Glycemia and fibrosis carried the prognostic signal in a national NHANES analysis
In a large U.S. population sample, a composite glucose–steatosis–fibrosis score tracked higher all-cause and cardiovascular mortality, but the extra steatosis component did not add independent predictive value once glycemia, fibrosis, and other covariates were taken into account. The pattern is clinically important because it separates metabolic liver fat from the parts of the phenotype that appear to drive risk more directly: impaired glucose regulation and advancing fibrosis.
Investigators analyzed 21,511 adults from NHANES 1999–2018 and linked outcomes through 2019. During follow-up, 3,317 participants died from any cause and 1,056 died from cardiovascular disease. Each 1-point increase in the GSF score was associated with higher all-cause mortality, with a hazard ratio of 1.200, and higher CVD mortality, with a hazard ratio of 1.273. Those are not subtle associations. At the population level, the score behaved as a marker of accumulating cardiometabolic injury, with mortality risk rising in step with worsening metabolic liver and glucose burden.
Why steatosis did not sharpen risk prediction
The analysis became more interesting once the components were separated. Steatosis itself was not independently associated with either mortality outcome after conditioning on glycemia, fibrosis, and covariates. In other words, once the burden of dysglycemia and fibrotic liver injury was accounted for, liver fat did not seem to contribute additional prognostic information. A one-sentence bedside takeaway follows from that result: fatty liver alone is not the whole story when mortality risk is being estimated in a cardiometabolic framework.
That makes biological sense. Hepatic steatosis often travels with insulin resistance, dyslipidemia, central adiposity, and systemic inflammation, but it is the transition toward fibrosis and the persistence of abnormal glucose handling that more closely reflect end-organ injury and long-term vascular stress. Fibrosis captures cumulative tissue damage and remodeling, while dysglycemia signals the metabolic milieu that accelerates atherosclerosis, renal decline, and adverse cardiac remodeling. A score that includes steatosis may describe disease burden, yet the incremental bedside value depends on whether that component changes risk classification beyond the stronger drivers already in play.
A simpler score performed as well as the steatosis-inclusive model
The study also tested a related glycemia–fibrosis score in a CKM common-complete internal benchmark. There, the simpler score performed at least as well as GSF for discrimination and Brier score. That finding matters because it challenges the assumption that adding more variables automatically improves prediction. In this setting, the added steatosis term did not improve the ability to separate higher-risk from lower-risk participants or meaningfully tighten calibration.
For clinicians, the implication is practical. If the goal is mortality stratification within the cardiovascular-kidney-metabolic framework, a model that centers on glycemia and fibrosis may be sufficient for many uses, while a steatosis-inclusive score may describe phenotype more completely without improving predictive utility. The distinction is especially relevant in primary care and preventive medicine, where every added variable must earn its place by improving decision-making, not just mathematical completeness.
High scores tracked shorter survival over 10 years
Participants classified with high versus low GSF had higher 10-year all-cause and cardiovascular risk and shorter restricted mean survival time. RMST adds a time-based perspective that hazard ratios alone cannot fully capture. It asks a simpler question: how much life time was lost over the observed horizon? In this cohort, the answer pointed in the same direction as the hazard models, reinforcing the idea that a higher metabolic-fibrotic burden translates into measurable survival loss across a decade.
These findings fit the broader CKM concept, in which metabolic dysfunction, kidney disease, liver injury, and cardiovascular disease reinforce one another rather than appearing as isolated problems. The current analysis places glycemia and fibrosis near the center of that loop. Liver steatosis still matters as a marker of metabolic stress, but its prognostic weight appears to be carried largely through the pathways that worsen glucose control and scar the liver over time.
What the data can and cannot support
The study was built on NHANES 1999–2018, a nationally representative survey design with survey-weighted Cox models pooled across 50 multiply imputed datasets. That approach strengthens the population relevance of the findings and helps address missing covariates, but it remains an observational secondary analysis of de-identified data. It can identify associations, compare prediction rules, and rank markers against one another, yet it cannot establish whether changing the score would change mortality.
The article also does not support a claim that steatosis is clinically irrelevant. Rather, it shows that in this specific mortality framework, steatosis did not add independent information beyond glycemia and fibrosis. That is a narrower and more useful conclusion. In risk models that already contain strong metabolic and fibrotic signals, liver fat may be redundant for prediction even if it remains biologically important as part of the disease process.
The final message is disciplined and unsentimental: in a national U.S. cohort, glycemia and fibrosis carried the prognostic weight, while steatosis did not improve mortality stratification once those factors were included. The next step is not more elaboration of the same score, but external validation in other cohorts and testing whether simpler CKM-oriented tools perform similarly in routine care.