About the Fibrosis-4 Index (FIB-4)
FIB-4 estimates the probability of advanced liver fibrosis from four routine values — age, AST, ALT and platelet count — with no imaging or biopsy required. Below 1.45 advanced fibrosis is unlikely (90% negative predictive value in the derivation cohort); above 3.25 it is likely (97% specificity, 65% positive predictive value); between the two the result is indeterminate and a second test is needed. In MASLD many services use lower cut-offs of 1.3 and 2.67, and every published cut-off loses specificity above age 65 because age sits in the numerator.
Formula
FIB-4 = (age in years × AST) ÷ (platelets × √ALT)- age
- Years. In the numerator, so the score rises with age alone.
- AST
- U/L, in the numerator.
- ALT
- U/L, entering as a square root in the denominator.
- platelets
- ×10⁹/L, in the denominator. Numerically identical to ×1,000/µL, so no conversion is needed between those two units.
- There are no floors, caps or rounding rules — unlike MELD, every value enters as measured.
- The units are the conventional ones and need no conversion: AST and ALT in U/L, platelets in ×10⁹/L (the same number as ×1,000/µL).
- Because ALT is under a square root, halving the ALT raises the score by about 41%, not 100%.
Interpreting the result
Treat FIB-4 as a triage instrument with three exits, not as a measurement. Below the lower cut-off, advanced fibrosis is unlikely and the patient can be rechecked in one to three years rather than investigated — that is the result the score exists to produce. Above the upper cut-off, refer, because specificity is high. In between, the score has told you nothing and the pathway must continue to elastography or a proprietary panel; roughly a third of patients land here, and treating an indeterminate result as normal is the commonest way the score is misused. Always ask whether age or a non-hepatic cause of thrombocytopenia is doing the work before acting on a raised value.
| Score | Band | What it means | Action |
|---|---|---|---|
| < 1.45 | Low risk | Advanced fibrosis unlikely — 90% negative predictive value, 70% sensitivity in the derivation validation set | Reassess in 1–3 years; manage metabolic risk factors |
| 1.45–3.25 | Indeterminate | The score does not discriminate here. Around a third of patients fall in this band | Proceed to a second-line test — transient elastography, ELF or equivalent. Do not treat as normal |
| > 3.25 | High risk | Advanced fibrosis likely — 97% specificity, 65% positive predictive value | Refer to hepatology; begin assessment for cirrhosis complications |
Scroll the table sideways for every column.
What the FIB-4 Index needs (4 inputs)
- Age (years)
- Age is in the numerator, so FIB-4 rises with age independently of any liver disease. This is the score's most important quirk and the reason it produces so many false positives in older patients.
- AST (U/L)
- Aspartate aminotransferase, also in the numerator. Any cause of transaminitis — recent alcohol, muscle injury, statins, acute hepatitis — inflates the score without indicating fibrosis.
- ALT (U/L)
- Alanine aminotransferase, in the denominator as a square root. Its position means a rising AST:ALT ratio drives the score up, which is the pattern seen as fibrosis progresses.
- Platelet count (×10⁹/L)
- In the denominator, acting as a marker of portal hypertension. Any other cause of thrombocytopenia — immune thrombocytopenia, marrow disease, hypersplenism from another cause, some drugs — will raise the score spuriously.
Units. No conversion is needed for this score. AST and ALT are in U/L everywhere, and platelets reported as ×10⁹/L and as ×1,000/µL are the same number — 150 ×10⁹/L is 150 ×1,000/µL. Do not multiply or divide between those two; doing so is a well-recognised source of hundred-fold errors.
What it returns
- FIB-4 index
- A continuous value, usually reported to two decimal places. It is a probability estimate, not a fibrosis stage — there is no FIB-4 value that equals F3.
- Risk band
- Low, indeterminate or high, against the cut-offs in use. The indeterminate band is not a reassuring result; it is an instruction to do a second test.
How it is calculated
FIB-4 came out of a multivariate logistic regression on the APRICOT trial cohort, in which platelet count, age, AST and INR were the variables independently associated with fibrosis stage. The published index keeps age, AST and platelets and substitutes ALT for INR, arranged so that the two features of progressive fibrosis — a falling platelet count from portal hypertension, and a rising AST relative to ALT — both push the score up. It is not derived from a mechanistic model of the liver; it is a fitted combination of markers that happen to move together as fibrosis advances, which is why it behaves badly whenever one of them moves for another reason.
Facts & figures
| Population | Lower cut-off | Upper cut-off | Source of the thresholds |
|---|---|---|---|
| HIV/HCV co-infection (original) | 1.45 | 3.25 | The APRICOT derivation cohort |
| MASLD / NAFLD | 1.3 | 2.67 | Widely adopted in metabolic liver disease pathways |
| Age over 65 | 2.0 | 2.67 | A raised lower threshold is commonly recommended because the standard one has poor specificity in this group |
Scroll the table sideways for every column.
Which cut-offs you use changes the answer for a substantial number of patients. State the pair you applied when recording a FIB-4.
| Metric | Value |
|---|---|
| AUROC for Ishak stage 0–3 vs 4–6 | 0.765 |
| NPV at < 1.45 (validation set) | 90%, with 70% sensitivity |
| PPV at > 3.25 | 65%, with 97% specificity |
| Patients who could have avoided biopsy | 71% of the validation group |
| Correctly classified outside the indeterminate range | 87% of the 198 patients with FIB-4 outside 1.45–3.25 |
Scroll the table sideways for every column.
Evidence
Derivation — APRICOT trial cohort
2006 · n = 832832 of the 868 patients enrolled in the APRICOT trial of HIV/HCV co-infection, after 36 were excluded for uninterpretable histology. Patients were grouped by Ishak stage (0–1, 2–3, 4–6) and randomly split into a training set of 555 and a validation set of 277. Multivariate regression identified platelet count, age, AST and INR as independently associated with fibrosis.
AUROC 0.765 for distinguishing Ishak stage 0–3 from 4–6. At the < 1.45 cut-off, negative predictive value 90% with 70% sensitivity; at > 3.25, positive predictive value 65% with 97% specificity. Liver biopsy could have been avoided in 71% of the validation group.
Adoption in metabolic liver disease
2023FIB-4 has since been validated across MASLD/NAFLD, chronic hepatitis B and C, and alcohol-related liver disease, and is embedded as the first step of most primary-care and diabetes case-finding pathways. Lower cut-offs of 1.3 and 2.67 are widely used in MASLD.
Consistently good at excluding advanced fibrosis and poor at confirming it, which is why guideline pathways use it as a rule-out and route everything above the lower cut-off to a second test.
How it compares
FIB-4 Index vs APRI
FIB-4 is generally the better first-line rule-out and the one guideline pathways specify — APRI is simpler but uses your laboratory's AST upper limit of normal, which makes it less portable between services.
Both are non-invasive indices built from routine bloods, and both are good at excluding advanced fibrosis and mediocre at confirming it. FIB-4 adds age and ALT and was derived against Ishak staging; APRI is AST relative to its own upper limit of normal, divided by platelets, which means the result depends on a reference range that differs between laboratories and between sexes. In practice they are often calculated together, and agreement between them is more reassuring than either alone.
FIB-4 Index vs NAFLD Fibrosis Score
Use FIB-4 first because it needs nothing beyond a standard blood panel; NFS adds BMI, diabetes status and albumin, which makes it more specific to metabolic liver disease but harder to run automatically over existing lab data.
The NAFLD Fibrosis Score was built specifically in NAFLD populations and includes clinical variables FIB-4 ignores, so it can perform better in that setting. FIB-4's advantage is operational: it can be calculated retrospectively from any historic full blood count and liver panel, which is what makes population case-finding possible at all. Most MASLD pathways therefore lead with FIB-4 and reserve NFS or elastography for the indeterminate group.
FIB-4 Index vs Transient elastography (FibroScan)
Elastography is the more accurate test and the usual next step after an indeterminate or high FIB-4 — but it needs a machine, an operator and a fasted patient, which is exactly why FIB-4 comes first.
Liver stiffness measurement discriminates advanced fibrosis considerably better than any blood index and can give a numeric stiffness in kPa. It is also confounded by obesity, ascites, acute hepatitis and food intake, and it is not available in primary care at scale. The standard pathway is sequential rather than competitive: FIB-4 to exclude, elastography to confirm.
Pearls & pitfalls
- Age is in the numerator, so FIB-4 rises with age by itself. A healthy 75-year-old can exceed the standard lower cut-off with entirely normal liver tests, which is why a raised threshold of about 2.0 is recommended over 65.
- The indeterminate band is not a negative result. Around a third of patients land in it, and stopping there is the single commonest misuse of the score.
- Any non-hepatic cause of thrombocytopenia inflates the score — immune thrombocytopenia, marrow infiltration, drugs. Check the platelet count has a hepatic explanation before referring.
- Do not use it during acute hepatitis or a recent alcohol binge. Transaminases are then unrepresentative of chronic fibrosis and the score is meaningless.
- ALT sits under a square root, which blunts its influence. A very low ALT with a high AST produces a strikingly high score, and that AST:ALT pattern is doing most of the work.
- Platelets in ×10⁹/L and ×1,000/µL are the same number. There is no conversion to do, and 'converting' between them is a real source of hundred-fold errors.
- It is a probability, not a stage. There is no FIB-4 value that corresponds to F3, and reporting one as though there were overstates what the index can do.
Critical actions
- Act on the indeterminate band — arrange elastography or an equivalent second-line test rather than filing the result as normal.
- Repeat rather than treating it as a one-off. FIB-4 is a monitoring tool in metabolic liver disease; a single low value at one point in time is weak reassurance over a decade.
- State which cut-off pair you used, because the MASLD and original thresholds classify a substantial number of patients differently.
- Look for a non-hepatic cause of a low platelet count before referring on a high score.
- Above the upper cut-off, start assessing for cirrhosis complications — varices and hepatocellular carcinoma surveillance — rather than only referring.
- Address the metabolic drivers at every level of the score; a low FIB-4 in a patient with obesity and type 2 diabetes is not a reason to do nothing.
Why this score exists
FIB-4 was developed in HIV/HCV co-infected patients, for a specific and narrow purpose: reducing the number of liver biopsies in a population where biopsy was routine and carried real risk. The authors were explicit that the index was for distinguishing advanced from non-advanced fibrosis, not for staging, and the derivation grouped patients into broad Ishak bands rather than individual stages for exactly that reason. Its subsequent career as a population screening tool in metabolic liver disease is a considerable extrapolation from that origin — a different disease, a different age distribution, and a screening rather than a triage-to-biopsy context. The cut-offs having to be re-tuned for MASLD and for older patients is the visible consequence of that.
About the creator
First author, 2006 derivation study
Derived FIB-4 in patients with HIV/HCV coinfection; it has since become the most widely used first-line fibrosis index across liver disease of any cause.
Limitations
- Derived in HIV/HCV co-infection, then applied at scale in metabolic liver disease — a different population, different age distribution and different purpose from the one it was validated for.
- Age in the numerator makes it unreliable above 65, where the standard lower cut-off generates many false positives.
- It cannot stage fibrosis, only estimate the probability that fibrosis is advanced.
- Roughly a third of patients fall in the indeterminate band, where the index provides no information.
- Any non-hepatic cause of thrombocytopenia or transaminitis distorts it, and the score cannot tell you that has happened.
- Performance is poor in acute hepatitis and unreliable soon after heavy alcohol intake.
- Multiple competing cut-off pairs are in use, so two services can classify the same patient differently without either being wrong.
If you are the patient
FIB-4 is a number worked out from four things already measured in a normal blood test — your age, two liver enzymes (AST and ALT), and your platelet count. It estimates how likely it is that your liver has significant scarring, without needing a scan or a biopsy. A low result means significant scarring is unlikely, and your team will usually just recheck it in a year or two. A high result means scarring is likely and you will be referred to a liver specialist. A middle result is common and means the test could not tell either way, so you need a second test such as a FibroScan — this is not bad news, it just means more information is needed. One thing worth knowing: because your age is part of the calculation, the number naturally rises as you get older even if your liver is fine, so an older person's slightly raised result is often not a liver problem at all. Your team takes that into account.
Frequently asked questions
What is the FIB-4 index?#
FIB-4 is a non-invasive score that estimates the probability of advanced liver fibrosis from age, AST, ALT and platelet count. It is calculated as (age × AST) ÷ (platelets × √ALT) and is used as the first-line test to rule out advanced fibrosis in chronic liver disease.
What is a normal FIB-4 score?#
A FIB-4 below 1.45 makes advanced fibrosis unlikely, with a 90% negative predictive value in the derivation cohort. In metabolic liver disease many services use 1.3 instead, and over the age of 65 a threshold nearer 2.0 is often recommended because age alone raises the score. There is no single 'normal' value independent of age and the cut-offs in use.
What does a FIB-4 of 2.0 mean?#
It falls in the indeterminate band between 1.45 and 3.25, where the index does not discriminate. It is not a reassuring result and it is not a diagnosis — it means a second-line test such as transient elastography is needed. About a third of patients land in this range.
What FIB-4 score indicates cirrhosis?#
None, strictly. FIB-4 estimates the probability of advanced fibrosis (Ishak 4–6, which includes cirrhosis) rather than diagnosing it. Above 3.25 advanced fibrosis is likely, with 97% specificity, but the positive predictive value was only 65% in the derivation cohort — so a high score warrants referral and confirmation, not a diagnosis.
How accurate is FIB-4?#
In the APRICOT derivation cohort the AUROC was 0.765 for distinguishing Ishak stage 0–3 from 4–6. It is much better at excluding advanced fibrosis than confirming it: 90% negative predictive value below 1.45, but only 65% positive predictive value above 3.25. That asymmetry is why it is used as a rule-out test.
Why is FIB-4 unreliable in older patients?#
Because age sits in the numerator of the formula, so the score rises with age whether or not the liver is diseased. A healthy person over 65 can exceed the standard 1.45 cut-off on age alone, which is why a higher lower threshold of around 2.0 is commonly recommended in that group.
Which FIB-4 cut-offs should be used in MASLD?#
Most metabolic liver disease pathways use 1.3 and 2.67 rather than the original 1.45 and 3.25. The lower rule-out threshold increases sensitivity in a population where the prior probability of advanced fibrosis is lower than in the HCV cohort the score was derived in.
Can FIB-4 replace a liver biopsy?#
Not in general, but it removes the need for one in many patients. In the derivation validation set, applying the cut-offs would have let 71% of patients avoid biopsy, and 87% of those falling outside the indeterminate range were correctly classified. Patients in the indeterminate band, or with discordant results, still need further testing.
References
Original / primary reference
Clinical practice guidelines
- EASL-EASD-EASO Clinical Practice Guidelines on the management of metabolic dysfunction-associated steatotic liver disease (MASLD). J Hepatol. 2024;81(3):492-542.
- Rinella ME, Neuschwander-Tetri BA, Siddiqui MS, et al. AASLD Practice Guidance on the clinical assessment and management of nonalcoholic fatty liver disease. Hepatology. 2023;77(5):1797-1835.