Steatotic Liver Disease Subtypes Linked to Sepsis Risk
Key Takeaways
- In a nationwide Korean cohort of 4,389,734 adults, first hospitalization for sepsis was more common in ALD than in non-SLD, at 4% versus 2%.
- On crude incidence, ALD had the highest sepsis burden, MASLD was intermediate, and MetALD was close to non-SLD.
- After multivariable adjustment, MASLD, MetALD, and ALD each remained significantly associated with higher sepsis hospitalization risk than non-SLD, with the largest increase in ALD.
- The adjusted ordering of ALD above MetALD above MASLD was generally maintained across most subgroup and sensitivity analyses, including in the analysis restricted to participants with a CCI score of 0.
Using the Korean National Health Insurance Service (NHIS) and National Health Screening Program (NHSP), investigators conducted a retrospective longitudinal nationwide cohort study of steatotic liver disease subtypes and sepsis that included 4,389,734 adults screened in 2012; follow-up for first hospitalized sepsis ran from 2013 through 2022 after a 1-year washout and 1-year lag, with mean follow-up of about 9 years. Steatotic liver disease (SLD) was defined operationally as a fatty liver index (FLI) of 30 or higher, with metabolic dysfunction-associated steatotic liver disease (MASLD), metabolic dysfunction and alcohol-related liver disease (MetALD), and alcohol-related liver disease (ALD) assigned by cardiometabolic risk factors together with sex-specific alcohol exposure and alcohol-related diagnostic codes; non-SLD was defined by FLI below 30. First hospitalized sepsis events were identified from hospitalization claims using International Classification of Diseases, 10th Revision (ICD-10) codes, and the fully adjusted Cox model accounted for demographic, lifestyle, and comorbidity variables.
During follow-up, 101,520 participants had a first sepsis hospitalization, representing 2% of the cohort. Crude incidence rates per 1,000 person-years were 2.20 for non-SLD, 3.11 for MASLD, 2.25 for MetALD, and 4.49 for ALD. In fully adjusted models, adjusted sepsis risk across MASLD, MetALD, and ALD remained higher than in non-SLD at 1.07 (95% CI 1.06-1.09), 1.11 (1.07-1.15), and 1.53 (1.48-1.59), respectively, so MetALD sat below MASLD on crude incidence but exceeded MASLD after adjustment.
Effect modification analyses varied by sex, age, income, smoking status, and physical activity, but the general ordering of ALD above MetALD above MASLD was maintained across most strata. The same overall pattern also held when investigators applied a stricter FLI threshold, refined the reference and outcome definitions, used competing-risk models, and limited the cohort to participants with a CCI score of 0. With follow-up restricted to 3 years, the MASLD and MetALD associations attenuated while ALD remained elevated.
Interpretation remains tied to operational definitions rather than direct tissue or imaging measures because SLD classification relied on FLI and sepsis ascertainment relied on hospitalized claims coded with ICD-10. Baseline alcohol intake, smoking, and physical activity were self-reported, subtype assignment was fixed at baseline despite possible changes over time, and residual confounding is inherent to an observational analysis of this scale. The findings came from a Korean population using population-specific operational criteria, so they should not be assumed to generalize automatically to North American populations.
The authors concluded that MASLD, MetALD, and ALD were each associated with higher risk of first hospitalization for sepsis than non-SLD, with the strongest association in ALD and an adjusted ordering of ALD, then MetALD, then MASLD. They added that distinguishing SLD subtypes may provide added context for sepsis risk assessment beyond comorbidity burden.
Clinician Questions
How were MASLD, MetALD, and ALD defined in this nationwide sepsis-risk cohort?
In this Korean cohort, steatotic liver disease was operationally defined by a fatty liver index of 30 or higher, and subtype assignment combined at least 1 cardiometabolic risk factor with alcohol exposure categories and alcohol-related diagnostic codes. MASLD referred to lower alcohol exposure in the setting of metabolic dysfunction, MetALD referred to increased alcohol exposure with metabolic dysfunction, ALD referred to higher alcohol exposure or alcohol-related codes, and non-SLD was defined by a fatty liver index below 30.
Why did MetALD show lower crude sepsis incidence than MASLD but higher adjusted risk?
Crude incidence reflects the unadjusted mix of age, sex, lifestyle factors, and comorbid conditions within each group, whereas adjusted models estimate relative risk after accounting for those baseline differences. Participants classified as MetALD were generally younger and had fewer comorbidities than participants with MASLD or ALD, so multivariable adjustment made the underlying sepsis association appear higher than the crude comparison alone suggested.
Which subgroup patterns stood out for sepsis risk across SLD subtypes?
Effect estimates varied by sex, age, income, smoking status, and physical activity. Relatively higher risks were seen among females, middle-aged adults, and lower-income groups within several SLD subtypes, particularly ALD, while smoking and physical-activity patterns were divergent and exploratory rather than proof of causal protection or harm.
How consistent was the sepsis signal when investigators changed follow-up or restricted the cohort?
The ALD-above-MetALD-above-MASLD pattern persisted when investigators used a stricter FLI definition, alternative reference and outcome definitions, competing-risk analyses, and restriction to participants with a Charlson Comorbidity Index score of 0. With only 3 years of follow-up, the MASLD and MetALD associations were no longer statistically significant, while ALD remained elevated.