PDF
Citation
Print
INTRODUCTION
Steatotic liver disease (SLD) is a leading chronic liver disease worldwide and is commonly associated with metabolic risk factors, including obesity, impaired glucose metabolism, hypertension (HTN), and dyslipidemia [1]. Data from the serial National Health and Nutrition Examination Survey (NHANES) dataset demonstrated that the prevalence of SLD rapidly increased from 20.0% (1988–1994) to 31.9% (2013–2016) fueled by an increase in metabolic diseases [2]. SLD affects the wellbeing of the patient and reduces the quality of life [3]. Furthermore, SLD results in a huge economic burden in the context of increased mortality [4]. The causes of death in patients with SLD vary by the disease state. Hepatic steatosis can result in the development of steatohepatitis, eventually leading to cirrhosis. Patients with SLD and cirrhosis have increased risk of liver-related events while those without cirrhosis have risk of cardiovascular (CV)-related events and non-hepatic cancer [5,6].
Recently, a new term, metabolic dysfunction-associated steatotic liver disease (MASLD), was introduced to improve our understanding of the disease [7]. MASLD was defined as SLD with the presence of at least one of five cardiometabolic risk factors. This definition has lowered the diagnostic threshold compared to that of metabolic dysfunction-associated fatty liver disease (MAFLD), which is diagnosed by the presence of at least two of five cardiometabolic risk factors. Moreover, metabolic alcohol-associated liver disease (MetALD) and “alcoholic liver disease (ALD) with metabolic dysfunction (MD)” have been introduced to describe patients with MASLD who have excessive alcohol consumption (MetALD: 140–350 g/week for females and 210–420 g/week for males; ALD with MD: >350 g/week for females and >420 g/week for males). This modification presents a clear classification of SLD and alcohol-induced liver diseases. However, the clinical outcomes of these patients remain unclear.
Therefore, we aimed to investigate the clinical outcomes of patients with MASLD, MetALD, and ALD with MD in terms of all-cause and cause-specific mortalities.
METHODS
Data source
We used healthcare reimbursement data from the Korean National Health Insurance Service (KNHIS). The KNHIS is a government-run insurer covering the entire South Korean population.
The database contains demographic and health characteristics of the population as well as diagnoses coded according to the 10th edition of the International Classification of Disease (ICD-10) and prescription of medications. Details on the KNHIS datasets have been published earlier [8,9]. We used nationwide death certificate data from the Korean National Statistical Office. This mortality database includes causes and dates of death based on ICD-10 codes.
This study was approved by the Institutional Review Board of the Samsung Medical Center (approval no. 2024-03-104), Seoul, Republic of Korea. Written informed consent was waived because all data was de-identified.
Study population and study design
A flowchart of the study population is displayed in Fig. 1. We screened 1,531,452 participants aged 40 to 79 years who underwent a Korean national health checkup between January 2013 and December 2014. We excluded participants who had been diagnosed with cancer (n=72,006), chronic viral hepatitis (n=86,086), or SLD without cardiometabolic risk factors (n=1,803), or whose data were missing (n=72,564). Ultimately, 1,298,993 participants were included in the analysis.
Measurements and definition of variables
Body mass index (BMI) was calculated as body weight (kg) divided by height (m) squared in meters. Diabetes mellitus (DM) was defined as a fasting blood glucose ≥126 mg/dL, or use of anti-diabetic medication or at least one claim using codes E10–14. HTN was defined as systolic/diastolic blood pressure (BP) ≥130/85 mm Hg or at least one claim per year using codes I10 or I11 or taking antihypertensive medication. Dyslipidemia was defined as triglyceride (TG) ≥150 mg/dL or at least one claim per year using code E78 or for the prescription of a lipid-lowering medication. Chronic kidney disease (CKD) was defined as an estimated glomerular filtration rate (eGFR) ≤60 mL/min/1.73 m2. Regular exercise was evaluated through the questionnaire and was defined as at least one of the following criteria: (1) 3 or more days/week of vigorous activity (causing extreme shortness of breath) for at least 20 min/day or (2) 5 or more days/week of moderate-intensity activity (causing significant shortness of breath) for at least 30 min/day.
Definition of SLD and types of MASLD
Hepatic steatosis was defined as SLD with a fatty liver index (FLI) ≥30. Additionally, FLI was calculated using TG, waist circumference, BMI, and gamma-glutamyl transferase (GGT) levels [10]. FLI is a well-validated fatty liver prediction model that has been validated against liver biopsy, abdominal ultrasonography, or controlled attenuation parameter in multiethnic populations, with a particular focus on Asians, displaying an area under the receiver operating characteristics of 0.87 [11,12].
MASLD was diagnosed when participants had SLD (FLI ≥30) and at least one of the following conditions: obese (BMI ≥23 kg/m2 or waist circumference ≥90 cm in males, and ≥80 cm in females); presence of impaired fasting glucose (IFG) (fasting glucose ≥100); treatment with anti-diabetic medication or diagnosed with diabetes; or evidence of MD, which is defined as at least one of the following: (1) BP ≥130/85 mm Hg or treatment with antihypertensive medication; (2) TG ≥150 mg/dL or treatment with lipid-lowering medication; and (3) high-density lipoprotein cholesterol (HDL-C) ≤40 mg/dL for males and ≤50 mg/dL for females.
The study participants were classified into the “no SLD” group, “pure MASLD” group, “MetALD” group, and “ALD with MD” group (Supplementary Table 1). To calculate alcohol consumption, one standard glass was calculated as 8 g of pure alcohol. Participants with an FLI <30 were classified as the “no SLD” group. Participants with MASLD and low alcohol consumption (<140 g/week for females and <210 g/week for males) were classified as the “pure MASLD” group. Participants with MASLD and high alcohol consumption were classified as the “MetALD” group or “ALD with MD” group according to alcohol consumption (MetALD: 140–350 g/week for females and 210–420 g/week for males; ALD with MD: >350 g/week for females and >420 g/week for males).
Definition of subtypes of MASLD
To determine which MD had a powerful impact on cause-specific mortality, we performed a subgroup analysis of participants with MASLD and only one MD. We divided the patients into five groups: obese MASLD, DM or IFG-MASLD, HTN-MASLD, dyslipidemia-MASLD, and low HDL-C-MASLD. We also analyzed all-cause and all cause-specific mortalities.
Outcomes
The study endpoints were all-cause mortality and cause-specific mortality, including liver-, cancer-, hepatocellular carcinoma (HCC)-, and CV-related mortalities. The cause of death was categorized using ICD-10 codes as liver-related mortality (K70– 76), cancer-related mortality (C00–97), HCC-related mortality (C22), and CV-related mortality (I00–I99).
Sensitivity analysis
To strengthen the robustness and investigate potential discrepancies in outcome, we repeated our main analysis using different SLD assessment tool—the simple nonalcoholic fatty liver disease (NAFLD) score with cut-off point of 8, alongside the FLI with cut-off point of 60. Details regarding the simple NAFLD score calculation can be found in Supplementary Table 2 [13].
Statistical analysis
Continuous variables were presented as means with standard deviation, and categorical variables are presented as frequencies (%). Continuous baseline characteristics were compared using one-way analysis of variance and categorical baseline variables were compared using the chi-square test or Fisher’s exact test. We performed Cox proportional hazard regression to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for all-cause and cause-specific mortalities. In the multivariable-adjusted analyses, model 1 was crude, model 2 was adjusted for age and sex, and model 3 was additionally adjusted for smoking status, income, regular exercise, fasting glucose, eGFR, alanine aminotransferase (ALT), HTN, BMI, prescriptions for dyslipidemia, oral hypoglycemic agents (OHAs) and insulin. Dyslipidemia medications include statins, ezetimibe, fibrates, omega-3 fatty acids and proprotein convertase subtilisin kexin 9 (PCSK9) inhibitors.
The level of significance was a P value <0.05. All analyses were run in R version 4.0.3 software (R Foundation for Statistical Computing, Vienna, Austria) and SAS Enterprise Guide 7.1 for Windows (SAS Institute, Cary, NC, USA).
RESULTS
The baseline characteristics of the study population are displayed in Table 1. Among 1,298,993 participants, no SLD group had 804,858 (61.96%) participants, pure MASLD group consisted of 430,081 (33.11%) participants, MetALD group comprised 51,032 (3.93%) participants, and ALD with MD group had 13,022 (1.00%) participants. The mean age was 54.68 years and the mean BMI was 23.98 kg/m2. Altogether, 625,314 (48.13%) participants were male, and 673,679 (51.86%) were female. Participants with pure MASLD, MetALD, and ALD with MD were more likely to be male, obese, smokers, have high levels of aspartate aminotransferase (AST), ALT, TG, and fasting blood sugar (FBS), and have more comorbidities such as DM and HTN compared to no SLD. Among them, ALD with MD group had the highest proportion of males, current smokers, patients with DM and HTN, and the highest levels of AST, ALT, TG, and FBS.
Risk of all-cause mortality according to MASLD and alcohol consumption
Among the 1,298,993 participants, 57,963 (4.46%) died during the mean follow-up duration of 9.04 years of which 1,282 (0.09%) participants died from liver-related diseases, 19,913 (1.53%) from cancer-related diseases, 2,066 (0.16%) from HCC-related and 10,832 (0.83%) from CV-related diseases. Table 2 displays the results of multivariable Cox regression analysis for the risk of all-cause mortality. Compared to no SLD, the risk of all-cause mortality was significantly increased in pure MASLD (HR, 1.31; 95% CI, 1.29 to 1.34), MetALD (HR, 1.27; 95% CI, 1.22 to 1.32), and ALD with MD (HR, 1.81; 95% CI, 1.68 to 1.93) groups in the crude model. After adjusting for age, sex, smoking status, income, regular exercise, fasting glucose, eGFR, ALT, HTN, BMI, prescriptions for dyslipidemia, OHA and insulin, pure MASLD (HR, 1.28; 95% CI, 1.26 to 1.31), MetALD (HR, 1.38; 95% CI, 1.32 to 1.44), and ALD with MD (HR, 1.80; 95% CI, 1.68 to 1.92) groups demonstrated a significantly higher risk of all-cause mortality compared to the no SLD group (model 3).
Risk of liver-related mortality according to MASLD and alcohol consumption
As demonstrated in Table 3, SLD and alcohol consumption were strongly associated with an increased risk of liver-related mortality. Pure MASLD, MetALD, and ALD with MD groups displayed a significantly higher liver-related mortality compared to the no SLD group with adjusted hazard ratios (aHRs) of 5.64 (95% CI, 4.81 to 6.60), 9.57 (95% CI, 7.82 to 11.70), and 21.68 (95% CI, 17.34 to 27.11), respectively (model 3).
Risk of cancer-related mortality and HCC-related mortality according to MASLD and alcohol consumption
Table 4 displays the association between the presence of MASLD and its types and cancer-related mortality. Furthermore, we also explored specifically focused on HCC-related mortality. Mortality risk due to cancer increased in pure MASLD (aHR, 1.17; 95% CI, 1.13 to 1.22), MetALD (aHR, 1.26; 95% CI, 1.18 to 1.36), and ALD with MD (aHR, 1.33; 95% CI, 1.18 to 1.51) groups, compared to the no SLD group (model 3). Similarly, HCC-related mortality increased in pure MASLD (aHR, 1.97; 95% CI, 1.75 to 2.20), MetALD (aHR, 2.39; 95% CI, 1.99 to 2.87), and ALD with MD (aHR, 2.35; 95% CI, 1.72 to 3.21) groups compared to the no SLD group.
Risk of CV-related mortality according to MASLD and alcohol consumption
Table 5 showed that SLD increased risk of CV-related mortality. However, among individuals with SLD, alcohol consumption and CV-related mortality showed J-shaped association. The pure MASLD group exhibited an increased risk of CV-related mortality with aHR of 1.30 (95% CI, 1.23 to 1.36, model 3) compared to no SLD group. Interestingly, the MetALD group displayed a lower magnitude of increased risk for CV-related mortality compared to the pure MASLD group (aHR, 1.22; 95% CI, 1.10 to 1.36). Conversely, the ALD with MD group demonstrated a higher risk of CV-related mortality (aHR, 1.44; 95% CI, 1.21 to 1.73).
Comparison of all-cause and cause-specific mortalities across MASLD subtypes
The number of participants with MASLD and only one MD was 31,909. As displayed in Supplementary Table 3, the “low HDL-C-MASLD” group exhibited the highest risk of all-cause mortality and all cause-specific mortalities except for CV-related mortality. In CV-related mortality, the “HTN-MASLD” group showed the highest risk. Of note, the “DM or IFG-MASLD” group had the second highest risk of all-cause mortality and cause-specific mortalities including liver-, cancer-, and HCC-related mortalities. Interestingly, the “obese MASLD” group displayed lower all-cause mortality and CV-related mortality compared to the “no SLD” group.
After multivariate adjustments (model 3), aHRs for all-cause mortality were 0.85 (95% CI, 0.78 to 0.94) in the obese MASLD, 1.38 (95% CI, 1.19 to 1.59) in the dyslipidemia-MASLD, 1.78 (95% CI, 1.56 to 2.04) in the HTN-MASLD, 2.21 (95% CI, 1.88 to 2.61) in the DM or IFG-MASLD, and 2.63 (95% CI, 1.75 to 3.96) in the low HDL-C MASLD groups.
During the follow-up, a total of 422 liver-related deaths occurred. The risk of liver-related mortality was aHR 1.44 (95% CI, 0.85 to 2.45) in the obese MASLD, 3.59 (95% CI, 1.99 to 6.47) in the dyslipidemia-MASLD, 7.37 (95% CI, 4.40 to 12.34) in the HTN-MASLD, 12.71 (95% CI, 7.89 to 20.47) in the DM or IFG-MASLD, and 26.46 (95% CI, 10.63 to 65.89) in the low HDL-C MASLD groups.
A total of 11,270 cancer-related deaths occurred over time. The adjusted HRs were 0.95 (95% CI, 0.83 to 1.09) in the obese MASLD, 1.21 (95% CI, 0.96 to 1.53) in the dyslipidemia-MASLD, 1.55 (95% CI, 1.22 to 1.96) in the HTN-MASLD, 1.99 (95% CI, 1.52 to 2.61) in the DM or IFG-MASLD, and 3.09 (95% CI, 1.75 to 5.46) in the low HDL-C MASLD.
Altogether, 5,995 CV-related deaths occurred. The incidence was higher in the “HTN-MASLD” group (aHR, 2.13; 95% CI, 1.58 to 2.88) compared to the “dyslipidemia-MASLD” (aHR, 1.56; 95% CI, 1.07–2.29) and “DM or IFG-MASLD” (aHR, 1.56; 95% CI, 0.84 to 2.61). Obese MASLD group showed lower risk of CV-related mortality (aHR, 0.70; 95% CI, 0.52 to 0.93) compared to the “no SLD group.”
Sensitivity analysis: comparison of all-cause and cause-specific mortalities stratified by simple NAFLD score or FLI with cut-off 60
Consistent with our primary analysis, similar trends in all-cause and cause-specific mortalities were observed when applying the simple NAFLD score and FLI with a cut-off 60 (Supplementary Table 4). Compared to no SLD group, pure MASLD, MetALD, and ALD with MD group showed increased risk of all-cause and cause-specific mortalities. And the MetALD group again displayed a lower risk of CV-related mortality compared to pure MASLD group.
DISCUSSION
As metabolic mechanisms are responsible for the development of SLD, a new nomenclature for MAFLD has been proposed to replace NAFLD. However, MAFLD does not comprise alcohol consumption, and the term ‘fatty’ is considered to be stigmatizing. Therefore, new terms, MASLD, MetALD, and ALD with MD have been proposed [7].
In this study, using KNHIS 2013 to 2014 data sets, we assessed the clinical outcomes of individuals with SLD who were classified as having MASLD, MetALD, or ALD with MD according to their respective definitions.
We demonstrated that individuals with SLD had a higher risk of all-cause and all cause-specific mortalities, including liver-, cancer-, HCC-related, and CV-related mortality than no SLD group. Except CV-related mortality, the risk of SLD was discovered to gradually increase from no SLD to MASLD, and further to MetALD and ALD with MD. Considering CV-related mortality, individuals with ALD with MD exhibited higher risk compared to MASLD, however individuals with MetALD showed lower risk of CV-related mortality compared to MASLD.
Hepatic steatosis is also common in South Korea, where the average BMI is relatively low compared to that in the West. In a recent nationwide cohort study, the prevalence of Korean adults with SLD was 15.5% in 2017, and the prevalence of SLD increased sharply in individuals with MD [14]. Moreover, MD and alcohol consumption, and not just BMI, are involved in the prevalence of SLD [1,15-17]. Hence, the new terminology of MASLD would allow for improved patient identification in individuals having SLD with low BMI.
In the present study, significant differences were observed in all-cause mortality between the “no SLD” group and MASLD, MetALD, and ALD with MD groups unlike in previous studies. Previous studies from the NHANES demonstrated no significant association between ultrasound-diagnosed MAFLD/NAFLD and increased all-cause mortality [18]. Furthermore, a previous study displayed that no significant association was identified between ultrasound-diagnosed MASLD and increased all-cause mortality [19]. However, their study population was relatively smaller than that of our study, and most participants were non-Hispanic White individuals. In this study, we included 430,081 patients with MASLD, most of whom belonged to the Asian population. Our study provides an opportunity for an improved understanding of racial differences in the clinical outcomes of MASLD.
In addition, in our study, patients with SLD displayed significantly increased liver-, cancer-, and HCC-related mortalities, consistent with previous studies [1,20-24]. Moreover, as alcohol consumption increases, liver-, cancer-, and HCC-related mortalities increase synergistically, as displayed in previous studies [25]. Hepatic steatosis results from an overwhelming accumulation of carbohydrates and fatty acids in hepatocytes. These metabolites induce hepatocellular stress and injury, leading to fibrogenesis and genomic instability, which, in turn, lead to HCC [26]. Alcohol is metabolized to acetaldehyde, which is responsible for oxidative stress and induction of cell death [27]. Our findings support the need for noninvasive risk stratification in patients at risk for hepatic injury in terms of metabolic risk factors and/or harmful alcohol consumption [28].
While all SLD groups (pure MASLD, MetALD, and ALD with MD) showed an increased risk of CV-related mortality compared to no SLD group, a J-shaped association was observed between alcohol consumption and CV-related mortality. Interestingly, the MetALD group showed a lower risk of CV-related mortality compared to the pure MASLD group, whereas the risk increased again in the ALD with MD group. These findings are consistent with a meta-analysis of 56 cohort studies, which demonstrated that alcohol consumption of less than 50 g/day was associated with a reduced risk of CV-related mortality, whereas alcohol consumption of more than 50 g/day increased the risk (relative risk, 0.73; 95% CI, 0.64 to 0.83) [8, 21,29-32]. Therefore, our study suggests that alcohol consumption below 50 g/day may be associated with a protective effect against CV-related mortality in individuals with SLD. However, further research is warranted to elucidate the underlying mechanisms.
Our study is important because the MASLD group displayed significant differences in clinical outcomes although the diagnostic threshold was lower than that in the MAFLD definition. Therefore, our results suggest that the change in definition from MAFLD to MASLD may result in the inclusion of a large number of individuals with SLD at a high risk of death, including liver-, cancer-, HCC-, and CV-related deaths.
Moreover, our subgroup analysis revealed the effects of each type of MD on cause-specific mortality. This provides an insight into the prognostic prediction of each MASLD subtype. In the subgroup analysis, the MASLD subgroup with DM or IFG had a significantly high risk of all-cause, liver-, HCC-, and cancer-related mortalities, although CV-related mortality was not statistically significant. These findings suggest that abnormal glucose metabolism is a major driver of mortality in patients with MASLD. Therefore, our findings emphasize the importance of monitoring for SLD in individuals with DM to reduce mortality risk. Additionally, the MASLD subgroup with low HDL-C levels was associated with the highest risk of allcause, cancer-, HCC-, and liver-related mortalities compared to other subtypes of MASLD. A recent meta-analysis of 37 prospective cohort studies demonstrated that low HDL levels were associated with high all-cause and cancer-related mortalities in the general population [33]. Low HDL-C levels have been suggested as a prognostic marker of disease progression and survival in patients with chronic liver disease [34]. However, in our study, the number of participants with “low HDL-C” levels was relatively small. Therefore, further research is needed to interpret these findings. Interestingly, the “obese MASLD” group displayed a lower risk of all-cause and CV-related mortalities; these patients were a subgroup of MASLD, with obesity, but they did not exhibit other cardiometabolic risk factors. Conversely, the “no SLD” group included people with cardiometabolic risk factors but without hepatic steatosis (Supplementary Table 1). Therefore our results should be interpreted with caution. However, numerous studies have observed that metabolically healthy people with obesity still have a higher risk of developing cardiometabolic diseases than metabolically healthy lean people [35]. Therefore, although they do not exhibit a high risk of mortality, the “obese MASLD” group indicates weight-loss interventions.
The strength of this study is that it is a large, representative nationwide cohort study with long-term follow-up. In addition, we adjusted for various variables that could affect mortality including demographic factors, comorbidities, medications such as insulin, OHA. To the best of our knowledge, this study is the first to analyze the association between MASLD and all-cause and cause-specific mortality. Furthermore, we examined all-cause mortality and cause-specific mortality stratified by MASLD subtype to unveil the differential impact of specific MD on cause-specific mortality. Our findings lay groundwork for future research to develop personalized treatment algorithms that manage MD and reduced mortality in individuals with SLD.
However, our study has certain limitations. First, we used FLI and not ultrasonography to define hepatic steatosis. Therefore, the detection rate of hepatic steatosis using ultrasound and clinical outcomes may differ. However, in previous studies, ruling out hepatic steatosis with a cut-off value of 30 based on the FLI displayed a sensitivity of approximately 90% [10-12,36,37]. Furthermore, previous study has demonstrated that the FLI is a more sensitive in detecting SLD compared to ALT or GGT [38]. Given the cost-effectiveness and easy accessibility, the use of noninvasive serum biomarkers can help narrow down patients who need further evaluations using ultrasonography.
Second, we aimed to investigate potential adjustments for sodium-glucose cotransporter 2 (SGLT2) inhibitors and glucagon-like peptide-1 (GLP-1) agonist. However, the only 79 individuals prescribed SGLT2 inhibitors and 114 prescribed GLP-1 agonist. Therefore, we were unable to include them as separate adjustment variables.
In conclusion, the new consensus criteria for SLD emphasize that individuals with MASLD are at an increased risk of all-cause, liver-, cancer-, HCC-, and CV-related mortalities. In addition, patients with MASLD and excessive alcohol consumption (MetALD or ALD with MD group) had a higher risk of all-cause and cause-specific mortalities including liver-, cancer-, and HCC-related mortalities compared to pure MASLD group. However, MetALD group displayed a lower risk of CV-related mortality compared to the pure MASLD group. Conversely, the ALD with MD group exhibited a higher risk. These findings provide a basis for future research on the prognosis of the new consensus criteria for MASLD.
XML Download