Subgroups of Steatotic Liver Disease Have Distinct Clinical Phenotypes

Abstract

Background and aims

Steatotic liver disease (SLD) describes a spectrum of liver disease caused by cardiometabolic risk factors (CMRF) and/or alcohol. We aimed to describe the effect of cumulative CMRF and alcohol in subgroups of SLD and compare clinical outcomes.

Methods

Patients from a single centre with biopsy proven SLD were retrospectively included. Patients were classified according to consensus definitions into three subgroups of SLD. The risk of liver-related death or liver transplantation during follow-up was analysed considering competing risks. Outcomes were tabulated for ordinal groups of CMRF and alcohol intake.

Results

726 patients were included: 516 (71%) had metabolic dysfunction-associated steatotic liver disease (MASLD), 85 (12%) MASLD with increased alcohol intake (MetALD), and 125 (17%) had ALD. Patients were followed up for a median of 60.5 months (IQR 29–84.5), during which time 64 (8.8%) patients died, including 23 liver-related deaths. Competing risk regression analysis showed that ALD was associated with higher risk of liver-related death sHR 8.47 (95% CI, 2.26–31.8, P = 0.002) compared to MASLD. The risk of major adverse liver outcomes or liver-related death increased with the number of CMRF and with alcohol, showing a synergistic effect of these factors on patient outcomes.

Conclusions

Amongst patients with SLD, patients with ALD have the greatest risk of adverse liver outcomes. Conversely, liver-related outcomes are less common in MASLD in the short to medium-term. This emphasises the need to identify and treat alcohol misuse as an important risk factor for adverse outcomes.

Graphical abstract

Alcohol consumption and metabolic risk factors, mainly obesity and diabetes, are the leading cause of chronic liver disease in the Western world.¹ Under the umbrella term of steatotic liver disease (SLD), these aetiologies share a similar histopathological spectrum, ranging from hepatic steatosis and steatohepatitis to liver cirrhosis and hepatocellular carcinoma (HCC). The evidence to date confirms that, despite these histological similarities, alcohol related liver disease (ALD) and non-alcoholic fatty liver disease (NAFLD) have distinct clinical courses, liver prognosis, and mortality rates.2, 3, 4

The World Health Organization reported that 5.1% of the global burden of disease and injury is attributable to alcohol, as measured in disability-adjusted life years.⁵ An unhealthy diet and excessive food intake can lead to overweight and obesity. This issue has grown to epidemic proportions, with over 4 million people dying each year as a consequence of overweight or obesity.⁶ Hagstrom et al., in a narrative review, examined the possible synergistic effect of alcohol intake and metabolic risk factors in ALD and NAFLD patients, suggesting that a combination of both might impact the risk of developing progressive liver disease.⁷ As the epidemic of both alcohol use disorder and obesity continue to rise, it is becoming increasingly important to delineate the natural history, outcomes of SLD, and to investigate the combined effect of alcohol and metabolic factors on disease progression.

Following a Delphi process involving multiple stakeholders to reduce stigma and health inequalities, the term “steatotic liver disease” (SLD) was selected as the nomenclature to encompass all causes of liver steatosis.⁸ SLD is an umbrella for three main subcategories: metabolic dysfunction-associated steatotic liver disease (MASLD), MASLD with increasing alcohol intake, referred to as MetALD and ALD. A novel category, MetALD, was introduced to capture individuals whose alcohol consumption exceeded the previously defined thresholds of NAFLD but had remained unclassified (140–350 g/week for women and 210–420 g/week for men).⁸ With the introduction of MetALD, there is an opportunity to study and learn about patients who have both metabolic and alcohol-related risk factors. Limited data are available on all-cause mortality and liver-related mortality of SLD subgroups, especially the MetALD subgroup. This emphasises the importance of studying the interaction between cardiometabolic risk factors (CMRF) and alcohol, and the need for better understanding of outcomes in SLD to gain better understanding of disease progression and management.

Methods

Study Design and Participants

This retrospective observational cohort study included adult patients (>16 years) undergoing a liver biopsy for staging of SLD disease in a single large tertiary centre (St James's University Hospital, Leeds) in the United Kingdom between 1999 and 2023. We included patients who are over 16 years old and not in education, as these patients are cared for in adult Hepatology services in the UK. Patients underwent liver biopsy for diagnostic reasons and/or staging of the disease as per local routine practice. Patients were included in the study if they had SLD confirmed by histopathology and available clinical/anthropometric information. Patients with liver disease were managed in accordance with established national and international guidelines. Patients with serological or histological evidence of other liver diseases, including viral hepatitis, autoimmune liver disease (autoimmune hepatitis, primary biliary cholangitis, primary sclerosing cholangitis), haemochromatosis, Wilson's disease, alpha-one antitrypsin deficiency, were excluded. Patients with decompensated chronic liver disease–as defined below–at the time of the index biopsy were removed from the analysis to allow identification of only incident cases of decompensated chronic liver disease. Patients were followed up from the point of biopsy until August 2024 or until death occurred. Follow-up information was collected by reviewing electronic patient records.

Assessment of Steatotic Liver Disease and Data Collection

Liver biopsies were reviewed by a team of specialist liver histopathologists and were reported according to the Kleiner-Brunt system.⁹ Information from the liver biopsy was recorded, including the presence of steatosis and the presence and severity of fibrosis (F0-4). We recorded the presence of CMRF in accordance with the recent multi-society statement on SLD: 1) body mass index (BMI) ≥25 kg/m²; 2) fasting serum glucose ≥5.6 mmol/L OR HbA1c ≥ 39 mmol/L OR type 2 diabetes OR treatment for type 2 diabetes; 3) blood pressure ≥130/85 mmHg OR specific antihypertensive drug treatment; 4) plasma triglycerides ≥1.70 mmol/L OR plasma HDL-cholesterol ≤1.0 mmol/L (males) and ≤1.3 mmol/L (females) OR lipid lowering treatment.⁸ Alcohol history was recorded by clinicians (within the year of the liver biopsy performed), and where the data were insufficient, this was confirmed via access to alcohol screening in primary care records (recorded as units/week or AUDIT-C). Alcohol use was reported as average units/week or standard drinks consumed on average in a week (number of glasses of wine, pints of beer/lager, or measures of spirit), converted to units/week. For data analysis purpose, we converted this into grams, where one UK unit is equivalent to 8 g of pure ethanol.¹⁰ Laboratory and transient elastography data were noted. Missing data are reported in the results section.

Outcome Assessments

Patients were classified according to the new SLD nomenclature (Supplementary Figure 1). Outcome data were recorded with regards to overall survival from the time of biopsy, cause of death (liver-related death, non-liver related death) based on a death certificate or information in patient electronic records, and the incidence of major adverse liver outcomes (MALO), including decompensation (defined as incident hepatic encephalopathy [HE], ascites, or portal hypertensive [variceal] haemorrhage), HCC or liver transplantation (LT). The presence of ascites was recorded based on clinical or ultrasound assessment. Variceal haemorrhage was confirmed by endoscopy. Hepatic encephalopathy was assessed routinely during clinical assessment.

Statistical Analysis

Simple descriptive statistics were used to describe the cohort. Categorical variables are reported as counts and percentages (%). Continuous variables are shown as median and interquartile range (IQR). Outcomes were analysed with Kaplan–Meier analysis to describe the risk of all-cause mortality. Competing risk analysis was used to describe the risk of liver-related death or transplant (Supplementary Figure 2), taking into account death from non-liver causes and controlling for baseline fibrosis stage and age. A similar analysis was conducted to describe the risk of incident MALO, with death as a competing event. These factors were defined a priori. For patients who did not have available records on cause of death, death was assumed to be liver-related death if patient had experienced an episode of decompensation in the three months prior to death. To support competing risk analysis, a pre-planned sensitivity analysis was done using Cox proportional hazard models to investigate the same outcomes. Only events that occurred after the index liver biopsy were included in the analysis. A subgroup analysis was performed based on histological fibrosis score.

To examine the effects of CMRF and alcohol intake on fibrosis and clinical outcomes, we followed the methodology of Marti-Aguado et al. to produce heatmaps illustrating the risk of fibrosis/outcomes, and used ordinal numbers of CMRF and alcohol thresholds in univariable and multivariable analysis.¹¹ For the purpose of heatmaps, CMRF were grouped as the presence of one, two, or three or more CMRF. Alcohol thresholds were based on the methods used by Marti-Aguado et al.¹¹ to examine fibrosis in MASLD and MetALD cohorts – very low MASLD (0–40 g/week), low MASLD (40–90 g/week), moderate MASLD (90–130 g/week in women; 90–200 g/week in men), and MetALD (130–350 g/week in women; 200–420 g/week in men).¹¹ To examine the effects of CMRF and alcohol intake on clinical outcomes, we used alcohol thresholds as defined in a multisociety Delphi consensus statement and alcohol use was categorised into four levels–low-MASLD (<140 g/week in women, <210 g/week in men), MASLD-predominant MetALD (140–245 g/week in women, 210–315 g/week in men), ALD-predominant MetALD (245–350 g/week in women, 315–420 g/week in men), and high-ALD (>350 g/week in women, >420 g/week in men).⁸ Statistical analyses were performed in R Software v 4.0.3 using the following packages: finalfit,¹² gtsummary,¹³ tidyverse,¹⁴ tidycmprsk,¹⁵ cmprsk,¹⁶ broom,¹⁷ patchwork,¹⁸ devtools,¹⁹ and survival.²⁰

Research Ethics

This research used routinely collected data without additional procedures beyond standard medical care. This research is exempt from specific research ethics approval, as we followed and adhered to the UK Health Research Authority guidance regarding use of anonymous routine data, and no additional procedures were necessary for this study. The research was conducted in accordance with both the Declarations of Helsinki and Istanbul.

Results

Patient Characteristics

In total, our cohort comprised 784 patients. Fifty-eight patients who were decompensated at the time of the index biopsy were excluded from further analysis, leaving 726 patients (Supplementary Figure 1). Most of the cohort had MASLD (516, 71%), 85 had MetALD (12%), and 125 had ALD (17%). Considering previous classifications of NAFLD and ALD, only a minority of patients were reclassified (85 of 726, 12%). A greater proportion of patients with ALD were reclassified (46 of 226, 20.4%) compared to patients with previous NAFLD (39 of 558, 7%). All newly reclassified patients were given a new diagnosis of MetALD. The median age was 53 (IQR 43–61) years, 63% were male, and 99 (14%) had cirrhosis (F4 stage) on biopsy. The characteristics of the cohort are shown in Table 1. Median follow up was 60.5 months (IQR 29.0–84.5) during which period 64 (8.8%) patients died, and 51 (7%) patients experienced MALO (decompensation, liver transplant, HCC). One patient had all three: decompensation, HCC, and underwent liver transplantation, six patients had HCC and decompensation.

Table 1.

Characteristics of the Whole Cohort Excluding Patients Who Were Decompensated at Baseline.

Characteristics	MASLD	MetALD	ALD
Number of patients	516	85	125
Age (years)	53.0 (43–61)	53.0 (47–61)	55.0 (48–62)
Male sex, n (%)	313 (61%)	63 (74%)	81 (65%)
BMI (kg/m2)	34.48 (31.02–38.95)	32.78 (30.0–35.91)	30.90 (27.41–35.0)a
Diabetes mellitus, n (%)	237 (46%)	31 (36%)	26 (21%)
Hypercholesterolaemia, n (%)	224 (44%)	32 (38%)	33 (26%)
Hypertension, n (%)	228 (44%)	46 (54%)	55 (44%)
Fibrosis stage	2 (1–3)	2 (1–3)	2 (1–3)
F0, n (%)	119 (23%)	6 (7.1%)	18 (14%)
F1, n (%)	144 (28%)	20 (24%)	22 (18%)
F2, n (%)	75 (14%)	17 (20%)	22 (18%)
F3, n (%)	134 (27%)	24 (28%)	26 (21%)
F4, n (%)	44 (8.9%)	18 (21%)	37 (30%)

Continues values are summarised as median (interquartile range) and categorical number as percentage. BMI – body mass index; MASLD – metabolic dysfunction-associated steatotic liver disease; MetALD – MASLD and increased alcohol intake; ALD – alcohol related liver disease.

^aMissing data for 1 (0.8%) patient.

Prevalence of Cardio-metabolic Risk Factors and Alcohol Intake

The distribution of CMRF was similar across SLD subgroups. The median number of CMRF was 2 (IQR 1–3) in patients with MASLD, MetALD, and ALD. One CMRF was present in 188 (26%) patients, two were present in 218 (30%) patients, three in 236 (33%) patients and only 15 (2.1%) presented with no CMRF (all these patients had ALD). Four CMRF were present in 69 patients (9.5%). Median alcohol intake in each subgroup varied from 0 g/week (0–37) in the group with MASLD, 270 g/week (240–324) in the group with MetALD, and 630 g/week (540–1080) in patients with ALD.

Morbidity After Biopsy

Incident MALO occurred in 51 (7%) individual patients. New decompensation during follow-up was observed in 43 (5.9%) patients, liver transplant in 3 (0.4%) patients, and HCC in 10 (1.4%) patients. The five-year cumulative incidence of MALO was 1.8% (0.80%, 3.6%) in MASLD, 7.1% (2.6%, 15%) in MetALD and 25% (17%, 33%) in ALD (Supplementary Figure 4B). The most frequent decompensating event was ascites (29 patients), followed by hepatic encephalopathy (14 patients) and variceal bleeding (11 patients) (Supplementary Table 1). Patients in the ALD group had a greater incidence of MALO (35 individuals, 28%) compared to MetALD (7 individuals, 8.2%) and MASLD (9 individuals, 1.7%), P < 0.001.

Mortality After Biopsy

During follow-up, 64 (8.8%) patients died. All-cause mortality was higher in patients with ALD than MetALD or MASLD (log rank test = 46.26, P < 0.001) (Supplementary Figure 3). Liver-related deaths accounted for 37.5% of observed deaths (n = 24) and deaths unrelated to liver disease for 62.5% (n = 40). At one year after index biopsy, no liver-related deaths had occurred in patients with MASLD or MetALD, the cumulative incidence of liver-related deaths was 4.8% (2.0%, 9.6%) in patients with ALD. At five years, the cumulative incidence of liver-related deaths was 1.2% (0.39%, 2.9%) in MASLD, 1.4% (0.12%, 6.8%) in MetALD, and 14% (8.3%, 21%) in ALD (Table 2 and Supplementary Figure 4A).

Table 2.

Cumulative Incidence of Liver-related Deaths and Major Adverse Liver Outcome at One Year, Five Years, and Ten Years After Index Biopsy.

Cumulative incidence of liver-related death or transplantation
	n	1 year	5 years	10 years
MASLD	516	0	1.2% (0.39%, 2.9%)	2.2% (0.66%, 5.6%)
MetALD	85	0	1.4% (0.12%, 6.8%)	4% (0.65%, 13%)
ALD	125	4.8% (2.0%, 9.6%)	14% (8.3%, 21%)	25% (16%, 36%)
Cumulative incidence of MALO
MASLD	516	0.21% (0.02%, 1.1%)	1.8% (0.80%, 3.6%)	4.1% (1.4%, 9.1%)
MetALD	85	1.2% (0.10%, 5.7%)	7.1% (2.6%, 15%)	16% (5.3%, 32%)
ALD	125	7.2% (3.5%, 13%)	25% (17%, 33%)	35% (24%, 45%)

MASLD – metabolic dysfunction-associated steatotic liver disease; ALD – alcohol related liver disease, MetALD – MASLD and increased alcohol intake, MALO – major adverse liver outcome.

Univariable competing risk analysis with Fine and Gray models showed that subgroup of SLD was significantly associated with the risk of liver-related death: sHR 17.7 (95% CI, 5.11–61.7, P < 0.001) in ALD, compared to MASLD (Table 3). After controlling for fibrosis stage, age, sex, diabetes mellitus, hypertension, and smoking the ALD cohort remained at significantly higher risk of liver-related death: compared to MASLD, the sHR for ALD was 8.47 (95% CI, 2.26–31.8, P = 0.002) (Table 3). The MetALD cohort had a higher risk of liver-related deaths compared to MASLD, but this was not statistically significant. These findings were confirmed in sensitivity analyses with Cox proportional hazard ratios (Supplementary Model 1 and Model 2).

Table 3.

Risk of Steatotic Liver Disease Mortality.

	Univariable			Multivariable
	sHR	95% CI	P value	sHR	95% CI	P value
MASLD	(ref)	–	–	(ref)	–	–
MetALD	4.78	0.98–23.2	0.053	3.27	0.66–16.2	0.15
ALD	17.7	5.11–61.7	<0.001	8.47	2.26–31.8	0.002
F0–F2	(ref)	–	–	(ref)	–	–
F3	5.25	1.58–17.4	0.007	5.95	1.78–20.0	0.004
F4	15.8	5.36–46.4	<0.001	11.0	3.26–37.1	<0.001
Age	1.02	0.99–1.06	0.19	1.02	0.97–1.06	0.48
Sex (male)	0.98	0.43–2.27	0.97	1.30	0.55–3.07	0.56
Smoking (never)	(ref)	–	–	(ref)	–	–
Smoking (current)	1.62	0.60–4.37	0.34	0.99	0.33–2.97	0.99
Smoking(ex-smoker)	1.32	0.51–3.40	0.56	1.37	0.48–3.92	0.56
DM (present)	0.36	0.13–0.97	0.043	0.17	1.48–1.41	0.18
HTN (present)	0.54	0.24–1.24	0.15	0.48	0.20–1.13	0.092

MASLD – metabolic dysfunction-associated steatotic liver disease; MetALD – MASLD and increased alcohol intake; ALD – alcohol related liver disease, F – fibrosis stage, DM – diabetes mellitus, HTN – hypertension.

Further, we assessed the risk of liver disease mortality only in patients with histological fibrosis score 3–4 and after controlling for age, sex, diabetes mellitus, hypertension, and smoking we found the sHR of 7.64 (95% CI, 1.96–29.8, P = 0.003) in ALD subgroup compared to MASLD subgroup (Supplementary Table 2). Analysis for patients with histological fibrosis score of 0–2 was not performed due to low number of liver-related outcomes.

We looked at patient overall survival and performed a subgroup analysis using Cox proportional hazard based on histological fibrosis stage. Multivariable analysis for SLD patients with histological fibrosis stage 0–2, controlling for age, sex, and smoking status, did not show a statistically significant difference in overall survival between MASLD, MetALD, and ALD subgroups (Supplementary Model 3). However, when we performed the same analysis for patients with histological fibrosis stage 3–4, the hazard ratio was higher (6.88, 95% CI, 3.19–14.85, P < 0.001) in ALD compared to MASLD subgroup (Supplementary Model 4).

Influence of Cardiometabolic Risk Factors and Alcohol Intake on Outcomes

Analysis of the joint effect of CMRF and alcohol intake on significant fibrosis confirmed that the presence of significant fibrosis increased with increasing CMRF and increasing alcohol intake, consistent with results from Marti-Aguado et al. (Supplementary Figure 5).¹¹ To further explore the interaction between risk factors in different subgroups of SLD, we considered clinical outcomes using alcohol thresholds based on the SLD nomenclature statement (defined in the methods section), which includes the ALD group. In general, the risk of MALO or liver-related death increased with the number of CMRF and with alcohol, showing a synergistic effect of these factors on patient outcomes. However, when considering ALD as well, the rate of adverse outcomes was dominated by alcohol intake, confirming a distinctly worse prognosis in this group as well as the deleterious effect of alcohol across all subgroups of SLD. Multivariable analysis showed MASLD-predominant MetALD and high-ALD alcohol consumption were independently associated with MALO/liver-related death but increasing numbers of CMRF in isolation was not linked to an increased risk of liver related deaths or MALO (Table 4). We have illustrated this using a heatmap (Figure 1).

Table 4.

Univariable and Multivariable Analysis for Identifying Factors Associated With Major Adverse Liver Outcomes And/or Liver-Related Mortality.

Parameter	Univariable		Multivariable
	OR (95% CI)	P value	OR (95% CI)	P value
Number of CMRF
1	(ref)	–	(ref)	–
2	0.28 (0.11–0.65)	P=0.005	0.27 (0.09–0.70)	P=0.009
3	0.53 (0.26–1.08)	P = 0.082	0.49 (0.20–1.14)	P = 0.101
4	0.51 (0.14–1.41)	P = 0.233	1.0 (0.25–3.29))	P = 0.996
Alcohol consumption
Low MASLD	(ref)	–	(ref)	–
MASLD-predominant MetALD	5.12 (1.34–16.42)	P=0.009	3.77 (0.95–12.69))	P=0.040
ALD-predominant MetALD	4.83 (1.04–17.03)	P=0.022	3.69 (0.75–13.97)	P = 0.071
High-ALD	23.05 (11.18–52.49)	P < 0.001	16.49 (7.25–40.93)	P < 0.001
Fibrosis stage
F0-2	(ref)	–	(ref)	–
F3	4.87 (2.18–11.63)	P < 0.001	4.56 (1.83–11.94)	P=0.001
F4	17.14 (7.99–40.09)	P < 0.001	9.54 (3.97–24.55)	P < 0.001

CMRF – cardiometabolic risk factors, MASLD – metabolic dysfunction associated steatotic liver disease, MetALD – MASLD and increased alcohol intake, ALD – alcohol related liver disease. Low MASLD alcohol consumption was defined as <140 g/week in women and <210 g/week in men; MASLD-predominant MetALD 140–245 g/week in women and 210–315 g/week in men; ALD-predominant MetALD 245–350 g/week in women and 315–420 g/week in men, high-ALD >350 g/week in women and >420 g/week in men. Multivariable logistic regression analysis was performed for major adverse liver event and/or liver related mortality.

Figure 1.

Impact of the interactions between cardiometabolic risk factors and alcohol intake on the prevalence of major adverse liver outcomes and/or liver-related deaths in MASLD, MetALD, and ALD cohorts. MASLD – metabolic dysfunction associated steatotic liver disease, MetALD – MASLD and increased alcohol intake, ALD – alcohol-related liver disease. Low MASLD alcohol consumption was defined as <140 g/week in women and <210 g/week in men; MASLD-predominant MetALD 140–245 g/week in women and 210–315 g/week in men; ALD-predominant MetALD 245–350 g/week in women, and 315–420 g/week in men, high-ALD >350 g/week in women and >420 g/week in men. Figures in the boxes represent proportion (%).

Discussion

In this study we present data of biopsy proven SLD that was reclassified according to the Delphi consensus and define outcomes according to the novel subgroups. In our cohort, reclassification was necessary in a minority of patients (n = 85). All reclassifications occurred in patients assigned a new diagnosis of MetALD and originated from both the previously assigned NAFLD and previously assigned ALD groups. Ninety-three percent of patients previously classified as NAFLD met the new MASLD definition. This is lower than reported in a recent population-based study from Younessi et al., which observed 99.8% of patients with NAFLD met the MASLD definition.²¹ The difference likely arises from our consideration of alcohol consumption, resulting in a proportion of our NAFLD cohort reclassified as MetALD.

Our study addresses an important gap in our understanding of the newly defined subgroups, as we show that outcomes, in terms of liver-related morbidity and mortality are worse in alcohol-related liver disease. We also show that the joint effect of CMRF and alcohol increase the risk of significant fibrosis and increase the risk of major liver-related outcomes and liver related mortality. Our data can help clinicians and researchers understand the interaction between cardiometabolic risk factors and alcohol, with reference to clinically relevant outcomes.

The study has several strengths. Firstly, the cohort was defined by liver histology, with all biopsies reviewed by experienced histopathologists, confirming the presence of steatosis and excluding co-factors from other liver diseases, which provides additional information to other similar studies that have used patient databases with only clinically proven diagnoses of MASLD, MetALD, and ALD. Furthermore, our electronic patient records encompasses both primary and secondary care that allowed us to capture all relevant clinical outcomes. Finally, we have a median follow-up exceeding 5 years (60.5 months) ensuring a comprehensive timeframe to capture clinical events relevant for prognostication.

Our cohort study defines outcomes in three SLD subgroups, with a special interest in a newly defined MetALD subgroup. The population is reflective of a ‘real-world’ group of patients who underwent liver biopsy during their clinical care of NAFLD or ALD. Similarly, a prospective study by Israelsen et al. looked at 446 patients with a history of excessive alcohol intake only and showed that the risk of decompensation as well as overall mortality increased in a stepwise manner from MASLD through MetALD to ALD.²² Li et al. discussed the new nomenclature and outcomes in each group. They analysed 7980 participants from NHANES III and found that patients with MetALD and ALD with metabolic dysfunction had a poorer prognosis.²³

Whilst MetALD patients accounted for only 12% of our SLD cohort, we clearly demonstrated that the 3 subgroups (MASLD, MetALD, and ALD) display distinct clinical outcomes. There is a significant difference in the rates of liver-related deaths and MALO between the subgroups. Our results corroborate previous work on the cumulative effect of alcohol on liver disease progression for patients with NAFLD.^24,25 We observed an increasing prevalence of cirrhosis, MALO, and liver-related deaths in ALD subgroup compared to MASLD and MetALD subgroups. A recent meta-analysis reported ‘moderate’ alcohol intake increased the risk for advanced fibrosis, with a pooled OR of 1.56 (95% CI, 1.08–2.26) and HR 1.39 (95% CI, 1.22–1.57), but defining the cumulative risk of moderate alcohol intake on progression of liver disease in patients with MASLD has previously been hampered by the wide heterogeneity of studies.²⁶

A further noteworthy observation from this cohort is the low number of outcome events observed in the MASLD subgroup. Specifically, the cumulative incidence of MALO and liver-related deaths at 5 years was 1.8% and 1.2% respectively. This aligns with previous studies describing the natural history of NAFLD and long term prognosis.^4,27

Conversely, we observed a significant burden of liver-related morbidity and mortality in the ALD group. The cumulative incidence of liver related death at 5 years was 1.4% and 14% for MetALD and ALD respectively. While, noteworthy, this is somewhat less significant than a recently published systematic review on global outcomes in ALD, which reported liver-related mortality of 21.6%.² In the multivariable analysis, the sHR of liver-related death when compared to MASLD was 8.47 (95% CI, 2.26–31.8, P = 0.002) for ALD. This again highlights the significant role of alcohol in the development and prognosis of liver disease and the necessity for dedicated attention to this subgroup of SLD. The sensitivity analysis of Cox proportionate hazard clearly states that ALD patients with fibrosis score 3–4 are at higher risk of mortality compared to MASLD patients; however, there was no statistically significant difference between these two groups in patients with fibrosis score 0–2. This highlights the need of addressing the risk factors in early fibrosis stage to stop or delay the progression to advanced liver fibrosis.

The UK Office for Health Improvement and Disparities recently published an update on liver disease profile, showing that the rate of premature deaths from liver disease increased by 42% from 2001 to 2023, and the number of premature deaths from alcohol-related liver disease increased by 3.6% in 2023 compared to 2022.²⁸ Considerable changes in drinking culture are required to try and reduce the disease burden from alcohol related liver conditions. Diaz et al. clearly showed that stricter alcohol-related public health policies correlate with declines in alcohol use disorder prevalence and mortality from ALD.²⁹

We were able to reproduce the results as per Marti-Aguado et al., using the same alcohol thresholds (very low MASLD, low MASLD, moderate MASLD and MetALD), and confirmed the joint effect of CMRF and alcohol in increasing the risk of significant fibrosis, confirming that our data were comparable to their cohort.¹¹ In addition, we demonstrated that the combined effect of CMRF and alcohol consumption increases the risk of major liver-related outcomes and liver-related mortality; however these clinical outcomes are predominantly driven by alcohol consumption. The results suggest that CMRF and alcohol synergise to increase the risk of liver histological severity. This confirms the message presented by the World Health Organization: “No level of alcohol consumption is safe for our health”.³⁰ Similarly, Anderson et al. clearly state that alcohol is a toxic, psychoactive substance that has been linked to multiple cancers, and increasing levels of use can lead to raised morbidity and mortality.³¹ Based on these findings, there is a need for future research exploring the interaction between CMRF and dynamic alcohol consumption patterns. This is especially relevant for the MetALD group, where both metabolic risk factors and alcohol consumption interplay with liver disease progression.

The study does have some limitations inherent to the retrospective observation design. Firstly, there is likely to be selection bias in patients undergoing liver biopsy with older, co-morbid patients less likely to be represented within the cohort. We also acknowledge that ALD patients undergoing the biopsy may have more advanced disease, potentially leading to some overestimation of adverse outcomes. We cannot account for potential variations in alcohol intake over time, as consumption was recorded at baseline at the time of the liver biopsy and not during follow-up. While self-reported alcohol intake provides valuable information, objective biomarkers such as phosphatidylethanol (PEth) or the Alcoholic Non-Invasive (ANI) index could have enhanced the accuracy of alcohol exposure assessment, particularly in detecting underreporting. Future similar studies incorporating PEth and ANI may provide a more comprehensive understanding of the impact of alcohol on SLD. It can be difficult to record changing alcohol intake in retrospective studies, and longitudinal alcohol use might be more suited to prospective studies where alcohol intake can be accurately recorded over time. We noted data from Wong et al. regarding longitudinal alcohol use, where people with a baseline low-risk AUDIT-C with subsequent increase in alcohol intake had higher risk of incident cirrhosis, whereas those who decreased their alcohol intake during the follow-up experienced a reduction in long-term risk of cirrhosis by 39% (HR 0.61; 95% CI, 0.45–0.83, P < 0.01). They also reported that concurrent high-risk use of alcohol is associated with 43% higher risk of cirrhosis.³² Zhang et al. presented data on the associations between dynamic changes in alcohol use and MASLD, where sustained moderate alcohol intake was associated with the risk of MASLD compared to sustained non-drinking. It also showed a trend towards a higher risk of MASLD in people with increased alcohol consumptions.³³

It remains unclear how clinicians should address alcohol use, as it can change dynamically, and it is unclear how we should assess patients who used alcohol excessively in the past but subsequently developed CMRF. CMRF themselves are dynamic in nature and can change over a period and influence the stage of SLD. The new SLD nomenclature has not addressed this issue, and we believe that historical alcohol use should be integrated into the nomenclature in the future. The increasing prevalence of CMRF has posed and interest in cohort of patients with ALD and CMRF. There is limited data on ALD and CMRF interplay and impact this has on disease outcomes. The current study did not focus on each of the CMRF in ALD cohort due to the limited number of included patients. However, the findings from this study could promote further research to better understand the impact each CMRF has on the cohort.

The introduction of MetALD provides an opportunity to better define impact of alcohol consumption within the context of MASLD. Our study provides information on alcohol and CMRF synergy that should be addressed when managing patients with SLD, especially in MetALD and ALD subgroups. The distinct clinical outcomes observed highlight the need for earlier interventions to screen and treat alcohol misuse to prevent liver disease and its progression to more advanced stages. The findings from this study will promote more research to better understand the interplay between CMRF and alcohol, as well as promote more research on longitudinal alcohol use.

CRedit authorship contribution statement

All authors contributed to the study conception and design. Richard Parker had the idea for this study and curated databases. Katrina Pekarska and Laura Burke did analyses with support from R. Parker and Ian Rowe. K. Pekarska and L. Burke wrote the initial draft of the manuscript. R. Parker and I. Rowe supported data analysis and interpretation and finalised the draft manuscript. Final version was approved by K. Pekarska, L. Burke, R. Parker and I. Rowe.

Funding

No financial support was received to complete this study or write this manuscript.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Ian Rowe reports a relationship with Roche that includes: consulting or advisory. Ian Rowe, Richard Parker reports a relationship with Norgine Ltd that includes: consulting or advisory. Ian Rowe reports a relationship with Boehringer Ingelheim Ltd that includes: consulting or advisory. Richard Parker reports a relationship with Durect Corporation that includes: consulting or advisory. Richard Parker reports a relationship with Novo Nordisk Inc that includes: consulting or advisory. Richard Parker reports a relationship with Orphalan Limited that includes: travel reimbursement. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Supplementary data

The following is the Supplementary data to this article: