Non-invasive tools for liver steatosis and steatohepatitis predict incidence of diabetes, cardiovascular disease and mortality 20 years later: The ATTICA cohort study (2002-2022).

BACKGROUND & AIMS: Non-alcoholic fatty liver disease (NAFLD) or, as recently renamed, metabolic dysfunction-associated steatotic liver disease (MASLD), has common metabolic pathways with diabetes and cardiovascular disease (CVD). Non-invasive tools (NITs) for liver steatosis and steatohepatitis (MASH) were studied as potential predictors of diabetes, cardiovascular disease (CVD) and mortality over a 20-year period.

METHODS

In 2001-02, 3042 individuals from the Attica region of Greece were recruited randomly, and were stratified by subgroups of sex, age and region to reflect the general urban population in Athens, Greece. Validated NITs for hepatic steatosis (Hepatic Steatosis Index (HIS), Fatty Liver Index (FLI), Lipid Accumulation Product (LAP), NAFLD liver fat score (NAFLD-LFS)) and steatohepatitis (Index of non-alcoholic steatohepatitis (ION), aminotransferase-creatinine-clearance non-alcoholic steatohepatitis (acNASH)) were calculated. Incidence of diabetes, CVD and mortality were recorded 5, 10 and 20 years later.

RESULTS

Within a 20-year observation period, the diabetes and CVD incidence was 26.3% and 36.1%, respectively. All hepatic steatosis and steatohepatitis NITs were independently associated with diabetes incidence. ION and acNASH presented independent association with CVD incidence [(Hazard Ratio (HR)per 1 standard deviation (SD) = 1.33, 95% Confidence Interval (95% CI) (1.07, 1.99)) and (HRper 1 SD = 1.77, 95% CI (1.05, 2.59)), respectively]. NAFLD-LFS which is a steatosis NIT indicating features of steatohepatitis, was linked with increased CVD mortality (HRper 1 SD = 1.35, 95% CI (1.00, 2.30)) and all-cause mortality (HRper 1 SD = 1.43, 95% CI (1.08, 2.01)). Overall, steatohepatitis NITs (i.e., ION and acNASH) presented stronger associations with the outcomes of interest compared with steatosis NITs. Clinically important trends were observed in relation to diabetes and CVD incidence progressively over time, i.e. 5, 10 and 20 years after baseline.

CONCLUSIONS

Easily applicable and low-cost NITs representing steatohepatitis may be early predictors of diabetes and CVD onset. More importantly, these NITs increased the attributable risk conveyed by conventional CVD risk factors by 10%. Thus, their potential inclusion in clinical practice and guidelines should be studied further.