Reading FIB-4 Correctly: A Triage Score, Not a Fibrosis Stage

The short answer#

FIB-4 is a sorting tool, not a verdict. It takes four numbers a primary care office already has (age, AST, ALT, and platelet count) and turns them into a probability that liver scarring has reached an advanced stage. In the pathway laid out by the American Association for the Study of Liver Diseases (AASLD), FIB-4 is the front door: a low result sends most people home reassured, while anything above the low threshold moves them to a sharper test such as elastography or the Enhanced Liver Fibrosis (ELF) panel. The score is excellent at ruling advanced fibrosis out and mediocre at ruling it in. That imbalance is not a defect. It is the entire reason the score comes first and is never used alone.

Key points#

• FIB-4 estimates the odds of advanced fibrosis; it does not measure scar tissue or assign a stage.
• Below 1.3 is a strong all-clear (negative predictive value near 90 percent). At or above 2.67 raises suspicion but confirms nothing.
• After age 65 the lower cutoff rises to 2.0, because age is baked into the formula and inflates the score in older adults.
• Every score above the low band gets a second, more specific test before any conclusion.
• Common conditions (low platelets, an acute liver insult, diabetes) can push the number the wrong way.

Why start with a cheap, imperfect score#

The population at risk is huge and mostly well. Most people with metabolic dysfunction-associated steatotic liver disease (MASLD) will never develop the kind of fibrosis that threatens the liver, yet the pool of people with fatty liver runs into the tens of millions. Sending all of them for imaging or biopsy is neither practical nor humane. FIB-4 answers the volume problem by reusing labs that are already on the chart, so it adds almost no cost. The AASLD guidance places it first for exactly that reason, and modeling summarized by the AASLD Liver Fellow Network suggests that leading with FIB-4 could avoid roughly 87 percent of confirmatory elastography studies.

Think of it less like a diagnostic test and more like a metal detector at an airport. It is tuned to wave the harmless majority through quickly, and to flag a smaller group for a closer look. A beep does not mean a weapon; it means someone should check. FIB-4 works the same way. A low score is a confident wave-through. A high score is a beep.

That trade is deliberate and asymmetric. A FIB-4 under 1.3 carries a negative predictive value close to 90 percent for advanced fibrosis, so a low score reliably clears people. A score at or above 2.67 carries only about an 80 percent positive predictive value, so a high score should raise the eyebrow, not close the case. The cheap test safely dismisses the low-risk crowd; the expensive tests are saved for whoever remains.

Where the age cutoff comes from#

Age sits inside the FIB-4 formula, which creates a quirk worth understanding before you trust any single number.

Because the score rises with age on its own, older adults drift upward even with a healthy liver. Past 65, specificity at the standard 1.3 cutoff falls below 30 percent, meaning the score flags far too many people who do not need referral. To correct for this, the guidance raises the lower threshold to 2.0 in adults over 65, so the indeterminate band shifts to 2.0 through 2.67. The mirror-image problem appears at the young end: under age 35 the score performs poorly, so a low result there should not be taken as the final word.

The pathway, one gate at a time#

The AASLD approach is a funnel, not a single reading. Each gate is more specific than the one before it.

First gate: FIB-4#

The standard bands are below 1.3 (low risk), 1.3 to 2.67 (indeterminate), and above 2.67 (high risk), with the older-adult adjustment noted above. Anyone in the low band is generally reassured and rechecked over time rather than referred.

Second gate: elastography or ELF#

Everyone not cleared at the first gate moves to a more specific test. Vibration-controlled transient elastography (VCTE, the FibroScan platform) measures liver stiffness: under 8 kPa argues against advanced fibrosis, 8 to 12 kPa is intermediate, and above 12 kPa points toward it. The ELF blood panel is a laboratory alternative with parallel cutoffs (below 7.7 low, 7.7 to 9.8 intermediate, above 9.8 high). Because both are more specific than FIB-4, they turn a soft positive into a firmer estimate and spare many people an unnecessary specialist visit.

Third gate: settle the disagreements#

When the second gate lands in the intermediate zone, or when two tests disagree, the pathway escalates to magnetic resonance elastography or specialist evaluation, with liver biopsy held in reserve for genuinely ambiguous cases. To confirm cirrhosis specifically, the guidance describes a sequential rule (FIB-4 above 3.48 followed by VCTE stiffness of 20 kPa or more) that reaches roughly 90 percent specificity. A 2025 Hepatology Communications analysis of a noninvasive pathway for suspected MetALD echoed this layered logic, reporting that a stepwise FIB-4 then elastography strategy kept the false negative rate low while sending only a minority of patients to specialty referral.

The ways FIB-4 misleads#

A first-line score earns its keep by being wrong in a predictable direction. Knowing that direction is the skill.

It reports odds, not anatomy. FIB-4 never touches scar tissue. It estimates the probability that scarring is advanced, and that probability moves whenever its inputs move.

Its inputs answer to other masters. Platelet count can drop for reasons that have nothing to do with the liver, which inflates the score. AST and ALT climb during any acute liver-cell injury, which can shove a stable liver into a falsely high band. Age lifts the number year by year on its own.

Some groups blur the signal. The guidance flags that noninvasive tests can perform less reliably in diabetes, one of the conditions that defines much of the at-risk population, and notes that the cutoffs themselves may need revision in type 2 diabetes.

Moving a cutoff relocates the error; it does not erase it. Raising the older-adult threshold to 2.0 trims false positives but accepts more false negatives, so a few true cases slip under the line. No single cutoff is generous to everyone.

None of this is a failure of the score. It is the reason FIB-4 is a filter rather than a finding. The number tells you who deserves a closer look, and the later gates do the staging.