFLD and AFLD with CAC progression. This analysis included all study participants who had baseline and at least one follow-up cardiac CT to measure CAC scores between 2011 and 2017 (n=23 320). Study participants have been recruited continuously into the study since 2011 and many of the participants recruited in more recent years did not have a second CAC score measurement included in the dataset we used. As a consequence, only 23 320 participants (21.1%) had a follow-up CAC score and were included in the investigation of the prospective association of FLD with CAC progression. We used linear mixed models with random intercepts and random slopes32 to estimate CAC scores and their progression over time adjusting for baseline potential confounders. Since CAC scores were markedly right-skewed, we transformed the scores into loge (CAC +1) as the outcome. Annual progression rate with 95% CIs was estimated while comparing EAC and no FLD, NAFLD and AFLD to the reference category (no EAC and no FLD).
Statistical analysis was performed using STATA V.15.0. All reported p values were two tailed, and comparisons with p<0.05 were considered statistically significant.
The mean age (SD) and mean BMI (SD) of 105 328 participants were 40.8 years (7.8) and 24.4 kg/m2 (3.3), respectively, and 77.5 per cent of participants were male (table 1). The prevalence of EAC and no FLD, NAFLD and AFLD were 9.6%, 32.6% and 7.9%, respectively. EAC with no FLD and AFLD were positively associated with current smoking. NAFLD and AFLD were positively associated with diabetes, hypertension, obesity and higher levels of BMI, BP, total cholesterol, LDL-C, glucose, TG, AST, ALT, HOMA-IR and hsCRP and inversely associated with HDL-C. GGT level was higher in EAC and no FLD, NAFLD and ALFD than in the reference category (no EAC and no FLD) with the highest level of GGT in AFLD. The prevalence of CAC score >0 was 12.3% overall, and its prevalence was progressively higher across FLD categories.
Table 2 shows the relationship between FLD categories and the presence of detectable CAC (>0) overall and in the non-obese and obese groups separately. Both types of FLD, including NAFLD and AFLD, were positively associated with the presence of CAC. After adjusting for age, sex, screening centre, year of screening examination, smoking status, physical activity, educational level, total calorie intake, BMI, family history of CVD, diabetes, hypertension, LDL-C and medication for dyslipidaemia, multivariable-adjusted OR (95% CIs) for detectable CAC comparing EAC with no FLD, NAFLD and AFLD to the reference category were 1.25 (95% CI 1.16–1.35), 1.10 (95% CI 1.05 to 1.16) and 1.20 (95% CI 1.11 to 1.30), respectively. AFLD was associated with higher CAC than NAFLD. In post hoc analysis, OR (95% CI) for detectable CAC comparing AFLD to NAFLD was 1.09 (95% CI 1.01 to 1.17) (p=0.021). In analyses with adjustment for waist circumference instead of BMI, we found similar results (online supplementary table 1).
Supplementary file 1
The associations between FLD categories and the presence of CAC tended to be slightly stronger in the non-obese than in the obese, and although there was a trend towards there being a significant difference by obesity status, these associations did not reach significance (p for interaction=0.088, table 2) even though obese FLD subjects showed unfavourable profiles of metabolic risk factors compared with non-obese FLD subjects (online supplementary table 2). For the non-obese group, multivariable-adjusted OR (95% CIs) for detectable CAC comparing EAC with no FLD, NAFLD, and AFLD to the reference category were 1.31 (95% CI 1.19 to 1.44), 1.10 (95% CI 1.02 to 1.18) and 1.25 (95% CI 1.10 to 1.43), respectively, while for the obese group, corresponding OR (95% CIs) were 1.11 (95% CI 0.98 to 1.27), 1.06 (95% CI 0.98 to 1.15) and 1.14 (95% CI 1.02 to 1.26), respectively.
Similarly, in sensitivity analysis using multinomial regression model, the multivariable-adjusted prevalence ratios comparing EAC with no FLD, NAFLD and AFLD to the reference category were 1.24 (95% CI 1.14 to 1.34), 1.11 (95% CI 1.05 to 1.18) and 1.21 (95% CI 1.11 to 1.31) for CAC score 1–100 and 1.32 (95% CI 1.12 to 1.56), 1.07 (95% CI 0.95 to 1.21) and 1.21 (95% CI 1.03 to 1.43) for CAC score >100, respectively (online supplementary table 3). In sensitivity analysis using Tobit regression model, multivariable-adjusted CAC score ratios (95% CIs) comparing EAC with no FLD, NAFLD and AFLD to the reference category were 1.68 (95% CI 1.42 to 1.98), 1.22 (95% CI 1.09 to 1.37) and 1.54 (95% CI 1.30 to 1.83), respectively (online supplementary table 4).
To explore whether the association between FLD categories and the presence of CAC was mediated by inflammation and insulin resistance, additional analyses adjusting for hsCRP and HOMA-IR were performed (table 2, model 3). The association of both NAFLD and AFLD with the prevalent CAC remained statistically significant. When we performed a further analysis in including participants with a history of CVD, results were similar to those of the analyses excluding participants with a history of CVD (online supplementary table 5).
Table 3 shows the association of FLD categories with presence of CAC according to degree of fibrosis based on FIB-4 index. Compared with the reference category (no EAC and no FLD), multivariable adjusted OR (95% CIs) for detectable CAC in low and intermediate/high FIB-4 among NAFLD cases were 1.09 (95% CI 1.03 to 1.15) and 1.14 (95% CI 1.01 to 1.29), respectively, whereas corresponding OR (95% CIs) among AFLD cases were 1.17 (95% CI 1.08 to 1.27) and 1.37 (95% CI 1.16 to 1.63), respectively. After further adjustment for hsCRP and HOMA-IR, the association between fibrosis scores and presence of CAC remained statistically significant in both NAFLD and AFLD groups. In a sensitivity analysis using the APRI, the associations between FLD and presence of CAC were similarly observed (online supplementary table 6 and 7).
Table 4 shows the association of FLD categories with presence of CAC according to severity of hepatic steatosis on ultrasonography. Compared with the reference category (no EAC and no FLD), multivariable adjusted OR (95% CIs) for detectable CAC in mild and moderate/severe steatosis among NAFLD cases were 1.09 (95% CI 1.03 to 1.15) and 1.14 (95% CI 1.01 to 1.29), respectively, whereas corresponding OR (95% CIs) among AFLD cases were 1.17 (95% CI 1.08 to 1.27) and 1.37 (95% CI 1.16 to 1.63), respectively.
In subgroup analyses other than obesity (online supplementary table 8), the association between FLD categories and CAC scores was stronger in younger individuals (age <40 years) (vs age ≥40 years; p for interaction <0.001). Otherwise, the associations between FLD categories and CAC scores were similar across participant subgroups with no significant interactions by sex (men vs women), current smoking (no vs yes), physical activity (no HEPA vs HEPA), HOMA-IR (<2.5 vs ≥2.5) and hs-CRP (<1.0 vs ≥1.0 mg/L).
Finally, we evaluated a prospective association of NAFLD and AFLD with CAC progression among 23 320 participants with baseline and follow-up cardiac CT (table 5). The median duration of follow-up was 3.0 years (IQR 2.0–4.2, maximum 6.7). The annual rates of CAC progression (95% CI) in no EAC and no FLD, EAC and no FLD, NAFLD, and AFLD were 5.1%, 8.2%, 9.2% and 12.3%, respectively. The multivariable adjusted ratio of progression rates comparing EAC and no FLD, NAFLD and AFLD to the reference category (no EAC and no FLD) were 1.03 (95% CI 1.02 to 1.04), 1.04 (95% CI 1.03 to 1.05) and 1.07 (95% CI 1.06 to 1.08), respectively. These associations were similar in non-obese and obese individuals. Further adjustment for HOMA-IR and hsCRP did not change the result.
In this large-scale study of 113 263 apparently healthy young and middle-aged men and women, both NAFLD and AFLD were significantly associated with a higher risk of prevalent subclinical coronary atherosclerosis compared with the reference (no EAC and no FLD). This association was observed in non-obese individuals, indicating that non-obese NAFLD and AFLD are also associated with a higher risk of atherosclerosis. The risk of subclinical atherosclerosis in FLD was also observed with mild and moderate/severe hepatic steatosis and with both low and higher degrees of fibrosis. Our data suggest that there was a slightly stronger association between AFLD and CAC than between NAFLD and CAC (see table 2, compared with NAFLD, OR (95% CIs) for AFLD and CAC was 1.09 (95% CI 1.01 to 1.17) (p=0.021)).
A slightly stronger risk of atherosclerosis with AFLD than with NAFLD seen in our study might reflect the fact that subjects in our cohort with AFLD have more advanced liver disease than subjects with NAFLD. Such speculation is supported by the recent evidence from a meta-analysis investigating the association between NAFLD and incident CVD.5 In this meta-analysis, the OR (95% CIs) for the association between more severe NAFLD and incident CVD events was 2.58 (95% CI 1.78 to 3.75) compared with 1.64 (95% CI 1.26 2.13) for the association between overall NAFLD and incident CVD. Similarly, a long-term follow-up study of patients with biopsy-proven NAFLD demonstrated an increased risk of CVD death in those with advanced fibrosis.33 In our study, there was limited power to study associations between liver fibrosis and CAC scores in subjects with AFLD and NAFLD, as very few subjects had advanced fibrosis. However, our data using FIB-4 or APRI scores show that there was a trend towards higher risk for prevalent atherosclerosis in subjects with evidence of liver fibrosis.
There are limited studies regarding the impact of AFLD on CVD although multiple studies have reported the association of NAFLD with clinical and subclinical CVD.4 5 7 A recent cohort study reported that in patients with type 2 diabetes who had severe AFLD or severe NAFLD (necessitating hospital admission or causing death), there was a greater risk of CVD mortality with ALD than with NAFLD.10 An earlier cross-sectional study of 265 patients with early liver disease showed higher carotid intima-media thickness in both AFLD and NAFLD patients compared with the reference (no FLD without alcohol history), but this study design was limited by lack of adjustment for confounders.9 Another cross-sectional study of 10 710 participants involved in a health check-up programme demonstrated that the estimated 10-year coronary heart disease risk based on Framingham risk scores was similarly higher in the AFLD and NAFLD groups compared with the no fatty liver group.8 In our study, individuals with AFLD showed a higher prevalence of unhealthy behaviours and CVD risk factors, but whether these behaviours or risk factor mediate an increase in risk of subclinical atherosclerosis is uncertain. Adjustment for those factors attenuated the association between AFLD and CAC scores, but these associations remained significant with AFLD, suggesting that AFLD, like NAFLD, is a metabolic liver disease that is associated with increased risk of CVD risk.
The mechanisms linking hepatic steatosis with atherosclerosis or CVD are not yet fully elucidated. Ectopic accumulation of fat in the liver can be an indicator of lipid overload34 and has been strongly associated with both hepatic and systemic insulin resistance.35 Hepatic steatosis has also been reported to be associated with individual CVD risk factors including diabetes, hypertension, impaired fasting glucose, low HDL-C and hypertriglyceridaemia, in accordance with our findings.36 37 However, the association of hepatic steatosis with subclinical atherosclerosis was not fully explained by those risk factors in our study. Indeed, hepatic steatosis is likely to be implicated in the interplay between insulin resistance, abnormal lipoprotein metabolism, low-grade inflammation, oxidative stress and unfavourable adipokine profiles.36 37 Hepatic steatosis has also been closely associated with altered secretory patterns of hepatokines and proatherogenic factors such as fibrinogen, plasminogen activator inhibitor-1 and other proinflammatory cytokines, all of which promote atherosclerosis.36
In the present study, a positive association between FLD category and prevalent CAC was more evident in individuals younger than 40 years (online supplementary table 6) than in the older age group. The reasons for this finding suggests that FLD may be more important contributor to subclinical atherosclerosis in younger than older populations. This is consistent with increasing prevalence of other CVD risk factors in older age groups. Due to the use of multiple comparisons, chance might be another possible explanation for the observed difference across subgroups.
We note that our study has some limitations. First, fatty liver was determined using US, which is less sensitive (60%–90%) when hepatic fat infiltration is below approximately 30%,38 but is widely used both clinically and in population-based studies due to its non-invasive nature and acceptable degree of diagnostic accuracy for steatosis.38 Additionally, in our study, there was limited power to study associations between liver fibrosis and CAC scores in subjects with AFLD and NAFLD, as very few subjects had evidence of advanced fibrosis from these scores. Second, behavioural factors such as smoking and alcohol use were assessed via a self-administered structured questionnaire used in health check-up programmes in Korea as part of the National Health Insurance plan.39 Measurement errors of these variables might introduce some degree of residual confounding, similar to most epidemiological studies. Finally, our results derived from a sample of relatively healthy young and middle-aged educated Koreans who participated in a health check-up programme and might not be generalisable to other ages and ethnic populations. However, our study population was mainly composed of healthy employees and their spouses without clinically manifest CVD, minimising the possibility of reverse causation and being less likely to be affected by biases related to comorbidities compared with studies conducted in higher risk populations.
In this large sample of young and middle-aged individuals, an increased risk of prevalent subclinical atherosclerosis was found in NAFLD and in AFLD. These associations were observed in non-obese and obese individuals, with mild and moderate/severe steatosis and with both low and intermediate/high fibrosis scores. Our findings suggest that AFLD is also a metabolic liver disease associated with increased risk of subclinical coronary atherosclerosis.
CDB is supported in part by grants from the Southampton National Institute for Health Research Biomedical Research Centre. This work was supported in part by the MRC-KHIDI UK-KOREA PARTNERING AWARD (Medical Research Council MC_PC_16016).
Contributors YC: study concept and design; acquisition of data; interpretation of data; drafting of the manuscript; and critical revision of the manuscript. SR: study concept and design; acquisition of data; analysis and interpretation of data; and critical revision of the manuscript. K-CS and SHW: interpretation of data smf and critical revision of the manuscript. YKC and HS: technical or material support; critical revision of the manuscript; and study supervision. ES: technical or material support; interpretation of data; and study supervision. H-NK,H-SJ, KEY and JA: acquisition of data; interpretation of data; and critical revision of the manuscript. CDB: study concept and design; interpretation of data; and critical revision of the manuscript.
Funding This study was supported by the National Research Foundation of Korea funded by the Ministry of Science, ICT & Future Planning (NRF-2017R1A2B2008401) and by the MRC-KHIDI UK-KOREA PARTNERING AWARD (Medical Research CouncilMC_PC_16016).
Competing interests None declared.
Patient consent Not required.
Ethics approval The study was approved by the Institutional Review Board of Kangbuk Samsung Hospital (IRB No. KBSMC 2018-01-018), which waived the requirement for informed consent as only de-identified data obtained as part of routine health screening exams were used.
Provenance and peer review Not commissioned; externally peer reviewed.
Correction notice This article has been corrected since it published Online First. The funding statement has been updated.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.