Five-Protein Plasma Panel Predicts MASLD Up to 16 Years Before Diagnosis
Researchers have identified a blood-based protein signature that flags risk for fatty liver disease up to 16 years before clinical detection, validated across more than 50,000 people in the UK and China.
What Was Studied and Why
Metabolic dysfunction-associated steatotic liver disease (MASLD)—the updated name for what was commonly called non-alcoholic fatty liver disease—is typically caught late. By the time imaging reveals hepatic steatosis or blood tests show elevated transaminases, the metabolic disruption driving it has usually been running for years. Existing screening tools (ultrasound, AST, ALT, GGT) are reactive by design: they detect damage already present, not the trajectory toward it.
Two papers published simultaneously in Nature Aging on 31 July 2026 address this gap using plasma proteomics—measuring hundreds of circulating proteins in blood to find patterns that precede disease. Both papers converge on the same core finding: a set of plasma proteins can stratify future MASLD risk long before conventional diagnostics would register anything abnormal.
The Five-Protein Panel
The first paper identifies five specific proteins as the predictive signature: FUOM, ACY1, KRT18, CDHR2, and GGT1. These were discovered and validated across multiple large population cohorts totalling over 50,000 participants drawn from UK and Chinese populations.
The second paper describes the same broad framework—an ultra-early proteomic signature predictive up to 16 years out—but does not independently enumerate the five proteins or provide separate effect sizes. Both papers were published on the same date in the same journal and appear to represent complementary reporting of the same research programme rather than two independent studies. Where the second paper differs, it is thinner on specifics, so the detail below draws primarily from the first.
How Well Does It Predict?
The five-protein panel alone predicted MASLD risk with reasonable accuracy at five years and moderate accuracy extending to 16 years. The model's area under the curve (AUC) varied across cohorts from 0.71 to 0.91—a meaningful spread that matters. An AUC of 0.71 is useful but imperfect; 0.91 is strong. This variation suggests the panel's performance is not uniform and likely depends on factors such as ethnicity, age range, and baseline metabolic phenotype.
When the protein panel was combined with routine clinical measurements—standard metabolic panels, liver enzymes, and similar data—predictive power improved substantially. The cross-population validation is an important signal that the proteins reflect genuine biology rather than cohort-specific noise, but the accuracy range also means clinicians and users should not expect uniform performance across all demographic groups.
What the Proteins Likely Reflect
The study identifies correlations, not causal mechanisms. The five proteins—spanning functions related to liver cell integrity (KRT18), bile acid and amino acid metabolism (ACY1, FUOM), cellular adhesion (CDHR2), and gamma-glutamyl transfer (GGT1)—collectively appear to capture metabolic dysfunction patterns that develop years before steatosis becomes imaging-visible.
It is not known from this research whether these proteins are directly involved in driving MASLD, whether they are downstream markers of a common upstream cause (insulin resistance, dyslipidaemia, chronic low-grade inflammation), or whether intervening on the proteins themselves would change disease course. The panel is a mirror of a pathological process already in motion—it does not identify the engine.
What This Means for Metabolic Health Tracking
For MyKine users already monitoring liver enzymes, lipid panels, fasting glucose, and insulin as part of a metabolic health protocol, this research reinforces why those markers matter: they reflect the same underlying dysfunction that the proteomic panel appears to capture even earlier.
If this panel reaches clinical practice, it would offer something not currently available—a blood test that identifies excess MASLD risk in your 40s or 50s, while the window for lifestyle and metabolic intervention is widest. The 16-year lead time is the most clinically meaningful aspect: it suggests there is a long phase of detectable metabolic derailment before hepatic damage becomes structural.
For users on protocols that affect hepatic lipid metabolism—including TRT, which can influence lipid profiles and insulin sensitivity, or GLP-1 receptor agonists, which reduce hepatic fat—a predictive biomarker of this kind could eventually help individualise monitoring intensity and justify earlier or more aggressive intervention before fibrosis develops. That remains speculative until prospective intervention trials exist; right now, the proteins predict, but there is no evidence yet that acting on the prediction changes outcomes.
Honest Evidence Gaps
Several limitations deserve direct acknowledgement rather than footnoting:
- No intervention data. This is a predictive study. There is no evidence that detecting this signature early and acting on it reduces MASLD incidence or severity compared with standard metabolic monitoring.
- AUC variability is real. A range of 0.71–0.91 across cohorts means the test performs meaningfully better in some populations than others. The reasons are not yet characterised.
- Mechanistic picture is absent. The study does not explain why these proteins precede MASLD or whether they are causal, consequential, or coincidental to the underlying metabolic process.
- Not a clinical tool yet. Plasma proteomics at this resolution is a research instrument. Clinical translation requires standardisation, cost-effectiveness analysis, and integration into care pathways that do not currently exist.
- The second paper is sparse on specifics. It supports the same broad conclusion but does not add independently verifiable effect sizes, cohort details, or mechanistic insight. Users should weight the detail-rich first paper more heavily.
Practical Takeaway
- This is real but early-stage science. A five-protein plasma panel predicts MASLD up to 16 years out across large, multi-population cohorts—that is a meaningful finding—but it is not a test you can order today, and no evidence yet shows early detection changes outcomes.
- Your current metabolic tracking is already capturing the same underlying signal. Fasting glucose, insulin, triglycerides, GGT, ALT, and body composition reflect the metabolic dysfunction the proteomic panel detects earlier. Keeping those in range remains the actionable step.
- Accuracy varies across groups. The AUC range (0.71–0.91) means population, age, and metabolic phenotype will affect how informative the panel is for any individual.
- Users on TRT or GLP-1 protocols should note the hepatic relevance. Both interventions intersect with the metabolic pathways involved in MASLD; a future predictive panel could sharpen monitoring decisions, but that application requires prospective evidence first.
- Watch for clinical validation studies. The meaningful next step is a prospective trial asking whether acting on an early proteomic signal reduces MASLD progression—that evidence does not exist yet.
Sources
- Plasma proteomics framework predicts metabolic dysfunction-associated steatotic liver disease up to 16 years before onset — Nature Aging (https://www.nature.com/articles/s43587-026-01183-2)
- Ultra-early proteomic signatures predict MASLD up to 16 years ahead — Nature Aging (https://www.nature.com/articles/s43587-026-01182-3)